Harmful Algal Blooms (Habs) and Desalination: A Guide To Impacts, Monitoring, and Management

Manuals and Guides 78
Harmful Algal Blooms (HABs)

and Desalination: A Guide to
Impacts, Monitoring, and Management
Edited by:
Donald M. Anderson, Siobhan F.E. Boerlage, Mike B. Dixon
UNESCO
Manuals and Guides 78
Intergovernmental Oceanographic Commission
Harmful Algal Blooms (HABs) and Desalination: A

Guide to Impacts, Monitoring and Management
Edited by:
Donald M. Anderson*
Biology Department, Woods Hole Oceanographic Institution
Woods Hole, MA 02543 USA
Siobhan F. E. Boerlage
Boerlage Consulting
Gold Coast, Queensland, Australia
Mike B. Dixon
MDD Consulting, Kensington
Calgary, Alberta, Canada
*Corresponding Author’s email: danderson@whoi.edu
UNESCO 2017
IOC Manuals and Guides, 78
Paris, October 2017
English only
The preparation of this manual was supported by the U.S. Agency for International
Development (USAID) through a contract to the Middle East Desalination Research Center
(MEDRC). Additional support for publishing came from the Intergovernmental
Oceanographic Commission (IOC) of UNESCO.
The Manual was prepared with support from multiple programs and agencies, including the
United States Agency for International Development, the Middle East Desalination Research
Center, and the Intergovernmental Oceanographic Commission. Neither these sponsors nor
the editors and contributing authors guarantee the accuracy or completeness of any
information published herein, and neither these sponsors nor the editors and contributing
authors shall be responsible for any errors, omissions, or damages arising out of the use of
this information. This work is published with the understanding that the sponsors, editors,
and contributing authors are supplying information but are not attempting to render
engineering or other professional services. If such services are required, the assistance of
appropriate professionals should be sought. Furthermore, the sponsors, editors, and authors
do not endorse any products or commercial services mentioned in the text.
The designations employed and the presentation of the material in this publication do not
imply the expression of any opinion whatsoever on the part of the Secretariats of UNESCO
and IOC concerning the legal status of any country or territory, or its authorities, or
concerning the delimitation of the frontiers of any country or territory.
For bibliographic purposes this document should be cited as follows:
Anderson D. M., S. F. E. Boerlage, M. B. Dixon (Eds), Harmful Algal Blooms (HABs) and
Desalination: A Guide to Impacts, Monitoring and Management. Paris, Intergovernmental
Oceanographic Commission of UNESCO, 2017. 539 pp. (IOC Manuals and Guides No.78.)
(English.) (IOC/2017/MG/78).
Cover photo: Chlorophyll concentrations captured by the MODIS Aqua sensor (NASA) from
February 6, 2016, showing typical bloom patterns for the Arabian Gulf, Sea of Oman region,
with low chlorophyll (dark blue) away from the coasts and high chlorophyll (orange and red)
in complex patches and filaments, particularly around the Arabian Peninsula. The patterns are
caused by surface transport and concentration of a harmful Cochlodinium bloom that is
widely dispersed in response to surface currents and eddies. (Courtesy of R. Kudela,
University of California, Santa Cruz and the National Aeronautics and Space Administration
(NASA). Back cover photo: Foaming observed in the rapid mix basins during pretreatment
at Tampa Bay Seawater Desalination Plant during an algal bloom event (Courtesy of Nikolay
Voutchkov, Water Globe Consultants, LLC 2009).
First published in 2017 by the United Nations Educational,

Scientific, and Cultural Organization
7, Place de Fontenoy, 75732 Paris 07 SP
© UNESCO 2017
Printed in Denmark
FOREWORD
Coastal development is progressing at a rapid pace and coastal populations are increasingly
vulnerable to sea-level rise, tsunamis, coastal erosion, storms, and other adverse
environmental phenomena. One of the objectives of the Intergovernmental Oceanographic
Commission of UNESCO (IOC) is to facilitate safety and security of people in the coastal
zone and at sea. For example, the IOC operates a quasi-global tsunami warning system and
strives to enable efficient coastal zone management that is fully informed of all major coastal
risks.
In a number of coastal countries with dry climates, the demands for food and water are to a
significant extent supported by seawater desalination activities. One marine hazard that is a
threat to the large and rapidly expanding desalination industry is harmful algal blooms
(HABs). The harm comes in part from algal production of neurotoxins as well as bad taste
and odor and skin-irritating compounds that may persist in the treated water. Another major
concern is the organic material produced by some algal blooms, as these compounds can clog
intake filters and foul membrane surfaces, greatly compromising plant operations. Expansion
of harmful algal events is inevitable given global trends in population, agriculture,
development and climate. With the already observed increase in the number of toxic and
harmful blooms, the resulting economic losses, the types of resources affected, and the
number of toxins and toxic species reported, HAB problems can only be expected to increase.
For more than 20 years IOC has offered leadership in capacity building and international
research cooperation in relation to harmful algae. The overall goal of the IOC Harmful Algal
Bloom Programme is to foster the effective management of, and scientific research on, HABs
in order to understand their causes, predict their occurrences, and mitigate their effects. The
programme activities include provision of information and expertise, training, and research to
improve understanding of harmful algae ecology. The design of efficient and effective HAB
monitoring programmes can minimize the impacts of HABs on drinking water and seafood
quality, thereby protecting public health and resources.
Despite the overall progress of HAB research, desalination plant operators and those who
design plants or advise plant managers have thus far had very little information and guidance
on how to manage and mitigate the effects of harmful algae on desalination operations. To
meet their needs, the Intergovernmental Panel on Harmful Algal Blooms (IPHAB)
established an international Task Team to address the effects of harmful algae on
desalination. The group was heavily involved in the organization of two major conferences
on the subject and was successful in reaching out to the expert community. The time has now
come to combine experience and publish this book: Harmful Algal Blooms (HABs) and
Desalination: A Guide to Impacts, Monitoring and Management.
The IOC is pleased to be a co-sponsor of this important effort and hopes this Guide can help
address the practical issues that harmful algae pose to desalination. At the same time the need
for continuing targeted research in this topic area must also be stressed. Improved HAB
forecasts for desalination plants will rely on improved coastal models, in combination with in
situ observations that can detect and quantify HAB cells and toxins, or satellite remote
sensing data to characterize the spatial extent and density of the blooms. A collaborative
program can be envisioned involving multiple desalination plants with the aim of
transitioning pilot HAB forecasting systems (like those described in this Guide) into
operational systems.
Vladimir Ryabinin
Executive Secretary, IOC
3
Acknowledgements
The Editors would like to thank the following people who assisted in the preparation of this
manual. These include: Shannon McCarthy for her efforts to initiate and obtain funding for
this project, Kevin Price for his administrative work, Karen Steidinger for reviewing the algal
species descriptions, Judy Kleindinst for major editorial and layout support, and Henrik
Enevoldsen for publication assistance.
4
PREFACE
Arid countries throughout the world are heavily reliant on seawater desalination for their
supply of drinking and municipal water. The desalination industry is large and rapidly
growing, approaching more than 20,000 plants operating or contracted in greater than 150
countries worldwide and capacity projected to grow at a rate of 12% per year for the next
several decades (http://www.desaldata.com; 2016). Desalination plants are broadly
distributed worldwide, with a large and growing capacity in what will be referred to as the
“Gulf” region throughout this manual. Here the Gulf refers to the shallow body of water
bounded in the southwest by the Arabian Peninsula and Iran to the northeast. The Gulf is
linked with the Arabian Sea by the Strait of Hormuz and the Gulf of Oman to the east and
extends to the Shatt al-Arab river delta at its western end.
One of the operational challenges facing the industry is also expanding globally – the
phenomena termed harmful algal blooms or HABs. Blooms are cell proliferations caused by
the growth and accumulation of individual algal species; they occur in virtually all bodies of
water. The algae, which can be either microscopic or macroscopic (e.g., seaweeds) are the
base of the marine food web, and produce roughly half of the oxygen we breathe. Most of the
thousands of species of algae are beneficial to humans and the environment, but there are a
small number (several hundred) that cause HABs. This number is vague because the harm
caused by HABs is diverse and affects many different sectors of society (see Chapter 1).
HABs are generally considered in two groups. One contains the species that produce potent
toxins (Chapter 2) that can cause a wide range of impacts to marine resources, including mass
mortalities of fish, shellfish, seabirds, marine mammals, and various other organisms, as well
as illness and death in humans and other consumers of fish or shellfish that have accumulated
the algal toxins during feeding. The second category is represented by species that produce
dense blooms - often termed high biomass blooms because of the large number of cells.
Cells can reach concentrations sufficient to make the water appear red (hence the common
term “red tide”), though brown, green and golden blooms are also observed, while many
blooms are not visible.
In this manual, we define toxic algae as those that produce potent toxins (poisonous
substances produced within living cells or organisms), e.g., saxitoxin. These can cause
illness or mortality in humans as well as marine life through either direct exposure to the
toxin or ingestion of bioaccumulated toxin in higher trophic levels e.g. shellfish. Non-
toxic HABs can cause damage to ecosystems and commercial facilities such as desalination
plants, sometimes because of the biomass of the accumulated algae, and in other cases due to
the release of compounds that are not toxins (e.g., reactive oxygen species, mucilage) but that
can still be lethal to marine animals or cause disruptions of other types.
Both toxic and non-toxic HABs represent potential threats to seawater desalination facilities.
Although toxins are typically removed very well by reverse osmosis and thermal desalination
processes (see Chapter 10), algal toxins represent a potential health risk if they are present in
sufficiently high concentrations in the seawater and if they break through the desalination
process. It is therefore important for operators to be aware when toxic blooms are near their
plants so they can ensure that the removal has indeed occurred (Chapter 3). High biomass
blooms pose a different type of threat, as the resulting particulate and dissolved organic
material can accelerate clogging of media filters or contribute to (bio)fouling of pretreatment
and RO membranes which may lead to a loss of production.
Impacts of HABs on desalination facilities are thus a significant and growing problem, made
worse by the lack of knowledge of this phenomena among plant operators, managers,
5
engineers, and others involved in the industry, including regulatory agencies. Recognizing
this problem, the Middle East Desalination Research Center (MEDRC) and the UNESCO
Intergovernmental Oceanographic Commission (IOC) organized a conference in 2012 in
Muscat, Oman, to bring HAB researchers and desalination professionals together to exchange
knowledge and discuss the scale of the problem and strategies for addressing it. One of the
recommendations of that meeting was that a “guidance manual” be prepared to provide
information to desalination plant operators and others in the industry about HABs, their
impacts, and the strategies that could be used to mitigate those impacts. With support from
the US Agency for International Development (USAID) and the IOC Intergovernmental
Panel for Harmful Algal Blooms (IPHAB), an editorial team was assembled and potential
authors contacted. For the first time, HAB scientists worked closely with desalination
professionals to write chapters that were scientifically rigorous yet practical in nature – all
focused on HABs and desalination. During the planning of this manual, it became clear from
an informal survey of the desalination industry that generally, HAB problems are far more
significant for seawater reverse osmosis (SWRO) plants than for those that use thermal
desalination. Both types of processes are very effective in removing HAB toxins (Chapter 10),
but the SWRO plants are far more susceptible to clogging of pretreatment granular media
filters and fouling of membranes by algal organic matter and particulate biomass.
Accordingly, the focus of this book is on SWRO, with only occasional reference to thermal
processes. Likewise, emphasis has been placed on seawater HABs, with reference to
estuarine and brackish-water HABs only when practices from those types of waters can be
informative or illustrative.
A brief synopsis of the book follows. Chapter 1 provides a broad overview of HAB
phenomena, including their impacts, the spatial and temporal nature of their blooms, common
causative species, trends in occurrence, and general aspects of bloom dynamics in coastal
waters. Chapter 2 describes the metabolites of HAB cells, including toxins, taste and odor
compounds. Methods for analyses are presented there, supplemented by detailed
methodological descriptions of rapid toxin screening methods in Appendix 2. As discussed in
Chapters 8 and 10, thermal and SWRO operations are highly effective in the removal of HAB
toxins, but plant personnel should have the capability to screen for these toxins in raw and
treated water to ensure that this removal has been effective. This would be critical, for
example, if the public or the press were aware of a toxic HAB in the vicinity of a desalination
plant intake and asked for proof that their drinking water is safe.
Currently, most desalination plants do not collect data on seawater outside their plants, so
they are generally unaware of the presence (now or anticipated) of a potentially disruptive
HAB. Chapter 3 provides practical information on the approaches to implementing an
observing system for HABs, describing sampling methods and measurement options that can
be tailored to available resources and the nature of the HAB threat in a given area. Appendix
4 provides more details on methods used to count and identify HAB cells during this process.
All are based on direct water sampling, but it is also possible to observe HABs from space –
particularly the high biomass events. Chapter 4 describes how satellite remote sensing can be
used to detect booms. The common sources of imagery (free over the Internet) are presented,
as well as descriptions of the software (also free) that can be used to analyze the satellite data.
It is relatively easy and highly informative for plant personnel to use this approach to better
understand what is in the seawater outside their plants. The cover of this guide provides a
graphic example of the incredible scale and resolution of this observational approach.
Chapter 5 discusses typical water quality parameters that are measured online or in feedwater
samples at desalination plants that could be used to detect blooms at the intake or evaluate
process efficiency in removing algal particulates and organics. Emerging parameters that also
6
show promise are examined to provide a resource for plant personnel. Chapter 6 looks at
desalination seawater intakes that are the first point of control in minimizing the ingress of
algae into the plant. A brief overview of siting considerations that may ultimately drive the
location of an intake is also provided.
One question asked frequently of HAB scientists is whether the blooms can be controlled or
suppressed in a manner analogous to the treatment of insects or other agricultural pests on
land. This has proven to be an exceedingly difficult challenge for the HAB scientific and
management community, given the dynamic nature of HABs in coastal waters, their large
spatial extent, and concerns about the environmental impacts of bloom control methods.
Chapter 7 presents a summary of the approaches to bloom prevention and control that have
been developed, and discusses whether these are feasible or realistic in the context of an
individual desalination plant.
Chapter 8 describes management strategies for HABs and risk assessment, including Hazard
Analysis Critical Control Point (HACCP) and Alert Level Framework procedures. Once a
HAB is detected, a wide range of approaches can be used to address the problems posed by
the dissolved toxins associated with those blooms. Chapter 9 presents many of these
pretreatment strategies and discusses their use in removing algal organic matter and
particulates to prevent filter clogging and membrane fouling. This is necessary to maintain
effective plant operation and avoid serious operational challenges for the reverse osmosis
step. The chapter covers common pretreatments such as chlorination/dechlorination,
coagulation, dissolved air flotation, granular media filtration, ultrafiltration, and cartridge
filtration, in addition to discussing issues experienced due to the inefficiencies of each
pretreatment on reverse osmosis.
Chapter 10 then addresses the important issue of HAB toxin removal during pretreatment and
desalination, and describes laboratory and pilot-scale studies that address that issue. Finally,
Chapter 11 provides a series of case studies describing individual HAB events at desalination
plants throughout the world, detailing the types of impacts and the strategies that were used
to combat them. These studies should be of great interest to other operators as they encounter
similar challenges.
The manual concludes with a series of appendices that provide images and short descriptions
of common HAB species (Appendix 1), rapid screening methods for HAB toxins (Appendix
2), methods to measure transparent exopolymer particles (TEP) and their precursors
(Appendix 3), methods to enumerate algal cells (Appendix 4), and reverse osmosis autopsy
and cleaning methods (Appendix 5).
Compilation of this manual was a major undertaking, requiring the cooperation of scientists
and engineers from multiple disciplines, including a number where interactions have been
rare in the past. We hope the accumulated material proves useful, and plan to keep this
document updated through time and readily available through the Internet. The Editors
welcome questions, comments, and suggestions that can make this compilation more useful
and accurate.
Donald M. Anderson
Woods Hole Oceanographic Institution
Siobhan F.E. Boerlage
Boerlage Consulting
Mike B. Dixon
MDD Consulting
7
Table of Contents
Foreword 3
Acknowledgements 4
Preface 5
Contributing Authors and Affiliations 13
Chapter 1. Harmful algal blooms Donald M. Anderson
1.1 Algal blooms 17
1.2 Harmful or toxic bloom species 20
1.3 Algal cell characteristics 22
1.4 Trends and species dispersal 32
1.5 Growth features, bloom mechanisms 34
1.6 Summary 41
1.7 References 42
Chapter 2. Algal issues in seawater desalination Philipp Hess, Loreen O. Villacorte,
Mike B. Dixon, Siobhan F.E. Boerlage, Donald M. Anderson, Maria D. Kennedy and Jan C.
Schippers
2.1 Introduction 53
2.2 Algal organic matter (AOM) and membrane fouling 53
2.3 Algal issues in thermal desalination plants 63
2.4 Marine and freshwater toxins 63
2.5 Taste and odor compounds 71
2.6 Detection techniques 71
2.7 Gaps and perspectives on analytical techniques 75
2.8 References 76
Chapter 3. Designing an observing system for early detection of harmful algal blooms
Bengt Karlson, Clarissa R. Anderson, Kathryn J. Coyne, Kevin G. Sellner, and Donald M.
Anderson
3.1 Introduction 89
3.2 Designing an observation system 90
3.3 Background information 90
3.4 Identifying existing infrastructure 93
3.5 Sampling methods 93
3.6 Identification and enumeration of HAB organisms 102
3.7 Satellite remote sensing 108
3.8 Transport and delivery of harmful algal blooms 108
3.9 Distributing warnings and information 110
3.10 Data storage and distribution 112
3.11 Facilities, equipment, and personnel 112
3.12 Summary 114
3.13 References 115
Chapter 4. Acquisition and analysis of remote sensing imagery of harmful algal blooms
Raphael M. Kudela, Richard P. Stumpf, and Peter Petrov
4.1 Introduction 119
4.2 Availability of data and software 121
9
4.3 Internet access to imagery 121
4.4 Software for processing satellite data 123
4.5 Algorithms used to detect blooms 124
4.6 Examples of algorithms 127
4.7 Summary for end-users 128
4.8 Useful links to satellite based ocean color data 130
4.9 References 130
Chapter 5. Harmful algal bloom-related water quality monitoring for desalination
design and operation Siobhan F.E. Boerlage, Loreen O.Villacorte, Lauren Weinrich, S.
Assiyeh Alizadeh Tabatabai, Maria D. Kennedy, and Jan C. Schippers
5.1 Intake feedwater characterization and water quality monitoring 133
5.2 Suitability of conventional online water quality parameters to detect
HABs 134
5.3 Overview of parameters to determine organic matter 136
5.4 Measuring biofouling potential 147
5.5 Fouling indices to measure particulate fouling potential 153
5.6 Summary 162
5.7 References 164
Chapter 6. Seawater intake considerations to mitigate harmful algal bloom impacts
Siobhan F.E. Boerlage, Thomas M. Missimer, Thomas M. Pankratz, and
Donald M. Anderson
6.2 Intake options for SWRO desalination plants 171
6.3 Surface intake and screen options 172
6.4 Subsurface intake options 181
6.4 Siting of desalination seawater intakes 197
6.5 Summary 199
6.6 References 200
Chapter 7. Bloom prevention and control Clarissa R. Anderson, Kevin G. Sellner, and
Donald M. Anderson
7.2 Bloom prevention 207
7.3 Bloom control 209
7.4 Summary 214
7.5 References 215
Chapter 8. World Health Organization and international guidelines for toxin control,
harmful algal bloom management, and response planning
Alex Soltani, Phillip Hess, Mike B. Dixon, Siobhan F.E. Boerlage, Donald M. Anderson,
Gayle Newcombe, Jenny House, Lionel Ho, Peter Baker, and Michael Burch
8.1 Guidelines and standards 223
8.2 Using guideline values 228
8.3 Australian drinking water guidelines regarding multiple treatment
barriers 228
8.4 Risk assessment for the presence of HABs 230
8.5 Alert level frameworks 240
8.6 Summary 247
8.7 References 247
10
Chapter 9. Algal biomass pretreatment in seawater reverse osmosis Mike B. Dixon,
Siobhan F.E. Boerlage, Nikolay Voutchkov, Rita Henderson, Mark Wilf, Ivan Zhu, S. Assiyeh
Alizadeh Tabatabai, Tony Amato, Adhika Resosudarmo, Graeme K. Pearce, Maria Kennedy,
Jan C. Schippers, and Harvey Winters
9.2 Chlorination in SWRO 252
9.3 Dechlorination in SWRO 257
9.4 Coagulation for DAF, DMF and UF pretreatment 257
9.5 DAF pretreatment for SWRO 272
9.6 Granular media filtration 278
9.7 Microscreens for membrane pretreatment 287
9.8 Microfiltration/ ultrafiltration 290
9.9 Cartridge filters for reverse osmosis pretreatment 297
9.10 Reverse osmosis 299
9.11 Summary of biomass removal in SWRO 307
9.12 References 308
Chapter 10. Removal of algal toxins and taste and odor compounds during desalination
Mike B. Dixon, Siobhan F.E. Boerlage, Holly Churman, Lisa Henthorne, and Donald M.
Anderson
10.2 Chlorination 316
10.3 Dissolved air flotation (DAF) 317
10.4 Granular media filters 317
10.5 Ultrafiltration/microfiltration 318
10.6 Reverse osmosis 320
10.7 Sludge treatment and backwash disposal 324
10.8 Toxin removal in thermal desalination plants 324
10.9 Chlorination prior to the distribution system 327
10.10 Summary 328
10.11 References 329
Chapter 11. Case histories for harmful algal blooms in desalination Siobhan F.E.
Boerlage, Mike B. Dixon, and Donald M. Anderson
11.2 Fujairah 2, United Arab Emirates – Effects of harmful algal blooms
on plant operations in 2008 and 2013 - Herve Faujour, Cyril de
Vomecourt, and Jérôme Leparc 347
11.3 Sohar, Oman – Harmful algal bloom impact on membrane pre-
treatment: challenges and solutions - Abdullah Said Al-Sadi and
Khurram Shahid 355
11.4 Barka 1, Oman - The impact of harmful algal blooms on the
performance stability of UF pretreatment - Graeme K. Pearce 367
11.5 Shuwaikh, Kuwait – Harmful algal bloom cell removal using dissolved
air flotation: pilot and laboratory studies - Robert Wiley, Mike Dixon,
and Siobhan F. E. Boerlage 383
11
11.6 La Chimba, Antofagasta, Chile – Oxygen depletion and hydrogen sulfide
gas mitigation due to harmful algal blooms - Walter Cerda Acuña, Carlos
Jorquera Gonzalez, and Victor Gutierrez Aqueveque 391
11.7 Mejillones, Chile – Operation of the ultrafiltration system during harmful
algal blooms at the Gas Atacama SWRO plant - Frans Knops and
Alejandro Sturnilio 399
11.8 Antofagasta, Chile - Abengoa water micro/ultrafiltration pretreatment
pilot plant - Francisco Javier Bernaola, Israel Amores, Miguel Ramón,
Raquel Serrano and Juan Arévalo 409
11.9 Tampa Bay, Florida (USA) – Non-toxic algal blooms and operation of the
SWRO plant detailing monitoring program for blooms - Lauren Weinrich 417
11.10 Jacobahaven, The Netherlands – Ultrafiltration for SWRO pretreatment: A
demonstration plant - Rinnert Schurer, Loreen O. Villacorte, Jan C.
Schippers, and Maria D. Kennedy 425
11.11 Barcelona, Spain - SWRO demonstration plant: DAF/DMF versus
DAF/UF - Joan Llorens, Andrea R. Guastalli and Sylvie Baig 439
11.12 Gold Coast, Queensland, Australia - Deep water intake limits
Trichodesmium ingress - Dianne L. Turner, Siobhan F. E. Boerlage
and Scott Murphy 447
11.13 Berlin, Germany – akvola: An integrated DAF-UF pilot - Johanna
Ludwig and Matan Beery 459
Appendix 1. Algal species potentially harmful to desalination operations

David G. Borkman and Donald M. Anderson 465
Appendix 2. Rapid screening methods for harmful algal bloom toxins

Donald M. Anderson, Maurice Laycock, and Fernando Rubio 485
Appendix 3. Methods for measuring transparent exopolymer particles and

their precursors in seawater
Loreen O. Villacorte, Jan C. Schippers, and Maria D. Kennedy 501
Appendix 4. Preservatives and methods for algal cell enumeration

Donald M. Anderson and Bengt Karlson 509
Appendix 5. Autopsy and cleaning of reverse osmosis elements affected by harmful

algal bloom-contaminated seawater
Nuria Peña, Steve Chesters, Mike Dixon, and Siobhan F. E. Boerlage 519
Index 527
12
Contributing Authors and Affiliations
Abdullah Said Al-Sadi

Majis Industrial Services, Sohar, Oman
E-mail: abdullah.al-sadi@miscoman.com
Tony Amato
Water2Water Consulting, England, UK
E-mail: water2water@btinternet.com
Israel Amores
Abengoa, Spain
E-mail: israel.amores.bautista@gmail.com
Clarissa R. Anderson
University of California, Santa Cruz, Santa Cruz, CA, USA
E-mail: clrander@ucsc.edu
Donald M. Anderson
Biology Department, Woods Hole Oceanographic Institution, Woods Hole, MA USA
E-mail: danderson@whoi.edu
Juan Arévalo
Abengoa, Spain
E-mail: arevalo.vilches@gmail.com
Sylvie Baig
Degrémont SA, Rueil-Malmaison Cedex, France
E-mail: Sylvie.baig@degremont.com
Peter Baker
South Australian Water Corporation, Adelaide, South Australia, 5000
Water Research Australia, Adelaide, South Australia, 5000
E-mail: peter.baker@sawater.com.au
Matan Beery
akvola Technologies, Berlin, Germany
E-mail: beery@akvola.com
Francisco Javier Bernaola
Abengoa, Spain
E-mail: patxobernaola@yahoo.com
David G. Borkman
Pausacaco Plankton, Saunderstown, RI USA
E-mail: David.Borkman@dem.ri.gov
Siobhan F.E. Boerlage
Boerlage Consulting, Gold Coast, Queensland, Australia
E-mail: desal@boerlageconsulting.com
Michael Burch
E-mail: mike.burch@sawater.com.au
Walter Cerda Acuña
Aguas Antofagasta, Anofagasta, Chile
E-mail: wcerda@aguasantofagasta.cl
Steve Chesters
Genesys International, Cheshire, United Kingdom
E-mail: schesters@genesysro.com
Holly Churman
Water Standard, Houston, TX, USA
E-mail: hchurman@waterstandard.com
13
Kathyrn J. Coyne
University of Delaware, Lewes, DE, USA
E-mail: kcoyne@udel.edu
Cyril de Vomecourt
Veolia Middle East, Dubai, United Arab Emirates
E-mail: Cyril.devomecourt@veolia.com
Mike B. Dixon
MDD Consulting, Kensington, Calgary, Alberta, Canada
E-mail: mikedixondesalination@gmail.com
Herve Faujour
E-mail: herve.faujour@veolia.com
Andrea R. Guastalli
University of Barcelona, Barcelona, Spain
E-mail: guastalli.a@gmail.com
Victor Gutierrez Aqueveque
E-mail: vgutierrez@aguasantofagasta.cl
Rita Henderson
School of Chemical Engineering, The University of New South Wales, Australia
E-mail: r.henderson@unsw.edu.au
Lisa Henthorne
Water Standard, Houston, TX, USA
E-mail: lhenthorne@waterstandard.com
Philipp Hess
IFREMER, Laboratoire Phycotoxines, 44311 Nantes, France
E-mail: Philipp.Hess@ifremer.fr
Lionel Ho
E-mail: lionel.ho@sawater.com.au
Jenny House
E-mail: jenny.house@sawater.com.au
Carlos Jorquera Gonzalez
E-mail: cjorquera@aguasantofagasta.cl
Bengt Karlson
Swedish Meteorological and Hydrological Institute, Gothenberg, Sweden
E-mail: Bengt.Karlson@smhi.se
Maria D. Kennedy
UNESCO-IHE Institute for Water Education, Delft, The Netherlands
E-mail: m.kennedy@un-ihe.org
Frans Knops
X-Flow BV / Pentair Water Process Technology BV,
Enschede, the Netherlands
E-mail: Frans.knops@pentair.com
Raphael M. Kudela
University of California, Santa Cruz, Santa Cruz, CA USA
E-mail: kudela@ucsc.edu
14
Maurice Laycock
Scotia Rapid Testing Ltd, Chester Basin, Nova Scotia, Canada
E-mail: mlaycock@bellaliant.net
Jérôme Leparc
Veolia Recherche and Innovation, Maisons Laffitte, France
E-mail: jerome.leparc@veolia.com
Joan Llorens
E-mail: jllorensl@ub.edu
Johanna Ludwig
E-mail: ludwig@akvola.com
Thomas M. Missimer
Florida Gulf Coast University, Fort Myers, FL USA
E-mail: tmissimer@fgcu.edu
Scott Murphy
Veoila Australia and New Zealand, Gold Cost, Australia
E-mail: scott.murphy@veolia.com.au
Gayle Newcombe
E-mail: gayle.newcombe@sawater.com.au
Thomas M. Pankratz
Water Desalination Report, Houston, TX USA
E-mail: tp@globalwaterintel.com
Graeme K. Pearce
Membrane Consultancy Associates Ltd, Reading, UK
E-mail: graemekpearce@btinternet.com
Nuria Peña
Genesys Membrane Products, S.L. Spain
E-mail: npena@genesysro.com
Peter Petrov
Kuwait Institute for Scientific Research (KISR)
E-mail: ppetrov@kisr.edu.kw
Miguel Ramón
Abengoa, Spain
E-mail: mramonmuro@gmail.com
Adhikara Resosudarmo
The University of New South Wales, Australia
E-mail: adhikara.resosudarmo@student.unsw.edu.au
Fernando Rubio
Abraxis LLC, Warminster, PA USA
E-mail: frubio@abraxiskits.com
Jan C. Schippers
E-mail: jancschippers@gmail.com
Rinnert Schurer
Evides Water Company, Rotterdam, the Netherlands
E-mail: r.schurer@evides.nl
Kevin G. Sellner
Chesapeake Research Consortium, Edgewater, MD, USA
E-mail: sellnerk@si.edu
15
Raquel Serrano
Abengoa, Spain
E-mail: raquelserrano@hotmail.com
Khurram Shahid
Water Solutions International Ltd, Gatwick, England
E-mail: kshahid@waterwastewater.uk
Alex Soltani
Alex Soltani Consulting, Calgary, Alberta, Canada
E-mail: alexander.soltani@outlook.com
Richard P. Stumpf
National Oceanographic and Atmospheric Administration, Silver Spring, MD USA
E-mail: Richard.Stumpf@noaa.gov
Alejandro Sturniolo
RWL Water Unitek, Mar del Plata, Argentina
E-mail:asturniolo@rwlwater.com
S. Assiyeh Alizadeh Tabatabai
E-mail: assiyeh.alizadeh@gmail.com
Dianne L. Turner
E-mail: dianne.turner@veolia.com.au
Loreen O. Villacorte
GRUNDFOS Holding A/S, Bjerringbro, Denmark (current affiliation)
E-mail: lvillacorte@grundfos.com
Nikolay Voutchkov
Water Globe Consulting, Winter Springs, FL, USA
E-mail: nvoutchkov@water-g.com
Lauren Weinrich
American Water, Voorhees, NJ USA
E-mail: Lauren.Weinrich@amwater.com
Robert Wiley
Leopold, a Xylem Brand, Zelienople, PA, USA
E-mail: Bob.Wiley.III@xyleminc.com
Mark Wilf
Mark Wilf Consulting, San Diego, CA, USA
E-mail: Mark.Wilf@rotechnology.net
Harvey Winters
Fairleigh Dickinson University, Teaneck, NJ USA
E-mail: harvey@fdu.edu
Ivan Zhu
Leopold, a Xylem Brand, Zelienople, PA USA
E-mail: ivan.zhu@evoqua.com
16
Harmful algal blooms
1! HARMFUL ALGAL BLOOMS
Donald M. Anderson1
1
Woods Hole Oceanographic Institution, Woods Hole, MA USA
!
1.1! Algal blooms ........................................................................................................................................... 17!
1.2! Harmful or toxic bloom species .............................................................................................................. 20!
1.3! Algal cell characteristics ......................................................................................................................... 22!
1.3.1! Toxins .............................................................................................................................................. 22!
1.3.2! Cell size ........................................................................................................................................... 25!
1.3.3! Cell wall coverings and surface charge ........................................................................................... 25!
1.3.4! Life histories .................................................................................................................................... 31!
1.4! Trends and species dispersal ................................................................................................................... 32!
1.5! Growth features, bloom mechanisms ...................................................................................................... 34!
1.5.1! Scale of blooms ............................................................................................................................... 34!
1.5.2! Cell growth ...................................................................................................................................... 35!
1.5.3! Bloom dynamics and coastal oceanography .................................................................................... 36!
1.5.4! Bloom initiation ............................................................................................................................... 36!
1.5.5! Bloom transport ............................................................................................................................... 36!
1.5.6! Fronts ............................................................................................................................................... 37!
1.5.7! Upwelling systems........................................................................................................................... 38!
1.5.8! Alongshore transport ....................................................................................................................... 38!
1.5.9! Vertical distributions ....................................................................................................................... 39!
1.6! Summary ................................................................................................................................................. 41!
1.7! References ............................................................................................................................................... 42!
1.1! ALGAL BLOOMS
Oceans and freshwater rivers, lakes, and streams teem with microscopic plants called algae
that capture the sun’s energy with their pigments and grow and proliferate in illuminated
surface waters, typically through simple cell division. These increases in abundance over
background levels are termed “blooms”, analogous to the growth and flourishing of terrestrial
plants. Many algal species are non-motile and thus their distributions are simply determined
by the movements of water. Even species that swim are not powerful enough to control their
location in most situations, so they too are generally dominated by the motion of waves,
currents, and tides (though the combination of swimming behaviour and water movement can
lead to dense cell aggregations and other spatial features (see section 1.5.3.6)). The
microscopic algae are called phytoplankton (drifting, single celled plants), to be distinguished
from their close relatives, the multi-cellular macroalgae or seaweeds. Algae of both types are
critical to life on earth, as they produce half of the oxygen we breathe and represent the base
of the aquatic food chain that provides substantial food for human society.
Among the many thousands of species of microalgae are a few hundred that cause harm in
various ways. Potentially harmful species are found in multiple phytoplankton groups. Many
are eukaryotes, (i.e., organisms with a nucleus and other organelles enclosed within
membranes) such as dinoflagellates, raphidophytes, diatoms, euglenophytes, cryptophytes,
haptophytes, pelagophytes, and chlorophytes. Some are prokaryotes, (i.e., single-celled
organisms such as cyanobacteria that lack a membrane-bound nucleus or other organelle).
While dinoflagellates comprise the majority of toxic harmful algal bloom (HAB) species in
the marine environment where seawater reverse osmosis (SWRO) plants are located, many of
the toxic species that pose a threat to drinking water supply in fresh- or brackish-water
systems are cyanobacteria.
17
Historically, blooms of harmful species are sometimes called “red tides”, as in some cases,
these microscopic cells increase in abundance until their pigments make the water appear
discolored and often red (Figure 1.1).
There are, however, blooms of
species that are orange or green or
brown, and others which do not
reach cell concentrations high
enough to discolor the water, but
which still cause harm. This harm is
sometimes because of the potent
toxins produced by those algae, but
in other cases, the harm derives
from the accumulated algal biomass
that can shade aquatic vegetation,
Figure 1.1. Water discoloration due to “red tide” in Texas. deplete oxygen as that biomass
Photo: Texas Department of Wildlife.
decays, and cause other societal or
ecosystem disruptions. The scientific community now uses the term ‘harmful algal bloom’ or
HAB to describe these phenomena. The term HAB is very broad and covers blooms of many
types, but HABs all have one unique feature in common - they cause harm. HABs are most
common in coastal marine ecosystems, but they also occur in the open ocean, and in brackish
or freshwater systems.
Toxic algal blooms are defined as those that produce potent toxins (poisonous substances
produced within living cells or organisms), e.g., saxitoxin. These can cause illness or
mortality in humans as well as marine life through either direct exposure to the toxin or
ingestion of bioaccumulated toxin in higher trophic levels e.g. shellfish. Non-toxic HABs can
cause damage to ecosystems and commercial facilities such as desalination plants, sometimes
because of the biomass of the accumulated algae, and in other cases due to the release
of compounds that are not toxins (e.g., reactive oxygen species, polyunsaturated fatty acids,
mucilage) but that can still be lethal to marine animals or cause disruptions of other
types. One prominent example of this latter mechanism relates to the high biomass that some
blooms achieve. When this biomass begins to decay, oxygen is consumed, leading to
widespread mortalities of all plants and animals in the affected area. These “high biomass”
blooms are sometimes linked to excessive pollutant inputs, but can also occur in relatively
pristine waters.
Six human poisoning syndromes are linked to the consumption of shellfish or fish
contaminated by HAB toxins (Table 1.1): amnesic shellfish poisoning (ASP), diarrhetic
shellfish poisoning (DSP), neurotoxic shellfish poisoning (NSP), paralytic shellfish poisoning
(PSP), azaspiracid shellfish poisoning (AZP), and ciguatera fish poisoning (CFP). The latter
is not a threat to desalination plants because the causative species, Gambierdiscus toxicus,
lives attached to seaweeds, dead coral, and other surfaces on the ocean bottom, and thus will
not be drawn into plant intake waters to any significant extent. Other threats to human health
are posed by HAB-derived aerosols that cause respiratory problems and water-borne
compounds that lead to skin irritation.
Macroalgae (seaweeds) are also considered HABs, as blooms of macroalgae have been
increasing and causing impacts of various types along many of the world’s coastlines.
Macroalgal blooms often occur in nutrient-enriched nearshore areas that are shallow enough
for light to penetrate to the sea floor. Booms of buoyant seaweeds can accumulate at the
water surface. Both types of blooms have a broad range of ecological and societal effects,
and often last longer than “typical” phytoplankton HABs. Some, like the spectacular “green
18
Table 1.1. Human illnesses associated with HABs.

Syndrome Causative organisms Toxins Route of Clinical manifestations
produced acquisition
Ciguatera fish Gambierdiscus toxicus Ciguatoxins, Toxin passed up Acute gastroenteritis,
poisoning and multiple maitotoxin marine food chain; paresthesias and other
(CFP) Gambierdiscus species illness results from neurological symptoms
eating large,
carnivorous reef
fish
Paralytic Alexandrium species, Saxitoxins Eating shellfish Acute paresthesias and

shellfish Gymnodinium harvested from other neurological
poisoning catenatum, Pyrodinium affected areas manifestations; may
(PSP) bahamense var. progress rapidly to
compressum, and others respiratory paralysis and
death
Neurotoxic Karenia brevis and Brevetoxins Eating shellfish Gastrointestinal and

shellfish others harvested from neurological symptoms;
poisoning affected areas; respiratory and eye
(NSP) toxins may be irritation with aerosols
aerosolized by wave
action
Diarrhetic Dinophysis species; Okadaic acid Eating shellfish Acute gastroenteritis

shellfish Prorocentrum lima and others harvested from
poisoning affected areas
(DSP)
Azaspiracid Azadinium spinosum Azaspiracids Eating shellfish Neurotoxic effects with

shellfish and others harvested from severe damage to the
poisoning affected areas intestine, spleen, and liver
(AZP) tissues in test animals
Amnesic Pseudo-nitzschia Domoic acid Eating shellfish (or, Gastroenteritis,

shellfish australis and others possibly, fish) neurological
poisoning harvested from manifestations, leading in
(ASP) affected areas severe cases to amnesia,
coma, and death
tides” of northeast China (Figure 1.2; Smetacek and Zingone 2013) are floating masses of
seaweed that may pose significant problems to power plants, desalination plants, and
recreational resources in some areas. Despite the long list of HAB impacts that are well
known and recurrent throughout the world, (e.g., Hallegraeff 1993; Landsberg 2002;
Anderson et al. 2012) new impacts are emerging. One current example is with desalination
plants. The global expansion of HABs due to pollution, coastal development, and other
factors (see section 1.4), is occuring at a time when there is also an increase in the
construction of seawater desalination plants. In 2015, there were more than 18,600 contracted
or commissioned desalination plants in more than 150 countries worldwide, and the
desalination market is forecast to grow by 12% per year (Virgili 2015). Interactions between
19
some of these plants and nearshore

HABs is inevitable. Concerns that arise
include the possible retention of algal-
produced toxins and taste and odor
compounds in treated water, as well as
the clogging of filters and fouling of
membranes. Algal biomass (i.e., the
solid or particulate component of an
algal bloom) and algal-derived
compounds (those dissolved in
seawater) can be seriously disruptive,
particularly to those plants that use
SWRO to produce fresh water. A recent
example is the bloom of Cochlodinium
Figure 1.2. Spectacular “green tide” in Qingdao China. polykrikoides in the Gulf 1 and Sea of
These annually recurrent, massive outbreaks result from
the growth and accumulation of the seaweed Ulva
Oman in 2008/2009 that affected a large
prolifera that originate far to the south of Qingdao, carried number of SWRO desalination plants,
to the region by ocean currents. Photo: D. Liu. closing some for as long as four months
(Richlen et al. 2010; Shahid and Al Sadi
2015). Since economic considerations are leading to a huge expansion in SWRO plants
compared to those that use thermal processes, we can expect many more impacts of HABs on
desalination plants than have been recorded thus far. It is also likely that species that are not
considered harmful to other sectors of society will be harmful to the desalination industry
simply because they produce disproportionally large amounts of dissolved organic materials
and suspended solids. With proper documentation of bloom events and communication
between HAB scientists and the desalination industry, a list of species that are prolific
producers of algal organic matter (and that are non-toxic) can be generated and used by
desalination plant operators to facilitate mitigation strategies.
1.2! HARMFUL OR TOXIC BLOOM SPECIES
Although many different phytoplankton and macroalgal species are now considered harmful,
this group still represents a small fraction of the many thousands of species of algae in the
ocean. Moestrup et al. (2017) list 144 toxic or harmful marine algal species. This list contains
species known to produce toxins as well as those that cause harm due to excessive biomass,
mucus production, or morphology, (spines etc.). Another 35 toxic cyanobacterial species are
listed, but these are predominantly from fresh water. The list, which is continually updated, is
available at: http://www.marinespecies.org/hab/index.php.
There is no list of species that have caused harm, or are likely to, at desalination plants. This
is in part because plants that have been affected by HABs often do not have the taxonomic
expertise to identify the organisms that are causing problems, and rarely do those plants send
bloom samples to the appropriate experts. All too often, plants experience problems from
algal blooms, but no identification of the causative algal species is made or publicized. In
hopes that this will change going forward, Chapter 3 provides guidance on how to collect
water samples for algal identification and counting, and Chapter 11 presents case studies of
algal bloom events and the steps taken to try to mitigate their impacts. Table 1.2 lists some of
1
Here the Gulf refers to the shallow body of water bounded in the southwest by the Arabian Peninsula and Iran
to the northeast. The Gulf is linked with the Arabian Sea by the Strait of Hormuz and the Gulf of Oman to the
east and extends to the Shatt al-Arab river delta at its western end.
20
the most common toxin-producing species, and Appendix 1 provides photographs and short
descriptions of some of these species, as well as others that might represent a threat to
desalination plants. With thousands of species of algae in the ocean, many of which form
blooms at one time or another, it is not possible to list all those that represent possible threats
to desalination plants. Thus, neither the list in Table 1.2 nor the species described in
Appendix 1 are comprehensive. Readers are urged to refer to the Web links provided in
Chapter 3 for other identifications. Readers are also urged to contact the Editor with
information on species that cause desalination plant problems in the future, as Appendix 1
and other parts of this manual will be updated periodically and information made available
online through MEDRC (http://www.medrc.org/) and IOC (http://hab.ioc-unesco.org/)
websites.
Table 1.2. Some toxic marine planktonic species of potential concern for SWRO operations.
(Adapted from Caron et al. 2010). This list is not comprehensive.
Microalgae Toxin(s) Poisoning Syndrome References
Diatoms
Pseudo-nitzschia spp. Domoic acid (DA) Amnesic Shellfish Subba Rao et al. (1988);
P. australis Poisoning (ASP) Bates et al. (1989); Martin et
P. brasiliana al. (1990); Buck et al.
P. caciantha Human effects (1992); Garrison et al.
P. calliantha •! Gastrointestinal (1992); Rhodes et al. (1996);
P. cuspidata symptoms Horner et al. (1997);
P. delicatissima •! Neurological Lundholm et al. (1997);
P. fraudulenta symptoms Rhodes et al. (1998); Trainer
P. fukuyoi •! Death et al. (2000, 2001); Baugh et
P. galaxiae Ecosystem effects al. (2006)
P. granii •! Marine mammal
P. kodamae mortalities
P. multiseries •! Bird mortalities
P. multistriata
P. plurisecta
P. pungens
P. pseudodelicatissima
P. seriata
P. subpacifica
P. turgidula
Dinoflagellates
Alexandrium spp. Saxitoxins (STXs) Paralytic Shellfish Sommer and Meyer (1937);
A. acatenella Poisoning (PSP) Gaines and Taylor (1985);
A. catenella1 Human effects Steidinger (1993); Scholin et
A. fundyense1(renamed •! Gastrointestinal al. (1994); Taylor and
A. catenella) symptoms Horner (1994); Jester (2008);
A. hiranoi •! Paralysis John et al. (2014); Usup et
A. ostenfeldii1 •! Death al. (2012); Prud'homme van
A. pacificum1 Ecosystem effects Reine WF. (2017)
A. australiense1 •! Marine mammal
Pyrodinium bahamense mortalities
Gymnodinium catenatum
Lingulodinium polyedrum
Gonyaulax spinifera Yessotoxins Human and ecosystem
Protoceratium reticulatum (YTXs) effects Holmes et al. (1967); Draisci
None reported, but et al. (1999a); Armstrong
animal bioassays show and Kudela (2006); Rhodes
toxicity et al. (2006); Howard et al.
(2007)
21
Table 1.2. (Continued)
Microalgae Toxin(s) Poisoning References

Syndrome
Dinoflagellates (Cont.)
Azadinium spp. Azaspiracids Azaspiracid Shellfish Satake et al. (1998);
A. spinosum (AZAs) Poisoning (AZP) James et al. (2003);
A. trinitatum Jauffrais et al. (2012);
A. cuneatum Human effects McCarron et al. (2009);
A. concinnum •! Gastrointestinal Ofuji et al. (2014);
A. dalienense symptoms Tillman et al. (2009,
A. poporum Ecosystem effects 2010, 2011, 2012)
A. obesum •! None reported
Raphidophytes Brevetoxins Neurotoxic Shellfish Loeblich and Fine

Chattonella marina (PbTxs); other fish- Poisoning (NSP) (1977); Hershberger
Fibrocapsa japonica killing toxins, et al. (1997); Gregorio
Heterosigma akashiwo possibly related to Human effects and Connell (2000);
fatty acids and •! Gastroenteritis Tyrell et al. (2002);
oxygen radicals •! Neurologic symptoms O’Halloran et al.
•! Respiratory irritation (2006)
and/or failure
Ecosystem effects
•! Marine mammal
mortalities
•! Fish mortality events
1
All members of the “tamarensis” complex of Alexandrium were recently reclassified by John et al. (2014). See
also Prud'homme van Reine WF. (2017).
As an alternative, Table 1.3 presents a list of algal species that have bloomed in the Arabian
Gulf, Sea of Oman, and Arabian Sea region, the global center of desalination activity. Once
again, this is not a comprehensive list, and only some of these species have been documented
to cause problems in desalination plants, but the list does show the diversity of organisms that
can achieve high biomass levels that probably would cause disruptions if those blooms
occurred near plant intakes. Unfortunately, for some bloom-formers the species designation
was not known or specified in the publications, so only genus names can be listed. In time, it
would be of great value to add resolution at the species level to tables such as this, as well as
more details about the cell size and cell wall characteristics of many of these species.
The best advice to plant operators seeking to mitigate the effects of a specific algal bloom is
to collect samples and identify the causative organism, hopefully to the species level, but at
least to genus. With some training and modest microscope facilities, this can be done on site
(Chapter 3). There are also outside experts and services that will do this type of work on
demand. The Intergovernmental Oceanographic Commission (IOC) Science and
Communication Centre on Harmful Algae, University of Copenhagen, Denmark can offer
assistance in identification of eukaryotic microalgae – see
http://hab.ioc-unesco.org/index.php?option=com_content&view=article&id=15&Itemid=0.
1.3! ALGAL CELL CHARACTERISTICS
1.3.1! Toxins
Among the thousands of species of microalgae in the ocean, only a hundred or so are toxic.
This means that most algal blooms at desalination plant sites are likely to be of concern
because of the algal biomass or associated organic products. Nevertheless, when toxic species
22
do occur, it is important to be aware of the dangers. Here common aspects of cell physiology
and toxin production are described. Details on the chemical structure and other properties of
HAB toxins are given in Chapter 2, which also discusses the levels of these toxins that pose
risks to human consumers. Chapter 10 discusses the removal of HAB toxins and taste and
odor compounds during SWRO and thermal desalination.
Table 1.3. Common bloom-forming species in the Arabian Gulf, Sea of Oman, and Arabian
Sea region. (Adapted from Al Shehhi et al. 2014; Al Azri et al. 2012).
Country Region Observed species References
India Arabian Sea Trichodesmium erythraeum, Noctiluca scintillans, D’Silva et al. (2012); Saeedi
Thalassiothrix longissima, Amphiprora sp., et al. (2011); Padmakumar
Thalassiosira sp., Fragilaria cylindrus et al. (2012); Joseph et al.
(2008); Krishnan et al.
(2007)
Bahrain Arabian Gulf Gonyaulax sp., Noctiluca sp.
KSA
Pakistan Arabian Sea Noctiluca scintillans, Prorocentrum minimum, Chaghtai and Saifullah
Phaeocystis sp. (2006); Saifullah (1979);
Rabbani et al. (1990);
Chaghtai and Saifullah
(2001)
Kuwait Arabian Gulf Noctiluca scintillans , Karenia sp., Gymnodinium Heil et al. (2001);
sp., Gymnodinium impudicum, Pryodinium Thangaraja et al. (2007);
bahamense, (Karenia selliformis, Prorocentrum Glibert et al. (2002); Al-
rathymum) Yamani et al. (2000)
UAE Arabian Gulf Noctiluca scintillans, Cochlodinium polykrikoides, Thangaraja et al. (2007);
Sea of Oman Trichdesmium erythaeum, Dinophysis caudata, Richlen et al. (2010); R.
Prorocentrum minimum, P. triestinum, P. balticum, Alshihi, pers. comm.; A.
P. micans, Coscinodiscus radiatus, Chaetoceros Rajan (pers. comm.)
peruvianus, C. compressus, C. curvisetus, C.
socialis, Cylindrotheca closterium, Guinardia
delicatula, Pseudo-nitzschia multiseries, P. pungens,
P. seriata, P. delicatissima, Skeletonema costatum,
Alexandrium sp., Amphidinium klebsii, Akashiwo
sanguinea, Ceratium furca, C. tripos, Dinophysis
miles, D. acuminata, Gonyaulax polygramma, G.
spinifera, Gonyaluax grindleyi, Gymnodinium
sanguinium, Peridinium quinquecorne,
Protoceratium reticulatum, Gyrodinium sp.,
Ostreopsis lenticularis, Dictyocha fibula,
Pyrodinium bahamense, Scrippsiella trochoidea,
Rhizosolenia setigera, Skeletonema costatum,
Leptocylindrus danicus, Bacteriastrum delicatulum
23

Country Region Observed species References
Oman Arabian Sea, Phaeocystis globosa, Nitzschia longissima,, Navicula Madhupratap et al. (2000);
Sea of Oman directa, Rhizosolenia spp., Chaetoceros didymus, Thangaraja et al (2007);
Noctiluca scintillans, Gymnodinium sp., Karenia sp., Morton et al. (2002); Al
Dinophysis sp., Trichodesmium sp., Coscinodiscus Azri et al. (2012); Tang et
sp., Ceratium furca, Prorocentrum arabianum, al. (2002); Al-Busaidi et al.
Prorocentrum minimum, Gymnodinium breve (2008); Al Gheilani et al.
(2011); Saeedi et al. (2011)
Qatar Arabian Gulf Pseudo-nitzschia spp., Alexandrium spp., Pyrodinium Al-Ansi et al. (2002)
bahamense
Alexandrium sp., Dinophysis sp., Pseudo-nitzschia
sp., Gymnodinium breve
Iran Arabian Gulf Karenia spp, Cochlodinium polykrikoides, Thangaraja et al. (2007);
Trichodesmium sp., Noctiluca scintillans, Navicula Fatemi et al. (2012)
sp.
___________________________________________________________________________
For virtually all HAB species, toxin production is a constitutive property of the cell, meaning
that if toxin is produced, it is present in all stages of growth; however, the amount of toxin in
a cell can vary dramatically with growth conditions. Some cells, such as Dinophysis species
that produce okadaic acid, for example, produce less toxin when they are actively dividing
(exponential phase growth) than when they are limited by some nutrient(s) and are in what is
termed “stationary phase” (Figure 1.3). The exact opposite occurs with other species, such as
those in the genus Alexandrium that produce saxitoxin. In those species, some of the highest
toxin production during a growth cycle occurs
when the cells are growing exponentially.
Furthermore, in those species, the amount of
toxin produced can vary with different types
of nutrient limitation. Cells that run out of
phosphorus, for example, produce much more
saxitoxin than those that are nutrient replete.
By reducing nitrogen supplies, the cells can
be made much less toxic.
Thus, the nutritional characteristics of the
Figure 1.3. Phases of algal growth in laboratory water in which the HAB is occurring can
batch culture. In lag phase (A), there is no growth influence levels of toxicity, sometimes as
after the initial inoculation; cells divide and increase much as 10-fold. Further complicating efforts
exponentially in phase B, then enter stationary to estimate the amount of toxin in a given
phase (C) when no growth occurs again because
nutrients or other growth factors are depleted or sub
HAB is the genetic heterogeneity between
optimal. Phase D represents death or mortality of strains of the same species. This means that
the culture. Modified from M. Komorniczak. strains of a given species isolated from
different locations, or from the same location
at different times, can vary dramatically in toxicity (sometimes 100 fold or more) even when
those strains are grown under identical conditions. Identifying the species and counting the
cells that are being drawn into a plant is a good start, but estimating the toxicity of those cells
is simply not reliable without detailed knowledge of the toxicity range of that species within
that region. In some cases, that information is published, but the range of values can often be
quite large for the reasons given above, and thus introduce considerable uncertainty into these
types of calculations. The best recommendation is to collect samples and either test them on
24
site using the simple toxin testing procedures described in Chapter 2 and Appendix 2, or to
send those samples to an appropriate testing laboratory for direct toxin analyses.
Recognizing that desalination plant operators might wish to estimate the amount of toxin
present in a specific bloom, Table 1.4 lists the maximum toxicity observed in laboratory
cultures of the major toxic HAB species. The table also provides estimates of the amount of
toxin that might be contained in a liter of water if blooms of 100,000, 500,000 and 5,000,000
cells/L of each species were present, if those cells contained the maximum level of toxicity
measured in laboratory cultures of that species. These are arbitrary cell concentrations
intended to represent small, moderate, and large blooms. This is a highly conservative way to
estimate the level of toxicity in a HAB, as the list of toxin values in the table is not
comprehensive, nor is the species list, and thus the data in Table 1.4 are presented here only
as a general guide.
1.3.2! Cell size
There are major physiological and morphological differences between species within a genus,
or species within different algal classes or phyla, and thus major differences in potential
impacts to desalination plants, or in the pretreatment options that might be effective in cell
removal. With respect to size, HAB cells can vary by a factor of more than 1000. Some, like
the brown tide organism, Aureococcus anophagefferens, are tiny 2-3 µm spheres, the upper
end of the range defining the picoplankton (0.2 – 2 µm), whereas at the other extreme there
are cells such as the prolific red tide-former Noctiluca scintillans that are 200 – 2,000 µm
(0.2 - 2 mm) in diameter. Many HAB cells, however, are in the 20 – 60 µm range. Some
HAB species occur as individual cells, occasionally paired with another as they divide,
whereas others form long chains or occur in colonies. Images and short descriptions of these
and other species described in this manual are given in Appendix 1.
1.3.3! Cell wall coverings and surface charge
Microalgal cells display a variety of surface features that can influence their resistance to
rupture (lysis), as well as their surface charge and thus their susceptibility for aggregation and
removal through processes such as dissolved air flotation (see Chapter 9). For example,
among the dinoflagellates, one of the major HAB classes, there are cells that have hard,
cellulose “plates” that are joined together like tiles to form a rigid cell wall, but there are also
many dinoflagellates that have soft, pliable, and easily ruptured cell walls. By definition, all
diatoms have silicified cell walls, but some have large, thick, and rigid coverings while others
have thin and fragile walls. More importantly, algal cells are often surrounded by various
organic molecules, such as nucleic acids, lipids, glycoproteins and carbohydrates (Dodge
1973). The surface charge of microalgae is thought to be generated by the hydrolysis or
ionization of these molecules (Maruyama et al. 1987). Ives (1956) was the first to determine
the surface charge of several freshwater species using electrophoresis and found that they
carried a negative charge; however, the author made no observations using flagellated species
because the swimming ability of these organisms interfered with their motion in the electric
field. Geissler (1958) also reported a negative charge on several freshwater diatoms and
confirmed the finding by observing the strong attachment of positively-charged dye particles
onto the cell surface. Tenney et al. (1969) demonstrated the binding of cationic polymers on
the cell surface and postulated that the association was electrostatic instead of chemical in
nature. Several authors have speculated that the surface charge of marine microalgae is also
negative (e.g. Yu et al. 1994). Sengco (2001) measured the electrophoretic mobility of nine
species of marine microalgae. All species displayed a slight electronegative charge ranging
from -0.19 to -0.57 x 10 -8 m2 s-1 V-1. Zeta potential ranged from -2.51 to -7.62 mV. These
25
Table 1.4. Maximum or representative toxicity for select HAB species. Also shown are water column toxin concentrations at three different cell
concentrations representative of potential blooms, calculated using the maximum toxin content estimates in cultures (typically measured under
nutrient-replete conditions). Toxin content varies with different types and degrees of nutrient limitation, and among strains of each of these
species, so this Table does not represent all possible situations.
Species Toxin Maximum toxin Maximum toxicity of dense bloom (µg/L) Reference
content 100,000 500,000 cells/L 5,000,000 cells/L
(pg/cell) cells/L
Alexandrium fundyense Saxitoxin 58.7 5.87 29.35 293.5 Anderson et al. 1994
(now A. catenella)
Alexandrium catenella Saxitoxin 18.3 1.83 9.15 91.5 Kim et al. 1993
(now A. pacificum),
Alexandrium minutum Saxitoxin 11.6 1.16 5.8 58 Chang et al. 1997
Alexandrium ostenfeldii Saxitoxin 217 21.7 108.5 1,085 Mackenzie et al. 1996
Spirolides 78.21 7.82 39.1 391 Gribble et al. 2005
Azadinium poporum Azaspiracid 0.02 0.0002 0.001 0.01 Krock et al. 2015
Azadinium spinosum Azaspiracid 0.1 0.001 0.01 0.1 Jauffrais et al. 2012
Dinophysis acuminata Okadaic acid, 58.8 5.88 29.4 294 Nagai et al. 2011
Dinophysistoxins 9.6 0.96 4.8 48
Pectenotoxins 73.3 7.33 36.65 366.5
Dinophysis acuta7 Okadaic acid, 51.8 5.18 25.9 259 Nielsen et al. 2013
Pectenotoxins 182 18.2 91 910
Dinophysis caudata Okadaic acid,

Dinophysistoxins
Pectenotoxins ~6003 60 300 3,000 Basti et al. 2015
Dinophysis fortii Okadaic acid, 49.4 4.94 24.7 247 Nagai et al. 2011
Pectenotoxins 191.1 19.1 95.5 955
content 100,000 500,000 cells/L 5,000,000 cells/L
(pg/cell) cells/L
Dinophysis norvegica Okadaic acid, 1.014 0.1 0.5 5 Rodriguez et al. 2015
Pectenotoxins 24.024 2.4 12 120
Dinophysis tripos Okadaic acid,

Dinophysistoxins 0.08 0.008 0.04 0.4 Nagai et al. 2013
Pectenotoxins 1236 123.6 618 6180
Gymnodinium Saxitoxin 101 10.1 50.5 505 Band-Schmidt et al. 2006

catenatum
Karenia brevis Brevetoxin 49 4.9 24.5 245 Corcoran et al. 2014
Karlodinium veneficum Karlotoxins 1.34 0.134 0.67 6.7 Fu et al. 2010
Ostreopsis siamensis Palytoxins 16 1.6 8 80 Suzuki et al. 2012
Pseudo-nitzschia Domoic acid 37 3.7 18.5 185 Garrison et al. 1992

australis
Pseudo-nitzschia Domoic acid .0095 0.00095 0.0048 0.048 Sahraoui et al. 2011
brasiliana
Pseudo-nitzschia Domoic acid 0.01 0.001 0.005 .05 Álvarez et al. 2009
calliantha
Pseudo-nitzschia Domoic acid 0.019 0.002 0.01 0.1 Trainer et al. 2009
cuspidata
Pseudo-nitzschia Domoic acid 0.12 0.012 0.06 0.6 Rhodes et al. 1998
delicatissima
fraudulenta
content
100,000 500,000 cells/L 5,000,000 cells/L
(pg/cell)
cells/L
Pseudo-nitzschia Domoic acid 0.00036 0.000036 0.00018 0.0018 Cerino et al. 2005
galaxiae
Pseudo-nitzschia granii Domoic acid 0.000004 0.0000004 0.000002 0.00002 Trick et al. 2010
Pseudo-nitzschia Domoic acid 67 6.7 33.5 335 Bates et al. 1999
multiseries
Pseudo-nitzschia Domoic acid 0.697 0.0697 0.349 3.49 Orsini et al. 2002
multistriata
Pseudo-nitzschia Domoic acid 0.0078 0.00078 0.0039 0.039 Moschandreou et al. 2010
pseudodelicatissima
pungens
Pseudo-nitzschia Domoic acid 33.6 3.36 16.8 168 Lundholm et al. 1994
seriata
Pseudo-nitzschia Domoic acid 0.09 0.009 0.045 0.45 Bill 2011
turgidula
Pyrodinium bahamense Saxitoxin 120 12 60 600 Usup et. al. 1994

1
(calculated from fmol/cell value - 692g/mole)
2
(DTX-1b)
3
(Calculated from plot)
4
(calculated from bulk analysis of plankton sample)
5
(lab incubation of nutrient enriched field sample)
Chapter 1 – Harmful algal blooms
data confirm the prediction that marine algal species, including the dinoflagellates, possess
negative surface charges like their freshwater counterparts (Maruyama et al. 1987; Shirota
1989; Yu et al. 1994). The magnitude of these charges was, however, small compared to
freshwater algae, as Ives (1956) reported a range of zeta potential between -7.6 mV to -11.6
mV at pH values from 7.2 to 8.8 in freshwater, while the values reported by Sengco (2001)
ranged from -2.5 to -7.7 mV. Zhu et al. (2014) measured the zeta potential of the
dinoflagellate Prorocentrum minimum to be about -4 mV at a moderate cell concentration.
A few of the species listed in Table 1.3 are worth highlighting because of the scale of their
impacts, or their prevalence in regions with significant desalination capacity, or simply
because they are prolific bloom formers. One of the most significant is Cochlodinium
polykrikoides, the organism that disrupted desalination operations at many plants in the
Arabian Gulf and Sea of Oman in 2008 and 2009 (Richlen et al. 2010; Shahid and Al Sadi
2015). First described from Puerto Rico in the Caribbean by Margalef (1961), the geographic
distribution of C. polykrikoides is widespread, and populations have been documented in
tropical and warm-temperate waters around the world, including the Caribbean Sea, eastern
and western Pacific Ocean, the eastern Atlantic Ocean, Indian Ocean, and Mediterranean Sea
(see Kudela et al. 2008; Matsuoka et al. 2008). This species has been spreading globally in
recent years and thus represents a significant threat to desalination operations worldwide.
This species does not produce a toxin that affects
humans, but it does produce massive, dense blooms
that cover large areas, frequently discolor the water,
kill coral reefs (Foster et al. 2011), and also has been
known to cause mass mortalities at fish farms and
other aquaculture facilities (reviewed in Kudela and
Gobler 2012). The mechanism(s) of fish or coral
mortality are not known, but Tang and Gobler (2009)
describe labile compounds similar to reactive oxygen
species (ROS). A significant amount of mucus or
Figure 1.4. Scanning electron mucilage is produced by C. polykrikoides (Figure 1.4),
micrograph of two cells of Cochlodinium
polykrikoides showing mucilage and this is undoubtedly one of the reasons it has been
excretions. Scale bar = 10 µm. Photo: S. so problematic at SWRO plants.
Morton.
Species within the genus Phaeocystis are well known
for their production of mono-specific (i.e., dominated by a single algal species), high-biomass
blooms worldwide (Schoemann et al. 2005). Among the 6 species in the genus, only 3 (P.
pouchetii, P. antarctica, P. globosa) have been reported as blooming species. Of particular
importance is the existence of a complex life cycle exhibiting alternation between small, free-
living cells 3–9 µm in diameter and gelatinous colonies usually reaching several mm. These
large colonies, consisting of thousands of small cells embedded in a polysaccharide matrix,
are not toxic. They are, however, significant threats to SWRO plants due to their high
particulate biomass and high organic content, and thus their potential to cause clogging of
filters and fouling of membranes.
Other dinoflagellate species, such as Gonyaulax hyaline and G. fragilis (e.g., Mackenzie et al.
2002; Sampedro et al. 2007) are noted for their massive mucilage production during blooms
in New Zealand and the Mediterranean Sea. The diatom Cylindrotheca closterium has been
linked to major mucilage events in the northern Adriatic Sea, stimulated by nutrient loadings
from the Po and other rivers (Ricci et al. 2014). Some of the mucilage events formed by
phytoplankton populations have been linked to high N/P ratios and increased stratification in
coastal waters, and thus are at least partially reflective of human influences on the nutrient
balance of coastal waters (Danovaro et al. 2009; Ricci et al. 2014).
29
Another diatom noted for mucilage production is Coscinodiscus wailesii, a species that has
been recorded worldwide and that causes blooms that harm shellfish and cultures of
macroalgae (e.g., Nagai et al. 1995), while also causing problems with commercial fisheries
operations due to net clogging. Its distribution, first restricted to the tropical Pacific and
western Atlantic oceans, has extended to Europe, the USA, and Japan in recent years. Some
of the damage from this species occurs when the mucilage aggregates, sinks, and covers the
seabed, where it can decay and cause anoxic conditions.
The filamentous blue-green alga Trichodesmium erythraeum is a common 'red tide' organism
in tropical and subtropical coastal waters. It can live as solitary cells or in floating colonies.
The colonies are visible to the naked eye and sometimes form extensive blooms. It is said
that the Red Sea derived its name from visible blooms of this organism – sometimes
described as "sea sawdust” or “sea straw". At the start of a bloom, the filaments usually
appear throughout the water column, but during late bloom stages, the development of strong
gas vacuoles causes Trichodesmium to rise to the surface of the water column. The alga is
perceived as a nuisance to swimmers on beaches and has significant impacts on recreation,
but harmful effects on humans or marine life have seldom been reported. Some species of
Trichodesmium have been reported to produce neurotoxins (e.g., Hawser et al. 1991; Kerbrat
et al. 2010, 2011). Colonies of Trichodesmium are capable of fixing atmospheric nitrogen
(i.e., obtaining nitrogen from N2 gas in seawater), which allows the alga to thrive under low
nutrient oceanic conditions. It is possible, however, that coastal nutrient pollution (especially
phosphates) can stimulate or prolong the blooms once they are washed inshore.
The final bloom-forming species to be highlighted here is Noctiluca scintillans, well known
for its production of vivid red or green tides (Figure 1.5) as well as intense blue-green
bioluminescence that lights up the water at night. This species occurs in two forms. Red
Noctiluca is heterotrophic (non-photosynthetic) and engulfs food from the water around it,
including, diatoms, other dinoflagellates, fish eggs and bacteria. In contrast, green Noctiluca
contains a photosynthetic symbiont (Pedinomonas noctilucae), but it also feeds on other
plankton when the food supply is abundant. Widely distributed throughout the world,
Noctiluca scintillans is often found along the coast in shallow areas of the continental shelf
where algal blooms occur that make up a large portion of this species’ diet (Harrison et al.
2011). Accordingly, Noctiluca blooms are often seen in areas where pollution and nutrient
enrichment due to human activities occur. Noctiluca is a large cell - roughly spherical,
ranging from 200 to 2,000 µm in diameter. Noctiluca does not appear to be toxic, but as it
feeds voraciously on phytoplankton, it accumulates and excretes high levels of ammonia into
the surrounding area, and some ecosystem impacts have been linked to that mechanism.
Figure 1.5. Green and red tides formed by Noctiluca. Photos: H. Gheilani and K.C. Ho.
30
Some report ammonia concentrations as high as 250 µg/L during Noctiluca blooms (G.
Hallegraeff, pers. comm.). This characteristic should be of interest to the desalination
industry because shock chlorination of water containing high levels of ammonia can lead to
production of the highly potent carcinogen N-nitrosodimethylamine (NDMA) (e.g., Mitch
and Sedlak 2002). Current NDMA guidelines for drinking waters are as low as 10 ng/L.
1.3.4! Life histories
A number of HAB species have dormant, cyst
stages in their life histories (Dale 1983) that are a
critical aspect of bloom initiation and decline.
These include Alexandrium spp., Pyrodinium
bahamense, Cochlodinium polykrikoides,
Gymnodinium catenatum, Chattonella spp.,
Pyrodinium bahamense, and Heterosigma
akashiwo. The highly resistant resting stages
remain in bottom sediments (sometimes
accumulating in high concentrations in areas
termed ‘seedbeds’; Anderson et al. 2014) when
conditions in the overlying waters are unsuitable
Figure 1.6. Life cycle diagram of Alexandrium
catenella. Stages are identified as follows: (1) for growth. When conditions improve, such as
vegetative, motile cell; (2) temporary or pellicle with seasonal warming, or simply after a certain
cyst; (3) ‘female’ and ‘male’ gametes; (4) period of dormancy or maturation, the cysts
fusing gametes; (5) swimming zygote or germinate, inoculating the water column with a
planozygote; (6) resting cyst or hypnozygote; population of cells that begins to divide asexually
(7,8) motile, germinated cell or planomeiocyte;
and (9) pair of vegetative cells following via binary fission to produce a bloom. As the
division. Redrawn from Anderson (1998). bloom progresses, vegetative growth ultimately
slows (typically due to the draw-down and
limitation of nutrients) and the cells undergo sexual reproduction, whereby gametes are
formed that fuse to form the swimming zygotes that ultimately become dormant cysts. Figure
1.6 shows the life history of Alexandrium catenella (formerly A. fundyense). Clearly, the
location of cyst seedbeds can be an important determinant of the location of resulting blooms,
and the size of the cyst accumulations can affect the magnitude of the blooms as well
(Anderson et al. 2014). Some cyst seedbeds can be enormous – two that were documented in
the Gulf of Maine, USA, are in excess of 22,000 km2, with total cyst abundances as high as
40 x 1016 cysts in the top cm of sediment alone (Anderson et al. 2014). Another way to view
these abundances is that many areas have in excess of 50 million cysts in one square meter of
bottom sediment. In many areas, however, the environmental regulation of cell division is
more important to eventual bloom magnitude than the size of the germination inoculum from
cysts.
Cysts are also important in species dispersal. Natural (storms or currents) or human-assisted
(e.g. via ballast water discharge or shellfish seeding) transport of cysts from one location to
another can allow a species to colonize a region and extend its geographic range. In 1972, a
tropical storm introduced Alexandrium catenella into southern New England waters from
established populations further to the north and east. Since that time, toxic blooms of the
species have become an annually recurrent phenomenon in the region. An example of
human-assisted species introductions is the appearance of Gymnodinium catenatum in
Tasmania in the 1970s, coincident with the development of a wood chip industry involving
commercial vessels and frequent ballast water discharge (McMinn et al. 1997).
31
In the context of desalination plant design and operations, it is important to recognize that
when a bloom of a cyst-forming species occurs in an area or region near a plant for the first
time, the species is likely to bloom again in that area in future years due to the deposition and
accumulation of cysts or resting stages. This is the manner in which some species disperse
and colonize new areas. In this context, the recent appearance of Cochlodinium polykrikoides
in the Arabian Gulf and Sea of Oman regions (Richlen et al. 2010) is indeed worrisome, as
that species has a cyst (Tang and Gobler 2012) and thus is likely to recur for many years.
1.4! TRENDS AND SPECIES DISPERSAL
The nature of the global HAB problem has changed considerably over the last several
decades. Virtually every coastal country is
now threatened by harmful or toxic algal
species, often in many locations and over
broad areas. Over 40 years ago, the problem
was much more scattered and sporadic.
Figure 1.7 shows the global expansion of the
PSP poisoning syndrome from 1970 to 2015
(caused by multiple Alexandrium species, as
well as Gymnodinium catenatum and
Pyrodinium bahamense). Figure 1.8 shows
the increase in the number of algal bloom
events along the coast of Oman from 1976 to
2011. Similar plots could be provided for
many parts of the world. Clearly, the number
of toxic or harmful blooms, the economic
losses from them, the types of resources
affected, and the number of HAB species
have all increased dramatically in recent
years. Disagreement only arises with respect
Figure 1.7. Expansion of the paralytic shellfish to the reasons for this expansion, of which
poisoning (PSP) problem over the past 35 years. there are several possibilities (Anderson
Sites with proven records of PSP-causing organisms 1989; Hallegraeff 1993).
are noted in 1970, and again in 2015. Source: US
National Office for Harmful Algal Blooms, Woods New bloom events may simply reflect
Hole Oceanographic Institution. indigenous populations that are discovered
because of improved chemical
detection methods, more
observers, and better scientific
communication. The discovery
of ASP toxins along the west
coast of the USA in 1991 (Work
et al. 1993) is a good example of
this, as toxic diatom species
were identified and their toxin
detected as a direct result of
communication with Canadian
scientists who had discovered
the same toxin and toxic species
four years earlier (Wright et al.
Figure 1.8. Time series of algal bloom events along the coast of
1989).
Oman. Modified from Al Azri et al. 2012.
32
Other ‘spreading events’ are most easily attributed to natural dispersal via currents, rather
than human activities. The first NSP event ever to occur in North Carolina was shown to be a
Florida bloom transported over 1500 km by the Gulf Stream – a natural phenomenon with no
linkage to human activities.
Many believe that humans have contributed to the global HAB expansion by transporting
HAB species in ship ballast water (e.g. Gymnodinium catenatum in Tasmania; McMinn et al.
1998) or by shellfish relays and seeding. Some types of species are easily transported -
especially those that produce resting stages, but long survival in the dark is also possible for
species that do not form cysts (e.g., Popels and Hutchins 2002). Furthermore, species that
bloom in high concentrations are more likely to be effectively transported since they can be
taken in as large populations in the ballast tank, leading to more survivors on discharge.
Human-mediated transport of HAB species is especially enhanced between sheltered areas,
such as harbors and aquaculture sites, where many dinoflagellates thrive. All these reasons
make many HAB species good candidates for transport, and the low number of invasions so
far demonstrated is most probably a conservative estimate of the real number.
Another factor underlying the global increase in HABs is that we have dramatically increased
aquaculture activities worldwide, and this inevitably leads to increased monitoring of product
quality and safety, revealing indigenous toxic algae that were probably always there.
Climatic changes can also affect HAB species distributions, either directly through
temperature variations or storms, or, indirectly, through periodic or long-term effects on
oceanographic conditions, e.g., changes in stratification or in water circulation patterns.
Nutrient enrichment is another major cause for the increasing frequency of HAB events in
some regions (GEOHAB 2006). Manipulation of coastal watersheds for agriculture, industry,
housing, and recreation has drastically increased nutrient loadings to coastal waters. Just as
the application of fertilizer to lawns can enhance grass growth, marine algae grow in
response to nutrient inputs to coastal waters. Fertilizer finds its way into lakes and oceans
through runoff from agricultural farms, golf courses, and suburban lawns. Other nutrients get
added from the atmosphere, soil erosion, upwelling, aquaculture facilities, and sewage plants.
Shallow and restricted nearshore waters that are poorly flushed appear to be most susceptible
to nutrient-related algal problems. Nutrient enrichment of such systems often leads to
excessive production of organic matter, a process known as eutrophication, and increased
frequencies and magnitudes of phytoplankton blooms, including HABs. There is no doubt
that this is occurring in certain areas of the world where pollution has increased dramatically.
It is real, but less evident in areas where coastal pollution is more gradual and unobtrusive.
A frequently cited dataset from an area where pollution was a significant factor in HAB
incidence is from the Inland Sea of Japan, where visible red tides increased steadily from 44
per year in 1965 to over 300 per year a decade later, matching the pattern of increased
nutrient loading from pollution (Figure 1.9). Effluent controls were instituted in the mid-
1970s, resulting in a 70% reduction in the number of red tides, a level that has persisted to
this day. A related data set for the Black Sea documents a dramatic increase in red tides up to
the mid-1990s, when the blooms began to decline. That reduction, which has also continued
to this day, has been linked to reductions in fertilizer usage in upstream watersheds by former
Soviet Union countries no longer able to afford large, state-subsidized fertilizer applications
to agricultural land (Anderson et al. 2002).
The Arabian Gulf, bordered by many desalination plants, is also experiencing problems with
nutrient pollution and eutrophication (Sheppard et al. 2010). It is thus no surprise that the
frequency of HABs is increasing in the Gulf, with examples being reported from almost all
33
areas (e.g., Thangaraja et al. 2007;

Al Azri et al. 2012; D’Silva et al.
2012; Saeedi et al. 2011). In
addition to the input from human
sewage and agriculture, a rise in
nutrient levels around fish cages
(mariculture) may be another
contributing factor to this increase.
In Kuwait Bay, for example,
mariculture led to a HAB incident
that killed caged and wild fish in
1999 (Al-Yamani et al. 2000).
The global increase in HAB
Figure 1.9. Time series of red tides and industrial production in phenomena is thus a reflection of
the Seto Inland Sea, Japan. Decreases in pollution inputs were two factors – an actual expansion of
mandated by the Seto Inland Sea Law in 1973, which was the problem due to a variety of
followed by a significant decrease in red tides thereafter,
including toxic blooms with fisheries impacts. Redrawn from
factors, and an increased awareness
Okaichi (2004). of its size or scale. It is expanding
due to natural processes as well as
through human activities such as pollution and ballast water-related species introductions, but
improved methods and enhanced scientific inquiry have also led to a better appreciation of its
true size. The fact that part of the expansion is a result of increased awareness or improved
detection capabilities should not negate our concern – HABs are a serious and growing
problem affecting many sectors of society and industry, including desalination.
1.5! GROWTH FEATURES, BLOOM MECHANISMS
1.5.1! Scale of blooms

HABs vary dramatically in their spatial and temporal scales. Some are small and localized,
occurring within embayments, harbors, or other coastal features, and covering small areas
(Figure 1.10, left). These types of blooms are often relatively short-lived, lasting a month or
two at most, terminating due to simple transport processes that carry the population out of an
area, or to environmental factors such as nutrient limitation or losses from the zooplankton
and other grazers that consume the HAB cells as food. At the other extreme, some HABs are
massive, extending hundreds of kilometers along coastlines, sometimes visible from space
Figure 1.10. A small, localized algal bloom in Norway, and a massive bloom in the Gulf of California, seen from
space. Photos: E. Dahl and NASA respectively.
34
(Figure 1.10, right). These large coastal blooms can move with winds, tides, and currents
such that the impacts not only extend spatially over large stretches of a coast, but also can be
sustained for weeks, months, or even years.
One feature that characterizes many HABs is that they are "patchy" in their distribution, both
horizontally and vertically. On a large scale, this is exemplified in Figure 1.10, an image
from a satellite that shows patches with dense accumulations of cells, interspersed with other
areas where the concentrations are much lower. Patchiness can also occur at much smaller
scales where the high concentrations of cells can be a kilometer or less in size.
The relevance of patchiness to desalination plant operators is that because of it, algal biomass
conditions can change dramatically from one day to another or one hour to another as patches
of HAB cells encounter the intake with normal currents and wind-driven flow. Pretreatment
processes that work one day may not work the next, or vice versa. In some cases, a
continuous pretreatment effort may not be warranted, as some days will have far fewer cells
than other times. In an ideal world, the operator would have bloom distribution data at
sufficient resolution to allow a pretreatment program to be undertaken that is responsive to
present conditions, but that also reflects knowledge of what will be coming into the plant in
the near future. In this regard, there are autonomous instruments of various types that can
provide information on algal species and their abundance on a real-time basis, (see Chapter
3) so that operators can adjust their pretreatment strategies accordingly. Another variant of
the patchiness concept is presented in section 1.5.3.6 where the vertical distribution of HAB
cells is discussed.
1.5.2! Cell growth
HABs are typically caused by single-celled algae that increase in abundance through a
process called binary fission – one cell divides to form two cells, those two cells become four,
four become eight, and so on. This is the “exponential growth” that is depicted in Figure 1.3.
The rate of cell division varies dramatically among species, with most dinoflagellates taking
1-3 days to divide under good growth conditions, whereas diatoms and other species can
divide several times in a single day. In some cases, HAB species are able to form nearly
mono-specific populations, benefitting from mechanisms such as grazer avoidance (through
swimming, migration patterns, or even cell morphology that discourages predators), or the
inhibition of grazers or other competing algal species through the release of allelopathic
substances – a type of inter-species chemical warfare likened to a “watery arms race”
(Smetacek 2001). At other times, HAB species are simply a minor component of a complex
community of co-occurring microscopic plants and animals, but they make themselves
known through their toxins or other harmful traits.
There are no easy generalizations about the cell concentrations that might be encountered
during blooms. At barely detectable concentrations (102–103 cells/L), some harmful species
can have dramatic effects; this is the case of highly toxic species like the PSP-producers
Pyrodinium bahamense var. compressum and Alexandrium catenella, which may also form
much denser blooms (106–107 cells/L). Other species, including many that are less toxic or
non-toxic, easily reach these same concentrations, while small forms such as Aureococcus
anophagefferens, a brown tide organism, can exceed 109 cells/L. Vertically migrating cells
like Cochlodinium polykrikoides can aggregate in dense patches at the water surface that
exceed 1010 cells/L. In contrast, the DSP-producing species of the genus Dinophysis rarely
reach concentrations higher than 104–105 cells/L.
A common misconception is that HABs are caused by the explosive growth of a single
species that then rapidly dominates the water column. Given the above information, however,
35
it should be clear that it is only necessary to have conditions that favor the growth of a
moderately large population of a given species and the proper hydrographic and
meteorological conditions to permit the accumulation of those cells. In other words, coupled
with normal or average growth rates, winds, tides, currents, fronts, and other hydrographic
features can combine with organism behavior (e.g., vertical migration) to create discrete
“patches” of cells (blooms) at all scales.
1.5.3! Bloom dynamics and coastal oceanography
Though the number of HAB species is only a small fraction of the many algal species in the
ocean, their diversity in terms of shape (morphology), physiology, and ecology is very large.
In effect, there are hundreds of common HAB species and they differ in their growth
characteristics and in the types of blooms they cause and the waters they inhabit. Given this
diversity of species and habitats in which they occur, there exist few unifying principles that
explain bloom dynamics in all habitats. A few common processes and mechanisms for bloom
development and accumulation are highlighted below.
1.5.4! Bloom initiation
HABs can be initiated from cells present at low concentrations, sometimes persisting in the
background for months or years before a bloom develops (the “hidden flora” concept). Other
HABs are delivered to a location by tides and
currents (see below). Still other HABs are
initiated from resting cysts (Figure 1.11) that
germinate from bottom sediments, significantly
impacting many aspects of HAB phenomena.
In those instances, cyst or spore germination
provides the inoculum for blooms, and the
transformation back to the resting state can
remove substantial numbers of vegetative cells
from the population and act as a major factor in
bloom decline. Cysts are also important for
population dispersal, including through ballast
water transfer. They permit a species to survive
Figure 1.11. Satellite image of sea surface through adverse conditions, and since sexuality
temperature, showing the warm Gulf Stream
(blue) off the coast of North Carolina, USA. is typically required for their formation, they
Several filaments of the Gulf Stream extend facilitate genetic recombination. They can even
toward shore. These carried a toxic population of be major sources of toxin to shellfish and other
Karenia brevis that originated in the Gulf of benthic animals.
Mexico, >1500 km away. Photo: NASA.
1.5.5! Bloom transport

Once a population is established, its geographic extent and biomass are affected by physical
controls such as the transport and accumulation of cells in response to water flows (e.g.,
Franks and Anderson 1992), by the swimming behavior of organisms (Kamykowski 1995)
and by the maintenance of suitable environmental conditions (including temperature and
salinity, stratification, irradiance, and nutrient supply). These factors all interact to determine
the timing, location, and ultimate biomass achieved by the bloom, as well as its impacts.
Physical processes that are likely to influence the population dynamics of HAB species are
operative over a broad range of spatial and temporal scales. Large-scale circulation patterns
(Figure 1.10) affect the distribution of water masses and their associated HABs. Some HABs
are delivered into a specific region via advection (transport) after developing elsewhere. In
36
such cases, the population increases can be significant and alarming, but should be attributed
to transport and not to in situ growth. Eddies (spinning masses of water) or filaments from
the open ocean (Figure 1.11) can, for example, impinge on shelf regions, transporting HABs
and nutrients to nearshore waters. This type of transport has been invoked for the delivery of
the Florida red tide organism Karenia brevis to nearshore waters from an offshore zone of
initiation (Walsh et al. 2006). Another prominent example is the wind-driven delivery of
offshore Dinophysis acuminata cells into Bantry Bay in southwest Ireland (Raine et al. 2010).
1.5.6! Fronts
Physical processes at intermediate scales can lead to the formation of convergence zones,
fronts, and upwelling. Fronts in the ocean (Figures 1.12, 1.13) are directly analogous to the
more familiar weather fronts where two
air masses meet, but in the ocean,
representing the boundary between two
water masses.
There are many types of fronts in the
ocean. A tidal front forms when the
water is well mixed in a shallow zone
due to tidal energy near the bottom and
Figure 1.12. Accumulation of HAB cells near a front. This winds at the surface, whereas the water
schematic demonstrates how cells can accumulate at a overlying deeper areas further offshore
frontal convergence, with a surface manifestation of the
bloom at the frontal convergence (right panel), and a
remains stratified (layered), and
subsurface extension of the bloom that extends below the therefore has a different density
surface along the sloping pycnocline (density structure. The front is the interface
discontinuity). Adapted from Franks 1992. between these two water masses.
Another common front is found within
or outside of estuaries where low
salinity, upstream waters meet higher
salinity coastal waters.
Just as storms and other dynamic
features are commonly found at fronts
in the atmosphere, oceanic fronts are
frequently the site of enhanced algal
biomass. This is the result of the
interaction between physical processes
Figure 1.13. Algal bloom showing accumulation at a front
such as upwelling, shear, and turbulence,
(straight lines parallel to shoreline), and in windrows or
streaks of cells aligned with the wind, and perpendicular to and physiological processes such as
shore. Photo: M. Godfrey. swimming and enhanced nutrient uptake
by the algal cells. One example is the
linkage between tidally generated fronts and the sites of massive blooms of the toxic
dinoflagellate Gyrodinium aureolum (now called Karenia mikimotoi) in the North Sea
(Holligan 1979). The pattern generally seen is a high concentration of cells at the surface of
the frontal convergence, contiguous with a subsurface cell maximum which follows the
sloping interface between the two water masses beneath the stratified side of the front (Figure
1.12). Foam and other debris also accumulate at the frontal interface, making it easy to see.
The surface signature of the chlorophyll or algal cell maximum at the front (sometimes a
visible red tide) may be 1–30 km wide. Cell concentrations are generally lower and more
uniform on the well-mixed, typically offshore side of the front. If located offshore, the bloom
at a front may be harmless, but when movement of the front and its associated cells brings
37
HAB populations into contact with fish and other susceptible resources, including
desalination plants, negative impacts can result.
Frequently, wind will create patterns called “windrows”. Winds that blow steadily across the
ocean, and the small waves that such winds generate, can create vortices, or rotating cells, in
the surface waters. These vortices align in the direction of the wind, and are made visible by
streaks of foam, seaweed, debris, and algal cells. Figure 1.13 shows an algal bloom at a
frontal feature, with windrows extending offshore.
1.5.7! Upwelling systems
A common mechanism that leads to a widespread and sustained HAB is when there is a
source population of cells living in offshore waters, with periodic delivery of those cells to
the nearshore zone as a result of winds and currents (GEOHAB 2005). This is often related to
processes called upwelling and upwelling relaxation or downwelling. Large upwelling
systems tend to occur along the eastern boundaries of the world’s oceans, such as along the
west coasts of the Americas, the continental shelves of northwest and southwest Africa, and
the western edge of the Iberian Peninsula. Smaller-scale upwelling can occur in virtually all
coastal waters if the wind blows in the proper direction to move surface waters away from the
coast. In these upwelling systems, the winds force offshore transport of surface water due to a
process called Ekman transport
that moves water 90 degrees to the
right of the wind direction. This
offshore flow is compensated for
by the onshore flow and
“upwelling” of deep, colder,
nutrient-rich water into coastal
zone (Figure 1.14). The boundary
along the coast between the
nearshore upwelled water and the
warmer adjacent ocean surface
water is usually a front (Figure
1.12) that has high biological
Fig. 1.14. Schematic of coastal downwelling, driven by productivity, including HABs,
alongshore winds that force surface waters offshore, and deeper fueled by the upwelled nutrients.
waters onshore. Image: Lichtspiel.
This feature can then be delivered
to shore as a result of the
relaxation or reversal of the winds, allowing offshore waters and the associated high biomass
populations to move to shore very quickly. The 2008/2009 Cochlodinium bloom in the
Arabian Gulf and Sea is thought to be an example of the upwelling process. Satellite imagery
shows large patches of cells throughout much of the eastern Arabian Gulf and Sea of Oman
during upwelling intervals (see Chapter 4), with these populations transported to shore where
they had devastating impacts (Zhao and Ghedira 2014).
1.5.8! Alongshore transport
The preceding sections describe mechanisms that concentrate cells at certain hydrographic
features and deliver them to shore via cross-sure transport (i.e. perpendicular to shore). A
related mechanism is alongshore transport whereby an established HAB is carried along a
coastline by currents, sequentially causing impacts at different sites. Many HABs originate in
one location and are delivered to other areas through this type of transport. These transport
events can span hundreds of km, and can last months. Upwelling and downwelling winds will
affect the location and timing of impacts on shore, acting on blooms as they move down the
38
coast. The appearance of HABs can be sudden and unexpected at times, and in some of these
instances, alongshore transport is the explanation. Operators should learn the nature of the
transport pathways in the vicinity of their plant intakes to better anticipate and respond to
HABs of this type.
An example of this type of transport is found with blooms of the PSP-toxin producing
Alexandrium catenella in the Gulf of Maine (USA). In that region, the bloom populations are
found within a low-salinity coastal current that travels along the coast (Franks and Anderson
1992). This water mass provides a suitable growth habitat for the Alexandrium cells as its
southerly transport is regulated by river runoff, Coriolis forces (due to the Earth’s rotation),
wind stress and other large-scale hydrographic forcings. Of these factors, wind appears to be
particularly important in determining transport variability in surface waters. Downwelling-
favorable winds (from the north or northeast) compress the plume against the coast and
accelerate it alongshore, while an upwelling favorable wind (from the south) thins the low-
salinity surface plume as it is moved offshore. Upwelling winds also retard the southward
progression of the plume and its associated cells. In the downwelling case, the bloom would
progress rapidly down the coast, whereas with upwelling, the bloom would move more
slowly. The general velocity of alongshore transport of this bloom and Alexandrium cells
during downwelling is roughly 0.2 m/s, or 17 km/d or 120 km in a week.
1.5.9! Vertical distributions
An important feature of some HAB events, particularly in the context of desalination, is that
the algal cells are often not uniformly distributed in the vertical direction (GEOHAB 2008).
Typically, the “water column” in nearshore waters has a stratified structure whereby there is
a surface layer that is warmer or fresher, and thus more buoyant than deeper waters. That
layer is often stirred by winds, and is thus termed the "mixed layer". Algal cells and other
constituents of the plankton that are non-motile are distributed uniformly through the mixed
layer. Species that swim, however, such as many HAB species, can override that random
mixing process and form aggregations at specific depths – sometimes at the surface,
sometimes deeper. In some instances, these subsurface layers of cells do not move vertically
because they are linked to some feature, such as the interface or density discontinuity
between the mixed layer and the deeper waters (called the pycnocline) below which nutrients
are typically higher than in the surface. Cells residing at that location would have ready
access to high nutrients. For example, off the French coast, a thin layer of dinoflagellates,
including the HAB species Dinophysis acuminata, is frequently observed in the proximity of
the pycnocline (Gentien et al. 1995). Thin, subsurface layers have been observed at scales as
small as 10 cm in the vertical and as large as 10 km in the horizontal.
Some subsurface cell aggregations do move up and down, typically on a daily basis, due to a
process called diel vertical migration. These migrating populations often reside in surface
waters during the day to harvest the sunlight for photosynthesis, and then swim to the
pycnocline and below to take up nutrients at night. This strategy can explain how dense
accumulations of cells can appear in surface waters that are devoid of nutrients and which
would seem to be incapable of supporting such apparent, prolific growth. In truth, the growth
is not rapid; the cells are aggregating, not growing fast. This behavior also explains the
disappearance, and the reappearance of cells or surface bloom patches on a daily basis.
Typically, vertical movement down to 10, 15 or even 20 m have been observed, depending
on the organism and depth of the mixed layer or water column. An example of the extent and
timing of vertical movement is seen in Park et al. (2001) who observed the behaviour of
Cochlodinium polykrikoides, the same species that caused desalination plant disruptions in
2008 and 2009 in the Arabian Gulf and Sea of Oman regions (Richlen et al. 2010; Shahid and
39
Al Sadi 2015). In the daytime, between 1100 and 1600 hrs, the species concentrated near the
surface at depths less than 2m (Figure 1.15). The highest surface cell concentrations were
observed at approximately 1600 hrs, when > 60% of the cells in the population were found in
the top meter, forming a visible red tide. Thereafter, the population began to migrate
downwards, leading to a distinct subsurface maximum at 1700 hrs. The species arrived at the
bottom (15 m in this study) by 2000 hrs, but cells were somewhat dispersed in the bottom
layers. The upward migration of the species from the bottom began around 0600 hrs, before
sunrise, and the population was concentrated back in the service layer by 1100 hrs once again.
This pattern is depicted in Figure 1.15.
Swimming speeds of C. polykrikoides determined from this study were ~3 m/h, which is fast
Figure 1.15. Temporal changes in the vertical distribution of Cochlodinium polykrikoides in Kwangyang
Bay, Korea. Relative densities (colors) are the ratios of cell counts at each depth to the average number of
cells in the water column. Modified from Park et al. 2001.
compared to many dinoflagellates, which typically swim ~ 1 m/h. Figure 1.16 shows the
hypothetical depth that different species of algae can reach under the assumption that they
swim downward for 10 hours. The fastest swimmer, and deepest migrator, is Cochlodinium
polykrikoides, but a number of other prolific bloom formers (e.g., Lingulodinium polyedrum,
Prorocentrum micans, Heterosigma akashiwo) or toxin producers (e.g., Alexandrium
catenella, A. minutum, Dinophysis acuta) all can descend 15 - 20 m in a day.
Vertical migration is a behavioral feature of some HABs that should be of concern to
desalination plants in the context of intake design, as well as routine operations. Near-surface
intakes (i.e., those with intake channels or intake pipes at depths of 4-5 m or less) would see
increased numbers of cells during daylight hours and low numbers at night. Deep intakes (i.e.,
those at 8-10 m or deeper) would see just the opposite – maximum numbers of cells of
vertically migrating species at night. One practical application of this knowledge would be
that in a major bloom of a migrating species (as in the 2008/2009 Cochlodinium HAB in the
Arabian Gulf and Arabian Sea), there might well be times of the day (nighttime) when the
plant could be operated without taking in too many cells. Likewise, this means that
40
pretreatment filtration rates that are effective at one time of day might not work well at other
times, and therefore that actions might need to be varied to compensate for the large
differences in the number of cells present in the intake waters at different times of the day.
Figure 1.16. Depth (m) which a phototrophic, motile algal species can reach when it swims downward
for 10 h. This depth was calculated by multiplying the maximum swimming speed (mm/sec.) of a given
species by 36,000 sec. Cp: Cochlodinium polykrikoides, Ga: Gymnodinium aureolum, Lp: Lingulodinium
polyedrum, Gp: Gonyaulax polygramma, Da: Dinophysis acuta, Ht: Heterocapsa triquetra, Am:
Alexandrium minutum, Aa: Alexandrium affine, At: Alexandrium tamarense, Cfur: Ceratium furca, Pm:
Prorocentrum micans, Kb: Karenia brevis, Ha: Heterosigma akashiwo, St: Scrippsiella trochoidea,
Dacu: Dinophysis acuminata, As: Akashiwo sanguinea, Pd: Prorocentrum donghaiense, Gc:
Gymnodinium catenatum, Ac: Amphidinium carterae, Eg: Eutreptiella gymnastica, Km: Karenia
mikimotoi, Ao: Alexandrium ostenfeldii, Ca: Chattonella antiqua, Pmin: Prorocentrum minimum, Pt:
Prorocentrum triestinum, Acat: Alexandrium catenella, Ct: Ceratium tripos, Cfus: Ceratium fusus, Tel:
Teleaulax sp., Rs: Rhodomonas salina. H.J. Modified from Jeong et al. 2015.
1.6! SUMMARY
Harmful algal blooms are increasing in frequency and magnitude in many parts of the world,
and one of the sectors of society that is being increasingly affected is the desalination
industry. Given trends in the development of that industry, as well as the global expansion of
the HAB problem, impacts will continue to occur, and are likely to increase. Desalination
plant operators and managers are urged to take a more active role in determining the nature of
the algal populations that are in the waters near their intakes, as this can directly help with
identification of timely and appropriate mitigation strategies. One of the many challenges
desalination plant managers face is that HABs are incredibly diverse in terms of toxicity, cell
size, morphology, and bloom dynamics, and this diversity needs to be recognized when
developing and implementing monitoring and mitigation strategies. Partnerships between
regional HAB scientists and desalination operators and managers are encouraged as these
will help the managers understand the nature of the problems they are facing.
41
1.7! REFERENCES
Al-Ansi, M. A., Abdel-Moati, M. A. R., Al-Ansari, I. S., 2002. Causes of fish mortality along
the Qatari waters (Arabian Gulf). International Journal of Environmental Studies. 59(1),
59–71.
Al-Azri, A., Piontkovski, S., Al-Hashmi, K., Al-Gheilani, H., Al-Habsi, H., Al-Khusaibi, S.,
and Al-Azri, N. 2012. The occurrence of algal blooms in Omani coastal waters. Aquatic
Ecosystem Health and Management 15 (supplement 1), 56-63.
Al-Busaidi, S. S., Al-Rashdi, K. M., Al-Gheilani, H. M., and Amer, S. 2008. Hydrographical
observations during a red tide with fish mortalities at Masirah Island, Oman. .
Agricultural and Marine Sciences 13, 63–72.
Al-Gheilani, H. M., Matsuoka, K., Al-Kindi, A. Y., Amer, S., and Waring, C. 2011. Fish kill
incidents and harmful algal blooms in Omani waters. Sultan Qaboos University
Research Journal-Agricultural and Marine Sciences 16, 23-33.
Al Shehhi, M. R., Gherboudj, I. and Ghedira, H. 2014. An overview of historical harmful
algae blooms outbreaks in the Arabian Seas. Marine Pollution Bulletin 86(1), 314-324.
Al-Yamani, F., Al-Ghunaim, D. V., Subba R. N., Khan, M., Al-Ghool, M., Muruppel, S., and
Al-Qatma, L. M. 2000. Fish Kills, Red Tides, and Kuwait’s Marine Environment.
Kuwait Institute for Scientific Research, Kuwait. 202 pp.
Álvarez, G., Uribe, E., Quijano-Scheggia, S., López-Rivera, A., Mariño, C., and Blanco, J.
2009. Domoic acid production by Pseudo-nitzschia australis and Pseudo-nitzschia
calliantha isolated from North Chile. Harmful Algae 8(6), 938-945.
Anderson, D. M. 1989. Toxic algal blooms and red tides: a global perspective. In: T. Okaichi,
D. M. Anderson, and T. Nemoto (Eds.), Red Tides: Biology, Environmental Science and
Toxicology. pp. 11-16. Elsevier.
Anderson, D. M. 1998. Physiology and bloom dynamics of toxic Alexandrium species, with
emphasis on life cycle transitions. NATO ASI Series G Ecological Sciences 41, 29-48.
Anderson, D. M., Glibert, P. M., and Burkholder, J. M. 2002. Harmful algal blooms and
eutrophication: nutrient sources, composition, and consequences. Estuaries 25(4), 704-
726.
Anderson, D. M., Kulis, D. M., Doucette, G. J., Gallagher, J. C., and Balech, E. 1994.
Biogeography of toxic dinoflagellates in the genus Alexandrium from the northeastern
United States and Canada. Marine Biology 120(3), 467-478.
Anderson, D. M., Keafer, B. A., Kleindinst, J. L., McGillicuddy Jr, D. J., Martin, J. L.,
Norton, K., Pilskaln, C. H., Smith, J. L., Sherwood, C. R., and Butman, B. 2014.
Alexandrium fundyense cysts in the Gulf of Maine: Long-term time series of abundance
and distribution, and linkages to past and future blooms. Deep-Sea Research Part II 103,
6-26.
Armstrong, M. and Kudela, R. 2006. Evaluation of California isolates of Lingulodinium
polyedrum for the production of yessotoxin. African Journal of Marine Science 28 (2),
399–401.
42
Band-Schmidt, C., Bustillos-Guzmán, J., Morquecho, L., Gárate-Lizárraga, I., Alonso

Rodríguez, R., Reyes-Salinas, A., and Luckas, B. 2006. Variations of PSP toxin profiles
during different growth phases in Gymnodinium catenatum (Dinophyceae) strains
isolated from three locations in the Gulf of California, Mexico. Journal of Phycology
42(4), 757-768.
Basti, L., Uchida, H., Matsushima, R., Watanabe, R., Suzuki, T., Yamatogi, T., and Nagai, S.
2015. Influence of temperature on growth and production of Pectenotoxin-2 by a
monoclonal culture of Dinophysis caudata. Marine Drugs 13(12), 7124-7137.
Bates, S. S., Bird, C. J., Defreitas, A. S. W., Foxall, R., Gilgan, M., Hanic, L. A., Johnson, G.
R., McCulloch, A. W., Dodense, P., Pocklington, R., Quilliam, M. A., Sim, P. G., Smith,
J. C., Rao, D. V. S., Todd, C. D., Walter, J. A., and Wright, J. L. C. 1989. Pennate
diatom Nitzschia pungens as the primary source of domoic acid, a toxin in shellfish from
eastern Prince Edwards Island, Canada. Canadian Journal of Fisheries and Aquatic
Sciences 46(7), 1203–1215.
Bates, S. S., Hiltz, M. F., and Leger, C. 1999. Domoic acid toxicity of large new cells of
Pseudo-nitzschia multiseries resulting from sexual reproduction. Canadian translations
of fisheries and aquatic sciences/Traductions canadiennes en sciences halieutiques et
aquatiques. Ottawa ON, (2261), 21-26.
Baugh, K. A., Bush, J. M., Bill, B. D., Lefebvre, K. A., and Trainer, V. L. 2006. Estimates of
specific toxicity in several Pseudo-nitzschia species from the Washington coast, based
on culture and field studies. African Journal of Marine Science 28(2), 403–407.
Bill, B. D. 2011. Carbon and nitrogen uptake of toxigenic diatoms: Pseudo-nitzschia
australis and Pseudo-nitzschia turgidula. Ph.D. thesis, San Francisco State University,
San Francisco CA USA.
Buck, K. R., Uttal-Cooke, L., Pilskaln, C. H., Roelke, D. L., Villac, M. C., Fryxell, G. A.,
Cifuentes, L., and Chavez, F. P. 1992. Autecology of the diatom Pseudo-nitzschia
australis, a domoic acid producer, from Monterey Bay, California. Marine Ecology
Progress Series 84(3), 293–302.
Caron, D. A., Garneau, M. È., Seubert, E., Howard, M. D., Darjany, L., Schnetzer, A.,
Cetinić, I., Filteau, G., Lauri, P., Jones, B. and Trussell, S. 2010. Harmful algae and their
potential impacts on desalination operations off southern California. Water Research
44(2), 385-416.
Cerino, F., Orsini, L., Sarno, D., Dell’Aversano, C., Tartaglione, L., and Zingone, A. 2005.
The alternation of different morphotypes in the seasonal cycle of the toxic diatom
Pseudo-nitzschia galaxiae. Harmful Algae 4(1), 33-48.
Chaghtai, F. and Saifullah, S. M. 2001. Harmful algal bloom (HAB) organisms of the north
Arabian sea bordering Pakistan. Pakistan Journal of Botany 33, 69–75.
Chaghtai, F. and Saifullah, S. 2006. On the occurrence of green Noctiluca scintillans blooms
in coastal waters of Pakistan, North Arabian Sea. Pakistan Journal of Botany 38, 893-
898.
Chang, F. H., Anderson, D. M., Kulis, D. M., and Till, D. G. 1997. Toxin production of
Alexandrium minutum (Dinophyceae) from the Bay of Plenty, New Zealand. Toxicon
35(3), 393-409.
43
Corcoran, A. A., Richardson, B., and Flewelling, L. J. 2014. Effects of nutrient-limiting

supply ratios on toxin content of Karenia brevis grown in continuous culture. Harmful
Algae 39, 334-341.
D’Silva, M. S., Anil, A. C., Naik, R. K., D’Costa, P. M. 2012. Algal blooms: a perspective
from the coasts of India. Natural Hazards 63, 1225–1253.
Dale, B. 1983. Dinoflagellate resting cysts: “benthic plankton”. Survival strategies of the
algae. 69-136.
Danovaro, R., Umani, S. F. and Pusceddu, A. 2009. Climate change and the potential
spreading of marine mucilage and microbial pathogens in the Mediterranean Sea. PLoS
One 4(9), e7006.
Dodge, J. D. 1973. The Fine Structure of Algal Cells. Academic Press, London. pp. 21-50.
Draisci, R., Ferretti, E., Palleschi, L., Marchiafava, C., Poletti, R., Milandri, A., Ceredi, A.,
and Pompei, M., 1999. High levels of yessotoxin in mussels and presence of yessotoxin
and homoyessotoxin in dinoflagellates of the Adriatic Sea. Toxicon 37, 1187–1193.
Fatemi, S. M. R, Nabavis, M. B., Vosoghi, G., Fallahi, M., and Mohammadi, M. 2012. The
relation between environmental parameters of Hormuzgan coastline in Persian Gulf and
occurrence of the first harmful algal bloom of Cochlodinium polykrikoides
(Gymnodiniaceae). Iranian Journal of Fisheries Sciences 11(3), 475-489.
Foster, K. A., Foster, G., Tourenq, C., and Shuriqi, M. K. 2011. Shifts in coral community
structures following cyclone and red tide disturbances within the Gulf of Oman (United
Arab Emirates). Marine Biology 158, 955–968.
Franks, P. J. S. 1992. Sink or swim: Accumulation of biomass at fronts. Marine Ecology
Progress Series. 82(1), 1-12.
Franks, P. J. S. and Anderson, D. M. 1992. Alongshore transport of a toxic phytoplankton
bloom in a buoyancy current: Alexandrium tamarense in the Gulf of Maine. Marine
Biology 112, 153-164.
Fu, F. X., Place, A. R., Garcia, N. S., and Hutchins, D. A. 2010. CO2 and phosphate
availability control the toxicity of the harmful bloom dinoflagellate Karlodinium
veneficum. Aquatic Microbial Ecology 59(1), 55-65.
Gaines, G., and Taylor, F. J. R., 1985. An exploratory analysis of PSP patterns in British
Columbia: 1942–1984. In: Anderson, D.M., White, A.W., Baden, D.G. (Eds.), Toxic
dinoflagellates. Elsevier Science, New York, pp. 439–444.
Garrison, D.L., Conrad, S.M., Eilers, P.P., and Waldron, E.M. 1992. Confirmation domoic
acid production by Pseudonitzchia australis (Bacillariophyceae) cultures. Journal of
Phycology 28(5): 604-607.
Geissler, U. 1958. Das Membranpotential einiger Diatomeen und seine Bedeutung fuer die
lebende Kieselalganzelle. Mikroskopie 13, 145-172.
GEOHAB. 2005. Global Ecology and Oceanography of Harmful Algal Blooms, GEOHAB
Core Research Project: HABs in Upwelling Systems. G. Pitcher, T. Moita, V. Trainer, R.
Kudela, P. Figueiras, T. Probyn (Eds.) IOC and SCOR, Paris and Baltimore. 82 pp.
GEOHAB. 2006. Oceanography of Harmful Algal Blooms, Harmful Algal Blooms in
Eutrophic Systems. P. Glibert (Ed.) IOC and SCOR, Paris and Baltimore. 74 pp.
44
Glibert, P. M., Landsberg, J. H., Evans, J. J., Al-Sawari, M. A., Faraj, M., Al-Jarallah, M. A.,
Haywood, A., Ibrahem, S., Klesius, P., Powell, K., and Shoemaker, C. 2002. A fish kill
of massive proportion in Kuwait Bay, Arabian Gulf, 2001: the roles of bacterial disease,
harmful algae, and eutrophication. Harmful Algae 1, 215–231.
Gribble, K. E., Keafer, B. A, Quilliam, M. A., Cembella, A. D., Kulis, D. M., Manahan, A.,
and Anderson, D. M. 2005. Distribution and toxicity of Alexandrium ostenfeldii
(Dinophyceae) in the Gulf of Maine, USA. Deep Sea Research Part II 52(19), 2745-
2763.
Gregorio, D. E. and Connell, L. 2000. Range of Heterosigma akashiwo (Raphidophyceae)
expanded to include California, USA. In: Ninth International Conference on Harmful
Algal Blooms, Tasmania, Australia (abstract).
Hallegraeff, G. M. 1993. A review of harmful algal blooms and their apparent global increase.
Phycologia 32(2), 79-99.
Harrison, P. J., Furuya, K., Glibert, P. M., Xu, J., Liu, H. B., Yin, K., Lee, J. H. W.,
Anderson, D. M., Gowen, R., Al-Azri, A. R. and Ho, A. Y. T. 2011. Geographical
distribution of red and green Noctiluca scintillans. Chinese Journal of Oceanology and
Limnology 29(4), 807-831.
Hawser, S. P., Codd, G. A., Capone, D. G., and Carpenter, E. J. 1991. A neurotoxic factor
associated with the bloom-forming cyanobacterium Trichodesmium. Toxicon 29(3), 277-
278.
Heil, C. A., Glibert, P. M., Al-Sarawl, M. A., Faraj, M., Behbehani, M., and Husain, M. 2001.
First record of a fish-killing Gymnodinium sp. bloom in Kuwait Bay, Arabian Sea:
chronology and potential causes. Marine Ecology Progress Series 214, 15–23.
Hershberger, P. K., Rensel, J. E., Postel, J. R., and Taub, F.B. 1997. Heterosigma bloom and
associated fish kill. Harmful Algae News 16(1), 4.
Holligan, P. M. 1979. Dinoflagellate blooms associated with tidal fronts around the British
Isles. In Toxic Dinoflagellate Blooms: Developments in Marine Biology. D. L. Taylor
and H. H. Seliger (Eds.) pp 249-256. Elsevier, New York.
Holmes, R. W., Williams, P. M., and Eppley, R. W. 1967. Red water in La Jolla, 1964–1966.
Limnology and Oceanography 12(3), 503–512.
Horner, R. A., Garrison, D. L., and Plumley, F. G. 1997. Harmful algal blooms and red tide
problems on the U.S. west coast. Limnology and Oceanography 42(2), 1076–1088.
Howard, M. D. A., Cochlan, W. P., Ladizinsky, N., and Kudela, R. M. 2007. Nitrogenous
preference of toxigenic Pseudo-nitzschia australis (Bacillariophyceae) from field and
laboratory experiments. Harmful Algae 6(2), 206–217.
Ives, K. J. 1956. Electrokinetic phenomena of planktonic algae. Proceedings of the Society
for Water Treatment and Examination 5, 41-58.
James, K. J., Sierra, M. D., Lehane, M., Brana, M. A, and Furey, A. 2003. Detection of five
new hydroxyl analogues of azaspiracids in shellfish using multiple tandem mass
spectrometry. Toxicon 41(3), 277–283.
Jauffrais, T., Kilcoyne, J., Sechet, V., Herrenknecht, C., Truquet, P., Herve, F., Berard, J. B.,
Nulty, C., Taylor, S., Tillman, U., Miles, C. O., and Hess, P. 2012. Production and
isolation of Azaspiracid-1 and -2 from Azadinium spinosum culture in pilot scale
photobioreactors. Marine Drugs 10(12), 1360–1382.
45
Jeong, H. J., Lim, A. S., Franks, P. J., Lee, K. H., Kim, J. H., Kang, N. S., Lee, M. J., Jang, S.
H., Lee, S. Y., Yoon, E. Y., Park, J. Y., Yoo, Y. D., Seong, K. A., Kwon, J. E., and Jang,
T. Y. 2015. A hierarchy of conceptual models of red-tide generation: nutrition, behavior,
and biological interactions. Harmful Algae 47, 97-115.
Jester, R. 2008. An investigation into the prevalence of Alexandrium derived toxins in marine
food webs. Ph.D., University of California, Santa Cruz, Santa Cruz, 116 pp.
John, U., Litaker, R. W., Montresor, M., Murray, S., Brosnahan, M. L., and Anderson, D. M.
2014. Formal revision of the Alexandrium tamarense species complex (Dinophyceae)
taxonomy: the introduction of five species with emphasis on molecular-based (rDNA)
classification. Protist 165, 779-804.
Joseph, T., Shaiju, P., Laluraj, C. M., Balachandran, K. K., Nair, M., George, R., Nair, K. K.,
Sahayak, S., and Prabhakaran, M. P. 2008. Nutrient environment of red tide-infested
waters off south-west coast of India. Environmental Monitoring and Assessment 143,
355–361.
Kamykowski, D. 1995. Trajectories of autotrophic marine dinoflagellates. Journal of
Phycology 31(2), 200-208.
Kerbrat, A. S., Amzil, Z., Pawlowiez, R., Golubic, S., Sibat, M., Darius, H. T., Chinain, M.
and Laurent, D. 2011. First evidence of palytoxin and 42-hydroxy-palytoxin in the
marine cyanobacterium Trichodesmium. Marine Drugs 9(4), 543-560.
Kerbrat, A. S., Darius, H. T., Pauillac, S., Chinain, M., and Laurent, D. 2010. Detection of
ciguatoxin-like and paralysing toxins in Trichodesmium spp. from New Caledonia
lagoon. Marine Pollution Bulletin 61(7), 360-366.
Kim, C. H., Sako, Y., and Ishida, Y. 1993. Comparison of toxin composition between
populations of Alexandrium spp. from geographically distant areas. Nippon Suisan
Gakkaishi 59, 641–646.
Krishnan, A. A., Krishnakumar, P. K., and Rajagopalan, M. 2007. Trichodesmium
erythraeum (Ehrenberg) bloom along the southwest coast of India (Arabian Sea) and its
impact on trace metal concentrations in seawater. Estuarine, Coastal and Shelf Science
71, 641–646.
Krock, B., Tillmann, U., Potvin, E., Jeong, H. J., Drebing, W., Kilcoyne, J., Al-Jorani, A.,
Twiner, M. J, Göthel, Q., and Köck, M. 2015. Structure elucidation and in vitro toxicity
of new Azaspiracids isolated from the marine dinoflagellate Azadinium poporum.
Marine Drugs 13(11), 6687-6702.
Kudela, R. M. and Gobler, C. J. 2012. Harmful dinoflagellate blooms caused by
Cochlodinium sp.: Global expansion and ecological strategies facilitating bloom
formation. Harmful Algae 14, 71-86.
Kudela, R., Ryan, J., Blakely, M., Lane, J., Peterson, T. 2008. Linking the physiology and
ecology of Cochlodinium to better understand harmful algal bloom events: a
comparative approach. Harmful Algae 7, 278–292.
Loeblich III, A. R. and Fine, K. E. 1977. Marine chloromonads: More widely distributed in
neritic environments than previously thought. Proceedings of the Biological Society of
Washington 90(2), 388–399.
46
Lundholm, N., Skov, J., Pocklington, R., and Moestrup, O. 1994. Domoic acid, the toxic
amino acid responsible for amnesic shellfish poisoning, now in Pseudonitzschia seriata
(Bacillariophyceae) in Europe. Phycologia 33(6), 475-478.
Lundholm, N., Skov, J., Pocklington, R., and Moestrup, O. 1997. Studies on the marine
planktonic diatom Pseudo-nitzschia. 2. Autecology of P. pseudodelicatissima based on
isolates from Danish coastal waters. Phycologia 36, 381–388.
Mackenzie, L., White, D., Oshima, Y., and Kapa, J. 1996. The resting cyst and toxicity of
Alexandrium ostenfeldii (Dinophyceae) in New Zealand. Phycologia 35(2), 148-155.
MacKenzie, L., Sims, I., Beuzenberg, V. and Gillespie, P. 2002. Mass accumulation of
mucilage caused by dinoflagellate polysaccharide exudates in Tasman Bay, New
Zealand. Harmful Algae, 1(1), 69-83.
Madhupratap, M., Sawant, S., and Gauns, M. 2000. A first report on a bloom of the marine
prymnesiophycean, Phaeocystis globosa from the Arabian Sea. Oceanologica Acta
23(1), 83–90.
Margalef, R. 1961. Hidrografia y fitoplancton de un area marina de la costa meridional de
Puerto Rico. Investigación pesquera 18, 76–78.
Martin, J. L., Haya, K., Burridge, L. E., and Wildish, D. J. 1990. Nitzschia
pseudodelicatissima - a source of domoic acid in the Bay of Fundy, eastern Canada.
Marine Ecology Progress Series 67(2), 177–182.
Maruyama, T., Yamada, R., Usui, K., Suzuki, H. and Yoshida, T. 1987. Removal of marine
red tide plankton with acid-treated clay. Nippon Suisan Gakkaishi, 53(10), 1811-1819.
Matsuoka, K., Iwataki, M., and Kawami, H. 2008. Morphology and taxonomy of
chainforming species of the genus Cochlodinium (Dinophyceae). Harmful Algae 7, 261–
270.
McCarron, P., Kilcoyne, J., Miles, C. O., and Hess, P. 2009. Formation of Azaspiracids-3, -4,
-6, and -9 via decarboxylation of carboxyazaspiracid metabolites from shellfish. Journal
of Agricultural and Food Chemistry 57(1), 160–169.
McMinn, A., Hallegraeff, G. M., Thomson, P., Jenkinson, A., and Heijnis, H. 1997. Cyst and
radionucleotide evidence for the recent introduction of the toxic dinoflagellate
Gymnodinium catenatum into Tasmanian waters. Marine Ecology Progress Series 161,
165-172.
Mitch, W. A. and Sedlak, D. L. 2002. Formation of N-nitrosodimethylamine (NDMA) from
dimethylamine during chlorination. Environmental Science and Technology 36(4), 588-
595.
Moestrup, O., Akselmann, R., Fraga, S., Hansen, G., Hoppenrath, M., Iwataki, M., Komarek,
J., Larsen, J., Lundholm, N., and Zingone, A. (Eds.) 2017. IOC-UNESCO Taxonomic
Reference List of Harmful Micro Algae.
Morton, S. L., Faust, M. A., Fairey, E. I., and Moeller, P. D. R. 2002. Morphology and
toxicology of Prorocentrum arabianum sp. nov., (Dinophyceae) a toxic planktonic
dinoflagellate from the Gulf of Oman, Arabian Sea. Harmful Algae 1(4), 393–400.
Moschandreou, K. K., Papaefthimiou, D., Katikou, P., Kalopesa, E., Panou, A., and
Nikolaidis, G. 2010. Morphology, phylogeny and toxin analysis of Pseudo-nitzschia
pseudodelicatissima (Bacillariophyceae) isolated from the Thermaikos Gulf, Greece.
Phycologia 49(3), 260-273.
47
Nagai, S., Hori, Y., Manabe, T., and Imai, I. 1995. Morphology and rejuvenation of
Coscinodiscus wailesii Gran (Bacillariophyceae) resting cells found in bottom sediments
of Harima-Nada, Seto Inland Sea. Bulletin of the Japanese Society of Scientific Fisheries
61, 179-185.
Nagai, S., Suzuki, T., and Kamiyama, T. 2013. Successful cultivation of the toxic
dinoflagellate Dinophysis tripos (Dinophyceae). Plankton and Benthos Research 8(4),
171-177.
Nagai, S., Suzuki, T., Nishikawa, T., and Kamiyama, T. 2011. Differences in the production
and excretion kinetics of okadaic acid, dinophysistoxin-1, and pectenotoxin-2 between
cultures of Dinophysis acuminata and Dinophysis fortii isolated from western Japan.
Journal of Phycology 47(6), 1326-1337.
Nielsen, L. T., Krock, B., and Hansen, P. J. 2013. Production and excretion of okadaic acid,
pectenotoxin-2 and a novel dinophysistoxin from the DSP-causing marine dinoflagellate
Dinophysis acuta–effects of light, food availability and growth phase. Harmful Algae 23,
34-45.
Ofuji, K., Satake, M., McMahon, T., James, K. J., Naoki, H., Oshima, Y., and Yasumoto, T.
2014. Structures of azaspiracid analogs, azaspiracid-4 and azaspiracid-5, causative
toxins of azaspiracid poisoning in Europe. Bioscience, Biotechnology and Biochemistry
65(3), 740–742.
O’Halloran, C., Silver, M., Holman, T., and Scholin, C. 2006. Heterosigma akashiwo in
central California waters. Harmful Algae 5 (2), 124–132.
Okaichi, T. 2004. Red tide phenomena. In: Okaichi, T. (Ed.), Red Tides, Kluwer Academic
Publishers, London, pp. 7-60.
Orsini, L., Sarno, D., Procaccini, G., Poletti, R., Dahlmann, J., and Montresor, M. 2002.
Toxic Pseudo-nitzschia multistriata (Bacillariophyceae) from the Gulf of Naples:
morphology, toxin analysis and phylogenetic relationships with other Pseudo-nitzschia
species. European Journal of Phycology 37(2), 247-257.
Padmakumar, K. B., Menon, N. R., and Sanjeevan, V. N. 2012. Is occurrence of harmful
algal blooms in the exclusive economic zone of India on the rise? International Journal
of Oceanography.
Park, J., Jeong, M., Lee, J., Cho, K. J, and Kwon, O. S. 2001. Diurnal vertical migration of a
harmful dinoflagellate, Cochlodinium polykrikoides (Dinophyceae) during a red tide in
coastal waters of Namhae Island, Korea. Phycologia 40, 292–297.
Popels, L. C., and Hutchins, D. A. 2002. Factors affecting dark survival of the brown tide
alga Aureococcus anophagefferens (Pelagophyceae). Journal of Phycology 38(4), 738-
744.
Prud'homme van Reine, W. F. 2017. Report of the Nomenclature Committee for Algae: 16 –
On proposals to amend the Code. Taxon 6 (1): 197-8.
Rabbani, M. M., Rehman, A. U., and Harms, C. E. 1990. Mass mortality of fishes caused by
dinoflagellate bloom in Gwadar Bay, southwestern Pakistan. In: Graneli, E., Sundström,
B., Edler, L., and Anderson, D. M. (Eds.), Toxic Marine Phytoplankton. Elsevier
Science Publishing, New York, US, pp. 209–2014.
48
Raine, R., McDermott, G., Silke, J., Lyons, K., Nolan, G. and Cusack, C. 2010. A simple
short range model for the prediction of harmful algal events in the bays of southwestern
Ireland. Journal of Marine Systems 83(3), 150-157.
Rhodes, L., McNabb, P., de Salas, M., Briggs, L., Beuzenberg, V., and Gladstone, M. 2006.
Yessotoxin production by Gonyaulax spinifera. Harmful Algae 5 (2), 148–155.
Rhodes, L., Scholin, C., and Garthwaite, I. 1998. Pseudo-nitzschia in New Zealand and the
role of DNA probes and immunoassays in refining marine biotoxin monitoring
programmes. Natural Toxins 6(3-4), 105-111.
Rhodes, L., White, D., Syhre, M., and Atkinson, M. 1996. Pseudo-nitzschia species isolated
from New Zealand coastal waters: domoic acid production in vitro and links with
shellfish toxicity. T. Yasumoto, Y. Oshima, Y. Fukuyo (Eds.), Harmful and Toxic Algal
Blooms, Intergov. Oceanogr. Comm. UNESCO, Paris, pp. 155–158.
Ricci, F., Penna, N., Capellacci, S., and Penna, A. 2014. Potential environmental factors
influencing mucilage formation in the northern Adriatic Sea. Chemical Ecology 30, 364–
375.
Richlen, M. L., Morton, S. L., Jamali, E. A., Rajan, A., and Anderson, D. M. 2010. The
catastrophic 2008–2009 red tide in the Arabian gulf region, with observations on the
identification and phylogeny of the fish-killing dinoflagellate Cochlodinium
polykrikoides. Harmful Algae 9, 163–172.
Rodríguez, L. P., González, V., Martínez, A., Paz, B., Lago, J., Cordeiro, V., Blanco, L.,
Vieites, J. M. and Cabado, A. G. 2015. Occurrence of lipophilic marine toxins in
shellfish from Galicia (NW of Spain) and synergies among them. Marine Drugs, 13(4),
1666-1687.
Saeedi, H., Kamrani, E., and Matsuoka, K. 2011. Catastrophic impact of red tides of
Cochlodinium polykrikoides on the razor clam Solen dactylus in coastal waters of the
Northern Persian Gulf. Journal of the Persian Gulf 2(6), 13–20.
Sahraoui, I., Bates, S. S., Bouchouicha, D., Hadj Mabrouk, H., and Sakka Hlaili, A. 2011.
Toxicity of Pseudo-nitzschia populations in Bizerte Lagoon, Tunisia, southwest
Mediterranean, and first report of domoic acid production by P.brasiliana. Diatom
Research 26, 293–303.
Saifullah, S. M. 1979. Occurrence of dinoflagellates and distribution of chlorophyll A on
Pakistan shelf. In: Taylor, D. L. and Seliger, H. H. (Eds.), Toxic Dinoflagellate Blooms.
Elsevier, North Holland, New York, pp. 203–208.
Sampedro, N., Arin, L., Quijano, S., Rene, A., and Camp, J. 2007. Mucilage event associated
with Gonyaulax fragilis in NW Mediterranean Sea. Harmful Algae News 33, 10-11.
Satake, M., Ofuji, K., Naoki, H., James, K. J., Furey, A., McMachon, T., Silke, J., and
Yasumoto, T. 1998. Azaspiracid, a new marine toxin having unique spiro ring
assemblies, isolated from Irish mussels, Mytilus edulis. Journal of the American
Chemical Society 12(38), 9967–9968.
Schoemann, V., Bequevort, S., Stefels, J., Rousseau, V., and Lancelot, C. 2005. Phaeocystis
blooms in the global ocean and their controlling mechanisms: a review. Journal of Sea
Research 53(1), 43-66.
49
Scholin, C. A., Herzog, M., Sogin, M., Anderson, D. M., 1994. Identification of group-and
strain-specific genetic markers for globally distributed Alexandrium (Dinophyceae). Part
II. Sequence analysis of a fragment of the LSU rRNA gene. Journal of Phycology 30(6),
999–1011.
Sengco, M. R. 2001. The aggregation of clay minerals and marine microalgal cells:
physicochemical theory and implications for controlling harmful algal blooms. Ph.D.
Dissertation, Massachusetts Institute of Technology/Woods Hole Oceanographic
Institution Joint Program in Oceanography, 237 pp.
Shahid, K., and Al Sadi, A. 2015. Algae: a tide of trouble. Desalination and Water Reuse,
May-June, 35-37.
Sheppard, C., Al-Husiani, M., Al-Jamali, F., Al-Yamani, F., Baldwin, R., Bishop, J., Benzoni,
F., Dutrieux, E., Dulvy, N. K., Durvasula, S. R. V. and Jones, D. A. 2010. The Gulf: a
young sea in decline. Marine Pollution Bulletin 60(1), 13-38.
Shirota, A. 1989. Red tide problem and countermeasures (2). International Journal of
Fisheries and Aquatic Studies, 195-223.
Smetacek, V. 2001. A watery arms race. Nature 411(6839), 745-745.
Smetacek, V. and Zingone, A. 2013. Green and golden seaweed tides on the rise. Nature,
504(7478), 84.
Sommer, H. and Meyer, K.F. 1937. Paralytic shellfish poisoning. Archives of Pathology 24,
560–598.
Steidinger, K. A. 1993. Some taxonomic and biologic aspects of toxic dinoflagellates. In:
Falconer, I. (Ed.), Algal Toxins in Seafood and Drinking Water. Academic Press,
London, pp. 1–28.
Subba Rao, D. V., Quilliam, M. A., and Pocklington, R. 1988. Domoic acid – a neurotoxic
amino acid produced by the marine diatom Nitzschia pungens in culture. Canadian
Journal of Fisheries and Aquatic Sciences 45(12), 2076–2079.
Suzuki, T., Watanabe, R., Uchida, H., Matsushima, R., Nagai, H., Yasumoto, T., Yoshimatsu,
T., Sato, S. and Adachi, M. 2012. LC-MS/MS analysis of novel ovatoxin isomers in
several Ostreopsis strains collected in Japan. Harmful Algae 20, 81-91.
Tang, D. L., Kawamura, H., and Luis, A. J. 2002. Short-term variability of phytoplankton
blooms associated with a cold eddy in the northwestern Arabian Sea. Remote Sensing of
Environment 81, 82–89.
Tang, Y. Z. and Gobler, C. J. 2009. Characterization of the toxicity of Cochlodinium
polykrikoides isolates from Northeast US estuaries to finfish and shellfish. Harmful
Algae 8(3), 454-462.
Tang, Y. Z. and Gobler, C. J. 2012. The toxic dinoflagellate Cochlodinium polykrikoides
(Dinophyceae) produces resting cysts. Harmful Algae 20, 71-80.
Taylor, F. J. R. and Horner, R. A. 1994. Red tides and other problems with harmful algal
blooms in Pacific Northwest coastal waters. In: Review of the Marine Environment and
Biota of Strait of Georgia, Puget Sound and Juan de Fuca Strait. Canadian Fisheries and
Aquatic Science Technical Report 1948, pp. 175–186.
Tenny, M. W., Echelberger, W. F., Schuessler, R. G., and Pavoni, J. L. 1969. Algal
flocculation with synthetic organic polyelectrolytes. Applied Microbiology 18, 965-971.
50
Thangaraja, M., Al-Aisry, A., Al-Kharusi, L., 2007. Harmful algal blooms and their impacts
in the middle and outer ROPME sea area. International Journal of Oceans and
Oceanography 2, 85– 98.
Tillman, U., Elbrachter, M., John, U., and Krock, B. 2011. A new non-toxic species in the
dinoflagellate genus Azadinium: A. poporum sp. nov. European Journal of Phycology 46,
74–87.
Tillman, U., Elbrachter, M., John, U., Krock, B., and Cembella, A. D. 2010. Azadinium
obesum (Dinophyceae), a new nontoxic species in the genus that can produce azaspiracid
toxins. Phycologia 49, 169–182.
Tillmann, U., Elbrachter, M., Krock, B., John, U., and Cembella, A. D. 2009. Azadinium
spinosum gen. et sp. nov. (Dinophyceae) identified as a primary producer of azaspiracid
toxins. European Journal of Phycology 44, 63–79.
Tillman, U., Soehner, S., Nezan, E., and Krock, B. 2012. First record of the genus Azadinium
(Dinophyceae) from the Shetland Islands, including the description of Azadinium
polongum sp. nov. Harmful Algae 20, 142–155.
Trainer, V. L., Adams, N. G., Bill, B. D., Stehr, C. M., Wekell, J. C., Moeller, P., Busman,
M., and Woodruff, D. 2000. Domoic acid production near California coastal upwelling
zones, June 1998. Limnology and Oceanography 45(8), 1818–1833.
Trainer, V. L., Adams, N. G., and Wekell, J. C. 2001. Domoic acid producing Pseudo-
nitzschia species off the U.S. west coast associated with toxification events. In:
Hallegraeff, G. M., Blackburn, S. I., Bolch, C. J., and Lewis, R. J. (Eds.), Harmful Algal
Bloom 2000. Intergovernmental Oceanographic Commission of UNESCO, Paris, pp.
46–49.
Trainer, V. L., Hickey, B. M., Lessard, E. J., Cochlan, W. P., Trick, C. G., Wells, M. L.,
MacFadyen, A. and Moore, S. K. 2009. Variability of Pseudo‐nitzschia and domoic
acid in the Juan de Fuca eddy region and its adjacent shelves. Limnology and
Oceanography 54(1), 289-308.
Trick, C. G., Bill, B. D., Cochlan, W. P., Wells, M. L., Trainer, V. L., and Pickell, L. D. 2010.
Iron enrichment stimulates toxic diatom production in high-nitrate, low-chlorophyll
areas. Proceedings of the National Academy of Sciences 107(13), 5887-5892.
Tyrell, J. V., Connell, L. B., and Scholin, C. A. 2002. Monitoring for Heterosigma akashiwo
using a sandwich hybridization assay. Harmful Algae 1(2), 205–214.
Usup, G., Ahmad, A., Matsuoka, K., Lim, P. T, and Leaw, C. P. 2012. Biology, ecology and
bloom dynamics of the toxic marine dinoflagellate Pyrodinium bahamense. Harmful
Algae 14, 301-312.
Usup, G., Kulis, D. M., and Anderson, D. M. 1994. Growth and toxin production of the toxic
dinoflagellate Pyrodinium bahamense var. compressum in laboratory cultures. Natural
Toxins 2(5), 254-262.
Virgili, F. 2015. GWI Q4 desalination market review and forecast points to some
improvement in contracted capacity. pp. 12, 13: IDA News Nov./Dec. 2015.
International Desalination Association.
51
Walsh, J. J., Joliff, J. K., Darrow, B. P., Lenes, J. M., Milroy, S. P., Dieterie, D. A., Chen, F.
R., Vargo, G. A., Weisburg, H. R., Fanning, K. A., Muller-Karger, F. E., Whitledge, T.
E., Stockwell, D. A., Tomas, C. R., Villareal, T. A., Jochens, A. E. 2006. Red tides in
the Gulf of Mexico: where, when and why? Journal of Geophysical Research C11003, 1-
46.
Work, T. M., Barr, B., Beale, A. M., Fritz, L., Quilliam, M. A., and Wright, J. L. 1993.
Epidemiology of domoic acid poisoning in brown pelicans (Pelecanus occidentalis) and
Brandt's cormorants (Phalacrocorax penicillatus) in California. Journal of Zoo and
Wildlife Medicine, 54-62.
Wright, J. L. C., Boyd, R. K., De Freitas, A. S. W., Falk, M., Foxall, R. A., Jamieson, W. D.,
Laycock, M. V., McCulloch, A. W., McInnes, A. G., Odense, P., Pathak, V. P., Quilliam,
M. A., Ragan, M. A., Sim, P. G., Thibault, P., Walter, J. A., Gilgan, M., Richard, D. J.
A., and Dewar, D. 1989. Identification of domoic acid, a neuroexcitatory amino acid, in
toxic mussels from eastern Prince Edward Island. Canadian Journal of Chemistry 67,
481-490.
Yu, Z., Zou, J., and Ma, X. 1994. Applications of clays to removal of red tide organisms I.
Coagulation of red tide organisms with clays. Chinese Journal of Oceanology and
Limnology 12(3), 193-200.
Zhao, J. and Ghedira, H. 2014. Monitoring red tide with satellite imagery and numerical
models: a case study in the Arabian Gulf. Marine Pollution Bulletin 79(1), 305-313.
Zhu, I. X., Bates, B. J. and Anderson, D. M. 2014. Removal of Prorocentrum minimum from
seawater using dissolved air flotation. Journal of Applied Water Engineering and
Research 2(1), 47-56.
52
Algal issues in seawater desalination
2 ALGAL ISSUES IN SEAWATER DESALINATION
Philipp Hess1, Loreen O. Villacorte2,3, Mike B. Dixon4, Siobhan F.E. Boerlage5,

Donald M. Anderson6, Maria D. Kennedy2, and Jan C. Schippers2
1
2
IHE-Delft Institute for Water Education, Delft, The Netherlands
3
4
5
6
Biology Department, Woods Hole Oceanographic Institution, Woods Hole, MA 02543 USA

2.1 Introduction ............................................................................................................................................. 53
2.2 Algal organic matter (AOM) and membrane fouling.............................................................................. 53
2.2.1 AOM release and composition ........................................................................................................ 54
2.2.2 AOM classification .......................................................................................................................... 54
2.2.3 Transparent Exopolymer Particles (TEP) ........................................................................................ 56
2.2.4 Fouling issues in SWRO plant during HAB events ........................................................................ 57
2.2.4.1 Biofouling in SWRO due to AOM deposition ........................................................................ 58
2.2.4.2 Particulate/Organic fouling of MF/UF during HABs .............................................................. 59
2.2.5 Summary and outlook...................................................................................................................... 62
2.3 Algal issues in thermal desalination plants ............................................................................................. 63
2.4 Marine and freshwater toxins .................................................................................................................. 63
2.4.1 Background...................................................................................................................................... 63
2.4.2 Chemistry and source organisms ..................................................................................................... 64
2.5 Taste and odor compounds ...................................................................................................................... 71
2.6 Detection techniques ............................................................................................................................... 71
2.6.1 Geosmin and methylisoborneol ....................................................................................................... 72
2.6.2 Cylindrospermopsin ........................................................................................................................ 73
2.6.3 Saxitoxins and tetrodotoxins ........................................................................................................... 73
2.6.4 Domoic acid..................................................................................................................................... 73
2.6.5 Microcystins and nodularin ............................................................................................................. 73
2.6.6 Azaspiracids, brevetoxins, ciguatoxins, cyclic imines, okadaic acid and dinophysistoxins ........... 75
2.7 Gaps and perspectives on analytical techniques ..................................................................................... 75
2.8 References ............................................................................................................................................... 76

2.1 INTRODUCTION
Once harmful algal blooms (HABs) reach a desalination plant, they can cause significant
operational issues and potential health concerns for consumers. These issues stem from two
factors – first, the algal cells produce organic matter that can cause filter clogging and
membrane fouling, and secondly, some cells produce toxic substances or taste and odor
compounds. This chapter first explains the mechanisms for cellular release of organic matter,
the types of matter that are produced, and the relative contribution of each type of matter to
fouling mechanisms. It then describes the wide range of toxins that are produced by HABs,
their mode of toxicity, and analytical methods for detecting them. While taste and odor
compounds are non-toxic, they are included in this chapter as they can create customer
perception issues and distrust in the water supply system.
2.2 ALGAL ORGANIC MATTER (AOM) AND MEMBRANE FOULING
Natural organic matter in seawater is comprised of a diverse mixture of particulate, colloidal

and dissolved organic substances which may originate from autochthonous (local) and
allochthonous (external) sources. Algal blooms are believed to be responsible for about half
of the autochthonous organic matter input to the earth’s oceans (Field et al. 1998). These
53
algal-derived substances are collectively known as algal (or algogenic) organic matter
(AOM). AOM can cause (directly or indirectly) operational problems in membrane-based
desalination plants, affecting both the pretreatment processes and reverse osmosis membrane
units. This section reviews the latest knowledge on the occurrence, composition, and
characteristics of AOM from the perspective of seawater desalination.
2.2.1 AOM release and composition
Algal blooms are often responsible for the highest annual pulses of natural organic matter in
seawater. A spike (>50% increase) in total organic carbon (TOC) concentration has been
recorded in coastal seawater during algal blooms (e.g., Petry et al. 2007); however, AOM
produced during algal blooms may vary substantially in terms of concentration, composition
and characteristics, depending on the causative species and environmental conditions. AOM
is generally released into seawater through metabolic excretion of active algal cells or
through autolysis of damaged or dying cells. Active cells may excrete AOM in response to
low-nutrient stress, unfavorable environmental conditions (e.g., light, pH and temperature) or
invasion by bacteria or viruses (Leppard 1993). Several species of algae may also release
AOM even under fairly favorable conditions (Fogg 1983).
Excessive production of AOM due to depletion of specific nutrients (e.g., phosphorus (P),
nitrogen (N) and silicon (Si)) and pathogen invasion have been linked to the occurrence of
marine mucilage (Mingazzini and Thake 1995). This phenomenon is characterized by the
appearance of a sporadic but massive accumulation of gelatinous material at and below the
water surface. Severe mucilage events occasionally occur in the North Sea, Adriatic Sea, and
other parts of the Mediterranean Sea, and undoubtedly elsewhere in the world as well, though
unreported. The proliferation of smaller mucilaginous aggregates known as “marine snow”
has been reported in most oceanic and marine systems.
Specific conditions at the seawater intake and through the pretreatment processes may also
induce further release of AOM. For instance, the addition of oxidizing or biocidal agents such
as chlorine has been shown to cause damage to algal cell walls and membranes, resulting in
the release of intracellular AOM (Daly et al. 2007). Moreover, exposure to hydrodynamic
shear stress (e.g., valves and pumps) may cause breakage of soft-walled algal cells, releasing
AOM which is normally stored inside the cells (Ladner et al. 2010; Voutchkov 2010).
The chemical composition of AOM usually includes proteins, polysaccharides, nucleic acids,
lipids and other dissolved organic substances (Fogg 1983). For some algae, particularly
diatoms such as Chaetoceros affinis, extracellular polysaccharides may comprise up to 90%
of AOM released (Myklestad 1995). A major fraction of AOM is very sticky and is thought
to cause operational issues in SWRO plants during blooms (Villacorte et al. 2015a). These
substances are widely known as transparent exopolymer particles or TEPs (see Section 2.3.3
and Appendix 3).
2.2.2 AOM classification
Based on the mechanism of release by algae, AOM can be classified into two major groups,
namely: (1) extracellular organic matter (EOM) - organic substances released through
metabolic activity of algae, and (2) intracellular organic matter (IOM) - substances released
through autolysis and/or during the process of cell decay. EOM substances can be either dis-
54
crete or attached (bound) to the algal cell as coatings. Discrete or free EOMs comprise
mainly polysaccharides and tend to be more hydrophilic while bound EOM comprise more
protein compounds and tend to be more hydrophobic (Qu et al. 2012; Henderson et al. 2008).
Figure 2.1. Graphical illustrations of how AOMs are released into seawater by algae at
different phases of a bloom and in response to the availability of essential nutrients.
On the other hand, IOMs comprise mainly low molecular weight polymers released from the
interior of damaged (e.g., broken cell wall), dying or decaying cells, which may include
toxins as well as taste and odor compounds (see Section 2.3). Considering the conditions of
how they are released, the contribution of IOM to the total AOM production is expected to
increase during the stationary-death phase of the bloom (Figure 2.1).
AOM components can be also classified in terms of their molecular weight or size. The low
molecular weight components of AOM include humic acid-like substances, nucleic acids,
lipids, toxins, taste/odor compounds and other organic acids (Figure 2.2). In principle, these
compounds also fall under the IOM classification because they are part of the intracellular
components of an algal cell. High molecular weight AOM comprises protein and
polysaccharide compounds, including TEP and their precursors. AOM typically cover a wide
size spectrum, ranging from less than 1 nm to more than 1 mm. Based on their size cut-off,
granular media filtration (GMF), microfiltration (MF), and ultrafiltration (UF) are expected to
remove only part of high molecular weight AOM (Figure 2.2). Nanofiltration (NF) is
expected to completely remove this fraction as well as part of the low molecular weight
AOM while complete removal of AOM can be achieved by reverse osmosis.
55
Figure 2.2. Classification of AOM components based on their molecular weight and size: (top chart) major
components of AOM which belong to the low and high molecular fraction; (bottom chart) the size spectrum of
major AOM components, in comparison with other suspended/dissolved matter in seawater and the size cut-off
of relevant filtration methods (Villacorte 2014).
2.2.3 Transparent Exopolymer Particles (TEP)

TEPs are organic substances usually associated with algal blooms in both fresh and marine
aquatic environments. These amorphous substances have been observed in various shapes
(e.g., strings, disks, sheets or fibers) and sizes, ranging from a few nanometers in diameter up
to hundreds of µm long as in the mucilage aggregates previously described (Passow 2000). In
the ocean, TEPs are mainly produced by phytoplankton (micro-algae) and bacterioplankton
but they may also originate from macro-algae, and shellfish (Passow 2002). Algae can
directly release TEPs through shedding of cell mucus/coatings (Figure 2.3) or through
disintegration of large algal colonies (Kiørboe and Hansen 1993).
By definition, the term ’TEP’ refers to substances (including their associated components)
stainable by Alcian Blue, a cationic dye, and are larger than 0.4 µm - they were originally
discovered through retention on 0.4 µm pore size membrane filters (Alldredge et al. 1993).
TEPs are not solid particles, but rather an agglomeration of particulate and colloidal
hydrogels.
Colloidal AOM (1-10 kDa) may agglomerate to form TEP, first by spontaneous assembly of
56
free fibrils through alignment on hydrophobic surfaces, then through annealing and gelation
to form micro-hydrogels, and eventually TEP (>0.4 µm) through aggregation (Verdugo et al.
2004). These sub-micron components (<0.4 µm) which have similar chemical properties as
TEP are collectively known as TEP precursors (Passow 2000).
Figure 2.3. TEP release by marine algae through shedding of cell mucus or membrane coatings. Optical
microscope images of Alcian Blue stained (a) Alexandrium tamarense, (b) Lepidodinium chlorophorum and
(c) Chaetoceros affinis. Photos: (a) and (c) Villacorte et al. (2015a); (b) Claquin et al. (2008).
TEPs and their precursors are generally sticky, highly hydrophilic and comprise mainly
negatively-charged polysaccharides and glycoprotein. Their stickiness and anionic charge are
mainly attributed to the presence of sulfate half ester (R-OSO3-) functional groups (Mopper et
al. 1995). In aquatic systems, they contain more than 99% water, which means they are
neutrally buoyant and can bulk-up to more than 100 times their solid volume (Azetsu-Scott
and Passow 2004). Moreover, they can be associated with or tend to absorb proteins, lipids,
dissolved trace elements and metals in the water (Passow 2002). This makes them a nutritious
platform and hotspot for bacterial activity (Berman and Holenberg 2005). Such
characteristics have led some researchers to suspect that they may have an important role in
the formation of aquatic biofilms. In 2005, Berman and Holenberg proposed the potential role
of TEP and their precursors as a major initiator of biofilm leading to biofouling in reverse
osmosis membranes (Berman and Holenberg 2005). Consequently, various studies were
conducted to investigate the link between these substances and biofouling. Moreover, it was
also demonstrated in lab- and pilot scale studies using seawater that they can cause severe
organic fouling in micro-/ultra-filtration membranes (e.g., Villacorte et al. 2015b; Ladner et
al. 2010; Kim and Yoon 2005).
Since the discovery of TEP in the early 90’s, a number of methods have been developed and
further modified to monitor TEP concentrations in the aquatic environment. The latest
development of these methods and their potential application in monitoring the impact of
HABs on SWRO plants and their removal are discussed in Chapter 5. Methods to measure
TEP and TEP precursors are given in Appendix 3.
2.2.4 Fouling issues in SWRO plant during HAB events
The high AOM concentration in the raw water during algal blooms can cause fouling issues
in both the pretreatment and RO systems of a SWRO desalination plant. RO membranes are
primarily designed to remove dissolved constituents in the water, in particular, inorganic ions
(dissolved salts). Membrane systems are vulnerable to fouling and clogging. Fouling occurs
due to deposition of particulates and/or growth of bacteria to form a biofilm on the membrane
surface, resulting in an increase in hydraulic resistance. Usually this resistance is
compensated by increasing the feed pressure and ultimately membranes are chemically
cleaned in place (see Appendix 5). Clogging of a membrane system is due to fouling of the
57
spacer of spiral wound SWRO elements or bundles of hollow fiber MF/UF membranes
resulting from particulate matter deposition and/or biomass formation. To avoid frequent
(chemical) cleaning, RO systems are generally preceded by a pretreatment process to
minimize the particulate, organic and biofouling potential of the feed water. Conventional or
advanced pretreatment techniques can be applied in seawater SWRO plants. Conventional
techniques comprise various types of granular media filters. Frequently, ‘in line coagulation’
(addition of a coagulant in front of the filters) is applied to improve the hydraulic
performance of the filter as well removal of smaller particles and organic matter. In some
designs, these filters are preceded by coagulation/sedimentation or coagulation/flotation
systems. Advanced pretreatment techniques include the application of micro- or ultra-
filtration membranes. Dissolved air flotation (DAF) is often employed as an additional
pretreatment step before GMF and/or MF/UF in seawater RO systems with the aim of
making the pretreatment more reliable and robust.
Typically, algae and AOM are (partially) captured by the pretreatment process before the
seawater is fed to the SWRO system. This means the pretreatment system will be the first to
be exposed to AOM fouling. Clogging of granular media filters (GMF – the most widely
used SWRO pretreatment technique) was one of the identified causes of SWRO plant
shutdown during the severe HAB in the Gulf of Oman in 2008-2009 (Richlen et al. 2010;
Berktay 2011) due to severely reduced operation times from 24 hours to 2 hours. GMF
typically accumulate materials larger than 10 µm (Ripperger et al. 2012) which may include
algal cells and large AOM (Figure 2.2). Considering the relatively high filtration rate (5-
10 m/h) in GMF, a sudden spike in algae and AOM concentration will rapidly clog the
interstices of the filter media and eventually form a slimy cake layer on the surface of the
filter bed. In this scenario, the filtrate production of the GMF will rapidly decline, which
would then require frequent backwashing and coagulant addition and hence, longer system
downtime. In addition a significant part of the AOM can pass through these filters, resulting
in pretreated water with high fouling potential (e.g., high silt density index) which is
unacceptable for RO operation.
Several SWRO plants have MF/UF pretreatment systems that can remove algae and much
smaller AOM than GMF (see Figure 2.2). During MF/UF pretreatment of algal bloom-
impacted seawater, algal cells and AOM can block or foul the pores and eventually form a
cake/gel layer on the surface of the membrane. These will cause rapid decline in the
membrane permeability. More frequent hydraulic backwashing, chemical enhanced
backwashing or chemical cleaning-in-place (CIP) is therefore required to recover the initial
permeability, resulting in longer system downtime. Some MF/UF plants apply in-line
coagulation, by dosing ferric salts, in front of the systems, to control the rate of non-
backwashable fouling.
If AOM is not effectively removed by the pretreatment processes, it can accumulate to form a
heterogeneous and compressible cake/gel layer on the surface of RO membranes. This can
result in lower permeability and higher differential pressure along the RO feed channel (Her
et al., 2004). The accumulated sticky substances may further initiate or promote particulate
and biological fouling by enhancing deposition of bacteria and other particles from the feed
water to the RO membrane and spacers (Berman and Holenberg 2005).
2.2.4.1 Biofouling in SWRO due to AOM deposition
Sticky and high charge-density TEPs produced during HABs can adhere and accumulate on
the surface of the SWRO membranes and spacers. The accumulated TEPs may serve as a
“conditioning layer” – a good platform for effective attachment and initial colonization by
bacteria which may then accelerate biofilm formation (Berman and Holenberg 2005; Li et al.
58
2015; Villacorte 2014). TEPs are also (partially) biodegradable and may serve as a substrate
for bacteria (Alldredge et al. 1993; BarZeev and Rahav 2015). Recently, Berman and co-
workers proposed a “revised paradigm” of aquatic biofilm formation facilitated by TEPs
(Berman and Holenberg 2005; Bar-Zeev et al. 2012). As illustrated in Figure 2.4, colloidal
and particulate TEPs and protobiofilms in surface water can initiate, enhance, and possibly
accelerate biofilm accumulation in RO modules.
Figure 2.4. Schematic illustration of the possible contribution of (a) colloidal biopolymers, (b) TEP, and (c)
protobiofilm (suspended TEP with extensive microbial outgrowth and colonization) in the initiation of
aquatic biofilms. A number of planktonic bacteria (first colonizers) can attach (d) reversibly on clean
surfaces or (e) irreversibly on TEP-conditioned surfaces. When nutrients are not limited in the water, (f)
contiguous coverage of mature biofilm can develop within a short period of time (minutes to hours). Biofilm
accumulation can cover a significant surface area of a (g) spiral wound membrane. Operational issues will
occur when substantial accumulation (h) obstructs permeate and feed channel flow. Photos and descriptions
adapted from (a-f) Bar-Zeev et al. (2012a), (g) Villacorte et al. (2009) and (h) Villacorte (2014).
Since bacteria require nutrients for energy generation and cellular synthesis, essential
nutrients such as biodegradable or assimilable organic carbon (BDOC or AOC), phosphates
and nitrates are likely the main factors dictating the formation and growth rate of biofilm in
RO modules. During the peak of an algal bloom, some of these essential nutrients may be
limited (e.g., phosphate, nitrates, silica) due to uptake by algae. However, when the bloom
reaches the death phase, algal cells start to disintegrate while releasing some of these
nutrients. Hence, biofouling initiated or enhanced by AOM will likely occur in SWRO some
time (depending on available nutrients) after the termination of an algal bloom. So far, the
role of AOM on membrane biofouling has only been illustrated in lab-scale experiments but
autopsies of biofouled membrane elements from both brackish and seawater RO plants have
shown the ubiquitous presence TEP among the biofilm accumulations (Figure 2.4; Villacorte
et al. 2009; Villacorte 2014).
2.2.4.2 Particulate/Organic fouling of MF/UF during HABs
Various marine laboratory- and pilot-scale studies have demonstrated the effect of algal
blooms on MF/UF operation. High molecular weight AOM (e.g., algal-derived biopolymers)
59
were identified to be the main causes of membrane fouling, more so than the algal cells
themselves (Kim and Yoon 2005; Ladner et al. 2010; Schurer et al. 2013; Villacorte et al.
2015b); however, a synergistic effect between algal cells and AOM may intensify the rate of
fouling in MF/UF membranes.
AOM such as TEPs are very sticky and can adhere strongly to the surface and pores of the
membranes, rendering conventional hydraulic backwashing ineffective in recovering the
initial membrane permeability (Figure 2.5). This scenario has been reported in recent studies
(e.g., Qu et al. 2012; Schurer et al. 2013; Villacorte 2014), signifying that AOM
accumulation not only causes rapid increase in operating pressures to maintain flux, but also
can increase non-backwashable or hydraulically irreversible fouling in MF/UF.
Figure 2.5: Fouling in MF/UF due to accumulation of algae and AOM and implications to membrane
performance. During filtration, algal cells and AOM retained by the membrane form a (compressible)
cake/gel layer which then causes rapid decline in permeability. This layer will only be partially removed
during backwashing due to the gluey nature of some large AOMs (e.g., TEPs), resulting in progressively
lower permeability (higher feed pressure if operated at constant flux) in the succeeding filtration cycles
(non-backwashable fouling). Modified from Villacorte 2014.
When deposition of algae and AOM on the MF/UF membrane surface is relatively uniform,
the impact on membrane permeability and backwashability can be explained with known
fouling mechanisms, illustrated in Figure 2.6 and further discussed below.
a) Membrane cake and pore constriction. (Figure 2.6a). Colloidal AOMs can enter into
the narrow pores of MF/UF membranes, some of which will adsorb to the pore wall and
eventually cause partial blockage of permeate flow (Herman and Bredee 1936). This can
cause rapid increase in trans-membrane pressure (TMP) during the initial stage of
filtration. Algal cells and large AOM can form a cake/gel layer on the surface of the
membrane. Colloidal AOMs and other colloids will then fill up the large interstitial voids
of the cake, narrowing the voids in the process. This may result in substantial increase in
cake resistance due to the gradual reduction in cake porosity. During backwashing, the
sticky AOM accumulated inside the membrane pores may not be completely removed,
resulting in only partial recovery of the initial membrane permeability.
60
Figure 2.6. Possible fouling mechanisms involved due to uniform deposition of AOM and small algal cells
in MF/UF. Each process is described in detail in the text. Modified from Villacorte (2014).
b) Substantial loss in effective filtration area (Figure 2.6b). Colloidal AOM can
accumulate inside membrane pores while algal cells and large AOM can accumulate at
the entrance of the pores. In both cases, some pores may be completely blocked by the
material and the active filtration area (membrane surface porosity) is substantially
reduced, resulting in higher localized flux for the remaining active pores (Herman and
Bredee 1936). An increase in flux can cause proportional increase in membrane resistance
to filtration. Additionally, non-backwashable fouling can occur if the foulants blocking
the pores are not effectively removed by hydraulic backwashing.
c) Incomplete cake/gel removal during backwashing (Figure 2.6c). Since algal cells
(typically range from 2 µm to 2 mm) and a substantial fraction of AOM are much larger
than the pores of commercial MF/UF membranes (<1 µm), cake/gel formation can be
mainly responsible for the increase in TMP. The accumulated AOM (like TEPs) are
typically sticky and tend to adhere strongly to the membrane surface. During
backwashing, a layer of the cake may remain on the surface of the membrane, which will
then cause additional filtration resistance in the subsequent filtration cycle.
d) Compression of accumulated cake/gel (Figure 2.6d). Filter cake/gel comprising AOM
and algal cells (soft-bodied) can be compressed due to localized increase of flux. Such
localized increase in flux may be a consequence of pores narrowing and/or completely
blocking as described above and hence occurs in combination with these fouling
mechanisms.
Theoretically, small particles (e.g., small algal cells, colloidal TEPs) deposit uniformly along
the capillary length whereas large particles (e.g., large algal cells and large particulate TEPs)
tend to deposit non-uniformly (Panglisch 2003; Lerch et al. 2007). Modeling the transport of
algal cells during dead-end filtration through a capillary membrane has shown that cells
smaller than 5 μm tend to deposit evenly along the capillary length while cells larger than
25 μm tend to accumulate at the segment of the capillary with the lowest axial flow e.g.,
61
dead-end (Figure 2.7; Villacorte 2014). Considering that most bloom-forming algae in
seawater are larger than 5 μm while TEPs and marine snow are in the range of hundreds of
micrometers, it is expected that cake thickness is not uniform and filtration resistance can
significantly vary along the length of a capillary membrane during severe HABs. Moreover,
very large algal cells (e.g., Noctiluca scintillans) or algal aggregates/colonies (e.g.,
Phaeocystis globosa) may partially or fully plug the entrance of the capillary channels, which
may then limit the productivity of the affected capillaries, resulting in substantial loss in
overall permeability of the membrane module (Figure 2.7d; Heijman et al. 2007). Applying
micro-screens with openings of 50 -150 µm upstream of UF membranes (current practice)
can eliminate the possibility of capillary plugging. Increased capillary diameter, higher
backwash flux/frequency or forward water/air flushing can potentially mitigate capillary
plugging issues during HABs.
Figure 2.7. Graphical illustration of foulant accumulation in inside-out, dead-end capillary MF/UF
membrane during filtration of algal bloom impaired seawater: axial and radial flow through (a) clean
membrane and fouled capillary membrane with (b) uniform accumulation, (b) non-uniform accumulation
and (d) entrance blockage. Figures modified from Panglisch (2003).
Pressure-driven capillary UF membranes have reportedly exhibited some degree of fouling

during algal bloom periods due to the high concentrations of algal cells and associated
organic matter. This is discussed in the case studies in Chapter 11 for the Jacobahaven
demonstration plant on the North Sea and the Sohar plant on the Gulf of Oman. HABs at the
Jacobahaven Plant impaired operation of UF membranes with chemically enhanced
backwashing (CEBs) required as frequently as once in 6 hours (see Chapter 11-10). Similarly,
at the Sohar plant (Chapter 11-3) with pressure driven, outside-in microfiltration membranes
(PDO), shorter periods between CEB were required.
2.2.5 Summary and outlook
A spike in AOM concentration during algal blooms at a seawater desalination plant can
adversely affect the operation of both pretreatment and RO systems. The high molecular
weight fraction of AOM may enhance clogging of GMF and significantly increase non-
backwashable fouling in MF/UF pretreatment. AOM, specifically TEP and their precursors,
may initiate and accelerate biofouling in RO membranes. To ensure continuous operation in
SWRO plants prone to algal blooms, the intake and/or the pretreatment system should be
62
reliable and robust to ensure continuous operation at design capacity and to minimize
breakthrough of AOM to the downstream SWRO system. It is also advantageous to have
monitoring instruments that can detect algal biomass (and algal species known to produce
TEP and other organic material in high quantities) and alert operators that intervention or
oversight is needed.
2.3 ALGAL ISSUES IN THERMAL DESALINATION PLANTS
Multi stage flash (MSF) and multi effect distillation (MED) are the most commonly
employed methods to desalinate seawater for municipal use and for drinking water supply in
the Middle East. MSF and MED thermal desalination account for 60% of the total seawater
desalinated capacity in the region (20 Mm3/d) (DesalData 2015). Thermal desalination plants
have been known to be impacted by blinding of intake screens by seaweed (macroalgae) but
generally are very forgiving of source water quality (Boerlage and Nada 2014). Unlike
SWRO desalination plants, phytoplankton blooms seldom affect thermal desalination plants.
During the prolonged 2008 algal bloom in the Gulf, the high concentration of the marine
dinoflagellate Cochlodinium polykrikoides had limited impact on thermal plants. At Fujairah
1 in the UAE, the MSF plants continued to operate without issue while the SWRO
desalination plant had to be shut down for more than one week. A minor shut down of
thermal desalination plants did occur in Sharjah, UAE (less than 24 hours) due to odor issues
associated with the product water. This was overcome through increased chlorination.
Another thermal desalination plant (MED) at Kish Island in Iran reported a higher seawater
pH during the bloom which required additional treatment measures to prevent alkaline
scaling during this period.
2.4 MARINE AND FRESHWATER TOXINS
2.4.1 Background
Marine and cyanobacterial toxins have been identified globally in various coastal
environments. Most of these toxins have been identified due to (i) poisoning events following
the consumption of fish or shellfish or (ii) harmful effects through direct contact or exposure
to aerosols. Therefore, the risks classically described for these compounds mostly relate to
acute toxicity. However, acute poisoning from toxins following the consumption of
desalinated drinking water has not yet been reported globally. The absence of such reports
may relate to the absence of actual poisoning events or may reflect the typically experienced
under-reporting of such events. In view of the risks these toxins pose, this section describes
HAB toxins, their pharmacological activity, and methods of analysis. Risk assessment is
presented in Chapter 8.
The following toxin groups have been identified for inclusion into this section: anatoxins
(ATXs), azapiracids (AZAs), ß-methylamino alanine (BMAA), brevetoxins (BTXs),
ciguatoxins (CTXs), cyclic imines (gymnodimines (GYM), spirolides (SPX), pinnatoxin
(PnTX)), domoic acid (DA), microcystins (MCs), nodularins (NODs), okadaic acid and
analogues (OA), palytoxins (PLTXs) incl. ovatoxins (OvTXs), saxitoxins (STXs),
tetrodotoxins (TTXs) and trichotoxins (TRXs). Additional to this group of toxins are two
taste and odor compounds, methyl isoborneol (MIB) and geosmin (GSM) that are non toxic,
but can be a source of customer complaints. Geosmin is included in the information below in
order of increasing molecular weight. Removal of these toxins is discussed in Chapter 10
with practical cases from the laboratory and full-scale plants. Note that not all of these toxins
are direct threats to desalination plants, but all are presented here for completeness. For each
toxin group there is a short description of the chemical nature of the compound and the
63
structure of a key compound from each group, as well as a listing of the algal or
cyanobacterial species producing the toxins.
The toxins range dramatically in their polarity and interactions with water – described here in
terms of lipophilicity (Table 2.1). Some are hydrophilic (soluble in water) and some
lipophilic (soluble in fats, oils, etc.). Molecular weights range from just over 100 to close to
3500 Daltons (Table 2.1). Toxins from both pelagic (water column) and benthic (seafloor or
epiphytic) micro-algae are considered since intake pipes of desalination plants can be close to
surface and close to the seafloor. Likewise, some freshwater toxins are included because
these can be found in brackish water, and because there is increasing evidence of them being
washed into nearshore coastal waters via rivers.
Some HAB toxins are ubiquitous around the planet, e.g. DA, OA, STX and some cyclic
imines (e.g. SPXs). Others are predominantly found in tropical and sub-tropical latitudes, e.g.
CTXs, PLTXs and OvTXs. BTXs are typically only found in the Gulf of Mexico, and rarely
in New Zealand, while the extent of the problems with some groups is not yet entirely clear,
e.g. AZAs are distributed globally but most poisoning events have been reported from Irish
shellfish. Most of the cyanobacterial toxins (e.g. MCs and NODs) have been reported to be of
terrestrial or brackish water origin. Some of the toxins, however, such as homo-anatoxin a
(homo-ATX-a) and trichotoxin, originate from benthic and pelagic marine organisms,
respectively. Cyanobacterial toxins rarely occur in open seawater.
Assessment of potential public health problems requires the detection and quantitation of the
HAB toxins in both intake and drinking water. To assess ecotoxicological problems it may
also be necessary to analyze them in the concentrated waste streams from desalination plants.
Classical detection methods for marine biotoxins have been based on whole animal assays,
e.g. intraperitoneal (i.p.) mouse bioassays. Such assays typically do not possess detection
limits (LODs) sufficiently low to detect the levels occurring in seawater or drinking water.
Therefore, methods described here include physico-chemical methods of analysis (HPLC-
UV/FLD/MS), antibody-based (e.g. ELISAs) or functional assays (PP2A, receptor-based
assays). Recognizing that desalination plants are unlikely to have direct access to this type of
sophisticated analytical equipment, Appendix 2 provides detailed instructions for some
relatively simple, antibody-based screening methods for HAB toxins.
2.4.2 Chemistry and source organisms
Beta-methyl-amino alanine (BMAA) is a small amino acid (Figure 2.8) that has been
implicated in a disease referred to as amyotrophic lateral sclerosis
(ALS), following its discovery in Guam. As an amino acid,
BMAA is a hydrophilic compound and has been shown to occur in
association with proteins; it is still unclear whether this bonding is
due to incorporation into proteins or due to nonspecific adsorption
or inclusion.
Even though initially reported to be widely produced by marine
and freshwater cyanobacteria, recent evidence points towards
Figure 2.8. ß-methyl- production in marine diatoms, some of which can be dominant
amino alanine (BMAA).
species, e.g. Chaetoceros spp. (Jiang et al. 2014; Réveillon et al.
2015).
Anatoxin-a (ATX-a) is a potent neurotoxin that has been related to deaths of animals, (e.g.
dogs), that have consumed contaminated surface waters from freshwater lakes or streams
(Figure 2.9). The cyanobacterium Anabaena circinalis is a common producer. Chemically,
ATX-a is a bicyclic secondary amine and belongs to the homotropane family (Wonnacott and
64
Gallagher 2006). While pinnamine had been isolated from a marine bivalve, Pinna muricata
(Takada et al. 2000) and the actual biological origin has not yet been elucidated, homo-ATX-
a has only been identified in a
benthic, mat-forming marine
cyanobacterium, Hydrocoleum
lyngbyaceum (Méjean et al.
2010).
Figure 2.9. Homotropanes: anatoxin-a (ATX-a) and its

methylated analogue homo-anatoxin-a (h-ATX-a) and pinnamine.
Table 2.1. Characteristics of marine and freshwater biotoxins: chemical formula, molecular
weights, lipophilicity, toxicity and mode of action. (Note: geosmin is non-toxic) FW =
predominantly freshwater origin; M = marine origin; FW + M = found in both freshwater and
marine systems.
Toxin Source Chemical class Formula Molecular LipophilicityToxicity Mode of
weight action
BMAA FW + M amino acid C4H10N2O2 118.1 hydrophylic amyotrophic unknown
lateral
sclerosis
anatoxin-a FW bicyclic amine C10H15NO 165.1 hydrophilic fast acting Na channel

alkaloid neurotoxin
geosmin FW bicyclic alcohol C12H22O 182.3 lipophilic odor olfactive

disturbance
saxitoxin FW + M alkaloid C10H17N7O4 299.1 hydrophilic fast acting Na channel

neurotoxin
domoic acid M cyclic amino acid C15H21NO6 311.1 hydrophilic neurotoxin glutamate
agonist
trichotoxin M chlorinated phenyl- C20H27CLO 318.2 lipophilic neurotoxin unknown

alkene
tetrodotoxin M alkaloid C11H17N3O8 319.1 hydrophilic neurotoxin Na channel
gymnodimine M cyclic imine, C32H45NO4 507.3 lipophilic fast acting Na channel

macrocycle neurotoxin
13desmethyl- M cyclic imine, C41H61NO7 691.4 lipophilic fast acting Na channel

spirolide C macrocycle neurotoxin
pinnatoxin G M cyclic imine, C42H63NO7 693.5 lipophilic fast acting Na channel

macrocycle neurotoxin
okadaic acid M polyether C44H68O13 804.5 lipophilic diarrhetic PP2a

toxin inhibitor
nodularin FW pentapeptide C41H60N8O10 824.4 lipophilic liver- PP2a

damaging inhibitor
azaspiracid M polyether C47H71NO12 841.5 lipophilic diarrhetic unknown

toxin
brevetoxin-B M polyether C50H70O14 894.5 lipophilic diarrhetic Na channel

neurotoxin
microcystin-LR FW + M heptapeptide C49H74N10O12 994.5 lipophilic liver- PP2a

damaging inhibitor
ciguatoxin M polyether C60H85O16 1061.6 lipophilic diarrhetic Na

neurotoxi channel
n
palytoxin M polyether C129H223N3O54 2677 amphiphilic neurotoxin Na K-
ATPase
maitotoxin M polyether C164H258O68S2 3380 amphiphilic neurotoxin Ca-

channel
65
Saxitoxin (STX) and tetrodotoxin (TTX). Saxitoxin (Figure 2.10) and its analogues are very
potent neurotoxins that induce symptoms in humans within minutes after consumption of
contaminated shellfish; severe poisoning may lead to rapid death in patients (Rossini and
Hess 2010). Tetrodotoxins
have a different chemical
structure from STX (Figure
2.10), but act in a very
similar fashion by blocking
sodium ion channels. Effects
in humans are also very
similar in that rapid death can
occur as a result of paralysis.
Tetrodotoxin has long been
Figure 2.10: Saxitoxin and tetrodotoxin. known as the causative agent
in puffer fish poisoning
(Fuchi et al. 1988; Kodama et
al. 1983). Saxitoxins are a family of toxins based on a tetrahydropurine skeleton. The
tetrahydropurine group renders the molecule highly water-soluble. To date, 57 analogs have
been reported in cyanobacteria, marine dinoflagellates, and in molluscs (Wiese et al. 2010).
STX analogues do not exhibit a strong ultraviolet (UV) absorbance or fluorescence. They are
typically stable to heat treatment up to 100°C. Different acid and base treatments will lead to
various transformations. In particular, all C11-epimeric pairs (e.g. GTX2 and 3 or GTX1 and
4) will interconvert and equilibrate to a constant ratio at high pH. Similarly, carbamoyl and
sulfocarbamoyl analogues will convert to decarbamoyl analogues through cleavage of the
carbamoyl-ester group at pH > 9 (e.g. C1 to dc-GTX2 and C2 to dc-GTX3). Under acidic
conditions, the carbamoylester is relatively stable but the sulphate ester will be cleaved to
convert sulfocarbamoyl-groups into the more toxic carbamoyl groups (e.g. C1 to GTX2 and
C2 to GTX3). These transformations are important because the STX analogues can differ by
well over an order of magnitude in potency.
TTX and analogs have recently been found as contaminants in bivalves in temperate waters,
i.e. the English Channel (Turner et al. 2015). These authors have also been able to
demonstrate that bacteria associated with the same shellfish are capable of biosynthesizing
TTXs. The association of tetrodotoxins with the marine dinoflagellate Prorocentrum and
accumulation in bivalve mollusks, recently discovered in the Mediterranean (Vlamis et al.
2015), suggest that microalgal blooms may well act as carriers for TTX-producing bacteria.
Domoic acid (DA) Due to the common occurrence of one of its source organisms (the diatom
Pseudo-nitzschia spp.), DA occurs throughout the world. As DA has weak diarrheic
properties, the seaweed Chondria armata (which also produces DA) has been used in Japan
as an anti-worming agent; however, the severe poisoning of over 100 people following
consumption of DA-containing mussels 1987 in Canada, including 3 fatalities, stopped this
practice. DA is a small cyclic amino acid, with three carboxylic acid groups (Figure 2.11).
These groups are responsible for its solubility in
water and its relatively high polarity. The acid
constants (pKas) of the three carboxylic acids and
the cyclic amino group have been determined
using NMR techniques by Walter et al. (Walter et
al. 1992). Although numerous isomers and several
analogues have been reported, so far only DA and
its C5-diastereomer have been shown to be of Figure 2.11. Domoic acid.
66
toxicological relevance (Rossini and Hess 2010). DA transforms into its diastereomer
through heat or long-term storage (Quilliam et al. 1995) and analysis has focused on
determination of the sum of these two isomers as best estimate of the total toxicity. A
conjugated double bond in the aliphatic side chain allows for detection of DA by UV
absorbance and both UV and MS detection are commonly used for the physico-chemical
determination of DA (Hess et al. 2005). The conjugated double bond also leads to light-
sensitivity and is the cause of radical-mediated oxidative metabolism.
Domoic acid has been reported in a wide variety of seafood organisms, including mussels,
scallops and anchovies. As a contaminant in shellfish tissues, DA is heat stable and cooking
does not destroy the toxin. Its stability under various conditions has been studied, and storage
of raw or autoclaved tissues only resulted in approximately 50% degradation of the toxin
after 5 months (McCarron et al. 2007).
Trichotoxin (TRI) Cytotoxicity of trichotoxin is ca. 1000-fold less than STX, but negative
effects of Trichodesmium sp. have been reported on marine fauna and humans, so the toxin
has to be considered (Schock et al. (2011)). TRI is a small, lipophilic, organic molecule
(Figure 2.12) that has been isolated as an oil
from Trichodesmium thiebauthii, a
ubiquitous nitrogen-fixing marine
cyanobacterium (Schock et al. 2011). This
species is pelagic and known to form
extensive and dense blooms in tropical and
subtropical areas. Some claim the Red Sea
derived its name from extensive blooms of
this organism.
Figure 2.12. Trichotoxin isolated from a field sample
Cyclic imines: pinnatoxins (PnTXs),
of Trichodesmium thiebauthii. gymnodimines (GYM) and spirolides
(SPX). These compounds are all classified
as fast acting toxins (FTAs) due to the rapid death of mice following intraperitoneal injection.
Acute poisoning in humans has not been proven, however, despite an initial suspicion
following consumption of the mussel Pinna muricata (Zheng et al. 1990). Even though acute
toxicity has not been demonstrated to
date, the risk of long-term exposure to
sub-lethal doses is of concern given
these toxins capacity to cross the
intestinal and blood–brain barriers,
and their high affinities for human
neuronal nicotinic acetylcholine
receptors (Aráoz et al. 2015).
The group of cyclic imine toxins was
discovered due to their rapid response
in the lipophilic mouse bioassay (Hu
et al. 1995; Seki et al. 1995; Uemura
et al. 1995). They all have the
chemical functional groups of a cyclic
imine and a macrocycle (Figure 2.13);
breakage of either ring will result in
loss of toxicity. Spirolides are Figure 2.13. Cyclic imine toxins: pinnatoxin G, gymnodimine,
13-desmethyl spirolide C.
produced by Alexandrium ostenfeldii
67
and Alexandrium peruvianum (Hu et al. 1995; Van Wagoner et al. 2011); gymnodimines are
produced by Karenia selliformis and A. peruvianum (Miles et al. 2003; Seki et al. 1995; Van
Wagoner et al. 2011). All of these organisms are marine, pelagic dinoflagellates. In contrast,
pinnatoxins are produced by Vulcanodinium rugosum, a semi-benthic dinoflagellate (Hess et
al. 2013; Rhodes et al. 2011; Rhodes et al. 2010; Selwood et al. 2010). Chromatographic
behavior suggests intermediate lipophilicity and studies using passive samplers or seawater
pre-concentration with lipophilic resins demonstrate that detectable concentrations are
dissolved in seawater following HAB occurrences (Fan et al. 2014; Fux et al. 2009; Garcia-
Altares et al. 2014; Rundberget et al. 2009; Zendong et al. 2014).
Microcystins and nodularins are of intermediate lipophilicity. They are produced by fresh-
and brackish water cyanobacteria (Figure 2.14). Like okadaic acid, these compounds inhibit
phosphoprotein phosphatases and have been linked to liver damage in humans. Microcystins
are found in coastal and
fresh water environments,
either accumulated in
shellfish or directly in the
water (Morais et al. 2008;
Amorim and Vasconcelos
1999; Kohoutek et al. 2010,
Kudela 2011; Vasconcelos
1995, 1999). Transfer from
freshwater to coastal marine
waters and subsequent
uptake by coastal mammals
has recently been shown by
Californian researchers
(Gibble and Kudela 2014;
Figure 2.14. Microcystin-LR and nodularin. Miller et al. 2010). Analogs
of these groups are
numerous and a Norwegian group has recently reported a large number congeners of
microcystins (Ballot et al. 2014; Miles et al. 2013a; Miles et al. 2012; 2013b).
Okadaic acid (OA) and dinophysistoxins (DTXs). OA and DTXs are phosphoprotein
phosphatase inhibitors and potentially tumor promoters and have been responsible in many
areas for diarrhetic shellfish poisoning (EFSA, 2008b). Chemically, OA and DTXs belong to
the polyether family and
possess a carboxylic acid
group rendering them
somewhat water soluble
(Figure 2.15). Numerous
analogs have been reported
but all are based on these
Figure 2.15. Okadaic acid and dinophysistoxins-1 and -2: OA (R1 = CH3,
R2 and R3 = H), DTX1 (R1 and R2 = CH3, R3 = H), DTX2 (R1 and R2 =
three main skeletons. OA
H, R3 = CH3). had initially been
discovered as a bioactive
metabolite in a marine sponge of the genus Halichondria, but was rapidly also attributed to
the benthic dinoflagellate Prorocentrum lima (Murakami et al. 1982; Tachibana et al. 1981).
Around the same time a closely related compound, dinophysistoxin-1 (DTX1), was described
as a metabolite of the pelagic dinoflagellate Dinophysis following a major series of human
shellfish poisoning (Murata et al. 1982; Yasumoto et al. 1978). Since then, many species of
68
the genera Dinophysis and Prorocentrum have been described in all seas and oceans and
most of these are considered ubiquitous and toxic in all areas (Henrichs et al. 2013;
Hoppenrath et al. 2013; Reguera et al. 2012). Solubility and persistence of OA and DTXs in
seawater have been shown for several weeks to months after blooms via field studies using
passive samplers (Fux et al. 2009; MacKenzie et al. 2004; Zendong et al. 2015a).
Azaspiracids (AZAs) are another diarrhetic shellfish poisoning group that were discovered
following human poisoning from consumption of mussels (Mytilus edulis) produced in
Ireland (McMahon and Silke 1996; Satake et al. 1998a). Toxicity in humans clearly targets
the digestive tract, but the mechanism of action has not yet been elucidated despite many
efforts (EFSA 2008a; Hess et al. 2015; Twiner et al. 2014). Almost 40 analogs of this
polyether have been described, and the distribution of the toxins and causative organisms has
been shown to be ubiquitous (Hess et al. 2014; Tillmann et al. 2014; Twiner et al. 2014).
The AZA producing organisms are
all small, pelagic dinoflagellates
belonging to the closely related
genera of Amphidoma and
Azadinium. Consistent with its polar
functional groups (carboxylic acid
and secondary amine), AZA is
Figure 2.16. Azaspiracid. soluble in seawater (Fux et al. 2009).
Brevetoxin (BTX) toxicology is
complex because this family of compounds causes toxicity from consumption of
contaminated seafood as well as from direct contact with seawater or inhalation of spray from
seawater. Despite their documented relation to harmful microalgae in US since the 1960s
(Spikes et al. 1968), controversy existed until recently as to which were the toxicologically
most relevant analogs (Bottein et al. 2007; Bottein et al. 2010; Henri et al. 2014). Thus, a
compound-specific maximum permissible
limit has not yet been agreed upon, and
furthermore, risk assessment and
management at international level will
continue to remain very difficult for this
toxin group (EFSA 2010; Lawrence et al.
2011).
BTXs are polyethers with contiguously
fused rings which make the molecule
somewhat more rigid than other polyethers,
e.g. OA and AZAs (Figure 2.17). BTXs are
much more lipophilic than most previously
described toxins and little is known about
their absolute environmental dissolved
concentration, even when passive samplers Figure 2.17. Brevetoxins: lipophilic polyethers with
or very sensitive methods have been contiguously fused rings. Two main structural skeletons
developed to detect them in seawater can be distinguished with analogues mainly changing at
(Kulagina et al. 2006a; Shea et al. 2006). the position indicated with R.
The causative organism is Karenia brevis, a
major, pelagic bloom-forming dinoflagellate, which has undergone a number of taxonomic
revisions; synonyms include Gymnodinium breve and Ptychodiscus brevis. Distribution has
69
mostly been reported from the Gulf of Mexico region, but one major event also occurred in
New Zealand, suggesting a wider distribution than initially believed (Ishida et al. 2004).
Ciguatoxin (CTX) and maitotoxin (MTX). CTXs are among the most toxic compounds
known and current US-FDA guideline values in fish are not to exceed 0.01 µg P-CTX1
eq. kg-1 fish flesh. Like BTXs, they belong to the family of polyethers with contiguously
fused rings (Figure 2.18). Both are produced by dinoflagellates belonging to the tropical
marine genera Gambierdiscus and Fukuyoa (Litaker et al. 2010). These are epiphytic or
benthic dinoflagellates, meaning that they live attached to surfaces on the sea floor. They are
responsible for the syndrome called ciguatera fish poisoning (CFP) which is a major source
of illness in tropical countries dependent upon reef fish for protein. Gambierdiscus species
also swim, and thus can be drawn into a desalination plant intake under certain situations,
though it seems unlikely that large numbers of cells would be encountered in this way.
Furthermore, although some species or strains produce ciguatoxins directly, the major
metabolites of Gambierdiscus species are metabolized to the much more potent ciguatoxins
following consumption by fish. Ciguatoxins and maitotoxins are not likely to be a concern to
desalination plants.
Figure 2.18. Ciguatoxin (CTX) and maitotoxin (MTX) are lipophilic polyethers with contiguously fused
rings (similar to BTXs). CTX4B is shown here. CTXs are among the most lipophilic compounds while
MTX is an amphiphilic polyhydroxy- polyether with two sulphate-groups.
Palytoxin (PLTX) and ovatoxin (OVTX). PLTX is atypical of most known marine toxins in
that it poses risks to humans through multiple routes of exposure (oral, inhalational, and
dermal). Palytoxins have been associated with human deaths following consumption of fish
(Onuma et al. 1999) and with respiratory and dermatological syndromes from exposure
through household aquarium supplies (Cortini et al. 2015; Davey et al. 2015) or
environmental exposure to bathers and beachgoers (Funari et al. 2015; Tartaglione et al.
2015). These compounds are amongst the largest non-proteinaceous natural molecules
(Figure 2.19) and have very high intrinsic toxicity. Palytoxin and its analogs ostreocins and
ovatoxins are produced by zooanthids, e.g. Palythoa spp. (Kimura et al. 1972),
dinoflagellates, e.g. Ostreopsis spp. (Usami et al. 1995) and potentially cyanobacteria
(Kerbrat et al. 2011). PLTXs are found in dinoflagellates distributed throughout tropical and
sub-tropical habitats, as well as in temperate waters of the Mediterranean and Adriatic Seas.
Their chemistry and pharmacology have been recently reviewed (Carmen Louzao et al. 2015;
Ciminiello et al. 2015).
70
Although the algal

source for palytoxins
(Ostreopsis species) are
benthic organisms, they
also do occur in dense
blooms in shallow
coastal waters, with
accumulations of cells
embedded in mucilage
(Funari et al. 2015,
Tartaglione et al. 2015)
that could be of concern
to desalination plants. Of
particular interest are the
Figure 2.19: Palytoxins are amongst the largest non-proteinaceous natural observations of acute
molecules and have very high intrinsic toxicity.
toxicity following
aerosol exposures to
these blooms. One noteworthy example occurred in 2005 when ~ 200 beach-goers
experienced symptoms of rhinorrhea, cough, mild dyspnea, bronchoconstriction, and fever
that coincided with a bloom of Ostreopsis ovata along the Mediterranean coast near Genoa,
Italy (Ciminello et al. 2006). Altogether, over 650 cases have now been reported throughout
the northern Mediterranean and Adriatic seas in association with exposure to waters
containing Ostreopsis ovata. The concentrations of PLTX and/or PLTX-like compounds
required to cause these effects through inhalational, dermal, and ocular exposures are still
unknown.
2.5 TASTE AND ODOR COMPOUNDS
Geosmin (GSM) and methylisoborneol (MIB). Geosmin and MIB are both non toxic
volatiles produced by cyanobacteria and marine species. Geosmin is a bicyclic alcohol
(Figure 2.20) with a distinct earthy
flavor and aroma produced by a
type of actinobacteria, and is
responsible for the earthy taste of
beets and a contributor to the strong
scent that occurs in the air when
rain falls after a dry spell of weather
or when soil is disturbed. In
Figure 2.20 Structures of geosmin (left) and methylisoborneol chemical terms, it is a lipophilic
(right). compound and an analogue of
decalin. Its name is derived from
the Greek geo- "earth" and osmin- "smell". Cyanobacteria are also major producers of
geosmin and MIB, another compound potentially adding to poor smelling drinking water
(Polak and Provasi 1992; Suurnäkki et al. 2015).
2.6 DETECTION TECHNIQUES
There is an increasing range of analytical methods available for the detection and
quantification of marine and cyanotoxins, and they vary greatly in the manner of detection,
the information they provide and level of sophistication (Botana 2014; Harada et al. 1999;
Lawrence et al. 2011; Meriluoto and Codd 2005; Nicholson and Burch 2001). For
71
convenience, geosmin and MIB will be included in this section, even though they are not
toxins.
As mentioned above, assays based on whole, live animal exposure are excluded from this
discussion due to their lack of sensitivity for desalination processes. In some cases, assays
based on immortalized cell lines are also available for screening. A comprehensive discussion
of the range of cell-based screening assays used to detect cyanotoxins is given in the Water
Quality and Treatment Research Report 60 (Froscio et al. 2008). Such cellular approaches
have also been developed for many marine biotoxins (Canete et al. 2010; Canete and Diogene
2010; Ledreux et al. 2012; Ledreux et al. 2010); however, the techniques have been only
rarely validated for seawater matrix (Kulagina et al. 2006a). Still, they may be used for the
estimation of toxin concentrations present in concentrated phytoplankton from seawater.
Similarly, some lateral flow immunoassays have been developed for DA, OA, STXs, and
other HAB toxins (Laycock et al. 2010; McLeod et al. 2015; Vale et al. 2009); some of these
may also be used for analysis of concentrated phytoplankton from seawater. Appendix 2
provides protocols for the use of some of these as screening assays.
Quantitative techniques available include immunological or biochemical screening
techniques based on enzyme-linked immunosorbent assays (ELISA) or enzyme activity
(protein phosphatase inhibition, PPIA) assays respectively. Some techniques, here referred to
as assays, will give a sum response for all compounds, either related to the sum of
concentrations present (immunological assays) or relating to the sum of toxic equivalents
present (functional assays). Other methods, mainly those based on separation by gas- or
liquid chromatography with various detectors will give results on individual compounds for
which the sum toxicity present needs to be calculated via multiplication with toxic
equivalence factors, specific to each compound.
One technique, liquid chromatography coupled to tandem mass spectrometry has been
extensively used for all biotoxin compound groups except the very volatile geosmin and MIB.
Even though it is an expensive and sophisticated technique, it has also been adapted for
detection and quantitation of multiple groups of toxins in a single analysis (Brana-Magdalena
et al. 2014; Fux et al. 2007; Quilliam et al. 2001; These et al. 2011; van den Top et al. 2011;
Zendong et al. 2015b); however, sensitivity of the technique by itself is not good enough for
direct analysis of seawater and thus, pre-concentration or other sample pretreatment must
typically be used to achieve required detection limits for analysis in sea- or drinking water;
this has recently been effectively demonstrated for okadaic acid group toxins (Zendong et al.
2015a).
A summary of analytical techniques that are available for different classes of toxins and their
detection limits are given in Table 2.2. For the techniques described in the table, the detection
limits may vary depending upon the standards that are available and instrumentation used. A
range of other methods used within various research laboratories for screening and analysis
includes ELISA methods for microcystins (Appendix 2), neuroblastoma cytotoxicity assay,
saxiphilin and single-run HPLC methods for saxitoxins. The following section gives a brief
overview of major methods available for individual compound groups.
2.6.1 Geosmin and methylisoborneol
The chemical procedures used to analyze organic taste and odor compounds in water must be
very sensitive, because many of these substances can be detected by sensory analysis (i.e. the
human nose) at ng/L levels. The most common method currently used for quantitative
analysis is gas chromatography combined with mass spectrometry (GC/MS). As these
compounds need to be detected at very low concentrations, a pre-concentration method often
72
is required. The most important methods used for the pre-concentration step are summarized
below.
Closed-loop stripping analysis (CLSA) has been widely used for the analysis of non-polar
volatile organic compounds of intermediate molecular weight, at the ng/L to µg/L level. The
compounds are stripped from the water by a recirculating stream of air and then adsorbed
from the gas phase onto a few milligrams of activated carbon. They are then extracted from
the carbon with a few µL of carbon disulphide for direct analysis. This method can be applied
to both raw and treated waters. The main advantage of the method is that it does not require
further concentration of the solvent. Prior to the widespread adoption of solid phase micro-
extraction (see below) this method was considered the standard for isolation of MIB and
geosmin (Krasner et al. 1983). The limit of detection (LOD) for this method is usually
reported as 1-2 ng/L.
Solid phase microextraction (SPME) is simpler and more cost-effective than CLSA, and has
thus gained popularity in recent years (Huang et al.; 2004). The LOD for this method is usually
reported as 1-2 ng/L for geosmin and slightly higher for MIB at 4ng/L.
2.6.2 Cylindrospermopsin
The method recommended for cylindrospermopsin is an HPLC method with SPE pre
concentration (Nicholson and Burch 2001; Metcalf et al. 2002). A protein synthesis inhibition
assay has been developed for cylindrospermopsin (Froscio et al. 2001).
2.6.3 Saxitoxins and tetrodotoxins
The analytical methods available for saxitoxins are continuously evolving and are based upon
either high performance liquid chromatography and fluorescence detection or mass spectral
detection (LC/MS/MS). Internationally, the only technique recognized by the Association of
Official Analytical Chemists (AOAC) for analyzing saxitoxins from shellfish (where they are
commonly found) other than mouse bioassay is a technique based upon liquid
chromatography with pre-column derivatization (Nicholson and Burch 2001; Lawrence et al.
2005). This technique is not yet widely used for analysis of cyanobacterial material. Similarly,
TTXs may be detected using LC-MS/MS (Boundy et al. 2015, Turner et al. 2015).
2.6.4 Domoic acid
This toxin is one of the rare compounds where detection of a single entity is sufficient to
characterize the risk. Thus, several methods have been developed and validated, or cross-
validated (Hess et al. 2001; Kleivdal et al. 2007; Quilliam et al. 1995). More recent
developments have also allowed for a significant lowering of detection limits that permit
detection of relevant levels (Table 2.3), with LODs sufficiently low to ascertain relevant
levels in purified drinking water.
2.6.5 Microcystins and nodularin
Congener-independent immunoassay techniques have recently been developed for
microcystin and nodularin (Fischer et al. 2001; Samdal et al. 2014). These techniques have
the most appropriate detection and quantitation limits. It is important to select the appropriate
analytical method for each situation, which may change regionally. For example, the
technique considered most suitable to monitor microcystins in relation to the Australian
Drinking Water Guidelines is high performance liquid chromatography with photo diode
array detection or mass spectral detection (HPLC-PDA or HPLC-MS) (Nicholson and Burch
2001.)
73
Table 2.2. Toxin or taste and odor compounds, limits of detection (LOD) or quantitation
(LOQ) in seawater. When an assay was not developed for seawater analysis, the LOD and
LOQ values were estimated from the working range.
Quantitative
Toxins detection LOD/LOQ Reference
techniques
Geosmin, methyl- GC-MS(/MS) 2 ng/L / 6 ng/L Huang et al. 2004
isoborneol
Domoic acid Biosense ELISA 3 µg/kg / 11 µg/kg § McLeod et al. 2015;
Trainer et al. 2009
LC-FLD (direct 15 ng/L / 45 ng/L in Devez and Delmas 2013
injection) seawater
LC-MS/MS 15 ng/L / 45 ng/L in Mafra Jr et al. 2009
seawater
LC-MS/MS 30 ng/L / 100 ng/L Wang et al. 2007a
LC-MS/MS (SPE- 20 ng/L (LOD) de la Iglesia et al. 2008
disks)
LC-UV 43 ng/L / 130 ng/L Mafra Jr et al. 2009
3
Saxitoxin H-STX-RBA 45 µg/kg / 126 µg/kg § van Dolah et al. 2012
Abraxis ELISA 200 µg/kg shellfish McLeod et al. 2015
(LOQ) §
Neuronal network 76 pM in seawater Kulagina et al. 2006a
(LOD)
LC-FLD > 2000 nM (LOD) *§ Dell’Aversano et al. 2005
LC-MS/MS > 4000-6000 nM Dell’Aversano et al. 2005
(LOD) *§
Cylindrospermopsin ELISA 0.5 µg/L Froscio et al. 2001
HPLC 1 µg/L Metcalf et al. 2002
Microcystins ELISA 40 ng/L (LOQ) in Samdal et al. 2014
drinking water
PPIA 100 ng/L Carmichael et al. 1999
Radioactive PP2a 50 pM (LOD) Serres et al. 2000
binding assay
HPLC-PDA 0.1 µg/L Ho et al. 2006
LC-MS/MS (MYC- 2.5 ng/L (LOQ) Wang et al. 2007b
LR)
Nodularins ELISA 50 ng/L Samdal et al. 2014
PPIA 100 ng/L Nicholson and Burch 2001
Radioactive PP2a 40 ng/L / 120 ng/L Serres et al. 2000
binding assay
HPLC 0.5 µg/L Nicholson and Burch 2001
74
Table 2.2 (Continued)

Quantitative
Toxins detection LOD/LOQ Reference
techniques
Okadaic acids Abraxis ELISA 100 µg/kg shellfish § McLeod et al. 2015
PPIA 63 pg/mL (LOD) in Tubaro et al. 1996
aqueous solution
Radioactive PP2a 200 pM (LOD) Serres et al. 2000
binding assay
LC-MS/MS (without 0.16 ng/mL (LOD) § Brana-Magdalena et al.
preconcentration) 2014
LC-MS/MS (with HP- 0.2 ng/L seawater Zendong et al. 2015a

20 pre-concentration) (LOQ)
Azaspiracids ELISA 57 µg/kg LOQ) § Samdal et al. 2015
LC-MS/MS (without 0.4 µg/kg (LOD) § Zendong et al. 2015b
preconcentration)
Neuronal network 0.5 nM LOD solution Kulagina et al. 2006b
(IC50=2nM)
13-desmethyl LC-MS/MS (without 0.15 µg/kg (LOD) § Zendong et al. 2015b
Spirolide C preconcentration)
PnTX-G LC-MS/MS (without 0.1 µg/kg (LOD) § Zendong et al. 2015b
preconcentration)
*Per analog (sum of toxic equivalents may be significantly higher)
§ Not validated for seawater matrix but for shellfish matrix
2.6.6 Azaspiracids, brevetoxins, ciguatoxins, cyclic imines, okadaic acid and

dinophysistoxins
These toxins are all lipophilic toxins and may be detected by LC-MS/MS (Plakas et al. 2008;
Yogi et al. 2014, Zendong et al. 2015b), preferentially following pre-concentration with
passive sampling resins (Zendong et al. 2015a). Alternative techniques such as ELISAs exist
for some groups (e.g. OA and AZA groups) but are not necessarily more sensitive (Table 2.3).
2.7 GAPS AND PERSPECTIVES ON ANALYTICAL TECHNIQUES
Improvements are direly required for a methodology allowing for the detection and
quantitation of large numbers of toxins in seawater and drinking water. Most of the currently
available techniques have been developed for detection of algal toxins in shellfish and the
concentration levels are typically 100 – 1000 fold higher in this matrix compared to seawater
or drinking water. Pre-concentration techniques using resins, either in situ or in the laboratory,
have recently been shown to be an effective approach; however, none of these methods have
been brought to validation at the interlaboratory level.
A further need to implement regular testing of seawater and drinking water in desalination
plants would be proficiency testing for this matrix. Currently, proficiency testing for algal
toxins in shellfish matrices is available internationally via a registered provider: Quality
Assurance in Marine Environmental Matrices (QUASIMEME 2015). A scheme for seawater
and drinking water matrices could be added through this provider, subject to an expert
laboratory providing test materials and analytical services to characterize such test materials.
75
A promising area that is developing rapidly is the application of molecular techniques

(quantitative PCR) for determination of genes for toxin production. Among the algal toxins
from diatoms and dinoflagellates, this approach has only been applied to STX thus far, as the
toxin-producing genes are not known for most of the other toxins. Still, this approach will
only apply to detection of toxin-producing algae, not the toxins themselves.
2.8 REFERENCES
Alldredge, A. L., Passow, U., and Logan, B. E. 1993. The abundance and significance of a
class of large, transparent organic particles in the ocean. Deep-Sea Research I: 40, 1131–
1140.
Amorim, Á. And Vasconcelos, V. 1999. Dynamics of microcystins in the mussel Mytilus
galloprovincialis. Toxicon 37, 1041-1052.
Aráoz, R., Hess, P., Pelissier, F., Benoit, E., Servent, D., Zakarian, A., and Molgó, J. 2015.
Cyclic imine toxins: From shellfish poisoning to neuroscience: The case of acyl
derivatives. Biochemical Pharmacology 97, 622.
Azetsu-Scott, K. and Passow, U. 2004. Ascending marine particles: Significance of
transparent exopolymer particles (TEP) in the upper ocean. Limnology and
Ballot, A., Sandvik, M., Rundberget, T., Botha, C. J., and Miles, C. O. 2014. Diversity of
cyanobacteria and cyanotoxins in Hartbeespoort Dam, South Africa. Marine and
Freshwater Research 65, 175-189.
Bar-Zeev, E., Berman-Frank, I., Girshevitz, O., and Berman, T. 2012. Revised paradigm of
aquatic biofilm formation facilitated by microgel transparent exopolymer particles.
Proceedings of the National Academy of Sciences 109(23), 9119-9124.
Bar-Zeev, E. and Rahav E. 2015. Microbial metabolism of transparent exopolymer particles
during the summer months along a eutrophic estuary system. Frontiers in Microbiology
6:403. doi: 10.3389/fmicb.2015.00403.
Berktay, A. 2011. Environmental approach and influence of red tide to desalination process
in the middle-east region. International Journal of Chemical and Environmental
Engineering 2(3), 183-188.
Berman, T. and Holenberg, M. 2005. Don't fall foul of biofilm through high TEP levels.
Filtration & Separation 42(4), 30-32.
Botana, L.M. 2014. Seafood and Freshwater toxins: pharmacology, physiology, and detection.
Third Edition. CRC Press, Boca Raton, FL, USA.
Boerlage, S. F. E. and Nada, N. 2014. Algal toxin removal in seawater desalination processes,
In Proceedings of European Desalination Society, Cyprus.
Bottein, M.-Y. D., Fuquay, J. M., Munday, R., Selwood, A. I., van Ginkel, R., Miles, C. O.,
Loader, J. I., Wilkins, A. L., and Ramsdell, J. S. 2010. Bioassay methods for detection of
N-palmitoylbrevetoxin-B2 (BTX-B4). Toxicon 55, 497-506.
Bottein, M.-Y. D, Wang, Z., and Ramsdell, J. S. 2007. Intrinsic potency of synthetically
prepared brevetoxin cysteine metabolites BTX-B2 and desoxyBTX-B2. Toxicon 50,
825-834.
76
Boundy, M. J., Selwood, A. I., Harwood, D. T., McNabb, P. S., and Turner, A. D. 2015.
Development of a sensitive and selective liquid chromatography–mass spectrometry
method for high throughput analysis of paralytic shellfish toxins using graphitised
carbon solid phase extraction. Journal of Chromatography A 1387, 1-12.
Brana-Magdalena, A., Leao-Martins, J. M., Glauner, T., Gago-Martinez, A. 2014.
Intralaboratory validation of a fast and sensitive UHPLC/MS/MS method with fast
polarity switching for the analysis of lipophilic shellfish toxins. Journal of AOAC
International 97, 285-292.
Canete, E., Campas, M., de la Iglesia, P., and Diogene, J. 2010. NG108-15 cell-based and
protein phosphatase inhibition assays as alternative semiquantitative tools for the
screening of lipophilic toxins in mussels. Okadaic acid detection. Toxicology in Vitro 24,
611-619.
Canete, E., Diogene, J. 2010. Improvements in the use of neuroblastoma x glioma hybrid
cells (NG108-15) for the toxic effect quantification of marine toxins. Toxicon 55, 381-
389.
Carmen Louzao, M., Fraga, M., and Vilariño, N. 2015. Pharmacology of palytoxins and
ostreocins, Phycotoxins. John Wiley & Sons, Ltd, pp. 113-135.
Carmichael, W. W. and An, J. 1999. Using an enzyme linked immunosorbent assay (ELISA)
and a protein phosphatase inhibition assay (PPIA) for the detection of microcystins and
nodularins. Journal of Natural Toxins 7: 377-385 .
Ciminiello, P., Dell'Aversano, C., Fattorusso, E., Forino, M., Magno, G. S., Tartaglione, L.,
Grillo, C., and Melchiorre, N. 2006. The Genoa 2005 outbreak. Determination of
putative palytoxin in Mediterranean Ostreopsis ovata by a new liquid chromatography
tandem mass spectrometry method. Analytical Chemistry, 78(17), 6153-6159.
Ciminiello, P., Dell'Aversano, C., and Forino, M. 2015. Chemistry of palytoxin and its
analogues, Phycotoxins. John Wiley & Sons, Ltd, pp. 85-111.
Claquin, P., Probert, I., Lefebvre, S., AND Veron, B. 2008. Effects of temperature on
photosynthetic parameters and TEP production in eight species of marine microalgae.
Aquatic Microbial Ecology 51, 1-11.
Cortini, E., Lonati, D., Petrolini, V. M., Giampreti, A., Tubaro, A., and Locatelli, C. A. 2015.
Dangerous palytoxin exposure after boiling coral. Clinical Toxicology 53, 339-339.
Daly, R., Ho, L., and Brookes, J. 2007. Effect of chlorination on Microcystis aeruginosa cell
integrity and subsequent Microcystin release and degradation. Environmental Science
and Technology 41, 4447-4453.
Davey, M. P., McKeown, N., and Hendrickson, R. 2015. Palytoxin yoxicity in a coral
enthusiast and his family from unmanipulated coral. Clinical Toxicology 53, 776-777.
de la Iglesia, P., Giménez, G., and Diogène, J. 2008. Determination of dissolved domoic acid
in seawater with reversed-phase extraction disks and rapid resolution liquid
chromatography tandem mass spectrometry with head-column trapping. Journal of
Chromatography A 1215, 116-124.
Dell’Aversano, C., Hess, P., and Quilliam, M. A. 2005. Hydrophilic interaction liquid
chromatography–mass spectrometry for the analysis of paralytic shellfish poisoning
(PSP) toxins. Journal of Chromatography A 1081, 190-201.
DesalData, 2015. https://www.desaldata.com/, accessed 3 November 2015.
77
Devez, A. and Delmas, D. 2013. Selective liquid chromatographic determination of trace

domoic acid in seawater and phytoplankton: improvement using the o-
phthaldialdehyde/9-fluorenylmethylchloroformate derivatization. Limnology and
Oceanography: Methods 11, 327-336.
EFSA, 2008a. Marine biotoxins in shellfish - Azaspiracid group, Scientific Opinion of the
Panel on Contaminants in the Food chain, adopted on 9 June 2008. EFSA Journal 723,
1-52.
EFSA, 2008b. Marine biotoxins in shellfish - Okadaic Acid and analogues, Scientific
Opinion of the Panel on Contaminants in the Food chain, adopted on 27 November 2007.
EFSA Journal 589, 1-62.
EFSA, 2010. Marine biotoxins in shellfish - Emerging Toxins: Brevetoxin-group, Scientific
Opinion of the Panel on Contaminants in the Food chain; adopted on 5 July 2010. EFSA
Journal 8, 1-29.
Fan, L., Sun, G., Qiu, J., Ma, Q., Hess, P., and Li, A. 2014. Effect of seawater salinity on
pore-size distribution on a poly(styrene)-based HP20 resin and its adsorption of
diarrhetic shellfish toxins. Journal of Chromatography A 1373, 1-8.
Field, C. B., Behrenfeld, M. J., Randerson, J. T., and Falkowski, P. 1998. Primary production
of the biosphere: Integrating terrestrial and oceanic components. Science 281, 237-240.
Fischer, W. J., Garthwaite, I., Miles, C. O., Ross, K. M., Aggen, J. B., Chamberlin, A. R.,
Towers, N. R., Dietrich, and D. R. 2001. Congener-independent immunoassay for
microcystins and nodularins. Environmental Science and Technology 35, 4849-4856.
Fogg, G. E. 1983. The ecological significance of extracellular products of phytoplankton
photosynthesis. Botanica Marina 26, 3–14.
Froscio, S., Fanok, S., King, B., and Humpage, A. R. 2008. Screening assays for water-borne
toxicants research report 60. CRC for Water Quality and Treatment.
Froscio, S. M., Humpage, A. R., Burcham, P. C. and Falconer, I. R. 2001. Cell-free protein
synthesis inhibition assay for the cyanobacterial toxin cylindrospermopsin.
Environmental Toxicology 16: 408–412.
Fuchi, Y., Morisaki, S., Nagata, T., Shimazaki, K., Noguchi, T., Ohtomo, N., and Hashimoto,
K. 1988. Determination of tetrodotoxin in puffer fish and shellfish by high performance
liquid chromatography. Shokuhin Eiseigaku Zasshi 29, 306-312.
Funari, E., Manganelli, M., and Testai, E. 2015. Ostreospis cf. ovata blooms in coastal water:
Italian guidelines to assess and manage the risk associated to bathing waters and
recreational activities. Harmful Algae 50, 45-56.
Fux, E., Biré, R., and Hess, P. 2009. Comparative accumulation and composition of
lipophilic marine biotoxins in passive samplers and in mussels (M. edulis) on the West
Coast of Ireland. Harmful Algae 8, 523-537.
Fux, E., McMillan, D., Biré, R., and Hess, P. 2007. Development of an ultra-performance
liquid chromatography-mass spectrometry method for the detection of lipophilic marine
toxins. Journal of Chromatography A 1157, 273-280.
Garcia-Altares, M., Casanova, A., Bane, V., Diogene, J., Furey, A., and de la Iglesia, P. 2014.
Confirmation of pinnatoxins and spirolides in shellfish and passive samplers from
catalonia (Spain) by liquid chromatography coupled with triple quadrupole and high-
resolution hybrid tandem mass spectrometry. Marine Drugs 12, 3706-3732.
78
Gibble, C. M. and Kudela, R. M. 2014. Detection of persistent microcystin toxins at the land–
sea interface in Monterey Bay, California. Harmful Algae 39, 146-153.
Harada, K. I., Kondo, F., and Lawton, L. 1999. Toxic cyanobacteria in water. A guide to their
public health consequences, monitoring and management, in: Chorus, I., Bartram, J.
(Eds.). World Health Organisation, pp. 369-405.
Heijman, S. G. J., Vantieghem, M., Raktoe, S., Verberk, J. Q. J. C., and van Dijk, J. C. 2007.
Blocking of capillaries as fouling mechanism for dead-end ultrafiltration. Journal of
Membrane Science, 287(1), 119-125.
Henderson, R. K., Baker, A., Parsons, S. A., and Jefferson, B. 2008. Characterisation of
algogenic organic matter extracted from cyanobacteria, green algae and diatoms. Water
Research 42:3435-3445.
Henri, J., Leighfield, T. A., Lanceleur, R., Huguet, A., Ramsdell, J. S., and Fessard, V. 2014.
Permeability of dihydro- and cysteine-brevetoxin metabolites across a Caco-2 cell
monolayer. Harmful Algae 32, 22-26.
Henrichs, D. W., Scott, P. S., Steidinger, K. A., Errera, R. M., Abraham, A., and Campbell, L.
2013. Morphology and Phylogeny of Prorocentrum texanum sp nov (Dinophyceae): A
new toxic dinoflagellate from the Gulf of Mexico coastal waters exhibiting two distinct
morphologies. Journal of Phycology 49, 143-155.
Her, N., Amy, G., Park, H.-R., and Song, M. 2004. Characterizing algogenic organic matter
(AOM) and evaluating associated NF membrane fouling. Water Research 38, 1427–
1438.
Herman, P.H., Bredee, H.L. 1936. Principles of the mathematical treatment of constant
pressure filtration. Journal of the Society of Chemical Industry 1-4.
Hess, P., Abadie, E., Herve, F., Berteaux, T., Sechet, V., Araoz, R., Molgo, J., Zakarian, A.,
Sibat, M., Rundberget, T., Miles, C. O., and Amzil, Z. 2013. Pinnatoxin G is responsible
for atypical toxicity in mussels (Mytilus galloprovincialis) and clams (Venerupis
decussata) from Ingril, a French Mediterranean lagoon. Toxicon 75, 16-26.
Hess, P., Gallacher, S., Bates, L. A., Brown, N., and Quilliam, M. A. 2001. Determination
and confirmation of the amnesic shellfish poisoning toxin, domoic acid, in shellfish from
Scotland by liquid chromatography and mass spectrometry. Journal of the AOAC
International 84, 1657-1667.
Hess, P., McCarron, P., Krock, B., Kilcoyne, J., and Miles, C. O. 2014. Azaspiracids:
Chemistry, Biosynthesis, Metabolism, and Detection, Seafood and Freshwater Toxins.
CRC Press, Boca Raton, Florida, USA, pp. 799-822.
Hess, P., Stobo, L. A., Brown, N, A., McEvoy, J. D. G., Kennedy, G., Young, P. B., Slattery,
D., McGovern, E., McMahon, T., and Gallacher. S. 2005. LC-UV and LC-MS methods
for the determination of domoic acid. Trends in Analytical Chemistry 24, 358-367.
Hess, P., Twiner, M., Kilcoyne, J., and Sosa, S. 2015. Azaspiracid Toxins: Toxicological
Profile, in: Gopalakrishnakone, P., Haddad Jr, V., Kem, W.R., Tubaro, A., Kim, E.
(Eds.), Marine and Freshwater Toxins. Springer Netherlands, pp. 1-19.
Ho, L., Onstad, G., von Gunten, U., Rinck-Pfeiffer, S., Craig, K., and Newcombe, G. 2006.
Differences in the chlorine reactivity of four microcystin analogues, Water Research
40(6), 1200-1209.
79
Hoppenrath, M., Chomérat, N., Horiguchi, T., Schweikert, M., Nagahama, Y., and Murray, S.
2013. Taxonomy and phylogeny of the benthic Prorocentrum species (Dinophyceae)—
A proposal and review. Harmful Algae 27, 1-28.
Hu, T., Curtis, J. M., Oshima, Y., Quilliam, M. A., Walter, J. A., Watson-Whright, W. M.,
and Wright, J. L. C. 1995. Spirolides B and D, two novel macrocycles isolated from the
digestive glands of shellfish. Journal of the Chemical Society, Chemical
Communications, 2159-2161.
Huang, Y., Ortiz, L., Garcia, J., Aguirre, P., Mujeriego, R., and Bayona, J. M. 2004. Use of
headspace solid-phase microextraction to characterize odour compounds in subsurface
flow constructed wetland for wastewater treatment. Water Science and Technology 49,
89-98.
Ishida, H., Nozawa, A., Nukaya, H., Rhodes, L., McNabb, P., Holland, P. T., and Tsuji, K.
2004. Confirmation of brevetoxin metabolism in cockle, Austrovenus stutchburyi, and
greenshell mussel, Perna canaliculus, associated with New Zealand neurotoxic shellfish
poisoning, by controlled exposure to Karenia brevis culture. Toxicon 43, 701-712.
Jiang, L. Y., Eriksson, J., Lage, S., Jonasson, S., Shams, S., Mehine, M., Ilag, L. L., and
Rasmussen, U. 2014. Diatoms: A novel source for the neurotoxin BMAA in aquatic
environments. PLoS One 9.
Kerbrat, A. S., Amzil, Z., Pawlowiez, R., Golubic, S., Sibat, M., Darius, H. T., Chinain, M.,
and Laurent, D. 2011. First evidence of palytoxin and 42-hydroxy-palytoxin in the
marine cyanobacterium Trichodesmium. Marine Drugs 9, 543-560.
Kim, S. H. and Yoon, J. S. 2005. Optimization of microfiltration for seawater suffering from
red-tide contamination. Desalination 182(1–3), 315–321.
Kimura, S., Hashimoto, Y., and Yamazato, K. 1972. Toxicity of the zoanthid Palythoa
tuberculosa. Toxicon 10, 611.
Kiørboe, T. and Hansen, J. L. 1993. Phytoplankton aggregate formation: Observations of
patterns and mechanisms of cell sticking and the significance of exopolymeric
material. Journal of Plankton Research, 15(9), 993-1018.
Kleivdal, H., Kristiansen, S. I., Nilsen, M. V., Goksoyr, A., Briggs, L., Holland, P., and
McNabb, P. 2007. Determination of domoic acid toxins in shellfish by Biosense ASP
ELISA - A direct competitive enzyme-linked immunosorbent assay: Collaborative study.
Journal of the AOAC International 90, 1011-1027.
Kodama, M., Ogata, T., Noguchi, T., Maruyama, J., and Hashimoto, K. 1983. Occurrence of
saxitoxin and other toxins in the liver of the pufferfish Takifugu pardalis. Toxicon 21,
897-900.
Kohoutek, J., Maršálek, B., and Bláha, L. 2010. Evaluation of the novel passive sampler for
cyanobacterial toxins microcystins under various conditions including field sampling.
Analytical and Bioanalytical Chemistry 397, 823-828.
Krasner, S. W., Hwang, C. J., and McGuire, M. J. 1983. A standard method for quantification
of earthy-musty odorants in water, sediments, and algal cultures. Water Science and
Technology 15, 127-138.
Kudela, R. M. 2011. Characterization and deployment of Solid Phase Adsorption Toxin
Tracking (SPATT) resin for monitoring of microcystins in fresh and saltwater. Harmful
Algae 11, 117-125.
80
Kulagina, N. V., Mikulski, C. M., Gray, S., Ma, W., Doucette, G. J., Ramsdell, J. S., and
Pancrazio, J. J. 2006a. Detection of marine toxins, brevetoxin-3 and saxitoxin, in
seawater using neuronal networks. Environmental Sciemce amd Technology 40, 578-583.
Kulagina, N. V., Twiner M. J., Hess, P., McMahon, T., Satake, M., Yasumoto, T., Ramsdell,
J. S., Doucette, G. J., Ma, W., and O'Shaughnessy, T. J. 2006b. Azaspiracid-1 inhibits
bioelectrical activity of spinal cord neuronal networks. Toxicon 47, 766-773.
Ladner, D. A., Vardon, D. R., and Clark, M. M. 2010. Effects of shear on microfiltration and
ultrafiltration fouling by marine bloom-forming algae. Journal of Membrane Science
356, 33-43.
Lawrence, J., Loreal, H., Toyofuku, H., Hess, P., Iddya, K., and Ababouch, L. 2011.
Assessment and management of biotoxin risks in bivalve molluscs. FAO Fisheries and
Aquaculture Technical Paper No. 551, 337 pages.
Lawrence, J. F., Niedzwiadek, B., and Ménard, C. 2005. Quantitative determination of
paralytic shellfish poisoning toxins in shellfish using prechromatographic oxidation and
liquid chromatography with fluorescence detection: Collaborative study. Journal of the
AOAC International 88, 1714-1732.
Laycock, M. V., Anderson, D. M., Naar, J., Goodman, A., Easy, D. J., Donovan, M. A., Li, A.
F., Quilliam, M. A., Al Jamali, E., and Alshihi, R. 2012. Laboratory desalination
experiments with some algal toxins. Desalination 293, 1-6.
Laycock, M. V., Donovan, M. A., and Easy, D. J. 2010. Sensitivity of lateral flow tests to
mixtures of saxitoxins and applications to shellfish and phytoplankton monitoring.
Toxicon 55, 597-605.
Ledreux, A., Serandour, A. L., Morin, B., Derick, S., Lanceleur, R., Hamlaoui, S., Furger, C.,
Bire, R., Krys, S., Fessard, V., Troussellier, M., and Bernard, C. 2012. Collaborative
study for the detection of toxic compounds in shellfish extracts using cell-based assays.
Part II: application to shellfish extracts spiked with lipophilic marine toxins. Analytical
and Bioanalytical Chemistry 403, 1995-2007.
Ledreux, A., Thomazeau, S., Catherine, A., Duval, C., Yepremian, C., Marie, A., and
Bernard, C. 2010. Evidence for saxitoxins production by the cyanobacterium
Aphanizomenon gracile in a French recreational water body. Harmful Algae 10, 88-97.
Leppard, G. G. 1993. Organic flocs in surface waters: their native state and aggregation
behavior in relation to contaminant dispersion. In: Rao, S.S. Lewis (Ed.), Particulate
Matter and Aquatic Contaminants, Boca Raton, FL, pp. 169– 195.
Lerch, A., Uhl, W., and Gimbel, R. 2007. CFD modelling of floc transport and coating layer
build-up in single UF/MF membrane capillaries driven in inside-out mode. Water
Science and Technology: Water Supply 7(4), 37-47.
Li, S., Winters, H., Villacorte, L. O., Ekowati, Y., Abdul-Hamid, E., Kennedy, M. D., and
Amy, G. L. 2015. Compositional similarities and differences between Transparent
Exopolymer Particles (TEP) from two Marine Bacteria and two Marine Algae:
Significance to Surface Biofouling. Marine Chemistry 174, 131–140.
Litaker, R. W., Vandersea, M. W., Faust, M. A., Kibler, S. R., Nau, A. W., Holland, W. C.,
Chinain, M., Holmes, M. J., and Tester, P. A. 2010. Global distribution of ciguatera
causing dinoflagellates in the genus Gambierdiscus. Toxicon, 56(5), pp.711-730.
81
MacKenzie, L., Beuzenberg, V., Holland, P., McNabb, P., and Selwood, A. 2004. Solid
phase adsorption toxin tracking (SPATT): a new monitoring tool that simulates the
biotoxin contamination of filter feeding bivalves. Toxicon 44, 901-918.
Mafra, Jr, L. L., Léger, C., Bates, S. S., and Quilliam, M. A. 2009. Analysis of trace levels of
domoic acid in seawater and plankton by liquid chromatography without derivatization,
using UV or mass spectrometry detection. Journal of Chromatography A 1216, 6003-
6011.
McCarron, P., Burrell, S., and Hess, P. 2007. Effect of addition of antibiotics and an
antioxidant on the stability of tissue reference materials for domoic acid, the amnesic
shellfish poison. Analytical and Bioanalytical Chemistry 387, 2495-2502.
McLeod, C., Burrell, S., and Holland, P. 2015. Review of the currently available field
methods for detection of marine biotoxins in shellfish flesh. Report FS102086 UK-FSA
Crown Copyright, website access.
McMahon, T. and Silke, J. 1996. West coast of Ireland; winter toxicity of unknown aetiology
in mussels. Harmful Algae News 14, 2.
Méjean, A., Peyraud-Thomas, C., Kerbrat, A. S., Golubic, S., Pauillac, S., Chinain, M., and
Laurent, D. 2010. First identification of the neurotoxin homoanatoxin-a from mats of
Hydrocoleum lyngbyaceum (marine cyanobacterium) possibly linked to giant clam
poisoning in New Caledonia. Toxicon 56, 829-835.
Meriluoto, J., and Codd, G. A. 2005. Toxic cyanobacterial monitoring and cyanotoxin
analysis. Åbo Akademi University Press, Turku, Finland.
Metcalf, J. S., Beattie, K. A., Saker, M. L., and Codd, G. A. 2002 Effects of organic solvents
on the high performance liquid chromatographic analysis of the cyanobacterial toxin
cylindrospermopsin and its recovery from environmental eutrophic waters by solid phase
extraction. FEMS Microbiology Letters 216(2), 159.
Miles, C. O., Sandvik, M., Haande, S., Nonga, H., and Ballot, A. 2013a. LC-MS analysis
with thiol derivatization to differentiate [Dhb(7)]- from [Mdha(7)]-Microcystins:
Analysis of cyanobacterial blooms, planktothrix cultures and european crayfish from
Lake Steinsfjorden, Norway. Environmental Science and Technology 47, 4080-4087.
Miles, C. O., Sandvik, M., Nonga, H. E., Rundberget, T., Wilkins, A. L., Rise, F., Ballot, and
A. 2012. Thiol derivatization for LC-MS identification of Microcystins in complex
matrices. Environmental Science and Technology 46, 8937-8944.
Miles, C. O., Sandvik, M., Nonga, H. E., Rundberget, T., Wilkins, A. L., Rise, F., and Ballot,
A. 2013b. Identification of microcystins in a Lake Victoria cyanobacterial bloom using
LC–MS with thiol derivatization. Toxicon 70, 21-31.
Miles, C. O., Wilkins, A. L., Stirling, D. J., and MacKenzie, A. L. 2003. Gymnodimine C, an
isomer of gymnodimine B, from Karenia selliformis. Jouranl of Agricultural and Food
Chemistry 51, 4838-4840.
Miller, M. A., Kudela, R. M., Mekebri, A., Crane, D., Oates, S. C., Tinker, M. T., Staedler,
M., Miller, W. A., Toy-Choutka, S., Dominik, C., Hardin, D., Langlois, G. W., Murray,
M., Ward, K., and Jessup, D. A. 2010. Evidence for a novel marine harmful algal bloom:
cyanotoxin (Microcystin) transfer from land to sea otters. PLoS One 5, e12576.
Mingazzini, M. and Thake, B. 1995. Summary and conclusions of the workshop on marine
mucilages in the Adriatic Sea and elsewhere. Science of the Total Environment, 165:9-14.
82
Mopper, K., Zhou, J., Sri Ramana, K., Passow, U., Dam, H. G., and Drapeau, D. T. 1995.
The role of surface-active carbohydrates in the flocculation of a diatom bloom in a
mesocosm. Deep-Sea Research Part II, 42(1), 47-73.
Morais, J., Augusto, M., Carvalho, A. P., Vale, M., and Vasconcelos, V. M. 2008.
Cyanobacteria hepatotoxins, microcystins: bioavailability in contaminated mussels
exposed to different environmental conditions. European Food Research and
Murakami, Y., Oshima, Y., and Yasumoto, T. 1982. Identification of okadaic acid as a toxic
component of a marine dinoflagellate Prorocentrum lima. Nippon Suisan Gakkaishi 48,
69-72.
Murata, M., Shimatani, M., Sugitani, H., and Oshima, Y. T. Y. 1982. Isolation and structural
elucidation of the causative toxin of the diarrhetic shellfish poisoning. Bulletin of the
Japanese Society for the Science of Fish 48, 549-552.
Myklestad, S. M. 1995. Release of extracellular products by phytoplankton with special
emphasis on polysaccharides. Science of the Total Environment 165, 155-164.
Nicholson, B. and Burch, M. 2001. Evaluation of analytical methods for the detection and
quantification of cyanotoxins in relation to Australian drinking water guidelines.
NHMRC, Water Services Association of Australia and the Cooperative Research Centre
for Water Quality and Treatment report, National Health and Medical Research Council
of Australia, Canberra, Australia.
Onuma, Y., Satake, M., Ukena, T., Roux, J., Chanteau, S., Rasolofonirina, N., Ratsimaloto,
M., Naoki, H., and Yasumoto, T. 1999. Identification of putative palytoxin as the cause
of clupeotoxism. Toxicon 37, 55-65.
Panglisch, S. 2003. Formation and prevention of hardly removable particle layers in inside-
out capillary membranes operating in dead-end mode. Water Science and Technology:
Water Supply 3 (5-6), 117-124.
Passow, U. 2000. Formation of transparent exopolymer particles (TEP) from dissolved
precursor material. Marine Ecology Progress Series 192, 1-11.
Passow, U. 2002. Transparent exopolymer particles (TEP) in aquatic environments. Progress
in Oceanography 55(3), 287-333.
Petry, M., Sanz, M. A., Langlais, C., Bonnelye, V., Durand, J.-P., Guevara, D., Nardes, W.
M., and Saemi, C. H. 2007. The El Coloso (Chile) reverse osmosis plant. Desalination
203(1-3), 141-152.
Plakas, S. M., Jester, E. L. E., El Said, K. R., Granade, H. R., Abraham, A., Dickey, R. W.,
Scott, P. S., Flewelling, L. J., Henry, M. S., Blum, P., and Pierce, R. H. 2008.
Monitoring of brevetoxins in the Karenia brevis bloom-exposed Eastern oyster
(Crassostrea virginica). Toxicon 52, 32-38.
Polak, E. H. and Provasi, J. 1992. Odor sensitivity to geosmin enantiomers. Chemical Senses
17, 23-26.
Qu, F., Liang, H., He, J., Ma, J., Wang, Z., Yu, H., and Li, G. 2012. Characterization of
dissolved extracellular organic matter (dEOM) and bound extracellular organic matter
(bEOM) of Microcystis aeruginosa and their impacts on UF membrane fouling. Water
Research 46, 2881-2890.
QUASIMEME, 2015. http://www.quasimeme.org/,. website accessed 31 December 2015.
83
Quilliam, M. A., Hess, P., and Dell'Aversano, C. 2001. Recent developments in the analysis
of phycotoxins by liquid chromatography-mass spectrometry, Mycotoxins and
Phycotoxins in Perspective at the Turn of the Century, pp. 383-391.
Quilliam, M. A., Xie, M., and Hardstaff, W. R. 1995. A rapid extraction and cleanup
procedure for the liquid chromatographic determination of domoic acid in unsalted
seafood. Journal of AOAC International 78, 543-554.
Reguera, B., Velo-Suárez, L., Raine, R., and Park, M. G. 2012. Harmful Dinophysis species:
A review. Harmful Algae 14, 87-106.
Réveillon, D., Abadie, E., Séchet, V., Masseret, E., Hess, P., and Amzil, Z. 2015. β-N-
methylamino-l-alanine (BMAA) and isomers: Distribution in different food web
compartments of Thau lagoon, French Mediterranean Sea. Marine Environmental
Research 110, 8-18.
Rhodes, L., Smith, K., Selwood, A. I., McNabb, P., Molenaar, S., Munday, R., Wilkinson, C.,
and Hallegraeff, G. M. 2011. Production of pinnatoxins E, F and G by scrippsielloid
dinoflagellates isolated from Franklin Harbour, South Australia. New Zealand Journal of
Marine and Freshwater Research 45, 703-709.
Rhodes, L., Smith, K., Selwood, A. I., McNabb, P., van Ginkel, R., Holland, P. T., and
Munday, R. 2010. Production of pinnatoxins by a peridinoid dinoflagellate isolated from
Northland, New Zealand. Harmful Algae 9, 384-389.
catastrophic 2008–2009 red tide in the Arabian Gulf region, with observations on the
identification and phylogeny of the fish-killing dinoflagellate Cochlodinium
polykrikoides. Harmful Algae 9(2), pp.163-172.
Ripperger, S., Gösele, W., and Alt, C. 2012. Filtration, 1. Fundamentals. Ullmann's
Encyclopedia of Industrial Chemistry. pp. 677–709.
Rossini, G. P. and Hess, P. 2010. Phycotoxins: chemistry, mechanisms of action and shellfish
poisoning. Exs 100, 65-122.
Rundberget, T., Gustad, E., Samdal, I. A., Sandvik, M., and Miles, C. O. 2009. A convenient
and cost-effective method for monitoring marine algal toxins with passive samplers.
Toxicon 53, 543-550.
Samdal, I. A., Ballot, A., Lovberg, K. E., Miles, C. O. 2014. Multihapten approach leading to
a sensitive ELISA with broad cross-reactivity to microcystins and nodularin.
Environmental Science and Technology 48, 8035-8043.
Samdal, I. A., Lovberg, K. E., Briggs, L. R., Kilcoyne, J., Xu, J., Forsyth, C. J., and Miles, C.
O. 2015. Development of an ELISA for the Detection of Azaspiracids. Journal of
Agriculture and Food Chemistry 63, 7855-7861.
Satake, M., Ofuji, K., James, K., Furey, A., and Yasumoto, T. 1998a. New toxic event caused
by Irish mussels, in: Reguera, B., Blanco, J., Fernandez, M.L., Wyatt, T. (Eds.), Harmful
Algae. Xunta de Galicia and Intergovernmental Oceanographic Commission of
UNESCO pp. 468-469.
Schock, T. B., Huncik, K., Beauchesne, K. R., Villareal, T. A., and Moeller, P. D. R. 2011.
Identification of Trichotoxin, a novel chlorinated compound associated with the bloom
forming cyanobacterium, Trichodesmium thiebautii. Environmental Science and
84
Schurer, R., Tabatabai, A., Villacorte, L., Schippers, J. C., and Kennedy, M. D. 2013. Three
years operational experience with ultrafiltration as SWRO pretreatment during algal
bloom. Desalination and Water Treatment 51 (4-6), 1034-1042.
Seki, T., Satake, M., MacKenzie, A. L., Kaspar, H. F., and Yasumoto, T. 1995.
Gymnodimine, a new marine toxin of unprecedented structure isolated from New
Zealand oysters and the dinoflagellate, Gymnodinium sp. Tetrahedron Letters 36, 7093-
7096.
Selwood, A. I., Miles, C. O., Wilkins, A. L., van Ginkel, R., Munday, R., Rise, F., and
McNabb, P. 2010. Isolation, structural determination and acute toxicity of pinnatoxins E,
F and G. Journal of Agriculture and Food Chemistry 58, 6532-6542.
Serres, M. H., Fladmark, K. E., and Doskeland, S.O. 2000. An ultrasensitive competitive
binding assay for the detection of toxins affecting protein phosphatases. Toxicon 38,
347-360.
Shea, D., Tester, P., Cohen, J., Kibler, S., and Varnam, S. 2006. Accumulation of brevetoxins
by passive sampling devices. African Journal of Marine Science 28, 379-381.
Spikes, J. J., Ray, S. M., Aldrich, D. V., Nash, J. B. 1968. Toxicity variations of
Gymnodinium breve cultures. Toxicon 5, 171-174.
Suurnäkki, S., Gomez-Saez, G. V., Rantala-Ylinen, A., Jokela, J., Fewer, D. P., and Sivonen,
K. 2015. Identification of geosmin and 2-methylisoborneol in cyanobacteria and
molecular detection methods for the producers of these compounds. Water Research 68,
56-66.
Tachibana, K., Scheuer, P. J., Tsukitani, Y., Kikuchi, H., Van Engen, D., Clardy, J.,
Gopichand, Y., and Schmitz, F. J. 1981. Okadaic acid, a cytotoxic polyether from two
marine sponges of the genus Halichondria. Journal of the American Chemical Society
103, 2469-2471.
Takada, N., Iwatsuki, M., Suenaga, K., and Uemura, D. 2000. Pinnamine, an alkaloidal
marine toxin, isolated from Pinna muricata. Tetrahedron Letters 41, 6425-6428.
Tartaglione, L., Dell’Aversano, C., Mazzeo, A., Forino, M., Wieringa, A., and Ciminiello, P.
2015. Determination of palytoxins in soft coral and seawater from a home aquarium.
comparison between Palythoa- and Ostreopsis-related inhalatory poisonings.
Environmental Science and Technology 50(2):1023-30.
These, A., Klemm, C., Nausch, I., and Uhlig, S. 2011. Results of a European interlaboratory
method validation study for the quantitative determination of lipophilic marine biotoxins
in raw and cooked shellfish based on high-performance liquid chromatography–tandem
mass spectrometry. Part I: collaborative study. Analytical and Bioanalytical Chemistry
399, 1245-1256.
Tillmann, U., Salas, R., Jauffrais, T., Hess, P., and Silke, J. 2014. AZA: The producing
organisms - Biology and trophic transfer, in: Botana., L.M. (Ed.), Seafood and
Freshwater Toxins. CRC Press, Boca Raton, Florida, USA, pp. 773-798.
Trainer, V. L., Wells, M. L., Cochlan, W. P., Trick, C. G., Bill, B. D., Baugh, K. A., Beall, B.
F., Herndon, J., and Lundholm, N. 2009. An ecological study of a massive bloom of
toxigenic Pseudo-nitzschia cuspidata off the Washington state coast. Limnology and
Oceanography 54, 1461-1474.
85
Tubaro, A., Florio, C., Luxich, E., Sosa, S., DellaLoggia, R., and Yasumoto, T. 1996. A
protein phosphatase 2A inhibition assay for a fast and sensitive assessment of okadaic
acid contamination in mussels. Toxicon 34, 743-752.
Turner, A. D., Powell, A., Schofield, A., Lees, D. N., and Baker-Austin, C. 2015. Detection
of the pufferfish toxin tetrodotoxin in European bivalves, England, 2013 to 2014.
Eurosurveillance 20, 2-8.
Twiner, M.J., Hess, P., and Doucette, G. J. 2014. Azaspiracids: Toxicology, Pharmacologiy,
and Risk Assessment, Seafood and Freshwater Toxins. CRC Press, Boca Raton, Florida,
USA, pp. 823-856.
Uemura, D., Chou, T., Haino, T., Nagatsu, A., Fukuzawa, S., Zheng, S., and Chen, H. 1995.
Pinnatoxin A: a toxic amphoteric macrocycle from the Okinawan bivalve Pinna
muricata. Journal of the American Chemical Society 117, 1155-1156.
Usami, M., Satake, M., Ishida, S., Inoue, A., Kan, Y., and Yasumoto, T. 1995. Palytoxin
analogs from the dinoflagellate Ostreopsis siamensis. Journal of the American Chemical
Society 117, 5389-5390.
Vale, P., Gomes, S. S., Lameiras, J., Rodrigues, S. M., Botelho, M. J., and Laycock, M. V.
2009. Assessment of a new lateral flow immunochromatographic (LFIC) assay for the
okadaic acid group of toxins using naturally contaminated bivalve shellfish from the
Portuguese coast. Food Additives and Contaminants Part A-Chem. 26, 214-220.
van den Top, H. J., Gerssen, A., McCarron, P., and van Egmond, H. P. 2011. Quantitative
determination of marine lipophilic toxins in mussels, oysters and cockles using liquid
chromatography-mass spectrometry: inter-laboratory validation study. Food Addit
Contam Part A Chem Anal Control Expo Risk Assess 28, 1745-1757.
van Dolah F. M., Fire S. E., Leighfield T. A., Mikulski C. M., and Doucette G. J. 2012.
Determination of paralytic shellfish toxins in shellfish by receptor binding assay:
collaborative study. Journal of AOAC Internationa. 95, 795-812.
Van Wagoner, R. M., Misner, I., Tomas, C., and Wright, J. L. C. 2011. Occurrence of 12-
methylgymnodimine in a spirolide-producing dinoflagellate Alexandrium peruvianum
and the biogenetic implications. Tetrahedron Letters 52, 4243-4246.
Vasconcelos, V. M., 1995. Uptake and depuration of the heptapeptide toxin microcystin-LR
in Mytilus galloprovincialis. Aquatic Toxicology 32, 227-237.
Vasconcelos, V. M. 1999. Cyanobacterial toxins in Portugal: effects on aquatic animals and
risk for human health. Brazilian Journal of Medical and Biological Research 32, 249-
254.
Verdugo, P., Alldredge, A. L., Azam, F., Kirchman, D. L., Passow, U., and Santschi, P. H.
2004. The oceanic gel phase: a bridge in the DOM–POM continuum. Marine Chemistry
92, 67-85.
Villacorte, L. O. 2014. Algal blooms and membrane-based desalination technology. ISBN
978-1-138-02626-1, CRC Press/Balkema, Leiden.
Villacorte, L. O., Ekowati, Y., Neu, T. R., Kleijn, J. M., Winters, H., Amy, G., Schippers J.
C., and Kennedy M. D. 2015a. Characterisation of algal organic matter produced by
bloom forming marine and freshwater algae. Water Research 73, 216–230.
86
Villacorte, L. O., Ekowati, Y., Winters, H., Amy, G., Schippers, J. C., and Kennedy, M. D.
2015b. MF/UF rejection and fouling potential of algal organic matter from bloom-
forming marine and freshwater algae. Desalination 367, 1-10.
Villacorte, L. O., Kennedy, M. D., Amy, G. L., and Schippers, J. C. 2009. The fate of
Transparent Exopolymer Particles (TEP) in integrated membrane systems: Removal
through pretreatment processes and deposition on reverse osmosis membranes. Water
Research 43 (20), 5039-5052.
Voutchkov, N. 2010. Considerations for selection of seawater filtration pretreatment
system. Desalination 261 (3), 354-364.
Vlamis, A., Katikou, P., Rodriguez, I., Rey, V., Alfonso, A., Papazachariou, A., Zacharaki, T.,
Botana, A. M., and Botana, L. M. 2015. First detection of Tetrodotoxin in greek shellfish
by UPLC-MS/MS potentially linked to the presence of the dinoflagellate Prorocentrum
minimum. Toxins 7, 1779-1807.
Walter, J. A., Leek, D. M., and Falk, M. 1992. NMR study of the protonation of domoic acid.
Canadian Journal of Chemistry 70, 1156-1161.
Wang, Z., King, K. L., Ramsdell, J. S., and Doucette, G. J. 2007a. Determination of domoic
acid in seawater and phytoplankton by liquid chromatography-tandem mass
spectrometry. Journal of Chromatography A 1163, 169-176.
Wang, J., Pang, X., Ge, F., and Ma, Z. 2007b. An ultra-performance liquid chromatography–
tandem mass spectrometry method for determination of microcystins occurrence in
surface water in Zhejiang Province, China. Toxicon 49, 1120-1128.
Wiese, M., D'Agostino, P. M., Mihali, T. K., Moffitt, M. C., and Neilan, B. A. 2010.
Neurotoxic alkaloids: Saxitoxin and its analogs. Marine Drugs 8, 2185-2211.
Wonnacott, S. and Gallagher, T. 2006. The chemistry and pharmacology of anatoxin-a and
related homotropanes with respect to nicotinic acetylcholine receptors. Marine Drugs 4,
228-254.
Yasumoto, T., Oshima, Y., and Yamaguchi, M. 1978. Occurrence of a new type of shellfish
poisoning in the Tohoku district. Bulletin of the Japanese Society for the Science of Fish,
1249-1255.
Yogi, K., Sakugawa, S., Oshiro, N., Ikehara, T., Sugiyama, K., and Yasumoto, T. 2014.
Determination of toxins involved in ciguatera fish poisoning in the Pacific by LC/MS.
Journal of AOAC International 97, 398-403.
Zendong, Z., Abadie, E., Mazzeo, A., Hervé, F., Herrenknecht, C., Amzil, Z., Dell’Aversano,
C., and Hess, P. 2015a. Determination of the concentration of dissolved lipophilic algal
toxins in seawater using pre-concentration with HP-20 resin and LC-MS/MS detection,
in: MacKenzie, L. (Ed.), 16th International Conference on Harmful Algae 27th-31st
October 2014. Cawthron Institute, Nelson, New Zealand and International Society for
the Study of Harmful Algae, Wellington, New Zealand.
Zendong, Z., Herrenknecht, C., Abadie, E., Brissard, C., Tixier, C., Mondeguer, F., Séchet,
V., Amzil, Z., and Hess, P. 2014. Extended evaluation of polymeric and lipophilic
sorbents for passive sampling of marine toxins. Toxicon 91, 57-68.
87
Zendong, Z., McCarron, P., Herrenknecht, C., Sibat, M., Amzil, Z., Cole, R. B., and Hess, P.
2015b. High resolution mass spectrometry for quantitative analysis and untargeted
screening of algal toxins in mussels and passive samplers. Journal of Chromatography A
1416, 10-21.
Zheng, S. Z., Huang, F. L., Chen, S. C., Tan, X. F., Zuo, J. B., Peng, J., Xie, R. W. 1990. The
isolation and bioactivities of pinnatoxin. Chinese Journal of Marine Drugs 33, 33-35.
88
Observing systems for early detection of HABs
3 DESIGNING AN OBSERVING SYSTEM FOR EARLY DETECTION

OF HARMFUL ALGAL BLOOMS
Bengt Karlson1, Clarissa R. Anderson2, Kathryn J. Coyne3, Kevin G. Sellner4, and Donald M.
Anderson5
1
2
3
University of Delaware, Lewes, DE USA
4
Chesapeake Research Consortium, Edgewater, MD USA
5

3.1 Introduction......................................................................................................................................... 89
3.2 Designing an observation system ....................................................................................................... 90
3.3 Background information ..................................................................................................................... 90
3.3.1 Characterizing the physical and chemical environment ................................................................ 91
3.3.2 Characterizing phytoplankton community composition ................................................................ 91
3.4 Identifying existing infrastructure ...................................................................................................... 93
3.5 Sampling methods .............................................................................................................................. 93
3.5.1 Sampling from shore or from vessels ............................................................................................ 93
3.5.1.1 Fixation procedures for plankton samples ........................................................................... 96
3.5.2 Water transparency – Turbidity tubes and Secchi discs ................................................................ 97
3.5.3 Chlorophyll-a and other photosynthetic pigment .......................................................................... 98
3.5.3.1 In vivo and in situ chlorophyll fluorescence ......................................................................... 98
3.5.4 Automated water sampling ............................................................................................................ 99
3.5.5 Sampling using fixed platforms ................................................................................................... 100
3.5.6 Ships of opportunity and Ferrybox systems................................................................................. 101
3.5.7 Flow-through systems on land or on buoys ................................................................................. 102
3.5.8 In-water optical instrumentation for detecting HABs .................................................................. 102
3.6 Identification and enumeration of HAB organisms .......................................................................... 102
3.6.1 Essential information for identification phytoplankton ............................................................... 103
3.6.1.1 Books for identification of harmful algae and phytoplankton ........................................... 103
3.6.1.2 Web sites with information on harmful algae and phytoplankton ..................................... 104
3.6.2 Light microscopy ......................................................................................................................... 104
3.6.3 Fluorescence microscopy ............................................................................................................. 105
3.6.4 Electron microscopy .................................................................................................................... 105
3.6.5 Imaging flow cytometry ............................................................................................................... 106
3.6.6 Molecular techniques ................................................................................................................... 107
3.7 Satellite remote sensing .................................................................................................................... 108
3.8 Transport and delivery of harmful algal blooms .............................................................................. 108
3.8.1 Empirical and numerical models ................................................................................................. 108
3.8.2 High-resolution circulation models to resolve flow near intakes ................................................ 109
3.8.3 An example of a regional HAB forecast system .......................................................................... 110
3.9 Distributing warnings and information ............................................................................................. 110
3.10 Data storage and distribution ............................................................................................................ 112
3.11 Facilities, equipment, and personnel ................................................................................................ 112
3.12 Summary ........................................................................................................................................... 114
3.13 References......................................................................................................................................... 115
3.1 INTRODUCTION
Harmful algal blooms (HABs) are a serious and growing threat to many desalination plants. It
is therefore important to limit the impact from HABs by preventing blooms from reaching
seawater reverse osmosis (SWRO) plants in the first place, while also mitigating their effects
through pretreatment and other actions within the plant once intake has occurred.
89
In this chapter, traditional and emerging technologies in the field of HAB detection and
monitoring are summarized. Also advice on designing “observing systems” for early
detection or characterization of algal blooms is provided. These systems will vary
dramatically in terms of the number of parameters to be measured, the number of stations,
frequency of sampling and instruments used - all determined by desalination plant budgets
and personnel skills, the nature of the HAB threat for a given plant or region, and other such
considerations. An observing system might be as simple as visual observations of the color or
nature of the intake water, or as complex as a moored array of autonomous sensors outside
the plant, or weekly surveys from small vessels to determine what algal species and blooms
are in the intake area or surrounding waters, and thus likely to impact the plant.
There are a number of factors that complicate the design of an observing system. One is the
diversity of HAB species. Potentially harmful phytoplankton are found in many groups
(mainly eukaryotes) such as dinoflagellates, raphidophytes, diatoms, euglenophytes,
cryptophytes, haptophytes, pelagophytes, and chlorophytes (see Chapter 1), but prokaryotes,
(cyanobacteria) are also a concern. While dinoflagellates comprise the majority of toxic HAB
species in the marine environment where desalination plants are located, many of the toxic
species that pose a threat to drinking water supply in fresh- or brackish-water systems are
cyanobacteria.
A second factor is that phytoplankton distribution in the sea is not uniform vertically or
horizontally in space or in time. This is termed “patchiness” and results from the interaction
between physical and biological processes. Examples are presented later in this chapter. The
simultaneous use of multiple monitoring methods is therefore often necessary to characterize
the species composition and extent of blooms, but even then, a full picture of the distribution
of a HAB may not be achievable.
3.2 DESIGNING AN OBSERVATION SYSTEM
In the context of providing observations of the water and plankton that can guide desalination
operations and plant siting, a HAB observing system can be very informative in many
locations. The main goal of such a system is to provide information for actions (rapid
response) to avoid or minimize operational disruptions and damage to desalination plants.
Prior to the design and construction of a plant, a HAB observing system can be used to gather
information on the nature and function of the regional oceanographic system, its role in HAB
occurrence, and the historical patterns and extent of HAB events. This can be used to provide
input on where to place, and how to design, water intake systems, as well as highlighting the
types of pretreatment equipment that might be needed in order to minimize damage from
HABs during operation. Figures 3.1 – 3.3 show some features to be considered in this regard.
3.3 BACKGROUND INFORMATION
To design an observation system, it is necessary to gather background information on the

occurrence of harmful algae in relation to local physical and chemical conditions. Existing
information should be used when possible (likely from monitoring programs run by
government, industry, or academic institutions), but often a pilot study may be required. A
physical oceanographic model describing regional hydrodynamics surrounding the
desalination facility would be a valued asset to any observing system. These are often
developed by universities and other academic institutes, as well as government agencies, and
are sometimes utilized in studies of brine dispersion and recirculation during plant design.
Models in HAB monitoring and management are discussed in section 3.8.1.
90
3.3.1 Characterizing the physical and chemical

environment
The geographic position and depth of the water inlet
of a desalination plant is one factor that will
influence the design of the observing system. Local
and larger scale current conditions as well as
seasonal fluctuations in water column stratification
are important physical parameters to assess or
monitor. Questions about the physical and chemical
environment that should be considered during the
design of an observing system include: do the HABs
develop locally or do currents transport them to the
area (Figure 3.1)? What are the dominant sources of
nutrients available for local algal growth – pollution
discharges from nearby population centers for
Figure 3.1. A schematic drawing example, or natural sources through the circulation of
illustrating the effect of local currents on water masses? And how dynamic is the hydrographic
the selection of sampling locations. In the
system outside the plant – are water masses and their
yellow rectangle, tides dominate the
currents on a 24-hour cycle. Water from associated blooms moving rapidly along the coast, or
locations 3, 4, and 5 all pass location 4. is it a more gradual and constant flow? These and
The northward current dominates over other example questions that need to be answered
longer time scales, and since blooms before an observing system is designed are listed in
develop in the South, observations at
station 1 may be of great importance.
Table 3.1.
Distributions of currents and circulation patterns in
the area of the facility should be determined, and if
possible, models used to estimate particle delivery to
the plant’s intake under normal weather patterns and
during/following major meteorological events.
Bottom sediment type and depths relative to the
facility intake location should also be known to
minimize bottom-derived sediment intake but also to
assess the potential for blooms derived from
resuspension of HAB cysts or spores up-current of
Figure 3.2. This schematic shows a the plant.
strongly stratified water column. Water
intakes at two different depths are depicted 3.3.2 Characterizing phytoplankton community
for the desalination plant. Bloom A in near composition
surface water is reaching the plant through
the black intake while subsurface blooms The phytoplankton community in a given region
(bloom B) can be taken in through the red,often consists of hundreds of different species, with
deeper intake. that community composition changing through time.
Only a fraction of these species are potentially
harmful. When a HAB organism reaches high biomass levels and becomes the main species
present, it can cause problems due to that biomass, but for some species, a relatively low
number of cells can still present operational concerns for a desalination plant if toxins
deleterious to health are produced. The amount of toxin that might be present in blooms of
different sizes is discussed in Chapter 1, and Chapter 10 evaluates the risk associated with the
small amounts of residual toxins that might be present after desalination has occurred. HAB
91
Table 3.1. Questions about the physical and chemical environment that should be considered
when designing a HAB observation system.
Questions Data needed
What distance may HABs be transported Local and regional data on current speed
during a tidal cycle? and direction at depths where HABs occur.
Are there currents transporting HABs to the Data from in situ instruments, e.g. an ADCP
location of the water intake of the (Acoustic Doppler Current Profiler).
desalination plant? Simulations from a physical oceanographic
model developed and verified for the area
are also useful.
Is the water stratified during the whole year Depth profiles of salinity and temperature
or part of the year? together with measurements of chlorophyll
fluorescence, a proxy for phytoplankton
biomass. Measurements are usually made
from research vessels using a CTD, an
instrument used to determine depth profiles
of conductivity and temperature together
with other parameters. Conductivity and
temperature are used to calculate salinity
(calculated using the practical salinity scale).
Also moored depth profiling platforms are
available providing information on
subsurface algal blooms in near real time.
Are there short-term events that may favor Background data on air temperature,
HAB-development? precipitation, river flow, wind speed and
direction, cloud cover. A meteorological and
hydrological institute may provide the data.
Are there nutrients supporting HAB growth Data on concentrations of inorganic

available during the whole year or part of the nutrients, i.e. phosphate, silicate, nitrate and
year? ammonium from the surface mixed layer.
Water sampling and chemical analysis in a
laboratory on ship or on land should be
carried out by laboratories specialized in
saline samples. The samples do not preserve
well and should be analyzed within a few
hours after collection or frozen for later
analysis. Also, riverine input of nutrients is
important.
92
toxins are sufficiently well removed that toxins in treated

water should not be a concern, but they are an operational
issue that should be identified and monitored when such a
threat is present.
Since many of the non-harmful and harmful algal species
are similar in appearance (morphology) it is necessary to
have trained personnel identifying the organisms. Semi-
automated systems exist, as described below, but staff or
outside experts with knowledge of the instruments and
phytoplankton taxonomy are needed to set up the systems
and evaluate the results. Other information on the local
phytoplankton community can often be obtained from
existing monitoring programs in a region, perhaps
conducted by a state, province, or municipality. Some of
Figure 3.3. Typical vertical
distribution of algae in the water the questions and required data relative to phytoplankton
column. The graph illustrates a population dynamics are listed in Table 3.2.
common vertical distribution with a
sub-surface maximum of chloro-
3.4 IDENTIFYING EXISTING INFRASTRUCTURE
phyll or phytoplankton biomass In many cases, existing sampling infrastructure may be
(often 10-20 m deep), which may
used when setting up a HAB observation system. An
move vertically to the surface and
back (termed migration), depending oceanographic laboratory with facilities for working with
on the day-night cycle. Water phytoplankton nearby the sampling sites is ideal. If there
sampling and automated are already on-going marine monitoring programs, they
observations should account for this can be adapted to undertake HAB work. Ships of
heterogeneity. opportunity, e.g. ferries, with a stable timetable, may be
used for automated sampling. It is useful to investigate if there are buoys or permanent
structures (e.g. pilings) in the area that can be used for mounting automated sensors and water
sampling devices. Although existing buoys may not be available for mounting sensors and
water sampling devices, colocation of HAB buoys is useful to avoid problems with fishing
and traffic of merchant vessels. These approaches are described in more detail below.
3.5 SAMPLING METHODS
3.5.1 Sampling from shore or from vessels

The most basic sampling procedure for observing algal bloom species that cause problems for
desalinations plants is to collect a water sample and analyze it with a microscope. This can
complement online, continuous analyses, such as chlorophyll fluorescence, discussed below.
The recommended frequency for sampling is once per week, as phytoplankton can grow and
accumulate very rapidly (some can double their cell concentrations in a day or less). If
resources are limited, bi-weekly sampling can provide reasonable protection. The number of
sampling locations and the depth of sampling will depend on the local conditions and
available funding and staff. One approach is to sample at the seawater intake, but that gives
little advance notice or information about the geographic extent of the bloom. At the other
extreme, ship-based surveys can be conducted, covering an area several km or more from the
intake. During a HAB event, increased sampling frequency and spatial coverage can be very
informative, as it can reveal the spatial extent of a bloom, its vertical distribution, and other
factors that can help the plant anticipate future impacts and potential treatments.
Nearshore samples can be taken from land, but the sampling point must be before the waves
break. A jetty, dock, or other extended feature can be used for that purpose. Both dedicated
ships, i.e. research vessels, and other boats may be used. Ships should be of a size suitable for
93
work in rough weather and have room for both the crew and technical personnel. For this
type of nearshore sampling, the ships should be fitted with CTD (conductivity, temperature,
depth profilers) or other such devices to measure water column structure. The CTD should
include a sensor for chlorophyll fluorescence if possible (see Table 3.1). A laboratory area on
the ship is important for filtering of samples and other activities.
Table 3.2. Questions regarding HAB species and bloom dynamics that need to be addressed
when designing an observation system. Some of the information is likely available in other
institutions and should be explored prior to initiating the observing system.
Question Data needed
Which HAB species occur in the area? Abundance and distribution of phytoplankton, in
What is the temporal and spatial distribution of general, and of HAB species, in particular.
HAB species? Frequent (e.g. weekly) water sampling and
microscopy-based analyses of the samples by
Do HABs develop upstream of the desalination skilled personnel. In addition, automated
plant? analyses using imaging flow cytometry and/or
During what time of year do the HABs occur? genetic methods can be useful. Surveys including
What is the background composition of the water sampling at multiple locations are needed
phytoplankton community in the area? to document the spatial distribution of HAB
species. Data on current speed and direction
support the design of the surveys.
What are the ecological and bloom dynamics of Long-term observations and experimental work
the local HAB species? are needed to characterize the ecology of HAB
Can the HAB species regulate their position in species. This may be outside the scope of the
the water column? observation system, but some observations (e.g.,
vertical swimming behavior) are relatively
What is the growth rate of the HAB species? simple to make, and are important for minimizing
HAB intake.
Do the HAB-species produce resting stages? Resting stages (cysts or spores) should be
What is the distribution of these? documented from observations or the scientific
literature. If a common HAB-organism in the
region produces resting stages, a distributional
survey may be useful, as this can guide
understanding of the timing and location of
blooms.
Do the HAB species produce toxins that may Toxins produced by the species in the area. Field
cause health problems for humans? samples of phytoplankton should be analyzed for
toxin content using methods described in Chapter
2. Once the local HAB species are identified,
known toxin profiles are likely available in the
scientific literature.
Are there local nutrients to support HABs? Concentrations of inorganic nutrients (see Table
1). This information can help explain the
frequency and size of HABs in the area.
94
Water samples should be collected for laboratory analysis of phytoplankton and chlorophyll -
a, and if possible also for inorganic nutrients, oxygen, and other parameters. In waters
beyond the intake area, the focus should be on the surface (0 - 1m), mixed layer, as described
in Chapter 1. Fixed depth sampling (e.g. 0, 10, 20 m, and a near sea floor sample,) can be
informative, but this will depend on the local conditions, and whether there is a need for that
degree of vertical resolution. Otherwise, a surface sample is all that is needed. If a more
comprehensive measurement is needed that accounts for vertically migrating cells, water
samples can be collected from individual depths using Niskin bottles, and these can then be
pooled for later phytoplankton analyses, with one count to characterize the entire water
column or mixed layer. Integrated hose sampling (see below) is another useful approach that
is ideal for keeping the number of samples low while sampling the surface mixed layer.
Where possible, the depth of the maximum chlorophyll fluorescence (typically determined
with a vertical profiling instrument) should be sampled directly for phytoplankton analysis.
Water sampling devices are needed, regardless of the platform from which the samples are
taken. Bucket samples at the very surface of the water can be used, but also can sometimes be
misleading, so ideally, a surface sample should be collected 1 m or so below the actual
surface using a Niskin-style bottle (Figure 3.4). There are simpler sampling devices like the
Ruttner sampler and advanced types like the GoFlo bottles. The Niskin and GoFlo samplers
may be mounted on special racks called rosette samplers to facilitate sampling at multiple
depths on a single cast (i.e. a lowering of the bottle and associated instruments on a cable)
from the ship. These bottles are cocked open during descent, and are commonly released by
either a weight that is dropped down the line or wire once the desired depth is reached, or by
a computer when the bottles are mounted on a rosette (Figure 3.4).
Figure 3.4. Water sampling devices. Left: Individual Niskin-type bottles mounted in a rosette for
sampling at multiple depths on a single cast; right: a water sampling device of the Ruttner type. Photo: B.
Karlson.
Since phytoplankton are often not distributed uniformly in the water column (see Chapter 1),
hoses can be used to sample the mixed layer at the surface of the water column, e.g. 0-10 m.
A 10 m-long segment of hose or silicone tubing (Figure 3.5) can be lowered through the
water with a weight on the bottom end. Sometimes a valve can be attached to the top end.
When no valve is present, a string or a line attached to the weight can then be pulled to the
surface, being careful that water is not lost during this process. By removing the bottom end
95
of the hose from the water first, no sample is lost, and the contents can then be emptied into a
bucket. If a valve is used, it needs to be closed once the hose has been lowered, as this will
help to retain the sample on retrieval. If need be, segments of hose can be added or subtracted
to give the appropriate depth for sampling.
Plankton nets (Figure 3.6) can also be used to
sample the phytoplankton community, but
this type of collection is not quantitative. It is
a good way to collect a large amount of
biomass to see if HAB species are present,
even if they are at low cell concentrations. It
could be used to collect sufficient material for
toxin analysis, for example. For general
surveys, plankton nets with 20 - 25 µm mesh
diameter are commonly employed. If smaller
species are to be monitored, a mesh size from
Figure 3.5. A hose used for phytoplankton 5 to 10 µm should be used to ensure that both
sampling. The valves are open when lowering the
nanoplankton (2-20 µm) and microplankton
hose into the water. The top valve is closed before
lifting the tube out of the water. Photo: B. Karlson. (20-200 µm) are sampled. Picoplankton (0.2 –
2 µm) are too small to be collected by a net.
The net is towed vertically up and down
through the water column and the material
contained in the end container is poured into
a sampling bottle. The net can also be trailed
alongside or behind a slowly moving boat to
collect surface plankton. The planktonic
material collected in the cup at the bottom of
the net should not be preserved, and should
Figure 3.6 A plankton net, used to collect large be kept in the cold and dark until analyzed.
amounts of biomass. Note that these types of This material can be concentrated further
samples are not quantitative. using a filter and extracted for toxin analysis.
Examination of a live sample facilitates
identification of species, since it is often easier to make a species identification on a cell that
is swimming or that has its normal pigmentation.
Equipment such as water samplers, hoses, and sample bottles should be rinsed in fresh water
and dried before storage for future use. Drying should be rapid to prevent unwanted algal
growth within the hoses and tubes. Plankton nets should be rinsed thoroughly with fresh
water to remove all plankton cells that may be attached to the net. At the same time, the nets
should be checked to ensure there are no tears or holes. Plankton nets should be hung up to
dry in an area protected from direct sunlight, and sharp or pointed objects. Plankton nets
should be washed regularly in soapy water, at least once a year. They should be soaked for
one day and then rinsed in abundant fresh water. After rinsing, they should be placed in fresh
water for one day and then dried and stored.
3.5.1.1 Fixation procedures for plankton samples
There are a number of different fixation methods for phytoplankton, but for routine
monitoring programs, Lugol’s is the preferred preservative. The recipe for the acidic form is
given here, but note that if the fluorescent dye called Calcofluor is to be used to delineate the
thecal plates of some dinoflagellates (a very useful tool for species identification: Fritz and
Treimer 1985; Edler and Elbrächter 2010; Andersen 2010), neutral Lugol’s is needed.
96
Acid Lugol’s is made by dissolving 100 g potassium iodide (KI) in 1L of distilled water, then
50 g crystalline iodine (I2) is dissolved in this solution followed by the addition of 100 mL of
glacial acetic acid. This produces about a 3% solution. For neutral Lugol’s, the acid is
omitted. Lugol’s should be stored in the dark or in a brown bottle as the iodine is light
sensitive and will degrade. It should also be stored with a tight fitting lid and kept away from
live sample areas (e.g. the general culture environment).
For cultures, add 1 drop of 3 % Lugol’s solution to 1 mL of a culture, and for field samples,
10 drops per 200 mL of sample or until the color of weak tea. Overuse of Lugol’s will cause
some delicate flagellate species to over stain, lose flagella, or break up entirely.
3.5.2 Water transparency – Turbidity tubes and Secchi discs

An algal bloom with a high biomass decreases water transparency. Perhaps the simplest way
to measure turbidity is using a turbidity tube. This is a tall glass or plastic cylinder with a
white or black and white disc at the bottom, and is useful when a desalination plant does not
have easy access to a dock or small boat, or if waters are shallow. The tubes are available
commercially or can be easily constructed from common laboratory supplies. Water is poured
into the tube until the disc at the bottom is no longer visible. For turbidity tubes which have a
turbidity scale marked on the side, read the number on the nearest line to the water level. This
is the turbidity of the water. If the tube does not have a scale marked, measure the distance
from the bottom of the tube to the water level with a tape measure and look up or calculate
the turbidity o f the water sample using the instructions provided with the tube.
A related way to measure water
transparency and detect blooms is to use a
Secchi disc (Figure 3.7). This can be
purchased or made by hand. A weighted,
white, circular disc, usually 30 cm in
diameter, is lowered from a boat or a dock
using a thin rope with markings every
meter or every half meter until the disc is
not visible. Then the disc is raised until it
is barely visible. The distance from the
sea surface to the disc is called the Secchi
depth. It is important to measure the
Figure 3.7. A Secchi disc is used to measure water Secchi depth on the side of the boat or the
transparency. Photo: B. Karlson. dock that is in the shadow or has the least
sun glint. From small boats it is
recommended to use an aquascope to minimize effects of reflections. Some scientists
working in freshwater prefer the disc be divided into quarters painted alternately black and
white. This is not a standard Secchi disc and should be avoided, at least in the sea. It is
important to carry out the Secchi depth measurements in a consistent way, e.g. carrying out
the measurements at certain time of day, and to collect water samples and net samples at the
same time. The amount of suspended particles from sediments influences water transparency
making the intrepretation of the Secchi depth difficult during and after high wind events and
close to river mouths. The Secchi depth is related to the attenuation coefficient which may be
calculated if the light field is measured at several depths in the water column. This may be
carried out e.g., using a light meter mounted on a CTD. A reference light meter mounted in
air is also needed.
97
3.5.3 Chlorophyll-a and other photosynthetic pigment

Chlorophyll-a (chl a) is the main photosynthetic pigment in most algae, and therefore it can
be used as a proxy for phytoplankton biomass. Since chl a content is not a constant fraction
of phytoplankton biomass, this proxy must be used with caution. Light history and nutrient
conditions and other factors may influence the chl a content of microalgae.
Chl a is often estimated using water sampling and subsequent filtering and extraction of the
pigment, which is measured using a spectrophotometer or a laboratory fluorometer. The most
common way to separate phytoplankton cells from seawater is to filter the seawater sample to
concentrate all the particles. The filters are then soaked in a solvent (typically 90% acetone)
that will extract the pigments from the cells. This extract can then be measured in a
fluorometer (to detect chlorophyll fluorescence) or a spectrophotometer (to detect light
absorbance by chlorophyll). Details of the fluorometric method can be found in Welschmeyer
(1994).
A more exact method for use on water samples is High Performance Liquid Chromatography
(HPLC), which separates the different photosynthetic pigments before they are quantified.
HPLC is considered by many to be the new standard for chl a analysis. HPLC also gives
information on pigments such as chl b, chl c1, c2, c3, carotenoids and other accessory pigments.
Some of these pigments are specific for certain phytoplankton groups, e.g. peridinin for most
dinoflagellates. Thus HPLC analysis gives what is called chemotaxonomic information on
the phytoplankton community. It is, however, unlikely that a desalination plant will have an
HPLC available for this type of measurement, so the fluorometric method (above) or in vivo
fluorescence (below) are recommended.
3.5.3.1 In vivo and in situ chlorophyll fluorescence
The chlorophyll in live phytoplankton produces red fluorescence when exposed to light, (e.g.
sunlight or the blue excitation light in fluorometers). Fluorometers mounted on CTDs and
other in situ instruments that are lowered through the water column are often called in situ
fluorometers. These may also be mounted on oceanographic buoys or in Ferrybox systems on
ships of opportunity. The fluorescence is calibrated against measurements of chl a in samples
from the same location, making the fluorescence an easy-to-measure proxy for phytoplankton
biomass. Indeed, many desalination plants have fluorometers mounted within their plants,
measuring online chlorophyll fluorescence continually. Although this gives no information
about the species of algae or the other
Chl. fluorescence, arbitrary units
pigments that are present, it does give

an approximate indication of algal
biomass. The same is true for in situ
fluorescence measurements.
Some limitations of the approach
should be noted, however. First – the
relationship between chl a and
fluorescence is not constant across all
phytoplankton species, nutritional
Date in June 2002 conditions, and times of sampling.
Figure 3.8. Variability of in vivo chl a fluorescence Some cells are large, and some small,
measured at approximately 2 m depth using the and thus chl a will vary accordingly.
oceanographic buoy Läsö E. in the Kattegat, near the Likewise, cells that are nutrient- or
North Sea in 2002. The night to day ratio is about 2-3.
Figure: B. Karlson.
light-limited can have lower chl a
content than the same cells under more
98
favorable conditions. Furthermore, chl a fluorescence is influenced by the light history of the
organisms. The nighttime (nocturnal) to daytime ratio of chl a fluorescence of the same
phytoplankton community may vary by a factor of 2-3. In Figure 3.8, data on hourly
measurements of chl a fluorescence at approximately 2 m depth in the Kattegat, adjacent to
the North Sea, are presented. Note the low daytime values and the high nocturnal values. It is
likely that the same phytoplankton community was present day and night. Since nocturnal
data are the most consistent, it is recommended to use only the night time chl a fluorescence
data for near surface sensors.
Despite all of these limitations and caveats, when in vivo fluorescence measurements are high
in an area relative to past measurements, this is indicative of a major algal bloom, and thus
can be used to guide pretreatment options. Additional information on the identification and
abundance of the algal species causing
the fluorescence would be even more
informative. Methods to obtain that type
of data manually or using autonomous
instruments are given elsewhere in this
chapter.
3.5.4 Automated water sampling

To achieve cost efficient observations of
HAB-organisms, automated sampling
may be used. There are commercially
available, refrigerated water sampling
devices that hold 24 one-liter samples
(Figure 3.9). These types of samplers
Figure 3.9. An automated water-sampling device, which is
part of a Ferrybox-system in the Baltic Sea. Photo: B. are used in water treatment facilities and
Karlson. also in Ferrybox systems on ships and
in flow through systems on land. It is
useful to have two water sampling
devices at each location; one is used for
live and the other for preserved samples.
For the latter, preservatives such as
Lugols or formalin are put in the bottles
such that the organisms are instantly
preserved when the sample is added. If
Lugol’s iodine solution is used as
preservative, the sampling device will
turn brownish. Sampling may be
programmed for certain hours or
locations. Another option for automated
water sampling is in situ systems. At
present there are few of these systems
available commercially. Samples are
collected in plastic bags prefilled with
preservative. These devices are used, for
example, in a monitoring program using
Figure 3.10. Mooring designs. The black rectangles
represent sensors. In the depth profiling designs (G-K) and
oceanographic buoys in the United
the one with a pump (E) only one sensor package is needed Kingdom.
to cover a large depth interval.
99
3.5.5 Sampling using fixed platforms

To achieve high-frequency sampling at static locations, oceanographic buoys and other fixed
platforms such as pilings, oil platforms, and wind turbines may be used for mounting sensors
and automated water-sampling devices. Figure 3.10 shows a schematic of the different types
of mooring designs that could be considered. A disadvantage in mounting automated
sampling devices on buoys is that the samples need to be brought to the laboratory for
analyses. This is useful for research, but may be less useful for near real-time observing
systems. If sufficient power is available, e.g. in cabled ocean observatories, advanced
instruments such as Imaging FlowCytobot (Sosik and Olson 2007 and Olson and Sosik 2007)
or the FlowCam and other automated laboratories may be used to obtain data in near real
time. These are discussed in section 3.6.6. Buoys may be serviced at sea, but it is often cost
effective to carry out service and calibrations on land and to have two systems, one in
operation and one being serviced or ready to be deployed. In Figure 3.11, examples of fixed
platforms are presented, and Figure 3.12 shows instruments that can do vertical profiling.
Figure 3.11. Examples of instrumented oceanographic buoys operated by the Swedish Meteorological and
Hydrological Institute. The most important parts, i.e. the underwater sensors, are not shown. Left: the coastal
Koster fjord buoy in the Skagerrak, Sweden designed by Techworks Marine Ltd. Ireland, middle: the offshore
Huvudskär buoy in the Baltic Sea designed by Axys Tecnologies Inc. Canada, and right: the offshore Läsö
buoy in the Kattegat, between Sweden and Denmark, designed by Fugro-Oceanor, Norway. Photos: Fredrik
Waldh, Henrik Lindh/Per Olsson and Bengt Karlson.
Figure 3.12. Examples of automated depth-profiling instrument platforms. All can be fitted with
sensors useful for HAB observations. Left: the ArvorC from NKG, France, (http://www.nke-
instrumentation.fr, middle, the Thetis from Wetlabs Inc., USA (http://www.wetlabs.com) and right: the
Wirewalker. Photo: R. Kudela).
100
3.5.6 Ships of opportunity and Ferrybox systems

Ships of opportunity and Ferrybox systems are cost-efficient platforms for collecting
information on near-surface HABs, but require a
ship owner’s willingness to provide free access to
their vessel to install and service the instruments.
The systems are often mounted on ferries, but are
also found on other merchant and research
vessels. Ships that have stable timetables and
cross an area-of-interest frequently (e.g. every
other day) are the most suitable. The owner of the
ship should allow two holes (or through-hull
fittings) in the ship at 3-4 m depth, one for a
seawater inlet and the other for seawater
Figure 3.13. Screen shot of a Ferrybox
system in operation. Continuous values of discharge. Inside the ship, a small, but dedicated
temperature, salinity, turbidity and area for the Ferrybox system is needed to mount
chlorophyll fluorescence are shown. Photo: a pump that will not damage delicate
D.M. Anderson. phytoplankton, a de-bubbling device, an
automated cleaning system, sensors, and water
sampling devices. Data on chlorophyll fluorescence, turbidity, salinity, temperature,
dissolved oxygen, and phycocyanin, are collected continuously every few hundred meters
(Figure 3.13). Water samples are collected and archived when the ship passes predefined
locations. Ideally, an Internet connection makes it possible to send data in near-real time.
Sending data while the ship is in harbor and within reach of low cost wireless communication
may be sufficient. A service team collects water samples while the boat is docked in the
harbor and transports samples to a laboratory for analysis. During the visit to the ship, sensors
are cleaned and other maintenance is conducted. Karlson et al. (2016) describes a Ferrybox
system in some detail (Figure 3.14.)
Figure 3.14. Components of a Ferrybox system: Left: sensors for conductivity, oxygen, chlorophyll
fluorescence (proxy for phytoplankton biomass), phycocyanin fluorescence (proxy for cyanobacteria biomass),
and turbidity Photos: Bengt Karlson; right: an example of a fully configured, commercially available Ferrybox
system: http://www.4h-jena.de/.
101
3.5.7 Flow-through systems on land or on buoys

Systems very similar to Ferrybox systems may also be mounted on land or in large buoys. It
would be highly informative if this type of sensor and automated sampling device were
mounted in the water flow that leads to a desalination plant. The system should be indoors or
at least in an environment protected from rain and sea spray. Other options are to mount the
system on the dock or on a pier accessible to a service team that would collect samples and
maintain the system. It is important to avoid locations where sediments are suspended
regularly, e.g. through ship traffic. Seawater should be pumped to the sensors using pumps
that do not damage the phytoplankton, particularly if archived samples are to be collected for
microscopic species analysis; large peristaltic pumps are commonly used for this. Flow-
through systems on land are also useful when connected to imaging flow cytometry
instrumentation for automated identification and enumeration of HAB organisms, especially
if both power and a fast internet connection are available (further discussion below).
3.5.8 In-water optical instrumentation for detecting HABs

Submersible optical instruments can be deployed on a range of platforms from moorings to
autonomous underwater vehicles (AUVs) and provide information on the in situ constituents,
from phytoplankton biomass to detrital particles. Obtaining HAB-specific signals is more
difficult but could be accomplished with multivariate approaches that combine inherent (e.g.
phytoplankton absorption and backscatter) and apparent (e.g. diffuse attenuation coefficient,
remote-sensing reflectance) optical properties with information from other sensors, such as
temperature, salinity, and nutrients. As with any moored instrument, however, the constant
threat of biofouling requires continual maintenance and creative solutions to clearing organic
matter that will attach to any sampling device.
3.6 IDENTIFICATION AND ENUMERATION OF HAB ORGANISMS
There are multiple reasons to monitor the species of algae that are in the intake waters for a
desalination plant. Knowledge of which species are present makes it possible to anticipate
pretreatment or operational strategies. Some species are toxic, so it is important to identify
them and to ensure that membranes and other removal processes are functioning properly. To
document the safety of the desalinated drinking water, it may be necessary to make toxin
measurements in the intake and drinking water when major blooms of toxic HABs are
detected. Non-toxic species also need to be documented, as they can cause clogging or
fouling at low and high cell densities; some species are prolific producers of organic
materials, and others are not. It is thus important to know not only what species and cell
concentrations are present in intake waters, but also to have records of what happened in the
plant in the past when that species was present. Detailed record keeping of species,
concentrations, impacts and treatment strategies should thus be a standard operating
procedure. Unfortunately, this is not currently done at many desalination plants, as there is
seldom appropriate expertise for identifying and counting algal species, and often, there is no
appreciation of the significance and utility of such information at the managerial level. The
following sections provide information on this important aspect of algal monitoring.
To identify and to determine the abundance of phytoplankton, cells are typically counted
under a microscope or using automated cell counters, described below. It is fairly easy to
count cells (see Appendix 4), though identification to the species level can be challenging. As
described below, there are a number of online resources, and many HAB or phytoplankton
experts throughout the world who can help. The unit for abundance is usually cells per liter
(cells/L) or cells per milliliter (cells/mL). To determine the abundance of harmful organisms
and the biodiversity of samples, organisms should be identified to the species level if possible.
102
This is often done for the most abundant 5 or 10 taxa, with other less plentiful species noted,
but not counted. Microscopy is the classic method and includes light microscopy,
fluorescence microscopy and electron microscopy. The latter is necessary if the identification
of smaller cells is needed or if it is otherwise difficult to identify an organism at the species-
level. Microscopy and molecular methods for quantitative phytoplankton analyses are
described in a UNESCO – IOC Handbook (Karlson et al. 2010).
3.6.1 Essential information for identification phytoplankton

It would be quite useful for a plant operator to have knowledge of the species that are in the
intake waters or those surrounding the plant. To be able to correctly identify organisms that
cause problems for desalination plants, it is necessary to have access to personnel who know
how to identify those organisms. The IOC Harmful Algal Bloom Centre, University of
Copenhagen, Denmark, arranges courses in microalgae identification that can train staff. In
some cases, local universities can arrange courses on the topic. Identification guides and
taxonomic keys are available as books but also web sites provide useful information. A list of
these is provided below. Older books may be difficult to find and are not listed.
3.6.1.1 Books for identification of harmful algae and phytoplankton
Bérard-Therriault, L., Poulin, M., Bossé, L. 1999. Guide d'identification du
phytoplancton marin de l'estuaire et du golfe du Saint-Laurent incluant également
certains protozoaires. Canadian Special Publication of Fisheries and Aquatic Sciences
No. 128. 387 pp.
Fukuyo, Y., Takano, H., Chikara, M., Matsuoka, K. 1990. Red Tide Organisms in
Japan: An Illustrated Taxonomic Guide. Uchida Rokakuho Co. Ltd., Tokyo, Japan, 407
pp.
Hallegraeff, G. M. 1991, Aquaculturists' Guide to Harmful Australian Microalgae.
Fishing Industry Training Board of Tasmania Inc., CSIRO Division of Fisheries, Hobart,
Tasmania, Australia, ISBN 0-643-05184-8, 111 pp.
Hoppenrath, M., Elbrachter, M., Drebes, g. 2009. Marine Phytoplankton. Selected
Microphytoplankton Species from the North Sea Around Helgoland and Sylt., Kleine
Senckenberg-Reihe 49, Stuttgart, Germany, ISBN 978-3-510-61392-2, 264 pp.
Horner, R.A. 2002. A Taxonomic Guide to Some Common Marine Phytoplankton,
Biopress Limited, Bristol, England, ISBN 0-948737-65-4, 195 pp.
Lassus, P. Chomérat, N., Hess, P., and Nézan, E. 2016. Toxic and Harmful Microalgae
of the World Ocean. International Society for the Study of Harmful Algae/
Intergovernmental Oceanographic Commission of UNESCO, Denmark (2016). IOC
manuals and Guides 68. (Bilingual English/French)
Al-Kandari, M., Al-Yamani, F.Y., Al-Rifaie, K. 2009. Marine Phytoplankton Atlas of
Kuwait’s Waters. Kuwait Institute for Scientific Research, Safat, Kuwait, ISBN 99906-
41-24-2, 351 pp. http://www.issha.org/Welcome-to-ISSHA/Web-shop/Toxic-and-
Harmful-Microalgae-of-the-World-Ocean
Kraberg, A., Baumann, M., Dürselen, C.D. 2010. Coastal Phytoplankton: Photo Guide
for Northern European Seas. Verlag Dr. Friedrich Pfeil, München, Germany, ISBN 978-
3-89937-113-0, 204 pp.
Larsen, J., Moestrup O. 1989. Guide to Toxic and Potentially Toxic Marine Algae. The
Fish Inspection Service, Ministry of Fisheries, Copenhagen. 61 pp.
103
Omura, T., Iwataki, M., Borja, V.M., Takayama, H., Fukyo, Y. 2012. Marine
Phytoplankton of the Western Pacific. Kouseisha Kouseikaku, Tokyo, 160 pp.
Thomsen, H. A. 1992. Plankton i de indre danske farvande. Havforskning fra
Miljøstyrelsen, Nr. 11. Copenhagen.
http://www.mst.dk/Publikationer/Publikationer/1992/11/87-7810-034-8.htm
Throndsen, J., Hasle, G.R. and Tangen, K. (2007). Phytoplankton of Norwegian Coastal
Waters, Almater Forlag, 341 pp.
Tomas, C. (Editor) 1997. Identifying Marine Phytoplankton. Academic Press, San Diego.
858 pages.
3.6.1.2 Web sites with information on harmful algae and phytoplankton
IOC-UNESCO Taxonomic Reference List of Harmful Micro Algae:
http://www.marinespecies.org/hab
AlgaeBase: http://algaebase.org
World Register of Marine Species (the parts on algae are based on AlgaeBase):
http://marinespecies.org
Nordic Microalgae: http://nordicmicroalgae.org/
Center of Excellence for Dinophyte Taxonomy: http://www.dinophyta.org/
Phytoplankton guide (Louisiana Universities Marine Consortium):
http://phytoplanktonguide.lumcon.edu/
Phyto'pedia: http://www.eos.ubc.ca/research/phytoplankton/
PlanktonNet: http://planktonnet.awi.de/
3.6.2 Light microscopy

The recommended method for enumerating and identifying most HAB organisms is the
Utermöhl sedimentation chamber technique (Figure 3.15). This is also the most common
method for quantitative analysis of phytoplankton. Step-by-step instructions on how to use
the Utermöhl method are found in Edler and Elbrächter (2010). A description is also found in
Appendix 4, along with a
description of the use of an
alternative counting chamber
called the Sedgewick Rafter
slide. With the Utermöhl
method, organisms are
concentrated through
sedimentation. An inverted
Figure 3.15. The Utermöhl method for settling and counting algae. microscope with high quality
Left: concentration of samples using sedimentation chambers; right: optics is necessary to carry out
an inverted microscope. Photos: Å. Edler, B. Karlson.
the analyses (Figure 3.15).
Qualified training in
phytoplankton identification, taxonomy and systematics is very important to ensure high
quality results. A disadvantage with the method is that organisms smaller than about 5-10 µm
cannot readily be identified to the species level. Another problem is that a relatively small
volume (10-20 mL) is most often analyzed, although 50-100 mL chambers are also routinely
used. This means that rare species may be overlooked. This is unlikely to be a problem for
monitoring blooms affecting desalination plants, as it will be blooms at high cell
104
concentrations that are of primary concern. A useful tool when counting phytoplankton
samples and working with the resulting data is the free software Plankton Toolbox available
at http://nordicmicroalgae.org/tools (Karlson et al. 2015).
Although it is possible to do basic cell identification and enumeration using simple
microscopes, this process will be easier if a microscope fitted with contrast enhancement
equipment is available, either phase contrast or differential interference contrast (DIC, often
termed Nomarski). Epifluorescence is also useful. Objectives should include 4-5x, 10x, 20x,
40x, and for high-magnification work, 100x. With oculars of 10x this results in a
magnification of 40-1000x.
3.6.3 Fluorescence microscopy
Fluorescence microscopy is a useful tool to enumerate organisms that dominate algal blooms,
making it possible to differentiate particles that are algae from other organisms or detritus
that lack photosynthetic pigments. Figure 3.16 shows the natural colors of autofluorescence
in a sample. Fluorescence microscopy is also used with samples treated with chemicals that
are used in a diagnostic fashion, e.g.
fluorescent RNA-probes (described
below), calcofluor to stain the cell wall
features of the many dinoflagellates with
cellulose cell walls, and DAPI (4',6-
diamidino-2-phenylindole) to reveal cell
nuclei.
Samples for fluorescence microscopy are
most often concentrated by filtering and
then examined with inverted or
conventional microscopes equipped with
specific excitation lamps and filters. A
Figure 3.16. Autofluorescence in phytoplankton sample.
The red cells contain chlorophyll and the yellow cells combination of staining with calcofluor
phycoerythrin. The green cell is non-photosynthetic. and light microscopy with the Utermöhl
Photo: B. Karlson. method is presented in Edler and
Elbrächter (2010). Analysis of autotrophic
picoplankton (0.2-2 µm) is often carried
out using fluorescence microscopy.
Autofluorescence from phycobilins in
Synechococcus-type cyanobacteria
facilitates analysis.
3.6.4 Electron microscopy

Electron microscopy is a costly and time-
consuming method. It is used to identify
many small phytoplankton organisms to
the species level, or to visualize small,
Figure 3.17. Right: a scanning electron micrograph of distinctive features on larger cells (Figure
Protoceratium reticulatum. The arrow shows the ventral 3.17). Scanning electron microscopy
pore used to help identify this species. To the left, the (SEM) and transmission electron
same species as photographed using light microscopy. A
the motile, vegetative stage cell; B) the resting cyst.
microscopy (TEM) should only be used
Photos: M. Kuylenstierna. as an occasional complement to light
microscopy and imaging flow cytometry.
105
3.6.5 Imaging flow cytometry

Flow cytometers are particle counters that were originally developed for counting and
differentiating blood cells. Most models use one or a few lasers for creating light that is used
for exciting fluorescent particles. In addition, light scattering properties of particles are used
to differentiate cells. For phytoplankton research, they were first mainly used for pico- and
nanoplankton (0.2-2 and 2-20 µm respectively) and the fluorescent and scattering properties
of the algae were used to differentiate the algae to a very rough group level. Later, imaging
flow cytometers were developed, now available as in situ instruments (Sosik and Olson 2007;
Olson and Sosik 2007). In the latter instruments, fluorescence of chlorophyll is commonly
used to trigger a camera and all particles that fluoresce, i.e., algal cells, are documented in a
digital image. Automated image analysis is used to identify the organisms and to measure
size, etc. The software must first be developed (‘trained’) by experts on the local
phytoplankton community, but thereafter the instrument can run autonomously, taking
Figure 3.18. Imaging flow cytometers. Left to right: Imaging FlowCytobot (IFCB) from McLane
Laboratories, Inc., the FlowCam from FluidImaging Inc., and the CytoSense from CytoBuoy.
thousands of images every minute, and classifying the organisms to genus and even to the
species level. When new species are observed, new training sets are developed and added to
the software. There are currently at least three imaging flow cytometers available
commercially that are useful for phytoplankton analyses (Figure 3.18). Some are available
both as desktop and as in situ instruments deployable on oceanographic buoys. These sensors
could be placed online within a desalination plant to record the abundance and identity of
algal species in the intake waters, providing a high-frequency, high-resolution record of the
species that can be compared to the plant’s operational data to identify problem species, and
to guide pretreatment strategies. An example of the output from the Imaging FlowCytobot
(IFCB) is shown in Figure 3.19.
Imaging flow cytometers:
FlowCam: http://www.fluidimaging.com/products-particle-vision-pv-series.htm
CytoSense: http://www.cytobuoy.com/
Imaging FlowCytobot: (IFCB): http://www.mclanelabs.com/master_page/product-
type/samplers/imaging-flowcytobot
106
Figure 3.19. Example output from the autonomous IFCB. From left to right, the large cells on the left
are microzooplankton grazers, the rectangular cells are diatoms and the round cells are the toxic HAB
dinoflagellate Alexandrium fundyense. Photo: M. Brosnahan.
3.6.6 Molecular techniques

Advancements in molecular technology offer efficient and powerful alternatives to
microscopic methods for detection and enumeration of HAB species in natural assemblages
(reviewed in Sellner et al. 2003; Kudela et al. 2010). The application of molecular methods
for HAB monitoring is particularly valuable for species that cannot be reliably identified by
light microscopy due to small size and/or the lack of distinguishing morphological features
(e.g. Coyne et al. 2001), or for sensitive detection of invasive species (e.g. Cary et al. 2014).
In addition, fragile HAB species that disintegrate or distort in the presence of chemical
fixatives may be underestimated by microscopy, but can be detected and accurately
quantified using molecular methods (Doll et al. 2014).
It is unlikely that a desalination plant would have the capabilities to undertake most of the
molecular techniques discussed here, but the methods are presented to demonstrate what can
be done by outside laboratories, and to help explain the technology that is being incorporated
into some of the new sensors and instruments that may be used by desalination plants at some
point in the future.
Molecular approaches for detection of harmful algae are based on species-specific differences
in nucleic acid (DNA or RNA) sequences. These methods often target the small or large
subunit of the ribosomal RNA (rRNA) genes, which are present in high copy number (tens to
tens of thousands) in the genomes of all organisms. The rRNA gene sequences contain
conserved and variable regions, allowing development of molecular assays for different
taxonomic levels of distinction, ranging from strains and species to genera and classes of
phytoplankton, or encompassing the entire prokaryotic or eukaryotic communities. The
design and in silico validation of molecular assays is facilitated by the vast amount of rRNA
107
sequence information currently available in public databases from a broad range of species
(Hugerth et al. 2014).
While most molecular techniques still require collection of water for laboratory analysis,
autonomous platforms such as the Environmental Sample Processor (ESP; Figure 3.20) and
the Autonomous Microbial Genosensor have been
designed to conduct automated molecular assays for rapid
in situ detection of HAB species (Scholin et al. 2009;
Scholin 2010; Yamahara et al. 2015). Two-way wireless
communication allows data access and remote control of
sampling times. These devices collect particulates by
filtration for molecular detection of a range of microbial
species, including HABs and pathogens, as well as
antibody-based detection of HAB toxins in an enzyme-
linked immunosorbent assay (ELISA) format (Doucette et
al. 2009). Initially designed to be moored at a fixed
location changes in technology have allowed ESP
deployment in the deep sea (Ussler et al. 2013) and on
Figure 3.20. The Environmental gliders (Seegers et al. 2015) for additional flexibility. One
Sample Processor (green canister), important limitation of the ESP technology is that it
an autonomous “laboratory in a can” detects “target” organisms, i.e. those for which molecular
that can detect and enumerate HAB probes have been developed and which are incorporated
species and measure toxins. Photo: into the ESP assay system. This means that it could be
Woods Hole Oceanographic
Institution. useful in detecting known toxic HAB species, for example,
but not for identifying new or unknown species.
3.7 SATELLITE REMOTE SENSING
In the context of desalination and the development of observing capabilities for HABs,
satellite remote sensing has great potential to be of use to operators. A comprehensive
overview is provided in Chapter 4.
3.8 TRANSPORT AND DELIVERY OF HARMFUL ALGAL BLOOMS
Desalination plants would benefit greatly from forecasts of algal bloom transport and landfall,
but such capabilities typically require numerical models of coastal hydrography. These are
typically far beyond the technical or financial resources of many individual plants, but
regional approaches to this type of technology are being explored, and thus the fundamentals
of such systems are described here.
3.8.1 Empirical and numerical models

Technological advances have expanded capabilities for research and monitoring of HABs,
but the blooms will always be under sampled because of the large space and time scales over
which they occur. As a result, models are being used to help extrapolate and interpret these
sparse observations. These include empirical and numerical models. An example of an
innovative and useful empirical model is that of Raine et al. (2010) who developed a model
for predicting Dinophysis blooms on the southwestern coast of Ireland based on the wind
index as a proxy for wind-driven exchange of water and HAB probability onto the shelf. This
empirical approach was successful because these blooms occur during summer when offshore
water is advected by onshore winds into the highly-stratified nearshore environment
(downwelling – see Chapter 1). The model has improved understanding of the dynamics of
diarrhetic shellfish poison (DSP) intoxications that greatly impact the shellfish in Bantry Bay.
108
Hydrodynamic circulation models are commonly used to track and visualize bloom
formation and duration as well as to understand the physical processes controlling
phytoplankton bloom dynamics. Models with varying levels of sophistication have been
developed. Some are purely three-dimensional physical models capable of resolving
hydrography, and into which HAB cells can be introduced as passive particles. In the event
that a particular HAB can be identified through detection methods discussed above,
algorithms can be used to predict its trajectory and map the bloom using Lagrangian Particle
Transport (LPT). This can be coupled to either a 2D or 3D circulation model. LPT is widely
used in oil spill tracking and studies of fish or shellfish larval transport and is now seeing
growing popularity for HAB risk management. LPT can be a powerful tool for estimating the
timing and spatial impact of HAB landfall because many blooms originate offshore and are
moved into the regions where most intakes to desalination plants would be located (either
surface or subsurface). This is the approach used in a HAB forecasting system developed for
Karenia brevis blooms in the Gulf of Mexico in which HAB forecasts are made twice weekly
during bloom events (Stumpf et al. 2009). Blooms are detected using SeaWiFS and MODIS
imagery, and bloom transport is then predicted using hydrographic modeling with passive
particle transport. Vélo-Suarez et al. (2010) determined the physical processes responsible
for the demise of a Dinophysis acuminata bloom, illustrating the importance of retention-
dispersion patterns driven by the physics of the bay. Another study used a combination of
models to track particles and identify the dominant sites of discharge along the Saudi Coast
of the Red Sea (Zhan et al. 2015). This had direct implications to blooms as well as sediment
transport. In each of these cases, particle transport models provided crucial spatial
information about bloom or sediment transport that could potentially serve as an early
warning to desalination plants.
The next step in sophistication and complexity is to couple a detailed biological submodel
(one that incorporates cyst germination, cell growth, nutrient uptake, mortality and
swimming behavior) to a hydrographic model, as has been done for Alexandrium dynamics
in the Gulf of Maine region in the US (McGillicuddy et al. 2005; He et al. 2008). This level
of modeling is species-specific, and not generally appropriate for desalination plants,
however, at least at this point in time.
An alternative but practical approach to biophysical modeling blends empirical and dynamic
methods to leverage the power of both simple and complex modeling approaches. In
California, where seawater desalination plants already exist and are being revitalized or built
for the first time, there is a great need to understand toxin production and transport of HABs
nearshore. To this end, a HAB forecasting system has been developed to predict domoic acid-
producing Pseudo-nitzschia blooms from empirical HAB models that are computed routinely
in near real-time from satellite ocean color parameters and physical output (salinity and
temperature) in a physical oceanographic model (Anderson et al. under review).
3.8.2 High-resolution circulation models to resolve flow near intakes
High-resolution hydrodynamic models for studying physical features ~10-50 km in size, such
as eddies, plumes and fronts, are becoming more accessible as technological advances reduce
computing times and costs. Many of these models were initially developed to study patterns
of fish larval recruitment or runoff close to shore. Models with a degree of physical
complexity are useful for determining the initiation and termination of eddies and their
movement near intake systems. Additional complexity in the form of coupled
biological/ecosystem models provides specific guidance on where nutrient loading is highest
in order to better inform the placement of intakes, initial design strategies or early warning of
109
bloom conditions. Such an approach was demonstrated for the Karkheh Reservoir in Iran
(Afshar et al. 2012).
3.8.3 An example of a regional HAB forecast system

Efforts are underway to combine the technologies described above into forecast systems that
would be of value to multiple desalination plants within large regions. In a pilot project that is
underway at this time, coordinated by the Middle East Desalination Research Center
(MEDRC), an observation and forecast system is being developed that can provide plant
operators with a broader view of the environment around their plants, allowing them to
anticipate algal blooms that are approaching, thereby allowing more adaptive management
and informed decision-making (D.M. Anderson, K. Price, unpub. data). This capability is
being developed through a pilot-project early warning system that involves: 1) development
of satellite remote sensing indices for HABs along the coasts of Oman and the United Arab
Emirates; 2) refinement and expansion of a high-resolution numerical model of regional
hydrography and circulation; 3) combination of satellite bloom data with the hydrographic
model to predict the transport of blooms to the area of desalination intakes; and 4)
development of a web-based portal to provide data and forecasts to plant operators and other
users. The remote sensing and modeling approach to be taken here is similar to that used in
the US to forecast HABS in the Gulf of Mexico and in Lake Erie (Stumpf et al. 2009; Wynne
et al. 2011).
For this project, an Arabian Gulf - Sea of Oman atmosphere-ocean forecast system was
developed. The system includes 1) a high-resolution weather forecast model (WRF) and 2)
the Arabian Gulf - Sea of Oman hydrographic model (named AGSO-FVCOM). WRF is
driven by a global weather forecast model. AGSO-FVCOM is a regional ocean model with a
computational domain that has been expanded to cover the entire Arabian Gulf - Sea of
Oman region.
A Lagrangian tracking model has also been implemented for forecasting HAB trajectories.
An example of the forecasting approach and the types of data that can be generated is given
in Figure 3.21. Satellite imagery provided by ROPME revealed an algal bloom event in the
vicinity of the Barka desalination plant north of Muscat, Oman on November 1, 2015. The
imagery was further processed and digitized so that it could be represented as a field of
particles. The AGSO-FCVOM model was then run in particle tracking mode to forecast
where the bloom might move over the next several days. The model run started on November
1, and produced a series of images (Figure 3.21) extending through 4 November, suggesting
that the bloom patches were moving away from the Barka site. The forecasting model
successfully predicted the general offshore movement of the bloom patches, as confirmed in
subsequent satellite imagery.
This is an example of the manner in which the combination of remote sensing and numerical
modeling can be used to forecast algal blooms that might affect desalination plants. The
regional pilot project will end in 2016, but may be extended. The concept can be readily
applied to many other parts of the world, particularly those where multiple desalination plants
are located in relatively close proximity to each other.
3.9 DISTRIBUTING WARNINGS AND INFORMATION
Many plants will operate bloom observation systems independently, and thus may retain the
data for their own internal use. Ideally, however, the information should also be broadly
distributed since HABs may cover large sea areas, cross national borders, and affect
neighboring desalination plants. All data collectors will benefit from receiving data produced
by other data collectors. An e-mail listserv or regional HAB web page would facilitate such
110
information transfer and dialogue. The web site could also contain information of an
educational nature, so that the public and others can find information (such as the causes of
HABs and summaries of the effectiveness of toxin removal during desalination) to alleviate
their concerns. It is of course important to avoid false positives, i.e. warnings that are wrong.
An example may be that a HAB is observed near the location of a desalination plant and a
warning is issued within the region. However, the HAB is transported away due to shifts in
currents. False positives are inevitable and information to operators and the public describing
this trajectory and likely diversion would be needed. In addition to updated information on
the common web site, yearly reports on the HAB-events in the area should be produced and
archived, as such historical information will be very valuable through time. Figure 3.22
shows a schematic of this data collection, analysis, and information flow.
Figure 3.21. Real-time demonstration of the particle tracking module of the AGSO-FVCOM forecast system
described above. A dense algal bloom was detected in satellite imagery (A) along the Oman coast near the
Barka desalination plant (blue circle) on November 1, 2015 (chlorophyll depicted, with highest concentrations
in red and yellow); B: The densest portion of the bloom was digitized and converted to passive particles (red
circles) for November 1 AM. The forecast model was run with real-time and forecast weather and current
patterns, showing the projected position of the bloom on November 1 PM (C), November 2 AM and PM (D
and E), and November 4 (F) as the bloom was dispersed and transported away. Note that the scale of panels A
and F are different from those for B-E. The green boxes in A and F depict the areas imaged in B-E. Source: D.
M. Anderson, R. Kudela, and R. Ji, unpub. data.
111
Figure 3.22. Illustration of flow of information from data sources to end users, national and international
databases.
3.10 DATA STORAGE AND DISTRIBUTION
All data produced should be stored and their quality controlled. Ideally, the data should be
made available freely to the scientific community. The UNESCO-IOC-Global Ocean
Observing System (GOOS) and its regional organizations provide an umbrella-organization
for this. IOC is the Intergovernmental Oceanographic Commission; it supports the Harmful
Algal Event Database (http://haedat.iode.org) and the International Oceanographic Data and
Information Exchange (IODE, http://iode.org). ICES (www.ices.dk) provides a system for
handling quantitative physical, chemical, and biological data (including phytoplankton) for
the North Atlantic. Similar systems exist or are being set up for other areas.
3.11 FACILITIES, EQUIPMENT, AND PERSONNEL
Facilities and equipment needed for a HAB-observation system are listed in Table 3.3, with
some suggested priorities. However, facilities and priorities will vary dramatically among
plants, depending upon available resources and the magnitude of the perceived HAB threat.
The number of personnel involved in HAB observing efforts will vary dramatically among
desalination plants, as this once again will be determined by the nature of the HAB threat as
well as the funding resources available. In some cases, existing staff can be assigned
additional duties to undertake the sampling and analyses after adequate training. In others, a
single individual might be assigned HAB monitoring responsibilities, with the mandate to
draw upon additional plant personnel when more hands are needed. Local or international
HAB experts (e.g. taxonomy, bloom dynamics, HAB observing systems, remote sensing,
toxin analysis) should be identified and their contact details kept in a suitable archive so they
can be of assistance in training or during outbreaks.
112
Table 3.3. An overview of facilities and equipment needed for a HAB observing system. A
simple list of priorities is included. Many smaller standard items such as filtering equipment
and plankton nets are not listed.
Priority 1-3, 1
Facility/equipment Comment
being highest
Laboratory 1
Small research vessel to sample 1 This could be chartered as needed.
beyond the intake waters
CTD or multi-parameter sonde 1
with in situ fluorometer and
turbidity sensors
Microscopes and appropriate 1
cell counting chambers and
slides
Manuals and guides for 1 Many of these are available online, but hard
identifying algal species, copies near the microscope are still useful.
especially HAB species
Water sampling equipment for 1
intake water and broader-scale
surveys
Computers and software for 1 These are needed for general data collection
data storage, analysis and and analysis, but could also be used to
visualization. download and analyze satellite imagery.
Laboratory supplies and simple 1 This would be for toxin screening only.
toxin test kits Samples with positive results should be sent to
expert laboratories for confirmation. The need
for this capability would depend upon the
potential for toxic HABs in an area.
Routine access to satellite 1 Could develop in-house remote-sensing
imagery and trajectory models analysis capability, or could outsource. HAB
forecast models would be very useful, but need
regional cooperation.
Flow through system with 2 Can provide a steady flow of data over large
sensors and water sampling areas and time if a repeatable transect can be
devices, e.g. Ferrybox (for use monitored, as with ferry routes.
on research vessels or ships of
opportunity).
Oceanographic buoys 2 Buoys situated within and outside the intake
area would be highly informative with the
appropriate sensors, but costly to maintain.
Imaging Flow Cytometer 2 This would be very useful for continuous,
autonomous monitoring of the seawater intake
waters, but it is expensive.
113

Facility/equipment Priority 1-3, 1 Comment
being highest
Equipment and supplies (or 2 These are useful measurements, especially
service) for analyzing TEP and when paired with phytoplankton counts in the
other organic molecules intake water – helps to specify the species and
(Appendix 3) cell concentrations that disrupt operations or
that require specific pretreatments.
Equipment (or a service) for 3 This should be a low priority unless there is a
analyses of algal toxins (LC- significant risk for biotoxin contamination of
MS) water used for human consumption. These
types of analyses are typically outsourced.
Equipment (or a service) for 3 This provides useful information relevant to
analyses of inorganic nutrients HAB bloom growth and persistence, but
typically are outsourced.
Electron microscope 3 This is a low priority and should be available at
local universities.
3.12 SUMMARY
There are numerous techniques to detect HABs in order to react in time to minimize their
effects on desalination plants. To accomplish this goal, local and regional monitoring of
phytoplankton composition and biomass are needed. To carry out such monitoring, a
combination of methods is recommended. A comprehensive monitoring program might
include:
1. Long-term funding and staffing commitments

2. Appropriate facilities and equipment in the form of laboratory, shiptime, sampling
equipment, microscope, analysis instruments
3. Frequent water sampling near or at the desalination plant. This could include:
a. Manual sampling in water flowing into the plant or sampling from ships and
boats in nearby waters.
b. Automated sampling in water flowing into a plant or on fixed platforms or
ships of opportunity outside the intake.
c. Analysis of the composition and biomass of the phytoplankton by microscopy
and/or imaging flow cytometry. Molecular techniques or SEM may be
considered to complement the other methods when high-level species
identification is needed.
d. Continuous online analysis of chlorophyll-a, as a proxy for phytoplankton
biomass
4. Automated measurements of in-water bio-optical and physical parameters, e.g.
chlorophyll fluorescence, temperature and salinity, from fixed platforms such as
moorings or by using Ferrybox systems. Biofouling of sensors must be considered
when deciding frequency of service.
5. The use of ocean color data from satellite remote sensing. Algorithms for chlorophyll-
a, a proxy for phytoplankton biomass, are useful and there are also algorithms for
detecting certain phytoplankton groups in high biomass algal blooms.
6. The use of physical oceanographic models, verified for the geographic area in focus,
to produce forecasts regarding advection of HABs.
114
7. Quality control and storage of data as well as distribution of data to regional and
global data centers.
8. Interpretation of data and visualization of results.
9. Information officers who distribute warnings and information.
If funding is limited, regular water sampling using water bottles and plankton nets with
subsequent microscope analysis of the HAB-species should be conducted. In addition, water
transparency or turbidity should be measured using a Secchi disk or turbidity tube. This
probably requires a part-time technician and linkages to a phytoplankton identification expert.
However, such a low cost approach would miss many blooms and be very dependent on the
availability of the key staff. In practice such a system could benefit from a larger regional
program that detects and forecasts HABs within a region (using remote sensing and
numerical modeling, for example), with minimal staffing at each desalination plant to
respond to advancing HABs.
3.13 REFERENCES
Afshar, A., Saadatpour, M., and Marino, M. A. 2012. Development of a complex system
dynamic eutrophication model: application to Karkheh reservoir. Environmental
Engineering Science 29, 373–385.
Andersen, P. 2010. Filtering – calcofluor staining – quantitative epifluorescence microscopy
for phytoplankton analysis. in: Microscopic and molecular methods for quantitative
phytoplankton analysis, Eds. Karlson, B., Cusack, C. and Bresnan, E. Intergovernmental
Oceanographic Commission Manual and Guides 55, 31-35.
Anderson, C. R., Kudela, R. M., Kahru, M., Chao, Y., Rosenfeld, L. K., Bahr, F. L.,
Anderson, D. M., and Norris, T. A. Accepted for publication. Towards an operational
HAB forecast system for California. Harmful Algae.
Bosch van Drakestein, F. G. 2014. Modeling of tides and spreading of saline water in the
Arabian Gulf and in the coastal area of Abu Dhabi with a flexible mesh model. Master
thesis 67 pp.
Cary, S. C., Coyne, K. J., Rueckert, A., Wood, S. A., Kelly, S., Gemmill, C. E. C., Vieglais,
C., and Hicks, B. J. 2014. Development and validation of a quantitative PCR assay for
the early detection and monitoring of the invasive diatom Didymosphenia geminata.
Harmful Algae 36, 63–70.
Coyne, K. J., Hutchins, D. A., Hare, C. E., and Cary, S. C. 2001. Assessing temporal and
spatial variability in Pfiesteria piscicida distributions using molecular probing
techniques. Aquatic Microbial Ecology 24, 275–285.
Doll, C., Main, C. R., Bianco, C., Coyne, K. J., and Greenfield, D. I. 2014. Comparison of
sandwich hybridization assay and quantitative PCR for the quantification of live and
preserved cultures of Heterosigma akashiwo (Raphidophyceae). Limnology and
Oceanography - Methods 12, 232–245.
Doucette, G. J., Mikulski, C. M., Jones, K. L., King, K. L., Greenfield, D. I., Marin, R.,
Jensen, S., Roman, B., Elliott, C. T., and Scholin, C. A. 2009. Remote, subsurface
detection of the algal toxin domoic acid onboard the Environmental Sample Processor:
assay development and field trials. Harmful Algae 8, 880–888.
115
Edler, L. and Elbrächter, M. 2010. The Utermöhl method for quantitative phytoplankton
analysis. in: Microscopic and molecular methods for quantitative phytoplankton analysis,
Eds. Karlson, B., Cusack, C., and Bresnan, E. Intergovernmental Oceanographic
Commission Manual and Guides 55, 13-20.
Fritz, L. and Triemer, R. E. 1985. A rapid simple technique utilizing Calcofluor White M2R
for the visualization of dinoflagellate thecal plates. Journal of Phycology 21, 662-664.
He, R., McGillicuddy, D. J., Keafer, B. A., and Anderson, D. M. 2008. Historic 2005 toxic
bloom of Alexandrium fundyense in the western Gulf of Maine: 2. Coupled biophysical
numerical modeling. Journal of Geophysical Research: Oceans 113, 1978–2012.
Hugerth, L. W., Muller, E. E., Hu, Y. O., Lebrun, L. A., Roume, H., Lundin, D., Wilmes, P.,
and Andersson, A. F. 2014. Systematic design of 18S rRNA gene primers for
determining eukaryotic diversity in microbial consortia. PLoS One 9: e95567.
Karlson, B., Cusack, C., and Bresnan, E. (editors) 2010. Microscopic and molecular methods
for quantitative phytoplankton analysis. Paris, UNESCO. IOC Manuals and Guides 55,
110 pages.
http://hab.ioc-
unesco.org/index.php?option=com_oe&task=viewDocumentRecord&docID=5440
Karlson, B., Godhe, A., Cusack, C., and Bresnan, B. 2010. Overview of methods for
quantitative phytoplankton analysis in: Microscopic and molecular methods for
quantitative phytoplankton analysis, Eds. Karlson, B., Cusack, C., and Bresnan, E.
Intergovernmental Oceanographic Commission Manual and Guides 55, pp 5-12.
Karlson, B., Andreasson, A., Johansen, M., Mohlin, M., Skjevik, A-T., and Strömberg, P.
2015. Plankton Toolbox – open source software making it easier to work with plankton
data, A. Lincoln MacKenzie [Ed]. Marine and Freshwater Harmful Algae 2014.
Proceedings of the 16th International Conference on Harmful Algae. Cawthron Institute,
Nelson, New Zealand and the International Society for the Study of Harmful Algae
(ISSHA), pp 194-197.
Karlson, B., Andersson, L. S., Kaitala, S., Kronsell, J., Mohlin, M., Seppälä, J., and Wranne,
A. W. 2016. Oceanographic variability in near surface waters of the Baltic Sea and the
Kattegat – a comparison of Ferrybox data vs. monitoring data from research vessels.
Journal of Marine Systems 162, 98–111.
Kudela, R. M., Howard, M., Jenkins, B. D., Miller, P. E., and Smith, G., J. 2010. Using the
molecular toolbox to compare harmful algal blooms in upwelling systems. Progress in
Oceanography 85, 108–121.
McGillicuddy, Jr., D. J., Anderson, D. M., Lynch, D. R., and Townsend, D. W. 2005.
Mechanisms regulating large-scale seasonal fluctuations in Alexandrium fundyense
populations in the Gulf of Maine: Results from a physical-biological model. Deep-Sea
Res. II 52, 2698–2714.
Olson, R. J., and Sosik, H. M. 2007. A submersible imaging in flow instrument to analyze
nano and microplankton: Imaging FlowCytobot. Limnology and Oceanography -
Methods 5, 195–203.
Raine, R., McDermott, G., Silke, J., Lyons, K., and Nolan, G. 2010. A simple short range
model for the prediction of harmful algal events in the bays of southwestern Ireland.
Journal of Marine Systems 83, 150-157.
116
Scholin, C. A. 2010. What are“ ecogenomic sensors?” A review and thoughts for the future.
Ocean Science 6, 51–60.
Scholin, C., Doucette, G., Jensen, S., Roman, B., Pargett, D., Marin III, R., Preston, C., Jones,
W., Feldman, J., and Everlove, C. 2009. Remote detection of marine microbes, small
invertebrates, harmful algae, and biotoxins using the environmental sample processor
(ESP). Oceanography 22, 158-167.
Seegers, B. N., Birch, J. M., Marin, R., Scholin, C. A., Caron, D. A., Seubert, E. L., Howard,
M. D., Robertson, G. L., and Jones, B. H. 2015. Subsurface seeding of surface harmful
algal blooms observed through the integration of autonomous gliders, moored
environmental sample processors, and satellite remote sensing in southern California.
Limnology and Oceanography 60, 754–764.
Sellner, K. G., Doucette, G. J., and Kirkpatrick, G. J. 2003. Harmful algal blooms: causes,
impacts and detection. Journal of Industrial Microbiology and Biotechnology 30, 383–
406.
Sosik, H. M. and Olson, R. J. 2007. Automated taxonomic classification of phytoplankton
sampled with imaging in flow cytometry. Limnology and Oceanography - Methods 5,
204–216.
Stumpf, R. P., Tomlinson, M. C., Calkins, J. A., Kirkpatrick, B., Nierenberg, K., Currier, R.,
and Wynne, T. T. 2009. Skill assessment for an operational algal bloom forecast system.
Journal of Marine Systems 76, 151-161.
Ussler, W., Preston, C., Tavormina, P., Pargett, D., Jensen, S., Roman, B., Marin, R., Shah, S.
R., Girguis, P. R., Birch, J. M., Orphan, V., and Scholin, C. 2013. Autonomous
application of Quantitative PCR in the deep sea: In situ surveys of aerobic
methanotrophs using the deep-sea Environmental Sample Processor. Environmental
Science and Technology 47, 9339-9346.
Velo-Suárez, L., Reguera, B., González-Gil, S., Lunven, M., Lazure, P., Nézan, E., and
Gentien, P. 2010. Application of a 3D Lagrangian model to explain the decline of a
Dinophysis acuminata bloom in the Bay of Biscay. Journal of Marine Systems 83, 242–
252.
Welschmeyer, N. A. 1994. Fluorometric analysis of chlorophyll a in the presence of
chlorophyll b and pheopigments. Limnology and Oceanography 39(8), 1985-1992.
Wynne, T. T., Stumpf, R. P., Tomlinson, M. C., Schwab, D. J., Watabayashi, G. Y., and
Christensen, J. D. 2011. Estimating cyanobacterial bloom transport by coupling remotely
sensed imagery and a hydrodynamic model. Ecological Applications 2, 2709-2721.
Yamahara, K. M., Demir-Hilton, E., Preston, C. M., Marin, R., Pargett, D., Roman, B.,
Jensen, S., Birch, J. M., Boehm, A. B., and Scholin, C. A. 2015. Simultaneous
monitoring of faecal indicators and harmful algae using an in-situ autonomous sensor.
Letters in Applied Microbiology 61, 130–138.
Zhan, P., Yao, F., Kartadikaria, A. R., Viswanadhapalli, Y., Gopalakrishnan, G., and Hoteit, I.
2015. Far-field ocean conditions and concentrate discharges modeling along the Saudi
Coast of the Red Sea. In: Intakes and Outfalls for Seawater Reverse-Osmosis
Desalination Facilities, Springer. pp. 501–520.
117
Remote sensing imagery of HABs
4 ACQUISITION AND ANALYSIS OF REMOTE SENSING

IMAGERY OF HARMFUL ALGAL BLOOMS
Raphael M. Kudela1, Richard P. Stumpf2, and Peter Petrov3

1
2
National Oceanographic and Atmospheric Administration, Silver Spring, MD USA
3
Regional Organization for the Protection of the Marine Environment, Kuwait

4.1 Introduction ...........................................................................................................................................119
4.2 Availability of data and software ..........................................................................................................121
4.3 Internet access to imagery .....................................................................................................................121
4.3.1 Access to L1, L2 and L3 images ...................................................................................................121
4.3.2 Data subscription services .............................................................................................................123
4.4 Software for processing satellite data....................................................................................................123
4.4.1 Hardware requirements .................................................................................................................124
4.4.2 SeaDAS software...........................................................................................................................124
4.4.3 BEAM and SNAP software ...........................................................................................................124
4.5 Algorithms used to detect blooms .........................................................................................................124
4.5.1 Atmospheric correction .................................................................................................................124
4.5.2 Algorithms .....................................................................................................................................125
4.5.2.1 Standard algorithms ...............................................................................................................125
4.5.2.2 High biomass algorithms .......................................................................................................126
4.5.2.3 Noctiluca algorithms ..............................................................................................................126
4.5.2.4 Trichodesmium algorithms ....................................................................................................127
4.6 Examples of algorithms .........................................................................................................................127
4.7 Summary for end-users .........................................................................................................................128
4.8 Useful links to satellite based ocean color data .....................................................................................130
4.9 References .............................................................................................................................................130
4.1 INTRODUCTION
Remote sensing was long considered an obvious tool for studying the distribution of harmful
algal bloom (HAB) organisms over larger spatial and shorter time scales than is possible with
ship-based sampling (Tester et al. 1991; Keafer and Anderson 1993). Legacy and next-
generation instrumentation and sensors, including SeaWiFS, MODIS, MERIS, and the OLCI
sensor on Sentinel-3, are dramatically improving the ability to determine constituents in the
coastal ocean. Satellite altimeters and scatterometers also provide geophysical fields such as
dynamic height (current patterns) and local winds (e.g. upwelling indices). Currently,
MODIS Aqua and VIIRS are still operational, while the replacement for MERIS, OLCI, is
now operational.
In some regions, remote sensing has already become a valuable tool for helping to predict the
onset, location, and transport of HABs. For example, in the Florida Shelf and Gulf of Mexico,
SeaWiFS and MODIS imagery has been incorporated into the U.S. NOAA HAB Bulletin
reports to identify potential red tide events, while feature-tracking has been used to follow
the spatial transport of these events (e.g. Tester et al. 1991; Tester and Steidinger 1997).
Progress has also been made on the use of inherent optical properties, derived from ocean
color inversion algorithms, to identify functional phytoplankton groups based on
fundamental biophysical properties (e.g. Lohrenz et al. 2003; Schofield et al. 1999).
Although multi-spectral scanners (e.g. MODIS) can be used to detect the reflectance of
chlorophyll a and other pigments with some accuracy, these efforts have been constrained by
the inability of the sensors to discriminate phytoplankton populations at the species level.
119
This is, of course, a fundamental requirement of HAB programs. Instead, progress has been
made by first linking specific water masses to HAB organisms and then identifying and
tracking that water mass with an appropriate remote sensing technique. In particular,
remotely-sensed sea surface temperatures (SST) have been used to follow the movement of
fronts, water masses, or other physical features where HAB species accumulate. A
fundamental problem for identifying HAB events, however, is that the imagery is still limited
to identification of chlorophyll or other biomass proxies rather than individual organisms (at
the genus or even functional group level).
Satellite imagery by itself will simply not provide the specificity needed to identify particular
organisms. Recent advances have begun to extend our ability to use remote sensing beyond
simple bulk chlorophyll measurements, however. For example, considerable work has gone
into identifying phytoplankton functional groups, or groupings of optically similar organisms
such as diatoms, dinoflagellates, and coccolithophorids. In some specific cases, optical
estimates (either from in-water measurements or remote sensing) can be used to identify
particular organisms, as some have unique optical properties. This includes Karenia brevis,
Trichodesmium spp., and cyanobacterial (blue-green) algal blooms (Alvain et al. 2008;
Stumpf et al. 2003; Westberry et al. 2005; Wynne et al. 2008). While diatoms and
dinoflagellates are very similar optically, and both can cause high biomass events, there
appear to be enough differences to discriminate between dinoflagellate- and diatom-
dominated surface waters as well (Dierssen et al. 2006; Palacios 2012).
In addition to the limitations of optical methods (including remote sensing) for the
identification of specific HAB organisms, another problem arises when imaging high
biomass blooms. When the biomass exceeds ~50 mg/m3 total chlorophyll, standard satellite
algorithms (e.g. MODIS OC3 or MERIS Algal-2) often fail because the water-leaving
radiances are high enough to trigger atmospheric correction failures. This results in
consistent underestimates of high biomass events in coastal waters. This can be remedied
relatively easily by the use of non-standard ocean color products. For example, Kahru and
Mitchell (2008) showed that the 250 m resolution bands on the MODIS satellite can be used
to develop a “particle index” that closely tracks red tides, while also providing the highest
possible spatial resolution. Hu et al. (2005) advocated the use of fluorescence bands for the
same reason; a second advantage is that only chlorophyll-containing particles strongly
fluoresce, solving the issue of working in optically complex coastal waters. Chen et al.
(2009) extended this by using multiple bands (fluorescence line height (FLH), backscatter,
etc) to develop a “machine learning” algorithm that can detect red tides. Given enough data it
is also possible to develop region-specific algorithms that work better than the global
methods (Kahru et al. 2012).
To summarize, using modern methods and data freely available from several ocean color
sensors, it is currently possible to identify high biomass HAB events (e.g., red tides),
although this requires application of non-standard products. The biomass estimates can be
further categorized into phytoplankton functional types, potentially useful for identifying
subclasses of blooms such as high biomass dinoflagellate events. These methods require
more effort and access to some laboratory or field optical measurements to parameterize the
models. It is not currently possible (and is unlikely to become possible) to identify species of
algae from space. When combined with other data streams such as currents, field
measurements, and in-water monitoring programs, unusual events can be identified, tracked,
and the subsequent impacts predicted if there are independent means of identifying the
organisms. This is most effective when remote sensing is combined with in-water
observations as part of an ocean observing program (see Chapter 3; Frolov et al. 2013;
Kudela et al. 2013).
120
4.2 AVAILABILITY OF DATA AND SOFTWARE
A large amount of satellite data on ocean colour is freely available for users, though some
expertise is needed to use and interpret it. Downloading data from the web sites of the ESA
(European Space Agency) and NASA (U.S. National Aeronautics and Space Administration)
is straightforward. Free software such as SeaDAS and SNAP makes it feasible to work with
ocean color data on personal computers, and the NASA Ocean Biology Processing Group
offers a forum for technical support. Data suitable for automated downloading are also
available from sites such as the U.S. National Oceanic and Atmospheric Administration
(NOAA) via ERDDAP. Here the focus is on the most common satellites and sensors
described above. As of 2016, this limits data primarily to NASA MODIS (and VIIRS)
sensors, and ESA MERIS (and OLCI) sensors. Expert users may also take advantage of
Landsat8 (US Geological Survey), Sentinel-2 (ESA), and the various sea surface temperature
sensors (e.g. AVHRR). The European Space Agency also supports SNAP for use with the
Sentinel platform; it is based on the same system as SeaDAS. Some example links for those
data products are provided here, but exhaustive descriptions of all sensors and products is
beyond the scope of this chapter.
4.3 INTERNET ACCESS TO IMAGERY
The simplest way to identify blooms with satellite imagery is to take advantage of standard
(global) processing that makes both data and imagery available via web browsers. As
described below, these standard products include RGB, chlorophyll, nFLH, and other
products such as light attenuation depth, particulate backscattering (a useful indicator of
particle load), colored dissolved organic material, and various other products. Data are
typically divided into categories. Level-1 (L1) are “raw” data, suitable for reprocessing by
the end-user. Level 2 (L2) include derived products such as chlorophyll, and have been
atmospherically corrected. The L2 files may also be projected to a standard map. Level 3
(L3) are binned in space, time, or both. L3 imagery is often at reduced spatial resolution
(typically 4 or 9 km). Standard L2 products are typically at 1 km resolution, and with some
sensors (MODIS, MERIS) reprocessing of L1 data can generate imagery at 250-300 m
resolution.
4.3.1 Access to L1, L2 and L3 images
NASA provides two portals that make it easy to access L1, L2 and L3 data. The WorldView
site (http://worldview.earthdata.nasa.gov ) provides a graphical interface that can display
many types of satellite (and other) data using a graphical user interface. This is an excellent
tool for quickly examining recent imagery, but the data are spatially binned and thus provide
limited spatial resolution. Figure 4.1 provides a snapshot of MODIS chlorophyll for the same
region used in section 4.3.
The NASA Ocean Biology Processing Group, OBPG, (http://oceancolor.gsfc.nasa.gov/cms/)
provides the same data as WorldView, along with many other satellite products including
MODIS Aqua and Terra, MERIS, and VIIRS. From this site it is possible to view L1, L2, and
L3 imagery at various resolutions, and to download data directly for further processing.
Figure 4.2 provides a screenshot of similar data as shown in Figure 4.1, but for the entire
globe. These data (Figure 4.2) can be directly downloaded.
The ERDDAP site (http://coastwatch.pfeg.noaa.gov/erddap/griddap/documentation.html)
provides access to a subset of the same data provided by WorldView and OBPG, but is set up
primarily for machine-to-machine access. Using this site it is possible to set up automated
extraction of a particular region, to download large amounts of data, and to create time-series.
121
Figure 4.1. A screen shot of the NASA WorldView site.
Figure 4.2. A screen shot of the NASA OBPG site showing L3 chlorophyll at 4 km resolution for 15
January 2016.
Finally, the Giovanni site, run by NASA (http://giovanni.sci.gsfc.nasa.gov/giovanni/)

provides many options for accessing NASA data, including generation of individual images,
time-series, and other specialized analyses. At this time (July 2017) Giovanni has not moved
most reprocessing of NASA ocean color data to the system. While Giovanni only provides
limited data (i.e. L3 data), it makes it simple to extract time series of several standard
products for a given location. All of the processing and data extraction is completed on
NASA servers, and the end-user is given both graphical images and the option of
downloading the original data. Figure 4.3 provides an example of a chlorophyll time-series
extracted from the Fujairah, UAE desalination intake site, for 1998-2015.
122
Figure 4.3. Time-series of chlorophyll extracted from the Fujairah, UAE desalination plant intake location.
The inset shows the full range of the data, while the main graph was truncated at 30 mg/m3.
4.3.2 Data subscription services

In addition to viewing/accessing data interactively, it is possible to define a specific region or
regions and set up a data subscription with the NASA OBPG system. This requires the user
to register with NASA, but the service is free and the registration process is straightforward.
Access to the service is at this website:
http://oceancolor.gsfc.nasa.gov/sdpscgi/public/subscriptions_home.cgi
Once registered, the user can request either a non-extracted or extracted region globally.
Currently MODIS Aqua, MODIS Terra, and VIIRS are available. Start and end dates are
selected, along with a geographical region. Various products are available (depending on the
sensor) including chlorophyll, nFLH, SST, and RGB. The user can request images (emailed
to your account), data files, or both. For data, the user can further specify L1 or L2 files. This
is particularly useful because the data will always be processed with the most current version
of the OBPG processing routines.
4.4 SOFTWARE FOR PROCESSING SATELLITE DATA
NASA recently (~2015) changed formats from Hierarchical Data Format (HDF) to netCDF
4.0. ESA also generates data as netCDF 4.0. The advantages of these formats are that they are
“self-contained”, including metadata and data within one “container”. Any software that can
access netCDF or HDF files can be used to process satellite imagery, including (e.g.) Python
and MATLAB, other programming languages, or specialized software designed to work
directly with the imagery.
Freely available, commonly used, and highly recommended satellite processing software are
provided by both ESA and NASA. These include SeaDAS software (NASA) and SNAP
123
software, which recently replaced BEAM (Brockman Consulting, created for ESA). The most
recent versions of SeaDAS are based on BEAM, so the two software sets are somewhat
interchangeable. SeaDAS provides extra processing options specific to NASA processing
(the OBPG), while BEAM/SBAP provides tools specific to ESA processing (MERIS, OLCI,
Sentinel-2).
4.4.1 Hardware requirements
Both SeaDAS and BEAM operate on standard personal computers (Windows, OS X, Linux).
The only requirement is that Java is available. Satellite data are often very large, however, so
it is recommended that a dedicated hard-drive be available for more extensive processing.
Both SeaDAS and BEAM process much of the data “on the fly”, as well, so they can tax
systems with limited amounts of RAM. For routine processing 4-8 GB RAM is usually
sufficient. For processing of time-series or very large (high resolution) imagery, it may be
necessary to access up to 32 GB RAM, particularly if the computer is also being used for
other tasks. Both programs provide rich processing capability that is beyond the scope of this
Chapter. Either can be used to visualize and process satellite data obtained at L0 (completely
unprocessed), L1, L2, or L3 levels.
4.4.2 SeaDAS software
SeaDAS is provided for free by the NASA OBPG group. It can be downloaded directly from
NASA and comes precompiled for various operating systems. SeaDAS can be installed as a
GUI (basic use) and as processing code (expert use) for processing of satellite data using
low-level scripts. NOTE: the Windows version can be used to visualize data, but does not
include the low-level processing programs.
The SeaDAS development released SeaDAS 7.4 in March 2017, which is built atop a
modified version of BEAM. The science processing code has been updated to reflect changes
recently implemented in production, providing bug fixes and support for the R2014.0
reprocessing for SeaWiFS. As long as the most recent version of SeaDAS is used, processing
should be identical to the NASA OBPG products.
4.4.3 BEAM and SNAP software
BEAM is an open-source toolbox and development platform for viewing, analyzing and
processing remote sensing raster data. Originally developed to facilitate the utilization of
image data from Envisat's optical instruments, BEAM supports a growing number of other
raster data formats such as GeoTIFF and NetCDF as well as data formats of other Earth
Observation sensors such as MODIS, AVHRR, AVNIR, PRISM, and CHRIS/Proba. Various
data and algorithms are supported by dedicated extension plug-ins. The primary tool is
VISAT - an intuitive desktop application to be used for visualization, analyzing and
processing of remote sensing raster data. As with SeaDAS, access to low-level processing
scripts are also available for expert users. BEAM was replaced by SNAP, which is
functionally similar but supports the most recent satellite platforms and processing methods.
4.5 ALGORITHMS USED TO DETECT BLOOMS
4.5.1 Atmospheric correction

Many regions where desalination is used, such as the Arabian Sea and Sea of Oman, are
subject to severe dust and other atmospheric conditions that cause problems for the standard
processing provided by NASA and ESA. These atmospheric correction issues are
exacerbated by high-biomass events, which often trigger correction failures (Loisel et al.
2013). It is possible to recover much of the “lost” data by switching to non-standard
atmospheric correction. This is time-consuming and requires optimization of the
124
methodology. A more straightforward approach is to use algorithms that rely upon the shape
of the ocean color spectra (spectral shape algorithms) that are applied with simple
atmospheric corrections, or with no atmospheric correction at all. These rely on fairly
standard products from the satellite processing, but are not routinely available without end-
user processing. Some of the algorithms below take advantage of these methods, but
“standard” products using NASA and ESA atmospheric correction are also discussed.
4.5.2 Algorithms
There is no HAB-specific remote sensing algorithm, but there are several methods that work
well for identifying high-biomass bloom events. Most of these belong to a class called
“spectral shape” algorithms. In contrast to chlorophyll methods, which generally use band
ratios (typically the ratio of blue to green light), spectral shape methods rely on changes that
occur over 3 or more bands (colors). The advantage of these algorithms is that they are much
less sensitive to atmospheric correction issues, since it is the shape rather than the absolute
values that identify the property of interest. These are also sometimes called linear baseline
algorithms since, functionally, they are often calculated as the height of a peak above a
baseline of two other wavelengths. This is how both FLH (fluorescence line height) and MCI
(maximum chlorophyll index) are determined. Table 4.1 provides a list of algorithms
commonly used for red tide detection.
Table 4.1. Commonly used remote sensing algorithms. The first three are available without
additional end-user processing; the remainder require some expertise.
Target Method Reference
Biomass Chlorophyll Standard product
Chlorophyll fluorescence Fluorescence line height Standard product
(FLH), normalized
fluorescence line height
(nFLH)
True-color image Red-Green-Blue (RGB), Standard Product
Enhanced Red-Green-Blue
(ERGB)
High biomass Maximum chlorophyll index Gower et al. 2005, Ryan et
(MCI), Red band difference al. 2014; Amin et al. 2012;
(RBD), maximum peak Matthews et al. 2012
height (MPH)
High biomass 250 m band subtraction Kahru et al. 2008
Floating Algae Floating Algae Index (FAI) Hu, 2009
Noctiluca Spectral Shape Astoreca et al. 2005
Trichodesmium Spectral Shape Hu et al. 2010
4.5.2.1 Standard algorithms

These algorithms are standard products provided by NASA and ESA, and do not require any
special processing or effort. They are commonly available from multiple locations on the
Internet.
125
4.5.2.2 High biomass algorithms

This group of algorithms relies upon changes in spectral shape as phytoplankton biomass
increases. As described in Ryan et al. (2014), the primary issue with these algorithms is that
they are sensitive to the biomass concentration. Below ~25 mg/m3 chlorophyll, MODIS FLH
works well. Above about 50 µg/L, MERIS MCI works well. It is possible to develop a single
algorithm that works for all biomass levels (Ryan et al. 2014) given higher resolution spectral
data, such as is available from HICO, but this will not be routinely available for 5-10 years,
when the next-generation satellites launch. An alternative to these methods is to create an
image of scattering (particles) from the “green” part of the spectrum. This is how the band-
subtraction method (Kahru 2008) works. The advantage of this method is that it can use the
high-resolution 250 m bands from MODIS, and is straightforward. This method can also be
applied to either atmospherically corrected or non-atmospherically corrected data. The
Floating Algal Index (FAI) is another variation on these methods that takes advantage of
reflectance in the near-infrared caused by surface scums or floating algae.
4.5.2.3 Noctiluca algorithms
The heterotrophic dinoflagellate genus Noctiluca is a relatively common bloom-forming
organism in many areas of the world, including the Gulf1 and Sea. It occurs as both “red” and
“green” varieties, with the less common green variety colored by a prasinophyte symbiont
(Harrison et al. 2011). Red Noctiluca is the unpigmented heterotrophic version, but it often
discolors the water (reddish, or “tomato soup” color) due to a combination of its ingested
prey items, internal symbionts, and high reflectance in the red and near-infrared. Remote
sensing has been used
successfully in previous
studies (e.g. Gomes et al.
2008; Piontkovski et al.
2011) to infer bloom
dynamics of Noctiluca by
combining general products
such as chlorophyll
climatologies and anomalies
with in-water data and
observations of currents, sea
surface temperature, and
mixed-layer depth.
Red Noctiluca has somewhat
unique properties (Figure
4.4), particularly the strong
absorption feature between
480-530 nm leading to a
sharp increase in reflection
from 520-580 nm (similar to
Figure 4.4. Red Noctiluca exhibits unusual optical properties, including the “red edge” effect in kelp
the inflection at around 530 nm, the sharp increase from 480-580 nm, and higher plants). It also
and the extremely high reflectance in the red and near-infrared. Adapted exhibits very strong
from Astoreca et al. 2005.
reflectance in the red and
1
126
near-infrared (NIR; 650-850 nm). While this makes it relatively easy to identify Noctiluca
from high-resolution in-water or even airborne data, the lower spatial and spectral resolution
of most satellites is problematic. The strong NIR reflectance is qualitatively similar to
suspended sediments, making it easy to misinterpret imagery, while the edge effect in the
~580 nm region requires increased spectral resolution.
4.5.2.4 Trichodesmium algorithms
Colonial cyanobacterium (blue-green algae) Trichodesmium blooms (mostly Trichodesmium
erythraeum) occur regularly in the Arabian Sea, Sea of Oman, Indian Ocean, Gulf of Mexico,
and Atlantic, and have been proposed to serve as a significant nitrogen source for some
regions. Detection of Trichodesmium blooms from space has been of interest since the 1980s.
Early attempts used empirical algorithms developed for the Coastal Zone Color Scanner
(CZCS). More recent efforts focused on the inherent and apparent optical properties (spectral
absorption, backscattering, and reflectance) of Trichodesmium and on the development of
empirical (Subramaniam et al. 2002) and semianalytical algorithms (Westberry et al. 2005)
for application to multispectral data from SeaWiFS and MODIS.
These algorithms were developed for optically simple open ocean waters, and frequently
over- and underestimate Trichodesmium blooms in more complex coastal waters. To address
this problem, Hu et al. (2010) proposed to use a spectral shape algorithm derived from the
same processing used for the Floating Algal Index (FAI). The algorithm is based on the
unique scattering and absorption
properties of Trichodesmium, but
is more complicated than simple
linear baseline methods in that it
uses multiple wavelengths in the
visible. While it works well with
atmospherically corrected data, as
with most spectral shape
approaches it is relatively robust to
errors in the correction, since the
diagnostic is the shape of the
spectra rather than the absolute
values of the wavelengths. Hu et al.
(2010) demonstrated that this
method works well in coastal
waters, although it does require
Figure 4.5. Spectral remote sensing reflectance from the MODIS manual processing of the data
sensor for Trichodesmium on the West Florida Shelf (WFS), with (necessary for inspection of the
corresponding high-resolution data from the Florida Keys spectral shape). The processing
collected on 7/1/1997. For comparison, Sargassum is shown from
the Western Gulf of Mexico (WGOM). All of these blooms show
steps start with simple atmospheric
up as positive FAI (as would dense blooms of Noctiluca and correction of the satellite data to
Cochlodinium). Trichodesmium has a unique spectral shape, remove the effects of Rayleigh
however, with a high-low-high-low-high pattern at 469-488-531- scatter. The FAI is then calculated.
555 nm (MODIS bands, dashed circles). Other sensors with bands For pixels with positive FAI, the
in this range would detect the same pattern. Figure adapted from
Hu et al. 2010.
spectral shape is then examined
(Figure 4.5).
4.6 EXAMPLES OF ALGORITHMS
The algorithms discussed above can generally be divided into two categories. Those that are
provided by NASA and ESA and do not require any extra effort, and those that require the
127
end-user to conduct additional processing. In practice it is common to use several algorithms

(images) and to compare the results from each to develop an informed understanding of the
dynamics of a region. Here an example from the Arabian Sea and Sea of Oman is used to
highlight how these algorithms compare. Abbreviations are given in Table 4.1.
Figure 4.6 provides an image from the Oman/UAE region for 23 December 2008, using the
NASA MODIS Aqua satellite. At the time, a massive red tide of the dinoflagellate
Cochlodinium had extended throughout the region. Standard products available from NASA
include truecolor, or RGB (panel A), Chlorophyll (panel C), Sea Surface Temperature (panel
D; note that MODIS Terra was used, and that some data are missing, denoted by the black
region), and normalized Fluorescence Line Height, nFLH (panel F). End-user processing of
the files produced an Enhanced RGB image (panel B), the RBD (panel E) and FAI (panel G)
images, and spectra (panel H) were extracted using the SeaDAS processing program.
Starting with the RGB and ERGB, it can be seen that it was a cloud-free day with little to no
dust in the atmosphere. Shallow regions where bottom-reflectance occurs can be seen in the
Strait of Hormuz and around the UAE coastline—these areas show up as “bright” areas in the
water, and can result in artificially high (false) chlorophyll values. The ERGB does better at
highlighting the extent of the bloom (compare A, B, C). In the chlorophyll image (C) there
are several regions with missing data (white). Since there were no clouds, this indicates a
failure of the chlorophyll algorithm, typically in very high (red tide) patches. This is also
apparent in nFLH (F). In contrast, RBD (E) does not have those missing data; this is
particularly important along the coast where impacts are likely to occur, since relying solely
on the chlorophyll (or nFLH) imagery would suggest that there are no data available.
Comparing C and F, some places where there is supposedly high chlorophyll have little or no
fluorescence, suggesting that the “chlorophyll” was contaminated by bottom reflectance.
Finally, FAI (G) picks out the most intense surface patches of chlorophyll. Two regions are
circled (dashed lines) and the spectra are shown in panel H. The southern patch, offshore of
Oman, shows the characteristic up-down-up spectral shape of Trichodesmium, suggesting
that in addition to the Cochlodinium red tide along the coast, there were also patches of
floating algae offshore. The other region (north) has a spectra indicative of dinoflagellates,
with a pronounced green peak and another red/near-infrared peak.
These images show why it is useful to compare several products in order to understand the
dynamics of the region. Any single product (algorithm) provides similar patterns, but does
not provide all the information available from the satellites. Of course, data are only available
when it is not cloudy, and it is critical to have local validation of the satellite products to
ensure that interpretation is correct.
4.7 SUMMARY FOR END-USERS
Numerous free data products are available that provide useful and relevant information for
tracking algal blooms in coastal waters, with many new sensors coming online. For new users,
an excellent starting-point is one of the web-based systems to routinely visualize the region
of interest and become familiar with the general oceanographic patterns and data availability.
From there, basic data analysis, such as generation of time-series for a given location, can be
attempted. If in-water or plant-based data are available it is straightforward to extract data
using, for example, the Giovanni website to explore correlations between satellite
observations and local conditions. As the end-user becomes more familiar with the available
128
Figure 4.6. MODIS Aqua data from 23 December 2008 during a red tide event. A: RGB; B: ERGB; C:
chlorophyll; D: SST; E: RBD; F: nFLH; G: FAI; H: spectra from the circled region in G. See the text for a full
description.
129
data and limitations, the final step is to acquire and process data directly, using either
standard algorithms, or taking advantage of the tools available in both SeaDAS and BEAM
for non-standard algorithms. An excellent resource for this last step are self-guided tutorials
available online (https://www.youtube.com/results?search_query=seadas). At the time of
publication, the main satellite processing tools (SeaDAS, BEAM, SNAP) are all based on the
same underlying computer code, so an end-user familiar with SeaDAS will quickly be able to
use any of these packages. With the basic information provided in this chapter, an end-user
can quickly progress from casual use of satellite images to routine production of regionally-
adjusted datasets suitable for research and monitoring.
4.8 USEFUL LINKS TO SATELLITE BASED OCEAN COLOR DATA
General
European Space Agency: http://sentinel.copernicus.eu
NASA: http://oceancolor.gsfc.nasa.gov/cms/
USGS: http://earthexplorer.usgs.gov
Data Access
NASA WorldView: https://worldview.earthdata.nasa.gov
NASA Ocean Biology Processing Group: http://oceancolor.gsfc.nasa.gov
NASA Data Subscriptions:
http://oceancolor.gsfc.nasa.gov/sdpscgi/public/subscriptions_home.cgi
Old Giovanni Site: http://gdata1.sci.gsfc.nasa.gov/daac-
bin/G3/gui.cgi?instance_id=ocean_8day
New Giovanni Site: http://giovanni.sci.gsfc.nasa.gov/giovanni/
NOAA ERDDAP:
http://coastwatch.pfeg.noaa.gov/erddap/griddap/documentation.html
SST and other Data Visualization: http://podaac-tools.jpl.nasa.gov/soto-
2d/soto.html?layers[]=jpl_ourocean_l4___sst___36000_x_18000___daynight&date=
2016-01-11
SST Data: https://podaac.jpl.nasa.gov/GHRSST
Software
NASA SeaDAS: http://seadas.gsfc.nasa.gov
ESA BEAM: http://www.brockmann-consult.de/cms/web/beam/
ESA SNAP: http://step.esa.int/main/toolboxes/snap/
4.9 REFERENCES
Alvain, S., Moulin, C., Dandoneau, Y., and Loisel, H. 2008. Seasonal distribution and
succession of dominant phytoplankton groups in the global ocean: A satellite view.
Global Biogeochemical Cycles, 22: GB3001.
Amin R., Zhou J., Gilerson A., Gross B., Moshary F., and Ahmed, S. 2012. Novel optical
techniques for detecting and classifying toxic dinoflagellate Karenia brevis blooms
using satellite imagery. Optics Express, 17: 9126-9144.
130
Astoreca, R., Rousseau, V., Parent, J.-Y., Lancelot, C., Ruddick, K., and Van Mol, B. 2005.
Optical properties of algal blooms in an eutrophicated coastal area and its relevance to
remote sensing. Proc. SPIE 5885, doi:10.1117/12.615160.
Chen, W. W., Hall, L. O., Goldgof, D. B., Soto, I. M., and Hu, C. 2009 Automatic red tide
detection from MODIS satellite images. Proceedings of the 2009 IEEE International
Conference on Systems, Man and Cybernetics, 1864-1868.
Dierssen, H., Kudela, R., Ryan, J., and Zimmerman, R. 2006. Red and black tides:
Quantitative analysis of water-leaving radiance and perceived color for phytoplankton,
colored dissolved organic matter, and suspended sediments. Limnology and
Oceanography 51, 2646-2659
Frolov, S., Kudela, R. M., and Bellingham, J. G. 2013. Monitoring of harmful algal blooms
in the era of diminishing resources: A case study of the U.S. West Coast. Harmful Algae,
21-22, 1-12.
Gomes, H. R., Goes, J. I., Prabhu Matondkar, S. G., Garab, S. G., Al-Azri, A. R. N., and
Thoppil, P. G. 2008. Blooms of Noctiluca miliaris in the Arabian Sea—An in situ and
satellite study. Deep-Sea Research 55, 751-765.
Gower, J., King, S., Borstad, G., and Brown, L. 2005. Detection of intense plankton blooms
using the 709 nm band of the MERIS imaging spectrometer. International Journal of
Remote Sensing, 26, 2005-2012.
Harrison, P. J., Furuya, K., Glibert, P. M., Xu, J., Liu, H. B., Yin, K., Lee, J. H. W.,
Anderson, D. M., Gowen, R., Al-Azri, A. R., and Ho, A. Y. T. 2011. Geographical
distribution of red and green Noctiluca scintillans. Chinese Journal of Oceanology and
Limnology 29, 807-831.
Hu, C., Muller-Karger, F. E., Taylor, C. J., Carder, K. L., Kelble, C., Johns, E., and Heil, C.
A. 2005. Red tide detection and tracing using MODIS fluorescence data: A regional
example in SW Florida coastal waters. Remote Sensing of Environment 97, 311-321.
Hu, C. 2009. A novel ocean color index to detect floating algae in the global oceans. Remote
Sensing Environment 113, 2118-2129.
Hu, C., Cannizzaro, J., Carder, K. L., Muller-Karger, F. E., and Hardy, R. 2010. Remote
detection of Trichodesmium blooms in optically complex coastal waters: Examples with
MODIS full-spectral data. Remote Sensing of Environment 114, 2048-2058.
Kahru, M., and Mitchell, B. 2008. Ocean color reveals increased blooms in various parts of
the world. EOS 89:170-171.
Kahru, M., Kudela, R. M., di Lorenzo, E., Manzano-Sarabia, M., and Mitchell, B. G. 2012.
Trends in the surface chlorophyll of the California Current: Merging data from multiple
ocean color satellites. Deep-Sea Research II 77-80, 89-98.
Keafer, B. A., and Anderson, D. M. 1993. Use of remotely-sensed sea surface temperatures
in studies of Alexandrium tamarense bloom dynamics. In: Smayda, T. J., Shimizu, Y.
(eds) Toxic phytoplankton blooms in the sea. Elsevier Science, Amsterdam, pp 763-768.
Kudela, R. M., Frolov, S. A., Anderson, C. R., and Bellingham, J. G.. 2013. Leveraging
ocean observatories to monitor and forecast harmful algal blooms: A case study of the
U.S. West Coast. IOOS Summit. Available at http://www.iooc.us/summit/white-paper-
submissions/community-white-paper-submissions/
131
Lohrenz, S. E., Weidemann, A.D., and Tuel, M. 2003. Phytoplankton spectral absorption as
influenced by community size structure and pigment composition. Journal of Plankton
Research 25(1), 35-61.
Loisel, H., Vantrepotte, V., Jamet, C., and Dat, D. N. 2013. Challenges and new advances in
ocean color remote sensing of coastal waters. In: Topics in Oceanography, E
Zmbbianchi (Ed.), InTech.
Matthews, M.W., Berard, S., and Robertson, L. 2012. An algorithm for detecting trophic
status (chlorophyll-a), cyanobacterial-dominance, surface scums and floating vegetation
in inland and coastal waters. Remote Sensing of Environment, 124, 637-652.
Palacios, S. L. 2012. Identifying and tracking evolving water masses in optically complex
aquatic environments. Ph.D. Thesis, University of California, Santa Cruz, 207 pp.
Piontkovski, S., Al-Azri, A., and Al-Hashmi, K. 2011. Seasonal and interannual variability of
chlorophyll-a in the Gulf of Oman compared to the open Arabian Sea regions.
International Journal of Remote Sensing 32, 7703-7715.
Ryan, J. P., Davis, C. O., Tufillaro, N. B., Kudela, R. M., and Gao, B.-C. 2014. Application
of the hyperspectral imager for the coastal ocean to phytoplankton ecology studies in
Monterey Bay, California. Remote Sensing 6, 1007-1025.
Schofield, O., Grzymski, J., Bissett, W. P., Kirkpatrick, G. J., Millie, D. F., Moline, M., and
Roesler, C. S. 1999. Optical monitoring and forecasting systems for harmful algal
blooms: possibility or pipe dream? Journal of Phycology, 35(6), 1477-1496.
Stumpf, R. P., Culver, M. E., Tester, P. A., Tomlinson, M., Kirkpatrick, G. J., Pederson, B.
A., Truby, E., Ransibrahmanakul, V., and Soracco, M. 2003. Monitoring Karenia brevis
blooms in the Gulf of Mexico using satellite ocean color imagery and other data.
Harmful Algae 2, 147-160.
Subramaniam, A., Brown, C. W., Hood, R. R., Carpenter, E. J., and Capone, D. G. 2002.
Detection of Trichodesmium blooms in SeaWiFS imagery. Deep-Sea Research II 49,
107-121.
Tester, P. A., and Steidinger, K. A. 1997. Gymnodinium breve red tide blooms: Initiation,
transport, and consequences of surface circulation. Limnology and Oceanography
42,1039-1051.
Tester, P. A., Stumpf, R. P., Vukovich, F. M., Fowler, P. K., and Turner, J. T.1991. An
expatriate red tide bloom: Transport, distribution, and persistence. Limnology and
Westberry, T. K., Siegel, D. A., and Subramaniam, A. 2005. An improved bio-optical model
for the remote sensing of Trichodesmium spp. blooms. Journal of Geophysical Research
110: C06012.
Wynne, T. T., Stumpf, R. P., Tomlinson, M. C., Warner, R. A., Tester, P. A., Dyble, J., and
Fahnenstiel, G. L. 2008. Relating spectral shape to cyanobacterial blooms in the
Laurentian Great Lakes. International Journal of Remote Sensing 29, 3665-3772.
132
HAB-related monitoring for desalination design and operation
5! HARMFUL ALGAL BLOOM-RELATED WATER QUALITY

MONITORING FOR DESALINATION DESIGN AND OPERATION
!
Siobhan F.E. Boerlage1, Loreen O. Villacorte2,3, Lauren Weinrich4, S. Assiyeh Alizadeh
Tabatabai2,, Maria D. Kennedy2, and Jan C. Schippers2
1
2
UNESCO-IHE Delft Institute for Water Education, Delft, The Netherlands
3
4
American Water, Voorhees, NJ USA
!
5.1! Intake feedwater characterization and water quality monitoring ..........................................................133!
5.2! Suitability of conventional online water quality parameters to detect HABs .......................................134!
5.2.1! Temperature ...................................................................................................................................135!
5.2.2! Salinity (conductivity) ...................................................................................................................135!
5.2.3! pH ..................................................................................................................................................135!
5.2.4! Dissolved oxygen ..........................................................................................................................135!
5.2.5! Turbidity ........................................................................................................................................136!
5.3! Overview of parameters to determine organic matter ...........................................................................136!
5.3.1! Advanced methods to determine algal organic matter ..................................................................142!
5.3.1.1! Liquid chromatography – organic carbon detection (LC-OCD) ...........................................142!
5.3.1.2! Transparent exopolymer particles (TEP) ...............................................................................145!
5.4! Measuring biofouling potential .............................................................................................................147!
5.4.1! Assimilable organic carbon ...........................................................................................................148!
5.4.2! Membrane fouling simulator .........................................................................................................152!
5.5! Fouling indices to measure particulate fouling potential ......................................................................153!
5.5.1! Silt Density Index ..........................................................................................................................153!
5.5.2! Modified Fouling Indices ..............................................................................................................158!
5.5.2.1! Modified Fouling Index-0.45.................................................................................................158!
5.5.2.2! Modified Fouling Index-UF ..................................................................................................159!
5.6! Summary ...............................................................................................................................................162!
5.7! References .............................................................................................................................................164!
5.1! INTAKE FEEDWATER CHARACTERIZATION AND WATER QUALITY MONITORING
Characterization of the raw seawater at plant intakes and monitoring to detect poor water
quality events including harmful algal blooms (HABs) is critical throughout the lifetime of a
desalination plant. HABs can result in a substantial increase in the organic and solids load in
the seawater feed to be treated at a desalination plant. This may result in an increase in the
clogging of granular media filters and accelerated particulate and/or (bio)fouling of
pretreatment and reverse osmosis membranes (see Chapter 2). Other feedwater quality
changes may be observed during or following a HAB event, such as a reduction in dissolved
oxygen levels and continued high concentration of organics due to decomposition of algal
matter by bacteria when the algal bloom degrades. Seawater in areas that are prone to algal
blooms or silt inflow etc. may require additional pretreatment if events are frequent and/or of
long duration (Chapter 9).
Depending on the project structure and site, an extensive seawater quality assessment study
may be conducted prior to plant design to provide the raw seawater quality design envelope,
select pretreatment and/or to obtain environmental permits. The study may include a review
of historical seawater quality such as the frequency and severity of algal blooms,
hydrodynamic conditions at the intake area and consider diffuse and point pollution sources
133
around the intake area which may impact water quality (e.g., ballast water exchange in port
areas that may promote HABs).
Subsequently, during plant operation, water quality monitoring of the intake feedwater
(online or through sampling) is essential for process control and contractual purposes to:
•! confirm the influent seawater is within the raw water design envelope in which
case the plant is required to meet production and product water quality
specifications;
•! allow analysis of data relative to baseline data collected prior to design to identify
seasonal or diurnal trends and to detect poor water quality events such as HABs;
•! optimize pretreatment processes in response to a deterioration in feedwater quality
due to events such as a HAB in order to maintain production and water quality
targets; and
•! allow the operational performance of downstream plant unit processes to be
assessed against the intake values e.g. dissolved air flotation (DAF), granular
media filtration (GMF), seawater reverse osmosis (SWRO), residual handling.
During a HAB event, desalination plant operators require methods to measure the
concentration of algae and algal organic matter (AOM), its fouling constituents, and any
increases in membrane fouling potential and other HAB associated water quality changes in
the raw water and treatment process streams. This allows operators to respond to a bloom in a
timely manner to optimize plant operation to avoid disruption to supply.
This chapter examines the ability of conventional water quality parameters routinely used in
desalination feedwater monitoring during design, piloting, and SWRO plant operation to
detect HABs and characterize water quality during a bloom. Although the focus is on water
quality parameters monitored at the intake, key parameters commonly used within SWRO
desalination plants to assess the reduction in the particulate fouling potential (i.e. the Silt
Density Index (SDI)) and organic load of the raw water (e.g. total organic carbon) through
pretreatment are discussed. More sophisticated and lesser known techniques that are under
development to directly measure AOM or assess the biofouling and/or particulate fouling
potential of feedwater are presented to provide an overview to plant operators when
evaluating additional tests during a HAB event. Applications of these techniques are
presented to illustrate their use.
Methods which directly identify the presence of HAB species in the seawater feed through
algal cell identification and enumeration or an increase in algal productivity (such as
chlorophyll-a measurements or advance warning through remote sensing) are discussed in
Chapters 3 and 4, respectively.
5.2! SUITABILITY OF CONVENTIONAL ONLINE WATER QUALITY PARAMETERS TO DETECT
HABS
Online instrumentation typically installed at a SWRO desalination plant intake to

continuously monitor feedwater quality may include temperature, conductivity, dissolved
oxygen (DO), pH, turbidity, residual chlorine and dissolved hydrocarbons. These online
parameters can be monitored, trended and viewed in the control room to assess raw feedwater
quality and the impact of water quality changes on the efficiency of unit processes such as
pretreatment and desalination. None of the aforementioned parameters are specific to algal
blooms. Changes in the core physiochemical parameters monitored (temperature,
conductivity, DO, pH, and turbidity) can be caused by other factors such as pollution events
and/or marine hydrodynamics, thus the interpretation of these water quality variables can be
134
complex. At best they can indirectly indicate conditions that favor a bloom, the presence of
algal blooms, or associated with the termination of a bloom, such as DO depletion as
discussed below.
5.2.1! Temperature
Temperature, a key desalination water quality process parameter, is trended at almost all
desalination plants. Some algal species are known to bloom under specific ranges of
temperature, e.g., Trichodesmium, commonly responsible for red tide outbreaks in the Gulf,
favors seawater temperatures ranging from 20 – 34˚C (Thangaraja et al. 2007 cited in Zhao
and Ghedira 2014). Alternatively, changes in temperature may indicate downwelling or
upwelling events that can bring nutrients or established blooms to the plant intake.
5.2.2! Salinity (conductivity)
Salinity at plant intakes is typically calculated from online conductivity measurements using
a correlation of total dissolved solids (TDS) and conductivity.1 Many algal species have a
broad tolerance for salinity, particularly those endemic to estuarine regions. Nevertheless,
there are salinity ranges that are optimal for a HAB, as well as those that are too high or low
for rapid growth. A particular salinity range in combination with temperature may promote
growth of a specific bloom-forming algal species or break the dormancy of an algal cyst
(Chapter 1). Alternatively, a change in salinity outside the tolerance range of an algal species
may result in the termination of a HAB or changes in the species assemblage in a bloom
community. Hence, monitoring changes in salinity and temperature at a plant intake may be
useful where an algal blooming species routinely occurs.
5.2.3! pH
Monitoring of pH to detect blooms is complex and typically not useful for the purposes of
detecting or characterizing HABs. The pH of seawater is around 8.2 and does not normally
vary substantially due to the vast buffering capacity of the seawater bicarbonate-carbonate
system. As a result of estuarine input, however, pH may increase or decrease depending on
the salinity and evaporative effects on the estuary. Dense algal blooms in shallow coastal
areas with limited tidal exchange may also cause a local pH change – leading to both
increases and decreases in pH. As algae consume carbon dioxide during photosynthesis
during daylight hours, removing it from the water, less carbonic acid dissociates and a pH
rise can occur. During the night, when there is no photosynthesis, CO2 is released by
respiring cells, leading to a decrease in pH. When the bloom ends and the algal biomass
degrades, CO2 is again released and pH decreases. Diurnal changes in pH may subsequently
occur due to bacterial decomposition and aerobic respiration.
5.2.4! Dissolved oxygen
As with pH, diurnal changes in online DO may also reflect algal photosynthesis and
respiration as well as tidal exchange, which can replenish DO levels. Elevated DO may be
observed in the photic zone (or sunlight zone) through photosynthesis. In contrast, a rapid
decrease in DO may occur in stratified coastal water due to the respiratory activity of a bloom,
or to bacterial decomposition of algal biomass as blooms age and decay, sinking to the
seabed and sometimes leading to hypoxic conditions (<0.5 mg/L) or even anoxia. Hence, a
decline in online DO may be observed depending on the depth of the intake, and DO trends
may be used to indicate a HAB. Anoxic events due to DO depletion were observed at the La
Chimba SWRO desalination plant, leading to the presence of hydrogen sulphide in the
1
Caution should be taken when comparing this salinity with that measured by conductivity, temperature, and
depth (CTD) sensors at sea which calculate salinity using the Practical Salinity Scale (Boerlage 2012).
135
feedwater due to the proliferation of sulphate-reducing bacteria that prolifereate under low
DO conditions (see Chapter 11, section 11.6). Low DO can also be caused by pollution
events and therefore needs to be assessed in conjunction with other water quality testing such
as SDI, total organic carbon, and cell counts.
5.2.5! Turbidity
Turbidity, based on the amount of visible light (400 – 700 nm) scattered by particles in
solution, generally increases with a greater load of suspended solids. While turbidity provides
some information on particle concentration and water clarity 2 , measurements can be
inaccurate at both high and low levels. Scattered light is the aggregate response for all
particles; the physical properties of particles such as color, shape, size distribution and
numbers all affect the light scattering properties of a solution. When particles are in the
wavelength range of visible light, turbidity is at a maximum (Edzwald and Tobiason 2011).
In contrast, turbidity meters have been reported to be insensitive to small colloidal particles -
particles with a diameter less than half the wavelength of visible light (0.2 µm) will not
produce significant scatter (Kremen and Tanner 1998). Therefore, turbidity will decrease for
smaller particles due to poor light scattering, while if the total particle mass remains constant,
turbidity will also decrease for larger particles due to the decreased number or concentration
(Edzwald and Tobiason 2011). Moreover, as noted by Tabatabai (2014a), transparent
exopolymer particles (TEP), a component of AOM, do not absorb visible light. In some cases,
high chlorophyll correlates well with high turbidity, but this is not always the case, as
suspended materials like sand or clay will have similar light-scattering properties as algal
cells. Turbidity is thus only a general indicator of algal blooms and cannot be relied upon to
detect a bloom or an increase in particles associated with a bloom. Other confirmatory
measures are needed such as those provided by chlorophyll sensors, or through direct cell
counts.
5.3! OVERVIEW OF PARAMETERS TO DETERMINE ORGANIC MATTER
Total organic carbon (TOC) and dissolved organic carbon (DOC) are common measures of
the concentration of organics at desalination plant intakes and are used to assess the
efficiency of pretreatment processes in removing organics. Ultraviolet absorption at 254 nm
(UV254) and the related specific ultraviolet absorbance (SUVA) are used to a lesser extent.
The maximum, average, or median concentration of these parameters may be provided as
part of the raw seawater design envelope to characterize the organic load in the seawater to
be desalinated.
Standards from ASTM International and Standard Methods for the Examination of Water and
Wastewater (APHA 2012) are often recommended by membrane manufacturers for the
analysis of these aforementioned parameters. In addition, these protocols may be specified in
design and/or operation and maintenance (O&M) contracts for analysis of the saline feed and
desalination process streams (see Table 5.1). In general, these parameters can be quickly and
reliably determined by laboratories experienced in the analysis of saline matrices with some
parameters determined by on site laboratories. These parameters are routinely measured on a
weekly or monthly basis at the seawater intake depending on the site (e.g. to monitor the
potential for HABs and/or pollution or to check the feedwater is within design specifications).
Monitoring may also occur upstream and downstream of pretreatment processes to assess
performance or in the RO feedwater for compliance with membrane guarantees. The
2
Secchi disks - another method to measure water clarity are discussed in Chapter 3.
136
frequency of monitoring may increase when a bloom event is forecast or during a bloom to
adjust operating parameters for process steps accordingly.
The aforementioned techniques measure aggregate organic matter and therefore provide no
specific information as to the composition or concentration of potential AOM foulants
produced during an algal bloom. TOC measures both organic matter derived from natural
processes such as HABs, bacteria, riverine flushing or through direct anthropogenic input.
Therefore, spikes in feedwater TOC may be due to an algal bloom and/or pollution or other
events. Moreover, increases in TOC are not always observed during a bloom event.
Feedwater TOC increased significantly in the Red Sea off the coast of Saudi Arabia during
an algal bloom (pers. com. N. Nada) and also in the source seawater during testing of Long
Beach Water Department’s demonstration seabed infiltration gallery during blooms (see
Chapter 6 Section 6.4.1.6). Yet, in other cases, TOC has not significantly increased during a
bloom. The latter may be attributed to underestimating TOC if floatable organics are not
captured during sampling or sample homogenization (Table 5.1). Measuring TOC removal to
assess the efficiency of pretreatment processes is also inaccurate due to the difficulties in
measuring low-level TOC residuals in seawater process streams. In high temperature
catalytic oxidation measurement of organic carbon in seawater, the high salt concentration (in
the range of 30,000 to 45,000 mg/L) compared to a few mg/L of organic carbon, results in
low accuracy and high limits of detection for TOC measurements. Consequently, TOC results
are often interpreted in conjunction with chlorophyll-a and algal counts (where available) and
other more standard plant water quality monitoring parameters such as SDI, turbidity, total
suspended solids, and DO to assess the occurrence of algal blooms in the intake water.
Although, UV254 and TOC can be determined online, these instruments are not frequently
installed at SWRO plants. As with TOC, UV254 is an aggregate parameter, but only for
selected organic constituents such as lignin, humic acids, and various aromatic compounds
which strongly absorb UV radiation (APHA 5910B). SUVA, the quotient of UV254 and DOC,
provides an indication of dissolved natural organic matter measured by UV254 compared to
the overall dissolved organic concentration and can be used to indicate whether the dissolved
organic matter is primarily derived from natural processes occurring at the intake rather than
anthropogenic sources. Algal organic matter may contain some UV-absorbing compounds,
but their proportion among AOM components significantly varies among species releasing
them. Moreover, carbohydrates (i.e., polysaccharides produced by algae and marine bacteria)
do not absorb UV (APHA 5910B) and therefore UV254 cannot be relied upon to monitor or
measure AOM in the raw water or treatment process streams.
More sophisticated techniques to characterize the composition and concentration of AOM in
seawater have been developed or are under development such as liquid chromatography -
organic carbon detection (LC-OCD) to determine biopolymers or measurement of TEP,
(compounds demonstrated to promote fouling in SWRO and ultrafiltration (UF) – see
Chapter 2) and other AOM components as shown in Figure 5.1. While these methods offer
more targeted information (and higher sensitivity) of AOM constituents and potential
foulants in a feedwater, the degree of difficulty and cost in determining them is
correspondingly higher. As yet, samples need to be sent to specialized laboratories with
experience in determining these parameters, which are limited in number, resulting in delays
to obtain results. Hence, at present they cannot be employed directly as a trigger to alert a
plant of a bloom in the incoming feedwater or to adjust process parameters during plant
operation. Nonetheless, these parameters are expected to be a key factor in developing an
understanding in controlling AOM fouling in seawater UF and RO systems as they allow
quantification of specific foulant components of AOM not detectable by standard monitoring
parameters, including those components which are likely to cause membrane fouling.
137
Table 5.1 provides a summary of conventional analytical techniques used in desalination to

determine organic matter in seawater along with the more advanced techniques to determine
natural organic matter (NOM) and AOM, comparing the principle of each method,
interferences, organic matter fraction identified, and operator skill required for the test. The
more advanced techniques are discussed in detail in the following sections.
It should be noted that all these methods were originally developed for characterization of
NOM regardless of its origin (microbial or terrestrial input); however, during a HAB a
significant fraction of NOM will be comprised of AOM. TEP present in seawater can be a
mixture of those produced by bacteria, HABs, and shellfish. While there is no technique to
distinguish TEP based on their origin, during a HAB most TEP will be generated by algae.
Following the collapse of an algal bloom the succession of bacterial species which can thrive
on decaying AOM may release organic matter extracellularly including TEP and contribute
to the organic load in the source seawater.
Figure 5.1. Major components of aquatic organic matter and corresponding analytical techniques for
their identification and quantification. Legend: LC is liquid chromatography with inline detectors for
organic carbon (OCD), UV absorbance at 254 nm (UVD,) and organic nitrogen (OND). FEEM is
fluorescence excitation-emission matrices. TEP refers to transparent exopolymer particles measured
with a 0.4 µm or 10 kDa membrane.
138
Table 5.1. Comparison of conventional and advanced parameters to measure organic matter in seawater and feedwater.
Parameter Standard / Reference Basis of Method Organic matter (OM) Interferences / Issues Analysis time /
identified Operator skill
Conventional water quality parameters
a
Total organic matter ASTM D2579 , D4129, Homogenize sample, acidify and Aggregate parameter 510 B – potential loss of 1-2 hours including
(TOC) mg/L D4839, APHA5310B, sparge to remove inorganic dissolved and floating organic matter, preparation / routine
5310C carbon. Oxidation of organic particulate organic particle size limited by inexpensive test.
matter by (i) high temperature carbon. injection syringe Standard test for most
combustion or (ii) persulfate-UV needle≈500µm 5310C – high saline water
or heated-persulfate. Detection Non specific OM test. turbidity, high chloride laboratories.
of CO2 via NDIRb or requires modification of
conductivity. Limit of detection Natural and method.
defined by instrument and anthropogenic in origin.
method.
Dissolved organic carbon As above Filtration through 0.45 µm filter Dissolved organic 5310C – high chloride As above.
(DOC) mg/L and analysis of filtrate as above. carbon <0.45µm. Non requires modification of
specific OM test. method.
Natural and
anthropogenic in origin.
Ultraviolet absorption at APHA 5910B Filtration through 0.45µm. Specific to aromatic Turbidity, UV absorbing Less than one hour
254nm (UV254) cm-1 Absorbance at 254 nm measured organics. Carbohydrates inorganics (ferrous iron, high routine inexpensive
with UV/Vis spectrophotometer. (e.g. polysaccharides) bromide concentrations). test.
and carboxylic acids do
not absorb UV light and
are not measured.
Specific ultraviolet APHA Calculated from UV254/DOC As above. As per TOC test.
absorbance (SUVA) 5910B/APHA5310
L/mg.m
Advanced organic matter (OM) characterization methods
Liquid chromatography – Huber et al. 2011 Size-exclusion chromatography Chromatographable Chromatographic columns Expensive technique
organic carbon detection followed by in line detectors for hydrophilic DOC may adsorb or trap which requires
LC-OCD (i) organic carbon, (ii) UV fractions such as hydrophobic and/or high specialized equipment
absorbance at 254 nm and (iii) biopolymers (proteins molecular weight OM and high degree of
organic nitrogen. Area and polysaccharides), components (biopolymers). operator skill; sample
integration of identified humic substances, Particulate OM and large measurement and
chromatogram peaks using building blocks, low molecular TEP > 0.45 µm analysis time is up to 5
customized software. Pre- molecular weight acids excluded due to inline hours. Shipping of
filtration of sample through and neutrals. filtration through 0.45 µm sample is location
0.45µm filter. Limit of detection filter (the standard method dependent.
is fraction specific and may be for LC-OCD) or >2 µm
affected by water matrix. without filtration. No
standard method, nor
calibration when bypassing
the 0.45 µm filter.
Fluorescence excitation- Coble et al. 1993 Sample is excited to a specific Humic-like and protein- Samples should be diluted Inexpensive technique
emission matrices wavelength at which AOM like compounds below 1 mg C/L. Very that requires specialized
(F-EEM) fluorophores absorb light and <0.45µm Fluorescence sensitive to sample equipment and medium
subsequently emit the light at index (FI)c contamination. Does not degree of operator skill;
longer wavelength. This classification: FI = 1.7- detect non-fluorophore OM 0.5-1 hour of analysis
technique is performed using a 2.0 (microbial origin); F components. In principle, it time.
spectrofluorometer across a = 1.3-1.4 (terrestrial covers only humic and
spectrum of light wavelengths. origin) fulvic-like components, part
The acquired data is then plotted of biopolymers (proteins)
in a 3D fluorescence contour for and part of TEP
analysis. (glycoproteins).
Advanced organic matter (OM) characterization methods (continued)
Transparent exopolymer Passow and Alldredge Retention on 0.4-µm filter, TEP (>0.4µm) Overestimation of results Medium degree of
particles (TEP0.4µm) 1995 staining with Alcian blue (pH due to interference of operator skill and
2.5), sulfuric acid digestion and dissolved ions. Exclusion of standard lab equipment
absorbance measurement at the colloidal components required, risk during
787 nm. Calibration with (TEP precursors). A concentrated acid
Xanthan gum. proposed modification for handling, 2-3 hours of
salinity control by rinsing analysis time.
with ultrapure water has Calibration performed
been introduced (Villacorte for every batch of dye
et al. 2015c). solution takes 4-5
hours.
TEP10kDa Villacorte et al. 2015c Retention on 10 kDa membrane, TEP (>0.4 µm) and their Overestimation of results Medium degree of
resuspension in ultrapure water precursors (10kDa- due to release of intracellular operator skill and
by sonication, Alcian blue (AB) 0.4µm) AOM during the sonication specific lab equipment
staining, removal of precipitates step. Such release may vary required. 3-4 hours of
through 0.1-µm filter and significantly with species of analysis time.
absorbance measurement of algae. Calibration performed
residual AB in the filtrate at for every batch of dye
610 nm. TEP concentration is solution takes 4-5
based on reduction of AB hours.
absorbance and calibration with
Xanthan gum standard.
a
Historical method remains in use but withdrawn from ASTM
b
NDIR - non dispersive infrared analyzer
c
Fluorescence index (FI) = ratio of fluorescence intensity at emission wavelength of 450 nm to that at 500 nm obtained at excitation wavelength of 370 nm.
It should be noted that a loss in concentration due to adsorption of AOM components to

sample bottle walls and/or degradation by bacteria can be an issue for all the analytical
methods mentioned in Table 5.1 (TOC, UV, LC-OCD, FEEM, TEP). Therefore, samples
should be cooled (typically at 5oC) and analyzed as soon as possible after collection.
Concentration loss may vary from sample to sample. It is generally acceptable to follow the
standard protocol for preservation and transport for DOC/TOC samples for all these
analytical methods.
5.3.1! Advanced methods to determine algal organic matter
Advanced methods to determine NOM include FEEM, LC-OCD and TEP; the latter two can
provide more information on the composition and concentration of fouling AOM compounds
produced during a bloom. FEEM analysis provides insight into the presence and
concentration of humic-like, fulvic-like and protein-like organic matter. As such, only a
fraction of AOM, i.e. (glyco)proteins, can be determined by FEEM analysis resulting in an
underestimation of the concentration and composition of organic matter during a HAB (see
Figure 5.1 and Table 5.1). In contrast LC-OCD covers a wider range of NOM (and
consequently AOM) fractions in terms of molecular weight, aromaticity and protein content,
while TEP methods provide more information on a subset of biopolymers i.e. glycoproteins
and acidic polysaccharides.
5.3.1.1! Liquid chromatography – organic carbon detection (LC-OCD)
Liquid chromatography-organic carbon detection (LC-OCD) is a semi-quantitative technique
for identifying and measuring different components of NOM in aquatic environments. The
LC-OCD technique combines the physical separation capabilities of liquid chromatography
with mass balancing on the basis of chromatographable dissolved organic carbon (CDOC) for
identification and measurement of various fractions of NOM of different molecular weight.
Figure 5.2 shows where the LC-OCD technique stands in the suite of analytical tools
available for characterization of NOM. The technique was developed by Huber and co-
100 workers based on the
TOC
Gräntzel thin film
LC0OCD
reactor for high
Mass!Balancing!Potential!(in!%)
75 ! COD sensitivity carbon

detection in the early
Phenol!
index 1990s (Huber and
Frimmel 1991).
50 Single!compound!
Aggregate!parameters
analysis Further developments
of this technique have
HPLC led to substantial
25 reduction in footprint
! LC0MS
of the instrument (60 x
GC0MS 60 cm) and the
0 introduction of a
0 25 50 75 100 multi-detector system.
Compound!Identification!Potential!(in!%) Current generations of
Figure 5.2. Position of LC-OCD in the suite of analytical tools for natural LC-OCD utilize size-
organic matter characterization (modified from www.doc-labor.de). Legend: exclusion chroma-
HPLC is high performance liquid chromatography, LC-MS is liquid
chromatography coupled with mass spectrometry and GC-MS is gas
tography (SEC) and
chromatography coupled with mass spectrometry. inline organic carbon
detection, UV254
142
detection and organic nitrogen detection (Huber et al. 2011). The detectors measure signal
response of organic carbon, UV and organic nitrogen as a function of retention time of the
organic component in the chromatographic column. DOC is determined using a bypass mode
on the instrument.
Separation is achieved by differential exclusion of organic matter fractions through diffusion
of hydrophilic dissolved organic carbon molecules (with 0.45 µm prefiltration) into resin
pores of the column beads. In principle, larger molecules elute first as they cannot penetrate
deep into the pores of the beads, while smaller molecules diffuse into the pores and elute later.
Consequently, low molecular weight organics have higher retention time than compounds
with high molecular weights. A customized software program (CHROMCalc) is used for data
processing. Concentrations of organic carbon and organic nitrogen for different fractions is
obtained by area integration of the chromatograms and with reference to calibration with
standard organic compounds (International Humic Substances Society standards- Humic and
Fulvic acids) (Huber et al. 2011). An example of a chromatogram is given in Figure 5.3 for
North Sea water. Chromatographable dissolved organic matter is fractionated based on
molecular weight, and to some extent in terms of major functional groups based on UV254
absorbance into five classes of compounds. The high molecular weight fraction (>>1 kDa)
comprises non-UV absorbing biopolymers such as proteins and polysaccharides. The low
molecular weight fractions (<1.2 kDa) comprise UV-absorbing humic substances and
building blocks as well as biogenic substances such as low molecular weight organic acids
and neutrals. A description of chromatographable dissolved organic matter fractions resolved
by LC-OCD is presented in Table 5.2. Hydrophobic compounds e.g., natural hydrocarbons
and sparingly soluble humics, do not elute from the column and are therefore excluded from
the chromatograms and are referred to as non chromatographable hydrophobic organic
carbon (HOC). HOC is determined as the difference between DOC and CDOC.
Table 5.2. Description of dissolved organic matter fractions measured by LC-OCD (DOC-
Labor; Huber et al. 2011)
Organic fraction Typical size range Typical composition
(Da)
Biopolymers > 20,000 Very high in MW, hydrophilic, not UV-absorbing;
typically polysaccharides, but may also contain proteins,
amino sugars, polypetides (quantified on basis of OND),
and aminosugars
Humic substances 500 – 1200 Humic and fulvic acids

(HS)
Building blocks 300 – 500 Sub-units of HS and considered to be natural breakdown
products of humics through weathering and oxidation
LMW neutrals < 350 Low molecular weight, weakly or uncharged hydrophilic
or slightly hydrophobic (amphiphilic) compounds appear
in this fraction, e.g., mono-oligosaccharides, alcohols,
aldehydes, ketones, amino acids
LMW acids < 350 Aliphatic, LMW monoprotic organic acids co-elute due
to an ion chromatographic effect. A small amount of HS
may fall into this fraction and is subtracted on the basis of
SUVA ratios
LMW is low molecular weight
143
Figure 5.3. Typical signal responses generated by organic carbon detector (OCD) and UV254nm detector
(UVD) in coastal North Sea water samples and the assignment of the different organic matter fractions. The
dip in OCD chromatogram within the LMW neutrals fraction is due to salinity, typical in seawater samples.
Adapted from Villacorte et al. 2010.
High resolution LC-OCD, a recent development of DOC-Labor, provides better separation of

biopolymers from humic substances using two columns instead of one. The first is only used
to separate humic substances and low molecular weight compounds, whereas the separation
of high molecular weight material is done on the second column. This gives semi-quantitative
information on the molecular weight distribution of biopolymers into four fractions namely,
1000 kDa – 2 µm, 100-1000 kDa, 10-100 kDa, and < 10 kDa. Fractions are quantified based
on area integration (area boundaries are defined by pullulan3 standards). However, resolution
is poor for the fraction < 10 kDa, as this fraction is superimposed by the building blocks and
low molecular weight acids fractions. For this fraction the quantification is rather arbitrary
and bias may well exceed 50%.
LC-OCD has been used to fractionate NOM in feedwater, brine and RO permeate and can be
used to characterize and quantify AOM released during blooms in freshwater and seawater.
Outside algal bloom periods, NOM in coastal seawater is mainly composed of low molecular
weight aromatic compounds (e.g., humic substances) (Jeong et al. 2013). Algal-derived
organic matter mainly comprises high molecular weight, hydrophilic, non-UV absorbing
compounds, i.e., polysaccharides and proteins (Tabatabai 2014a; Villacorte et al. 2015a). A
3
Pullulans are non-ionic extracellular polysaccharides excreted by the fungus Aureobasidium pullulans.
144
substantial increase in the biopolymer fraction is observed during severe algal blooms as
shown in Figure 5.3 where an increase in biopolymer concentration was observed; from
140 ppb C in February 2009 outside of a bloom to 277 ppb C during a spring bloom in March
2009. Considering that high molecular weight AOM has been shown to cause irreversible
fouling of low pressure membranes (microfiltration and ultrafiltration) and is likely to deposit
and/or accumulate on SWRO membranes, monitoring the biopolymer fraction of organic
matter in seawater is a promising indicator of organic and biological fouling potential of algal
bloom-impacted waters.
The amount of proteins and polysaccharides in algal biopolymers can be estimated using LC-
OCD. The organic carbon concentration of protein can be estimated based on the organic
nitrogen content of the biopolymer fraction. Protein concentration is calculated by assuming
that all organic nitrogen detected by the organic nitrogen detector between 25 and 42 minutes
retention time are bound to proteinaceous compounds. Typically, proteins contain 14.5 -
17.5% nitrogen and 49.7 - 55.3 % carbon (Rouwenhorst et al. 1991). Hence, the C:N ratio of
the protein fraction of biopolymers can be estimated as 3:1. The estimated protein
concentration in mg C/L is calculated by multiplying the organic nitrogen concentration (in
mg N/L) by 3. From there, the polysaccharide concentration is calculated by subtracting the
organic carbon concentration of protein from that of the biopolymer concentration.
Since AOM may comprise large macromolecules (e.g., TEPs), LC-OCD analyses should be
preferably performed without 0.45 µm inline filtration of samples – a standard pretreatment
protocol for LC-OCD analysis (Villacorte et al. 2015b): however, removing the pretreatment
step has been shown to cause clogging issues in the size exclusion columns. The theoretical
maximum size when performing chromatography without sample pre-filtration is 2 µm,
which is based on the pore size of the frit at the column entrance (S. Huber pers. com.).
5.3.1.2! Transparent exopolymer particles (TEP)
Since the discovery of TEPs more than two decades ago, various quantification methods have
been developed, all of which are based on staining with cationic Alcian blue (AB) dye. This
particular dye is known to be highly selective and forms insoluble complexes with TEP that
cannot be easily reversed by subsequent treatment. In aqueous solutions without extra
electrolytes, AB specifically binds with functional components such as acidic
polysaccharides, glycoproteins and proteoglycans, resulting in the formation of neutral
precipitates (Ramus 1977). AB does not react with nucleic acids and neutral biopolymers.
The staining ability of AB depends on the type and density of anionic functional groups
associated with TEP in the sample and to a large extent on the pH and ionic strength of the
sample solution (Horobin 1988). In high ionic strength solutions, AB molecules
spontaneously precipitate due to interactions with dissolved salts resulting in the formation of
flocs not associated with TEP. This is considered the main drawback of the application of AB
staining for TEP measurements in seawater. To minimize measurement artifacts due to
coagulation, AB staining solutions should always be pre-filtered and should not be directly
applied to solutions with high salinity.
So far, five methods have been developed to quantify TEP and their precursors. The first TEP
method is based on optical microscopic enumeration. This method provides useful
information on the size-frequency distribution of TEP in seawater (Alldredge et al. 1993), but
it is laborious, complicated, time consuming and is not always feasible, especially for
samples with low concentration and smaller size range (<2 µm) of TEP. All the succeeding
methods based on semi-quantitative spectrophotometric techniques were able to address these
issues. The method by Passow and Alldredge (1995), referred hereafter as TEP0.4µm, has been
widely used in various scientific investigations, but additional time-consuming pretreatment
145
techniques (e.g., bubble adsorption, laminar shear) are needed to measure TEP precursors
(Zhou et al. 1998; Passow 2000). Further modification of this method using smaller pore size
filters (e.g., TEP0.1µm or TEP0.05µm) was later introduced to measure part of the TEP
precursors (Villacorte et al. 2009; 2010). The alternative methods introduced by Arruda-
Fatibello et al. (2004) and Thornton et al. (2007) are capable of measuring both TEP and their
precursors in one single analysis; however, the former is only applicable in freshwater
samples while the latter requires a dialysis step (performed for a couple of days) for saline
samples. Furthermore, the method introduced by Thornton et al. (2007) is only accurate for
samples with high concentration of TEP and their precursors.
The latest method, referred hereafter as TEP10kDa, developed by Villacorte et al. (2015c)
specifically for desalination applications aims to address the major practical issues associated
with the previous methods (e.g., salinity, exclusion of TEP precursors), but also allows
measurement of low concentration of TEP through the introduction of a concentration step
(i.e., filtration through 10 kDa membrane). As such, it allows analyses of samples with a
wide range of TEP + precursors concentrations (down to <0.1 mg Xeq/L) in seawater. In
principle, this method also enables size fractionation of TEPs in seawater by making use of
membranes with different pore sizes during the extraction step.
As discussed above, some of the TEP methods may not be suitable for seawater application
due to potential artefacts formed with high salinity. Although TEP has been identified as a
likely cause or initiator of biofouling in SWRO membranes during algal blooms (See Chapter
2), TEP monitoring is still not widely conducted in SWRO plants. Nevertheless, as TEP is
gaining increasing attention in this regard, there is a clear need for a reliable method to
measure these substances in seawater. The two methods considered to be most feasible for
SWRO applications (i.e., TEP0.4μm, TEP10kDa) are described in detail in Appendix 3.
From the perspective of HAB monitoring in SWRO plants, the TEP10kDa method is
complimentary to the more established and widely accepted TEP0.4µm method. TEP0.4µm
measures TEP while TEP10kDa can measure both TEP and most (if not all) of their precursors.
TEP0.4µm is a more rapid and less laborious method than TEP10kDa, which means it is ideal for
routine or high frequency TEP monitoring in intake seawater. This was demonstrated in the
Jacobahaven demonstration plant in the Netherlands where TEP0.4µm was monitored for three
years and a correlation between TEP and fouling rates in the UF pretreatment system was
observed (see Chapter 11.10 Table 11.10.4).
To assess the removal of TEP and their precursors over the treatment processes, TEP10kDa
measurement is more appropriate because it covers the wider size spectrum of TEP. This was
illustrated when both TEP0.4µm and TEP10kDa testing was conducted along with biopolymer
measurements (LC-OCD analysis) during a summer algal bloom (dominated by green algae)
at a low salinity lake water desalination plant with extensive pretreatment. Measurements
were taken of the raw water at the intake and after various pretreatment processes,
microstrainer, coagulation, sedimentation and rapid sand filtration, granular activated carbon
filtration and ultrafiltration and in the RO permeate. The concentration of TEP10kDa was very
high during the bloom in the raw lake water (see Figure 5.4) demonstrating that the test could
be used to detect a HAB while quantifying the fouling AOM component in the bloom.
Furthermore, the results illustrate the importance of measuring both TEP and their precursors
when evaluating the efficiency of the pretreatment systems in removing algal-derived
foulants from the RO feedwater. It revealed that TEP precursors were the dominant fraction
as compared to TEP. While coagulation-sedimentation-sand filtration proved successful in
almost completely removing the large TEP (TEP0.4μm) component, TEP precursors (TEP10kDa)
and smaller biopolymers remained in the water. A substantial fraction of TEP precursors and
146
biopolymers were also removed by UF. Although substantially reduced, some of these
potential foulants might still reach the SWRO system, even after advanced pretreatment. In
general, both TEP0.4µm and TEP10kDa concentrations demonstrated significant correlations
with biopolymer concentration. In combination with particulate fouling indices (to be
described in Section 5.5) and LC-OCD (Section 5.3.1.1), the application of at least one of
these TEP methods can be crucial in developing strategies to minimize fouling issues in
SWRO plants during HABs.""""""""
Figure 5.4. TEP (TEP0.4µm), TEP + precursor (TEP10kDa) and biopolymer concentrations measured in
samples collected over the treatment processes of an RO plant treating lake water suffering from algal
blooms. Legend: coag = coagulation + flocculation; sed = sedimentation; RSF = rapid sand filtration.
Source: Villacorte et al. 2015c.
5.4! MEASURING BIOFOULING POTENTIAL
Biofouling refers to the growth of a biofilm on a membrane and/or feed spacers due to the
attachment of microorganisms, principally bacteria and subsequent growth with the release of
biopolymers as a result of microbial activity. The biofilm can lead to a decline in normalized
permeate flux or increased differential pressure across membrane elements. As a result of this
fouling, operating pressures need to be increased to maintain production and/or ultimately
membranes will need to be cleaned to restore permeability, all of which will increase
operating costs. If the membranes cannot be cleaned, membrane replacement will be required
which is both costly and time consuming and will result in loss of production.
The TEP component of AOM can in particular potentially initiate and enhance biofouling in
RO systems (Berman and Holenberg 2005). Due its sticky nature, TEP can adhere and
accumulate on the surface of RO membranes and spacers. As TEP accumulate Bar-Zeev et al.
2012 proposed the TEP may serve as a “conditioning layer” - a platform for effective
attachment and initial colonization by bacteria - which may then accelerate biofilm formation
in RO membranes. As with assimilable organic carbon compounds (AOC - the easily
biodegradable portion of TOC), TEP may be partially degradable and may later serve as a
substrate for bacterial growth (Alldredge et al. 1993). The biodegradability of organic matter
in a SWRO feedwater may therefore increase due to AOM produced during a HAB or
through oxidation of AOM or other organic matter in the seawater by chlorination. The
addition of some antiscalants which are designed to be biodegradable to minimize impacts of
the RO brine discharge can also cause an increase in AOC (Boerlage 2001; Weinrich et al.
2015).
Biofouling remains the least understood of the RO fouling mechanisms, in part because
simple, reliable and fast tests are not readily available to measure the biofouling potential of
147
RO feedwater (Schneider et al. 2012). Tests to determine the biofouling potential could be
used to assess the increase in the biofouling potential of seawater from AOM and the
efficiency of pretreatment to reduce it. For instance, the presence of trace AOCs in seawater
can promote biofouling so by measuring AOC treatment plant operators would gain insight
for managing pretreatment in terms of biofouling potential (Weinrich et al. 2016). To support
this effort, an AOC test specifically for seawater was developed and later used for monitoring
pretreatment effects on biofouling potential in full scale RO systems with both subsurface
(beach well) and open intake seawater feed (Schneider et al. 2012; Weinrich et al. 2015).
Only limited data are available for the correlation between growth potential levels and
biofouling rate in full scale plants during algal blooms. This is mainly due to the relatively
recent development of appropriate AOC assays for use in seawater studies. Moreover these
tests require sophisticated equipment and are not currently deployed as an online test which
could offer immediate results. In order to gain an understanding of the fouling potential of a
feedwater without using a bioassay such as the AOC test, the Membrane Fouling Simulator
(MFS) test was developed (Vrouwenvelder et al. 2006). In this test, fouling in a spiral wound
element is simulated and monitored by the development of the head loss across the spacer.
The AOC and MFS tests are discussed further below.
5.4.1! Assimilable organic carbon
5.4.1.1 AOC method development
A test was developed to determine the biofilm regrowth formation potential of drinking water
in water distribution networks according to the growth response of a bacterial inoculum to the
amount of easily assimilable organic carbon (AOC) (van der Kooij 1978; van der Kooij et al.
1982). Initially, the test was based on the growth of the bacterial Pseudomonas fluorescence
strain P17 in a sample with the AOC concentration expressed as µg acetate-C/L, since the
test is calibrated with sodium acetate solutions of different concentrations. The strain P.
fluorescence P17 is capable of utilizing a variety of easily biodegradable compounds. As
some compounds formed by oxidation in the water treatment process such as ozonation
cannot be degraded by P17, the Spirilium spp. strain NOX was added to the inoculum (van
der Kooij et al. 1982) and more recently, a special test making use of Flavobacterium
johnsoniae has been developed to take into account less biodegradable organic compounds
e.g. biopolymers (Sack et al. 2011). Further improvements to the AOC test include increasing
the incubation test temperature from 15°C to 25°C and the inoculum from 500 to 104
CFU/mL (LeChevallier et al. 1993). Both of these adjustments reduced the time for the
culture to reach stationary phase which was generally between two to four days. Other
attempts to simplify the method used adenosine triphosphate (ATP) instead of determining
plate counts (LeChevallier et al. 1993), but problems with commercial ATP measurement
reagents discouraged adoption of this technique (Haddix et al. 2003). Nonetheless, both plate
count and ATP methods are complex requiring weeks of turnaround time before results are
available. In addition, the methods are costly because of the technical labor involved in
assaying ATP levels from filter-concentrated cells or in spread plating samples and
determining plate CFU counts.
A novel approach to simplify the method was achieved by producing bioluminescent strains
of P-17 and NOX test bacteria through mutagenesis with luxCDABE operon fusion and
inducible transposons (Haddix et al. 2004), and that method was further developed to include
the use of a high sensitivity 96-microtiter plate luminometer (Weinrich et al. 2009). Using
luminescence for bacterial growth estimation instead of the traditional plate counts reduces
labor for preparing media and turnaround time to two days, which was an improvement on
traditional methods taking three weeks or more. Moreover, the method provides information
148
regarding the kinetics of substrate utilization as well as the AOC concentration, in acetate
carbon equivalent units which is consistent with previous methods. The bioluminescent fresh
water method has been used for numerous projects that investigate biological filtration
effectiveness and distribution system biological stability in both drinking water and
reclaimed water matrices (Evans et al. 2013, Weinrich et al. 2010). Evaluations conducted
previously indicate that the genetically modified P17 and NOX-strains tolerate salinity up to
5,000 mg/L TDS in the freshwater bioluminescent-AOC test, and are therefore not useful for
seawater monitoring. Researchers have been developing methods for seawater application to
address the need for managing biofouling and mitigating its costly consequences. The
following sections describe tests suitable for seawater desalination plants.
5.4.1.2 Seawater AOC tests
An AOC test for seawater is used to assess the microbial growth potential in SWRO plants.
Since P17 and NOX strains cannot survive in water with salinity greater than that of
reclaimed water or brackish water (<5,000 mg/L TDS), luminescence assays have been
developed using Vibrio bacteria.
The genus Vibrio contains biofilm-forming species that have been detected on a biologically-
fouled SWRO membrane (Zhang et al. 2011). Weinrich et al. (2011) developed a seawater
AOC test using a naturally occurring bioluminescent marine organism, Vibrio harveyi
(ATCC®, 700106™). The V. harveyi seawater AOC is applicable to salinities between 20,000
– 35,000 mg/L TDS but has been applied in seawater with higher salinity from the Gulf and
the Gulf of Oman. Briefly, the seawater AOC test consists of inoculating the sample with V.
harveyi (from an overnight culture prepared at 30˚C). The inoculated samples are then
transferred to a 96-well microtiter plate and a sensitive microtiter plate-luminometer is
programmed to read the plate every two to four hours. The growth profile is monitored until
the stationary phase (Nmax) in which all substrate has been consumed by the test bacteria. The
rate of utilization (µmax) can be determined using Monod kinetics. Typically, the test duration
with V. harveyi is about one day. Maximum luminescence measured at the stationary phase is
converted into an AOC concentration with a 10 – 500 µg-C/L standard curve of acetate
carbon equivalents. A full description is published in Weinrich et al. 2011, and Schneider et
al. 2012. The AOC test for V. harveyi in seawater has been applied to monitor biofouling
potential at bench scale and full scale SWRO plants for pretreatment monitoring, (Section
5.4.1.3) as well as in HAB events (see Chapter 11 Section 11.9).
Another AOC test based on the direct measurement of bioluminescence using Vibrio fisheri
MJ-1 was developed by Jeong et al. (2013) and is reported to estimate the AOC concentration
within one hour. It is similar to the V. harveyi test described above but estimates AOC
concentration with a 10 – 100 µg-C standard curve of glucose for V. fischeri. Recent research
studies have been published using this method for monitoring AOC removal in SWRO
pretreatment using granular activated carbon deep bed filtration (Jeong et al. 2013), and
submerged membrane adsorption bioreactors which were operated with 2.4 – 8.0 g of
powdered activated carbon per cubic meter of seawater (Jeong et al. 2014). The latter was
shown to reduce biofouling for SWRO by adsorption and biodegradation of AOC and low
molecular weight organic matter.
While these bioluminescence AOC tests using specific bacteria are faster, they may not
measure high molecular weight biopolymers generated during a HAB. Vibrio harveyi utilizes
compounds in the 100 - 350 Da range including disaccharides trehalose and cellobiose
(Baumann et al. 1984) and low molecular weight monosaccharides, amino and carboxylic
acids, alcohols and aldehydes (Weinrich et al. 2011). Biopolymers have much higher
molecular weights and have been defined by size exclusion to be greater than 1 kDa (see
149
Figure 5.1) and in the range of 10 kDa defined by the TEP10kDa method or 20 kDa (Huber et
al., 2011 in LC-OCD). Under the conditions defined in the seawater AOC test, the test
organism is unlikely to assimilate high molecular weight biopolymers. Specifically for
SWRO treatment, biopolymers create conditions that enable bacterial attachment and
increase biofouling potential. Therefore, knowing the AOC concentration would provide
guidance on whether nutrients are present for the bacteria to proliferate and lead to biological
fouling.
The luminescent AOC tests also require equipment that may not be typically used in a water
quality lab and would need to be purchased, such as a luminometer for measuring bacterial
growth, a water bath for temperature control, and an autoclave for sterilizing media.
Consumables for the test such as test tubes, microtiter plates, and filters are inexpensive.
Glassware used for AOC tests must be carefully prepared to minimize cross contamination of
trace amounts of organic carbon. Alternatively, glassware is commercially available that
claims to be AOC-free. Bacterial contamination is also a concern for AOC assays; however,
salinity conditions in the seawater test would deter contamination from bacteria common in
the human environment that are not halophilic or halotolerant. Assay samples should be
analyzed as soon as possible after collection because of the highly biodegradable nature of
the organic matter to be measured. If samples are to be shipped to the laboratory, then they
should be cooled and shipped on ice overnight to inhibit biological activity in the sample
during transit.
5.4.1.3 Application of seawater AOC test
Recently, the biofouling potential in numerous full-scale SWRO plants worldwide and
extensively at the Tampa Bay Seawater Desalination Plant (TBSDP) was examined using the
V. harveyi seawater AOC method. The AOC test was also used to assess feedwater quality to
the plant, the impacts of pretreatment, and the biofouling potential in the SWRO feed.
The first study was at TBSDP when a non-toxic HAB contributed to periodic operational
challenges. AOC at TBSDP has been investigated during normal operation and during a
period of algal growth. The plant experienced foaming in the pretreatment basins, shortened
diatomaceous earth (precoat) run times and reduced production capacity. The algal species
Ceratium furca and Phaeocystis sp. were found in filter backwash media and were thought to
contribute to the operational issues. At the same time, TOC was 7.6 mg/L at the plant influent
and average AOC (from three consecutive days) was 360 ± 180 µg acetate C per L. AOC was
also measured after chlorine dioxide treatment and increased by 65% compared to the raw
water. TOC levels in TBSDP raw water were generally between 4-6 mg/L at the intake and
are variable. During normal operation, AOC has been determined to be less than 30 µg/L
(Weinrich et al. 2015) and up to 225 µg/L (Schneider et al. 2012). While the AOC test may
not measure high molecular weight biopolymers as discussed earlier, the AOC increase may
have occurred through shear of algal cells and release of low molecular weight AOM or from
bacterial oxidation of organic matter into easily biodegradable low molecular weight
compounds. In this case, the AOC test may measure the additional nutrients during a HAB
which can be utilized by bacteria in a biofilm on the RO membrane. The results are further
discussed in Chapter 11 Section 11.9.2. The plant has demonstrated AOC removal by the
diatomaceous earth filters in the same studies.
In another study, plant personnel at the Al Zawrah plant in UAE described evidence of an
algal bloom in May 2012 when the raw water pH decreased from the pH 8.1, normally
observed, to pH 7.5. This was accompanied by elevated organic matter measured in the raw
water as AOC (220 µg/L) and TOC (2.9 mg/L). During two other sampling events with
reportedly normal conditions, the concentration of organics in the raw seawater was below
150
detection for AOC (<10 µg/L) and 60% less for TOC at 1.2 ± 0.04 mg/L in two sampling
events from July and November 2012 (Weinrich et al. 2015).
These results demonstrate that AOC can increase at the seawater intake during algal bloom
periods compared to periods when water quality and algae levels are normal. The seawater
intake type also has important impacts on AOC levels. Organic matter measured as AOC and
TOC was found to be lower in plants that have subsurface intakes (e.g. beach wells)
compared to plants having surface intakes. Figure 5.5 shows AOC and TOC levels measured
between 2009-2010 at various desalination plants (adapted from Schneider et al. 2012). AOC
varied from 75 – 221 µg/L in surface intakes compared to <10 µg/L to 22 µg/L for beach
well intakes measured during that study (Schneider et al. 2012). Hypotheses predicting that
AOC was linked to biofouling of RO membranes in previous studies (including Franks et al.
2006, Fujiwara and Matsuyama 2008) were substantiated in recent evaluations of AOC
concentrations near 50 µg/L in the RO feed that was linked with biofouling, increases in RO
differential pressure and decreases in specific flux (Weinrich et al. 2015). Furthermore,
pretreatment chemicals such as antiscalants and some oxidants increase AOC within the
pretreatment process as mentioned previously, thereby, increasing the biofouling potential of
RO feedwater (Weinrich et al. 2015, Weinrich et al. 2011, Schneider et al. 2012,
Vrouwenvelder et al. 2000). An example is shown in Figure 5.5 for the Fujairah 1 SWRO
desalination plant in the UAE which dosed antiscalant to the SWRO membrane feed.
Antiscalants may contribute to AOC directly, or when dosed in pretreated water carrying a
chlorine residual. Sequential addition of antiscalant followed by cartridge filtration and then
sodium bisulfite (to reduce the oxidation reduction potential) presents an opportunity for
chlorine to react with the antiscalant. In this scenario, byproducts such as AOC may be
formed, or may be present as impurities in the antiscalant itself (Weinrich et al. 2015).
By measuring AOC
during an algal bloom,
the bio-fouling potential
could be observed and
quantified for operating
records; both at the
intake and at pre-
treatment locations (e.g.
after oxidation, before
and after filtration).
Assessing the impact of
chemical, pressurization
Figure 5.5. Organic carbon in raw and SWRO feedwater from SWRO plants or mechanical forces on
in Spain (SP), France (FR), the United Arab Emirates (UAE) and United AOC and the availability
States of America (USA) measured as TOC and AOC (adapted from
Schneider et al. 2012). DMF refers to dual media filtration and DE
of AOC would quantify
diatomaceous earth. biofouling potential.
There is a risk that shear
forces from feed pumps and valves could lyse algal cells and release soluble, intracellular
organic matter (Ladner et al. 2010) but the risk is reduced if low shear valves are selected.
Lysing algal cells (either through oxidation, pressurization or mechanical shear) is likely to
release low molecular weight algal toxins (depending on the species) as well as organic
matter that is highly biodegradable. For the latter issue, this release would create conditions
sufficient for bacterial proliferation on RO membranes thereby increasing biofouling
potential.
151
HABs would be best managed by SWRO plants that can effectively minimize nutrients
during routine (non-HAB) operations. Nutrient limitation begins with identifying nutrient
sources by testing antiscalants and other dechlorinating chemicals for impurities such as
AOC or phosphate. While costly capital improvements could be an option (e.g. changing the
intake source, installation of granular media filters and coagulation, conversion to biological
filtration), modifying the chlorination strategy may be an effective solution for reducing
biodegradable byproducts and bacterial growth. Plant operators need more specific water
quality information for pretreatment optimization and biofouling reduction outside of typical
bulk organic measurements or SDI. Monitoring AOC regularly and balancing pretreatment
may be one solution; another would be sidestream piloting using the Membrane Fouling
Simulator Test (see section 5.4.2). Granular media-based biological filtration is generally
effective in drinking water applications and could be an option for SWRO as discussed in
Chapter 9.6. Limiting AOC and removing it during pretreatment through biodegradation
mechanisms on filter media and by coagulation would lower the biofouling potential of the
RO feedwater carried to RO membranes. Removing this source of biological fouling material,
measured as AOC, would reduce the potential for microbial growth or at least delay it on the
RO membranes and associated operational impacts of biological fouling. Future work needed
at SWRO plants would be to identify the increase of AOC caused by blooms of specific algal
species in order to gain an understanding of their contribution to biofouling potential.
5.4.2! Membrane fouling simulator
Biofouling in spiral wound RO elements usually manifests in increasing head loss across the
spacer. In full scale plants, the head loss across the first stage is commonly monitored with
pressure sensors, since biofouling tends to occur primarily in the first 10 to 20 cm of the first
element. It is possible to monitor the head loss across the first element in order to modify the
pretreatment process and as a criterion for chemical cleaning.
In pilot tests, monitoring the increase of differential pressure along the feed channel over
time as an index is very useful in testing different pretreatment schemes and conditions. It is,
however, very costly due to the length of operations and/or the need for several pilot plants.
For this reason the MFS was developed (Vrouwenvelder et al. 2006). In a MFS, biofouling
along the feed channel is simulated. This device is constructed of two stainless steel plates
containing sample coupons of membrane, feed and product spacer. It is equipped with
connectors for feed and brine flow and sometimes for product flow as well. Head loss
development is accurately monitored using pressure sensors.
Villacorte (2014) demonstrated, by making use of the MFS, that AOM produced by
laboratory cultivated marine algae (Chaetoceros affinis in synthetic seawater) accelerated
biofilm growth and resulted in a rapid increase in the feed channel pressure drop. The tests
were performed by running two MFS in parallel, one initially with a RO membrane slightly
pre-fouled with AOM and another one with a clean membrane (control). Both MFS were fed
with 100,000 cells/L of marine bacteria Vibrio harveyi for about 24 hours and then fed with
synthetic seawater spiked with dissolved nutrients (0.1 mg acetate-C/L, 0.02 mg N/L and
0.01 P/L) for 10 to 21 days. An exponential increase of pressure drop was observed for MFS
pre-fouled with AOM with up to 1000% increase in only a span of 6 days. In comparison, the
control membrane only showed 250% increase in a span of 17 days.
While the MFS allows the biofouling potential of feedwater to be measured quite accurately
it remains a lengthy test. Therefore, the development of quick AOC tests that incorporate
biopolymers into the AOC measurements would be of great value to plant operators to
optimize pretreatment during a HAB.
152
5.5! FOULING INDICES TO MEASURE PARTICULATE FOULING POTENTIAL
The deposition of algal particulates (e.g. TEP and TEP precursors) on RO membranes during
a bloom, along with inorganic and organic colloids, bacteria and other materials, can lead to a
decline in normalized permeate flux and an increase in head loss across spacers (or
membrane bundles). This type of fouling, often referred to as particulate fouling, may
exacerbate other types of fouling (e.g. biofouling). Reliable methods to predict the particulate
fouling potential of feedwater are important in preventing and diagnosing fouling at the
design stage and for monitoring pre-treatment performance during full-scale plant operation.
Turbidity and the Silt Density Index (SDI) are universally applied for this purpose as they
often form the basis of RO membrane guarantees. Turbidity however, like particle counting,
can only indicate the concentration or mass of particles in feedwater, but provides no
information on the resistance of these particles when they deposit on a membrane. Similarly,
measurements of total suspended solids (TSS), while important for solids loading in design
and operation, will not provide any information on particulate fouling.
Fouling indices, of which the Silt Density Index is the most commonly used in practice, are
designed as quick filtration tests to simulate RO membrane fouling and thereby, determine
the particulate fouling propensity of a feedwater. The lesser known Modified Fouling Indices
(MFI) are increasingly applied, in particular in research projects, pilot and lab/bench scale
studies. While both indices are not specific for algal-related particulate material, an increase
in the values above background may occur due to HABs at the intake. Table 5.3 provides a
summary of the SDI and MFI indices, comparing the principle of each method, interferences,
particulate fraction identified and operator skill required for the test. The fundamental
background of the SDI and MFI fouling indices is provided in subsequent sections.
Applications of the indices in measuring algal-laden feedwater are given to illustrate
advantages and limitations of these indices. The performance of parameters such as turbidity
and chlorophyll-a to compliment and interpret SDI measured at the seawater intake are also
discussed.
5.5.1! Silt Density Index
The SDI, developed by the Du Pont Company at the request of the Bureau of Reclamation
(Du Pont 1972), has been universally applied in the desalination industry for the last 40 years
to assess the particulate fouling tendency of feedwater. ASTM International standardized the
test protocol in 1995 as ASTM D4189, reapproving it most recently in 2014 (ASTM D4189-
07(2014)).
The SDI test consists of passing a feedwater through a 0.45µm microfiltration membrane in
dead-end flow at a constant pressure (207 kPa) and determining the membrane filter-plugging
rate. The SDIT 4is calculated from Equation 1.
& ti #
$1 - t ! 100
%PF % T"
SDI T (%/min) = = (1)
T T
where ti is the time to collect an initial sample (normally 500 ml) filtered through the
membrane, tT is the time taken to collect a second sample after a total elapsed filtration time
(T) of 5, 10 or 15 minutes, and %PF is the percentage plugging factor. The ASTM specifies
the %PF should not exceed 75% when conducting the test. If so, the SDI should be
4
The SDI means the percentage flux decline per minute. Dimensions of the SDI test are %/min; by convention
the SDI is commonly reported as dimensionless.
153
Table 5.3. Comparison of Silt Density Index and Modified Fouling Indices to measure particulate fouling in seawater and feedwater.
Parameter Standard / Basis of Method Organic matter (OM) Interferences / Issues Analysis time /
Reference identified Operator skill
Silt Density Index ASTM D4189 Employs 47mm diameter flat 0.45µm >0.45 µm Inaccurate at high particle 5 – 15 minutes. Simple
(%/min) microfiltration membrane. Measured in Not specific to HAB concentration such as algae- routine inexpensive
constant pressure mode. particulate material. rich seawater. Not based on a test.
Measures filter plugging after 5, 10 or 15 Measurement will filtration mechanism.
minute interval. include algal cells and Not linear with particle
some algal debris. concentration.
No standard correction
method for temperature,
pressure or membrane
related properties.
Not applicable for UF
permeate.
Modified Fouling ASTM D8002 Employs 47 mm diameter flat 0.45µm >0.45 µm Not applicable for UF 30 – 60 minutes.
Index-0.45 (s/L2) microfiltration membrane. Measured in Not specific to HAB permeate. More complex to
constant pressure mode. MFI determined from particulate material. calculate MFI-0.45.
cake filtration region in t/v vs V graph. MFI- Measurement will
0.45 corrected to standard reference conditions include algal cells and
of temperature, pressure and membrane area. some algal debris.
Modified Fouling Not an ASTM Employs 25mm diameter flat sheet Not specific to HAB Test duration dependent
Index-UF (s/L2) standard ultrafiltration membranes of 10 – 100 kDa particulate material. on test flux in constant
MWCO1. Measurement will flux mode. 30 minutes
Measured in constant flux mode at 10 – include algal cells and at 250 L/m2h and
300 L/m2h (Salinas 2011) or constant pressure some algal debris. several hours at
mode. Depending on MWCO 15 L/m2h (Salinas-
MFI-UF corrected for temperature, pressure of test membrane, may Rodriguez 2011).
and membrane area in both modes. For the include TEP and TEP Longer in constant
MFI-UF in constant flux mode – the precursors. pressure mode.
recommended test flux is the same as target More complex to
MF/UF plant. For RO plants a flux correction calculate MFI-UF.
method is under development. Alternatively,
the same flux as for TEP measurements
(60 L/m2h) could be applied.
1
molecular weight cut off (MWCO)
determined after 10 or 5 minutes. If the PF still exceeds 75% after only 5 minutes of filtration,
the ASTM recommends another method be employed to analyze for particulate matter.
SDI is one of the key parameters to assess the particulate fouling potential of raw water and
monitor the efficiency of RO pretreatment processes over time in design and operation of
SWRO desalination plants. In some cases SDI may be provided in the raw seawater quality
design envelope. For plant operation, SDI15 limits are often specified for RO feedwater in RO
membrane guarantees (e.g. the SDI15 will not exceed 5, and be below 3 (or 4) for 90% of the
time) and other plant performance contracts. These limits may be linked to turbidity
monitoring in contracts as turbidity can be continuously monitored online. Automatic online
SDI analyzers (not continuous) are available which can be input into plant control systems so
SDI can be routinely monitored in the control room with SDI alerts and alarms allowing
operators to respond to changes in influent water quality, including HAB events.
Despite the well-documented limitations of the SDI (Schippers and Verdouw 1980; Kremen
and Tanner 1998; Boerlage et al. 2000; Boerlage 2008), it remains a mainstay in the
desalination industry due to its simplicity. Key limitations summarized in Table 5.3, include
that the SDI is not based on a filtration mechanism and therefore cannot be used as the basis
of a model to predict pressure increase in RO systems. This was again recently demonstrated
in practice by Jin et al. (2017) in a one year study measuring the SDI of RO feedwater which
included algal bloom events at a full scale SWRO plant employing DAF and UF pretreatment.
No correlation was found between the SDI and differential pressure increments in the RO
systems.
Nonetheless, increasing SDI values may be the first sign of a HAB at an intake in the absence
of changes in other parameters such as turbidity and chlorophyll-a. The increase in SDI is
due in part to the retention of marine algal cells by the smaller pores (0.45µm) of the test
membrane through size exclusion. For instance, Cochlodinium polykrikoides, the dominant
species present in the notorious 2008 bloom in the Gulf, is 20 - 40 µm in size for individual
cells, but forms chains that are much longer. Smaller algal related matter such as algal
detritus from ruptured cells comprising cell walls, flagella, organelles, dissolved and
particulate intra- or extracellular AOM may also be captured to some extent through a variety
of mechanisms resulting in higher SDI values. Partial blocking of membrane pores can
reduce the effective pore size of the test membrane. Smaller particles may also be trapped in
the cake formed on the test membrane where the cake has smaller interstices than the SDI
membrane pores. However, smaller pore size membranes are required to measure the more
fouling colloidal particles such as TEP and TEP precursors.
Turbidity may not register an increase during a bloom due to the deficiencies of turbidity
measurements in detecting HAB cells and small colloidal particles as discussed previously in
Section 5.2.5. In particular, particle sizes smaller than 0.2 µm may not be measured due to
limited light scattering (Kremen and Tanner 1998). Moreover, very small AOM-related
particles such as TEP are transparent and are therefore “invisible” to turbidity meters.
Results trialing chlorophyll-a measurement using fluorescence as a proxy for algal biomass
to compliment SDI and online water quality testing at the seawater intake have been mixed.
As with turbidity, spikes in SDI may not coincide with an increase in chlorophyll-a. There
are reports of no significant increase in chlorophyll-a above background being observed,
despite elevated algal cell concentrations up to 5 million cells/L accompanying the observed
increase in SDI. This is consistent with the discussion in Chapter 3 which describes
limitations to measuring chlorophyll-a to estimate biomass using fluorescence. The
relationship between chlorophyll-a fluorescence and cell biomass is not constant across all
phytoplankton species, nutritional conditions, and times of sampling. The large Noctiluca
155
scintillans dinoflagellate is a notable example; it ranges up to 2 mm in size and while it

releases ammonia into seawater and therefore can be harmful, it lacks chloroplasts and
therefore its chlorophyll content is low (Pool et al. 2015). In this instance, SDI may increase
along with ammonia concentrations (if measured) in the intake seawater, while chlorophyll
would show no change. Other factors influencing chlorophyll-a include nutrients, and light
history, with limitations often resulting in lower chlorophyll-a content than the same cells
under more favorable conditions (see Chapter 3). If chlorophyll-a is measured then it is
recommended that only the night time data be used, as nocturnal chlorophyll-a fluorescence
is more consistent.
Other options that could be used in conjunction with SDI are regular microscopic
examination and counting of cells, or some of the newer automated phytoplankton-
identifying instruments, such as the Imaging FlowCytobot or FlowCam (see Chapter 3). They
can also be automated, and provide more information about species and cell numbers.
Routine, online use of these automated biosensors would allow operators to generate a long-
term, high-resolution database of algal species and concentrations that are associated with
plant disruptions, and an associated record of which pretreatment strategies worked or failed.
Visual examination of the SDI test membrane may provide additional information on
feedwater constituents and indicate changes in quality such as an unusual color of the deposit
on the SDI membrane e.g. a red filter deposit from red blooming algae. Closer examination
of membranes by electron microscopy may be used in combination with algal counting to
identify species or bloom types. This is illustrated in electron micrographs of SDI test
membranes of raw water off the coast of Chile where diatoms and coccoliths are visible in
Figure 5.6 (left) and an intact coccolithopore (Emiliania huxleyi) cell in Figure 5.6 (right)
(Petry et al. 2007).
Figure 5.6. Electron microscopy of SDI test membranes showing the presence of algae and algal
particulate debris (left) and close up (right) of a coccolithopore cell. Photo: Petry 2007.
Severe HAB events can significantly increase the fouling potential of seawater at open
intakes due to the increase in algal biomass to the point that the SDI may not be measurable
due to rapid plugging of the SDI test membrane. During the prolonged 2008 HAB event in
the Gulf, algal cell counts of 11 to 21 million cells/L were recorded in surface waters near
Fujairah (Richlen et al. 2010). As a result, TSS increased to 30 mg/L on occasion, compared
to the median TSS of 5 mg/L; the SDI test was not informative, as it is limited to low fouling
feedwater. Despite the ASTM recommendation that the SDI is only employed for low
turbidity water (< 1 NTU) and for water that will not result in a %PF of > 75%, these
guidelines are widely ignored in practice. Furthermore, raw seawater SDI5 values often
156
exceed 75% PF and the industry then often measures the SDI3 (not a specified ASTM SDI
test interval).
Not surprisingly, even the SDI3 was reported to spike above 25 (75% PF limit) on several
occasions during the 2008 bloom in the seawater around Fujairah (see Chapter 11 Section
11.2). While these high SDI3 can be useful in indicating the possible presence of a bloom in
the incoming feedwater when trended against typical feedwater data, operators should be
aware these SDI values will underestimate the fouling potential of the raw water as the SDI is
not linear with particle concentration. Instead, the SDI is limited mathematically to a
maximum value of 6.7, 10 and 20 for a filtration time of 15, 10 or 5 minutes, respectively.
Using the ASTM criterion of 75% maximal plugging, the values are 5, 7.5 and 15
respectively. The asymptotic behavior of SDI with increasing particle concentration as it
approaches these limits was demonstrated experimentally by Schippers and Verdouw (1980)
by measuring SDIT for a series of diluted formazine (a model colloid) solutions. Figure 5.7
shows SDI15 as a function of formazine with the accompanying %PF, and ASTM-
recommended 75%PF
limit and SDI15 test limit.
As the SDI approaches
its limit, it is obviously
easier to obtain
repeatable SDI results
but the SDI will
underestimate fouling
and become increasingly
inaccurate at higher %PF
(Boerlage 2008). Hence,
the typical SDI15 limit
set by membrane
manufacturers is < 3 to 4
depending on the
feedwater source, which
is equivalent to <5 to
60% PF and is in the
more linear range of
Figure 5.7.
Figure 5.7. SDI and %PF of diluted formazine solution demonstrating the
15
non linearity of SDI with colloidal concentration. SDI15 and 75% PF limits
In summary, increases in
indicated (SDI data from Schippers and Verdouw 1980). SDI5 (or SDI3) can
indicate the presence of HABs at the intake, while increases in SDI15 downstream can
indicate the failure of pretreatment steps and that operator action is required; however, high
SDI3 and SDI15 values such as those observed during the 2008 Gulf bloom would have
increasingly underestimated the fouling potential of these desalination process streams.
Moreover, when assessing process performance it should be remembered that the SDI cannot
be directly compared for different filtration intervals e.g. SDI5 for raw water and SDI15 after
pretreatment or when measured at different temperatures (the SDI test applies no temperature
correction for differing feedwater temperature). Other factors which influence the SDI such
as test membrane porosity etc. are discussed in Boerlage (2008).
157
5.5.2! Modified Fouling Indices

5.5.2.1! Modified Fouling Index-0.45
The MFI-0.45 was developed by Schippers to overcome the limitations of the SDI (Schippers
and Verdouw 1980) and thus has the potential to be of value in detecting a HAB at an intake
and optimizing the efficiency of pretreatment processes. The MFI-0.45 has recently been
approved by ASTM International as a standard (D8002-15) to indicate the fouling potential
of feedwater due to particular matter. Automatic analyzers are commercially available that
can simultaneously measure MFI-0.45 and SDI online, but they are not as widely used as the
SDI analyzers.
Unlike the SDI, the MFI-0.45 is based on a filtration mechanism (cake filtration) and is linear
with particle concentration. As with the SDI, the MFI-0.45 is determined in dead-end flow
under constant pressure using similar equipment as the SDI. The MFI-0.45 is determined in a
plot of t/V vs V (where V is filtrate volume and t filtration time) from the gradient of the
linear region of minimum slope where cake filtration occurs (refer to Schippers and Verdouw
1980 for more information). Schippers defined the MFI at reference pressure and temperature
values of 2 bar (ΔPo) and 20°C (η20°), respectively and for the area of the 0.45 µm membrane
(Ao=13.8×10-4 m2). Incorporating the Carman-Kozeny relationship for idealized spherical
particles to calculate the specific resistance of the cake deposited on the MFI-0.45 membrane
yields the following equation for the MFI-0.45 at standard reference conditions:
$ 20 C 90(1 - " ) C b
!
MFI = 2
(2)
# p d p2 " 3 !Po A 0
where (Cb) is the concentration of particles in the feedwater, (ρp) is the density of particles
forming the cake, (ε) cake porosity and (dp) particle diameter. From this equation the
pronounced effect of decreasing particle size in increasing the MFI can be seen.
The MFI can be used as an index to characterize the fouling potential of a feedwater
containing particles, as it is a function of the dimension and nature of the particles forming a
cake on the membrane, and is directly correlated to particle concentration in a feedwater. For
feedwater containing algae, the MFI-0.45 could provide information about the difference in
net fouling potential (cake permeability) due to differing algal cell size (dp) and cell
abundance (Cb) of bloom species that can vary significantly. For instance, two algal bloom-
forming species found in the Gulf and Gulf of Oman are Cochlodinium polykrikoides (with a
size range of 20 – 40 µm and a maximum abundance up to 20 million cells/L in the 2008
HAB) compared to the much larger Noctiluca scintillans dinoflagellate which formed less
dense blooms with only up to 68,500 cells/L measured during HAB events in the Gulf (Al
Shehhi et al. 2014). The MFI-0.45 is expected to be higher for the smaller Cochlodinium
polykrikoides as the MFI is inversely proportional to particle diameter squared (see equation
2).
A clear advantage of the MFI-0.45 is that it could be used to measure the fouling potential of
low and high fouling feedwater and therefore assess the efficiency of pretreatment processes
during a HAB. Data available from the Jacobahaven SWRO demonstration plant (see Case
Study 11.10 for more information), where both the MFI-0.45 and SDI15 were measured
during an algal bloom pre- and post-UF are shown in Figure 5.8 (from Al Hadidi 2011). Both
158
the SDI15 and MFI-0.45 values of the UF permeate are below one following ferric
coagulation and filtration through the 150 kDa UF membranes. However, the SDI15 of the UF
feed was highly fouling with the 75% PF exceeded on both days. In fact, the plugging of the
SDI test membrane was reported to be so rapid that even the 75% PF was exceeded when
measuring SDI3.! Therefore, the fouling potential of the UF feed is underestimated and the
performance of the UF step cannot be accurately determined by the SDI test. In comparison
the MFI-0.45 is not limited to low turbidity or low fouling feedwater. While the ASTM MFI-
0.45 standard does not recommend the test is used for UF permeate it can measure the
efficiency of the UF step in removing fouling particles captured by the 0.45 µm test
membrane during the bloom. In this study Al Hadidi estimated a particle removal of 99.947±
0.053% based on the average MFI-0.45.
Assuming that cake filtration is the dominant mechanism in RO membrane fouling, a MFI
model was developed to predict flux decline or pressure increase to maintain constant
capacity in RO systems. However, the predicted rate of fouling based on MFI-0.45
measurements was much lower than observed in practice. For instance, given a MFI of 1 s/L2,
a fouling rate of 15% was predicted to occur within several hundreds of years. It was
Figure 5.8. SDI15 (A) and MFI-0.45 (B) measurements at the Jacobahaven SWRO Plant in May 2010 of the
UF feed (seawater filtered through 50 µm screens with ferric coagulant added) and post filtration through
150 kDa UF membranes (data from Al Hadidi 2011).
concluded that particles smaller than those captured by the MFI-0.45 (and SDI) test
membranes were responsible for the fouling observed in RO due to their much higher cake
resistance (Schippers and Verdouw 1980).
5.5.2.2! Modified Fouling Index-UF
The MFI-UF was initially developed to include smaller particles using a reference
ultrafiltration membrane of 13 000 Da (13 kDa) molecular weight cut off in both constant
flux and constant pressure modes (Boerlage et al. 2000, 2002, 2004). Much higher MFI-UF
values were measured as predicted. By employing UF membranes the MFI test could be
extended to measure UF permeate. However, more accurate fouling prediction is expected
with the MFI-UF measured in constant flux mode. In this case, the MFI is determined from
the linear region, where cake filtration occurs, in a plot of applied transmembrane pressure
over time (see Boerlage (2004) for derivation of the MFI in constant flux mode). Salinas–
Rodriguez et al. (2011; 2015) further developed the MFI-UF test in constant flux mode to use
smaller disposable UF test membranes (25 mm) with variable MWCO (10-100 kDa) where
filtration flux can range between 10 L/m2h to 300 L/m2h. Research efforts trialling the MFI-
UF on seawater during algal blooms or in laboratory studies with AOM for various purposes
are described below.
159
The use of lower MWCO membranes means that smaller fouling TEP and TEP precursors
present in an algal bloom could be included
in the MFI measurement as well as larger
algal cells and detritus. Villacorte (2014)
calculated the theoretical MFI-UF for
several bloom-forming algal species, based
on their cell size and concentration, and
found that it was significantly lower than
that measured in natural seawater by the
MFI-UF using 150 kDa test membranes.
The substantial difference was attributed
mainly to the presence of TEP precursors
and TEP in seawater retained by the test
membranes and which are not considered
when calculating the theoretical MFI-UF.
Furthermore, the TEP precursors (measured
by TEP10kDa) were found to be the dominant
fraction of TEP during a bloom (see Figure
5.4). As discussed by Villacorte (2014) TEP
are gel-like and capable of squeezing
through the interstitial voids between algal
cells accumulated on the MFI-UF
Figure 5.9. Scanning electron micrograph of the MFI-
UF PAN 13 kDa membrane showing pore size membrane, resulting in a more-fouling cake
comparison to MFI-0.45 membrane pore. Tight UF layer due to the higher resistance. Therefore,
membranes in the MFI test will capture some of the smaller MWCO MFI-UF test membranes
fouling TEP precursors (< 400 nm) in size in addition on the order of 10 kDa would capture these
to larger TEP. Modified from Boerlage 2008.
TEP precursors and their associated fouling
potential. This is depicted in Figure 5.9
which shows an electron micrograph of the pores of the earlier 13 kDa MFI-UF reference
membrane, which were around 1000 times smaller than the pores of the MFI-0.45 membrane.
Some of the TEP precursors (ranging in size from a few nm up to 0.4 µm) as well as TEP
(0.4 µm up to 1000 µm) which cause fouling on both UF and RO (refer Chapter 2) should
now be captured in the MFI-UF test for similar small MWCO test membranes.
Villacorte (2014) therefore trialed the MFI-UF (in constant flux) using the smaller 10 kDa
MWCO membrane to measure the fouling potential of the raw water during algal blooms at
five different RO plants desalinating water of various salinities, including lake, river, and
seawater and after pretreatment processes (Figure 5.10). The relationship between the MF-
UF and the concentration of fouling AOM constituents (TEP0.4µm, TEP10kDa and biopolymers)
determined in the treatment process stream was examined. Results showed a higher
correlation between MFI-UF and the TEP10kDa component (R2>0.65) of AOM than between
the MFI-UF and concentration of biopolymers or the larger AOM components measured by
TEP0.4µm. This demonstrated that the MFI-UF could be used to measure the fouling
constituents of AOM during HABs and that the TEP precursors can most likely influence the
fouling propensity of the water more than other types of organic matter.
160
Figure 5.10. The membrane-fouling potential (MFI-UF10kDa) of 26 water samples collected during the
treatment processes of five plants (lake, river, reservoir and seawater sources) plotted against (a) TEP0.4µm,
(b) TEP10kDa and (c) biopolymer concentrations, respectively. Modified from Villacorte 2014.
In another study, pretreatment efficiency was assessed using a range of MWCO membranes
in the MFI-UF test (constant flux) at the Jacobahaven SWRO demonstration plant (see
Chapter 11.10) which routinely experiences seasonal algal blooms in the European spring
(Salinas-Rodriguez 2015). Initially, in the spring and summer of 2009, MFI-UF
measurements were conducted only using the larger MWCO test membrane of 100 kDa.
Subsequently, in spring 2010 additional MFI-UF measurements were conducted across the
plant using smaller MWCO test membranes of 50 and 10 kDa. The MFI-UF fouling potential
measured using the larger 100 kDa membranes is summarized in Table 5.4. Chlorophyll-a
measurements at the plant intake close to the time of the MFI-UF tests are included in Table
5.4 (from Figure 11.10.6 in Chapter 11.10).
Table 5.4. Membrane fouling potential based on MFI-UF measurements with 100 kDa test
membranes (data from Salinas-Rodriguez (2015) and chlorophyll-a (data from Fig 11.10.6 in
Chapter 11.10 courtesy of R. Schurer) for the raw source water at the Jacobahaven SWRO
Demonstration Plant.
MFI-UF Chlorophyll-a
(s/L2) (µg/L)
23 April 2009 4310 3.4
28 April 2009 4840 6.3
16 June 2009 3800 1.0
2 July 2009 2950 1.0
6 July 2009 2840 1.0
10 May 2010 25,340 3.9
For the Jacobahaven plant, spikes in chlorophyll-a generally indicated algal blooms at the
intake. However, the fouling potential of the seawater in spring 2010 was very high based on
the MFI-UF, more than five times that measured during a previous bloom in April 2009. This
was not mirrored by a similarly high chlorophyll-a measurement in the raw feedwater. Algal
speciation varied during blooms (see Chapter 11, section 11.10) and as discussed above, the
chlorophyll-a concentration varies with a myriad of factors including the bloom species and
cell size. Hence, while MFI-UF measurements are not specific to algal particulate matter,
they can provide operators with more reliable information as to the potential fouling impact
of an algal bloom at the intake.
When Salinas-Rodriguez (2015) used the smaller MWCO test membrane of 10 kDa, thereby
potentially capturing the smaller algal-derived biopolymers (TEP precursors) as suggested by
Villacorte (2014), the fouling potential of the raw water in May 2010 more than doubled
relative to the MFI-UF with 100 kDa membrane (Figure 5.11). MFI-UF results for the three
161
different MWCO membranes across the plant allow the removal efficiency of pretreatment
processes for particles of
various sizes, captured by the
different test membranes, to be
assessed. For example, the
fouling potential of the
seawater during the May 2010
algal bloom was reduced
following coagulation and
ultrafiltration (nominal
MWCO of 150 kDa) by 94%,
93%, and 88% for 100, 50, and
10 kDa MFI-UF test
membranes, respectively. As
Figure 5.11. Effect of pretreatment on MFI-UF at the Jacobahaven biopolymers can vary upward
demonstration plant using test membranes of 100, 50 and 10 kDa size
of 1 – 20 kDa some will pass
at 250 L/m2h (data from Salinas-Rodriguez 2015).
through the UF membrane
with a much larger MWCO. This is reflected in the lower reduction in fouling potential for
the smaller 10 kDa membrane. This means that some of the more fouling AOM such as TEP
precursors may reach the RO membranes.
The MFI-UF was also trialed to determine the optimal coagulant dose to reduce the fouling
potential of seawater containing AOM. Tabatabai (2014b) conducted MFI-UF experiments in
constant flux using a larger 150 kDa MWCO test membrane in laboratory experiments for
seawater solutions containing algal organic matter (0.5 mg/L as biopolymers). MFI-UF
measurements showed the addition of 5 mg/L Fe reduced the fouling potential seven fold in
the seawater with no measureable reduction when the coagulant dose was doubled (see
Chapter 9 for further information). It would be worthwhile to repeat such an experiment
using a 10 kDa membrane in the MFI-UF test to assess the optimal coagulant dose for the
more fouling TEP precursors.
As there is currently no ASTM standard for the MFI-UF test, it has been applied in the field
in both constant pressure and constant flux mode and for membranes of varying MWCO. Jin
et al. (2017) evaluated multiple MFI-0.45/MFI-UF tests for RO feedwater in a one year study
which included algal bloom events at a full scale SWRO plant employing DAF and UF
pretreatment. The MFI tests were conducted in constant pressure mode with decreasing
MWCO membranes in series to interpret fouling potential through size fractionation; MFI-
0.45, MFI-UF 100 kDa and MFI-UF 10 kDa and results correlated to RO differential pressure.
During algal blooms, all MFI values significantly increased and generally reflected the
variation in differential pressure with the MFI 100 kDa more closely related to differential
pressured variation (Jin et al. 2017).
In conclusion, the above MFI-UF results to monitor the particulate fouling potential of
feedwater and assess pretreatment for removal of smaller fouling particles (including TEP
precursors) have so far proved promising. Moreover, the MFI-UF can be used in combination
with other MFI-UF tests of varying MWCO either in series or in parallel to compare the
efficiency of pretreatment processes or pretreatment trains for the removal of selected particle
sizes.
5.6! SUMMARY
HABs can result in a substantial increase in the organic and solids load in the raw source
water to be treated at a desalination plant and may lead to overloading of pretreatment
162
systems and membrane fouling. SWRO designers and operators therefore require methods to
determine the concentration of AOM, its fouling constituents, and any increases in the
particulate and biofouling fouling potential or other HAB-associated water quality changes.
Such methods allow HABs to be monitored and detected in the raw water so that treatment
processes can be optimized during a bloom event to maintain plant production and water
quality targets.
Temperature, conductivity, pH and turbidity are often continuously monitored at plant
intakes, in addition to analysis of SDI, TOC and TSS to characterize feedwater quality. None
of the conventional online parameters can definitively detect HABs as they are not specific to
algal blooms. Changes can be caused by other factors such as pollution events and/or marine
hydrodynamics. The interpretation of a water quality variable can thus be complex.
Nonetheless, these measurements may indicate conditions that promote a bloom, such as
temperature and salinity or indirect impacts from HABs such as low DO following
decomposition of a dense bloom. In conjunction with other conventional water quality tests
such as SDI, the standard online water quality parameters can be useful in indicating a
deterioration in feedwater quality due to HAB events, and can provide timely and valuable
information that action is required.
Measuring TOC to detect AOM and for process control is generally unreliable. Spikes in
TOC may be derived from both natural processes such as HABs and/or through
anthropogenic input. Measuring TOC removal to assess the efficiency of pretreatment
processes is also inaccurate due to the difficulties in measuring low level TOC residuals in
seawater process streams. Of the conventional water quality parameters used in desalination,
the SDI, despite its well known limitations, has proven useful in detecting algal blooms at the
intake compared to other parameters including turbidity and chlorophyll-a (determined via
fluorescence). Notwithstanding, care should be taken in interpreting results, as the SDI test
was not designed for high fouling feedwater such as algae-laden seawater. As a result, it can
underestimate the fouling potential of feedwater. Moreover, it does not include the smaller,
more fouling AOM produced during a bloom.
Recently, more sophisticated tests have been developed to determine constituents of AOM
which may better indicate the biofouling and particulate fouling potential of seawater and
process streams during a bloom. Villacorte (2014) demonstrated that algal-derived TEP and
biopolymers can promote fouling of both pretreatment and SWRO membranes and developed
tests to measure the concentration of larger TEP (TEP0.4μm) and smaller TEP precursors
(TEP10kDa) in seawater. Applications of both tests showed the smaller and more fouling TEP
precursors dominated AOM during a bloom and allowed the differences in the efficiency of
pretreatment in removing these algal-derived foulants to be distinguished. For instance, UF
was found to be superior in removing TEP precursors in comparison to conventional
coagulation-sedimentation-sand filtration.
AOM generated during a HAB may lead to an increase in the biodegradability of organic
matter in seawater and serve as a substrate for bacterial growth causing biofouling. A
recently developed AOC test, based on luminescence, to measure the biofouling potential of
seawater using Vibrio harveyi which utilizes low molecular weight organics, demonstrated an
increase in AOC during a HAB event compared to normal operating conditions. Other AOC
tests are under development to incorporate high molecular weight organic compounds such as
biopolymers generated during a bloom. As with AOC, the determination of TEP and the
MFI-UF have both proven valuable in laboratory and field trials. They offer distinct
advantages over the SDI. The MFI-UF has been applied to optimize coagulant dose to reduce
the fouling potential of feedwater containing AOM and to investigate pretreatment efficiency
163
during a bloom. Promising results were found when using a 10 kDa membrane in the MFI-
UF test. A high correlation was found between the MFI-UF and TEP10kDa, indicating the
MFI-UF10kDa could be used to investigate the fouling potential of feedwater containing the
smaller high fouling TEP precursors across a plant during a bloom.
While the AOC, TEP and MFI-UF tests offer more targeted information on AOM
constituents and the potential biofouling and particulate fouling potential of a feedwater
during a bloom than conventional parameters used in the industry, the degree of difficulty in
determining them is correspondingly higher. The tests require skilled analysts and specialized
equipment. At present, TEP, AOC and MFI-UF cannot be directly employed as a trigger to
alert a plant of a bloom in the incoming feedwater or to adjust process parameters during
plant operation. Nonetheless, these parameters are expected to be key factors in developing
our understanding of AOM fouling in seawater UF and RO systems, and therefore in efforts
to control them.
5.7! REFERENCES
Al-Hadidi, A. M. M. 2011. Limitations, Improvements, Alternatives for the Silt Density

Index, Enschede, Gildeprint Drukkerijen.
Al Shehhi, M. R., Gherboudj, I. and Ghedira, H. 2014. An overview of historical harmful
algae blooms outbreaks in the Arabian Seas. Marine Pollution Bulletin 86(1), 314-324.
Alldredge, A. L., Passow, U., and Logan, B. E. 1993. The abundance and significance of a
class of large, transparent organic particles in the ocean. Deep Sea Research Part I:
Oceanographic Research Papers 40(6), 1131-1140.
Arruda-Fatibello, S. H. S., Henriques-Vieira, A. A., and Fatibello-Filho, O. 2004. A rapid
spectrophotometric method for the determination of transparent exopolymer particles
(TEP) in freshwater. Talanta 62(1), 81-85.
APHA. 2012. Standard methods for the examination of water and wastewater, 21st edition.
American Public Health Association, Washington, D.C.
ASTM D2579-93e1, Standard Test Method for Total Organic Carbon in Water (Withdrawn
2002), ASTM International, West Conshohocken, PA.
ASTM D4129-05(2013), Standard Test Method for Total and Organic Carbon in Water by
High Temperature Oxidation and by Coulometric Detection, ASTM International, West
Conshohocken, PA.
ASTM D4189-07(2014), Standard Test Method for Silt Density Index (SDI) of Water,
ASTM International, West Conshohocken, PA.
ASTM D4839-03(2011), Standard Test Method for Total Carbon and Organic Carbon in
Water by Ultraviolet, or Persulfate Oxidation, or Both, and Infrared Detection, ASTM
International, West Conshohocken, PA.
ASTM D8002-15e1 (2015), Standard Test Method for Modified Fouling Index (MFI-0.45) of
Water, ASTM International, West Conshohocken, PA.
Bar-Zeev, E., Berman-Frank, I., Girshevitz, O., and Berman, T. 2012. Revised paradigm of
aquatic biofilm formation facilitated by microgel transparent exopolymer particles.
Proceedings of the National Academy of Sciences 109(23), 9119-9124.
Baumann, P., Furniss, A. L., and Lee, J. V. 1984. Genus I, Vibrio pacini 1854, 411 AL. In:
Krieg, N. R. and Holt, J. G. (eds) Bergey’s Manual of Systematic Bacteriology, Vol. 1.
Baltimore, Williams & Wilkins, p. 518-538.
164
Berman, T., and Holenberg, M. 2005. Don't fall foul of biofilm through high TEP levels.
Filtration and Separation 42, 30-32.
Boerlage, S. F. E. 2001. Scaling and particulate fouling in membrane filtration systems.
Swets & Zeitlinger Publishers, Lisse: pp. 242.
Boerlage, S. F. E. 2008 Understanding the Silt Density Index and Modified Fouling Indices
(MFI0.45 and MFI-UF). Desalination and Water Reuse Quarterly, May –June, 12-21.
Boerlage, S. F. E. 2012. Measuring salinity and TDS of seawater and brine for process and
environmental monitoring—which one, when? Desalination and Water Treatment 42(1-
3), 222-230.
Boerlage, S. F. E, Kennedy, M., Aniye, M., Abogrean, E., El-Hodali, D., Tarawneh, Z., and
Schippers, J. 2000. Modified Fouling Index - ultrafiltration to compare pretreatment
processes of reverse osmosis feedwater. Desalination 131, 201-214.
Boerlage, S. F. E., Kennedy, M., Aniye, M., Tarawneh, Z., Schippers, J. 2004. The Modified
Fouling Index-ultrafiltration in constant flux mode. Desalination 161, 103-113.
Boerlage, S. F., Kennedy, M. D., Dickson, M. R., El-Hodali, D. E. and Schippers, J. C. 2002.
The modified fouling index using ultrafiltration membranes (MFI-UF): characterisation,
filtration mechanisms and proposed reference membrane. Journal of Membrane Science,
197(1), pp.1-21.
Coble, P. G., Schultz, C. A. and Mopper, K. 1993. Fluorescence contouring analysis of DOC
intercalibration experiment samples: a comparison of techniques. Marine Chemistry
41(1-3), 173-178.
Du Pont Permasep Issued in 1972 in Technical Bulletin 300 Standard Millipore Test for
Determining Fouling Characteristics of Permeator Feed Supply.
Edzwald, J. K., and Tobiason, J. E. 2011. Chapter 3. Chemical Principles, Source Water
Composition, and Watershed Protection, In: Water Quality and Treatment: A Handbook
on Drinking Water, 6th edition. J. K. Edzwald, (Ed). American Water Works
Association, McGraw-Hill.
Evans, P. J., Smith, J. L., LeChevallier, M. W., Schneider, O. D., Weinrich, L. A., and
Jjemba, P. K. 2013. A Monitoring and Control Toolbox for Biological Filtration. Water
Research Foundation: Denver, CO. pp.1-92.
Franks, R., Wilf, M., Voutchkov, N., Murkute, P., and Kizer, J. 2006. A pilot study using
seawater reverse osmosis membranes in combination with various pretreatments to meet
the challenges of Pacific seawater desalination.
http://hydranautics.com/docs/papers/NewFolder/APILOTSTUDYUSINGSEAWATERR
EVERSEOSMOSISMEMBRANES.pdf. (Accessed 6 July 16).
Fujiwara, N., and Matsuyama, H. 2008. Elimination of biological fouling in seawater reverse
osmosis desalination plants. Desalination 227, 295–305.
Haddix, P. L., Shaw, N. J., and LeChevallier, M. W. 2003. Development of a simple assay.
American Water, Voorhees, NJ.
Haddix, P. L., Shaw, N. J., and LeChevallier, M. W. 2004. Characterization of
bioluminescent derivatives of assimilable organic carbon test bacteria. Applied and
Environmental Microbiology 70, 850–854.
Horobin, R. W. 1988. Understanding histochemistry: Selection, evaluation and design of
biological stains. Wiley, New York:
165
Huber, S. A., and Frimmel, F. H. 1991. Flow injection analysis of organic and inorganic
carbon in the low-ppb range. Analytical Chemistry 63, 2122-2130.
Huber, S. A. 2009. Personal Communication.
Huber S. A., Balz A., Abert, M., and Pronk, W. 2011. Characterisation of aquatic humic and
non-humic matter with size-exclusion chromatography – organic carbon detection –
organic nitrogen detection (LC-OCD-OND). Water Research 45(2), 879-885.
Jeong, S., Naidu, G., Vigneswaran, S., Ma, C. H., and Rice, S. A. 2013. A rapid
bioluminescence-based test of assimilable organic carbon for seawater. Desalination 317,
160-165.
Jeong, S., Rice, S. A., and Vigneswaran, S. 2014. Long-term effect on membrane fouling in a
new membrane bioreactor as a pretreatment to seawater desalination. Bioresource
Jin, Y., Lee, H., Jin, Y. O. and Hong, S., 2017. Application of multiple modified fouling
index (MFI) measurements at full-scale SWRO plant. Desalination 407, 24-32.
Kremen, S. S. and Tanner, M. 1998. Silt density indices (SDI), percent plugging factor (%
PF): their relation to actual foulant deposition. Desalination 119(1), 259-262.
Ladner, D. A., Vardon, D. R., and Clark, M. M. 2010. Effects of shear on microfiltration and
ultrafiltration fouling by marine bloom-forming algae. Journal of Membrane Science
356(1), 33-43.
LeChevallier, M. W., Shaw, N. E., Kaplan, L. A., and Bott, T. L. 1993. Development of a
rapid assimilable organic carbon method for water. Applied and Environmental
Microbiology 59, 1526–1531.
Nada, N., Personal Communication, 2012.
Passow, U. and Alldredge, A.L. 1995. A dye-binding assay for the spectrophotometric
measurement of Transparent Exopolymer Particles (TEP). Limnology and
Passow, U. 2000. Formation of transparent exopolymer particles (TEP) from dissolved
precursor material. Marine Ecology Progress Series 192, 1-11.
Petry, M., Sanz, M. A., Langlais, C., Bonnelye, V., Durand, J. P., Guevara, D., Nardes, W. M.
and Saemi, C. H. 2007. The El Coloso (Chile) reverse osmosis plant. Desalination
203(1-3), 141-152.
Pool, S. S., Krembs, C., Bos, J., and Sackmann, B. 2015. Physical, Chemical, and Biological
Conditions during Noctiluca Blooms in an Urban Fjord, Puget Sound,
https://fortress.wa.gov/ecy/publications/SummaryPages/1503040.html. (Accessed July
2016).
Ramus, J. 1977. Alcian Blue: A quantitative aqueous assay for algal acid and sulfated
polysaccharides. Journal of Phycology 13, 345-348.
catastrophic 2008-2009 red tide in the Gulf region, with observations on the
identification and phylogeny of the fish-killing dinoflagellate Cochlodinium
polykrikoides. Harmful Algae 9, 163-172.
166
Rouwenhorst, R. J., Jzn, J. F., Scheffers, A., and van Dijken, J. P. 1991. Determination of
protein concentration by total organic carbon analysis. Journal of Biochemical and
Biophysical Methods 22, 119-128.
Sack, E. L., van der Wielen, P. W., and van der Kooij, D. 2011. Flavobacterium johnsoniae
as a model organism for characterizing biopolymer utilization in oligotrophic freshwater
environments. Applied and Environmental Microbiology 77(19), 6931-6938.
Salinas-Rodriguez, S. G. 2011. Particulate and organic matter fouling of SWRO systems:
characterization, modelling and applications. Doctoral dissertation, UNESCO-
IHE/TUDelft, Delft.
Salinas-Rodriguez, S. G., Amy, G. L., Schippers, J. C., and Kennedy, M. D. 2015. The
Modified Fouling Index Ultrafiltration constant flux for assessing particulate/colloidal
fouling of RO systems. Desalination 365, 79-91.
Schippers, J. C., and Verdouw, J. 1980. The Modified Fouling Index, a method of
determining the fouling characteristics of water. Desalination 32, 137-148.
Schneider, O. D., Giraldo, E., Weinrich, L., Salinas, S., and Kennedy, M. 2012. Investigation
of Organic Matter Removal in Saline Waters by Pretreament. Water Research
Foundation: Denver, CO.
Tabatabai, S. A. A. 2014a. Coagulation and Ultrafiltration in Seawater Reverse Osmosis
Pretreatment UNESCO-IHE/TU Delft, Delft.
Tabatabai, S. A. A., Schippers, J. C., and Kennedy, M. D. 2014b. Effect of coagulation on
fouling potential and removal of algal organic matter in ultrafiltration pretreatment to
seawater reverse osmosis. Water Research 59, 283-294.
Thangaraja, M., Al-Aisry, A. and Al-Kharusi, L. 2007. Harmful algal blooms and their
impacts in the middle and outer ROPME sea area. International Journal of Oceans and
Oceanography, 2(1), pp.85-98.
Thornton, D. C. O., Fejes, E. M., Dimarco, S. F., and Clancy, K. M. 2007. Measurement of
acid polysaccharides (APS) in marine and freshwater samples using alcian blue.
Limnology and Oceanography: Methods 5, 73–87.
Van der Kooi, D. 1978. The occurrence of Pseudomonas spp. in surface water and in tap
water as determined on citrate media. Antonie van Leeuwenhoek, Journal Microbiology.
43(2), 187-197.
Van der Kooij, D., Visser, A., and Hijnen, W. A. M. 1982. Determining the concentration of
easily assimilable organic carbon in drinking water. Journal American Water Works
Association 74, 540-545.
Villacorte, L. O. 2014. Algal blooms and membrane-based desalination technology. PhD
Thesis UNESCO-IHE/TUDelft, ISBN 978-1-138-02626-1, CRC Press/Balkema, Leiden.
Villacorte, L. O., Ekowati, Y., Neu, T. R., Kleijn, J. M., Winters, H., Amy, G., Schippers, J.
C. and Kennedy, M. D. 2015a. Characterisation of algal organic matter produced by
bloom forming marine and freshwater algae. Water Research 73, 216–230.
Villacorte, L. O., Ekowati, Y., Winters, H., Amy, G., Schippers, J. C., and Kennedy, M. D.
2015b. MF/UF rejection and fouling potential of algal organic matter from bloom-
forming marine and freshwater algae. Desalination 367, 1-10.
167
Villacorte, L. O., Ekowati, Y., Calix-Ponce, H. N., Amy, G. L., Schippers, J. C., Kennedy, M.
D. 2015c. Improved method for measuring transparent exopolymer particles (TEP) and
their precursors in fresh and saline water. Water Research 70, 300–312.
Villacorte, L. O., Kennedy, M. D., Amy, G. L., and Schippers, J. C. 2009. The fate of
Transparent Exopolymer Particles (TEP) in integrated membrane systems: removal
through pretreatment processes and deposition on reverse osmosis membranes. Water
Research 43(20), 5039-5052.
Villacorte, L. O., Schurer, R., Kennedy, M. D., Amy, G. L., and Schippers, J. C. 2010.
Removal and deposition of Transparent Exopolymer Particles in a seawater UF-RO
system. IDA Journal of Desalination and Water Reuse 2(1), 45-55.
Vrouwenvelder, J. S., Manolarakis, S. A., Veenendaal, H. R., and Van der Kooij, D. 2000.
Biofouling potential of chemicals used for scale control in RO and NF membranes.
Desalination 132(1), 1-10.
Vrouwenvelder, J. S., Van Paassen, J. A. M., Wessels, L. P., Van Dam, A. F., and Bakker, S.
M. 2006. The membrane fouling simulator: a practical tool for fouling prediction and
control. Journal of Membrane Science 281(1), 316-324.
Weinrich, L. A., Giraldo, E., and LeChevallier, M. W. 2009. Development and application of
a bioluminescence-based test for assimilable organic carbon in reclaimed waters.
Applied and Environmental Microbiology 75, 7385–7390.
Weinrich, L. A., Jjemba, P. K., Giraldo, E., and LeChevallier, M. W. 2010. Implications of
organic carbon in the deterioration of water quality in reclaimed water distribution
systems. Water Research, 44(18), 5367-5375.
Weinrich, L. A., LeChevallier, M. W., and Haas, C. N. 2015. Application of the
bioluminescent saltwater assimilable organic carbon test as a tool for identifying and
reducing reverse osmosis membrane fouling in desalination. WaterReuse Research
Foundation, Alexandria, VA.
Weinrich, L., LeChevallier, M., and Haas, C. N. 2016. Contribution of assimilable organic
carbon to biological fouling in seawater reverse osmosis membrane treatment. Water
Research 101, 203-213.
Weinrich, L. A., Schneider, O. D., and LeChevallier, M. W. 2011. Bioluminescence-based
method for measuring assimilable organic carbon in pretreatment water for reverse
osmosis membrane desalination. Applied and Environmental Microbiology 77, 1148–
1150.
Zhang, M., Jiang, S., Tanuwidjaja, D., Voutchkov, N., Hoek, E. M., and Cai, B. 2011.
Composition and variability of biofouling organisms in seawater reverse osmosis
desalination plants. Applied and Environmental Microbiology 77(13), 4390-4398.
Zhao, J. and Ghedira, H. 2014. Monitoring red tide with satellite imagery and numerical
models: a case study in the Arabian Gulf. Marine Pollution Bulletin 79(1), 305-313.
Zhou, J., Mopper, K., and Passow, U. 1998. The role of surface-active carbohydrates in the
formation of Transparent Exopolymer Particles by bubble adsorption of seawater.
Limnology and Oceanography 43(8), 1860-18.
168
Seawater intake considerations to mitigate HAB impacts
6 SEAWATER INTAKE CONSIDERATIONS TO MITIGATE

HARMFUL ALGAL BLOOM IMPACTS
Siobhan F.E. Boerlage1, Thomas M. Missimer2, Thomas M. Pankratz3, and Donald M.

Anderson4
1
2
Florida Gulf Coast University, Fort Myers, Florida 33965-6565 USA
3
Water Desalination Report, Houston, TX, USA
4
Woods Hole Oceanographic Institution, Woods Hole, MA 02543 USA

6.1 Introduction ...........................................................................................................................................169
6.2 Intake options for SWRO desalination plants .......................................................................................171
6.3 Surface intake and screen options .........................................................................................................172
6.3.1 Onshore and offshore intake screening .........................................................................................173
6.3.1.1 Traveling Water Screens........................................................................................................174
6.3.1.2 Rotating Drum Screens ..........................................................................................................174
6.3.1.3 Velocity Caps.........................................................................................................................175
6.3.1.4 Passive Screens ......................................................................................................................175
6.3.1.5 Intake Head ............................................................................................................................176
6.3.2 Surface intake strategies to minimize harmful algal bloom (HAB) impacts.................................176
6.3.3 Deep-water intakes ........................................................................................................................178
6.4 Subsurface intake options......................................................................................................................181
6.4.1 Description of intake types with example installations .................................................................184
6.4.1.1 Conventional vertical wells ...................................................................................................184
6.4.1.2 Collector or Ranney Wells.....................................................................................................187
6.4.1.3 Angle wells ............................................................................................................................188
6.4.1.4 Horizontal wells (HDD) ........................................................................................................188
6.4.1.5 Beach gallery systems............................................................................................................189
6.4.1.6 Offshore or seabed gallery systems .......................................................................................190
6.4.1.7 New subsurface intake designs ..............................................................................................192
6.4.2 Subsurface intake performance for algae and NOM removal .......................................................193
6.4.3 Planning of desalination plants with subsurface intakes ...............................................................194
6.5 Siting of desalination seawater intakes .................................................................................................197
6.6 Summary ...............................................................................................................................................199
6.7 References .............................................................................................................................................200
6.1 INTRODUCTION
Seawater intakes are a key element in the design, construction and success of desalination
plants. Various intake options exist and are generally classified based on their abstraction
depth. Surface ocean intakes abstract seawater from the top of the water column or at depth,
while subsurface 1 intakes are embedded in the seabed or beach, thereby pre-filtering the
abstracted seawater. Location, intake type and depth are important determinants of water
quality. Intakes are also the first point of control in minimizing the ingress of algae into a
plant or where algal impacts first manifest.
Originally the more robust thermal desalination processes dominated the desalination market
where feedwater quality was not the primary driver in determining intake type or location.
Instead, feedwater supply was critical, as thermal plants were configured as cogeneration
power/desalination plants with common intakes with large volume requirements to generate
1
Note that ‘subsurface’ in this context differs from common oceanographic usage, in which the term refers to
waters just below the air/water interface.
169
both power and water. Intake and screening systems were often limited to shallow nearshore
intakes with screens sized to meet the necessary seawater quality for power plant, multi-stage
flash (MSF) and multi-effect distillation (MED) condenser tubes (Pankratz 2015). Macro-
algal seaweed species were initially a significant issue in thermal desalination plants,
completely blinding intake screens or clogging settling basins (Figure 6.1). In the mid-1970s,
the availability of MSF thermal
plants in Libya was dramatically
reduced to 100 days/year, with
seaweed blockage of the intake
pipes the third leading cause for
plant outages. At the Zuara plant
intake, up to 800 m3 of seaweed
was removed every second day
during winter when seaweed
became dislodged from the seabed
at the end of summer and during
storms (Kreshman 1985; 2001).
Due to advances in the design of
intake systems, the extent of
macro-algal intake blocking has
Figure 6.1. Dry seaweed extracted from the Zwitina been greatly reduced at thermal
desalination plant intake channel in Libya. Photo: Kershman desalination plants and now mainly
1985. results in short term outages.
Nowadays with seawater reverse osmosis (SWRO) dominating the desalination market,
microscopic algal species (phytoplankton) have been more problematic. Occasionally issues
have occurred at plant intakes when a high suspended solids load of phytoplankton and debris
have overloaded trash racks and/or clogged intake screens (Figure 6.2). In some cases, these
impacts have been severe. The notorious 2008/2009 bloom of Cochlodinium polykrikoides in
the Gulf of Oman resulted in the frequency of cleaning seawater intake screens at Sohar
increasing to every 4 hours (Sohar Case Study, Chapter 11). More often adverse impacts are
observed in downstream SWRO pretreatment processes or through the promotion of
(bio)fouling on membranes as microscopic algae and algal organic matter (AOM) pass
through conventional open intakes and screens.
Figure 6.2. Algae and other marine debris blocking the Traveling Water Screens (left) at a SWRO
desalination plant in the Indian Ocean and the screens following cleaning (right). Photos: Domingo Zarzo
Martinez, Valoriza Agua S.L.
170
The potential for phytoplankton and AOM to be entrained into SWRO plant intakes, the
focus of this chapter, varies greatly. In addition to the intake system design characteristics,
prevailing marine conditions, nutrient concentrations at the site, the type, motility and
concentration of the algal bloom species play a role. Intake characteristics are recognized to
have a significant effect on raw seawater quality and therefore the pretreatment processes
required, as well as limiting marine environment impacts which can be a major concern in
some projects. Consequently, more attention is given to the selection and location of intake
systems in SWRO feasibility studies and during design.
In areas prone to algal blooms, subsurface or open intakes abstracting seawater at depth are
often considered a solution to reduce the ingress of floating or surface-concentrated algal
blooms into desalination plant intakes. Subsurface intakes offer the advantage that they serve
both as a water intake and as pretreatment for a SWRO plant. The seawater is filtered during
passage through the strata of the subsurface intake, removing algae and natural organic
matter, including components of AOM by both physical and biochemical processes,
providing a high-quality feedwater, thereby potentially reducing or replacing conventional
pretreatment processes (Missimer et al. 2013; Rachman et al. 2014; Dehwah et al. 2015;
Dehwah and Missimer 2016). The effectiveness of these strategies for reducing the
entrainment of algae and associated AOM into an intake is discussed below. It should be
noted this is a little–studied area in the desalination industry, especially during algal bloom
events. Therefore, research on the removal of fractions of natural organic matter (NOM) such
as biopolymers produced by both bacteria and algae are examined here, as results may be
indicative of what can occur during an algal bloom. Finally, other factors such as engineering
constraints, environmental concerns, costs, construction time, and operability may ultimately
drive the selection and siting of an intake. A brief overview of approaches to determine the
seawater intake for a project is therefore provided in the last part of this chapter.
6.2 INTAKE OPTIONS FOR SWRO DESALINATION PLANTS
Seawater desalination plants require an intake system that is capable of reliably delivering the
seawater flow to meet production requirements. Secondly, and arguably equally important for
a SWRO plant, the intake ideally delivers water that is high quality and consistent over time -
free of pollutants with a low solids and organic load to minimize pre-treatment complexity
and chemical consumption. The latter reduces the generation of waste requiring disposal. The
key environmental concern in intake design is to reduce the potential for marine life mortality.
This may occur through impingement of marine life i.e. when larger organisms (typically
juvenile and adult stages) are trapped against an intake screen—and entrainment—when
smaller organisms (typically phytoplankton and early life stages— eggs and larvae) pass
through a screen into the process during intake operation.
Intakes therefore, not only play a significant role in SWRO plant capital and operating costs,
but designs are highly site-specific—possibly more so than any other aspect of the plant—
and have a considerable impact on the operational and environmental aspects of the plant.
Additionally, as the initial step in the pretreatment process, the intake effectiveness plays an
important role in determining the performance of downstream processes. Intakes are broadly
classified based on their abstraction depth and location from shore as described below and are
discussed more fully in the following sections with respect to preventing entrainment of
microscopic algae:
• Open ocean (or surface) intakes - where seawater is abstracted at the sea surface,
within 15 m of the surface (shallow intakes) or at water depths of greater than 15-
20 m. Intakes may be located onshore, nearshore or offshore. Feedwater to the plant
171
derived from surface intakes is dependent on the inherent seawater quality and will
fluctuate depending on prevailing marine and site conditions; and
• Subsurface intakes which abstract seawater from beneath the seabed or beach.
Filtration through the seabed strata generally provides a superior water quality due to
the removal of suspended solids, turbidity and some organics.
6.3 SURFACE INTAKE AND SCREEN OPTIONS
Large-scale desalination plants have traditionally employed open ocean surface intakes with
associated screening and chlorination plants. In most arrangements, the pump station and
screening chamber are located onshore and directly connected to the open ocean by means of
a concrete channel or jetty or an intake pipeline (or tunnel) which can extend hundreds of
meters into the sea (Figure 6.3).
For shallow offshore areas, it is
common to locate the raw water
intake structure well beyond the
surf zone, where it is less
vulnerable to damaging wave
action and turbidity entrainment.
In some instances, it may be up to
500 m or more from shore
depending on the bathymetry, to
enable the abstraction of water
from deeper, less environmentally
sensitive areas or to obtain more
consistent water quality that is less
susceptible to varying debris loads.
Note that being far from shore
does not necessarily reduce the
potential impact from algal
blooms, as many originate offshore
and are transported to the
nearshore waters by winds and
currents. Likewise, blooms near
the shore can be transported
Figure 6.3. Surface seawater intake alternatives – conventional further offshore by what are
shoreline intake through a lagoon or channel and offshore with termed “upwelling-favorable”
typical coarse and fine screening arrangements.
winds (Chapter 1).
To remove flotsam and larger debris, intake seawater is typically screened at the head of the
plant by coarse primary screening (bars) followed by finer secondary screening to protect
pumps and downstream processes. Screens need to be cleaned to remove pressure losses
caused by debris fouling and ensure flow, and/or sized with larger openings to avoid marine
build up. Typical screening options for desalination plants are discussed in more detail in
section 6.3.1.
In addition to screening of the seawater, chlorination is often applied at surface seawater
intakes to control marine biofouling on screens, piping, and pumps, although the practice
varies based on the desalination process employed and whether the intake also provides water
for cooling or other purposes. Thermal desalination plants generally add 1 to 2 mg/L of
chlorine continuously to maintain a 0.15 to 0.3 mg/L residual, and shock doses of up to
8 mg/L for 15-30 minutes several times a day. Most SWRO plants practice intermittent
172
chlorination/dechlorination with doses of up to 10 mg/L added for up to two hours on a daily,

weekly or biweekly basis. Chlorination is now most commonly used on a periodic basis
rather than continuously because it is known to cause biofouling of the downstream
membranes (Winters et al. 1997). As an additional measure, components of screens e.g. bars
may be constructed of alloys with biocidal properties such as cupronickel 90/10 to prevent
marine biogrowth.
Historically, large-scale thermal (MSF or MED) seawater desalination plants were coupled to
electric power plant to provide a steam source for the distillation process as cogeneration
plants. As power plants require large volumes of cooling water to condense power-cycle
steam, they are also able to share their seawater intake and screening infrastructure with the
desalination plant. Intake arrangements for such cogeneration plants are often open seawater
intakes of the channel or lagoon type which are connected to the screening chamber located
at the shoreline, or some distance
inland. The shared intake system,
based on that for power plant
cooling water, were developed
more than one hundred years ago
and typically consist of vertical
trash racks (100 mm spacing)
fitted with raking machines which
can remove floating debris and
filamentous algae. This is
followed by onshore screening
chambers, or wet wells, equipped
with mechanically cleaned,
traveling water screens or rotary
Figure 6.4. Traditional open seawater intake showing primary drum screens such as that shown
screening using trash racks followed by fine screening using in Figure 6.4.
Traveling Water Screens. Photo: Evoqua.
Initially, large capacity SWRO
plants followed this intake arrangement, particularly those co-located with electric power
plants or configured as hybrid desalination systems (i.e. thermal combined with SWRO
systems). Nowadays, the lower capital and operating costs of SWRO compete favorably with
thermal desalination processes, particularly for stand-alone desalination plants where no
existing intake or outfall exists. Even if an existing open ocean intake were available, the
SWRO process requires feedwater with a much lower level of suspended solids, both in
terms of particle size and volume, than thermal processes, which may necessitate a purpose-
built intake.
6.3.1 Onshore and offshore intake screening
Screen selection and configuration is influenced by a variety of factors such as type and
abundance of marine flora and fauna at site, impingement onto the screen, the risk of
entrainment into the intake, the type of pumps and pretreatment proposed downstream and its
ability to remove and handle solids. The most common techniques to mitigate the
impingement and entrainment of marine life is to lower the velocity of water through the
intake screen to less than 0.15 m/s and reduce the size of the screen openings to 1 mm or less,
respectively. Algae would be defined as entrainable organisms due to their size. Some
entrainable organisms have limited to no swimming ability and therefore lack the ability to
avoid the intake flow regardless of velocity (Hogan 2015).
173
Active (moving) or mechanical screens used for fine screening are located onshore in
concrete channels either at the far end of a forebay or a longer channel that extends out
beyond the surf zone. Alternatively, the screens may be installed in a wet well or pump
station that is connected to the sea by a pipe that extends out into the sea and terminated in a
coarse screened inlet head or a velocity cap. Unless the intake terminal of an offshore intake
is fitted with a passive (stationary) screen system, the onshore pump station should be
equipped with fine wire mesh screens to protect downstream pumps and pretreatment
equipment (Pankratz 2015). The mesh size of the mechanical screens generally depends on
the desalination process. In thermal MSF and MED plants the mesh openings range from 6 to
9.5 mm with smaller mesh openings of 0.5 to 5 mm sometimes used for MED plants as MED
needs finer filtration. For MED the allowable particle size for seawater going through the
spray nozzles is <0.5 mm.
6.3.1.1 Traveling Water Screens
Traveling Water Screens, also referred to as Traveling Band Screens, have been employed on
seawater intakes since the 1890s. The screens are equipped with revolving wire mesh panels
having 6 to 9.5 mm openings, although environmental regulations in the United States—
specifically §316(b) of the US EPA Clean Water Act—mandate that many intake screens
employ finer mesh screens with openings as small as 1.0 mm to minimize entrainment. The
screens are also usually designed so that
the maximum water velocity through
the screen is less than 0.15 m/s.
As the wire mesh panels revolve out
during flow, a high-pressure water spray
removes accumulated debris by washing
it into a trough for dewatering and
further disposal.
There are two distinct types of
Traveling Water Screens: a Through-
Flow screen in which the screening
Figure 6.5. Traveling Water Screens: Through-Flow, left panels are oriented perpendicular to the
and Dual Flow, right. Photos: Evoqua. flow with only the ascending panels
utilized as available screening area; and
the Dual Flow or central-flow type screen in which
the screening panels are oriented parallel to the flow,
utilizing both the ascending and descending panels
as active screening area (Figure 6.5). Besides
providing more active screening area per unit, a dual
flow traveling water screen virtually eliminates the
chances of ‘carry over’, where debris not removed
by the spray system would otherwise fall into the
screened water side of the unit and enter the pumps.
6.3.1.2 Rotating Drum Screens
Rotating Drum Screens are an alternative to
Traveling Water Screens, and consist of wire mesh
panels mounted on the periphery of a large cylinder
that slowly rotates on a horizontal axis (Figure 6.6). Figure 6.6. Rotating Drum Screen. Photo:
They are cleaned with a spray wash system similar Ovivo.
174
to Traveling Water Screens. Drum Screens may range up to 4 m in diameter and have similar
size openings to the traveling water screens.
6.3.1.3 Velocity Caps
As mentioned, above, the most common technique to minimize impingement mortality of
marine life is to reduce the through-screen velocity of an intake structure to ≤0.15 m/s
allowing fish to swim away from the currents generated at the intake (EPA 2014). An
alternative is to fit the vertical riser of an offshore open intake with a “velocity cap” which
acts as a behavioral deterrent to guide aquatic organisms away from the intake structure. A
velocity cap is a horizontal, flat cover located slightly above the terminus of the riser which
provides a narrow opening for the entrance of seawater. Intake water drawn through the
openings in the velocity cap is converted from vertical flow to horizontal flow into the pipe.
Rapid changes in horizontal flow will provide a physiological trigger in fish inducing an
avoidance response thereby avoiding impingement (EPA 2014). Some capped intake risers
operate at lower through-screen velocities (≤0.15 m/s), but may not function as an effective
fish diversion technology. Most velocity caps operate at higher entrance velocity with the
change in flow pattern created by a velocity cap operating at an entrance velocity of over
0.3 m/s, and as high as 0.9 m/s, triggering an
avoidance response mechanism in fish.
Extending the cap and riser lip by 1.5 times
the height of the opening has been shown to
result in a more uniform entrance velocity,
and improves the ability of fish to react and
avoid the intake (Figure 6.7). However, as
with all intake configurations, there are many
design issues that must be considered, and the
performance of a velocity cap may vary in
still water versus areas subject to tidal cross-
Figure 6.7. Cross-section of a Velocity Cap. flows. Virtually all velocity cap intakes
require some on-shore screening system,
usually a Traveling Water Screen or Rotating Drum Screen, to protect downstream pumps
and pretreatment equipment.
6.3.1.4 Passive Screens
A passive screen intake utilizes one or more fixed cylindrical screens (barrel screens)
manufactured of trapezoidal- or triangular-shaped ‘wedge wire’ bars arranged to provide 0.5
to 3.0 mm wide slotted openings. The screens are usually oriented on a horizontal axis with
the total screening area sized to maintain a through flow velocity of less than 0.15 m/s to
minimize marine life and debris impingement (Figure 6.8). Passive screens are best suited for
areas with ambient cross-flow currents that act to ‘self-clean’ the screen face, but may still be
impacted by the attachment of organisms such as coral barnacles, or shellfish. Systems may
also be equipped with an air backwash system to clear the screens when debris accumulation
occurs. In most cases, the screens are located at least one screen diameter from the seabed.
175
Figure 6.8. Seabed mounted passive screen left and bulkhead mounted passive screen right. Photos:
Johnson screens.
Passive screens have been used on many smaller plants around the world, particularly those
with nearshore intakes allowing the compressed air plants to be located on shore without
undue pressure losses in air transmission pipes. Passive screens have a proven ability to
reduce impingement—due to their low through-flow velocities—and entrainment—through
exclusion resulting from the narrow slot openings. Tests have shown that 1 mm openings are
highly effective for larval exclusion and may reduce entrainment by 80 % or more (Pankratz
2015.)
6.3.1.5 Intake Head
For an offshore intake that does
not employ passive screens or a
velocity cap, the intake pipe
terminus can be fitted with an
intake head that is designed
with a coarse bar or grid with
25 to 200 mm spacing.
Examples are shown in Figure
6.9.
6.3.2 Surface intake
strategies to minimize
harmful algal bloom (HAB)
impacts
Unlike subsurface intakes
where seawater withdrawal
takes place indirectly, beneath
the seabed, open seawater
intakes are directly exposed to
algal blooms and other natural
or anthropogenic increases in
debris loading that can inundate
an intake facility. Plants with
open seawater intakes must
Figure 6.9. Various offshore intake heads. From top left, clockwise therefore develop strategies—
to bottom left, courtesy of WaterSecure, Increa, Water Corporation, whether to change operating
and Abengoa.
tactics, reduce production or
176
shut down entirely—to deal with these inevitable occurrences.

Careful site selection is the first defense against algal blooms. Although occurrences may be
sporadic and difficult to predict, historical records that address the frequency and severity of
events, and the conditions that led up to the blooms should be considered and factored into
siting and operational strategies.
For offshore intakes, it may be possible to choose an intake location or the location of passive
screens so as to avoid areas of greatest potential algae concentrations and the entrainment of
algal blooms into the intake. For example, the intake point could be located in deeper water
and/or farther offshore as discussed in the next section or at a water depth where algae are
less likely to accumulate. Whereas, for nearshore intakes that include long approach channels
to the intake pumping station, algal concentrations approaching or within the channels could
be monitored in order to guide possible management actions such as the addition of
coagulation prior to ultrafiltration or modification of chlorination strategy. In situ chlorophyll
or optical sensors of several types are described in Chapter 3.
Adopting a lower through-screen velocity of ≤0.15 m/s is expected to have little impact on
the type of debris and suspended solids that are entrained into an intake, but this generally
results in slower debris build-up, which mechanical screens find easier to handle. Even so, an
algal bloom and the high suspended solids associated with it may result in debris loading
conditions that are higher than the screen capacity.
A velocity cap will also not be effective in reducing the entrainment of motile e.g.
dinoflagellates or non-motile algal blooms as the water flow at velocity caps is far faster than
even the strongest dinoflagellates can swim (approximately 0.08 cm/s). Also, algal cells are
not able to “sense” the presence of the intake to take evasive action. Thus, there will not be
any appreciable decrease in the intake of planktonic organisms during a bloom. At the least,
there may be limited reduction in the ingress of HABs as velocity caps limit the zone of
influence of the intake to the depth level at which the velocity cap is situated, thus entraining
only the algal cells present at that depth (EPA 2014).
The most common technique for mitigating entrainment is to reduce the size of the screen
openings, often to 2.0 mm or less such as those of passive screens; however, although screens
with fine openings generally allow the ingress of algal cells, the vast majority of which are
smaller than 1 mm, they are vulnerable to sudden plugging conditions that may occur when
large mats of particulates, such as those produced by algal blooms or flux of polyps (coral
spawning), are encountered.
One strategy to deal with algal blooms, should they enter the intake, is to operate all traveling
screens or drum screens continuously, (including those incorporated in design for redundancy
purposes) during those periods when blooms are most likely to occur. This ensures that
screens are kept clean and that a sudden surge of debris will not overwhelm a screen,
possibly causing operational problems. For installations anticipating higher debris loads as in
areas prone to algal blooms, it is usually possible to add a second spray header to facilitate
debris cleaning and/or an additional lifting shelf to accommodate increased debris volumes.
Onshore or offshore intakes equipped with mechanical screens can also select screening
equipment that is designed to handle higher debris loads. These options include finer screen
mesh, i.e. 2.0 to 3.0 mm versus 6.0 to 9.5 mm, higher pressure spray wash systems, auxiliary
lifting shelves on the wire mesh panels, and the addition of mechanical raking mechanisms
on coarse screens or trash racks. For locations with a likelihood of encountering large
quantities of macroalgae, such as kelp and seaweed, it may be necessary to use auxiliary
177
toothed lifting ledges or ‘kelp knives’ to ensure that the revolving screens can retain the
debris and convey it to the discharge trough.
Finally, chlorination at an intake may be suspended during an algal bloom as chlorination
breaks down NOM into easily degradable compounds, also known as assimilable organic
carbon (AOC), that serve as nutrients for the regrowth of bacteria and may actually increase
growth rates (see Chapter 5 for saline AOC test). In addition, chlorine may result in the lysis
of algal cells and release of AOM such as sticky transparent exopolymer particles (TEP) that
can promote biofouling or release of intracellular toxins. Retaining toxins inside the algal
cells will improve their removal in SWRO pretreatment processes.
6.3.3 Deep-water intakes
Intake depth is an important determinant of water quality. Increasing the intake depth and/or
increasing the distance from shore and from coastal influences or discharges is promoted as a
means to improve water quality and thereby reduce SWRO pretreatment requirements, filter
clogging and membrane fouling. As the total water depth increases, there is typically less
turbulence and less suspended solids due to wave action in the water column, and reduced
risk of accidental pollution from hydrocarbon spills or leakage from shipping, which
typically impact the surface layer. Conversely, strong tidal currents in some areas can create a
“benthic nepheloid layer” (BNL), or layer of re-suspended sediment and detritus near the
bottom. These can be large (tens of meters thick) and persistent features in areas with strong
tidal currents, and could affect water quality for near-bottom intakes. Fortunately, they are
easily detected using transmissometers and vertical profiling, and thus can be avoided in the
intake design phase.
At greater intake depth the seawater temperature is generally more constant, which is easier
for plant operation although exceptions may be found, e.g., the passage of seasonally
occurring internal waves can result in rapid significant temperature changes (6˚C) at deep
water intakes (Boerlage and Gordon 2011). Furthermore, the seawater may be colder at depth,
which necessitates an increase in RO feedwater pressure or more membranes to meet plant
capacity than is the case with warmer surface water. Thermoclines can also limit mixing and
impact water quality (see La Chimba Case Study, Chapter 11 section 11.6) as can seasonal
haloclines which may occur resulting in an increase in salinity of the seawater observed
(Boerlage and Gordon 2011). Such temperature and salinity changes may be detected by
vertical profiling prior to design.
Deep water intakes, in addition to reducing the suspended solids load, may reduce the organic
load of the raw water as seawater drawn from the surface or the upper levels of the water
column is where photosynthesis occurs. This is referred to as the photic zone - the depth of
which varies with turbidity and season but can extend 50 – 75 m or more. More important is
the mixed layer, created by winds, waves, and other surface stresses. The mixed layer is often
shallower than the photic zone. Algae, zooplankton, and larvae are often most abundant in
this layer and thus shallow intakes or channels may be more prone to algal blooms, as found
at the Sohar SWRO desalination plant (see Sohar Case Study, Chapter 11 section 11.2).
Therefore, abstracting seawater at depth (e.g. more than 15 to 20 m below sea surface and
typically below the mixed layer) is a strategy often put forward to reduce the ingress of algal
cells and AOM generated during a bloom into desalination plant intakes. The risk of
entraining algal cysts in sediments can also be minimized as screens for deep water intakes
are commonly located 1.5 to 4 m above the seabed to reduce sand and sediment entrainment;
however, for multiple reasons, dense concentrations of algae (cells) and algal-related detritus
including organics can be found well below the water surface as discussed below, causing
problems for even deep intakes.
178
First, the position of the peak algal concentration may vary over time in the water column.
For example, motile algal species may display diel vertical movement in the water column
whereby they migrate up to the surface during the day to access sunlight for photosynthesis,
and swim downwards to more nutrient rich waters late in the day and during night,
descending up to 20 m (Chapter 1).
This is illustrated in Figure 6.10
showing chlorophyll-a measurements
for the toxic dinoflagellate
Gymnodinium catenatum over five
days in 20 m water depth. Higher
chlorophyll is observed in the top 5 m
during the day, but the center of mass
of the bloom then moves to more
than 17 m depth at night as the algae
swim downwards in the water
column. Dinoflagellates, the
causative species for most toxic
Figure 6.10. Chlorophyll-a (fluorescence) profiles from the HABs, have the ability to swim using
Huon Estuary in Tasmania, Australia, showing diel vertical their flagella. Speeds of 1 m/h are
migration of the phytoplankton (dominated by Gymnodinium
typical, but Cochlodinium
catenatum) over the 20 m depth of the water column. On the
x-axis, white bars indicate the light period and black bars polykrikoides, the species that
indicate dark. Data from CSIRO Huon Estuary Study resulted in plant outages in 2008 and
(modified from Doblin et al. 2006). To abstract water with a 2009 in the Gulf and Sea of Oman
lower concentration of algae, a shallow intake may benefit regions was shown to reach
from operation at night (dashed line) and for a deep-water
intake during the day (unbroken line).
swimming speeds of 3 m/h (Chapter
1 Section 1.5.9).
Additionally, algal cells may move through the water column in response to nutrient supply
or during various growth stages. Some diatoms, for example, can alter their buoyancy
through adjustment of the composition in their vacuoles so that in favorable conditions they
are found mainly on or near the surface (Moore and Villareal 1996). When growth slows due
to nutrient limitation, they can adopt a “sink strategy” altering buoyance such that the weight
of their siliceous cell walls helps them to settle to deeper layers of the water column where
nutrients are more abundant. Finally, when algal cells age and die they lose their buoyancy
and contribute to the oceanic “snow” that falls slowly to the seabed.
Hence, algal cells (and the oceanic snow) may still be entrained into intake screens
depending on the intake depth and the migration or aggregation depth of the algal species that
is blooming. Consequently, resultant operational issues downstream in the SWRO treatment
process may manifest continuously during an algal bloom event, or only at night for deep
water intakes if the bloom species displays diel vertical migration. In the latter case, the
opposite would be found for shallow intakes, meaning that the maximum algal cells would be
found during daylight hours and lower numbers at night. Hence, monitoring of the RO
feedwater for algal cell abundance or proxies such as chlorophyll should be conducted over
24 hours, as discussed in Chapter 1.
Similarly, deep water intakes may or may not result in a reduction of the AOM fouling
propensity of a surface water intake as the distribution of AOM in the water column may
differ from that of the algal cells negating the effect of lower algal concentration at depth. In
other words, a major bloom in surface waters can generate a large amount of organic detritus
that falls into the intake zone of deep intakes. AOM comprises not only cell-bound
(intracellular) organic matter which may be released through cell lysis or decay but also
179
extracellular organic matter released into the seawater through metabolic excretion by live
algal cells (Chapter 2). Organic matter generated by algal blooms varies significantly in its
composition and molecular weight ranging in size from small molecular weight toxins to
high molecular weight biopolymers which includes the sticky TEP and TEP precursors which
may initiate and promote the formation of biofouling of SWRO membranes. This is discussed
in detail in Chapter 2.
There are no known studies in the desalination industry examining the distribution of algae
and AOM in the water column during an algal bloom. Some research has examined the
distribution of algae, bacteria and NOM with depth, where the NOM will include compounds
produced by both bacteria and algae such as TEP. TEP present in seawater can be a mixture
of those produced by bacteria, algal blooms, and shellfish (Chapter 2). Therefore, results may
be indicative of what may occur during an algal bloom. One study (Dehwah et al. 2015)
investigating the potential for deep water intakes in the Red Sea off the coast of Saudi Arabia,
reviewed the mechanisms leading to movement of TEP in the water column. Key factors
include sedimentation of TEP towards the seabed from the abiotic polymerization of
dissolved TEP precursors along with TEP in marine snow, whereas the buoyancy of TEP
would lead to the upward movement of
TEP. Dehwah et al. (2015) conducted
water sampling to determine the vertical
distribution of algae and NOM. Water
samples were collected at the surface and
at 10 m intervals to a depth of 90 m at
three sites 85 km north of Jeddah and
analyzed for total algae and bacteria. Total
NOM in these samples was characterized
Figure 6.11. Algal cell counts in the Red Sea at the by LC-OCD to provide information on
surface, at 10m and 20m water depth at three sites off fractions of NOM such as biopolymers.
the coast of Saudi Arabia (estimated from profiles Particulate TEP (size greater than 0.4 µm)
presented in Dehwah et al. 2015).
and colloidal TEP 2 (with size ranging
between 0.1 and 0.4 µm) were also
measured. Algal cell counts, total TEP (sum of particulate and colloidal TEP) and the
concentration of biopolymers estimated from profiles in this study are presented in Figures
6.11 and 6.12 for the surface, 10 m and 20 m depths to correspond with surface, shallow and
deep-water intakes.
Although, sampling was apparently not during an algal bloom event, the total algal
concentration was relatively high, as was the bacterial concentration (ranging from 350,000
to 450,000 cells/mL at the surface). Synechococcus, a marine cyanobacterium (and classed as
algae) ubiquitous in the ocean, accounted for more than half of the algal population, and
along with the other algal species, varied between the three sites and with depth. With a
distance of 3-4 km between sites A and C, the wide range in total algal cell counts between
the sites was not unexpected. Algal assemblages are often patchy in nature at the surface due
to random horizontal migration or drift produced by winds, shifting currents and tides
(Chapter 1). Similarly, algae are not uniformly distributed in the water column as discussed
above. Synechococcus, the dominant species in the study, is a small picoplankton genus
(< 2µm in size), and about 1/3 of its species are motile and move through the water column.
2
Colloidal TEP in the Dehwah et al. (2015) study was measured using the earlier method developed by
Villacorte et al. 2009 using a 0.1 µm test membrane. The latest TEP method employs a smaller 10 kDa
membrane which will capture much smaller TEP and TEP precursors present during an algal bloom (see
Chapter 5 and Villacorte et al. 2015).
180
Figure 6.12. The concentration of biopolymers and total TEP in the Red Sea at the surface, at 10m and 20m
water depth at three sites off the coast of Saudi Arabia (estimated from profiles presented in Dehwah et al.
2015).
As discussed by the authors of the study, no substantial reduction in algal counts was found
between the surface and the first 50 m depth, with peak concentrations observed at 50 m
(total algal counts of up to 900,000 cells/ml) instead of in surface layers. Algal concentration
did decrease after 50 m, with the minimum concentration found at 90 m. In contrast, although
total bacterial numbers also varied between sites, it declined with depth for each site with
peak concentrations observed in the first 10 m.
The largest fraction of NOM at the three Red Sea sites was the low molecular weight humic
acids; the concentration of biopolymers was substantially lower. The biopolymer fraction of
NOM measures proteins and polysaccharides, including TEP, of both algal and bacterial
origin (see Chapter 5). During an algal bloom event, spikes in biopolymers and TEP would
be attributable to the algae. The biopolymer fraction of NOM and particulate and colloidal
TEP varied between sites and with depth but biopolymers to a lesser extent (see Dehwah et al.
2015). The peak concentration of biopolymers was found in the top 10m layer at the three
sites in the Red Sea, corresponding to the highest concentrations of bacteria in the water
column. Particulate and colloidal TEP concentrations, while quite variable, showed a spike at
around 40 m water depth for two of the sites, reflecting the increase in algal concentrations
found 10 m deeper at 50 m water depth.
Consequently, Dehwah et al. (2015) concluded that a deep-water intake conferred no clear
improvement in water quality compared to a shallow intake. Since measurements in
Dehwah‘s study were not made during an algal bloom, algal counts and biopolymer
concentration may show a completely different pattern with depth for different algal species
during bloom events. In conclusion, there is no simple generalization favoring deep versus
shallow intakes in the context of HABs. This issue needs to be determined through sustained,
site-specific monitoring prior or during the design phase (Chapters 3 and 5). Furthermore, as
with all desalination projects, other factors may drive the intake type, location and depth (see
Sections 6.4.3 and 6.5). In the case of the Dehwah et al. (2015) study, a deep-water intake
was deemed not feasible due to other factors namely the construction and operational risks
for a deep water intake in that area.
6.4 SUBSURFACE INTAKE OPTIONS
Subsurface intake systems can be used to improve feedwater quality for SWRO desalination
plants (Missimer 2009; Missimer et al. 2013; Rachman et al. 2015; Dehwah et al. 2015;
Dehwah and Missimer 2016). There are several different types of subsurface intakes that can
be designed and constructed depending on the local hydrogeology at a given site. These
intake types can be subdivided into two categories, wells and galleries.
181
Well intakes include the following:

• conventional vertical wells (screen or open-hole completions);
• collector or Ranney wells;
• angle or slant wells; and
• horizontal wells (conventional utility type horizontal directional drilled (HDD)
systems or Neodren™ systems).
Gallery types include:

• beach galleries: and
• offshore or seabed galleries.
Detailed design methods and examples of subsurface intake utilization can be found in
Missimer (2009) and Missimer et al. (2013; 2015b). This includes approaches to borehole
completion, screen design, exploration and testing, and general use criteria.
Historically, subsurface intake systems have been employed by small- to medium-size
SWRO plants with capacities typically less than 15,000 m3/d. There are, however, several
new plants that are using subsurface intake systems that have higher capacities, and many
new plants are considering the use of subsurface intake systems. In fact, in the State of
California, where many SWRO projects are being investigated, a regulatory policy requires
SWRO plants to use subsurface intake systems unless they can prove that any potential
subsurface intake type is not technically feasible as described in Missimer (2015a) (revised
California Ocean Plan).
Most subsurface intakes function in a similar manner to river bank filtration used in drinking
treatment schemes in Europe and the USA, and dune infiltration practiced in the Netherlands.
Such filtration systems use the natural geological properties of sediments and rocks to strain
and/or biologically treat the raw water to remove organic matter, suspended solids and
dissolved organic matter. With the improvement in raw water quality, pretreatment
complexity and operational effort can be reduced. Almost all of the SWRO systems that use
surface intake systems utilize conventional pretreatment systems or membrane treatment as
shown in Schemes A and B of Figure 6.13, respectively, incorporating dissolved air flotation
in areas prone to algal blooms (Scheme C). Yet, despite extensive pretreatment, plants may
still encounter biofouling of the membranes. Ideally pretreatment could be reduced to fine
filtration and/or simply cartridge filtration with chemical addition limited to acid or anti-
scalant (Scheme D) for a well operated subsurface intake system. There is a significant
reduction in operational cost accompanying the use of this option, especially when the
primary process can be bypassed.
Indeed, there are a number of small to medium SWRO desalination plants in the Caribbean
and Malta which require only minimal pretreatment (bag and/or sand filters; WateReuse
2011). The majority of existing SWRO desalination plants using subsurface intakes however,
have an additional filtration step prior to SWRO (e.g. the Sur Plant in Oman and the
Uminonakamicchi Nata Seawater Desalination3 Plant in Japan).
Subsurface intakes are expected to attenuate feedwater quality during poor water quality
events in the source seawater. Recent studies (Rachman et al. 2014; 2015; Dehwah et al.
2015; Dehwah and Missimer 2016; Dehwah et al. 2016) have investigated the performance of
3
Commonly referred to as the Fukuoka Desalination Plant in the desalination industry
182
Figure 6.13. SWRO pretreatment schemes for conventional surface-water intake systems with the
goal of using alternative “d” with a subsurface intake system.
subsurface intake systems for the removal of natural particulate and dissolved organic matter,
such as algae, bacteria, various fractions of NOM and TEP and found them to successfully
remove these completely or a large degree, as discussed further in the following sections.
This is very important to SWRO plants that are located in regions subject to periodic bloom
events. SWRO plants have a history of operational problems or shut-downs during severe
blooms (Berktay 2011). Subsurface intake systems can continue to operate during algal
blooms depending on the intake type and duration of the event. While no literature has been
published for the operation of subsurface intake systems during major algal blooms, well
intakes located in area with frequent blooms have been operated during events with no
reported shutdown such as the Sur plant in Oman.
Subsurface intakes can be used to provide feedwater for virtually any capacity SWRO system
with significant savings achieved in terms of operating costs, which can be reduced by 5 to
35% (Missimer et al. 2013).
Potential cost reductions include:
• Lower capital costs for pretreatment processes due to improved raw water quality;
• Reduction in permitting costs, especially the investigation of impingement and
entrainment impacts;
• Elimination of chlorine and coagulant usage translating to a reduction in operating
costs;
• Reduction in costs associated with waste disposal e.g. elimination of marine debris
disposal from traveling screens and reduction in sludge generated during coagulation;
and
• Continuity of supply to meet contract targets.
183
The capital cost of these intakes can, however, be quite high and the construction complexity
can require extensive time periods to complete. In addition, subsurface intake systems require
considerable planning prior to development of tender documents in order to reduce contractor
bidding risk. Therefore, a full life-cycle cost analysis should be used to assess potential
reductions in the cost of water to consumers before a large-capacity subsurface intake system
is designed and constructed.
6.4.1 Description of intake types with example installations
Subsurface intakes are subdivided primarily into wells and galleries of varying design. There
are some new hybrid designs that also use the groundwater system as a primary filter. Each
of these intake types is described and an operating example is provided along with results
from research. There is no good example of a beach gallery system of large size, but several
small-capacity examples of a similar design are used in the Caribbean.
6.4.1.1 Conventional vertical wells
The most common subsurface intake used to supply feedwater to SWRO systems is the
conventional vertical well system. Wells are constructed as close to the shoreline as possible
to allow raw seawater to infiltrate through the
seabed into the aquifer with flow into the
pumped well (Figure 6.14). Well design and
capacity are based on the local hydrogeology.
Detailed design concepts, such as screen slot
size, are based on the specific size distribution
of the sediment and is covered in several
chapters of Missimer (2009).
Use of conventional wells is limited to small-
Figure 6.14. Conventional vertical wells located
to medium-capacity SWRO systems unless a
close to the shoreline. The produced water must
coastal aquifer containing a very high
come predominantly from the sea and not the
permeability can be used. Since vertical wells
landward direction. Figure: Missimer et al. (2013).
must be located very close to the surf zone on
the beach, they may not be a practical intake solution in heavily populated areas because of
the visual impacts on coastline or in areas where beaches are eroding; however, well systems
located near the shoreline will not clog during algal bloom events due to the self-cleaning
nature of the littoral zone wherein breaking waves move filtered debris laterally along the
shoreline.
Currently, the largest capacity vertical well intake system in the world feeding a SWRO plant
(80,200 m3 desalinated water/d) is located in Sur, Oman. The wellfield system, located in a
highly permeable aquifer, has a design capacity of roughly 160,000 m3/d, produced from 32
wells split into three clusters, with 5 Ha dedicated to beachwells (Craig 2012). The Sur
wellfield system has performed well and no clogging has been observed despite seasonally
high concentrations of algae. Water quality produced by the well system is excellent – high
SDI3 up to 27 in the source seawater have been reduced to SDI15 of 1.4 and is consistently
around 1. These low SDI15 results show that the intake indeed functions as pretreatment for
SWRO, as after filtration through the aquifer the raw water could be directly fed to the
SWRO system as it meets most RO manufacturer guarantee requirements for SDI15.
Nonetheless, there is pretreatment at the Sur plant, albeit limited to mono media (sand)
filtration with no coagulant addition.
Rachman et al. (2014, 2015) compared the removal of algae, bacteria and NOM at the Sur
SWRO plant and at three other plants operating with conventional well intake systems;
184
Jeddah in Saudi Arabia, Turks and Caicos in the Providenciales and Alicante in Spain. These
sites were selected to represent different geographic regions and geologies; the Gulf of Oman,
the Red Sea, Caribbean Sea, and Mediterranean.
Algae were almost completely eliminated in all vertical wells; this included small
picoplankton species such as Synechococcus (see section 6.3.3) irrespective of the
concentration in the associated seawater source. The Oman site in particular showed high
concentrations of this species, more than 80% higher than the other three sites ranging from
113,000 to 194,310 cells/mL on the two sampling dates. In addition, over 90% and up to 99%
of the bacterial population was removed by the subsurface intakes.
Rachman et al. (2014) characterized the NOM in the raw seawater at these sites and
following beach well abstraction by LC-OCD into the five fractions: biopolymers, humic
acids, building blocks, low molecular weight acids and low molecular weight neutrals. The
concentration of particulate TEP was also determined. During algal bloom events, the
concentration of biopolymers can peak. Sampling was not reported to coincide with a bloom
in the study and indeed the concentration of biopolymers was relatively low. Instead, humic
acid was found to be the major fraction in seawater at all sites. Removal of the NOM
fractions was found to be selective with the highest removal observed for the larger
molecular weight biopolymer fraction with complete to near complete removal at all sites.
Substantial removal of the smaller humic acid fraction (>50%) occurred, followed by
building blocks, and the light molecular weight organics. The particulate TEP removal rate
was, however, variable ranging from 34% up to 92% for the different vertical wells.
In another study, the performance of a beach well in the West Mediterranean was compared
to conventional and membrane SWRO pretreatment from other locations in the
Mediterranean and North Sea water (see Table 6.1). LC-OCD and the Modified Fouling
Index using UF membranes (MFI-UF) were employed to assess the reduction in NOM and
the particulate fouling potential, respectively, by beach well filtration and the various
pretreatment processes (Salinas-Rodríguez 2011; Lattemann et al. 2012). Both the MFI-UF
and the earlier MFI-0.45 test, using larger pore size (0.45 µm) microfiltration membranes,
were developed to measure the particulate fouling potential of RO feedwater (Chapter 5). The
MFI-0.45 was applied along with AOC to monitor the clogging potential of pretreated river
water to be infiltrated in artificial recharge wells due to the deposition of particles and
biogrowth, respectively (Schippers 1995). MFI-0.45 and AOC values below a threshold value
were expected to prevent clogging. In practice however, these parameters could not reliably
predict the clogging rate of recharge wells and low values did not preclude clogging
(Schippers 1995). The MFI-UF has not been trialed for predicting clogging of infiltration
wells.
LC-OCD results for the beach well showed that humic acid accounted for the major fraction
of NOM in the West Mediterranean seawater with biopolymers constituting only 10% of the
NOM. As with Rachman et al. (2014) the highest removal after passage through the seabed
was found for the larger molecular weight biopolymer fraction (70%) followed by building
blocks, neutrals, and finally humic acid, with only 9% removal (Salinas-Rodríguez 2011;
Lattemann et al. 2012). The performance of the beach well was superior to that of
conventional and membrane pretreatment in terms of removal of biopolymers where removal
was variable and ranged from 15% to 51% (see Table 6.1).
For the beach well, the MFI-UF was measured in constant pressure mode using both 30 and
10 kDa molecular weight cut off (MWCO) test membranes thus allowing smaller particles to
be captured than in the SDI and MFI-0.45 tests. This is especially so when using the smaller
10 kDa membrane in the MFI-UF test, which has shown a high correlation with the
185
concentration of smaller TEP (TEP10kDa) measured with a 10 kDa test membrane (Chapter 5).
Similar to the removal of organics, the beach well appears to be more efficient in removing
larger particles, as the removal efficiency was 15% higher for the larger 30 kDa MWCO test
membrane than for the 10 kDa membrane (35% removal). Moreover, the MFI-UF10kDa
measured in the beach well discharge was still high (7,300 s/L2), suggesting that particles,
including smaller TEP and TEP precursors, may remain in the seawater and thus could cause
downstream fouling.
The beach well achieved a higher removal of biopolymers than conventional and membrane
pretreatment. In reducing the particulate fouling potential, the beach well achieved a similar
reduction to coagulation and dual media filtration (MFI-UF10kDa removal of 40%). These
results are only indicative however, as the MFI-UF for the beach well was measured in
constant pressure mode while the pretreatment options were measured under constant flux
mode and therefore may not be directly comparable. A comparison with MFI-UF30kDa cannot
be made as it was not measured for all pretreatment processes.
Table 6.1. Performance of a beach well intake compared to SWRO plant pretreatment
processes in removing biopolymer and humic acid fractions of NOM, and reducing the
particulate fouling potential measured by the MFI-UF10kDa test (data from Salinas-Rodríguez
2011 and Lattemann 2012).
Humic
Biopolymer
Pretreatment/ acid MFI-UF10kDa2
Seawater removal
Intake removal (%)
(%)
(%)
West Beach well 70 9 35
Mediterranean
North In line coagulation 47 30 40

Mediterranean (ferric +polymer),
dual media filtration
East Coagulation, mono 32 6 52

Mediterranean media filtration
North West Ultrafiltration 15 1 68

Mediterranean 1
(outside-in)
North Sea Water Coagulation 51 1 88

(polyaluminum
chloride)
Ultrafiltration
(inside –out)
1
Outside in submerged membranes with no coagulation
2
Although, all MFI-UF were normalized to standard reference values of temperature, pressure and area
(Chapter 5), the MFI-UF for the beach well was measured in constant pressure mode (2 bar) while the
pretreatment options were measured under constant flux mode (250 L/m2h) and may not be directly comparable.
Results are therefore indicative of MFI-UF removal efficiency when comparing performance of the beach well
to other pretreatment options.
186
As expected, ultrafiltration with or without coagulation yielded the largest reduction in

particulate fouling potential (68 to 88%) due to smaller membrane pore sizes than the
interstices in dual media filtration or the seabed strata. The results presented in Lattemann et
al. (2012) while promising in
terms of biopolymer removal by
beach well filtration, are limited
and indicative only for the MFI-
UF results and warrant further
investigation.
The aforementioned studies show
that vertical wells reduce the
biopolymer fraction of NOM
(which includes TEP that can
Figure 6.15. Comparison of bacterial concentrations in promote biofouling of
surface seawater and well discharges for a SWRO plant with a membranes), in addition to
groundwater flow path averaging about 100 m. Figure:
Rachman et al. 2014. bacteria, algae, and particulate
matter. A typical removal
percentage for well intakes is
100% for algae and over 90% for
bacteria (Figure 6.15). Biopolymer
removal ranged from 70% (Table
6.1) up to nearly 100% in some
cases (Figure 6.16).
The degree of treatment provided
within a well intake system is
based upon a number of factors
including the flow path length
time, the type of geological media,
Figure 6.16. Comparison of NOM fraction for surface
seawater and well discharges for a SWRO plant in Jeddah in
the hydraulic retention time, the
Saudi Arabia with a groundwater flow path of about 200m. biochemical activity in the aquifer,
Note that the reduction of the biopolymer fraction, which and the local composition of the
contain sticky polysaccharides and proteins is nearly 100% at seawater. Rachman et al. (2014)
this location. Modified from Dehwah and Missimer 2016. found that the geological
characteristics of the site and
aquifer type (the Oman and the Turks and Caicos aquifers are limestone, while the Jeddah
system is siliciclastic) did not have a direct correlation to the removal rate of the organic
substances. Instead the flow path length and hydraulic retention time had a greater impact on
organic matter removal efficiency compared to the geology of the aquifer or specifically,
lithology. Therefore, a careful balance must be achieved wherein the wells are located close
enough to the shoreline to have most of the recharge from the sea, but sufficiently far away to
remove a significant percentage of the organic matter.
6.4.1.2 Collector or Ranney Wells
A collector well (Ranney well) is a specialized well type with a high unit production capacity
compared to most wells. It contains a central caisson with a diameter ranging from ~ 2 to 4 m
and a series of horizontal laterals to collect water from the penetrated aquifer (Figure 6.17).
Collector wells are commonly used to tap gravel units within aquifers underlying river
systems in the Midwest region of the United States and other geographic areas. These wells
can have capacities of over 51,400 m3/d (Missimer 1997; Missimer et al. 2013. There are a
187
few examples of the use of collector

wells for SWRO intakes; the largest
capacity system being the PEMEX
Salina Cruz refinery in Mexico which
has three wells with a capacity of
15,000 m3/d each (Voutchkov 2005).
6.4.1.3 Angle wells
Angle or slant well design and
construction is relatively new and is
being applied to SWRO plants under
design in California (Williams 2015).
An angle well is drilled from a
location on the beach at an angle so
that the screen section of the well is
located fully beneath the seabed and
seaward of the freshwater/seawater
interface (Figure 6.18). While no
large-scale slant well intake system is
currently operational, a detailed test
program over an extended period of
time (21 months) has been completed
Figure 6.17. Ranney or collector well used to obtain raw
for the Dana Point SWRO plant at
water from a fractured rock aquifer hydraulic connected to Doheny Beach in California
the sea. Images: Missimer 2009. (MWDOC 2014). The full-scale wells
are expected to have an installed
capacity of 113,600 m3/d (Williams
2015). The filtered water was reported
to have a very low SDI and turbidity
over the test period (MWDOC 2014).
However, the slant well began to draw
anoxic water enriched in dissolved iron
and manganese which may be
challenging for SWRO operation.
Aeration of the water during Figure 6.18. Angle well-constructed beneath the seabed
pretreatment could lead to oxidation of and seaward of the freshwater/seawater interface. Figure:
the iron and manganese into iron Missimer et al. 2013.
hydroxide and manganese dioxide
which may result in fouling of the RO membranes if not removed.
6.4.1.4 Horizontal wells (HDD)
Horizontal drilling and micro-tunneling, used to install pipes into the ground with minimal
surface disruption, are mature technologies employed in the utility field for over 50 years.
Horizontal wells have been designed and constructed in the petroleum field for many years
and have also have been used in remediation of groundwater contamination (Delhomme et al.
2005). These early wells have been constructed using conventional drilling technologies and
are completed with screens or screens covered with a geofabric.
188
A newer development of the horizontal technology involves the use of Neodren™ systems
(Peters et al. 2007). Construction of this
system occurs using a horizontal drilled
hole that emerges on the seafloor
(Figure 6.19). The well casing with the
attached, patented screened assembly is
then pulled back through the mudded
borehole and set in place. There are
several operating examples, the largest
of which is Alicante, Spain (Peters et al.
2007). Unfortunately, there is incorrect
capacity data contained in the literature
regarding the Alicante system which
was corrected in Rachman et al. (2014).
The current capacity is difficult to
assess based upon the combined use of
horizontal wells and a water tunnel. It
was reported to have a capacity of about
65,000 m3/d which is now likely to be
about 25,000 m3/d or less.
Figure 6.19. Horizontal well drilled beneath the seabed
(A). Note that many wells can be constructed from a Rachman et al. (2104) investigated the
common pad which saves considerable effort and cost inremoval of NOM, algae and bacteria by
siting them (B). Figures: Missimer et al. 2013. the NeodrenTM system in Alicante and
found a breakthrough of algae, low
removal of bacteria (41%) coupled to a higher concentration of the NOM fractions as
compared to the source seawater. The poorer-than-expected performance of the NeodrenTM
system may be a result of the direct inflow of seawater into one or more of the drains or into
vertical karst conduits that connect the natural seawater to the drain screens (Rachman et al.
2014).
Large-scale application of horizontal well technology has not yet been developed. While the
concept is attractive, there are issues with regard to methods that would have to be employed
to maintain the screens. Cleaning and repacking of the gravel could be very difficult based on
the distance from the shoreline. The operation of HDD systems have not been documented
during algal blooms.
6.4.1.5 Beach gallery systems
Gallery intakes use the concept of slow sand filtration by creation of an engineered filter that
can be located on the beach, near or
above the high tide line, within the
intertidal zone of the beach, or in the
seabed.
Beach gallery intake systems involve
the placement of an engineered filter
beneath the littoral zone of a natural
beach (Figure 6.20). Unlike slow sand
filters which operate under gravity,
beach gallery filters are pumped using
Figure 6.20. Beach gallery intake system directly beneath a series of collector screens underlying
the intertidal or surf zone. Figure: Missimer et al. 2013. the gravel and sand to abstract
189
seawater filtrate through the filter. Yet, they act like slow sand filters due to their low
filtration rates (Maliva and Missimer 2010). Slow sand filtration improves seawater quality
by removing particulate matter by straining and organic matter by biological treatment. With
low filtration rates and corresponding higher retention time in the filter, the assimilation of
organic compounds tends to improve. The advantage of the beach gallery system is that the
wave action occurring above the filter tends to keep it clean as particulate matter is removed
within the upper sand part of the filter which means the system is essentially self-cleaning.
This type of intake can be used on a moderate energy beach with typical wave heights being
between 0.5 and 1 m. A beach gallery intake system was recently explored for technical
feasibility at Huntington Beach, California. It was found not to be technically feasible due to
extreme rates of beach erosion and great complexity in construction. The construction in this
high-energy environment would require use of a tram system and would take 5 to 8 years to
complete based upon seasonal prohibitions on beach construction (Bittner et al. (2015).
At present, there are no large-capacity operating examples of a beach gallery intake system;
however, a trench intake system was operating in a similar manner on Useppa Island, Florida
for more than 20 years for supply to a SWRO plant with a capacity of approximately
400 m3/d. It is uncertain whether the trench or gallery system is still being used.
6.4.1.6 Offshore or seabed gallery systems
The seabed gallery or infiltration gallery is another intake type that can be used to supply
nearly any capacity desired (Missimer 2009). It consists of an engineered filter constructed in
the seabed offshore. In concept, it is
similar to a beach gallery and operates
similarly to a slow sand filter, but it
requires pumping and cannot operate
under a gravity condition.
The largest capacity seabed gallery
system operating today is located at
Fukouka, Japan and supplies the
Uminonakamicchi Nata Seawater
Desalination Plant (Figure 6.21). It has
a capacity of 103,000 m3/d and has
Figure 6.21. Seabed gallery system as constructed at operated with minimal maintenance for
Fukuoka, Japan. Image: Fukuoka District Waterworks over nearly 10 years achieving an SDI15
Agency, 2015. consistently below 2.5 in the filtered
water (Shimokawa 2012; Figure 6.22).
To investigate performance of a seabed
gallery, the Long Beach Water
Department designed and operated a
demonstration surf zone infiltration
gallery referred to as under ocean floor
intake system. The gallery comprised an
excavated pit filled with engineered
sand. Figure 6.23 shows filling of the
Figure 6.22. SDI15 data collected from the seabed gallery gallery using a temporary cofferdam
intake at the Fukuoka SWRO plant. Figure: Missimer et al. and the filtration area of the infiltration
2013.
gallery. Filtered water was collected
through a series of perforated laterals
190
Figure 6.23. Filtration area of under-ocean floor seawater intake (left hand side) and filling of infiltration
gallery with engineered sand during construction using a temporary cofferdam (right hand side) at Long
Beach (Zhang et al. 2011). Reprinted with permission from American Water Works Association.
(15 cm diameter V-wires with 0.13 cm slot openings) along the pit’s bottom (Allen et al.
2011). Infiltration rates varied between 2.9 and 8.8 m/d (Allen et al. 2009; Zhang et al. 2011).
The seabed gallery filtrate then passed through 100 µm or 5 µm cartridge filters. Three algal
bloom events occurred during operation of the seabed gallery that increased total organic
carbon (measurements available for only
two of the events) and turbidity in the
source seawater (Zhang et al. 2011) as
shown in Figs 6.24 and 6.25, respectively.
While infiltration through the seabed
gallery appeared to attenuate total organic
matter organic carbon (TOC) during these
events, turbidity was not always reduced.
In the previously mentioned study of
Salinas-Rodríguez (2011), NOM fractions
in the source seawater and filtrate from the
Figure 6.24. TOC in raw seawater and filtrate from the Long Beach demonstration gallery were
Long Beach under-ocean floor seawater intake (Zhang determined by LC-OCD. Although
et al. 2011). Reprinted with permission from American
Water Works Association.
sampling was not reported to coincide
with an algal bloom event, Salinas-
Rodríguez (2011) found significant
removal of biopolymers from the source
water (75%) by the seabed gallery and
close to 20% removal of humic acids
(19%). Cartridge filtration removed
biopolymers by a further 13%. This may
be due to a combination of adsorption onto
the cartridge filters, as biopolymers are
noted for being very sticky, followed by
degradation by bacteria in the cartridge
filters as the filtrate to the cartridge filters
was not chlorinated. Indeed, the lifetime of
the downstream cartridge filters was
Figure 6.25. Turbidity in raw seawater and filtrate from reported to be reduced to a week using the
the Long Beach under ocean floor seawater intake (Zhang seabed filtrate due to biofouling of the
et al. 2011). Reprinted with permission from American
Water Works Association.
cartridge filters (Carollo 2016). Iron and
191
manganese fouling of the cartridge filters also occurred (Zhang et al. 2012). Water quality
testing over time at the demonstration facility showed that it would not consistently meet
typical SDI15 and turbidity membrane guarantee requirements without further pretreatment
(Carollo 2016). This may be attributable to the fact that the sides of the seabed filter were not
sealed (T. Tseng, personal communication, 2017). Therefore, the engineered sand interfaced
directly with the native beach sand which allows native pore water, sediment, dissolved iron
and manganese to enter the gallery. Infiltration into the filter was indeed observed from the
sides as well as the top and bottom, albeit at different rates (T. Tseng, personal
communication, 2017). Sealing the sides and bottom of the filter would prevent this as then
the only water entering the system would have been seawater from the top where typically
the concentrations of iron and manganese in well oxygenated seawater are low.
A considerable amount of new research has been conducted on design and construction of
seabed galleries (Missimer et al. 2015b; Dehwah and Missimer 2017). Recent research
investigated the effectiveness of the active layer of a gallery intake system in improving
seawater quality in long term bench scale column experiments for two different media
(Dehwah and Missimer 2017). Silica and carbonate sand were tested in 1 m columns to
evaluate the removal of algae, bacteria, NOM and TEP over 620 days. The infiltration rate
was fixed at 5 m/d. The columns required several months to reach an equilibrium state, after
which there was a significant improvement in seawater quality. Nearly all of the algae, 87%
of the bacteria, 59% of the biopolymers, 57% of the particulate TEP and 32% of the colloidal
TEP were removed in the silica sand column in the last 330 days of operation. Within the
carbonate sand column in the same time period, removal of biopolymers, particulate and
colloidal TEP was higher at 75%, 66%, and 36% respectively but removal of bacteria was
lower at 74%. Although the bench scale test simulated a type of slow sand filtration, it was
found that a “schmutzdecke”4 did not form at the column surface and the columns did not
clog internally which is attributed to biochemical degradation of the organic materials.
6.4.1.7 New subsurface intake designs
Recently, new types of subsurface intake systems have been designed and constructed to
achieve a degree of pretreatment. Two systems, the water tunnel and the karst pit, are
interesting examples of hybrid subsurface intake systems.
The water tunnel system consists of a
horizontal tunnel that ranges from 2 to 4 m
in diameter and contains a series of vertical
collector screens that protrude into the roof
of the tunnel (Figure 6.26). This general
concept was originally developed for a
freshwater collection system in Louisville
wherein it was used to obtain water from a
shallow gravel unit lying beneath a river
(Missimer 2009). An example of the system
in in place at Alicante, Spain where it is
used as part of the intake system for a
SWRO plant and has a capacity of
Figure 6.26. Tunnel intake example. Photo: RBF 50,000 m3/d (Rachman et al. 2014; 2015).
Consulting, 2009. The tunnel system as described by Rachman
et al. (2014) comprises a 3.14 m diameter
4
thin biologically active top layer in slow sand filtration
192
tunnel, 1 km in length, oriented parallel to the shoreline located 14 m below sea level. The
water intake into the tunnel is provided by 104 pipe laterals, constructed perpendicular to
tunnel axis. The performance of the tunnel in removing particulate matter, algae, bacteria,
and NOM was compared to conventional vertical wells, including one at the same site in
Alicante in the study by Rachman et al. (2015: see discussion in Section 6.3.1.1). As with the
vertical wells, the tunnel system was found to be effective in virtually removing all algal cells.
The tunnel system also removed 71% of the bacteria and a significant amount of the organic
fractions of NOM (e.g. 90% of biopolymers) and TEP (84% and 55% removal of particulate
and colloidal TEP, respectively). Overall, the vertical wells gave higher removal of NOM.
The lower removal by the tunnel system is most likely caused by a shorter flowpath from the
seabed to the tunnel which provides less hydraulic retention time for organic carbon removal
(Rachman et al. 2015). The transport distance of seawater to the tunnel intake system was
much shorter compared to the transport distance from the sea to the vertical wells.
The karst pit concept was described by Pankratz (2015) for an intake system used in Curaçao
(Figure 6.27). The intake consists of a
surface excavation located about 100 m
from the shoreline. Feedwater is pumped
from the excavation which forces seawater
to filter through the limestone from the sea
into the excavated area. The walls of the
excavation contain prefabricated concrete
that contain perforations. The intake has a
capacity of 52,000 m3/d to meet the
requirements of a 26,000 m3/d SWRO plant.
6.4.2 Subsurface intake performance
for algae and NOM removal
In general, subsurface intake systems are
very effective at removing algae, a
considerable percentage of bacteria, and
Figure 6.27. Karst pit intake system. Photo: T. some percentage of the biopolymer fraction
Pankratz. of NOM, particulate and colloidal TEP
(Rachman et al. 2014, 2015; Dehwah et al.
2015, 2016; Dehweh and Missimer 2016; 2017). In the Rachman et al. (2014) study,
investigating the performance of conventional vertical wells, the tunnel and horizontal well
intake systems at Alicante, all of the subsurface intake types tested were effective with the
exception of the horizontal well intake system at Alicante, where there was breakthrough of
algae into some of the wells. The well intake systems showed 100% removal of algae during
transport through various types of coastal aquifers from the sea to the wells. A comparison of
three intake types at Alicante, Spain, showed that the vertical well system removed greater
amounts of bacteria, NOM, organic fractions, and TEP in comparison to the tunnel and
horizontal drain systems. The aquifer feeding the wells and tunnel did remove all of the algae,
but only 70% of the bacteria were removed in flow to the tunnel, and some biopolymers
entered the intake as 10% were not removed (Rachman et al. 2014).
The well system located at Sur, Oman, was particularly effective at removing algae. This site
is significant in that the Sea of Arabia has a high frequency of algal blooms which have
spilled into the Gulf with adverse operational impacts that have affected several SWRO
plants (Berktay 2011). During the operational period when the well intakes were being used,
it is likely that some type of algal bloom occurred; however, there is no specific
193
documentation concerning the algal concentrations occurring and any changes in the algae
occurrence within the production wells during a bloom. Based on the distances of the wells
from the sea ranging from 30 to 250 m and the flow path length through the aquifer, it is
highly unlikely that algae could actually enter the production wells, even during a bloom. In
addition, the removal of algae, biopolymers and TEP should reduce the biofouling potential
of a feedwater during a bloom event.
A considerable amount of research has been completed recently on the development of
gallery intake systems that can be used to produce a sufficient supply of feedwater even for
the largest SWRO plants (Missimer et al. 2015a; Dehwah and Missimer 2017). The seabed
gallery, in particular, appears to have the greatest potential number of applications. This
system operates similarly to a slow sand filter in which the seawater is filtered through an
engineering system with a hydraulic retention time of 6 to 8 hours. Experiments conducted
on the use of a slow sand filter as pretreatment for SWRO plants showed no penetration of
algae through the system and a very high rate of kainic acid removal which was used as an
algal toxin surrogate (Desormeaux et al. 2009). This research can be used as a proxy for the
operation of a seabed gallery system and will likely function in the same manner, although no
seabed gallery system has been operated throughout a major, high-biomass algal bloom, so
the need for additional on-land pretreatment of the gallery effluent is still open to question.
The processes of particulate and dissolved organics removal are similar to those occurring
during slow sand filtration with the exception that no biologically active “schmutzdecke”
layer forms during filtration at the sediment-water interface i.e. on the top surface of the filter
(Dehwah and Missimer 2017). Instead, all of the particulate organics and suspended
sediments are strained during transport and accumulate within the aquifer or upper 1 m of a
seabed filter. No reported clogging has been reported despite operation of wells and seabed
filters for up to 30 years (in the case of wells). Therefore, within groundwater systems, it is
assumed that biochemical processes are active in reducing particulate organics into dissolved
forms that move through the aquifer. This is suggested in some recently collected data by
increases in the concentration of light molecular weight neutrals in the well discharges
(Dehwah and Missimer 2017). Within seabed filters, the upper layer of the filter undergoes
bioturbation whereby organisms that derive nutrition from the sediments such as polychaete
worms and some molluscs assimilate organic material and small particulates, leaving beneath
rigid fecal pellets that act hydraulically similar to sand grains. The deposit feeders act to
prevent the building of a biological clogging layer at the sediment–water interface. Removal
of biopolymers and other fractions of NOM are likely caused by a variety of physical
(straining and adsorption) and biochemical breakdown processes (bacterial and geochemical).
Additional research is being conducted on these mechanisms in column and larger scale
experiments.
6.4.3 Planning of desalination plants with subsurface intakes
While subsurface intakes offer many advantages for SWRO plants in areas prone to blooms
they are not always feasible. Use of a particular type of subsurface intake is dependent on the
hydrogeologic and marine conditions of a specific site. In addition, local infrastructure also
plays an important role in the choice of intake type that can be used (e.g., availability of
specialized construction equipment, electric power availability to pipe and pump the raw
seawater). Since the use of subsurface intakes is related to the site-specific conditions, the
specification of an intake type requires planning and a certain level of pre-tender field
investigation.
It is prudent to perform feasibility level investigations of the coastal area in regions that are
planning use of a subsurface intake system to supply SWRO plants. Dehwah et al. (2014)
194
developed a coastal geomorphological mapping system that can be used for screening
purposes to assess which types of intakes can be used. The method considers both field
conditions and the capacity of the plant being considered for the Red Sea shoreline of Saudi
Arabia. The factors considered and areas mapped are shown in Table 6.2. The relationship
between the mapped coastline and the feasibility of using a specific intake type are given in
Table 6.3.
Table 6.2. Geomorphological classifications of the Red Sea coastline (Dehwah et al. 2014).
A. Sandy Beaches
A1 - Sandy beach with corresponding nearshore sand or slightly muddy sand, coral
reef complex offshore
A2 - Sandy beaches, restricted, with no reef
A3 - Offshore island with nearshore sandy sediments and reef
B. Rocky shorelines
B1 - Limestone rocky shoreline with corresponding nearshore sand, and offshore
coral reef complex
B2 - Limestone rocky shoreline with nearshore muddy sediments
B3 - Limestone rocky shoreline, nearshore deep water, no reef
B4 - Rocky headland with offshore rocky bottom, no reef
B5 - Rocky shoreline, wadi1 sediments nearshore, offshore reef
C. Wadi 1intersections
C1 - Wadi sediments (boulders, pebble, and gravel)at shoreline, variable sand, gravel
and mud offshore with no reef
C2 - Wadi shoreline sediments, nearshore marine hard ground, minor nearshore
sand , coral reef offshore
D. Sabkha2, lagoons, and mangrove

D1 - Coastal sabkha shoreline and nearshore muddy sediments
D2 - Muddy shoreline with lagoonal muddy sediments, nearshore sand and offshore
reef complex
D3 - Muddy shoreline /lagoon/ supra-tidal sabkha with no reef complex
D4 - Mangrove shoreline with nearshore muddy sediments
E. Others
E1 - Shoreline reef complex dropping to deep water in the nearshore off-reef area
E2 - Artificial channels or urban shoreline with artificially filled nearshore dropping
to deep water nearshore
E3 - Natural channel
1
Wadi- an ephemeral stream that flows only during flood conditions in arid regions
2
Sabkha -a supra-tidal to intertidal area wherein seawater is trapped during storms or high-tide events and the trapped water
evaporates to produce hypersaline conditions (salinity often over 250,000 mg/L), commonly with the precipitation of evaporite
minerals occurring on the Sabhka plain.
195
Table 6.3. Correlation between coastal environment and feasibility1 of using various
subsurface intakes along the Red Sea coastline (from Dehwah et al. 2014).
Intake Type Subsurface Intake System
Well / Gallery Well System Gallery system
Radial Beach Seabed
Environments Vertical Horizontal Angle
(collector) Gallery Gallery
A. Sandy Beaches
A1 1(b)2 3 2(b) 2(b) 1(d) 1(d)
A2 1(a) 3 2(b) 2(a) 4 1(c)
A3 1(a) 3 2(b) 2(b) 1(d) 1(d)
B. Rocky shorelines
B1 1(b) 3 1(b) 1(c) 1(c) 1(d)
B2 4 4 4 4 4 2(c)
B3 4 4 4 3 4 4
B4 4 4 4 4 4 4
B5 1(a) 3 2(b) 2(a) 2(c) 2(c)
C. Wadi intersections
C1 4 4 4 4 4 4
C2 1(b) 3 2(c) 2(b) 2(c) 2(c)
D. Sabkha, lagoons,
and mangrove
D1 4 4 4 4 4 4
D2 4 4 4 4 4 4
D3 4 4 4 4 4 4
D4 4 4 4 4 4 4
E. Others
E1 4 4 4 4 4 4
E2 4 4 4 4 4 4
E3 4 4 4 4 4 4
1
Feasibility factor: 1=Excellent, 2=Possible 3=Questionable, 4=Not feasible
2
Estimated Capacity (m3/d): a. Capacity <20,000, b. 20,000-50,000, c. 50,000-100,000, d. Any capacity
Mapping of the Red Sea coastline of Saudi Arabia in the study of Dehwah et al. (2013)
showed that the most favorable environments for use of subsurface intakes are;
• sandy beaches containing a low percentage of mud;
• limestone rocky shorelines with corresponding nearshore sand; and
• wadi sediments with low mud content.
Seabed galleries were found to be the preferred subsurface intake type for large-capacity
desalination plants based on the geology. Conventional wells or horizontal wells could be
used at shorelines containing limestone cliffs and reefs, but the relatively small thickness of
these deposits is a limitation on potential system capacity. Nearshore or coastal wadi
sediments not associated with a channel can also be used to develop low-capacity well intake
systems. Construction of subsurface intakes in environments where there is a high mud
concentration in the sediments and no water circulation (sabkha, lagoons, and mangrove) is
not desirable due to the potential for clogging of the filter. The high organic content and high
evaporation rate produce additional unfavorable conditions. All of the restricted and
nearshore muddy shorelines or mangrove coasts are not feasible for development of
subsurface intakes (D1, D2, D3 and D4).
This type of method can be applied to any coastal region. Feasibility will require additional
field work and preliminary engineering design with an economic analysis.
196
6.5 SITING OF DESALINATION SEAWATER INTAKES
Larger desalination plants are being built to take advantage of the economies of scale with
corresponding increases in intake flow rates. In addition, desalination projects are
increasingly being considered and implemented in regions where desalination was not
previously considered e.g. Australia and the USA. As a result, the development of these
plants face increased scrutiny by Governmental Agencies and the community, such as in
California. Consequently, siting of a plant and accompanying marine infrastructure are
important components of feasibility studies if the site is not already fixed.
The importance of the intake to the success of a desalination plant, particularly SWRO plants,
is sometimes overlooked in the development of a desalination project. Their design and
construction may represent one of the major risks to budget and schedule during the delivery
phase, especially in fast track projects, while access to good water quality seawater will
facilitate subsequent operation of a SWRO plant (e.g., Huntington Beach plant subsurface
intake system would require 5 to 8 years to construct; Missimer et al. (2015b)). In areas prone
to algal blooms, careful selection of the intake location, depth, and type can play a role in
minimizing their impacts and is the first defense against algal blooms. Intake siting studies
are ideally coupled to investigations to characterize seawater quality which, in addition to
field water quality sampling, includes a review of historical records to identify the frequency;
severity and duration of poor water quality events such as algal blooms (see Chapter 5).
Source water quality is, however, only one of a myriad of factors that need to be considered
in siting the intake. Other factors may drive the site selection process such as plant capacity,
seabed geomorphology and ecology, local environmental and marine regulations to name but
a few.
To satisfy all factors associated with the development of a new seawater intake (and brine
outlet), various approaches, varying in complexity, have been developed to identify and
assess the suitability of candidate seawater intake locations. Evaluation criteria can be
developed for engineering, environmental, and social aspects in siting studies. Separate sets
of evaluation criteria can be developed for the desalination plant and product water
conveyance pipeline to those of the intake (and brine outlet) as these project components
have different engineering, environment and social criteria and objectives. This also allows
comparison of varying arrangements of land based sites with marine infrastructure options.
Generic examples of evaluation criteria are given in Table 6.4 and are often more detailed
depending on the project requirements and specific to the region.
As site selection is often a challenging process, use of a multi-disciplinary team is
recommended to cover all competencies required in siting and designing desalination marine
infrastructure and to develop the evaluation criteria. A specialist in HABs is recommended in
areas prone to algal blooms, as these individuals can provide guidance on the likely location,
frequency, and type of blooms in an area, as well as information on water circulation and
transport. Various approaches can then be adopted to assess the criteria and rank sites.
Performance ratings can be assigned to the evaluation criteria ranging from - fatal flaw,
poorly suitable, moderately suitable through to highly suitable. If algal blooms are a known
water quality concern, considerations to minimize the ingress of cells can be built into the
ratings e.g. water depth, expected bloom transport pathways, or distance from shipping
activities such as ballast water exchange, which can result in the introduction of HAB and
other nuisance species to an area.
197
Table 6.4. Generic examples of engineering –infrastructure, environmental and social

evaluation criteria for siting a seawater intake.
Evaluation Criteria Objective

Infrastructure Shipping Activities associated with shipping do not
pose a risk to intake structure or impact
water quality. Intake structure does not
create a navigation risk.
Interference with Location of intake does not adversely affect

infrastructure current infrastructure or preclude future
developments.
Constructability Seabed characteristics do not influence

construction.
Distance to shore Minimize pipe length and thereby cost.

Adequate depth to avoid surface algal
blooms and ensure water quality.
Water Quality Minimize pretreatment requirements and

operational costs with high quality intake
water.
Maintenance Minimize risk associated with marine

maintenance activities.
Dredging activities Dredging activities do not create a risk to

inlet structure or impact water quality. Inlet
structure does not restrict maintenance of
dredging channels.
Environmental Marine vegetation Destruction of marine vegetation is

minimized e.g. sea grass.
Habitat Minimize loss or degradation of habitat of

rare, vulnerable and/or endangered species.
Conservation areas Minimize impacts on conservation areas e.g.

marine parks, reefs.
Social Aquaculture Minimize aquaculture impacts on

commercial fishing, fish and shellfish
farming etc.
Recreation Minimize the impact on marine recreational

activities e.g. boating, swimming and
fishing.
198
For instance, locating the seawater intake within 150 m of a shipping lane may be deemed a
fatal flaw by the team. Similarly, geomorphological environments that preclude the
development of a subsurface intake and only allow the option of a surface intake e.g. rocky
shorelines could be rated as poorly suitable (see geomorphological mapping system in
Section 6.4.3). Overall, performance ratings allow a comparative assessment of sites.
More sophisticated approaches may use a Multi Criteria Analysis (MCA) approach, whereby
the relative importance of each criterion within a set of environmental, social, and
engineering criteria is compared to other criteria, weighted and ranked, e.g. water quality
may be ranked first amongst the engineering criteria, especially if HABs are known to be a
persistent issue. Thereafter, the MCA can be linked to a Geographic Information System
(GIS) modelling and analysis platform to overlay geographic data sets to represent each of
the evaluation criteria. The GIS models can then be used to compile scores across all the
evaluation criteria and identify areas that are more suitable for the location of a seawater
intake. While these approaches aim to provide a balanced approach, some subjectivity is
unavoidable by the team with some criteria not directly measureable. Instead the experience
of the team is relied upon. Moreover, this technique can be limited by the accuracy and
currency of data sets used coupled to the fact that not all critical aspects that determine
suitability can be represented in a geographic format. Hence, results need to be verified and
validated with ground-truthing of sites.
While financial criteria may not be directly included in MCA, some criteria (e.g. engineering)
will require financial considerations of some aspects to weight the criteria. The final decision
on siting of marine infrastructure will be based on capital expenditure estimates plus
operation and maintenance expenditures at suitable sites plus constructability, permitting
requirements, construction and delivery schedules.
MCA has proven useful in selecting the location, depth and type of intake for SWRO
desalination plants. One such plant is the Gold Coast Desalination Plant which selected a
deep water intake (20 m), 1.5 km off shore based on a MCA, considering factors such as; cost,
risk, scheduled delivery window, environmental impact, community disruption, visual
amenity in addition to water quality (algal blooms were known to occur). The deep water
intake has been successful in providing good water quality and preventing the ingress of
Trichodesmium blooms which are the most frequent blooms found in the area (see Gold
Coast Case Study, Chapter 11).
6.6 SUMMARY
Intake channels and screens of surface intakes can be impacted by macroalgal (seaweed)
blooms, with the potential loss of plant availability whereas the impact of blooms of smaller
phytoplankton species at the intake is rare. Instead the microscopic algae generally pass
through intake screens, impacting downstream processes in SWRO plants. There are limited
opportunities to minimize the ingress of such blooms into a plant. Careful selection of the
location and depth of a surface intake are important considerations in areas prone to algal
blooms or the use of a subsurface intake.
Offshore velocity caps, commonly used for open-ocean intakes on large-capacity stand-alone
SWRO plants, have been proven successful in decreasing marine life impingement. They will
not however, directly reduce entrainment of free floating algae or motile algae as they cannot
detect the change in flow direction or swim away fast enough to avoid entrainment. Velocity
caps can reduce the likelihood of entrainment if they are located in an unproductive area, or
one with a lower number of drifting organisms. In addition, they could limit the zone of
199
influence of the intake to the depth level at which the velocity cap is situated, thereby
entraining only the algae and other plankton present at that depth.
Deep water open ocean intakes may be successful in avoiding some bloom-forming species,
but some species are motile or display diel vertical migration so that they move within the
water column and are not found solely within the surface, mixed layer. In the case of diel
vertical migration, a SWRO plant with a deep-water intake could change to intermittent
operation during a bloom, albeit reducing production, – operating during the daylight when
algae are more likely to be found at the surface. By the same reasoning, shallow intakes may
benefit from operating at night. However, the distribution of AOM may not reflect the
distribution of algal cells, as AOM can be extracellular and detrital in form, sinking to the
seabed or rising to the surface. Once algal cells are entrained into a plant, chlorination at the
intake could be suspended during a bloom event to limit the increase of AOC generated
through oxidation of organic matter, lysis of algal cells and release of AOM to reduce
downstream fouling and retain toxins intracellularly.
In addition to providing feedwater to a plant, subsurface intakes can act as an engineered
pretreatment step, performing similar to, or surpassing conventional or membrane
pretreatment to eliminate algae, a high percentage of bacteria, and a significant percentage of
natural organic matter (dissolved and colloidal), in particular sticky biopolymers from the
source seawater which can include AOM. Depending on the local hydrogeology and the
concentration of algae and duration of an algal bloom event, it is likely that most of the
subsurface intake systems would allow a SWRO plant to operate continuously during a
bloom without interruption. There is however a shortage of literature of SWRO plants
operating during a bloom examining the removal of algae and AOM.
Finally, intake type and location may be dictated by other project constraints such as
environmental regulations, budget or schedule. Various approaches can be used to balance all
competing factors in siting the intake for a SWRO desalination plant such as multi-criteria
analysis wherein considerations to minimize the ingress of algal blooms could be
incorporated.
6.7 REFERENCES
Allen, J., Cheng, R., Tseng, T., and Wattier, K. 2009. Update for the pilot and demonstration-
scale research evaluation of under-ocean floor seawater intake and discharge.
Proceedings of the AWWA Membrane Technology Conference, Memphis, Tennessee.
Allen, J. B., Cheng, R. C., Tseng, T. J. and Wattier, K. L., 2011. Update Evaluation of
Under-Ocean Floor Seawater Intake and Discharge. IDA Journal of Desalination and
Water Reuse 3(1), 19-25.
Boerlage, S. and Gordon, H. 2011. Assessing diffuser performance and discharge footprint
for the Gold Coast Desalination Plant. In: Proceedings of IDA World Congress, Perth,
Western Australia.
Berktay, A. 2011. Environmental approach and influence of red tide to desalination process
in the Middle East region. International Journal of Chemical and Environmental
Engineering 2(3), 183-188.
Bittner, R., Clements, J., Dale, L., Lee, S., Kavanaugh, M., and Missimer, T. M. 2015. Phase
2 report: Feasibility of subsurface intake designs for the proposed Poseidon water
desalination facility at Huntington Beach, California: Independent Scientific Advisory
Panel, California Coastal Commission and Poseidon Resources (Surfside) LLC Final
Report.
200
Carollo Consulting. 2016. Subsurface Desalination intake feasibility study. Technical

Memorandum No. 3 Basis of Design and initial screening 2016 Consultant’s report to
City of Santa Barbara, 338 pp.
Craig, K. 2012. Gold Coast, Sydney and Sur Desalination Plant Intake Overview. National
Centre of Excellence in Desalination. International Desalination Intakes and Outfalls
Workshop Proceedings. Adelaide, South Australia.
Dehwah, A. H., Al-Mashharawi, S., and Missimer, T. M. 2014. Mapping to assess feasibility
of using subsurface intakes for SWRO, Red Sea coast of Saudi Arabia. Desalination and
Water Treatment 52(13-15), 2351-2361.
Dehwah, A. H. A., Al-Mashharawi, S., Ng, K. C., and Missimer, T. M. 2016. Aquifer
treatment of seawater to remove natural organic matter before desalination.
Groundwater DOI:10.111/gwat.12476.
Dehwah, A. H. A., Li, S., Al-Mashharawi, S., Winters, H., and Missimer, T. M. 2015.
Changes in feedwater organic matter concentrations based on intake type and
pretreatment processes at SWRO facilities, Red Sea, Saudi Arabia. Desalination 360,
19-27.
Dehwah, A. H. A. and Missimer, T. M. 2016. Subsurface intake systems: Green choice for
improving feed water quality at SWRO plants, Jeddah, Saudi Arabia. Desalination, 88:
216-224. Dehwah, A.H., Al-Mashharawi, S., and Missimer, T.M. 2014. Mapping to
assess feasibility of using subsurface intakes for SWRO, Red Sea coast of Saudi Arabia.
Desalination and Water Treatment 52(13-15), 2351-2361.
Dehwah, A. H. A. and Missimer, T. M. 2017. Seabed gallery investigation of the water
pretreatment effectiveness of the active layer using a long-term column experiment,
Water Research 121, 95-108.
Delhomme, J.-P., Labregre, D., Rogala, J.–P., Mccann, D. 2005. Horizontal wells: A solution
for desalination feedwater intake and brine disposal. Proceedings, International
Desalination Association World Congress on Desalination and Water Reuse. Singapore.
Desormeaux, E. D., Meyerhofer, P. F., and Luckenbach, H. 2009. Results from nine
investigations assessing Pacific Ocean seawater desalination in Santa Cruz, California.
In: Proceedings of the International Desalination Association World Congress on
Desalination and Water Reuse, Atlantis, The Palm, Dubai, UAE: 7-12.
Doblin, M. A., Thompson, P. A., Revill, A. T., Butler, E. C., Blackburn, S. I., and
Hallegraeff, G. M. 2006. Vertical migration of the toxic dinoflagellate Gymnodinium
catenatum under different concentrations of nutrients and humic substances in culture.
Harmful Algae 5(6), 665-677.
EPA (United States Environmental Protection Agency) 2014. National pollutant discharge
elimination system—Final regulations to establish requirements for cooling water intake
structures at existing facilities and amend requirements at phase I facilities; Final Rule.
Federal Register 79(158).
Fukuoka District Waterworks Agency. 2015 Uminonakamichi Nata Seawater Desalination
Plant.
Kreshman, S. A. 1985. Seawater intakes for desalination plants in Libya. Desalination 55,
493-502.
201
Kreshman, S. A. 2001. 25 years of experience in operating thermal desalination plants.

Desalination 136(1), 141-145.
Hogan, T. W. 2015. Impingement and Entrainment at SWRO Desalination Facility Intakes.
Chapter 4. In: T. M. Missimer, B. Jones, and R. G. Maliva (Eds.). Intakes and outfalls
for seawater reverse osmosis desalination facilities: Innovations and environmental
impacts. Springer International Publishing: p. 57-78.
Lattemann, S., Rodriguez, S. G. S., Kennedy, M. D., Schippers, J. C., and Amy, G. L. 2012.
Environmental and performance aspects of pretreatment and desalination technologies.
Advances in Water Desalination 79-195.
Maliva, R. G., and Missimer, T. M. 2010. Self-cleaning beach gallery design for seawater
desalination plants. Desalination and Water Treatment 13(1-3), 88-95.
Missimer, T.M. 1997. Technical evaluation of Ranney collectors for raw water supply to
seawater reverse osmosis treatment facilities. In: Proceedings, International
Desalination Association World Congress on Desalination and Water Reuse. Vol. 1.
Madrid, Spain: 439-454.
Missimer, T. M. (with contributions by Ian Watson, Robert G. Maliva, and Tom Pankratz).
2009. Water supply development, aquifer storage, and concentrate disposal for
membrane water treatment facilities, 2nd edition: Methods in Water Resources
Evaluation Series No. 1. Schlumberger Water Services. Houston, Texas: 390 p.
Missimer, T. M. 2016. The California Ocean Plan: Does it eliminate the development of
large-capacity SWRO plants in California. In Proceedings of the American Membrane
Technology Conference & Exposition. San Antonio, Texas: 17 pp.
Missimer, T. M., Dehwah, A. H., Lujan, L., Mantilla, D., and Al-Mashharawi, S. 2015a.
Feasibility and Design of Seabed Gallery Intake Systems Along the Red Sea Coast of
Saudi Arabia with Discussion of Design Criteria and Methods Chapter 11. In: T. M.
Missimer, B. Jones, and R.G. Maliva (Eds.). Intakes and outfalls for seawater reverse
osmosis desalination facilities: Innovations and environmental impacts. Springer
International Publishing: p. 215-250.
Missimer, T. M., Ghaffour, N., Dehwah, A. H., Rachman, R., Maliva, R. G., and Amy, G.
2013. Subsurface intakes for seawater reverse osmosis facilities: Capacity limitation,
water quality improvement, and economics. Desalination 322, 37-51.
Missimer, T. M., Jones, B., and Maliva, R. G. 2015b. Intakes and Outfalls for Seawater
Reverse-Osmosis Desalination Facilities. Springer, New York: 544 pp.
Moore, J. K., and Villareal, T. A. 1996. Buoyancy and growth characteristics of three
positively buoyant marine diatoms. Marine Ecology Progress Series 132, 203–213.
Municipal Water District of Orange County January 2014. Final Summary Report Doheny
Ocean Desalination Project Phase 3 Investigation Extended Pumping and Pilot Plant
Test Regional Watershed and Groundwater Modeling Full Scale Project Conceptual
Assessment.
Pankratz, T. 2015. Overview of Intake Systems for Seawater Reverse Osmosis Facilities.
Chapter 1. In: T. M. Missimer, B. Jones, and R.G. Maliva (Eds.). Intakes and outfalls for
seawater reverse osmosis desalination facilities: Innovations and environmental impacts.
Springer International Publishing: p. 3-17.
202
Peters, T., Pintó, D., and Pintó, E. 2007. Improved seawater intake and pre-treatment system
based on Neodren technology. Desalination 203(1), 134-140.
Rachman, R., Dehwah, A. H., Li, S., Winters, H., Al-Mashharawi, S., and Missimer, T. M.
2015. Effects of well intake systems on removal of algae, bacteria, and natural organic
matter. Chapter 6. In: T. M. Missimer, B. Jones, and R. G. Maliva (Eds.). Intakes and
outfalls for seawater reverse osmosis desalination facilities: Innovations and
environmental impacts Springer International Publishing: p. 163-193.
Rachman, R. M., Li, S., and Missimer, T. M. 2014. SWRO feed water quality improvement
using subsurface intakes in Oman, Spain, Turks and Caicos Islands, and Saudi Arabia.
Desalination 351, 88-100.
RBF Consulting. 2009. Camp Pendleton desalination project feasibility study (California).
Consultant’s report to San Diego County Water Authority.
Salinas-Rodríguez Rodriguez, S. G. 2011. Particulate and organic matter fouling of SWRO
systems: Characterization, modelling and applications. Delft: CRC Press/Balkema.
ISBN:9780415620925.
Schippers, J. C., Verdouw, J. and Zweere, G. J. 1995. Predicting the clogging rate of artificial
recharge wells. Journal of Water Supply Research and Technology-Aqua, 44(1), 18-28.
Shimokawa, A. 2012. Fukuoka District desalination system with some unique methods.
National Centre of Excellence in Desalination, International Desalination Intakes and
Outfalls Workshop Proceedings. Adelaide, South Australia.
Villacorte, L.O., Ekowati, Y., Calix-Ponce, H. N, Schippers, J. C., Amy, G. L., and Kennedy,
M. D. 2015. Improved method for measuring transparent exopolymer particles (TEP)
and their precursors in fresh and saline water. Water Research 70, 300-312.
Villacorte, L. O., Kennedy, M. D., Amy, G. L. and Schippers, J. C. 2009. Measuring
transparent exopolymer particles (TEP) as indicator of the (bio) fouling potential of RO
feed water. Desalination and Water Treatment 5(1-3), 207-212.
Voutchkov, N. 2005. Thorough study is key to large beach-well intakes. Desalination and
Water Reuse Quarterly 17(1), 16-20.
WateReuse Association. 2011. Desalination Plant Intakes: Impingement and Entrainment
Impacts and Solutions. WateReuse Association White Paper.
Williams, D. E. 2015. Slant well intake systems: Design and construction. Chapter 13. In: T.
M. Missimer, B. Jones, and R. G. Maliva (Eds.). Intakes and outfalls for seawater
reverse osmosis desalination facilities: Innovations and environmental impacts. Springer,
International Publishing: p. 275-320.
Winters, H. 1997. Twenty years experience in seawater reverse osmosis and how chemicals
in pretreatment affect fouling of membranes. Desalination 110(1-2), 93-98.
Zhang, Y., Tseng, T. J., Andrews-Tate, C., Cheng, R. C., and Wattier, K. L. 2012. LBWD’s
Experience with Subsurface Intake as Pretreatment for Seawater Desalination. Water
Quality Technology Conference Toronto, ON, Canada.
Zhang, Y., Tseng, T. J., Cheng, R. C., and Wattier, K. L. 2011 Water Quality Impacted by
External Events in the Demonstration-Scale Under-Ocean Floor Seawater Intake System
Presented at Water Quality Technology Conference; Phoenix, Ariz. Reprinted with
permission from American Water Works Association.
203
Bloom prevention and control
7 BLOOM PREVENTION AND CONTROL
Clarissa R. Anderson1, Kevin G. Sellner2, and Donald M. Anderson3

1
2
Chesapeake Research Consortium, Edgewater MD USA
3
Woods Hole Oceanographic Institution, Woods Hole MA USA

7.1 Introduction ...........................................................................................................................................205
7.2 Bloom prevention ..................................................................................................................................207
7.2.1 Nutrient load reduction .................................................................................................................. 207
7.2.2 Nutrient load .................................................................................................................................. 207
7.2.3 Hydraulics..................................................................................................................................... 208
7.2.4 Mixing/destratification .................................................................................................................. 208
7.3 Bloom control ........................................................................................................................................209
7.3.1 Barley straw ................................................................................................................................... 209
7.3.2 Flocculation ................................................................................................................................... 209
7.3.3 Miscellaneous ................................................................................................................................ 211
7.3.4 Chemical additions ........................................................................................................................ 211
7.3.4.1 Copper sulfate ........................................................................................................................211
7.3.4.2 Hydrogen peroxide ................................................................................................................211
7.3.4.3 Sulfuric acid ...........................................................................................................................212
7.3.4.4 Potassium permanganate .......................................................................................................212
7.3.4.5 Chlorination ...........................................................................................................................212
7.3.5 Biological additions ....................................................................................................................... 212
7.3.5.1 Microbes ................................................................................................................................212
7.3.5.2 Competitors............................................................................................................................213
7.3.5.3 Grazers and trophic cascades .................................................................................................214
7.3.6 Combined methods/redundancy .................................................................................................... 214
7.4 Summary ...............................................................................................................................................214
7.5 References .............................................................................................................................................215
7.1 INTRODUCTION
Harmful algal blooms (HABs) are a serious and growing problem to many sectors of society,
including the desalination industry. The many problems that HABs present for seawater
reverse osmosis (SWRO) desalination plants include: 1) the production of dangerous toxins
that have the potential to contaminate treated water; 2) high algal biomass that clogs intake
filters; and 3) contributing to biofouling of equipment and SWRO membranes.
It is important to limit the impact from HABs by preventing blooms from reaching SWRO
plants in the first place, while also reducing their effects in the event that ingress to the plant
has occurred. Many of the management actions taken to respond to HABs can be termed
mitigation – i.e., dealing with an existing or ongoing bloom, and taking whatever steps are
necessary or possible to reduce negative impacts. Mitigation strategies can be classified into
two categories, precautionary impact preventions and bloom controls (Kim 2006; Anderson
2004). Precautionary impact preventions refer to monitoring, predictive, and emergent
actions - essentially actions taken to keep HABs from happening or from directly impacting a
particular resource. Several problems are immediately apparent in this regard. For one, we do
not have all of the knowledge we need about why HABs form in many areas, so it is
obviously difficult to regulate or control those factors. This argues for substantial and
sustained research on all aspects of HABs, including their ecology, physiology, and
oceanography. All too often managers and agency officials view these topics as fundamental
205
or basic science issues that have little direct practical utility, but in reality, such knowledge is
essential for the design and implementation of effective prevention strategies.
Another problem that arises with the concept of HAB prevention is that even if certain
environmental factors are known to influence the population dynamics of a specific HAB
organism, there are limitations on what can feasibly be done to modify or control those
factors. It might be known that a particular HAB is strongly influenced by the outflow of a
river system – that it is associated with a buoyant coastal current, for example - but are
unlikely to be able to justify the alteration of that river flow solely on the basis of HAB
prevention. As discussed below, it is nevertheless important to factor the possible impacts on
HABs into large-scale policy decisions on such topics as pollution reductions or alterations in
freshwater flows in response to agricultural and drinking water demands.
Obvious examples of impact prevention in the context of desalination are pretreatment
strategies that remove cells and the organic compounds they produce. These are described in
Chapter 9. In effect, these strategies are used to cope with HABs and to manage around them.
The question often arises, however, as to whether it is possible to be more pro-active. Can
something be done about blooms before they happen, or can something be done to destroy or
suppress them while they are occurring? These questions highlight the “control” aspects of
HAB management.
Bloom control is both challenging and controversial. The concept refers to actions taken to
suppress or destroy HABs, intervening directly in the bloom process. Curtailing or
suppressing the duration and magnitude of a HAB through physical, chemical, or biological
intervention are potential approaches, but this is one area where HAB science is rudimentary
and slow moving. Anderson (1997) highlighted the slow research progress on bloom control,
in contrast to aggressive policies to control pests and nuisance species in terrestrial
agriculture. A number of reasons were listed for the reticence or reluctance of scientists and
managers to explore and implement control strategies. These include:
• HABs are complex phenomena in highly dynamic environments. Many are large,
covering thousands of km2. Control strategies would be massively expensive and
logistically challenging.
• HABs are caused by algae from many phylogenetic clades (see Chapter 1), including
eukaryotes (armored and unarmored dinoflagellates, raphidophytes and diatoms,
euglenophytes, cryptophytes, haptophytes, pelagophytes, and chlorophytes) and
microbial prokaryotes (cyanobacteria that occur in both marine and freshwater
systems). Given this biodiversity, no single strategy or approach to bloom control or
suppression will apply to all harmful algae.
• HAB phenomena remain poorly understood, i.e.,“we can’t control what we don’t
understand”.
• Few, if any, countries have government agencies with the mandate to conduct
research or to implement strategies to control marine “pests”.
• The solutions may cause more damages than do the HAB problem being treated.
Each of these arguments has a counter argument, as discussed in Anderson (2004), but the
bottom line is that progress on bloom control has been slow, with advances being made by
only a few countries. The challenge is even more significant when viewed in the context of a
desalination plant. In the discussion that follows, traditional and emerging technologies in the
field of HAB mitigation and control are summarized in the context of their applicability to
HAB risk management at SWRO desalination plants. In doing this, it is recognized that
desalination plants are unlikely to undertake any large-scale bloom control or suppression
strategies outside their plants, given the cost, logistics, and uncertainty of such efforts. It may
206
be that bloom control would be considered at a small scale within an embayment or intake
lagoon, and thus it is important to know the various approaches that have been attempted in
different systems. This will also help operators address a very common question from the
public, or from plant management – “Is there anything we can do to control or stop this
bloom before it enters the plant?”
7.2 BLOOM PREVENTION
7.2.1 Nutrient load reduction

For many HABs, particularly those caused by freshwater or brackish water cyanobacteria or
nearshore or estuarine dinoflagellates, the reduction of nutrient inputs is an appropriate
strategy for preventing and/or limiting bloom magnitude. In the context of a desalination
plant, it is obviously not feasible for a plant to alter regional nutrient inputs. If HAB problems
persist in an area, the plant can, however, join with other advocates of pollution controls to
help reduce the nutrient loadings which often favor HAB development (Anderson et al. 2002;
Heisler et al. 2008).
7.2.2 Nutrient load
An example of major intervention in manipulating nutrient loads to coastal waters is the
brackish Baltic Sea. The Baltic has a long and on-going history of hypoxia and fish kills
associated with cyanobacterial blooms, such as Nodularia spumigena (Zillén et al. 2008).
Agriculture is the largest source of nitrogen entering the Baltic, but point source discharge of
sewage makes up a significant fraction of the load. To respond to these inputs, Sweden has
adaptively managed sewage outflow by intermittently releasing more N into surrounding
waters when there is a high risk of encouraging potentially toxic blooms of cyanobacteria
species, some that are N2-fixers (i.e., the cells use elemental nitrogen from the atmosphere to
form nitrate and thereby grow in waters that have low N:P ratios). If additional N is supplied
to the system, however, they are less likely to bloom (Elmgren and Larsson 2001). After the
Helsinki Convention of 1974, these Baltic blooms have been largely controlled by heavy
restrictions on land-based nutrient pollutants.
A second example is from Lake Erie, the shallowest, warmest, and most human-impacted of
the Laurentian Great Lakes in North America. While not a coastal system, its large size and
far-reaching impacts make it a good case study for marine HAB control in the context of
desalination plants. The predominant bloom species in this region is Microcystis aeruginosa,
a cyanobacterium that produces the hepatotoxin, microcystin. Importantly, and with direct
relevance to desalination, this freshwater toxin is now a known contaminant of coastal marine
waters as well due to its ability to move unaltered from watersheds to the ocean (Miller et al.
2010). Phosphorus abatement strategies in the late 1970s successfully suppressed blooms of
cyanobacteria in Lake Erie, but only until 1995 when an invasion of foreign mussels
(Dreissena polymorpha and D. bugensis, zebra and quagga mussels, respectively) opened an
ecological niche for Microcystis by selectively feeding on its competitors (Budd et al. 2001;
Juhel et al. 2006). As a result, a water treatment plant in Ohio, USA detected microcystin at
concentrations more than threefold higher than the WHO threshold of 1.0 part per billion
(ppb) in drinking water. This forced a shutdown of the municipal water supply (Henry 2013),
an event that was repeated in Toledo, Ohio in 2014 affecting nearly 500,000 people (Nelson
2015). Adaptive control of N and P akin to strategies in the Baltic Sea may be an effective
way to limit M. aeruginosa blooms in freshwater systems where drinking water is either
directly filtered or where adjacent coastal regions/SWRO plants may be affected by the toxin.
Ultimately, reductions in nitrogen and phosphorus must be considered as freshwater, brackish,
207
and oceanic system productivity (including HABs) responds spatially and seasonally to both
nutrients (Fisher et al. 1999; Conley et al. 2009; Smith and Schindler 2009).
7.2.3 Hydraulics
Decreasing residence times in some inland systems through alteration of river flows and
flushing rates can reduce blooms of cyanobacteria (Maier et al. 2004; Paerl 2014). Sellner et
al. (2015) documented the potential role of rapid flushing of ponds, lakes, or basins in
limiting recurrence of M. aeruginosa blooms in the coastal plain of Maryland, USA. The
process relies on elevated bottom shear stress to resuspend and then advect (transport)
overwintering populations of the Microcystis to systems downstream into waters less
favorable for growth. Similar hydraulic control of recently settled HAB populations might be
feasible for intake lagoons or holding ponds where treatment plant waters might be
manipulated to purge vegetative or resting stages of HAB taxa periodically from direct intake
into the SWRO plant.
7.2.4 Mixing/destratification
Several physical disturbance methods are now being tested that may not translate well to the
open, coastal zone, but
may be highly
applicable to intake
lagoons or holding
ponds given their
success in lakes and
fjords. These include
sediment capping with
chitosan - modified
sands (Pan et al. 2012),
dredging of nutrient-
rich waters to reduce
cyanobacterial bloom
initiation followed by
application of
®
Figure 7.1. SolarBee water circulator, drawing water from depth and Phoslock to trap and
circulating water horizontally near-surface by rotating paddles. Figure: sequester dissolved
Medora, Co. phosphorous (Lürling
and Faassen 2012), and
even solar-powered circulation to disturb cyanobacterial habitat in freshwater systems
(Figure 7.1; Hudnell et al. 2010).
Another novel and potentially environmentally benign approach to control blooms of cyst-
forming HAB species (e.g., Alexandrium) in shallow, localized systems is currently being
explored; the method might be applicable in holding ponds prior to SWRO intake. With this
method, manual mixing of bottom sediments buries cysts uniformly throughout the disturbed
layer, greatly reducing the number of cysts in the oxygenated surface layer, and thus the
potential inoculum for future HABs (D. Anderson, unpubl. data). Another unique approach
has been proposed, a ‘sediment-lift’ process (Imai et al. 2015) whereby bottom sediments
rich in diatom spores are pumped into nutrient-rich, mixed surface waters (Saiki Bay, Japan).
Dispersal of these resting stages in surface water may facilitate diatom growth rather than
growth of slower-growing dinoflagellates, preventing HAB impacts on local mariculture
operations. For constrained coastal bays with documented recurrent dinoflagellate blooms,
this approach might be feasible or at least explored in pilot studies. The risk to desalination
208
plants from such a strategy would be that the bloom that is facilitated or encouraged might
still be deleterious to operations if it reaches cell densities that can cause fouling.
7.3 BLOOM CONTROL
7.3.1 Barley straw

In lakes and ponds, deployment of barley straw (Figure 7.2) or dispersal of its extract can be
cost-effective alternatives to controlling
some HABs (Sellner et al. 2013; see also
references in Brownlee et al. 2003). It was
suggested that phenolic compounds in
barley straw (recently identified as
flavonoids, Xiao et al. 2013; Huang et al.
2015) are the main inhibitor of growth of
new dinoflagellate cells (i.e., algistatic)
(Terlizzi et al. 2002). Iredale et al. (2012)
showed that microbial degradation of the
barley straw releases hydrogen peroxide as
Figure 7.2. Barley straw bales distributed across well as inhibitory products from the lignin;
shoreline of drained Williston Lake, MD, USA. On
refilling the lake, the bales would be in the littoral zone
mechanical shearing is a possible solution
for slow decomposition and release of Microcystis- for accelerating this process if using fresh
inhibiting compounds. Photo: K. Sellner. rather than rotting barley straw. Note,
however, that decomposition of the barley
straw is light-dependent, and
decomposition products must accumulate
for maximum effect; hence, barley straw
efficacy requires long residence times for
straw breakdown, presumably not feasible
for high-volume SWRO intake systems.
Additionally and unfortunately, barley
straw may have limited use in coastal
marine environments where only a few
dinoflagellate species (Terlizzi et al. 2002;
Brownlee et al. 2003; Hagström et al.
2010) and halotolerant Microcystis
aeruginosa (K. Sellner, unpubl. data)
appear susceptible. The growth of some
dinoflagellate species may even be
stimulated by barley straw extract (Terlizzi
et al. 2002). That said, the heightened
specificity of the approach is appealing if
trying to avoid widespread ecological
consequences (Ferrier et al. 2005) and may
prove cost-effective in closed systems.
7.3.2 Flocculation
Clay minerals such as kaolinite and loess
Figure 7.3. Schematic diagram showing how dispersal compounds have been used effectively to
of a clay slurry can lead to particle flocculation in control HABs in Asia, Europe, and the
seawater, and scavenging and sedimentation of HAB
cells. Figure: D.M. Anderson. USA, and may be a viable solution for
suppressing blooms on fairly large scales in
209
the waters upstream or downstream of intakes. Liquid suspensions of the clay are sprayed
onto the surface layer of a bloom, resulting in scavenging and flocculation of algal cells
(Figures 7.3, 7.4), with 80 - 95% removal
efficiency of biomass from surface waters
in some cases (Sengco and Anderson 2004).
Phoslock®(lanthanum-modified bentonite)
and chitosan have both been applied to
cyanobacterial and Prymnesium blooms. In
the case of Phoslock®, which caps bottom
sediments and reduces phosphorous inputs
and concentrations and thus induces
phosphorous limitation in HAB species, pH,
growth phase, colony size, surface charge,
and chitosan quality also influence the
results (Sellner et al. 2013; Li et al. 2015).
Figure 7.4. Clay dispersal for HAB control in Korea. Increased ammonium regeneration is one
Photo: D.M. Anderson. possible negative outcome (Sellner et al.
2013), as this could further promote
cyanobacteria that respond to both P and N inputs (Paerl et al. 2011).
A concern about the use of chitosan is that bloom removal is often successful only at very
high clay and chitosan levels; Sellner et al. (2013) noted that concentrations of both materials
used in field intervention had to be 3-50x the levels suggested by Zou et al. (2006). Another
HAB former, Prymnesium, has been an expanding problem for fish aquaculture in Tasmania.
Body (2011) described successful removal of the flagellate cells from fish ponds using clays,
and current research (Seger et al. 2014) is focused on determining the most effective clays for
toxin removal. ‘Ball’ clay (20-80% kaolinite, 10- 25% mica and 6-65% quartz) has also
proved effective in severely reducing blooms of Pyrodinium bahamense var. compressum
and Gymnodinium catenatum, both marine dinoflagellates posing problems in Philippine
coastal waters. Removal efficiencies exceeded 90% (Padilla et al. 2007; Rivera 2015). A
number of other flocculants have been explored, including the wastewater treatment plant
flocculant poly-aluminum chloride (PAC, Sengco et al. 2001; Ghafari et al. 2009; Lu et al.
2015). In the latter study, Lu et al. (2015) found PAC-clay was effective in removing 90% of
a cultured Alexandrium tamarense population, 43–60% of total phosphorus and 17–30% of
total nitrogen, as well as most of the saxitoxins produced by the dinoflagellate. Extracts of
many angiosperm leaves and fruits (e.g., Li and Pan 2013; Wang et al. 2013; Tian et al. 2014)
have also been proposed as local, natural control agents in flocculation, with limited routine
use due to impracticalities of mass extraction, distribution, and application. It should be noted
that clay application can pose serious threats to benthic fauna, for example, molluscan
clearance rates (Frank et al. 2000; Shumway et al. 2003; Seo et al. 2008), juvenile clam
growth rates (Archambault et al. 2004), and burial of resident populations. Cranford and
Gordon (1992 in Hagström et al. 2010) reported “…extensive mortalities and/or significant
impact on somatic and reproductive tissue growth…” in Placopecten magellanicus when
exposed to 0.002-0.01 g bentonite L-1. Cuker (1993) reported altered pelagic food webs via
reduced visual predation of fish feeding on Chaoborus larvae in montmorillonite-treated
limnocorrals, leading to elevated midge larvae grazing on crustacean zooplankton. Rensel
and Anderson (2004) noted that a clay slurry of 200 g m-2 induced short-term coughing in
penned Atlantic salmon (Salmo salar). Further, acidified chitosan, one of the flocculants
noted above, has been shown to kill rainbow trout at concentration >0.038 ppm (Bullock et al.
2000). Bottom currents and depths and flushing rates in systems considering clay application
210
should be known and considered in selection of this option for open systems upstream of
SWRO intakes.
7.3.3 Miscellaneous
Ozonation, electrolysis, and ultrasound-induced cavitation have been proposed as a
mitigation option for small systems. All three generate free radicals, reducing harmful algae
and some toxins (e.g., Lürling et al. 2016). Several marine HA taxa have been lysed with
ozone (0.25-1 g O3 m-3) including K. brevis, Prorocentrum triestinum, Scrippsiella
trochoidea, Karenia digitale, and Amphidinium sp. (Schneider et al. 2003; Ho and Wong
2004; Oemcke et al. 2005). Brevetoxin was significantly reduced, but not eliminated, at 135
mg ozone L-1, a substantially higher oxidant level than would be found in commercial
ozonation (Schneider et al. 2003). Electrolysis of seawater yields hypochlorite, effective in
killing several dinoflagellates (see below and Jeong et al. 2002) and is used with clays to
flocculate C. polykrikoides in Korean waters (see Park et al. 2013).
Ultrasound, through the sonication of algal biomass, can kill cells, but this can release
dissolved organics, which can then be removed with addition of flocculant (e.g., Hakata et al.
2011). Ultrasound may be quite effective at controlling bloom growth (93.5% of M.
aeruginosa, Zhang et al. 2009) and removing toxins. Song et al. (2005) and Wu et al. (2011)
noted microcystin degradation, most effective at frequencies ~20kHz. In closed water
treatment plants such as SWRO plants, ultrasound might be an environmentally friendly
alternative to chemical control methods (Wu et al. 2011), but studies are needed to follow the
fate of toxins and released organic matter.
7.3.4 Chemical additions
7.3.4.1 Copper sulfate
McKnight et al. (1983) once called copper sulfate the “algicide of choice” for HABs in lakes
and reservoirs in the USA. This is because copper sulfate takes advantage of the natural
toxicity of cupric ions to phytoplankton (McKnight et al. 1983) and has been successful in
mitigating HABs in closed freshwater systems (recreational fountains, pools, ponds).
Cyanobacteria are particularly susceptible due to the inhibition of N2-fixation by CuSO4,
making it most effective in freshwater systems (Elder and Horne 1978). Copper sulfate with
chlorination is still used routinely to rid drinking water reservoirs of nuisance algae and
toxins (Zamyadi et al. 2012). In one case, it was added to a coastal marine system using crop-
dusting aircraft to combat Karenia brevis blooms in the 1950s (Rounsefell and Evans 1958).
It was also dispersed in brackish hybrid striped bass ponds resulting in fish kills and toxin in
pond waters; these negative effects were remedied by the subsequent addition of KMnO4 (see
below, Deeds et al. 2004). It is the collateral damage due to the non-specific toxicity of
copper to many marine organisms that poses severe constraints on its use in the open
environment.
7.3.4.2 Hydrogen peroxide
Additions of peroxide are also effective against several cyanobacteria (Lusty and Gobler
2017), M. aeruginosa (Lürling et al. 2014), and Planktothrix rubescens (Mattheiss et al.
2017) and P. agardhii (Matthijs et al. 2012) in lakes and small basins. Successful suppression
of an Alexandrium ostenfeldii bloom in a brackish water creek in The Netherlands (Burson et
al. 2014) as well as a microcosm “brown tide” of Aureococcus anophageferrens (Randhawa
et al. 2012) have also been noted. Importantly, peroxide additions do not seem to harm
macrofauna, and levels under 2.5 mg L-1 appear safe for herbivorous zooplankton resulting in
a final lake-wide application of 2 mg L-1 (Matthijs et al. 2012). Since eukaryotic
phytoplankton are not significantly affected by dilute peroxide, it is unclear how successful
211
this approach would be in coastal applications where the dominant HAB species are not
cyanobacteria. Higher concentrations would be harmful to all marine life, and as a result,
peroxide treatment is limited due to cost and hazardous chemical permitting (particularly in
the USA).
7.3.4.3 Sulfuric acid
Acid rain, volcanic activity, and mining are known sources of sulfuric acid to aquatic systems
and have been shown to be detrimental to phytoplankton populations via overall acidification
of surface waters (Geller et al. 1998). Enclosure experiments with toxic Prymnesium parvum
demonstrated the inhibitory effect of lowered pH on bloom development and toxin
(prymnesins) production after the addition of 0.1 N sulfuric acid (Prosser et al. 2012). Place
(A. Place, unpubl. data) has examined the effects of sulfuric acid additions to reduce pH of
surface waters surrounding local HABs temporarily, thereby increasing phytoplankton
susceptibility to flocculation and settling to the bottom.
7.3.4.4 Potassium permanganate
Potassium permanganate has been employed in wastewater treatment plants for decades as a
potent oxidizing agent to reduce biological and chemical oxygen demands in effluents
(reviewed by Scott and Ollis 1995). Use in the natural environment is a concern due to its
generic ability to oxidize all organic matter, i.e., HAB cells and all other biota. It has,
however, been used to control Karlodinium veneficum (an ichthyotoxin-producing
dinoflagellate) and its toxin in brackish hybrid striped bass ponds in Eastern Maryland, USA
(Deeds et al. 2004). In another spring-fed pond, near-bottom cold water anatoxin-producing
Planktothrix prolifera populations were effectively removed or severely reduced with bottom
water KMnO4 additions, thereby freeing the pond of detectable toxin, with only low cell
abundances (K. Sellner, unpubl. data) and no apparent impact on metazoans in the system.
The P. prolifera population has since been found at bloom levels under ice seven years post-
treatment.
7.3.4.5 Chlorination
Chlorination, through addition of NaOCl or other chloride compounds, is commonly used to
remove cyanobacteria and cyanotoxins from drinking water supplies, with the degree of toxin
degradation being toxin-specific (Zamyadi et al. 2012). Following experimental work with
electrolysis of seawater (yielding hypochlorite) and several marine dinoflagellates
(Gymnodinium catenatum, Cochlodinium polykrikoides, Akashiwo sanguinea, Lingulodinium
polyedrum, Prorocentrum micans, Alexandrium affine, and Gymnodinium impudicum), Jeong
et al. (2002) suggested effective NaOCl doses of 300-500 ppb for 10 min or 200-400 ppb for
1 h could minimize HAB exposures; however, free radicals may stress other biota in the
system (Brungs 1973). Chlorination to eliminate HAB toxins is discussed in Chapter 2.
7.3.5 Biological additions
7.3.5.1 Microbes
Viral and bacterially induced lysis (cell breakdown) are natural processes that regulate
phytoplankton communities and carbon flux (e.g., Fuhrman and Azam 1980; Salomon and
Imai 2006). Making use of this natural pathogenicity seems like a logical, cost-effective
solution to HAB control; however, the scientific community is skeptical of experiments that
introduce foreign, potentially invasive species, or have the potential to restructure natural
assemblages in an ecosystem irreversibly (Sanders et al. 2003; Secord 2003). Despite the
many laboratory studies demonstrating the harmful effects of heterotrophic bacteria on algal
species, Mayali and Azam (2004) argue that most field studies have failed to show
212
conclusively the causal relationship between the decline of a bloom in natural ecosystems
and the behavior of an introduced algicidal bacterium. Another major issue is that the
conversion from laboratory conditions to the natural environment is inherently complex
given the flexibility of predator-prey dynamics mediated by the presence or absence of other
algal species (Mayali and Azam 2004).
There are at least two studies indicating the benefit of natural plant-associated bacteria in
reducing HABs. Imai et al. (2012) and Onishi et al. (2014; Figure 1.6) noted lytic bacteria for
M. aeruginosa as well as H. akashiwo and Alexandrium tamarense in submersed
angiosperms and macroalgae, respectively, indicating potential natural control in several
Japanese bays (see Imai et al. 2014) and Puget Sound (Inaba et al. 2015).
HAB parasites are increasingly studied as bloom control agents, with Amoebophyra spp. the
most well-documented natural dinoflagellate control. The seminal work by Coats and
colleagues (e.g., Coats 1999) has stimulated other research, resulting in the identification of
several taxon-specific parasites that some suggest could be applied routinely to nearshore
zones where host dinoflagellates might be increasing, detected through satellite or other
remote detection systems (see Chapter 4). Jeong et al. (2003) have proposed using
heterotrophic dinoflagellates for control of natural HABs. Routine use of most biological
agents seems impractical at this time due to the costs associated with maintenance of the
parasite in culture, mass cultivation, manpower, and the diversity of HABs that could
frequent an area.
Although not identified for marine species, chytrid fungi have been shown to infect a wide
variety of phytoplankton in freshwater lakes, and while not necessarily host-specific, they
appear to prefer larger phytoplankton species (Kagami et al. 2007). These eukaryotic
parasites are ubiquitous in coastal marshes, and so far, only two genera of infecting fungi
have been studied for marine taxa (diatoms only), with little known regarding the magnitude
of their pathogenicity (Park et al. 2004). Interestingly, the major eukaryotic parasites
infecting marine dinoflagellates are other dinoflagellates (rather than fungi), such as
Amoebophyra spp. mentioned above. Jia et al. (2010) examined the effect of “white rot
fungus” (Trichaptum abietinum), a non-aquatic, wood-decay fungus often used to degrade
industrial pollutants, on several cyanobacterial cultures. Not only were the cultures destroyed
within 48 h, but it appeared that the fungus actively preyed on the algal cells and did not
seem to discriminate between species. Even more tantalizing is the complete degradation
after 12 h of microcystin in test M. aeruginosa cultures inoculated with mycelial pellicles. It
is important to note that white rot fungus has not been applied to larger reservoirs or drinking
water systems nor evaluated for environmental safety (Jia et al. 2012), although it has been
used to increase barley straw decomposition and inhibition of the cyanobacterium (Sellner et
al. 2015).
7.3.5.2 Competitors
Biological diversity may also be an important factor for keeping HAB species from gaining
dominance. Cardinale (2011) demonstrated that more diverse communities are naturally
buffered against nutrient enrichment relative to less diverse communities due to the enhanced
niche partitioning by benthic diatoms which increased nitrogen uptake. The promotion of
higher algal biodiversity and habitat preservation may thus be one method for facilitating
greater nutrient uptake capacity, particularly in protected environments where physical
advection processes do not dominate phytoplankton turnover rates. Allelopathic interactions
(e.g., Pratt 1966; Tang and Gobler 2011; Lim et al. 2014; Tang et al. 2014) introduced when
algae exude dissolved secondary metabolites (sometimes phycotoxins) into the environment
are an indicator of inter-specific competition for limiting resources (Graneli and Hansen
213
2006). It may also be that as species diversity increases, the ability of a given toxic species to
dominate its competitors is suppressed by the wider array of competitive strategies present in
the community. As marine ecosystem models become more sophisticated and include
realistic phytoplankton biodiversity (Follows et al. 2007; Goebel et al. 2010), varying
management strategies can be assessed in relation to the physical environment, community
composition, competitive interactions, and nutrient dynamics.
7.3.5.3 Grazers and trophic cascades
Algal proliferation is heavily regulated by grazing pressure from zooplankton, with grazing
and trophic cascades representing an often overlooked component of bloom development and
persistence (e.g., Verity and Smetacek 1996; Gobler et al. 2002; Turner and Graneli 2006;
Smayda 2008). There is evidence that eutrophication (i.e., nutrient enrichment) exerts an
indirect effect on zooplankton grazing efficiency. At higher nutrient levels, phytoplankton are
no longer suppressed by grazers (Kemp et al. 2001) and actually increase the production of
grazing deterrents (Mitra and Flynn 2006), a positive feedback that intensifies negative
impacts of HABs (Sunda et al. 2006). This idea of indirect effects is also consistent with a
simulation study by Daskalov (2002) who found that the increase in algal blooms in the
Black Sea was the result of intense overfishing of top predators, such as cetaceans and large
migratory fish. The overfishing resulted in decreased predator control of planktivorous fish,
thereby depleting zooplankton stocks and allowing phytoplankton to flourish. This ‘trophic
cascade’ was assisted by anthropogenic eutrophication that significantly relieved resource
limitation from the bottom-up (Daskalov 2002) and likely further suppressed grazing.
Another simulation study (Walsh et al. 2011) used a coupled physical-biological model of the
Chukchi/Beaufort Seas to illustrate that “fishing down of the food web” (sensu Pauly et al.
1998) and increased eutrophication in a scenario similar to that of the Black Sea supports
hypotheses of a regime shift favoring N2-fixers and HAB-forming dinoflagellates like the
saxitoxin-producing Alexandrium tamarense. Ultimately, the solution to this ecologically
complex interplay is similar to that discussed in Section 7.2.1, whereby reduction in nutrient
loads not only provides bottom-up regulation of blooms, but can lead to unexpected
consequences for other ecosystem components essential to bloom control.
7.3.6 Combined methods/redundancy
There are some examples of combinations of multiple interventions in bloom control. Sellner
et al. (2015) document the combined effects of hydraulic flushing and barley straw additions
in limiting growth of M. aeruginosa and toxin accumulation in a freshwater lake in
Maryland, USA. Flocculation, sedimentation, capping of bottom sediments, and additions of
toxin-utilizing Pseudomonas sp. in bloom control have been suggested for blooms of M.
aeruginosa in China (Li et al. 2015). As noted above, seawater electrolysis followed by clay
flocculation is an effective removal strategy for C. polykrikoides populations that threaten
fish pen mariculture in Korean (Park et al. 2013). Practical use of approaches like these near
SWRO intakes remains to be determined.
7.4 SUMMARY
There are numerous techniques to prevent and control HABs, though many remain
experimental or are practical only on a small scale. Most preventative measures involve
factors beyond the control of desalination plant operators, such as nutrient reduction from
regional watersheds or other point and non-point sources of nutrient discharge to the coastal
zone that can stimulate HABs. The increasing development of desalination plants should
serve to provide pressure on governmental entities (or favor government-private partnerships)
responsible for monitoring and controlling nutrient pollution given the need for safe, clean
214
drinking water. As covered in Chapter 3, monitoring methods that include satellite-based

bloom detection, and advanced models of HAB initiation and transport provide early warning
of delivery of potential toxins or bloom biomass to a plant’s intake system. While these
monitoring and modeling programs require significant financial investment to develop and
maintain, they allow operators to prepare for HAB events and potentially adjust intake
operations and plant pre-treatment technologies accordingly.
Fortunately, options remain for the control of HABs that have already entered a desalination
plant, and these are described in detail in Chapter 9. For plants with holding ponds or
reservoirs, intervening with physical disturbance techniques, oxidizing compounds (H2O2,
KMnO4) or clay minerals to remove HAB particles via flocculation are possibilities. Care
must be taken, however, as some turn particulate algal biomass (which is relatively easy to
remove with pretreatment) into dissolved organic compounds, which are much harder to
remove. Other post-intake technologies for SWRO include the application of copper sulfate
(cell lysis but little toxin degradation) followed by, or by themselves, the use of techniques
that generate oxidizing conditions (again peroxide or permanganate as well as chlorination,
electrolysis, ozonation, perhaps ultrasound), but post-treatment of these waters might be
required to mitigate possible damage to plant membranes.
High costs will likely prohibit the frequent use of these methods for mitigation, but several
could also be considered as options for minimizing HAB entry into a desalination plant,
particularly in the case of methods that are applicable to coastal environments or embayments,
such as hydrogen peroxide application or electrolysis-clay addition. The majority of methods
suggested in this chapter, however, are meant to give desalination operators an understanding
of the challenges associated with HAB prevention and control. Many are not practical, given
desalination–specific factors such as the high-volume intake requirements of plants, but
others may be of use, and may contribute to a suite of adaptive strategies to mitigate HAB
impacts on desalination operations.
7.5 REFERENCES
Anderson, D. M. 1997. Turning back the harmful red tide. Nature 388, 513-514.
Anderson, D. M. 2004. Prevention, control and mitigation of harmful algal blooms:
Multiple approaches to HAB management. Hall, S., Etheridge, S., Anderson, D.,
Kleindinst, J., Zhu, M., and Zou, Y. (Eds.), Harmful Algae Management and
Mitigation, Asia-Pacific Economic Cooperation (Singapore): APEC Publication #204-
MR-04.2. pp. 123-130.
Anderson, D. M., Glibert, P. M., and Burkholder, J. M. 2002. Harmful algal blooms and
eutrophication: Nutrient sources, composition, and consequences. Estuaries 25, 704–726.
Archambault, M. C., Bricelj, V. M., Grant, J., and Anderson, D. M. 2004. Effects of
suspended and sedimented clays on juvenile hard clams, Mercenaria mercenaria, within
the context of harmful algal bloom mitigation. Marine Biology 144, 553-565.
Body, A. 2011. Stable pond blooms through phosphate and pH control. Infofish International
1/2011, 22-25.
Brownlee, E. F., Sellner, S. G., and Sellner, K. G. 2003. Effects of barley straw (Hordeum
vulgare) on freshwater and brackish phytoplankton and cyanobacteria. Journal of
Applied Phycology 15, 525–531.
Brungs, W. A. 1973. Effects of residual chlorine on aquatic life. Journal of the Water
Pollution Control Federation 45, 2180-2193.
215
Budd, J. W., Drummer, T. D., Nalepa, T. F., and Fahnenstiel, G. L. 2001. Remote sensing of
biotic effects: Zebra mussels (Dreissena polymorpha) influence on water clarity in
Saginaw Bay, Lake Huron. Limnology and Oceanography 46(2), 213-223.
Bullock, G., Blazer, V., Tsukuda, S., and S. Summerfelt. 2000. Toxicity of acidified chitosan
for cultured rainbow trout (Oncorhynchus mykiss). Aquaculture 185, 273-280.
Burson, A., Matthijs, H. C., de Bruijne, W., Talens, R., Hoogenboom, R., Gerssen, A., Visser,
P. M., Stomp, M., Steur, K., and van Scheppingen, Y. 2014. Termination of a toxic
Alexandrium bloom with hydrogen peroxide. Harmful Algae 31: 125–135.
Cardinale, B. J. 2011. Biodiversity improves water quality through niche partitioning. Nature
472, 86–89.
Coats, D. W. 1999. Parasitic life styles of marine dinoflagellates. Journal of Eukaryote
Microbiology 46, 402-409.
Conley, D. J., Paerl, H. W., Howarth, R. W., Boesch, D. F., Seitzinger, S. P., Havens, K. E.,
Lancelot, C., and Likens, G. E. 2009. Controlling eutrophication: Nitrogen and
phosphorus. Science 323, 1014-1015.
Cuker, B. E. 1993. Suspended clays alter trophic interactions in the plankton. Ecology 74,
944-953.
Daskalov, G. M. 2002. Overfishing drives a trophic cascade in the Black Sea. Marine
Ecology Progress Series 225, 53–63.
Deeds, J. R., Mazzaccaro, A. P., Terlizzi, D. E., and Place, A. R. 2004. Treatment options for
the control of an ichthyotoxic dinoflagellate in an estuarine aquaculture facility: A case
study. Hall, S., Anderson, D., Kleindinst, J., Zhu, M., and Zou, Y. (Eds.), Proceedings of
the Second International Conference on Harmful Algae Management and Mitigation,
Asia-Pacific Economic Cooperation, APEC Publ. #204-MR-04.2, Singapore, pp. 177-
181.
Elder, J. F., and Horne, A. J. 1978. Copper cycles and CuSO4 algicidal capacity in two
California lakes. Environmental Management 2, 17–30.
Elmgren, R., and Larsson, U. 2001. Nitrogen and the Baltic Sea: managing nitrogen in
relation to phosphorus. The Scientific World Journal 1, 371–377.
Ferrier, M. D., Butler, B. R., Terlizzi, D. E., and Lacouture, R. V. 2005. The effects of barley
straw (Hordeum vulgare) on the growth of freshwater algae. Bioresource Technology 96,
1788–1795.
Fisher, T. R., Gustafson, A. B., Sellner, K., Lacouture, R., Haas, L. W., Wetzel, R. L.,
Magnien, R., Everitt, D., Michaels, B., and Karrh, R. 1999. Spatial and temporal
variation of resource limitation in Chesapeake Bay. Marine Biology 133, 763-778.
Follows, M. J., Dutkiewicz, S., Grant, S., and Chisholm, S. W. 2007. Emergent biogeography
of microbial communities in a model ocean. Science 315, 1843–1846.
Frank, D., Ewert, L., Shumway, S., and Ward, J. E. 2000. Effect of clay suspensions on
clearance rates in three species of benthic invertebrates. Journal of Shellfish Research 19,
663.
Fuhrman, J. A., and Azam, F. 1980. Bacterioplankton secondary production estimates for
coastal waters of British Columbia, Antarctica, and California. Applied and
Environmental Microbiology 39, 1085–1095.
216
Geller, W., Klapper, H., and Schultze, M. 1998. Natural and anthropogenic sulfuric
acidification of lakes. Geller, W., Klapper, H., and Salomons, W. (Eds.), Acidic Mining
Lakes. Springer, Berlin Heidelberg, pp. 3–14.
Ghafari, S., Aziz, H. A., Isa, M. H., and Zinatizadeh, A. A. 2009. Application of response
surface methodology (RSM) to optimize coagulation–flocculation treatment of leachate
using poly-aluminum chloride (PAC) and alum. Journal of Hazardous Materials 163,
650–656.
Gobler, C. J., Renaghan, M. J., and Buck, N. J. 2002. Impacts of nutrients and grazing
mortality on the abundance of Aureococcus anophagefferens during a New York brown
tide bloom. Limnology and Oceanography 47, 129–141.
Goebel, N. L., Edwards, C. A., Zehr, J. P., and Follows, M. J. 2010. An emergent community
ecosystem model applied to the California Current System. Journal of Marine Systems
83, 221–241.
Graneli, E., and Hansen, P. J. 2006. Allelopathy in harmful algae: A mechanism to compete
for resources? Granéli, E. and Turner, J. T. (Eds.), Ecology of Harmful Algae, Springer.
pp. 189–201.
Hagström, J. A., Sengco, M. R., and Villareal, T. A. 2010. Potential methods for managing
Prymnesium parvum blooms and toxicity, with emphasis on clay and barley straw: a
review. Journal of the American Water Resources Association 46, 187–198.
Hakata, Y., Roddick, F., and Fan, L. 2011. Impact of ultrasonic pre-treatment on the
microfiltration of a biologically treated municipal effluent. Desalination 283, 75–79.
Heisler, J., Glibert, P., Burkholder, J., Anderson, D., Cochlan, W., Dennison, W., Dortch, Q.,
Gobler, C., Heil, C., Humphries, E., Lewitus, A., Magnien, R., Marshall, H., Sellner, K.,
Stockwell, D., Stoecker, D., and Suddleson, M. 2008. Eutrophication and harmful algal
blooms: A scientific consensus. Harmful Algae 8, 3-13.
Henry, T. 2013. Carrol Township’s scare with toxin a “wake-up call.” Toledo Blade 15.
Ho, K. C. and Wong, Y. K. 2004. A study on the effectiveness of using ozone to mitigate
harmful algal blooms. Hall, S., Anderson, D., Kleindinst, J., Zhu, M., and Zou, Y. (Eds.),
Proceedings of the Second International Conference on Harmful Algae Management and
Mitigation, Asia-Pacific Economic Cooperation, APEC Publ. #204-MR-04.2, Singapore,
pp. 185-188.
Huang, H., Xiao, X., Ghadouani, A., Wu, J., Nie, Z., Peng, C., Xu, X., and Shi, J. 2015.
Effects of natural flavonoids on photosynthetic activity and cell integrity in Microcystis
aeruginosa. Toxins 7, 66-80.
Hudnell, H. K., Jones, C., Labisi, B., Lucero, V., Hill, D. R., and Eilers, J. 2010. Freshwater
harmful algal bloom (FHAB) suppression with solar powered circulation (SPC). Harmful
Algae 9, 208–217.
Imai, I., Imai, Y., Toda, T., Noda, M., and Miyamura, K. 2015. Manipulating phytoplankton
flora to harmless diatoms through germination of resting stage cells by sediment-lift to
euphotic layer in coastal sea. Proceedings of the 16th International Conference on
Harmful Algae, Abstract and presentation, Wellington, NZ.
217
Imai, I., Shimada, H., Shinada, A., Baba, K., Kanamori, M., Sato, M., Kuwahara, Y.,
Miyoshi, K., Tada, M., Hirano, K., Miyazono, A., and Itakura, S. 2014. Prediction of
toxic algal bloom occurrences and adaptation to toxic blooms to minimize economic loss
to the scallop aquaculture industry in Hokkaido, Japan. Trainer, V.L. and Yoshida, T.
(Eds.). Proceedings of a Workshop on Economic Impacts of Harmful Algal Blooms on
Fisheries and Aquaculture, PICES Scientific Report Number. 47, pp. 7-16.
Imai, Y., Ohgi, K., Mizuhara, S., Yamaguchi, A., and Kaerlyama, M. 2012. Experimental
study on control of Microcystis aeruginosa bloom by using of microorganisms in waters
of a reed community and a water plant zone. Abstract and presentation, 15th International
Conference on Harmful Algae, Gyeongnam, Korea.
Inaba, N., Onishi, Y., Ishii, K., Trainer, V. L., Wylie-Echeverria, S., Hatch, M. B. A., and
Imai, I. 2015. High density of algicidal and growth-inibiting bacteria against harmful
algal species detected from seagrass and macroalgae in Puget Sound. Abstract and
presentation, 16th International Conference on Harmful Algae, Wellington, NZ.
Iredale, R. S., McDonald, A. T., and Adams, D. G. 2012. A series of experiments aimed at
clarifying the mode of action of barley straw in cyanobacterial growth control. Water
Research 46, 6095–6103.
Jeong, H. J., Kim, J. S., Yoo, Y. D., Kim, S. T., Kim, T. H., Park, M. G., Lee, C. H., Seong,
K. A., Rang, N. S., and Shim, J. H. 2003. Feeding by the heterotrophic
dinoflagellate Oxyrrhis marina on the red-tide raphidophyte Heterosigma akashiwo: A
potential biological method to control red tides using mass-cultured grazers. Journal of
Eukaryotic Microbiology 50, 274-282.
Jeong, H. J., Kim, H. R., Kwang, I. K., Kwang, Y. K., Park, K. H., Seong, T. K., Yoo, Y. D.,
Song, J. Y., Kim, J. S., Seong, K. A., Pae, W. H. S. E., Lee, C. H., Huh, M. D., and Lee,
S. H. 2002. NaOCl produced by electrolysis of natural seawater as a potential method to
control marine red-tide dinoflagellates. Phycologia 41, 643-656.
Jia, Y., Han, G., Wang, C., Guo, P., Jiang, W., Li, X., and Tian, X. 2010. The efficacy and
mechanisms of fungal suppression of freshwater harmful algal bloom species. Journal of
Hazardous Materials 183, 176–181.
Jia, Y., Du, J., Song, F., Zhao, G., and Tian, X. 2012. A fungus capable of degrading
microcystin-LR in the algal culture of Microcystis aeruginosa PCC7806. Applied
Biochemistry and Biotechnology 166, 987–996.
Juhel, G., Davenport, J., O’Halloran, J., Culloty, S. C., O’Riordan, R. M., James, K. F., Furey,
A., and Allis, O. 2006. Impacts of microcystins on the feeding behaviour and energy
balance of zebra mussels, Dreissena polymorpha: A bioenergetics approach. Aquatic
Toxicology 79, 391–400.
Kagami, M., de Bruin, A., Ibelings, B. W., and Van Donk, E. 2007. Parasitic chytrids: their
effects on phytoplankton communities and food-web dynamics. Hydrobiologia 578, 113–
129.
Kemp, W. M., Brooks, M. T., and Hood, R. R. 2001. Nutrient enrichment, habitat variability
and trophic transfer efficiency in simple models of pelagic ecosystems. Marine Ecology
Progress Series 223, 73–87.
Kim, H. G. 2006. Mitigation and controls of HABs. Granéli, E. and Turner, J.T. (Eds.),
Ecology of Harmful Algae, Springer. pp. 327–338.
218
Li, L. and Pan, G. 2013. A universal method for flocculating harmful algal blooms in marine
and fresh waters using modified sand. Environmental Science and Technology 47, 4555–
4562.
Li, L., Zhang, H., and Pan, G. 2015. Influence of zeta potential on the flocculation of
cyanobacteria cells using chitosan modified soil. Journal of Environmental Sciences 28,
47–53.
Lim, A. S., Jeong, H. J., Jang, T. Y., Jang, S. H., and Franks, P. J. 2014. Inhibition of growth
rate and swimming speed of the harmful dinoflagellate Cochlodinium polykrikoides by
diatoms: Implications for red tide formation. Harmful Algae 37, 53–61.
Lu, G., Song, X., Yu, Z., Cao, X., and Yuan, Y. 2015. Environmental effects of modified
clay flocculation on Alexandrium tamarense and paralytic shellfish poisoning toxins
(PSTs). Chemosphere 127, 188–194.
Lürling, M. and Faassen, E. J. 2012. Controlling toxic cyanobacteria: Effects of dredging and
phosphorus-binding clay on cyanobacteria and microcystins. Water Research 46, 1447–
1459.
Lürling, M., Meng, D., and Faassen, E. J. 2014. Effects of hydrogen peroxide and ultrasound
on biomass reduction and toxin release in the cyanobacterium, Microcystis
aeruginosa. Toxins 6, 3260-3280.
Lürling, M., Waajen, G., and de Senerpont Domis, L. N. 2016. Evaluation of several end-of-
pipe measures proposed to control cyanobacteria. Aquatic Ecology 50, 499-519.
Maier, H. R., Kingston, G. B., Clark, T., Frazer, A., and Sanderson, A. 2004. Risk-based
approach for assessing the effectiveness of flow management in controlling
cyanobacterial blooms in rivers. River Research Applications 20, 459-471.
Mattheiss, J., Sellner, K. G., and Ferrier, D. 2017. Lake Anita Louise Peroxide Treatment
Summary – December 2016. CCWS Contribution #17-01, Center for Coastal and
Watershed Studies, Hood College, Frederick, MD. 9 pp.
Matthijs, H. C., Visser, P. M., Reeze, B., Meeuse, J., Slot, P. C., Wijn, G., Talens, R., and
Huisman, J. 2012. Selective suppression of harmful cyanobacteria in an entire lake with
hydrogen peroxide. Water Research 46, 1460–1472.
Mayali, X., and Azam, F. 2004. Algicidal bacteria in the sea and their impact on algal blooms.
Journal of Eukaryotic Microbiology 51, 139–144.
McKnight, D. M., Chisholm, S. W., and Harleman, D. R. 1983. CuSO4 treatment of nuisance
algal blooms in drinking water reservoirs. Environmental Management 7, 311–320.
M., Miller, W. A., Toy-Choutka, S., and Dominik, C. 2010. Evidence for a novel marine
harmful algal bloom: cyanotoxin (microcystin) transfer from land to sea otters. PLoS One,
5: e12576.
Mitra, A., and Flynn, K. J. 2006. Promotion of harmful algal blooms by zooplankton
predatory activity. Biology Letters 2, 194–197.
Nelson, L. A. 2015. Community assessment for public health emergency response
(CASPER) following detection of microcystin toxin in a municipal water system, Ohio,
September 2014. Proceedings of the 2015 CSTE Annual Conference. Cste.
219
Oemcke, D. J. and Van Leeuwen, J. H. 2005. Ozonation of the marine dinoflagellate alga
Amphidinium sp.—implications for ballast water disinfection. Water Research 39, 5119–
5125.
Onishi, Y., Mohri, Y., Tuji, A., Ohgi, K., Yamaguchi, A., and Imai, I. 2014. The seagrass
Zostera marina harbors growth inhibiting bacteria against the toxic dinoflagellate
Alexandrium tamarense. Fisheries Science 80, 353–362.
Padilla, L., San Diego-McGlone, M. L., and Azanza, R. 2007. Preliminary results on the use
of clay to control Pyrodinium bloom-a mitigation strategy. Science Diliman 18(1), 35-42.
Paerl, H. W. 2014. Mitigating harmful cyanobacterial blooms in a human- and climatically-
impacted world. Life 4, 988-1012.
Paerl, H. W., Hall, N. S., and Calandrino, E. S. 2011. Controlling harmful cyanobacterial
blooms in a world experiencing anthropogenic and climatic-induced change. Science of
the Total Environment 409, 1739–1745.
Pan, G., Dai, L., Li, L., He, L., Li, H., Bi, L., and Gulati, R. D. 2012. Reducing the
recruitment of sedimented algae and nutrient release into the overlying water using
modified soil/sand flocculation-capping in eutrophic lakes. Environmental Science and
Technology 46, 5077–5084.
Park, M. G., Yih, W., and Coats, D. W. 2004. Parasites and phytoplankton, with special
emphasis on dinoflagellate infections. Journal of Eukaryotic Microbiology 51, 145–155.
Park, T. G.,Lim, W. A., Park, Y. T., Lee, C. K., and Jeong, H. J. 2013. Economic impact,
management and mitigation of red tides in Korea. Harmful Algae 30S, S131-S143.
Pauly, D., Christensen, V., Dalsgaard, J., Froese, R., and Torres, F. 1998. Fishing down
marine food webs. Science 279, 860–863.
Pratt, D. M. 1966. Competition between Skeletonema costatum and Olisthodiscus luteus in
Narragansett Bay and in culture. Limnology and Oceanography 11, 447–455.
Prosser, K. N., Valenti, T. W., Hayden, N. J., Neisch, M. T., Hewitt, N. C., Umphres, G. D.,
Gable, G. M., Grover, J. P., Roelke, D. L., and Brooks, B. W. 2012. Low pH pre-empts
bloom development of a toxic haptophyte. Harmful Algae 20, 156–164.
Randhawa, V., Thakkar, M., and Wei, L. 2012. Applicability of hydrogen peroxide in brown
tide control - culture and microcosm studies. PLOS One
https://doi.org/10.1371/journal.pone.0047844
Rensel, J. E. J. and Anderson, D. M. 2004. Effects of phosphatic clay dispersal at two
divergent sites in Puget Sound, Washington. In: Steidinger, K.A., Landsberg, J.H.,
Tomas, C.R., and Vargo, G.A. (Eds.), Harmful Algae 2002. Florida Fish and Wildlife
Conservation Commission, Florida Institute of Oceanography and Intergovernmental
Ocenographic Commission of UNESCO, St. Petersburg Beach, Florida, pp. 522-524.
Rivera, P. P. L. 2015. Removal of harmful algal bloom (HAB)-forming organisms using ball
clay: Factors and effects of clay addition. University of Philippines. pp. 101.
Rounsefell, G. A., and Evans, J. E. 1958. Large-scale experimental test of copper sulfate as a
control for the Florida red tide. U.S. Fish and Wildlife Service Special Scientific Report.
Salomon, P. S. and Imai, I. 2006. Pathogens of harmful microalgae. Granéli, E. and Turner,
J.T. (Eds.), Ecology of Harmful Algae, Springer. pp. 271–282.
220
Sanders, N. J., Gotelli, N. J., Heller, N. E., and Gordon, D. M. 2003. Community disassembly
by an invasive species. Proceedings of the National Academy of Sciences USA, 100,
2474–2477.
Schneider, K. R., Pierce, R. H., and Rodrick, G. E. 2003. The degradation of Karenia brevis
toxins utilizing ozonated seawater. Harmful Algae 2, 101-107.
Scott, J. P. and Ollis, D. F. 1995. Integration of chemical and biological oxidation processes
for water treatment: review and recommendations. Environmental Progress 14, 88–103.
Secord, D. 2003. Biological control of marine invasive species: cautionary tales and land-
based lessons. Biological Invasions 5, 117–131.
Seger, A., Body, A., and Hallegraeff, G. M. 2014. Exploration of Phoslock™ clay in
mitigating Prymnesium parvum fish-killing algal blooms in aquaculture ponds.
Proceedings of the 15th International Conference on Harmful Algae, Changwon, South
Korea, October.
Sellner, K. G., Place, A. R., Paolisso, M., Gao, Y., Williams, E. and VanDolah, F. M. 2013.
Options in mitigating cyanobacterial blooms. Proceedings of the 7th U.S. Harmful Algal
Bloom Symposium, Abstract and presentation.
Sellner, K. G., Place, A. R., Williams, E., Gao, Y., VanDolah, E., Paolisso, M., Bowers, H.
A., and Roche, S. 2015. Hydraulics and barley straw (Hordeum vulgare) as effective
treatment options for a cyanotoxin-impacted lake. L. MacKenzie (Ed.), Proceedings of
the 16th International Conference on Harmful Algae, Cawthron Institute and
International Society for the Study of Harmful Algae, Vawthron, NZ. pp. 218-221.
Sengco, M. R., Li, A., Tugend, K., Kulis, D., and Anderson, D. M. 2001. Removal of red-and
brown-tide cells using clay flocculation. I. Laboratory culture experiments with
Gymnodinium breve and Aureococcus anophagefferens. Marine Ecology Progress Series
210, 41–53.
Sengco, M. R., and Anderson, D. M. 2004. Controlling harmful algal blooms through clay
flocculation1. Journal of Eukaryotic Microbiology 51, 169–172.
Seo, K. S., Lee, C. K., Park, Y. T., and Lee, Y. 2008. Effect of yellow clay on respiration and
phytoplankton uptake of bivalves. Fisheries Science 74, 120-127.
Shumway, S. E., Frank, D. M., Ewart, L. M., and Ward, L. E. 2003. Effect of yellow loess on
clearance rate in seven species of benthic, filter-feeding invertebrates. Aquaculture
Research 34, 1391-1402.
Smayda, T. J. 2008. Complexity in the eutrophication–harmful algal bloom relationship, with
comment on the importance of grazing. Harmful Algae 8, 140–151.
Smith, V. H. and Schindler, D. W. 2009. Eutrophication science: where do we go from here?
Trends in Ecology & Evolution 24, 201–207.
Song, W., Teshiba, T., Rein, K., and O'Shea, K. E. 2005. Ultrasonically induced degradation
and detoxification of microcystin-LR (cyanobacterial toxin). Environmental Science and
Sunda, W. G., Granéli, E., and Gobler, C. J. 2006. Positive feedback and the development
and persistence of ecosystem disruptive algal blooms. Journal of Phycology 42, 963–974.
Tang, Y. Z., and Gobler, C. J. 2011. The green macroalga, Ulva lactuca, inhibits the growth
of seven common harmful algal bloom species via allelopathy. Harmful Algae 10, 480–
488.
221
Tang, Y. Z., Kang, Y., Berry, D., and Gobler, C. J. 2014. The ability of the red macroalga,
Porphyra purpurea (Rhodophyceae) to inhibit the proliferation of seven common
harmful microalgae. Journal of Applied Phycology 27, 531–544.
Terlizzi, D. E., Ferrier, M. D., Armbrester, E. A., and Anlauf, K. A. 2002. Inhibition of
dinoflagellate growth by extracts of barley straw (Hordeum vulgare). Journal of Applied
Phycology 14, 275–280.
Tian, F., Zhou, J., Sun, Z., Cai, Z., Xu, N., An, M., and Duan, S. 2014. Inhibitory effects of
Chinese traditional herbs and herb-modified clays on the growth of harmful algae,
Phaeocystis globosa and Prorocentrum donghaiense. Harmful Algae 37, 153–159.
Turner, J. T. and Granéli, E. 2006. “Top-down” predation control on marine harmful algae,
Granéli, E. and Turner, J.T. (Eds.), Ecology of Harmful Algae, Springer. pp. 355–366.
Verity, P. G., and Smetacek, V. 1996. Organism life cycles, predation, and the structure of
marine pelagic ecosystems. Marine Ecology Progress Series 130, 277–293.
Walsh, J. J., Dieterle, D. A., Chen, F. R., Lenes, J. M., Maslowski, W., Cassano, J. J.,
Whitledge, T. E., Stockwell, D., Flint, M., and Sukhanova, I. N. 2011. Trophic cascades
and future harmful algal blooms within ice-free Arctic Seas north of Bering Strait: A
simulation analysis. Progress in Oceanography 91, 312–343.
Wang, L., Liang, W., Yu, J., Liang, Z., Ruan, L., and Zhang, Y. 2013. Flocculation of
Microcystis aeruginosa using modified larch tannin. Environmental Science and
Wu, X., Joyce, E. M., and Mason, T. J. 2011. The effects of ultrasound on cyanobacteria.
Harmful Algae 10, 738–743.
Xiao, X., Huang, H., Ge, Z., Rounge, T. B., Shi, J., Xu, X., Li, R., and Chen, Y. 2013. A pair
of chiral flavonolignans as novel anti-cyanobacterial allelochemicals derived from barley
straw (Hordeum vulgare): Characterization and comparison of their anti-cyanobacterial
activities. Environmental Microbiology doi:10.1111/1462-2920.12226.
Zamyadi, A., MacLeod, S. L., Fan, Y., McQuaid, N., Dorner, S., Sauvé, S., and Prévost, M.
2012. Toxic cyanobacterial breakthrough and accumulation in a drinking water plant: A
monitoring and treatment challenge. Water Research 46, 1511–1523.
Zhang, G., Zhang, P., and Fan, M. 2009. Ultrasound-enhanced coagulation for Microcystis
aeruginosa removal. Ultrasonics Sonochemistry 16, 334–338.
Zillén, L., Conley, D. J., Andrén, T., Andrén, E., and Björck, S. 2008. Past occurrences of
hypoxia in the Baltic Sea and the role of climate variability, environmental change and
human impact. Earth Science Reviews 91, 77–92.
Zou, H., Pan, G., Chen, H., and Yuan, X. 2006. Removal of cyanobacteria blooms in Taihu
Lake using local soils. II. Effective removal of Microcystis aeruginosa using local soils
and sediments modified by chitosan. Environmental Pollution 141, 201-205.
222
Guidelines for toxin control, HAB management, and response planning
8! WORLD HEALTH ORGANIZATION AND INTERNATIONAL

GUIDELINES FOR TOXIN CONTROL, HARMFUL ALGAL
BLOOM MANAGEMENT, AND RESPONSE PLANNING
Alex Soltani1, Philipp Hess2, Mike B. Dixon3, Siobhan F.E. Boerlage4, Donald M. Anderson5,
Gayle Newcombe6,7, Jenny House6,7, Lionel Ho6,7, Peter Baker6,7, and Michael Burch6,7
1
Alex Soltani Consulting, Calgary, Alberta, Canada
2
3
4
5
Woods Hole Oceanographic Institution, Woods Hole MA 02543 USA
6
7
!
8.1! Guidelines and standards ....................................................................................................................... 223!
8.1.1! Drinking water guideline values for harmful algal bloom (HAB) toxins .....................................223!
8.1.1.1! Fresh- and brackish-water toxins ........................................................................................... 223!
8.1.1.2! Marine toxins ......................................................................................................................... 224!
8.2! Using guideline values .......................................................................................................................... 228!
8.3! Australian drinking water guidelines regarding multiple treatment barriers ........................................ 228!
8.3.1! What is the multiple barrier concept? ............................................................................................228!
8.3.2! Monitoring as a barrier ..................................................................................................................229!
8.3.3! Multiple barriers in the context of HAB toxins .............................................................................230!
8.4! Risk assessment for the presence of HABs ........................................................................................... 230!
8.4.1! Risk assessment .............................................................................................................................231!
8.4.2! Assessing the likelihood of HABs in a water source ....................................................................234!
8.4.3! Hazard Analysis and Critical Control Point (HACCP) .................................................................235!
8.5! Alert level frameworks .......................................................................................................................... 240!
8.5.1! History of Alert Level Frameworks ..............................................................................................241!
8.5.2! Using an Alert Level Framework ..................................................................................................241!
8.5.3! Levels of the framework ................................................................................................................242!
8.5.3.1! Derivation and definition of the levels .................................................................................. 242!
8.5.3.2! Detection level ....................................................................................................................... 242!
8.5.4! Customer and media information dissemination ...........................................................................245!
8.6! Summary ............................................................................................................................................... 247!
8.7! References ............................................................................................................................................. 247!
8.1! GUIDELINES AND STANDARDS
8.1.1 Drinking water guideline values for harmful algal bloom (HAB) toxins
8.1.1.1! Fresh- and brackish-water toxins
Drinking water guidelines are designed to protect public health and the safety of drinking
water supplies by suggesting safe levels for constituents that are known to be hazardous to
health. The World Health Organization (WHO) Guidelines for Drinking Water Quality
(WHO 1996; 2004) represent a scientific consensus on the health risks presented by microbes
and chemicals in drinking water and are often used to derive guideline values for individual
countries, states or regions. The guideline values are to be used in the development of risk
management strategies and are associated with guidance on monitoring and management
practices.
Although no human deaths due to the consumption of cyanotoxins have been recorded,
cyanobacteria and their toxins remain a significant issue for the WHO, since extended
223
exposure may cause gastroenteritis, among other more serious health impacts
(NHMRC/NRMMC 2011). In addition, cyanotoxins are suspected to have resulted in
fatalities when introduced into the human body through routes other than ingestion, such as
through the use of toxin-containing water for renal dialysis (Jochimsen et al. 1998).
Motivated by growing concern over the presence of cyanotoxins in drinking water, the WHO
published an addendum to its Guidelines for Drinking Water Quality in 1998, which included
a guideline value for microcystin-LR (MCLR), an acutely toxic cyanotoxin (WHO, 1998).
The health-based guideline value for total (i.e., free plus cell-bound) concentration of MCLR
was set at 1 µg/L; however, the WHO emphasizes that the guideline value is only provisional,
since it only pertains to MCLR, and since the toxicity data for other cyanotoxins are still
being collected (WHO, 2004). According to the WHO, not enough data exist to allow
guideline values for other cyanotoxins to be developed (WHO 2004).
Concern over drinking water contamination by cyanotoxins has also grown among national
regulatory bodies, due to the increasing impact of anthropogenic activity on water resources,
as well as the improvement of analytical methods identifying and measuring cyanotoxins. For
example, Australian drinking water authorities have set a guideline value of 1.3 µg/L for
microcystins, expressed as MCLR. New Zealand has developed maximum allowable values
(MAVs) for several cyanotoxins, including anatoxin and anatoxin-A, cylindrospermopsin,
microcystins, nodularin, and saxitoxins. The US Environmental Protection Agency, on the
other hand, has yet to set any firm, enforceable maximum contaminant levels (MCLs) for
cyanobacterial toxins, and has only added cyanobacteria and their toxins to its candidate
contaminant list (CCL), which prioritizes contaminants for setting MCLs. In Canada, a
maximum acceptable concentration (MAC) of 1.5 μg/L has been developed for
cyanobacterial toxins, expressed as MCLR. Canada’s guideline was derived using tolerable
daily intake (TDI) values, determined using no-observed adverse effect levels (NOAEL),
which are based on human or animal toxicity studies. Brazil has developed guidelines for
three cyanobacterial toxins (microcystins, saxitoxins, and cylindrospermopsin), with
guideline values being set as 1.0 μg/L, 3.0" μg/L, and 15" μg/L, respectively. Several other
countries, however, still rely on the WHO provisional guideline of 1"μg/L MCLR.
A comprehensive summary of international guideline values for cyanobacterial toxins from
various countries worldwide are summarized in Table 8.1.
8.1.1.2! Marine toxins
No toxin guidelines currently exist for drinking water produced specifically from the
seawater that has been affected by HABs such as dinoflagellates and diatoms. As a result,
operators must look to existing guidelines for drinking water produced from fresh water
sources. In order to build knowledge of the potential for future guidelines, this chapter
discusses the existing guidelines internationally for cyanobacterial toxins from fresh water.
The WHO has a guideline specifically for microcystin-LR (1.0µg/L), a common analogue of
microcystin, a toxin produced by several species of cyanobacteria. Other guidelines around
the world (Table 8.1) have been determined in part by using this guidance. Relevant for
seawater desalination are the guidelines for saxitoxin set by New Zealand (1.0 µg/L) and
Brazil (3.0 µg/L). Outside the saxitoxin (STX) family, no guidelines exist as yet for other
224
Table 8.1. Guideline values or standards for cyanotoxins in drinking water from various countries. (From Newcombe et al. 2010, reproduced
with permission from Water Research Australia and updated in 2015.)!
Country Guideline Value Comments/Explanations

Argentina Under revision
Australia 1.3 µg/L total microcystins,
Australian Drinking Water Guidelines
guideline value
Brazil 1.0 µg/L for microcystins; Guideline values for microcystins, saxitoxins and cylindrospermopsin, along with biomass monitoring
3.0 µg/L for saxitoxins programs. Guideline value for microcystins adopted as mandatory. Guideline values for equivalents of
(equivalents); saxitoxins and for cylindrospermopsin included as recommendations. Use of algicides prohibited and
15 µg/L for cylindrospermopsin toxicity testing/toxin analysis when cell counts exceed 10,000,000 cells/L or 1 mm3 biovolume.
Canada 1.5 µg/L cyanobacterial toxins Canada uses guidelines as the standard of water quality. The guidelines are expressed with the unit of
as MCLR MAC Maximum acceptable concentrations (MAC). These are derived from tolerable daily intake (TDI),
which in turn are derived from a calculated no-observed adverse effect level (NOAEL) from data
from human or animal studies. To derive a MAC from a TDI, adjustments are made for average body
weight and drinking water consumption, as well as other considerations. ln terms of health, the
guidelines ensure that the MACs are far below exposure levels, at which adverse effects have been
observed. For the case of cyanobacterial toxins, the guideline is considered protective of human health
against exposure to other microcystins (total microcystins) that may also be present.
Czech Republic 1 µg/L MCLR Value as national legislation, follows WHO provisional guideline value.
China 1 µg/L MCLR WHO provisional guideline for microcystin-LR (MCLR)
France 1 µg/L MCLR Drinking water decree
Italy 1 µg/L MCLR WHO provisional guideline for MCLR used as a reference by local authorities.
Japan 1 µg/L MCLR WHO provisional guideline for MCLR
Korea 1 µg/L MCLR WHO provisional guideline for MCLR
Table 8.1 (Continued)
Country Guideline Value Comments/Explanations

New Zealand MAV for cyanobactcrial toxins: Maximum Acceptable Values (MAVs) for micro-organisms or organic determinants of health
- Anatoxin-a (S): 6.0 µg/L significance. MAVs are based on World Health Organization (WHO) Guidelines for Drinking Water
- Cylindrospermopsin: 1.0 µg/L Quality. They are the concentration of a determinant, which is not considered to cause any significant
- Microcystins: 1.0 µg/L risk to the consumer over a lifetime of consumption of water. The method of derivation varies
- Nodularin: 1.0 µg/L according to conditions and the way in that the determinant presents a risk. However, they are derived
- Saxitoxins: 1.0 µg/L with the use of a TDI. The MAVs are standards in New Zealand. The standards provide compliance
criteria and compliance is routinely monitored.
Norway 1 µg/L MCLR Provisional WHO guideline for drinking water adopted
Oceana Clean drinking water supply to all people main current focus
Poland 1 µg/L MCLR National legislation for guideline value in drinking water
South Africa 0 to 0.8 µg/L for MCLR Guideline levels for microcystins in potable water as a "Target Water Quality Range"
Spain 1 µg/L MCLR National legislation, maximum permissible amount in drinking water
Thailand No guideline currently Awareness for need for guidelines
USA None currently known Maximum Contaminant Levels (MCLs) are the highest level of a contaminant that is allowed in
drinking water. They are enforceable standards. Cyanobacteria and their toxins are listed as
microbiological contaminants on the contaminant candidate list (CCL). This means that they are
currently recognized as unregulated contaminants, but are known to occur in public water systems and
may require regulation under the Safe Drinking Water Act. A new draft recommends Health Alert
(HA) levels at or below 0.3 µg/L for microcystins and 0.7 µg/L for cylindrospermopsin in drinking
water for children pre-school age and younger (less than six years old). For school-age children
through adults, the recommended HA levels for drinking water are at or below 1.6 µg/L for
microcystins and 3.0 µg/L for cylindrospermopsin.
Uruguay Under revision
WHO 1 µg/L MCLR Refer to World Health Organization (WHO) Guidelines for Drinking Water Quality
commonly found seawater HAB toxins, such as the brevetoxins (BTX), okadaic acid (OA),
and domoic acid (DA).
Some guidelines, such as the Australian Drinking Water Guidelines (ADWG) provide
information and recommendations for several individual classes of toxins, but not necessarily
a specific concentration of toxin. In this case microcystins, nodularin, saxitoxins, and
cylindrospermopsin (NHMRC/NRMMC 2011). In the Australian case, a guideline value has
been recommended only for total microcystins. The guideline recommends that the
concentration of total microcystins in drinking water in Australia should not exceed 1.3 µg/L.
No guideline values could be set for concentrations of nodularin, saxitoxins, or
cylindrospermopsin in Australia due to the lack of adequate data. In relation to
lipopolysaccharides (LPS) produced by cyanobacteria, there is currently insufficient
information to carry out a critical assessment on occurrence and significance of LPS and so
no fact sheet has been produced. The most recent review of the ADWG has now led to the
recommendation that, although the strength of data is insufficient to establish a guideline
value for cylindrospermopsin in drinking water, a range of information can be used to
develop a ‘Health Alert’ value for cylindrospermopsin of 1 µg/L. The data used to develop
this health alert came from a range of Australian toxicological studies, one of which provided
sub-chronic oral doses of the toxin to mice and demonstrated responses to the toxins after an
extended trial of 11 weeks (Humpage and Falconer 2002). The recommendation of a health
alert acknowledges that health authorities should be notified if blooms of Cylindrospermopsis
raciborskii or other producers of this toxin are present in water supplies.
Given that data pertinent to drinking water produced from the sea are very rare, it will take
some time to establish guidelines for desalination. The process of establishing ‘Health Alert’
values could also be applied to relevant common seawater toxins or toxin families such as the
brevetoxins, saxitoxins, okadaic acid, and domoic acid.
Further considerations to assist in establishing meaningful guidelines for HAB toxins in
seawater desalination are the guidelines promulgated for shellfish harvesting based on acute
toxicity following ingestion. Guidelines for toxins derived from marine HABs exist in
Australia, New Zealand, and many other countries. The former are described by the
Australian and New Zealand Food Authorities (ANZFA) (Todd 2011). ANZFA sets the
following guideline values: paralytic shellfish poisoning (PSP or the saxitoxin family)
80 μg/100 g of edible shellfish, amnesic shellfish poisoning (ASP or domoic acid) 20 ppm,
and diarrhetic shellfish poisoning (DSP okadaic acid and other toxins) 20 μg/100 g
(Discussed further in Chapter 2). Although it is difficult to apply the latter values to drinking
water, one could consider the minimum mass of shellfish consumed in a meal, and then
compare that with a minimum volume of water consumed. For example, considering PSP,
consuming a minimum of 10 g of shellfish would correspond to ingestion of 8 µg of saxitoxin.
If one were to consume 100 mL of water containing the same 8 µg of saxitoxin, this would
correspond to a concentration of 80 µg/L: however, guidelines for saxitoxins worldwide
range from 1.0 to 3.0 µg/L, which suggests that it is far more prudent to work to these lower
values when producing drinking water. The oral LD50 for humans is 5.7 µg/kg, so the average
80 kg person would need to ingest 456 µg of saxitoxin for the dose to be fatal. One would
need to drink 5.7 L of water if guidelines for drinking water were based on shellfish
guidelines or 456 L of water at the New Zealand drinking water guideline. Considering that
small children are more susceptible to the toxin, as well as the fact that illness is common at
far lower saxitoxin doses, 1.0 µg/L is a far more reliable guideline.
227
8.2! USING GUIDELINE VALUES
The guideline value is important for water supply authorities, as this value sets the
concentration of toxin that is tolerable in drinking water, i.e. “at the tap”. For example,
ADWG for cyanobacterial toxins in fresh or brackish waters are not mandatory standards;
however, they provide a basis for determining the quality of water to be supplied to
consumers in all parts of Australia. In this circumstance the guideline level is effectively a
recommendation from the health authorities, although this situation is changing with the
introduction of more prescriptive drinking water standards in some jurisdictions. For some
water authorities in Australia the guidelines/standards become part of the de facto contractual
standards. They are therefore required to comply with the guideline values as part of their
standards of service.
For other countries the guideline level can be a standard that must be met and compliance
monitoring may be required.
8.3! AUSTRALIAN DRINKING WATER GUIDELINES REGARDING MULTIPLE TREATMENT
BARRIERS
8.3.1 What is the multiple barrier concept?

In the absence of guideline values for HAB toxins internationally, the best method for
ensuring that toxin does not enter the drinking water distribution system is by employing the
multiple barrier approach to treatment, which is well described in the Australian Drinking
Water Guidelines (NHMRC/NRMMC 2011). The concept, which is widely recognized as the
most effective approach for producing safe drinking water, aims to improve a water treatment
train’s reliability by employing multiple treatment processes in series. Providing multiple
treatment processes, or barriers, ensures that when one process fails or partially fails, the
performance reduction can be compensated for by other processes in the treatment train.
Ideally, all treatment barriers should be operating effectively to ensure that a water treatment
plant is performing optimally. Since it is likely that individual barriers are not effective 100
percent of the time, however (NHMRC/NRMMC 2011), employing multiple barriers against
contaminants ensures that produced drinking water remains of a high quality, even if one of
the treatment barriers becomes compromised. In addition, using multiple barriers expands the
range of contaminants removed by a water treatment train, also referred to as robustness,
since individual barriers typically target different groups of contaminants (Crittenden et al,
2005). The number of barriers required, as well as the types of barriers employed, depend
largely on how acute the source water contaminants are, with more acute contaminants
requiring more barriers (Crittenden et al. 2005). In addition, understanding both the
challenges that contaminants may present, as well as any vulnerabilities associated with the
treatment barriers, plays an important role in barrier selection (NHMRC/NRMMC 2011).
Although the multiple barrier concept places a great deal of emphasis on treatment processes,
other barriers that help ensure safe drinking water are equally important, such as source water
selection and protection, as well as safe water transportation and storage (Bixio and Wintgens
2006). ADWGs recognize the importance of these barriers, stressing that “understanding a
water supply system from catchment to consumer, how it works, and its vulnerabilities to
failure” are an integral part of the multiple barrier concept (NHMRC/NRMMC 2011).
Furthermore, ADWGs highlight the importance of including “mechanisms or fail-safes to
accommodate inevitable human errors without allowing major failures to occur” when
applying the multiple barrier concept (NHMRC/NRMMC 2011).
228
Seawater reverse osmosis (SWRO) plants are inherently a multi barrier treatment process for
certain feedwater contaminants. As discussed in Chapter 9, pre-treatment processes provide
an added barrier for toxin removal in addition to the core reverse osmosis (RO) treatment.
This allows intracellular toxin removal upstream of the RO step and thus less dependence on
this process as a sole toxin removal mechanism. In addition, many plants that commonly
experience HAB blooms are in the Gulf and have a full or partial second pass. Chlorination
of treated waters in the distribution system provides an added barrier for some toxins such as
saxitoxin (Laycock et al. 2012).
8.3.2 Monitoring as a barrier
Once barriers are in place to protect public health, regulatory bodies need to ensure that those
barriers are performing effectively. To this end, ADWGs identify drinking water quality
monitoring as the ultimate test to confirm whether employed treatment barriers, or other
preventative measures, are effectively protecting the public from consuming unsafe drinking
water (NHMRC/NRMMC 2011). Additionally, ADWGs emphasize that drinking water
quality monitoring should be performed on a frequent basis, so as to identify the failure of
any treatment processes, or other measures, in a timely manner (NHMRC/NRMMC 2011).
In addition to monitoring drinking water quality, operational monitoring of treatment
processes plays an important role in ensuring that safe water is delivered to consumers (refer
to Section 8.4.3 below – HACCP). According to ADWGs, operational monitoring aims to
ensure that all existing treatment barriers, and other water safety measures, are performing as
desired (NHMRC/NRMMC 2011). For example, monitoring of RO permeate for
conductivity can be used as a surrogate for failure of the RO (see Chapter 10). As with
drinking water quality monitoring, operational monitoring should be performed regularly so
that any detected process deficiencies can be corrected promptly. To this end, continuous,
online monitoring is the most desirable form of operational monitoring, and should be
implemented wherever practical (NHMRC/NRMMC 2011). This ensures that ineffective
treatment barriers, such as a damaged membrane, can be recognized and addressed
immediately, preventing unsafe drinking water from reaching the public. In the event of an
emergency, such as the discovery of a ‘bloom’ event in the source water, monitoring
frequency should be increased to ensure that barriers are effectively protecting consumers.
Validation monitoring is also vital to protecting the public from consuming unsafe drinking
water. Unlike operational monitoring, validation monitoring aims to assess newly added
treatment barriers, and ensure that they are performing effectively. For example, if a new
activated carbon column were to be added to a water treatment train, the column’s effluent
would need to be tested to confirm whether or not contaminants, such as toxins, are
continually removed over time. Validation monitoring is particularly important when adding
treatment processes that have not been implemented onsite previously, and have not been
tested by the manufacturer (NHMRC/NRMMC 2011). In such cases, validation monitoring
gives plant operators a chance to verify any assumptions regarding process performance, and
to ensure that all desired contaminants are effectively removed. Where a treatment process
has been tested by the manufacturer (i.e., pre-validated), validation monitoring should still be
performed; however, the treatment process need not necessarily be tested before it is brought
online in the water treatment plant, assuming that onsite water characteristics are not
substantially different from those used in manufacturer testing (NHMRC/NRMMC 2011).
All forms of monitoring play a significant role in managing risk in a waters supply system.
As a result, a monitoring strategy requires careful consideration of numerous factors,
including which variables should be monitored, what data should be collected, and how
collected data should be used to mitigate risk (NHMRC/NRMMC 2011). Furthermore, it is
229
important that a monitoring strategy be developed with mechanisms in place to incorporate

monitoring information into decision-making processes affecting the water supply system
(NHMRC/NRMMC 2011). Finally, for all forms of monitoring, it is essential that monitoring
data are compiled, analyzed, and reported in a timely manner to ensure that corrective
measures can be taken before unsafe drinking water is delivered to the public
(NHMRC/NRMMC 2011).
In summary, by checking that all treatment processes and other barriers are performing
effectively, monitoring in and of itself represents an additional barrier, preventing unsafe
drinking water from reaching consumers. Effective implementation of the multiple barrier
concept thus requires integrating an effective monitoring program with a reliable, robust
treatment train, along with sufficient preventative measures.
8.3.3 Multiple barriers in the context of HAB toxins
Employing multiple barriers against HAB toxins is a high priority, particularly due to their
acute toxicity. Such barriers can be comprised of a wide range of treatment processes,
preventative measures, and monitoring strategies.
In terms of treatment barriers, numerous processes have been developed to remove HAB
biomass and their toxins from drinking water (Chapters 9,10). These aim to remove HAB
cells (which can include cell-bound toxins), and can also remove dissolved (i.e.,
extracellular) toxins, depending on the treatment technology considered. For example,
treatment using coagulation and dual media filtration removes a significant portion of cell-
bound toxins from drinking water (Hrudey et al. 1999). Dissolved toxins, on the other hand,
are not effectively removed using pre-treatment processes and are better targeted using
reverse osmosis, oxidants, such as chlorine for some toxins, or activated carbon adsorption
(Hrudey et al. 1999; Laycock et al. 2012). Where multiple treatment processes are employed,
process interactions must be considered, since upstream treatment processes may result in
cell damage, which leads to significantly higher levels of dissolved toxins. For example,
chlorination of the intake will result in cell lysis and therefore higher concentration of
dissolved toxins. Therefore, treatment processes leading to cell damage are not recommended
unless a subsequent process is included to remove dissolved toxins (Newcombe et al. 2010).
A comprehensive monitoring strategy serves as the final barrier between HAB toxins and
consumers. Monitoring toxins can, however, be challenging, since instrument-based
analytical methods used to measure toxins are time expensive and consuming, required limits
of detection are extremely low, and analytical standards used to quantify many toxins are
lacking (WHO 2004). Consequently, analyzing toxins using sophisticated instruments such as
liquid chromatography/mass spectroscopy (LC/MS) and high performance liquid
chromatography (HPLC) is not likely for routine monitoring in desalination plants. On the
other hand, source water monitoring using rapid screening methods for toxins (see Appendix
2) presents a practical alternative. Steps might include visually checking source water for
indications that a HAB has occurred or may be imminent, or sampling of source waters to
screen for potentially toxic or harmful species (Chapter 3).
8.4! RISK ASSESSMENT FOR THE PRESENCE OF HABS
Risk assessment is the process of using available information to predict how often identified
hazards or events may occur and the magnitude of their consequences. Risk can be assessed
at two levels: maximum risk in the absence of preventative measures and residual risk after
consideration of existing preventative measures (Nadebaum et al. 2004).
230
8.4.1 Risk assessment

Formal risk assessment is often carried out under the auspices of official government
agencies. Such risk assessment also typically draws on knowledge from a group of scientists,
including chemists, toxicologists and statisticians. Risk evaluation is considered an iterative
process and includes the following steps:
(i)! Public call for data
(ii)! Hazard identification
(iii)! Exposure assessment
(iv)! Hazard characterization
(v)! Review of the status of detection and quantification methods
(vi)! Review of toxicological and epidemiological data
(vii)! Risk characterization
(viii)! Recommendations for management and monitoring
In the case of new or emerging risks, it is common that an initial risk assessment is carried
out, leading to a provisional management procedure. After some time, the effectiveness of
this procedure is then reviewed and, possibly in combination with new toxicology and
occurrence data, a new management procedure or trigger level may be put in place. Thus, risk
assessment should be considered to be an iterative process.
Based on previous risk assessments and current knowledge of HAB occurrence, it is clear
that algal toxins should be considered as an identified hazard in drinking water derived from
seawater (Caron et al. 2010; Laycock et al. 2012; Boerlage and Nada, 2014; Berdalet et al.
2015). Thus, the main objective of this section is to discuss potential exposure and assess risk.
For several toxin groups described above, such risk assessments have been carried out by the
European Food Safety Authority (EFSA 2008a, b, 2009a, b, c, d, e, f, g, 2010a, b, c, d; EFSA
(EFSA Panel on Biological Hazards (BIOHAZ) 2012; FSAI 2006; Lawrence et al. 2011), in
particular for exposure through the consumption of bivalve molluscs. Note that some
discrepancies between current legislation and risk evaluation for several toxins groups in
shellfish have been established in recent risk evaluation exercises by EFSA, as shown in the
summary report (Table 8.2). For drinking water, so far only microcystin-LR has been
formalized and regulated as a hazard by WHO (2004). This guide has also established 2 L as
a maximum quantity of water that can be consumed as one portion in a day.
The maximum amount of a toxin or contaminant that a human may be exposed to without ill
effects is referred to as the ARfD. ARfDs or safe doses have only been established for four
marine toxins (saxitoxin, domoic acid, okadaic acid, and azaspiracids by EFSA) and for
microcystin-LR (indirectly by WHO). Table 8.2 shows the current European Union (EU)
limits for selected algal toxins in shellfish meat, ARfD and recommended maximum levels
based on a representative portion of shellfish consumed. As not many reports are available
for the occurrence of toxins in seawater, it was not possible to make probabilistic exposure
assessments, so worst case assumptions of these five toxin groups used for assessment are
provided here. The data used were collated from laboratory culture data on toxic HAB
species described in Chapter 1, as well as from the studies described in Chapter 10 showing
99% rejection of dissolved toxins. The rationale for this approach is given below.
To assess risk, the amount of toxin that might be retained in treated water needs to be
compared to the ARfD or safe dose values. The two components to be considered in this
regard are the extracellular and intracellular toxins. Some studies report measurements of
both categories of toxin in seawater, but most do not. For saxitoxin, a very polar compound,
231
extracellular levels ranged from 12 to 31 µg/L in Alexandrium cultures studied by Lefebvre et

al. (2008), who also reported a maximum extracellular STX level of 0.8 µg/L during an
Alexandrium bloom in the field. Previous culture studies of cyanobacteria and dinoflagellates
have reported extracellular STX levels of 50 to 75 µg/L (Hsieh et al. 2000; Velzeboer et al.
2001). For domoic acid, a toxin actively excreted by diatoms, a dissolved concentration of
3.3 µg/L has been reported (Liefer et al. 2013; Trainer et al. 2009), but much higher levels
have also been observed. Kudela (pers. comm.) measured total and extracellular
concentrations of 100 and 50 µg/L domoic acid respectively during a massive Pseudo-
nitzschia bloom along the US west coast in 2014. Half of the measured toxin was
extracellular. One study estimated extracellular okadaic acid concentrations during blooms of
0.2 µg/L (Takahashi et al. 2007), whereas another study measured only low background
levels of okadaic acid in a coastal lagoon (Zendong et al. 2015). In contrast, Smith et al.
(2012) reported that 60% of the okadaic acid produced by Dinophysis cultures was
extracellular. Microcystin concentrations from terrestrial runoff in coastal water have been
recently estimated to be around 1.4 µg/L (Kudela 2011; Miller et al. 2010).
These are just a few of the many studies that report toxicity values for HAB species cultures
or for natural waters during blooms, and clearly there are many differences in the amounts of
toxin that might be in intra- versus extracellular form. Given the large variability in these
published data from different locations and conditions, the toxicity differences between
strains of a given species, the variation in toxicity that occurs with different forms of nutrient
limitation or environmental stress, and the different levels of population biomass achieved
under varying growth conditions (see Chapter 1), simplifying assumptions are needed to
constrain HAB risk assessments. Here a hypothetical approach is taken, in which laboratory-
derived cellular toxicity data for the most toxic species producing the major HAB toxins are
attributed to a dense bloom of those species. In Chapter 1 (Table 1.4), these calculations were
performed for three possible bloom types – 100,000, 500,000, and 5,000,000 cells/L,
representing moderate, large, and extreme HABs for most species. For this conservative
analysis, the extreme bloom level is used to estimate maximum exposure levels and safety
factors for treated water. For a bloom of this size, the highest calculated total toxin levels for
saxitoxin, domoic acid, okadaic acid, and azaspiracid are 600, 335, 294, and 0.1 µg/L
respectively (calculated in Chapter 2).
Chapters 2, 9, and 10 discuss the relative ease of removal of intracellular versus extracellular
toxins. A variety of pretreatment steps like ultrafiltration and dissolved air flotation (DAF)
will remove cells, and thus intracellular toxins, but they are not very effective in removing
the dissolved or extracellular fractions. It is clear that on occasion, 50% or more of the toxin
produced by HABs can be extracellular, and to this, one would add the toxin that is released
by physical and chemical processes during pretreatment. Again, in order to err on the
conservative side, it will be assumed that 100% of the toxin from a bloom of 5,000,000
cells/L will be dissolved and will reach RO membranes, where 99% removal will occur in a
single pass. Results are tabulated in Table 8.3. A similar calculation could be made for
thermal desalination systems, in which most cases there is no pretreatment and HAB cells
will be degraded by temperature and pressure, releasing toxin (Boerlage and Nada 2014). A
toxin removal of 99% in thermal systems could also be assumed for this calculation (toxin
removal efficiencies are discussed in Chapter 10).
232
Table 8.2. Current European Union (EU) limits for selected algal toxins in shellfish meat (= total flesh), acute reference doses (ARfD) and
recommended maximum levels based on a 400 g portion of shellfish consumed (adapted from (EFSA, 2009f)).
Exposure
Current EU Exposure by eating a 400 g Acute
Exposure by eating a corresponding to Ratio
limits in portion at the 95th percentile of reference Derived conc. in
Toxin group 400 g portion at the the ARfD by eating of
shellfish the concentrations in samples dose shellfish meat (B)
EU limit c a 400 g portion c B/A
meat (SM) a currently on the EU market (ARfDe)
Okadaic acid
160 µg OA 64 µg OA eq/person 96 µg OA eq/person 0.3 µg OA 45 µg OA eq/kg
(OA) and 18 µg OA eq/person 0.28
eq/kg SMa (1 µg OA eq/kg b.w.) (1.6 µg OA eq/kg b.w). eq/kg b.w. SM
analogues
64 µg AZA1
0.2 µg
Azaspiracids 160 µg AZA eq/person 16 µg AZA1 eq/person 12 µg AZA1 30 µg AZA1
AZA1 0.19
(AZA) eqc/kg SM (1 µg AZA1 eq/kg (0.3 µg AZA1 eq/kg b.w.) eq/person eq/kg SM
eq/kg b.w
b.w.)
320 µg STX eq/person
Saxitoxins 800 µg STX < 260 µg STX eq/person 0.5 µg STX 75 µg STX eq/kg
(5.3 µg STX eq/kg 30 µg STX eq/person 0.09
(STX) eq/kg SMb (< 4.3 µg STX eq/kg b.w.) eq/kg b.w SM
b.w.)
8 mg DA/person
Domoic acid 20 mg DA/kg 1 mg DA/person 30 µg
(130 µg DAd/kg b.w) 1.8 mg DA/person 4.5 mg DA/kg SM 0.23
(DA) SM (17 µg DA/kg b.w) DA/kg b.w
a"SM whole uncooked shellfish meat, For OA, dinophysistoxins and PTX, current regulation specifies a combination; however the CONTAM Panel concluded that PTX should be
considered separately."
b"eq. = equivalents: In the Commission Regulation (EC) No 853/2004 a limit value of 800 µg PSP/kg SM is given. In the EFSA opinion, the CONTAM Panel adopted this figure as
being expressed as µg STX equivalents/kg SM.

c"the 400 g portion has been derived from the European consumption databases, it represents the 95 %ile; The CONTAM Panel assumed that AZA equivalent should refer to AZA1
equivalents."
d"Applies to the sum of DA and epi-DA."
e
ARfD = Acute reference dose"
Table 8.3. Acute risk of poisoning from drinking desalinated water assuming 100% of the
toxin in a major HAB is dissolved or extracellular.
c
Toxin Theoretical MPL Extracellular Safety Safety factor Concentration to
safe dose a (µg/L) toxin factor, raw after 99% detect in drinking
(µg) concentration, intake rejection f water [µg/L]
g
(µg/L)
d
water e
Saxitoxin 30 15 600 0.025 2.5 0.15
Domoic acid 1,800 900 335 2.7 270 90
Okadaic acid 18 9 577 0.016 16 0.9
(and DTXs)
Azaspiracid 12 6 0.1 60 6000 0.6
b
Microcystin 2 1 1.4 0.7 71 0.1
a
The theoretical safe dose is calculated based on the ARfD by EFSA (Summary Opinion, 2009); it refers
to a 60kg person.
b
Safe dose for microcystin separately derived from WHO guideline. Extracellular concentration
estimated from Kudela 2011; Miller et al. 2010.
c
MPL = Maximum permissible level; based on a consumption of a 2-L dose of desalinated water.
d
Assumes that 100% of the total toxin from a dense bloom (5,000,000 cells/L) of the most toxic species
producing each toxin is extracellular.
e
The safety factor is the ratio of the MPL to the extracellular concentrations of each toxin. Safety factors
greater than 1 indicate a safety margin. Safety factors below 1 indicate a higher risk.
f
This assumes 99% rejection by the SWRO and thus considers only a single pass and no problems with
the membranes.
g
This concentration is arbitrarily set at 1/10th of the MPL or maximum permissible level
To explain the calculations in Table 8.3, a focus on saxitoxin is offered. The safe dose of
STX that may be consumed (30 µg) has been established from a number of cases reported for
shellfish poisoning (Table 8.2). Given the 2-L quantity of daily drinking water established by
WHO, a maximum permissible level of STX in drinking water of ca. 15 µg/L (15000 ng/L)
can be derived. (Note that some national guidelines have established lower, more
conservative health alert guide for STX based on discrepancies in safety factors to convert a
lowest observed adverse effect levels (LOAEL) to a NOAEL). If 100% of the toxin produced
by a 5,000,000 cells/L bloom of a saxitoxin-producing species were to occur near a
desalination plant, a safety factor of 2.5 can be calculated assuming that 100% of the toxin is
extracellular, and that there is 99% toxin removal during treatment. Risk would thus be
considered moderate. (Safety factors greater than 1 indicate a safety margin, safety factors
below 1 indicate a high risk). A similar moderate-risk safety factor is calculated for okadaic
acid, with other toxins at much safer levels. Given that many RO plants are two-pass systems,
the calculated residual risk will all drop dramatically (~100-fold if the system is full two pass,
care should be taken with partial two pass systems) after full treatment, giving some
reassurance on the risks posed by algal toxins in desalinated drinking water; however, the
uncertainty associated with these safety factor calculations is large.
The final column in Table 8.3 indicates the analytical levels of detection needed in order to
detect the maximum permissible level (MPL) of these toxins in treated water. This can be
helpful to plant operators in choosing analytical or screening
8.4.2 Assessing the likelihood of HABs in a water source
Information on the importance and interrelationship of environmental variables has been used
in a range of ways to determine the likelihood of the growth of HABs and the development of
blooms in particular areas. A range of approaches has been used to undertake risk
234
assessments and these have been variously termed ‘susceptibility’ or ‘vulnerability’

assessments. For marine HABs, there have been assessments prior to the construction of
plants as part of environmental impact statements and design risk assessments, but typically,
these will be based on site-specific factors such as the prior history of HAB occurrence or the
total phosphorous and nitrogen concentrations in the region. Another factor to consider is the
transport pathway that might bring established blooms from elsewhere to the vicinity of a
desalination plant. There is, however, no specific standard protocol or framework for risk
assessments for HABS and desalination similar to those that exist for freshwater
cyanobacteria. Each phytoplankton species has a different set of favorable conditions that
promotes its growth and reproduction, and many different species can potentially cause
problems with desalination operations (see Chapter 1). The two most important nutrients for
phytoplankton growth are the elements nitrogen (N) and phosphorus (P), which are found
naturally in aquatic environments in various concentrations. In freshwater, it has been
possible to identify concentration levels at which the risk of cyanobacterial blooms is rare,
infrequent, occasional, frequent and persistent, and frequent and persistent/strong. This risk
assessment approach was developed in Australia to assess water bodies for their
susceptibility to cyanobacterial contamination and is found in the National Health and
Medical Research Council (Australia) ‘Guidelines for Managing Risks in Recreational Water’
(NHMRC 2006). This technique is based upon empirical observations in Australian
reservoirs and from the range of literature studies on the variables influencing cyanobacterial
growth. No such compilation exists for marine algae.
8.4.3 Hazard Analysis and Critical Control Point (HACCP)
The fourth edition of the World Health Organization (WHO) drinking water guidelines
(WHO, 2011) advocates a risk management approach to water quality based on the multi-
barrier approach and the development of Water Safety Plans based on the Hazard Analysis
and Critical Control Point (HACCP) methodology, to assure safe drinking water. The
HACCP system was developed for NASA in 1959 to ensure food safety for their manned
space program as it was recognized that relying solely on end point testing was inadequate.
Instead HACCP was developed to take a systematic preventative approach to addresses
hazards through anticipation and prevention. Subsequently, HACCP evolved into a
universally accepted risk management framework by the Codex Alimentarius Commission
(Codex, 2003) established by the WHO and the United Nations Food and Agricultural
Organization for the implementation of their food standards program. Since its codification,
HACCP has been widely used in the food and beverage industry as a quality assurance
management strategy.
Havelaar (1994) discussed the application of the HACCP methodology for drinking water
supply. The HACCP system is now increasingly being incorporated into national drinking
water guidelines and regulations to assure safe drinking water. In the drinking water context,
Critical Control Points (CCP) are defined as process steps at which control can be applied
and are essential in preventing or eliminating the water quality health hazard or reducing it to
an acceptable level (WHO). For example, in surface water treatment, coagulation and media
filtration may be defined as CCPs for the removal of Giardia. In drinking water treatment
when the term ‘barrier’ refers to a CCP, it can be considered a critical barrier.
Due to the potency of the marine algal toxins, notably saxitoxin, there are clearly identifiable
microbiological hazards in seawater desalination with a potential detrimental impact on
human health. Boerlage and Nada (2014) described the use of HACCP in examining the fate
of marine toxins in desalination plants and the potential (residual) risk in desalinated drinking
water. The latter is also addressed in Chapter 9. The application of the HACCP framework
235
for controlling marine toxins in SWRO and thermal desalination plants and preventing the
risk of failure of CCP that may lead to toxins in the desalinated water is further described
below.
Within the HACCP framework defined by Codex Alimentarius, five preliminary steps are
detailed followed by the seven principles of HACCP, summarized in Figure 8.1. In
conducting a HACCP assessment for a seawater desalination plant, a multidisciplinary team
should be assembled covering a range of expertise including specialists in some of the
following areas; water quality, public health, desalination process engineers, plant operators,
maintenance managers, risk assessment specialists trained in HACCP, representatives from
the plant owner. In areas where HAB are prevalent, a HAB specialist would be recommended.
The team would be tasked to develop, verify and implement the HACCP plan. Ideally, the
team would meet prior to the detailed design process and therefore include designers. A
follow up workshop during commissioning and operation would help to continuously develop
and improve the HACCP plan.
Figure 8.1. HACCP Framework comprising five preparatory steps and the seven principles of HACCP.
The drinking water (product) would then be described, examining the source raw water,
specifications for chemicals used, treatment, storage and any standards the drinking water
must meet – i.e. plant water quality specifications, national and/or international drinking
water guidelines and its intended use – potable, industrial etc. For seawater desalination
plants, the source seawater may have been characterized in a seawater quality study as
described in Chapter 5 where the potential for HABs to occur would have been investigated.
In the final preparatory step, a process flow diagram (PFD) for the SWRO or thermal
desalination plant would be developed which is critical in proceeding through the formal
HACCP process. The process flow diagram would document the incoming raw seawater
catchment and all entry pathways into the process that could lead to contamination (e.g.,
coagulant addition, storages, unit process steps to treat the seawater and produce desalinated
drinking water, and waste treatment processes). The PFD needs to be verified as an accurate
236
representation of the plant treatment process and modified as required during the life of the
plant to document any changes.
In developing a HACCP plan, all potential chemical, physical and microbiological hazards
and hazardous events which may impact on human health are identified, their fate in the plant
process is determined using the PFD as a reference point and the hazards are inventoried.
Focusing on a toxic marine algal bloom event and the release of toxins (a microbiological
hazard), the risk associated with this event would be evaluated in the source water with
controls and no controls in place through a typical risk assessment framework. Assessing the
risk with no controls in place is essentially drinking seawater, which is highly unlikely,
(carrying with it its own risk), but it allows the maximum risk to be determined.
The assessment would semi-quantitatively rate the likelihood of the occurrence of toxins in
seawater based on categories, ranging from rare to almost certain, and their consequences,
ranging from insignificant to catastrophic, each with a score increasing in severity from 1-5.
A score of 1 for likelihood could be defined as a toxic HAB only occurring in ‘exceptional
circumstances’ (at <1% or once in a hundred years), where ‘almost certain’ might represent a
chance of >95%. A typical risk assessment matrix is shown in Table 8.4. The likelihood and
consequences of a water quality hazard can be subjective and interpreted differently. In such
a subjective case, the worse case scenario should be assumed.
As discussed above and in Boerlage and Nada (2014), it remains difficult to predict both the
likelihood of marine toxins occurring at a desalination plant intake and then being entrained
and thus their potential health consequences. Further, even if a bloom is detected at the intake
by online monitoring of water quality parameters (Chapter 5), or through remote satellite
sensing (Chapter 4), these methods cannot discriminate between toxic and non-toxic HABs.
Algal identification, enumeration, and toxin analyses (Chapters 2, 3 and Appendix 2) would
be required to confirm the toxicity of the bloom. Assessing the health consequences of a toxic
bloom with or without controls in place is even more difficult, as the potency of each marine
toxin varies. In addition, there are few water quality guidelines for marine toxins (see Section
8.1). Hence, it is difficult for an operator to assess human health consequences.
Consequences considered in addition to human health in such risk assessments would
typically include public perception, commercial, reputation, financial loss and legal. Other
hazards and/or hazardous events for a desalination plant would be assessed, rated and risks
prioritized for further evaluation. The net result of a water supply risk assessment would
likely classify marine toxins as “moderate to very high” and a significant risk prior to
treatment for plant operation and would therefore assigned priority for examination.
Table 8.4. Matrix scoring risk factors as low (L), moderate (M), high (H) and very high(VH).
Consequences
Insignificant Minor Moderate Major Catastrophic

(1) (2) (3) (4) (5)
Rare (1) H H VH VH VH
Likelihood
Unlikely (2) M H H VH VH
Possible (3) L M H VH VH
Likely (4) L L M H VH
Almost certain (5) L L M H VH
In the second part of the risk assessment, the residual risk of marine toxins in desalinated
drinking water with barriers (control measures) in place and the adequacy of these barriers
237
are determined. SWRO desalination plants are typically multi-barrier schemes, and there are
one or more barriers for algal cells and intracellular toxins (e.g., coagulation and granular
media filtration) to reduce the risk. The efficiency of intracellular toxin removal and these
barriers depends on: i) the effect of process parameters that may rupture the cell wall, such as
mechanical shear, pressurization, or chemical oxidation (chlorination); and ii) the inherent
nature and strength of the algal cell wall. Whereas, thermal plants typically only have
screening prior to the multi-stage flash distillation (MSF) and multiple-effect distillation
(MED) processes, and hence a single barrier. Should the HACCP team consider the residual
risk as unacceptable, additional barriers would be required to mitigate the residual risk.
Critical control points are then identified amongst the barriers, defined, as above, where
control can be applied at a step and which is essential in eliminating marine toxins or
reducing their concentration to an acceptable level. Critical control points can be a process
step, or a point in the system or procedure. Codex provides a decision tree approach (not
mandatory) to consistently and logically identify CCPs through a sequence of questions
shown in Figure 8.2.
Figure 8.2. Decision tree to identify critical control points (CCPs). Modified from Codex 2003.
238
None of the pretreatment barriers in a typical SWRO desalination plant scheme (i.e.,
conventional pretreatment coagulation and flocculation, granular media filtration followed by
cartridge filtration or ultrafiltration membrane pretreatment) would be deemed as critical
control points when answering the four questions in the Codex decision tree. While these
barriers may remove intact algal cells (and thus the intracellular toxins), none would act as a
reliable barrier to remove extracellular algal toxins, nor meet the definition of a CCP.
Hence, following the decision tree approach, the RO step is the first critical control point in
SWRO plants for removal of extracellular algal toxins. If a SWRO desalination plant
operates a second-pass brackish water RO system to desalinate the water further, then this
would be classed as the second critical control point. Similarly, for MSF and MED
desalination plants, the thermal desalination step would be the critical control point to prevent
algal toxins from contaminating drinking water.
Once CCPs have been determined, the HACCP framework requires the establishment of
critical limits and monitoring to assess whether the CCPS are under control. Critical limits
are defined as criteria that separate acceptability from unacceptability (i.e., safe and unsafe
operating conditions at a CCP). Maintaining operation within the critical limits demonstrates
a CCP is working and ensures safe drinking water. Breaches in a critical limit allow failures
of a CCP to be detected which would result in an increase in the residual risk and trigger
documented corrective actions, operating procedures etc. to bring the CCP back under
control. Alerts can be set which are more conservative than critical limits, thereby allowing a
timely response to rectify deviations prior to the critical limit being breached. Ideally,
monitoring of the CCP and its critical limits should be continuous, using online
measurements to allow monitoring of the CCP in real time and a rapid response by the duty
operator in the control room. Further, parameters tending towards loss of control can be
detected and provide an accurate record for verification.
Ensuring the integrity of both SWRO membranes and thermal desalination processes using
critical control points is paramount for the removal of marine toxins. Integrity loss for SWRO
membranes may occur due to membrane failure, defective interconnections, O-rings, etc.
Membrane integrity would be continuously monitored in SWRO plants by conductivity as a
surrogate for salinity to detect any leakage. Alert and critical limits, based on conductivity
would be defined specific to a plant, its design, level of instrumentation and treated water
storage. For example for a single pass system treating salinity of 35, an alert may be triggered
when the conductivity of the RO permeate from a single RO train exceeds 400 µS/cm for
more than 2 hours and an alarm when the combined RO permeate conductivity from several
trains exceeds 400 µS/cm for more than 2 hours.
Similarly, in MSF and MED systems direct carry over from the seawater to the distillate must
be limited to the greatest extent possible. Seawater and potential algal toxin carry over may
occur due to possible joint leakage or tube failure allowing bypassing of the separation
process or the displacement of demister pads (used to separate the entrained brine liquid
droplets entrained with the vapor and to allow the vapor to pass through the mesh). The
integrity of the tubes and joints can be checked and confirmed by hydro testing; however,
failures are readily identified by rapid increases in distillate conductivity as the rejection of
salts was demonstrated to be similar to marine toxins in the work of Laycock et al. (2012).
Should the conductivity increase in either a SWRO desalination or thermal plant approach,
the alert, or critical limits, documented corrective actions would be taken in line with the fifth
principle of the HACCP framework. Corrective actions would be established aimed at
identifying and eliminating the cause of the deviation in conductivity to bring the CCP back
under control. No unsafe desalinated water would be delivered should the CCP be out of
239
control. With duty operators appropriately trained in HACCP, the processes and instruments
would be directed to documented operating procedures associated with each alert and alarm.
For a SWRO plant, corrective measures could include verification that RO permeate
conductivity instruments were reading accurately using a handheld conductivity meter.
Corrective measures could also include checking differential pressure, recovery and
normalized flows. Should the conductivity reading be correct then conductivity profiling of
RO modules may be amongst the corrective measures outlined and/or the RO permeate could
be rejected and isolated and the RO train could be shut down. In the case of a prolonged HAB
event, the alert and alarm levels could be reduced further to ensure a higher salt removal
efficiency and corresponding removal of toxins. Corrective actions and reporting of CCP
alarms may be integrated into the plant’s incident and response plans.
The penultimate HACCP principle is to establish procedures to verify the HACCP system is
working effectively. This encompasses a range of tools from internal and external auditing of
procedures by qualified third parties, routine calibration checks of online CCP
instrumentation (conductivity), operational checks, random sampling and analysis, recording
CCP deviations, corrective actions, and follow up actions (Codex 2003).
The final HACCP principle requires accurate record keeping of HACCP documentation
sufficient to verify that the HACCP controls are in place and being maintained. These include
recording a hazard register, the risk assessment, determination of CCP, CCP monitoring
activities, deviations, and associated corrective actions such as verification procedures
performed and importantly modifications to the HACCP plan (Codex 2003). The HACCP
plan would be integrated into the plant’s operating systems.
8.5! ALERT LEVEL FRAMEWORKS
An ‘Alert Level Framework’ (ALF) is a monitoring and action sequence that operators can
use for a graduated response to the onset and progress of HABs in a seawater source. All
HABs should be treated with caution until the absence of toxicity in the treated water is
confirmed, or advice based upon past local knowledge indicates the absence of hazard.
The ALF described here is a generic model; however, it is possible to translate the format for
monitoring and management of HABs in waters used for other purposes such as recreation
and agriculture. The level thresholds, indicators, and actions for each specific area or plant
would be different and would need to be developed based upon appropriate local guidelines
and risk assessment procedures. The following framework acts as an example and basis for
further development by specific plant operations staff.
The ALF is a situation assessment tool based upon data from cell counts and other
observations that can be used in conjunction with the relevant drinking water guidelines for
toxins to assess the potential hazard from a bloom. The rationale for the use of cell counts to
prompt management actions is that, for most practical purposes, cell counting is the fastest
and most relevant way to detect algal-related water quality problems (Chapter 3). This is
because cell counting is not too difficult for plant staff to undertake, and provides relatively
rapid and cost-effective information. By contrast, reliable biotoxin testing is still sufficiently
complex such that most desalination plants will not have a local capability. If samples are
sent out for analyses, there can also be a slow turn-around time for results. Note, however,
that highly sensitive antibody-based kits are now available for many of the algal toxins (see
Appendix 2), so it is possible to do preliminary screening on site at a plant, with samples
showing positive results being sent off site to qualified analytical facilities for confirmation.
The cell counts are regarded as an indicator or "surrogate" for a potential toxin or
organic/fouling hazard. They do not replace toxin or organic carbon analyses, which are
240
required for health risk assessment or for pre-treatment decisions, but rather are used as
relatively conservative triggers in the management plan. The counts can be used to prompt
toxin or biomass screening, which can then be assessed to determine the hazard and risk to
plant operations. Cell counts are directly useful for determining the optimal operation of the
plant and for implementing mitigation strategies to accommodate the higher volumes of
biomass. If recorded properly and archived, they also provide useful information to guide
future operational decisions on pretreatment strategies.
The framework given below is developed from the perspective of the plant operator. The
circumstances and operational alternatives for use with the ALF will vary depending upon the
source of supply and water treatment barriers available. The associated monitoring program
for HABs will also be site- and season-specific. Further, the monitoring program will depend
upon the level of expertise of the operators, and on the degree of access to cell counting,
toxicity testing and analytical capacity for toxins, organic exudates, and other parameters.
The progress through this sequence, particularly in relation to consultation and warnings, will
vary depending upon the water source parameters such as the arrangement of the intake.
8.5.1! History of Alert Level Frameworks
The concept of the ALF was first developed for algal management in South Australia in 1991,
and modified and adopted nationally in 1992. It was subsequently adopted and used
internationally by the WHO as a model system for response to cyanobacterial blooms
(Bartram et al. 1999), and has been adapted by other users to incorporate recreational and
agricultural waters. The ALF given here is an updated version of the earlier Australian model
which now references the Australian Drinking Water Guidelines for microcystin toxins in
particular (NHMRC/NRMMC 2011). In this Chapter, the ALF concept has been adapted for
use in seawater desalination systems and has been expanded to consider biomass as well as
toxin. The generic Alert Level Framework described here was originally developed for
tracking populations of potentially toxic Microcystis aeruginosa using cell counts as a
surrogate for the toxin hazard. The use of ALFs in seawater desalination is a potential
management tool, but requires optimization to be plant specific.
8.5.2 Using an Alert Level Framework
The ALF follows the development of a potentially toxic or disruptive HAB through a
monitoring program with associated actions in four stages called Alert Levels. The actions
accompanying each level include additional sampling and testing, operational interventions,
consultation with health authorities and other agencies, and customer and media releases. The
sequence of Alert Levels is based upon initial detection of HAB cells at the Detection Level,
progressing to moderate HAB numbers at Level 1, where notification, operational readiness,
additional sampling, and assessment of toxicity may occur. For the next stage at Level 2, the
higher cell numbers can indicate the potential for the occurrence of toxins above guideline
concentrations (or potential guideline values). Alert Level 2 represents the point where the
operators may decide to take further action regarding operations within the plant to deal with
biomass. This would follow a full health assessment and depend upon circumstances such as
availability and performance of water treatment, consumption patterns, etc. It is possible of
course that an operator may decide to issue advice or a notice at cell numbers lower than that
equivalent to the guideline. The sequence can also continue to escalate to Alert Level 3 for
very high biomass in the seawater. This level represents the situation where the potential risk
of toxin in the seawater is significantly increased and treatment processes may break down
due to heavy biomass loads. Alert Level 1 and 2 ideally require an assessment of toxicity and
toxins in the seawater and assessment of both the drinking water and the performance of the
treatment system for toxin removal.
241
8.5.3 Levels of the framework

8.5.3.1 Derivation and definition of the levels
Cell toxin quotas in natural populations will be highly variable and the relationship between
toxin concentrations and cell numbers will not necessarily be valid for different species or
populations (see Chapter 1, Table 1.4); however, the cell number assumptions below are
regarded as reasonable for the purpose of preliminary hazard assessment in the absence of
toxin testing.
8.5.3.2!Detection level
This level encompasses the early stages of bloom development, where HABs are first
detected at low levels in raw water samples. The cell numbers for this level are somewhat
arbitrary, and as an example may be ~ 100,000 cells/L.
Taste and odors may become detectable in the supply, although this does not necessarily
indicate the presence of toxic HABs. If a routine monitoring program is not in place, this is
the appropriate time to collect and deliver samples to a laboratory for confirmation of the
presence of HABs. If there is no routine program the recommendation for monitoring is to
commence weekly sampling and cell counts at representative locations in the water body.
The presence of low population densities of HABs could still mean there is the potential for
the formation of localized surface scums or subsurface accumulations, and operators should
regularly inspect intake areas visually, or if not possible, inspect shoreline areas for scums or
discoloured water.
Alert Level 1
Alert Level 1 represents the level at which the HAB population has become established,
and localized high numbers may occur.
An example threshold for this level might be a cell number of 1,000,000 cells/L.
The definition for Level 1 is relatively conservative and has been chosen to indicate a
point that represents a cell density providing a buffer, or time margin, of a few days
before the guideline for toxin concentration in raw water might be exceeded (i.e. Level 2
conditions). This is based upon a population doubling rate of 4 days.
Alert Level 1 may require notification and consultation with health authorities and other
agencies for ongoing assessment of the status of the bloom. Although contact with health
authorities is recommended when this level is reached, it may not be required on a
weekly basis if local conditions deem this unnecessary. For instance, if the dominant
algal species present is not known to be a problem based on prior testing and experience,
this alert level can be adjusted to suit the local situation.
The requirement for toxicity assessment at this level will depend upon advice and
discussion with health authorities. It will also depend upon circumstances such as:
whether the HAB is a known toxic species, cell concentrations that might be dangerous
(see Table 8.3), past history of toxicity, nature of the supply and associated water
treatment, local sensitivity in relation to this supply, or other factors. This consultation
should be initiated as early as possible and continue after the results of toxicity testing
and toxin analyses become available.
The water (if a potentially toxic species is present) should be sampled to establish the
extent of the spread and patchiness of the bloom. Special samples (concentrated surface
layers and/or plankton net tow samples representative of the raw water intake) should be
242
collected and used for rapid toxin screening (Appendix 3), and/or dispatched for toxicity
testing or toxin analysis.
In consideration of biomass, this level may warrant operational actions such as analysis
of the pre-treatment performance or re-commissioning of the DAF system if this has
previously been bypassed. Operators should be made aware of the situation and
discussion regarding changes to pre-treatment made. Shock chlorination of the intake
should be suspended where possible. Other plant-specific actions should be discussed
that pertain to a readiness to act rapidly if Alert Levels escalate.
Alert Level 2
Alert Level 2 is the next stage at slightly higher cell numbers of potentially toxic HABs.
The threshold for Level 2 (in the absence of toxin information) is cell numbers and/or
biovolume that could indicate the potential for a toxin hazard at or above the guideline
level in the seawater if the population was highly toxic, and all toxins were released
from the cells and pre-treatment is ineffective for their removal. Some guideline values
for this determination are given in Table 8.3.
This level is characterised in general terms by an established bloom with moderately
high numbers showing a trend upwards over several successive samples at sampling
frequencies of at least twice per week. The HAB population is likely to have developed
to the extent that localised surface discoloration may be observed, though this does not
always happen.
An example threshold for this level might be a cell number of 2,000,000 cells/L.
Alert Level 2 represents the point where the operators and health authorities may need to
monitor the product water for relevant toxins to ensure toxin-free water is being
delivered to customers. It is also possible that an operator may decide to issue advice or
a notice at cell numbers lower than these thresholds, taking into account public safety.
It may be acceptable to continue to supply drinking water from the seawater even with a
positive toxicity result, dependent upon a risk assessment by the health authorities that
may recommend specific action to protect more susceptible population groups. As the
removal of toxin should be in excess of 99% using RO, some health authorities may
consider allowing continuation of supply. During this stage, direct toxin measurements
are critical for the decision process, emphasizing the needs to develop a capability for
reliable toxin screening by plant staff (Appendix 2), and to identify outside laboratories
that can rapidly do toxin analyses.
When considering the effect of biomass on the plant, the operational interventions at this
level are the same as those for Alert Level 1, with the addition of some minor changes to
operating parameters, such as backwash frequency of DMF and UF units. Diurnal
operation of the intake (off during the day and on at night) may provide some amount of
protection from the bloom, depending on the depth of the intake, and the swimming or
surface aggregation behavior of the HAB species (Chapters 1 and 6). Operation
conditions should be discussed to allow for changes, but have some operational capacity
in reserve to allow for Alert Level 3.
Alert Level 3
The cell number for Level 3 represents a minimum cell number that could produce ten
times as much toxin as any international guidelines. This describes an established toxic
bloom with high cell numbers and possibly localised surface accumulations. The
sampling program outside the plant (Chapter 3) will have indicated that the bloom is
243
widespread with no indication of a HAB population in decline in the short term.

Conditions in Level 3 are indicative of a significant increase in the risk of adverse
human health effects if water is treated by an ineffective pretreatment system.
An example threshold definition for Alert Level 3 is cell numbers of > 5,000,000 cells/L.
The cell count at Level 3 can be a trigger for the immediate notification of health
authorities, but this would only be in a situation where this has not occurred earlier (at
Level 1 or 2). This would occur where there was no prior information from an ongoing
monitoring program, treatment is limited, or its performance for toxin removal is
untested. This could be a scenario where a one-off sample or result is the initial
discovery of a major bloom in the source water. By definition, the circumstances for
Level 3 are that there is potential for adverse public health outcomes if these high
numbers are present in the source water or supply. When combined with failures of the
treatment system, a single pass RO system or thermal system, the population sensitivity,
and public water consumption patterns, a Level 3 incident has potential for serious
public health issues. High cell numbers also mean there is potential for much higher
localised concentrations, i.e. surface accumulations. Depending upon the position of the
intake, this could then mean that very high cell numbers could be entering the supply for
short periods and this may not be captured by the monitoring program.
In consideration of biomass removal, the full extent of the operational plan for HAB
mitigation should be deployed at this point. At the selected cell concentration for Level 3,
diurnal operation of the intake may no longer provide protection for the plant, again,
depending on the vertical migration behavior of the HAB species. Shock chlorination
should definitely be ceased, DAF should be fully operational where installed, and jar
tests undertaken to establish conditions for maximum cell removal without damage to
algae cells. If ultrafiltration is available, the backwash strategy should be reviewed and
pre-coagulation used if possible to maximize time between backwashes, avoid algal
organic material seeding biofouling on the RO, and to avoid irreversible fouling. If lower
fluxes can be used by operating more UF units in parallel, this should be undertaken. Jar
testing for GMF pre-coagulation should be undertaken and filter rates revised. Continual
analysis of RO normalized permeate flux, normalized salt passage and differential
pressure should be undertaken. The product water should be continually monitored at
this Alert Level for the specific toxins to confirm their removal in the interest of public
health.
In the unlikely event that seawater treatment barriers simultaneously fail and toxin
removal is unsatisfactory, toxins present in the product water at concentrations
significantly above suggested guidelines for Level 3 may result in the activation of a
contingency water supply plan that is appropriate for the operator and the system. This
may involve switching to an alternative drinking water supply for human consumption,
or in some circumstances the delivery of safe drinking water to consumers by tanker or
in bottles. More extensive media releases and provision of appropriate advice via direct
contact with customers may be necessary. Where advice is provided to the public
because of a toxin hazard to human health it may be appropriate to indicate that the
water would be suitable for purposes such as washing, laundry, toilet flushing. Complete
shutdown of a public drinking water supply because of a toxin hazard in source water is
not likely to be justified since potential hazards from disruption of supply (public
hygiene and fire-fighting, etc.) are likely to be worse than the toxin hazard.
Monitoring of the bloom in the intake and in coastal waters adjacent to the plant should
continue, to determine when it is in decline, so that normal operation and supply can be
244
resumed. Given the patchiness in marine HAB populations, monitoring at daily or every
other day intervals is sometimes warranted. Experience suggests that the cell
concentrations and thus the toxicity of a HAB population can change dramatically from
day to day, but it will likely take a week or two, and sometimes more to dissipate
completely.
The sequence of actions at Level 3 should follow through to deactivation of an
emergency with advice and media releases to confirm the decline of a bloom. It is
possible that the collapse of a bloom, or management action such as flushing and control
of surface accumulations, could lead to a rapid decline from Level 3 back to Level 1 or
beyond. Likewise, the sequence might escalate rapidly, bypassing Level 1 and 2, if
adequate monitoring and early warning information is not available. The collapse of a
bloom may be associated with the release of dissolved toxin into the water and the length
of time for toxins to degrade is variable. Generally, times to avoid toxin contamination
can vary from a minimum of several days to weeks depending upon the toxin and the
seawater.
A summary of Alert Level Frameworks is provided in Figure 8.3.
8.5.4 Customer and media information dissemination
Providing information to consumers and the media is an important aspect of managing water
quality problems associated with HABs. Information should be prompt and concise with
details regarding reasons for changes to supply and explanation for any differences in water
quality. It is important for all of the companies and agencies involved to provide coordinated
and consistent advice. It is also useful to have guidance from a HAB expert, in order to
address questions about the distribution and longevity of the bloom, or of the symptoms
associated with consumption of certain of the toxins.
The Alert Level Framework suggests a number of points where media releases could be
issued. These are in situations where consumers may experience changes in water quality, e.g.
due to changes in source water quality, switching to another source water, changes in
treatment, implementation of a contingency plan, or warning notices for recreational use of
the source water. It is always advisable to have a single, well-informed individual as the
spokesperson for the event. When too many people offer guidance to the media, the situation
can become confusing and inaccurate guidance released.
The approach to release information will depend on the nature of the supply and the problem.
For example, for SWRO plants with two passes and thus sophisticated multi-barrier treatment
infrastructure, it may not ever be necessary to advise consumers, as water quality changes
will not be evident in the product water. In those cases, however, it is important for the plant
to be aware of the algal bloom situation in the event that the public, government officials, or
journalists inquire about the safety of the drinking water after learning of the presence of a
dangerous HAB through other impacts (e.g., dead fish, poisonous shellfish). It is far better for
plant management to be able to say that they are aware of the situation and that tests have
been run to assure water quality, rather than to be ignorant of the situation and presume that
the RO and pretreatment processes are being effective. In circumstances with simple pre-
treatment and a single pass RO, as is often the case with low salinity feedwaters, or if the
bloom occurs in a multiple-use water resource (for instance those also used for recreation), it
is important to inform consumers of the extent of the problem as part of the management
strategy.
245
Figure 8.3. Flow chart for Alert Level Frameworks.
8.6! SUMMARY
This chapter has presented information regarding the risks presented by HAB toxins and how
designers and operators could use tools to consider these risks and mitigate them
appropriately. Many guidelines exist globally for HAB toxins in drinking water, although few
countries have guidelines that consider seawater toxins, apart from New Zealand and Brazil
that have guidelines for saxitoxin. To ensure complete removal of HAB toxins to a safe level
in a desalination plant, designers can consider the multi barrier concept to mitigate the risk of
unit process failure and toxin release into the distribution system. Risk assessments for HAB
toxins can be undertaken by several methods, one of which is the HACCP process. This
chapter also considered a practical assessment of toxin risks, calculating the residual risk
after treatment for several common seawater toxins. Finally, this chapter describes an Alert
Level Framework tool that can be used by operators to make rapid and effective decisions on
how to operate a plant during a HAB bloom according to cellular concentrations in the
seawater and thus react in an appropriate manner. By considering risks and mitigation options
using the tools presented here, designers and operators can keep drinking water supplies safe
and protect public health during toxic HABs.
8.7! REFERENCES
Bartram, J., Burch, M., Falconer, I. R., Jones, G., and Kuiper-Goodman, T. 1999. Chapter 6:
Situation Assessment, Planning and Management. In I. Chorus & J. Bartram (Eds.),
Toxic Cyanobacteria in Water: A Guide to Their Public Health Consequences,
Monitoring and Management. London: E & F Spon.
Berdalet, E., Fleming, L. E., Gowen, R., Davidson, K., Hess, P., Backer, L. C., Moore, S. K.,
Hoagland, P., and Enevoldsen, H. 2015. Marine harmful algal blooms, human health
and wellbeing: challenges and opportunities in the 21st century. Journal of the Marine
Biological Association of the United Kingdom FirstView, 1-31.
Bixio, D., and Wintgens, T. 2006. Water Reuse System Management Manual: AQUAREC:
Office for Official Publications of the European Communities.
Boerlage, S. F. E. and Nada, N. 2014. Algal toxin removal in seawater desalination processes,
In Proceedings of European Desalination Society, Cyprus.
Cetinić, I., Filteau, G., Lauri, P., Jones, B., and Trussell, S. 2010. Harmful algae and
their potential impacts on desalination operations off southern California. Water
Research 44, 385-416.
Codex Alimentarius. (2003) http://www.fao.org/fao-who-codexalimentarius/standards/list-of-
standards/en/?provide=standards&orderField=fullReference&sort=asc&num1=CAC/R
CP, World Health Organization and Food Administration Organization of the United
Nations.
Crittenden, J. C., Trussell, R. R., Hand, D. W., Howe, K. J., and Tchobanoglous, G. 2005.
Water Treatment: Principles and Design (Second ed.): Wiley.
EFSA. 2008a. Marine biotoxins in shellfish - Azaspiracid group, Scientific Opinion of the
Panel on Contaminants in the Food chain, adopted on 9 June 2008. EFSA Journal 723,
1-52.
247
EFSA. 2008b. Marine biotoxins in shellfish - Okadaic Acid and analogues, Scientific
Opinion of the Panel on Contaminants in the Food chain, adopted on 27 November
2007. EFSA Journal 589, 1-62.
EFSA. 2009a. Influence of processing on the levels of lipophilic marine biotoxins in bivalve
molluscs, Statement of the Panel on Contaminants in the Food Chain (Question No
EFSA-Q-2009-00203), Adopted on 25 March 2009. EFSA Journal 1016, 1-10.
EFSA. 2009b. Marine biotoxins in shellfish - Domoic acid group, Scientific Opinion of the
Panel on Contaminants in the Food chain; adopted on 2 July 2009. EFSA Journal 1181,
1-61.
EFSA. 2009c. Marine biotoxins in shellfish - Palytoxin Group, Scientific Opinion of the
Panel on Contaminants in the Food chain; adopted on 13 August 2009. EFSA Journal 7,
1-38.
EFSA. 2009d. Marine biotoxins in shellfish - Pectenotoxin group, Scientific Opinion of the
Panel on Contaminants in the Food chain; adopted on 27 May 2009. EFSA Journal
1109, 1-47.
EFSA. 2009e. Marine biotoxins in shellfish - Saxitoxin group; Scientific Opinion of the Panel
on Contaminants in the Food Chain, Adopted on 25 March 2009. EFSA Journal 1019,
1-76.
EFSA. 2009f. Marine biotoxins in shellfish - Summary Opinion, Scientific Opinion of the
Panel on Contaminants in the Food chain; adopted on 13 August 2009. EFSA Journal
1306, 1-23.
EFSA. 2009g. Marine biotoxins in shellfish - Yessotoxin group, Scientific Opinion of the
Panel on Contaminants in the Food chain, adopted on the 2 December 2008. EFSA
Journal 907, 1-62.
EFSA. 2010a. Marine biotoxins in shellfish - Cyclic Imines (spirolides, gymnodimines,
pinnatoxins and pteriatoxins), Scientific Opinion of the Panel on Contaminants in the
Food chain; adopted on 5 July 2010. EFSA Journal 8, 1-39.
EFSA. 2010b. Marine biotoxins in shellfish - Emerging Toxins: Brevetoxin-group, Scientific
Opinion of the Panel on Contaminants in the Food chain; adopted on 5 July 2010.
EFSA. 2010c. Marine biotoxins in shellfish - Emerging Toxins: Ciguatoxin-group, Scientific
Opinion of the Panel on Contaminants in the Food chain; adopted on 18 May 2010.
EFSA. 2010d. Statement on further elaboration on consumption figure of 400 g shellfish
meat on the basis of new consumption data, EFSA Panel on contaminants in the food
chain.; adopted on 31 July 2010. EFSA Journal 8, 1-20.
EFSA (EFSA Panel on Biological Hazards (BIOHAZ), E.P.o.C.i.t.F.C.C. 2012. Scientific
Opinion on the minimum hygiene criteria to be applied to clean seawater and on the
public health risks and hygiene criteria for bottled seawater intended for domestic use.
EFSA Journal 10, 2613 [2685 pages].
FSAI. 2006. Risk Assessment of Azaspiracids (AZAs) in Shellfish. A Report of the Scientific
Committee of the Food Safety Authority of Ireland (FSAI) August 2006, 39 pages.
Havelaar, A. H. 1994. Application of HACCP to drinking water supply. Food Control 5, 142-
152.
248
Hrudey, S., Burch, M., Drikas, M., and Ross, G. 1999. Chapter 9: Remedial Measures. In I.
Chorus & J. Bartram (Eds.), Toxic Cyanobacteria in Water: A Guide to Their Public
Health Consequences, Monitoring and Management. London: E & F Spon.
Hsieh, D., Huxtable, S., and Chang, G. 2000. Total production of C1/C2 by Alexandrium
tamarense. In Proceedings from the Ninth International Conference on Harmful Algal
Blooms; Hallegraeff, G., Ed.; University of Tasmania: Hobart, Tasmania.
Humpage, A. R., and Falconer, I. R. 2002. Oral Toxicity of Cylindrospermopsin: No
Observed Adverse Effect Level Determination in Swiss Albino Mice: CRC for Water
Quality and Treatment Research Report #13.
Jochimsen, E. M., Carmichael, W. W., An, J., Cardo, D. M., Cookson, S. T., Holmes, C. E.
and Barreto, V. S. T. 1998. Liver failure and death after exposure to microcystins at a
hemodialysis center in Brazil. New England Journal of Medicine, 338(13), 873-878.
Kudela, R. M., 2011. Characterization and deployment of Solid Phase Adsorption Toxin
Tracking (SPATT) resin for monitoring of microcystins in fresh and saltwater. Harmful
Algae 11, 117-125.
Laycock, M. V., Anderson, D. M., Naar, J., Goodman, A., Easy, D. J., Donovan, M. A., Li, A.
F., Quilliam, M. A., Al Jamali, E., and Alshihi, R. 2012. Laboratory desalination
Lawrence, J., Loreal, H., Toyofuku, H., Hess, P., Iddya, K., and Ababouch, L. 2011.
Assessment and management of biotoxin risks in bivalve molluscs. FAO Fisheries and
Aquaculture Technical Paper No. 551, 337 pages.
Lefebvre, K. A., Bill, B. D., Erickson, A., Baugh, K. A., O'Rourke, L., Costa, P. R., Nance,
S., Trainer, V. L. 2008. Characterization of intracellular and extracellular saxitoxin
levels in both field and cultured Alexandrium spp. samples from Sequim Bay,
Washington. Marine Drugs 6, 103-116.
Liefer, J. D., Robertson, A., MacIntyre, H. L., Smith, W. L., Dorsey, C. P., 2013.
Characterization of a toxic Pseudo-nitzschia spp. bloom in the Northern Gulf of Mexico
associated with domoic acid accumulation in fish. Harmful Algae 26, 20-32.
M., Miller, W. A., Toy-Choutka, S., Dominik, C., Hardin, D., Langlois, G. W., Murray,
M., Ward, K., Jessup, D.A. 2010. Evidence for a novel marine harmful algal bloom:
Cyanotoxin (microcystin) transfer from land to sea otters. PLoS One 5, e12576.
Nadebaum, P., Chapman, M., Morden, R., and Rizak, S. 2004. A Guide to Hazard
Identification and Risk Assessment for Drinking Water Supplies: CRC for Water
Quality and Treatment Research Report #11.
Newcombe, G., House, J., Ho, L., Baker, P., and Burch, M. 2010. Management Strategies for
Cyanobacteria (Blue-Green Algae): A Guide for Water Utilities: Cooperative Research
Centre for Water Quality and Treatment Research Report #74.
NHMRC. 2006. Guidelines for Managing Risks in Recreational Water. Canberra: National
Health and Medical Research Council.
NHMRC/NRMMC. 2011. Australian Drinking Water Guidelines Paper 6: National Water
Quality Management Strategy. Canberra: National Health and Medical Research
Council / National Resource Management Ministerial Council.
249
Smith, J. L., Tong, M., Fux, E., and Anderson D. M. 2012. Toxin production, retention, and
extracellular release by Dinophysis acuminate during extended stationary phase and
culture decline. Harmful Algae 19, 125–132.
Takahashi, E., Yu, Q., Eaglesham, G., Connell, D. W., McBroom, J., Costanzo, S., and Shaw,
G. R. 2007. Occurrence and seasonal variations of algal toxins in water, phytoplankton
and shellfish from North Stradbroke Island, Queensland, Australia. Marine
Environmental Research 64, 429-442.
Todd, K. 2011. Australian Marine Biotoxin Management Plan for Shellfish Farming:
Cawthorn Report #645.
Trainer, V. L., Wells, M. L., Cochlan, W. P., Trick, C. G., Bill, B. D., Baugh, K. A., Beall, B.
F., Herndon, J., and Lundholm, N. 2009. An ecological study of a massive bloom of
toxigenic Pseudo-nitzschia cuspidata off the Washington State coast. Limnology and
Velzeboer, R. M. A., Bater, P. D., and Rositano, J. 2001. Saxitoxins associated with the
growth of the cyanobacterium Anabaena circinalis (Nostocales, Cyanophyta) under
varying sources and concentrations of nitrogen. Phycologia 40, 305-312.
WHO. 1996. Guidelines for Drinking Water Quality: Health Criteria and Other Supporting
Information (Vol. 2). Geneva: World Health Organization.
WHO. 1998. Guidelines for Drinking Water Quality: Health Criteria and Other Supporting
Information - Addendum (Vol. 2). Geneva: World Health Organization.
WHO. 2004. Guidelines for Drinking Water Quality: Recommendations (Vol. 1). Geneva:
World Health Organization.
WHO. 2011. Guidelines for Drinking Water Quality, Fouth Edition. Geneva: World Health
Organization.
Zendong, Z., Abadie, E., Mazzeo, A., Hervé, F., Herrenknecht, C., Amzil, Z., Dell’Aversano,
C., and Hess, P. 2015. Determination of the concentration of dissolved lipophilic algal
toxins in seawater using pre-concentration with HP-20 resin and LC-MS/MS detection,
in: MacKenzie, L. (Ed.), 16th International Conference on Harmful Algae 27th-31st
October 2014. Cawthron Institute, Nelson, New Zealand and International Society for
the Study of Harmful Algae, Wellington, New Zealand.
250
Algal biomass pretreatment in SWRO
9 ALGAL BIOMASS PRETREATMENT IN SEAWATER REVERSE

OSMOSIS
Mike B. Dixon1, Siobhan F.E. Boerlage2, Nikolay Voutchkov3, Rita Henderson4, Mark Wilf5,
Ivan Zhu6, S. Assiyeh Alizadeh Tabatabai7, Tony Amato8, Adhikara Resosudarmo4, Graeme
K. Pearce9, Maria Kennedy7, Jan C. Schippers7, and Harvey Winters10
1
2
3
Water Globe Consultants, Winter Springs, Florida, USA
4
School of Chemical Engineering, The University of New South Wales, Australia
5
Mark Wilf Consulting, San Diego, California, USA
6
Leopold, a Xylem Brand, Zelienople, Pennsylvania, USA
7
8
Water2Water Consulting, England, UK
9
Membrane Consultancy Associates Ltd, Reading, UK
10
Fairleigh Dickinson University, Teaneck, New Jersey, USA

9.1 Introduction.......................................................................................................................................252
9.2 Chlorination in SWRO .....................................................................................................................252
9.2.1 Overview ......................................................................................................................................252
9.2.2 Chlorination during a HAB ..........................................................................................................254
9.2.3 Summary ......................................................................................................................................257
9.3 Dechlorination in SWRO ..................................................................................................................257
9.4 Coagulation for DAF, GMF, and UF pretreatment ..........................................................................257
9.4.1 Overview ......................................................................................................................................257
9.4.2 Type of coagulation feed systems ................................................................................................260
9.4.3 Coagulation operational considerations .......................................................................................262
9.4.4 Coagulation and flocculation for DAF pretreatment ...................................................................264
9.4.5 Coagulation for GMF pretreatment .............................................................................................265
9.4.6 Coagulation for MF/UF pretreatment ..........................................................................................266
9.5 DAF pretreatment for SWRO ...........................................................................................................272
9.5.1 Overview ......................................................................................................................................272
9.5.2 Fundamental principles of DAF...................................................................................................274
9.5.3 Process design of DAF systems ...................................................................................................275
9.5.4 DAF in SWRO pretreatment for removal of marine algae ..........................................................277
9.5.5 Optimization of process parameters for marine HABs ................................................................277
9.5.6 Summary ......................................................................................................................................278
9.6 Granular media filtration ..................................................................................................................278
9.6.1 Overview ......................................................................................................................................278
9.6.2 The filter operation cycle .............................................................................................................279
9.6.3 Single and two-stage filtration .....................................................................................................281
9.6.4 Gravity filters ...............................................................................................................................282
9.6.5 Pressure filters ...............................................................................................................................284
9.6.6 GMF filter performance ...............................................................................................................286
9.7 Microscreens for membrane pretreatment ........................................................................................287
9.7.1 Overview ......................................................................................................................................287
9.7.2 Types and configurations .............................................................................................................287
9.7.3 Algal bloom-related challenges ...................................................................................................288
9.7.4 Summary ......................................................................................................................................290
9.8 Microfiltration/ ultrafiltration ...........................................................................................................290
9.8.1 Overview ......................................................................................................................................290
9.8.2 MF/UF filtration modes ...............................................................................................................290
9.8.3 Fouling of MF/UF during algal blooms .......................................................................................292
9.8.4 Removal of HAB cells using MF/UF ..........................................................................................293
9.8.5 Summary ......................................................................................................................................296
251
9.9 Cartridge filters for reverse osmosis pretreatment............................................................................297

9.9.1 Overview ......................................................................................................................................297
9.9.2 Types and configurations .............................................................................................................297
9.9.3 Algal-bloom related challenges ...................................................................................................298
9.10 Reverse osmosis................................................................................................................................299
9.10.1 Overview .................................................................................................................................299
9.10.2 Raw water quality and pretreatment requirements for SWRO ...............................................299
9.10.3 Effect of algal blooms on SWRO operation ...........................................................................301
9.10.4 Ferric coagulant SWRO fouling during algal bloom events ...................................................301
9.10.5 SWRO Biofouling during and following algal bloom events .................................................302
9.10.6 SWRO operational strategies during a bloom .........................................................................306
9.10.7 Membrane cleaning .................................................................................................................306
9.10.8 Summary .................................................................................................................................307
9.11 Summary of biomass removal in SWRO..........................................................................................307
9.12 References.........................................................................................................................................308

9.1 INTRODUCTION
Harmful algal blooms (HABs) can result in a substantial increase in the organic and solids
load in the seawater feed to be treated at a desalination plant. In this chapter, the removal of
this material is addressed in the context of the multi-barrier treatment process for seawater
reverse osmosis (SWRO) as presented in Chapter 8 on risk management for HAB events.
While this chapter covers removal of non-toxic material, Chapter 10 builds upon these
principles and discusses the mechanisms and effectiveness for each barrier with respect to
toxin removal. This chapter covers only the main barriers used in the SWRO desalination
plants for HAB bloom risk mitigation, though the authors acknowledge that other niche
treatment barriers exist in SWRO systems. The treatment processes discussed here are
chlorination and dechlorination, dissolved air flotation (DAF), granular media filtration
(GMF), microscreens for microfiltration/ultrafiltration (MF/UF), MF/UF itself, cartridge
filtration and SWRO. Coagulation is discussed in general terms and then more specifically
for DAF, GMF, and MF/UF pretreatments. Each treatment process is broken down into a
discussion of how the process works and then how HAB cells affect the process operation.
Importantly, the chapter deals with how upstream actions can detrimentally affect
downstream treatment processes with respect to algal blooms.
In particular, this chapter discusses removal mechanisms for algal organic matter (AOM) and
how operational actions can prevent detrimental effects of AOM. As discussed in Chapter 2,
the chemical composition of AOM usually includes proteins, polysaccharides, nucleic acids,
lipids, and other dissolved organic substances. AOM compounds typically cover a wide size
spectrum, ranging from less than 1 nm to more than 1 mm. Based on their size cut-off, GMF
and MF/UF are expected to remove only part of high molecular weight AOM (as shown in
Chapter 2, Figure 2.2). SWRO is expected to achieve complete removal of AOM, but will
suffer from fouling issues if AOM is not removed upstream.
9.2 CHLORINATION IN SWRO
9.2.1 Overview
The disinfection properties of chlorine have been known for many years and routine use of
chlorine in water treatment processes began in the early 1900s; however, in the last decade
water chlorination has received criticism due to its production of disinfection byproducts.
In order to prevent marine growth, such as molluscs, in seawater intakes for both SWRO and
thermal plants, biocides such as chlorine, ozone, potassium permanganate, and hydrogen
peroxide can be used. The most widely used among them is chlorine, usually applied in one
252
of three forms: 1) chlorine gas; 2) calcium hypochlorite; and 3) sodium hypochlorite, with the
latter being the most typical (Bahamdan et al. 1999).
Chlorination in SWRO desalination plants is typically applied either at regular intervals
(intermittent/shock chlorination) or continuously (rare) and dosed at the intake structure with
additional dosing points downstream to maintain the required chlorine residual. Residual
concentration of chlorine in the feedwater is typically kept at 0.2 - 4 mg/L (Agus et al. 2009).
For intermittent chlorination, the duration of the dose may last from 30 to 90 minutes and is
typically undertaken every 1 to 2 weeks at a random time (Ferguson et al. 2011). Often this is
undertaken more frequently, up to daily.
The continuous chlorination method has suffered from criticism that it causes biofouling of
the reverse osmosis (RO) process unit. Appelgate et al. (1989) reported that chlorine
degrades humic acids and high molecular weight compounds present in coastal seawater to
smaller molecules that can be assimilated by bacteria. The chlorine suppresses bacterial
activity, but when the sodium metabisulfite (SMBS) is added to remove chlorine prior to the
RO, the surviving bacteria quickly take advantage of the nutrients generated by the
degradation of larger molecules and enter into a cycle of enhanced growth. The significant
increase in the biomass of bacteria after dechlorination causes slime development of biofilm
on the surfaces of pipes and RO membranes (Winters 1995). To overcome the problem,
alternate disinfectants have been used such as chlorine dioxide, chloramine and copper
sulfate (Winters and Isquith 1995) although these have not been widely adopted. Plants that
practice continuous chlorination now typically use very low doses (0.1 to 0.3 mg/L).
Shock chlorination, also referred to as intermittent chlorination, is undertaken to control
growth of marine life on pipelines and equipment that are constantly in contact with seawater.
Both continuous and shock chlorination inhibit but do not fully prevent the growth of marine
life so the presence of common oceanic foulants such as molluscan shells and barnacle
deposits is unavoidable. The growth rate can be controlled to manageable levels, however.
The effectiveness of shock chlorination can also be limited by the withdrawal of marine
creatures into their protective shells and their re-emergence when chlorine has dissipated,
especially when chlorine addition is undertaken at regular times rather than random intervals.
Shock chlorination of seawater collected by open ocean intakes will result in an increased
load of particulates on all solid removal processes due to the varied size range of solids
introduced. The suspended solids will be in the micron size range, such as finer silts through
to larger particles such as molluscan shells, byssal threads, and smaller debris like molluscan
tissues and larvae. The impact on a membrane filtration system can be minimized by
installing strainers or disk filters that would remove the bulk of the solids load created by the
chlorination process (Ferguson et al. 2011).
As an alternative to using chlorine, a low dose of chlorine dioxide can be compatible with
polyamide RO membranes, under certain conditions and to a certain extent. When generating
chlorine dioxide onsite, it can be contaminated with chlorine and thus dechlorination is still
necessary (Dow Water and Process Solutions 2015).
Intake chlorination can also produce carcinogenic compounds like trihalomethanes and
haloacetic acids, as well as other inorganic disinfection byproducts (DBPs) such as chlorite,
chlorate, bromate, and nitrogenous DBPs (N-DBPs). The latter generally form in much
smaller amounts than chlorinated DBPs, but have been a growing concern over the past
decade because of their greater health risk. High levels of nitrogen-containing compounds in
AOM, which increases dramatically during a bloom, can lead to the formation of significant
quantities of N-DBPs (Le Roux et al. 2015).
253
As bromide is present in seawater, hypobromus acid (HOBr) formation is favored over

hypochlorus acid (HOCl). HOBr has been found to react much more rapidly than HOCl with
organic (and inorganic) compounds. It is also reported that HOBr has been found to be 25
times stronger than HOCl in its halogen substitution. Therefore, further reaction through a
series of oxidations and reductions leads to the formation of the carcinogenic species bromate.
Where bromate becomes an issue, sodium hypochlorite generated from seawater should be
ceased and other alternatives used such as chlorine gas or calcium hypochlorite (Al-Rasheed
et al. 2009).
The concentration of the above DBPs in the final permeate/distillate should be controlled to
meet guideline levels prescribed by the WHO or other local agencies. A study by Le Roux et
al. (2013) showed that desalinated water produced by thermal MSF plants and SWRO plants
is drinkable and poses no threat to human health with respect to DBPs.
Destabilizing and inactivating algal cells through chlorination, which lyses the cells through
breakdown of the cell wall, can theoretically prevent biofouling on pre-treatment membranes
and RO if a suitable treatment strategy is employed, such as by using coagulation during
pretreatment to increase the percentage of AOM removed. Similar outcomes are observed
when other oxidative agents (i.e. ozone, permanganate and non-oxidizing biocides) are used
instead of chlorination (Heng et al. 2008).
9.2.2 Chlorination during a HAB
In addition to the applied concentration of chlorine, cell destruction efficiency is also
dependent on the duration of exposure and depends largely upon the cell wall thickness and
type. Thicker cell walls will take longer to degrade, therefore both the chlorine dose and the
residence time in the intake are important. Typical residence times for SWRO intakes are
highly variable (5-60 mins) and can depend upon factors such as the length of intake tunnels
and pump wells. Previous research using the marine dinoflagellate Prorocentrum, which has
a thick cellulose cell wall, showed that doses of 2-3 mg/L of chlorine could achieve complete
algal cell lysis within 24 hours (Resosudarmo et al. 2017). Total cell lysis also occurred in the
process. Even non-lethal doses of 0.1 – 0.5 mg/L were found to reduce the photosynthetic
ability of marine algae significantly without a high degree of cell lysis. While the mechanism
of chlorination of HAB cells is still being investigated, one study by Azanza et al. (2001)
showed that cells of the dinoflagellate Pyrodinium bahamense were degraded via rupturing of
the thecal plates and the release of mucilage.
There are also other benefits to chlorination beyond the destruction of algal cells during
HABs. The efficiency of coagulation-based processes (such as prior to DAF, GMF or UF)
increases significantly when combined with low doses of chlorine (0.1 - 0.5 mg/L). Previous
research has shown that up to 96% of algal cells can be removed when both processes are
combined (Shen et al. 2011). Chlorination was also found to reduce fouling of an outside-in
UF membrane based pretreatment systems (Xu et al. 2014). Once a dose above 1.5 mg/L was
applied, further increase in residual chlorine levels did not result in additional fouling
improvement.
The negative effects of chlorination are typically observed when excessive doses are applied
to the feed seawater during a HAB. High levels of chemical stress can result in the lysis of
algae cells and release of intracellular organic matter (IOM) comprising potential foulants,
taste and odor compounds and/or toxins. While the value will vary depending upon the algal
species in a particular bloom, due in part to differences in cell wall composition, previous
research with the marine alga Tetraselmis suecica showed that significant cell lysis occurred
at chlorine levels exceeding 5 mg/L (Resosudarmo et al. 2017). Another study using the
254
freshwater Microcystis aeruginosa cyanobacterium demonstrated that chlorine doses as low

as 0.8 mg/L were sufficient to induce cell lysis (Ma et al. 2012). In both studies, exceeding
the tolerable levels of residual chlorine resulted in large amounts of IOM release. In the case
of Microcystis, cell lysis also resulted in toxin release.
Analysis of the IOM released from Tetraselmis suecica cell lysis found that the majority was
small enough to pass easily through UF pretreatment membranes and potentially increase RO
biofouling, as depicted in Figure 9.1 (Resosudarmo et al. 2017). Other studies have pointed
out that AOM released by algae during a bloom can also contribute to long-term permeability
decline of pretreatment membranes and reduced cleaning effectiveness, as discussed in
Chapter 2 (Heng et al. 2008; Hung and Liu 2006). High molecular weight AOM comprises
protein and polysaccharide compounds, including transparent exopolymer particles (TEP)
and their precursors as discussed in Chapter 2.
Membrane
surface
Intact algal cells -‐ Intact algal cells prevent
IOM from passing through
the UF membrane pores.
-‐ Mucilage surrounding the
algal cells is retained.
-‐ Rapid transmembrane
pressure increase due to
compressible fouling layer.
Chlorination
-‐ Released IOM can now
transmit through the
membrane pores.
-‐ Mucilage now attaches to the
membrane surface,
preventing efficient cleaning
-‐ Reduced short-term fouling
Figure 9.1. Proposed effect of algal cell lysis on UF membrane fouling and rejection. Modified from
Resosudarmo et al. 2017.
It is therefore paramount that the correct residual chlorine dose is applied during algal blooms.
While this can be difficult due to the high number of existing algal species responsible for
blooms, it is important to note that lower chlorine doses combined with longer exposure
times will lead to less IOM release during treatment.
When long exposure is not feasible, it is possible to use a chlorine dose between 3-5 mg/L,
demonstrated to be sufficient to achieve destruction of many types of algae (Junli et al. 1997).
This also indicates that applying chlorination in a continuous manner during the pretreatment
stages may be more beneficial towards overall plant performance, as opposed to shock
chlorination at high concentrations. Continuous chlorination has, however, been extensively
shown to cause SWRO biofouling due to breakdown of organics into assimilable organic
carbon, which is more easily used by biofouling bacteria as a nutrient source (Dow Water and
Process Solutions 2015). Resosudarmo et al. (2017) showed that the increase in chlorine dose
255
greatly increased the amounts of fouling compounds such as biopolymers, building blocks,
low molecular weight (LMW) acids and LMW neutrals, with the greatest increase being for
the LMW acids (Figure 9.2). While biopolymers (mostly macro polysaccharide-like and
protein-like molecules) only appear to increase a small amount, they have been identified as
major foulants affecting membrane filterability (Zheng et al. 2010) and thus a small increase
has a major impact on fouling potential. One option to remove IOM may be the addition of
powdered activated carbon (PAC) to pretreatment systems as it has been shown to limit the
transmission of IOM to downstream RO membranes (Huang et al. 2015).
10
9 1mg/L
Concentration ratio (C/C0)
8 2.5mg/L
7
5.0 mg/L
6
10 mg/L
5
4 20 mg/L
3 40 mg/L
2
1
0
Biopolymer Building blocks LMW acid LMW neutrals
Figure 9.2. Effect of excessive chlorination on Tetraselmis suecica algal cells, showing a significant increase in
the low molecular weight (LMW) fractions. C/C0 is the concentration of compound in the chlorinated solution
divided by the concentration in the sample prior to chlorination, thus normalizing the conditions to the baseline
concentration. Modified from Resosudarmo et al. 2017.
Biofouling of SWRO membranes remains a significant detriment to successful plant
operation despite the use of chlorination. The bacteria responsible for forming biofilms can
survive chlorine addition through several possible mechanisms. Chlorine-resistant bacteria
may become individually encapsulated in response to chlorine and be protected from its
biocide effects or chlorine may promote bacterial aggregation, both of which would protect
cells from biocides (Appelgate et al. 1989). The formation of bacterial aggregates is a defense
mechanism in which only the outer cells are impacted by chlorine. The formation of bacterial
aggregates, while it decreases the number of colony forming units (cfu), enhances the
attachment of the aggregated bacteria to a surface to initiate biofilm formation (Mir et al.
1997).
Much of the organics formed during HABs act as chemical conditioning agents that modify
the RO membrane surface to allow for bacterial attachment. Therefore, there is an increase in
fouling potential during HABs and the use of increased concentrations of chlorine to cope
with these blooms will only enhance bacterial aggregation and mucoid development. This in
turn increases the rate of biofouling of the RO membranes. It is well understood that
aggregated bacterial cells are more capable of tolerating environmental stress, and that
survival of cells in aggregates promotes a highly clustered spatial distribution of bacteria on
surfaces.
Another aspect often overlooked is the potential negative impact on the integrity of pre-
treatment UF/MF membranes. While these membranes are assumed to be resistant, previous
research has shown that prolonged exposure to high chlorine levels can result in accelerated
256
membrane ageing (Regula et al. 2013). Membranes exposed to high levels of chlorine were
found to have lower permeability, hydrophilicity, and tensile strength. Therefore,
pretreatment membranes situated in plants frequently exposed to marine algal blooms and
high chlorine levels may have shorter than expected lifetimes.
9.2.3 Summary
Chlorination of the intake can lyse HAB cells, but this may complicate downstream processes
if not managed correctly. Shock chlorination leads to more aggressive lysis of HAB cells and
subsequent coagulation pretreatment processes may not remove all AOM prior to the RO
stage, causing biofouling. A strategy for avoiding cell lysis is to avoid shock chlorination
during a HAB. A low continuous dose (0.1-0.2 mg/L) of hypochlorite may be a better
approach to minimize the lysis of algal cells, while releasing some AOM to assist coagulation.
More research is required to ascertain the best approach. Care should be taken when
undertaking this approach to ensure RO biofouling is not inadvertently occurring. Further
complications in choosing a chlorination strategy when HAB species are toxic are discussed
in Chapter 10.
9.3 DECHLORINATION IN SWRO
Since polyamide RO membranes are susceptible to oxidative degradation from chlorine,

dechlorination of the RO feed water upstream of the RO membranes is necessary. This is
achieved by adding a reducing agent - typically SMBS. In theory, 1.34 mg of SMBS will
remove 1.0 mg of free chlorine. In practice, however, 3.0 mg of SMBS is normally used to
ensure complete dechlorination of 1.0 mg of chlorine (Dow Water and Process Solutions
2015). To the authors’ knowledge, no references were available at the time of publication
showing the direct impact of SMBS on HABs, but SMBS may lyse HAB cells; however, if
pretreatment is operated efficiently, very few HAB cells will be present entering the RO.
Given SMBS is routinely used to preserve RO elements for long term storage (Dow Water
and Process Solutions, 2015), the direct effect of SMBS on the surface of the RO membranes
may prevent biofouling to some degree.
9.4 COAGULATION FOR DAF, GMF, AND UF PRETREATMENT
9.4.1 Overview
The coagulation process is critical for removal of HAB cells and thus it is important to
optimize coagulant dose to obtain a high removal of particulates and AOM (via particle
destabilization and agglomeration or adsorption). As discussed in Chapter 2, high molecular
weight AOM such as biopolymers, particularly very sticky TEP, have been identified as the
main cause of membrane fouling rather than the algal cells themselves.
Critical to the downstream process is achieving a low iron residual to prevent iron fouling of
MF/UF and RO. This section examines process parameters that may be optimized for each
unit process downstream of the coagulant dosing point. For GMF/DAF/MF/UF, process
conditions such as optimization of pH, coagulant dose, mixing speed, and coagulation time
are discussed as well as associated operational parameters such as GMF filter rates, MF/UF
flux and DAF recycle rates. This section discusses optimization of backwash frequency, and
when this does not yield improved filtration conditions, how cleaning can be best undertaken
(such as in the case of MF/UF). These coagulation/flocculation strategies are illustrated by
the use of supporting research on simulated algal blooms to isolate and elucidate precise
mechanisms for scale-up and use in plant scenarios.
A coagulant (typically a hydrolyzed metal salt) with the opposite charge to a suspended
colloid is added to raw water to overcome the repulsive charge and "destabilize" a suspension.
257
In a colloidal suspension, particles will settle very slowly or not at all because the particles
carry the same surface charges that mutually repel each other; the coagulant accelerates this
particle settling process by neutralizing particle charge. For example, when colloidal particles
in source seawater typically are negatively charged, ferric chloride is added as a coagulant to
create positively charged ions (specifically cationic ferric chloride hydrolysis products),
which attract and ultimately neutralize the charge of the seawater’s suspended solid particles.
Once the repulsive charges are neutralized, the van der Waals force agglomerates the
particles and form micro floc. Conversely, flocculation involves the process of clumping the
small, destabilized micro flocs together into larger aggregates so that they can be more easily
separated from the water. Flocculation is a physical process and does not involve the
neutralization of charge. Coagulation may be used in conjunction with flocculation to assist
with suspended solids separation.
Ferric salts, ferric chloride, and ferric sulfate, are the best choice for seawater coagulation
(Edzwald and Haarhoff 2011). When ferric chloride is introduced in the seawater, both of
them form ferric hydroxide, which is a large, positively charged molecule that attracts and
coagulates predominantly negatively charged seawater suspended solids particles. While
aluminum sulfate and polyaluminum chlorides (PACls) have been studied extensively at
laboratory and pilot-scale in seawater RO pretreatment (Gabelich et al. 2006), they are not
used in full-scale applications, primarily due to the relatively high solubility of aluminum,
which may result in carryover and accumulation on RO membranes leading to aluminum
hydroxide fouling (Gabelich et al. 2005). Ferric chloride is less soluble over a wider pH
range, resulting in lower residual dissolved iron in RO feed water and less fouling problems.
Further, ferric hydroxide has a high ratio of cationic charge to total mass (Jamaly et al. 2014)
that makes hydrolysis products more reactive and adsorptive with emulsified and semi-
emulsified organic matter; e.g. oil and grease, natural and synthetic organic matter. The
settled sludge volume of the ferric hydroxide formed from ferric chloride is reportedly 30–
60% that of sulfate based coagulants (e.g. Fe2(SO4)3). Additionally, the sludge developed
from ferric chloride is generally much more dewaterable (CWT 2004).
The most important process parameters for coagulation are mixing intensity (G), flocculation
shear rate (product of mixing intensity and time, G x t), pH, and temperature. Camp and Stein
(1943) developed the basic theory of power input for mixing and defined G as the mean
velocity gradient, which is proportional to the square root of power dissipated per unit
volume of liquid. Upon coagulant addition, hydrolysis is instantaneous and the reactions that
lead to the formation of ferric hydroxide occur in the order of seconds. As such, mixing has
to ensure that the coagulant is fully dispersed within the liquid in the shortest time possible.
Flocculation occurs by particle collision through thermally induced Brownian motion,
stirring, or differential settling.
Notably, coagulation is electric-force driven attraction of the negatively charged particles of
the source water by the positively charged molecules of the coagulant, (which in the case of
ferric salts is a trivalent metal salt), which neutralizes the charge of suspended, colloidal, and
dissolved materials so these solids no longer repulse each other and subsequently are
removed by aggregation followed by sedimentation or flotation. The mechanisms of
destabilization for the negative particle charge of the suspended solids naturally occurring in
seawater were summarized previously by Crittenden et al. (2012):
1) compression of the electrical double layer;

2) adsorption and charge neutralization;
3) adsorption and inter-particle bridging; and
4) enmeshment in a precipitate or “sweep floc”.
258
It is unlikely that significant changes in ionic strength would occur due to coagulant addition.
Therefore, the compression of the double layer should not be the dominant mechanism in the
coagulation process, especially for seawater coagulation. Adsorption and inter-particle
bridging usually happen when nonionic polymers and high-molecular-weight polymers are
added. In the latter two mechanisms, surface charge plays an important role in dictating the
speed and effectiveness of the formation of larger particles. The speed and effectiveness of
the second mechanism – formation of larger particles by physical contact (enmesh/catch or
sweep flocculation) –depends mainly on the number of particles in the source water (i.e., the
turbidity/total suspended solids (TSS), concentration of the seawater) and their nature
(Edzward and Haarhoff 2011).
The dose of coagulant is determined by the content of solids in the source water, the
electrical charge of the solid particles, and desired removal, the concentration of algae and
particles, in addition to temperature, pH, alkalinity, and salinity, and other factors. In general,
the higher the negative electrical charge of the particles in the source water and the less algae
the seawater contains, the lower the coagulant dose needed. In addition, the higher the
content of mineral particles in the seawater (i.e. the higher the turbidity/TSS concentration)
the more coagulant will be needed to engage these particles in the formation of larger flocs.
If the source water particles have a strong negative charge, the dominating mechanism for
large floc formation is electric attraction – therefore, relatively low doses of coagulant could
achieve high coagulation effect. If the source particles do not have a strong charge, then the
dose of coagulant will mainly be driven by the content of mineral suspended solids and
natural organic matter (NOM) in the source water and the time the source water particles and
coagulant will have to get in physical contact with each other – i.e. to collide with each other
and stick together to form a larger floc. Recent research suggests that the dominant
mechanism for coagulation of the marine dinoflagellate Prorocentrum minimum was through
sweeping flocs, with charge neutralization constituting a critical step (Zhu et al. 2014).
Therefore, operators often incorrectly assume that if they add more coagulant to poorly
coagulating waters (i.e. waters containing particles with low or non-existent negative charge)
they will improve the coagulation and filtration process.
During hydrolysis of salts, a complex of polynuclear, positively charged species are formed
in a matter of seconds. During flocculation, colloidal particles and some fraction of dissolved
AOM may attach to the floc body, and are eventually retained in downstream DAF, GMF,
and UF processes.
Coagulation and flocculation are critical for DAF, GMF, and UF. During the DAF process,
compressed air is introduced into a recycle stream of clarified water, is dissolved, and
subsequently generates 10–100 µm bubbles when released through dispersion headers into a
DAF tank. Coagulated particles, such as algae, attach to the bubbles and float to the top of the
water column where they are mechanically or hydraulically removed. Two other removal
mechanisms are adsorption, where the particles stick to the media surface, and biological
removal, where soluble contaminants are removed through biological metabolism. GMF
typically accumulates materials larger than 10 µm (Ripperger et al. 2012), which may include
algal cells and large AOM (See Chapter 2, Figure 2.2). Since the size of some algae in
seawater can be smaller than this threshold, coagulation to increase algal cell size is of
critical importance to improve the removal efficiency of GMF. Unlike GMF, UF removes
particles through physical straining only.
Coagulant addition is accomplished ahead of the SWRO pretreatment sedimentation tanks,
dissolved air flotation units, GMF or MF/UF. The optimum coagulant dose is pH dependent
259
and should be established on site through jar or pilot testing to provide site-specific
conditions that will be encountered during plant operation. Practical experience indicates that
the optimum pH for coagulation of particles in saline waters is highly temperature dependent.
As the temperature decreases, the optimum pH for coagulation increases and vice versa. For
example, the optimum pH for a temperature of 10oC is 8.2, while for source water
temperature of 35oC, the optimum pH decreases down to 7.4 (Edzwald and Haarhoff 2012).
Other factors influencing pH adjustment are salinity and alkalinity. The acidity constants
K1sw and K2sw of carbonic acid are a function of salinity and temperature. For example, K1sw
for seawater salinity of 35,000 ppm is 10-5.99 at 10°C, and 10-5.76 at 35°C. These differ greatly
from fresh water constants (ionic strength approaching 0 M), which are 10-6.46 and 10-6.34.
Seawater alkalinity is mainly contributed by carbonate and bicarbonate in addition to borate.
The mechanisms proposed for coagulation of NOM are a chemical phase change or
precipitation by complexation with soluble metal species for pH < 6 and adsorption to and/or
enmeshment in metal hydroxide precipitates for pH > 6 (Dennett et al. 1996). In seawater,
dissolved metal speciation is affected by the high ionic strength, and optimum pH values for
organic matter complexation are higher than those reported in freshwater (Duan et al. 2002;
Edzwald and Haarhoff 2011). For lower temperature seawater (< 20°C), coagulation pH of
6.5 - 7 should be effective to maximize the availability of Fe(OH)2+; however, Henderson et
al. (2008a) demonstrated that AOM is characteristically different from NOM and therefore
existing knowledge on NOM coagulation may not be adequate to explain on AOM fouling
potential and removal in UF systems.
The formation of agglomerates in part comes about due to the negative charges (that naturally
occur on the surfaces of particles, including algae, in the untreated water) becoming
overcome by the addition of coagulants and sometimes polymers that neutralize surface
charges, encouraging closer contact and subsequent agglomeration. Similarly, colloidal and
dissolved organic matter, such as that produced by algae, can interact with coagulants,
undergoing a phase change as they grow to form larger particles. In freshwater, both ferric
and alum coagulants are commonly applied due to the production of positively-charged
hydrolysis products (Duan and Gregory 2003); however, in seawater applications, ferric salts
are the coagulant of choice as previously discussed (Edzwald and Haahoff 2011). Hence,
much of the research conducted to date has focused on ferric chloride, although other ferric
coagulants have been considered.
Algal cells are covered with AOM produced during metabolic activities. The charge of those
molecules is likely influenced by H+/OH− ions, providing opportunities to use pH
adjustment to control coagulation and optimize coagulant doses. Since cells are negatively
charged, adding hydrogen ions may neutralize the negatively charged functional groups (e.g.
phosphate and carboxyl). At pH lower than 5.5, cells can lyse under stress, releasing
intracellular substances that may not be fully removed by downstream pretreatment and may
contribute to RO membrane fouling (Zhu et al. 2014).
9.4.2 Type of coagulation feed systems
Chemical conditioning of the source seawater includes three key components: chemical feed
system, coagulation, and flocculation tanks. The purpose of coagulation tanks is to achieve
accelerated mixing of the coagulant with the source water and to neutralize the electric
charge of the source water particles and colloids. Subsequent agglomeration of the
coagulated particles into larger and easy to remove flocs is completed in flocculation tanks.
While coagulation is a relatively rapid chemical reaction, flocculation is much slower sand
typically requires longer contact time and mixing conditions. Therefore, coagulation and
260
flocculation system design requirements differ. The mixing intensity is defined by velocity
gradient G (1/s). The required mixing energy expressed as G × t (t, retention time, s) is
typically 4,000 - 20,000 for rapid mixing, and 30,000 - 80,000 for flocculation.
Several process conditions can be applied for coagulation prior to UF/MF, DAF, or GMF (as
illustrated in Figure 9.3).
-‐‑ Rapid mixing and no flocculation, i.e. inline coagulation;

-‐‑ Rapid mixing and flocculation;
-‐‑ Rapid mixing, flocculation and sedimentation (not common before MF/UF); and
-‐‑ Rapid mixing, flocculation and flotation.
Inline Coagulation
Static
Coagulant Mixing Coagulant
Sea Sea
Water Water
Mixing and Flocculation
Rapid Flocculation
Coagulant Mixing
Sea
Water
Mixing, Flocculation and Sedimentation
Rapid Flocculation Sedimentation

Coagulant
Mixing
Sea
Water
Mixing, Flocculation and Flotation
Rapid Flocculation Flotation

Coagulant
Mixing
Sea
Water
Figure 9.3. Schematic presentation of various process conditions for coagulant application
with or without solids separation processes prior to MF/UF systems (Tabatabai 2014).
261
The main purpose of the coagulant feed system is to achieve uniform mixing of the added
coagulant with the source water, which promotes accelerated attraction of the coagulant
particles to the source water solid particles (i.e. to facilitate efficient coagulation). The two
coagulant mixing systems most widely used in desalination plants are: in-line static mixers
(Figure 9.4) and mechanical (flash) mixers installed in coagulation tanks (Figure 9.5).
In-line static mixers have lower energy

and maintenance requirements and are
relatively easy to install. They typically
operate at a velocity range of 0.3 to 2.4
m/s and are designed to operate in plug-
flow hydraulics in order to provide
uniform mixing within the entire pipe
cross section.
Mechanical flash mixing systems
consist of a coagulation tank with one
Figure 9.4. In-line static mixer at the Adelaide SWRO or more mechanical mixers and
desalination plant prior to UF pretreatment. Approximate chambers. The coagulation tank is
flow through this mixer is 6,250 m3/h. typically designed for mixing energy G
x t = 4,000 to 6,000. This type of
mixing usually provides a more reliable
and consistent coagulation, especially
for desalination plants designed for
significant differences in minimum and
maximum plant production (i.e. more
than 1:10).
9.4.3 Coagulation operational
considerations
The increase of the coagulant dose 2-3
fold (usually applicable for
clarification) is a common practice
during algal bloom events and often
results in deterioration rather than
improvement of the downstream
clarification and/or filtration process, if
there are no clarification processes such
as sedimentation and DAF in front of
GMF and UF. Overdosing results in an
excessive quantity of coagulant, which
could have the undesirable effect of
increasing the stability of colloidal
particles and accelerating dispersion of
colloids This is due to reversal of
Figure 9.5. Flash mixers in a coagulation tank. surface charge, more specifically,
formation of high density positive
charges on the colloids’ surface and mutual electrostatic repulsion. As a result, higher silt
density index (SDI) and turbidity values may be found following filtration compared to
before coagulation. In addition, as the content of coagulant particles is significantly larger
than that of naturally occurring suspended solids, a lot of excessive unreacted coagulant
262
remains in the conditioned seawater. The larger flocs, coupled with inadequate mixing of
coagulant with source seawater and presence of unreacted coagulant, result in accelerated
clogging of the pretreatment filtration media and downstream cartridge filters (Figure 9.6)
and often cause heavy fouling of the RO membranes (Figure 9.7) during algal bloom events.
Figure 9.6. Coagulant accumulation on cartridge filters due to overdosing. Photo:

Voutchkov 2013.
Figure 9.7. Coagulant residue on the RO membrane feed due to overdosing. Photo:
Voutchkov 2013.
The effect of overdosing of ferric coagulant on the SDI can be recognized by visually
inspecting the SDI test membranes. In Figure 9.8, the first two SDI test membranes are
discolored as a result of coagulant overdosing, resulting in a feedwater with higher fouling
263
potential. Hence, they were measured at the 5-minute interval (SDI5 reading of 16.2 and 16.3)
compared to the third SDI membrane showing no discoloration and measured after 15
minutes (SDI15).
Figure 9.8. Iron accumulation on the first two SDI test membranes due to Coagulant Overdosing
compared to the SDI membrane on the right hand side. SDI measurements and time interval are
below the SDI membranes.
In such situations, a significant improvement in RO feedwater SDI during an algal bloom

event can be attained by simply reducing the coagulant feed dose or in case of poor mixing,
modifying the coagulant mixing system to eliminate the content of unreacted chemical in the
filtered seawater fed to the RO membrane system. Coagulation optimization for algal bloom
is further discussed in Sections 9.4.4, 9.4.5 and 9.4.6. Jar testing is recommended each time
an algal bloom occurs that causes the turbidity of the raw seawater to exceed 5 NTU or TSS
to exceed 10 mg/L.
9.4.4 Coagulation and flocculation for DAF pretreatment
Good coagulation chemistry is essential to obtain favorable attachment of algal cells to
bubbles generated in the DAF pretreatment system. Coagulation chemistry is the most
important operating control variable affecting flotation performance. Without coagulation,
the algal cells and other particles carry a negative charge. Since bubbles are also negatively
charged, resultant bubble attachment is poor. Good coagulation chemistry depends upon
using an appropriate coagulant dose and adjustment to a suitable pH. Optimum coagulation
conditions are those of coagulant dose and pH that produce flocs with charge within an
optimal operating window close to neutral, as measured using zeta potential, as this
minimizes the electrostatic barrier to contact resulting from surface charges (Henderson et al.,
2008b). This produces flocs with relatively high hydrophobicity, which minimizes metal
hydroxide precipitates. Metal coagulant hydroxides are hydrophilic, and therefore, when
sweep flocculation mechanisms dominate, bubble-particle attachment efficiency is
compromised. Charge neutralization conditions are therefore preferred in DAF, but can be
difficult to achieve, particularly in freshwater conditions, due to the narrow operating
window (Henderson et al. 2008c). These conditions cause high bubble attachment efficiency.
The increased ionic strength of seawater means that the zeta potential is not as extreme as in
freshwater systems due to electrical double layer compression that should in fact make
coagulation easier.
In a DAF system, removal of Prorocentrum minimum increased by ~5-10% in a study by Zhu
et al. (2014) at 60 and 50 mg/L ferric chloride doses, respectively when pH was adjusted to
264
6.4 and 6.3. Adjustment of pH maintained algal removal at over 90% even with 30 mg/L of
ferric chloride.
The formation of agglomerates in part comes about due to the negative charges that naturally
occur on the surfaces of the particles (or algal cells) and bubbles in the untreated water
becoming overcome by the addition of coagulants and sometimes polymers which neutralize
the surface charges encouraging close contact and bonding.
The need for effective surface charge neutralization and coagulation, whilst not peculiar to
DAF, requires the chemistry to be optimized both in terms of the coagulant dose and pH.
Ease of adjusting the pH by a mineral acid is dependent on the “buffer intensity”. This term
can be simply defined as the resistance of pH to change. The buffer intensity arises from the
ionic strength of both inorganic carbon and alkalinity or borate in freshwater or seawater,
respectively.
The resistance is greater in seawater and increases as the water temperature falls, resulting
potentially in higher doses of chemicals (e.g. sulfuric acid or hydrochloric acid when
downstream RO necessitates low sulfates) to achieve the optimal coagulation pH. Dose rates
should be calculated post jar testing.
As noted by Edzwald (2010), floc sizes in the range of 25–50 µm (pin point flocs) were of
the optimal size to achieve high floc-bubble collision efficiency and for separation of floc
bubble aggregates. The majority of marine bloom-forming algae (dinoflagellates, diatoms,
and cyanobacteria), fall into this size.
9.4.5 Coagulation for GMF pretreatment
Coagulation combined with granular media filtration is the most commonly used method for
seawater pretreatment at present. The severity of algal blooms in the area of the intake, as
well as the size and charge of the algae cells most commonly occurring during algal blooms,
have a significant impact on the sizing of these facilities and their efficient operation.
Conventional (or GMF) pretreatment technology is based upon removal of suspended solids
and some organics through coagulation and flocculation. The process is well established and
in the majority of cases is capable of producing SWRO feedwater of the required quality with
respect to suspended solids, SDI and turbidity and is thus very effective as a pretreatment.
The quality of the treated water varies significantly with quality of the raw water; however,
deterioration of raw water quality will affect operation of the filtration system. In
conventional pretreatment systems, coagulation is mainly applied to improve surface loading
rates and ensure that product water quality meets the requirements of RO membrane
manufacturers in terms of turbidity and SDI15. Coagulant dose in conventional SWRO
pretreatment systems may range from 0.5 to 10 mg Fe/L, although in some cases doses as
high as 20 mg Fe/L have been reported during poor water quality events (Lattemann 2010;
Edzwald and Haarhoff 2011).
Use of coagulants is critical for the effective and consistent performance of GMF
pretreatment filtration systems; however, if the source water contains low turbidity (< 0.5
NTU) and the prevailing size of particles is less than 5 µm (which is common for deep
intakes with low algal content), coagulant addition does not yield a significant improvement
in the GMF process. In this case, the addition of a minimal amount of coagulant (i.e.
0.5 mg/L or less) or even no coagulant addition is viable. In such conditions, it is critical to
have a prolonged period of coagulation and flocculation (i.e. coagulation and flocculation
times of 10 minutes or more), because for these particles, the main mechanism for floc
formation is physical contact rather than charge attraction (Voutchkov 2013).
265
9.4.6 Coagulation for MF/UF pretreatment

In contrast, to GMF and DAF, MF/UF systems do not rely on coagulation to enhance
permeate quality in terms of turbidity and SDI as particles as fine as 0.04 µm (MF
membranes) or 0.01 µm (UF membranes) can be removed without coagulation. UF
membranes are generally preferred over MF in SWRO pretreatment due to better removal of
particulate/colloidal organics, silt, and pathogens from seawater owing to their smaller pore
size (Voutchkov 2009). Operating without coagulant addition offers many advantages –
reducing process complexity and costs. Operators give preference to systems that require no
coagulant or if this is not achievable, minimum amounts of coagulant. Minimizing or
eliminating coagulant addition, while maintaining stable process performance and high
permeate quality, can be achieved by optimizing process parameters or applying alternative
coagulation process conditions. Operating without coagulant also avoids potential
environmental impacts associated with the use and disposal of pretreatment chemicals such
as coagulants, coagulant aids, and others. (WHO 2007). In areas with more stringent
legislation on brine discharge, such as Europe, Australia, and the USA, coagulant-rich waste
streams require extensive treatment and handling prior to discharge, which add a significant
cost component to the overall pretreatment process. In such cases backwash water containing
coagulant is treated separately (e.g. by gravity settling in lamella plate sedimentation tanks).
Supernatant can be either disposed with RO concentrate or recycled at the head of the
pretreatment. The coagulant-rich sludge retained in the sedimentation tank is often dewatered
onsite and transported to sludge treatment facilities or landfills (WHO 2007).
Potential impacts on the environment associated with the use and disposal of pretreatment
chemicals such as coagulants, coagulant aids, and others increase process complexity (WHO
2007). Coagulants and coagulant aids (high molecular weight organics, e.g. partially
hydrolyzed polyacrylamide) present in spent backwash water are typically discharged to the
ocean without treatment (Lattemann 2010). Ferric chloride has very low toxicity for marine
organisms; however, discharge may cause an intense discoloration of the reject stream (red
discoloration of the concentrate), which may increase turbidity and reduce light penetration,
or could bury sessile benthic organisms at the discharge site (Lattemann and Höpner 2008).
In some cases, coagulation is required upstream of MF/UF seasonally or in response to poor
water quality events to avoid excessive fouling. For example, coagulation may be necessary
if the source water contains NOM particles with strong negative charge that could be
coagulated easily and removed via filtration; 2) during heavy algal blooms (for AOM
removal); or 3) during oil spill events. Coagulant dosing prior to MF/UF filtration can greatly
reduce AOM passing through the pretreatment membranes, in particular UF membranes,
which can promote biofouling in the RO (see Chapter 2). Moreover, coagulation can reduce
pore blocking and/or surface attachment by sticky particles such as biopolymers produced
during an algal bloom, enhancing cake filtration. This will reduce non-backwashable fouling
of MF/UF membranes and pressure increase (Guigui et al. 2002; Choi and Dempsey 2004;
Schurer et al. 2013). If the source water is consistently high in suspended solids and organics,
then additional pretreatment steps may be applied upstream of the MF/UF to reduce loading
onto the membranes and to achieve higher membrane fluxes.
Coagulation is commonly applied using an inline mode in MF/UF systems for SWRO
pretreatment and sometimes with DAF ahead of the MF/UF. Inline coagulation is the
application of a coagulant without removal of flocs through a clarification step. Inline
coagulation may also be characterized by the absence of a flocculation chamber, as large floc
size for enhanced settling is not a requirement in UF systems. The absence of flocculation
and clarification steps in the overall process scheme results in lower investment cost for
266
inline coagulation as compared to conventional schemes that consist of coagulation/

flocculation/sedimentation/flotation. In most inline coagulation applications, mixing is
achieved in a static mixer or the wet well of the UF feed pump station. Flocculation may
occur during the mixing step (simultaneously with coagulant dispersion and hydrolysis), in
the piping network that carries the coagulated feedwater to the membranes, or within the UF
capillaries.
As discussed previously, Henderson et al. (2008a) demonstrated that AOM is
characteristically different from NOM and therefore existing knowledge on NOM
coagulation may not be adequate to explain the effect of coagulation on AOM fouling
potential and removal in MF/UF systems. Tabatabai (2014) therefore carried out a series of
laboratory-scale experiments to optimize AOM removal by pressure-driven inside–out UF
(150 kDa) membranes using inline coagulation and to compare removal with conventional
pretreatment. In these experiments, AOM was harvested from the marine diatom species
Chaetoceros affinis to simulate seawater bloom conditions, creating feed solutions at a
concentration of 0.5 mg C/L as biopolymers in synthetic seawater. The microalga
Chaetoceros was selected as it is known to produce large quantities of extracellular
polysaccharides throughout their growth cycle (Watt 1968; Dam and Drapeau 1995;
Myklestad 1995), making it a suitable choice for laboratory-scale production of AOM in
short-term experiments.
In this series of experiments, the optimal ferric coagulant dose and pH for removal of
biopolymer TEP0.4 (a fraction of biopolymer), and dissolved organic carbon (DOC) were
investigated (Figure 9.9). (See Chapter 5 for information on tests to measure biopolymers,
TEP, modified fouling index - UF (MFI-UF) and high resolution liquid chromatography –
organic carbon detection (LC-OCD)). As expected, more AOM was removed with increasing
ferric coagulant dose to a point at which there were diminished returns, which defines the
optimum dose (10 mg Fe/L). Two pH values were investigated, a pH typical of seawater (8),
and seawater acidified to a pH of 5, due to the two coagulation mechanisms explained
previously (Section 9.4.1). Laboratory-scale results from Tabatabai (2014) demonstrated that
pH did not greatly affect coagulation efficiency of AOM in terms of removal of biopolymers
(Figure 9.10).
Figure 9.9. Removal rates of biopolymers (left panel), TEP0.4 (settled samples; center panel;) and DOC as a
function of coagulant dose and pH (right panel). Figure: Tabatabai 2014.
Mixing intensity is also a factor for consideration in maximizing AOM removal and thus
reducing fouling on the RO membrane. In addition to AOM removal, the MFI-UF (see
Chapter 5, Section 5.5.2) enabled the effect of varying process parameters such as mixing
intensity and flocculation time on the fouling potential of the RO feedwater to be assessed. At
a coagulant dose of 1 mg Fe(III)/L (Figure 9.10), mixing intensity of 1100 s-1 resulted in
267
substantially lower MFI-UF values than for 100 s-1. Mixing time, however, did not affect
filterability of coagulated AOM flocs, as no difference was observed in MFI-UF values for
20 s versus 240 s of mixing
time.
The mode of coagulant
dosing was investigated by a
variety of pretreatment steps
in another series of
experiments using
Chaetoceros affinis, as
described previously
(Tabatabai 2014). High
resolution LC-OCD was
Figure 9.10. Effect of mixing intensity (G) and mixing time on fouling employed to investigate
potential of coagulated AOM for 1 mg Fe(III)/L (left panel) and 5 mg removal of biopolymer
Fe(III)/L (right panel). Figures: Tabatabai 2014. fractions to ascertain the best
coagulant dosing mode.
Figure 9.11. Biopolymer concentrations for fractions of different molecular weight as a function of coagulant
dose for (a) Mode A-coagulation(coag)/flocculation(flocc)/sedimentation(sed), (b) Mode B -
coag/flocc/sed/0.45 µum, and (c) Mode C-inline coag/UF. Figures: Tabatabai 2014.
Higher MWCO biopolymer fractions of the AOM were well removed using inline
coagulation with UF (150 kDa) and also in the experiments designed to simulate coagulation
with flocculation coupled to different levels of pretreatment (i.e. sedimentation and
sedimentation followed by 0.45-µm-membrane filtration (Figure 9.11).
As mentioned previously, coagulation can enhance cake filtration on the UF membrane
thereby, reducing non-backwashable fouling and pressure development in MF/UF systems.
The effect of coagulation on fouling propensity and removal of AOM in pressure driven
inside–out UF membranes from the laboratory-scale experiments of Tabatabai (2014) and the
findings from the Jacobahaven demonstration plant (see Case Study 11.10) where
Chaetoceros was also found are further discussed in the following sections.
a) Fouling potential and compressibility of AOM. Coagulation can reduce transmembrane
pressure (TMP) increase during filtration of algae-laden feedwater in pressure driven inside-
out (PDI) UF membranes. When coagulant is dosed to seawater containing high
concentrations of AOM, iron reacts with the biopolymers, changing the properties in such a
way that the fouling potential of AOM is improved (i.e. lower MFI-UF values; Figure 9.12
268
top) and the AOM cake/gel layer becomes less compressible (i.e. more linear TMP curves,
Figure 9.12 bottom). This effect was demonstrated for a synthetic seawater solution of AOM
harvested from C. affinis to simulate algal bloom conditions. Inline coagulation was
performed at different coagulant doses prior to filtration through PDI UF membranes with
nominal MWCO of 150 kDa.
0.07#
0.06# (b)#
0.05#
TMP#[bar]#
0.04#
0.03#
0.02#
0.01#
0.00#
0# 5# 10# 15# 20# 25#
Time#[min]#
0#mg#Fe/L# 0.1#mg#Fe/L# 0.5#mg#Fe/L# 1#mg#Fe/L#
2#mg#Fe/L# 5#mg#Fe/L# 10#mg#Fe/L#
Figure 9.12. Effect of inline coagulation on fouling potential as measured by MFI-UF (top panel) and
TMP development during filtration of algal laden seawater (0.5 mg C/L as biopolymers obtained from
Chaetoceros affinis) through 150 kDa UF membranes at 100 L/m2h. pH ranged from 8.0 to 7.1. (bottom
panel). Figures: Tabatabai et al. 2014.
At coagulant doses < 0.5 mg Fe/L (pH = 8.0), the concentration of positively charged iron
hydroxide flocs species is too small for any reaction to occur and hence AOM fouling
potential and compressibility were not affected. At low coagulant dose (< 1 mg Fe/L) in
natural seawater pH (~ 8), colloidal Fe-biopolymer complexes were formed and a slight
269
reduction in fouling potential and compressibility were observed. At coagulant doses of 0.5-1
mg Fe/L (pH 7-8), AOM adsorption on iron hydroxide precipitates occurred, resulting in the
formation of iron-biopolymer aggregates that were relatively large and less compressible. At
higher coagulant doses, the cake/gel layer properties tended toward iron hydroxide flocs.
Residual iron in all UF permeate samples was below the detection limit (20 µg Fe/L)
(Tabatabai et al. 2014).
b) Fouling reversibility. Coagulation can reduce the extent of hydraulically irreversible
fouling by AOM in PDI UF membranes. This was demonstrated at the Jacobahaven
demonstration-scale on North Sea water during several successive bloom periods (Schurer et
al., 2012, 2013). Ferric chloride dosed prior to the UF feed pump at an average dose of 0.5-
1.5 mg Fe/L stabilized UF operation and reduced the frequency of chemically enhanced
backwashing (CEB) at a nominal flux of 60 L/m2h (Case Study 11.10). The efficiency of
CEB in recovering membrane permeability was significantly reduced when coagulant was
applied, indicating UF fouling by residual iron. Under such conditions, membrane
permeability could only be restored by applying tailored CIP.
An alternative mode of coagulant application (i.e. coating) in seawater has shown promising
results at laboratory-scale in terms of UF hydraulic performance at very low coagulant dose
(0.5 mg Fe/L). In this process, a layer of preformed flocs of iron hydroxide (H2FeO3) is dosed
at the start of each filtration cycle to create a protective barrier that prevents the attachment of
sticky AOM (such as TEPs) to the membrane surface (Figure 9.13). The protective layer
should be formed in a short amount of time at the start of the filtration cycle in order to
prevent/minimize membrane-foulant interactions, and should not alter the intrinsic membrane
permeability. Thereafter, seawater is filtered directly through the coated membranes. At the
end of each filtration cycle, backwashing is applied whereby the coating layer containing
AOM (including sticky TEPs) is lifted off the membrane surface and flushed out. Handling
and treatment of spent coating material follows the same procedure as that of coagulated
sludge. For the process to be successful, the coating layer should be highly permeable, so as
not to reduce system efficiency, and easily backwashable.
Laboratory-scale experiments were conducted on feedwater containing AOM obtained from
Figure 9.13. Simplified schematic presentation of UF coating by iron

hydroxide particles to enhance hydraulic backwashing. Figure:
Tabatabai 2014.
Chaetoceros affinis (as described previously) in synthetic seawater (total dissolved solids
(TDS) = 35,000 ppm) to simulate bloom conditions in the North Sea. Coating suspensions
270
with a range of particle size were created by precipitation of iron hydroxide and subsequent
grinding at various intensities. Application of a coating layer prior to filtration of seawater
with high AOM concentration (0.2 – 0.7 mg C/L) stabilized operation of PDI UF membranes
with a nominal MWCO of 150 kDa, by significantly enhancing backwashability (Figure
9.14). Reducing particle size of the coating material to the submicron range (400-700 nm)
allowed for a significant reduction in coating dose. Coating with nanoparticles of iron
hydroxide, allowed for continuous stable operation at equivalent dose of 0.5 mg Fe/L. This is
a significant improvement to inline coagulation in terms of required coagulant dose.
coagulant dose. Furthermore, creating preformed flocs through precipitation and subsequent
grinding may reduce the risk of UF fouling by residual iron (Tabatabai 2014).
Figure 9.14. TMP development during filtration of algal laden seawater (0.5 mg C/L as
biopolymers obtained from Chaetoceros affinis) through 150 kDa UF membranes coated with
preformed iron hydroxide flocs at different equivalent dose, filtration flux 100 L/m2h. Figure:
Tabatabai et al. 2014.
At coagulant dose < 0.5 mg Fe/L (pH = 8.0), the concentration of positively charged iron
hydroxide flocs species is too small for any reaction to occur and hence AOM fouling
potential and compressibility were not affected. At low coagulant dose (< 1 mg Fe/L) and
natural seawater pH (~ 8), colloidal iron-biopolymer complexes were formed and a slight
reduction in fouling potential and compressibility was observed. At coagulant dose of 0.5-1
mg Fe/L (pH 7-8), AOM adsorption on iron hydroxide precipitates took place, resulting in
the formation of iron-biopolymer aggregates that were relatively large and less compressible.
At higher coagulant dose, the cake/gel layer properties tend toward the properties of iron
hydroxide flocs. Residual iron in all UF permeate samples was below detection limit (20µg
Fe/L) (Tabatabai et al. 2014).
c) Permeate quality. In SWRO plants, coagulant dosing prior to UF filtration can greatly
reduce AOM flux through the UF and the seeding of biofouling in the RO (see Chapter 2).
Commercially available PDI UF membranes based on polyethersulfone (PES) with nominal
MWCO of 150 kDa can remove up to 45% of algal biopolymers compared to 25% when
simulating conventional coagulation using 0.45 µm filtration (Figure 9.15). Inline coagulation
at 0.5 mg Fe/L enhanced biopolymer removal by approximately 20%. At 5 and 10 mg Fe/L,
biopolymer removal was further enhanced by 20%, resulting in a biopolymer removal of
approximately 85%. Inline coagulation/UF exhibited superior biopolymer removal compared
271
to conventional coagulation/0.45 µm
filtration at low coagulant dose (i.e., 0.5
mg Fe/L); however, this difference
became marginal at higher coagulant
dose, such that at 10 mg Fe/L no
difference was observed in biopolymer
con-centration between conventional
pretreatment and UF pretreatment.
Increase in dose shifts the predominant
coagulation mechanism to sweep floc
and inter-particle bridging, whereby
removal is enhanced by adsorption to
Figure 9.15. Removal of algal biopolymers as a function and/or enmeshment in precipitated iron
of coagulant dose for coagulation/flocculation/ sedimen- hydroxide. Biopolymer removal was
tation followed by 0.45 µm filtration and inline coagula- mainly through the removal of fractions
tion/UF. larger than 100 kDa (Tabatabai et al.
2014).
Biopolymers were mainly composed of larger molecular weight fractions; approximately
80% of the total was larger than 100 kDa. As a consequence, reduction in biopolymer
concentration was mainly due to the removal of compounds >100 kDa. Coagulation had a
strong impact on the removal of biopolymers larger than 100 kDa for UF and conventional
pretreatment simulated in laboratory scale experiments where removal was substantially
higher at higher coagulant dose. In contrast, to GMF and DAF, MF/UF systems do not rely
on coagulation to enhance permeate quality in terms of turbidity and SDI.
Higher MW biopolymer fractions of the AOM were well removed for inline coagulation with
UF (150 kDa) and in experiments designed to simulate coagulation with flocculation coupled
to different levels of pretreatment (e.g., sedimentation and sedimentation followed by 0.45
µm media filtration; see Figure 9.11 and Figure 2.2 in Chapter 2 for more details). Tabatabai
(2014) showed optimum removal doses for ferric coagulant and pH to remove biopolymer,
the fraction of biopolymers measured by TEP0.4 (see Chapter 5 Section 5.3.1.2) and DOC,
where feedwater concentrations were 0.55-0.63 mg/L, 0.27 mg/L and 1.7 mg/L respectively
(Figure 9.9). The optimum dose from this experiment was 10 mg/L of ferric coagulant at pH
8, although pH was far less important for optimization than the ferric coagulant dose.
9.5 DAF PRETREATMENT FOR SWRO
9.5.1 Overview
DAF has been used in drinking water treatment since the 1960s for the removal of low
density suspended solids and organics and for reducing turbidity. The performance of DAF is
dependent on the preceding agglomeration (coagulation/flocculation) step. Unlike
sedimentation however, where the aim is to generate large flocs to facilitate sedimentation,
DAF does not require large flocs, as removal is achieved by floating floc-bubble aggregates.
That coupled to technological developments has led to lower flocculation times, coagulant
consumption, and sludge production.
From its earliest use, DAF was found to be highly effective in treating a variety of algal rich
sources including freshwater and wastewater. For example, even as early as 1975, Hyde
(1975) reported that waters containing 30,000,000-150,000,000 cells/L could be treated using
flotation with flocculation periods of only 7-9 minutes. A consequence of these reduced
flocculation times is that there is a saving not only in terms of space but also in the cost of the
272
civil structure and mechanical and electrical equipment, as smaller flocculators can be used.
This work was further expanded and reported by Valade et al. (1996) and Edzwald et al.
(1999) confirming reduced flocculation times and, more significantly, that previously typical
clarification loading rates of 10 m/h (referred to as conventional rate DAF) could be
increased. Over the intervening years, further development leading to improved
understanding in the application of ever higher loading rates has continued, achieving 50 m/h
(referred to as high-rate DAF) with as little as 5 minutes of total flocculation time, as
reported by Amato et al. (2012), albeit at pilot scale. Practically, DAF is still commonly
designed with two stages of flocculation to optimize performance and prevent short circuiting
during mild algal blooms or normal operations, which usually need longer retention times,
especially if the source water contains relatively low turbidity.
In the 1990s, conventional-rate DAF was employed as part of the pretreatment scheme to
treat algal blooms at a small scale SWRO plant at the Gas Atacama power station in Chile
(see Case Study 11.7). High-rate DAF (Rictor /Aqua DAF) was later trialed at Taweelah in
the Gulf in a 2002 pilot plant study prior to two stage GMF. This demonstrated that the
required RO feedwater SDI15 could be obtained and that emulsified oil was removed in
spiked tests (Rovel 2003). DAF was not pursued further at that time as algal cell counts
remained < 100,000 cells/L. Subsequently, DAF and dissolved air flotation and filtration
(DAFF) were employed at the El Coloso (Chile) and Tuas 1 (Singapore) SWRO desalination
plants, respectively in the first large-scale applications of DAF/DAFF in SWRO pretreatment.
Algal blooms were the main driver for including high rate DAF in the pretreatment scheme
for the El Coloso plant. In the case of the Tuas 1 plant, DAFF (Figure 9.16) was primarily
installed because of the potential presence of
high solids (up to 60 mg/L) and oil (up to 10
mg/L), not for algal removal. Since, that time
DAF is increasingly being incorporated in
large-scale SWRO pretreatment schemes
prior to GMF or UF for treating algal blooms,
particularly in the Middle East, such as the
Shuwaikh plant in Kuwait (see Case Study
11.5) with a pretreatment capacity of 350,000
m3/d.
DAF is important in removing algal cells and
Figure 9.16. Dissolved air flotation and filtration reducing the suspended solids load for
(DAFF) installation at the Tuas, Singapore, SWRO downstream pretreatment processes, as the
desalination plant. Photo: PUB Singapore. DAF separation process is ‘gentle’ or low
shear, thereby reducing cell lysis and release
of fouling organics and algal toxins,
mitigating the risk of AOM fouling in UF and
RO. Removal of intact algae also assists in
reducing taste and odor issues associated with
algal blooms (see Chapter 10). The role of
DAF in seawater RO systems and important
parameters that can be optimized during an
algal bloom are discussed in more detail
Figure 9.17. DAF float layer when treating a below. Figure 9.17 shows a DAF float layer
cyanobacterial bloom. South Australian Water when treating a cyanobacterial (freshwater)
Corporation Bolivar WWTP DAFF plant. Photo: bloom.
Biomass lab, UNSW and SA Water.
273
9.5.2 Fundamental principles of DAF

The fundamental principle that has given rise to the development of DAF is that of enhancing
the natural buoyancy of the particulates carried within a fluid by attaching them to micro-
bubbles to encourage separation. The air used in the DAF process for freshwater applications
is normally dissolved under a pressure of 400-600 kPa into a proportion of the previously
clarified flow, termed the recycle (Edzwald 2010). The recycle rate applied will vary
depending on the nature of the flow to be treated and in freshwater may range between 8-
12%. The fundamental difference between fresh water and seawater (with all other conditions
being equal), is the higher salinity of seawater, typically in the range of 35,000-45,000 mg/l
total dissolved solids, which reduces the amount of air that can be dissolved. Henry’s
constant for the main gases that make up what is called air, (i.e. argon, nitrogen and oxygen)
are all higher in seawater by approximately 30% in all cases and all temperatures, meaning
they are all less soluble. This is sometimes referred to as the “salting out” affect and results in
either a requirement for an increase in pressure of ~30% or the recycle rate having to be
increased by ~20% (Haarhoff and Edzwald 2013). In practice, it is normal to increase these
two variables together. For example, if for fresh water a 10% recycle is used then, on a
seawater plant, the recycle rate would be increased to 12%. The pressurized recycle flow is
then passed through various types of pressure reducing devices such as needle valves or fixed
orifice nozzles resulting in the immediate release of a cloud of micro-bubbles within the
contact zone of the flotation tank. It is within this contact zone that the bubbles and flocs to
be removed are intermixed and floc-bubble aggregates are formed (Figure 9.18).
Historically there have been attempts to set feedwater quality limits (TSS, turbidity, algae
cells, oil, and grease). Jansenns and Buekens (1993) suggested that the application of DAF be
limited to the source water turbidity less than 100 NTU; however, the performance and
application of DAF is dependent on the nature of the solids or pollutants to be removed and
ensuring that the correct water chemistry conditions are applied during coagulation-
flocculation at all times to maximize removal efficiencies.
Qin
Qr
Figure 9.18. Schematic of a DAF system including dual stage flocculation.
DAF removal rates for algal cells are dependent on bloom cell concentrations - the higher the
cell concentration, the higher the removal rate, similar to that observed for TSS and turbidity
removal. For instance, DAF pilot trials treating waste stabilization pond effluent with algal
cell concentrations averaging 3´108 cells/L showed that at least 99% of cells could be
removed in the float when preceding coagulation-flocculation was optimized (Yap et al.
274
2012). In a seawater desalination context, however, pilot studies conducted using DAF have
typically been undertaken at low algal concentrations (e.g. less than 100,000 cells/L
(Bonnelye et al., 2004) and less than 3 µg/L chlorophyll a (Kim et al. 2011)) and therefore
removal rates are expected to be lower. The type of algae will also impact removal
efficiencies. For example, in drinking water applications the morphology (size and shape) of
algal cells have been shown to impact treatability. Spherical cells < 5 µm or needle-shaped
cells have been most difficult to remove by DAF, which is considered to be a result of the
cell morphology making coagulation more difficult combined with the fact that they are
likely to settle vertically, leaving only the narrow tip of the needle (< 5 µm) for collision by
the bubble to (Henderson et al. 2008c; Konno 1993). Some motile species have been
observed to swim out of flocs, meaning that flagellated species also tend to have a low
removal rate. A small amount of pre-oxidation can inactivate motile species (Henderson et al.,
2008c); however, great care is required to avoid excessive AOM release (see Section 9.1 on
chlorination/dechlorination). Furthermore, AOM that is released by algal cells has been
shown to impact treatability by DAF, as AOM concentration and character is species
dependent (Henderson et al. 2010; Villacorte et al. 2015). For example, AOM has been
observed to hinder coagulation by chelating coagulant, increasing the dose required, while
other studies have shown that if its character is enriched in biopolymers, it can act as a
bioflocculant and thus enhance flocculation via bridging mechanisms (Henderson et al. 2010;
Pivokonsky et al. 2016). Further investigations are required for seawater applications to
confirm that this also occurs for higher salinity feedwater.
9.5.3 Process design of DAF systems
The typical DAF tank is split into two primary sections: the “contact” and the “clarification”
zones. The first, as the name suggests, is where the air is released from the air-saturated
recycle flow through an arrangement of headers and nozzles or needle valves, forming a
profusion of micro bubbles which, as they rise, intimately mix and attach to the floc carried
through by the bulk flow. The total flow including micro bubbles and algal cell flocs
normally exit the contact zone over a baffle, generally referred to as the “incline baffle”,
which in practice can actually be vertical. The design of this baffle forming the downstream
boundary of the contact zone is critical to the design of the operation of the DAF tank as poor
“contact” in this first zone will result in poor performance overall. The actual average
retention time in the contact zone is typically ~60 seconds. The second zone in conventional
designs is typically ~80% of the total tank volume and is where the air bubble agglomerates,
with a density lower than water, formed via the contact zone are allowed time to rise to the
surface where algal bloom agglomerates accumulate as floated sludge (Figure 9.18). The
float is removed through a mechanical skimming unit (mechanical removal) or by solids
overflow to the collection through (hydraulic removal). Mechanical removal results in a
waste stream with solids concentration of 2 – 3%. The hydraulic removal produces
wastewater with lower solids concentration in the rate of 0.5 – 1%. Typically, DAF tanks are
covered to prevent disturbance of the float from wind and rain.
The excess air at this point (there should always be excess air) provides general buoyancy to
the floated sludge and together with the hydrodynamic flows set up by the design, acts as a
barrier preventing short circuiting and as a “filter”. This filter layer is what is sometimes
called the “whitewater” layer and comprises a range of bubbles sizes that are continually
moving both vertically and horizontally. It is the management through proprietary designs of
this whitewater layer that can impact the overall performance of the DAF system as the
“whitewater” layer serves a number of functions and is not simply providing air to float flocs
to the surface. It is for this reason that the use of the air:solids ratio to determine air dose (as
275
used in sludge thickening applications) is not appropriate for seawater conditions, a TSS of
several hundred is considered very high for SWRO pretreatment applications.
The size of bubbles regarded as ideal for freshwater DAF applications have traditionally been
in the range of 10-100 µm with most having sizes of 40-80 µm (Edzwald 2010). This appears
to be the case for desalination applications as Kim et al. (2011) also reported bubble sizes in
the range of 10-100 µm when using seawater. It was reported that the mean bubble size in
saline water was smaller than that in fresh water because of such factors as higher surface
tension, higher ionic strength, and higher density of seawater (Besson and Guiraud 2012).
The concentration of air bubbles (bubble density) will vary depending on a number of factors
including sizes. These factors can include, but not be limited to, the recycle rate (that can
vary typically in the range of 6-20% on a volumetric basis), recycle pressure, the saturator
efficiency and temperature. However, for a typical system delivering the equivalent of ~8 g
air/m3 of throughput, the number of bubbles can be in the range of 1.8 – 2.5 x 105/mL.
Smaller bubbles in seawater DAF systems may slightly offset the negative effects of salinity
on the air demand. The smaller the bubble size (and the lower the water temperature), the
slower the rise rate of the bubble, and thus, a larger flotation tank is required to allow bubbles
to reach the surface (Gregory et al. 1999).
There are various DAF designs on the market and these range from what some may describe
as horizontal, where the flow enters at one end (Figure 9.18) or from the center, and then
flows along the tank length or radius to the outlet normally via an underflow baffle or a series
of collector pipes. An alternative to this approach is combined dual media gravity filters with
DAF which include the proprietary systems such as CoCo™ and Enflo-Filt™ or generic type
called stack DAF, in-filter DAF or DAFF (Figure 9.19). These systems offer the end-user the
advantage of space savings; however, the operation of the DAF in terms of loading rate is
restricted by the limits placed on the filter and the physical property of an air bubble. Air
bubbles with average diameters of 40-60 µm would have a rise rate in the range of 3-7 m/h,
respectively (with large bubbles of
100 µm reaching rise rates of 20
m/h). This means that the higher
net flotation rates now being
utilized in high-rate DAF of 30-50
m/h cannot be used because of the
problems associated with air being
drawn into the filter bed causing
air blinding. There is also the lack
of available hydraulic driving
head required for flow to pass
through the filter to match the
higher DAF rates. Moreover, the
whole system (including the
flotation cell) will need to be
taken offline for 15 to 20 minutes
during a filter backwash event.
When no bloom, oil, or grease are
present in the feedwater, some
systems provide pipework to
bypass the DAF. Some DAFF
Figure 9.19. Combined DAF clarifier and granular media systems will operate in direct filter
filter. mode when blooms are not present.
276
The typical frequency of float removal tends to be site specific ranging from continuous to
intermittent. This is true regardless of whether the sludge removal method is mechanical or
hydraulic. The resulting sludge from the DAF tank can be dewatered and thickened
separately but is generally mixed with the settled GMF or UF backwash water before
thickening and dewatering by centrifuge or plate press. The clarified water following sludge
treatment is typically returned to sea with the SWRO brine.
When designing a DAF system according to hydraulic loading, the influent flow is
considered as the upstream flocculation system and excludes recycle flow. The tank area can
then be designed according to the preferred hydraulic loading rate. The area calculated is that
of the contact and clarification zones, whereby the velocity of the water flowing towards the
base of the tank must be less than the rising velocity of floc-bubble aggregates to ensure
separation (Edzwald et al., 2010). The area of the adjacent contact zone must be designed to
ensure a residence time of 1-2.5 minutes. It is not always clear whether the hydraulic loading
rate reported is for the clarification zone only.
9.5.4 DAF in SWRO pretreatment for removal of marine algae
A study by Haarhoff and Edzwald (2013), examining the differences in the application of
DAF in seawater compared to freshwater, found that contact and separations zones were not
significantly different. The largest difference was due to the lower solubility of air in
seawater compared to freshwater. The dynamic viscosity, density and surface tension are all
higher in seawater and though these differences are small at +8%, +3% and +1% respectively
at a salinity of 35 g/kg and temperature of 20 °C (Haarhoff and Edzwald 2013) as viscosity
cannot be ignored when designs are proposed where the operational load is high, i.e. typically
≥ 20 m/h. The reason why viscosity cannot be ignored at these elevated levels is because of
the additional drag on the floc-bubble aggregates and the overall supporting whitewater layer
as described by Amato et al. (2012). This ultimately impacts on the tank design, particularly
in terms of its overall depth, which needs to be slightly deeper when considering the
treatment of seawater. This will not always be necessary and will depend upon other factors
such as the method used to draw off the subnatant and any water quality requirements
imposed on the DAF product flow. At a minimum, an additional 30 cm of tank depth might
be needed, as reported by Amato et al. (2012).
The percentage removal of algae in fresh water applications is typically 90-99% (Yap et al.
2012; Zhu et al. 2014) while practical experience shows that when used for seawater
pretreatment, DAF systems usually yield significantly lower algal removal rates – 40 to 50%
(Voutchkov 2013). Actual removal can be as low as 40%; however, as it is dependent on the
cell concentration in the feed, the type and size of algal cells, and phase of growth (lag,
exponential or stationary phase), due to differing amounts of AOM in solution (Henderson et
al. 2008b). Oil and grease removal are difficult to quantify because of the three states in
which it may be found, i.e. suspended (easy), emulsified, and dissolved (both more difficult).
Therefore, in the case of oil, any removal rates need quantification by jar testing. Efficiency
of turbidity reduction is a very poor parameter as it is not unusual to have higher turbidities
post DAF than what may have been found in the raw. True turbidity removal efficiency can
only be reported when the feed to the DAF is measured after flocculation.
9.5.5 Optimization of process parameters for marine HABs
While removal of algal cells via DAF is important during a bloom, operators must also
consider the presence of associated AOM that may also deteriorate the water quality. This
AOM may interfere with coagulation or reach downstream RO membranes and cause
biofouling. It is therefore important that coagulation conditions are optimized, not just for
277
cell removal, but also for simultaneous AOM removal. Application of coagulation-
flocculation-DAF to remove AOM of larger molecular size (e.g., biopolymers and humic-
type substances), in combination with biofiltration in GMF will maximize organic matter
removal to protect the downstream SWRO membranes from accelerated biofouling (Shutova
et al. 2016; Naidu et al. 2013). (This is discussed further in Chapter 2 and Section 9.4).
9.5.6 Summary
DAF is useful in removing algal cells and reducing the suspended solids load for downstream
pretreatment, as the process is ‘gentle’ or low shear, preventing cell lysis and release of
fouling organics and algal toxins, thereby reducing the risk of AOM fouling of UF and RO
membranes. Practical full-scale experience to date with using DAF for pretreatment of
seawater is very limited and have shown a wide range of algal removal efficiency; however,
with optimization, DAF can be very effective as an SWRO pre-treatment, removing up to
99% of cells when preceding coagulation-flocculation is optimized. This removal efficiency
depends on many factors – one of the key factors is the content of suspended solids in the
source water. Designers of DAF for seawater plants should take care to consider salt water-
specific parameters such as the “salting out” affect that results in either a requirement for an
increase in pressure (~30%) or recycle rate (~20%). If the DAF is only operational
periodically, operators could consider bringing the DAF online while cell counts are low so
that the plant is fully operational when counts increase. This argues for plankton monitoring
in the vicinity of the plant (Chapter 3) and within the plant (Chapter 5) so that effective
actions can be taken sufficiently early to minimize clogging and fouling.
9.6 GRANULAR MEDIA FILTRATION
9.6.1 Overview
Conventional pretreatment using coagulation, flocculation, and filtration through granular
media is the most commonly used source water pretreatment process for SWRO desalination
plants today (other than cartridge filtration). This process includes filtration of the source
seawater through one or more layers of granular media (e.g. anthracite coal, silica sand,
garnet).
In the GMF process, suspended solids are removed through attachment to the filtration media
particles and through blockage/capture by the filtration cake. Organics may be removed by
biofiltration occurring in the filter (e.g., the utilization and breakdown of organic materials by
microbes). The preferred process of filtration is capture of suspended solids with bed
penetration (depth filtration) as opposed to surface filtration, since the latter results in a
significantly faster increase of pressure loss and therefore shorter filter runs. During algal
bloom events, coagulant overdosing may occur as described in Section 9.4.3 resulting in
accumulation of large flocs at the filter surface, blinding the filter bed if the blooms are very
intense.
Conventional filters used in SWRO pretreatment are typically rapid dual-media (anthracite
and sand) filters (DMF) in a single-stage configuration; however, in some cases where the
source water contains high levels of organics (total organic carbon (TOC) > 6 mg/L) and
suspended solids (monthly average turbidity > 20 NTU/TSS > 30 mg/L), two-stage filtration
systems are applied to achieve desired SDI levels. Under this configuration, the first filtration
stage is mainly designed to remove macroalgae, solids, and organics that are present in
suspended form. Often when a plant is subject to HABs, coagulation is employed in the first
stage filtration. The second-stage filters are configured to retain fine solids (including HAB
cells) and silt, and to remove a portion (20 to 50 %) of the soluble organics contained in the
saline water by biofiltration.
278
Depending on the driving force for water filtration, GMF systems are classified as gravity or
pressure filters. The main differences between the two are the head required to convey the
water through the media bed, the filtration rate, and the type of vessel used to contain the
filter media. Because of the high cost of constructing large pressure vessels with proper
wetted surfaces for corrosion resistance, pressure filters are typically used for small and
medium size capacity SWRO plants. Gravity pretreatment filters are used for both small and
large SWRO desalination plants.
9.6.2 The filter operation cycle
GMF is a cyclical process, which incorporates two sequential modes of operation: 1) source
water processing (filtration) mode; and 2) filter media backwash mode. During the filtration
cycle the water moves in the direction of decreasing size gradation of the media and solids in
the water are retained on and around the media grains.
As the feed water is filtered through the media, the content of solids and silt in this water
decreases. Well-operating filters typically remove 90 to 99 % of the solids and silt in the
source seawater (to an approximate size of 10 µm). Some of the marine microorganisms in
the source water are also retained on the filter media forming biofilm around the filtration
media granules. These microorganisms may consume a portion of the dissolved AOM from
the source seawater such as biopolymers and TEP through biofiltration. The organic load
removal efficiency of the filters is a function of four main factors: media depth, surface
loading rate, coagulant concentration, and temperature. Removal of organics by the filters
increases with depth and temperature and with the decrease of the filter-loading rate.
The solids retained in the pore volume between the filter grains reduce this volume over time
and create hydraulic losses through the filter media (filter bed resistance). Most filters used in
SWRO pretreatment operate at constant filtration rate, which means that the feed pressure of
these filters increases over the filtration cycle to compensate for the head losses in the filter
bed caused by accumulation of solids. Once the filter media head losses reach a certain preset
maximum level, the filter is taken out of service and media backwash is activated. Deeper
filters or larger surface area have larger capacity to retain solids and therefore, usually have
longer filtration cycles.
Typical parameters used to monitor pretreatment GMF performance are SDI, turbidity, TOC
and iron. Usually, turbidity of the feed and filtered seawater is measured continuously with
online turbidity meters. For larger plants SDI15 of the filtered water may also be measured on
line. The measurement frequency of TOC may be increased during an algal bloom event,
especially when TOC exceeds 2 mg/L, chlorophyll a increases over 1 µg/L and/or algal
counts exceed one or two million cells/L. Usually, if TOC, chlorophyll a or algal counts are
below these levels, the algal bloom is not expected to have a major impact on pretreatment
system operation and RO fouling. If these source water quality parameters exceed the above-
mentioned thresholds, usually plant operators institute algal bloom mitigation strategies such
as increasing the dose of coagulant fed to the source seawater, increasing acid addition in
order to decrease pH, and thereby enhance algal removal, and/or decrease surface loading
rate to enhance filter retention time and encourage biofiltration. (Note, however, that cell
counts can be deceiving – one million cells of a small species will constitute a much lower
cell volume or biomass than the same number of a larger species. Ideally, operators need to
learn the species and cell concentrations of those species that cause problems, and the
pretreatment strategies that were effective for those conditions. Simply stated - not all algal
blooms are the same, so local experience needs to be documented and applied.
279
GMF are typically backwashed using filtered seawater or concentrate from the SWRO
membrane system. Normally filter cell backwash frequency is once every 24 to 48 hours and
spent (waste) backwash volume is 2 to 10 % of the intake seawater. During the severe 2008
algal bloom in the Arabian Gulf, backwash intervals at the Fujairah 1 plant in the UAE were
dramatically reduced to 2 hours from 24 hours.
Use of SWRO concentrate instead of filtered effluent to backwash filter cells allows for a
reduction in backwash volumes and a reduction in the energy needed to pump source water to
the desalination plant; however, use of concentrate for filter backwash during algal blooms is
not recommended because the concentrate will have an elevated content of AOM and
biodegradable organics, which will not benefit the pretreatment process and may exacerbate
RO membrane fouling. Additionally, osmotic shock from the higher salinity may cause cell
lysis in some cases, releasing additional AOM.
During backwash of down-flow filters, the backwash water flows upwards through the filters,
scours the filter grains, removes the solids accumulated on the filter grains, expands the filter
bed, and transports the removed solids towards the backwash troughs. From experience, it is
known that backwashing of filter media grains smaller than 0.8 mm with water only is
inefficient. Therefore, a typical backwash regime currently includes a combination/sequence
of air and water washing. Air creates greater turbulence and enhances particle scrubbing. The
length of water and air backwashing cycles is a function of the solids content in the source
water and the depth of media bed and typically is between 5 and 15 minutes.
GMF pretreatment efficiency is very dependent on the presence of a biofilm and the
formation of a matrix of small particles around the granular filtration media that are needed
to remove fine particles. Formation of such biofilm and filtration matrix is referred to as
“filter cell maturation”. Every time the filters are backwashed for removal of the residuals
accumulated during the filtration process, a portion of the biofilm and solids matrix around
the filtration media grains are removed and as a result, when the filter is put back in service
after backwash, it usually does not produce pretreated water of quality compliant with the
target SDI and turbidity values. It usually takes between 15 and 45 minutes after a filter cell
is returned to service for the fine solids matrix to form to its previous level and for the
backwashed filter to begin producing pretreated water of adequate quality. During this filter
cell maturation period, the out-of-specification filtrate is usually discharged to the plant
outfall.
Filter media type, uniformity, size, and depth are of key importance for the performance of
pretreatment filters. Characteristics of media commonly used in SWRO desalination plants
are presented in Table 9.1.
Single media filters (mono media) are not commonly used in SWRO pretreatment because of
their limited ability to perform under varying source water conditions. In mono media filters,
the fine size filtration media particles tend to aggregate at the top of the bed after a number of
backwash runs. This reduces penetration of suspended solids and, therefore, mainly results in
surface bed filtration. Typically, such filters could be used for desalination plants with
subsurface intakes producing turbidity of < 2 NTU, TSS of < 5 mg/L and SDI15 < 5 (see the
Sur plant in Oman, Chapter 6). A large-scale application of single media (0.7 mm sand)
filters using an open onshore intake is at the Tampa Bay desalination plant in Florida (see
Case Study 11.9).
A graduation of the filtration bed from coarse to fine particles can be achieved in dual media
configuration by placing fine, high specific gravity filtration media as the lower filtration
layer and coarse, low specific gravity filtration media as a top layer. Filtration media
280
selection that provides coarse to fine filtration bed configuration includes 0.4 to 1 m
anthracite or pumice as a top layer and 0.4 to 2.0 m of silica sand as a bottom filtration layer.
DMF is the most commonly used GMF in SWRO plants worldwide. Deep DMF are often
used if the desalination plant filtration system is designed to achieve enhanced removal of
soluble organics from source water by biofiltration. In this case, the depth of the anthracite
level is enhanced to between 1.5 and 1.8 m. In comparison to mono media filters, DMF
systems operate at higher filtration rates, have longer filter run times, and produce less
backwash water.
Tri-media filters are not commonly used in SWRO pretreatment and are primarily used for
capturing small-size phytoplankton and fine silt that cannot be well retained by the top two
layers in DMF. Tri-media filters typically comprise 0.45 to 0.6 m of anthracite as the top
layer, which retains large-size algae (i.e. algae over 100 µm), 0.2 to 0.4 m of sand as a middle
layer to remove medium-size algae (20 to 100 µm) and 0.10 to 0.15 m of garnet or limonite
as the bottom layer. The third (garnet or limonite) layer of filtration is usually used only if the
source water contains a large amount of very fine silt or the source water intake experiences
algal blooms dominated by small algae (0.2 to 2 µm).
Since the cost of filter cells increases with depth, often instead of a deep, single tri-media
gravity filter, a combination of coarser media (anthracite-sand) gravity filter followed by a
pressure filter containing finer (sand and garnet) is used.
Most filters used in seawater pretreatment are down-flow filters. This flow direction allows
large algal particles to be retained at the top of the filter media and removed with the
backwash water with minimum breakage and release of organics. If upflow filtration is used,
algae contained in the source water are pressed against the filter media and unwanted
dissolved organics such as algal biopolymers may be released from the broken algal cells into
the filtered water, which is undesirable as it can exacerbate biofouling of the downstream
SWRO membranes.
Table 9.1. Typical media characteristics for GMF used in SWRO plants.
Typical Effective Specific Density

Media Type
Grain Size - mm tons/m3
Pumice 0.8 – 2.0 1.2
Anthracite 0.8 – 2.0 1.4 – 1.7
Silica Sand 0.4 – 0.8 2.60 – 2.65
Garnet 0.2 – 0.6 3.50 – 4.30
9.6.3 Single and two-stage filtration

Two-stage filtration is typically used when the source water contains high levels of turbidity
(usually above 20 NTU) and organics (TOC > 6 mg/L) for long periods of time (i.e.,
weeks/month). Such conditions occur in desalination plant intake areas exposed to prolonged
red-tide events (which sometimes could last for several months) or in river estuaries, which
are exposed to an elevated turbidity levels occurring during the wet-season of the year.
Two stage filtration systems typically consist of coarse (roughing) filters and fine (polishing)
filters operated in series. Usually the first stage filter is a mono-media type (i.e., coarse sand
or anthracite) or dual media while the second stage filter is configured as a DMF with design
criteria described in the previous section. The first (coarse-media) filter typically removes 60
to 80% of the total amount of solids contained in the source water and is designed to retain all
large debris and floating algal biomass. The second stage filter removes over 99% of the
281
remaining solids and fine silt as well as the microalgae contained in the source seawater,
typically producing effluent turbidity of less than 0.05 NTU.
Two stage filters have several advantages. The filtration process through the coarse media
filters not only removes large particulate foulants, but also enhances coagulation of the fine
particulates contained in the source water, which makes their removal in the second-stage
filters less difficult and allows the second-stage filters to be designed as shallow-bed rather
than deep-bed filters and to operate at higher surface loading rates. This benefit results in
reduced size of the DMF and in a lower total amount of coagulant needed to achieve the
same final filter effluent water quality, as compared to single-stage DMF.
Two other benefits of the two-stage filters are that: 1) they can handle larger fluctuations of
intake source water turbidity because of the larger total filter media volume/solids retention
capacity; 2) if the second stage filters are designed as deep-bed (rather than shallow bed)
filters they can achieve enhanced TOC and AOM removal by biofiltration. While deep,
single-stage dual media filters can typically reduce 20 to 30 % of the TOC contained in the
source seawater, the two-stage systems with deep second-stage filters can achieve 40 to 60 %
TOC removal, mainly due to enhanced fine particle coagulation and biofiltration.
It should also be pointed out that if the filters are designed to achieve TOC removal by
biofiltration, it would take at least four to six weeks for the filters to accumulate sustainable
biofilm on the surface of the filter media to yield steady and consistent filter performance and
TOC removal of 10 to 20%. If the source water temperature is relatively cold (i.e. below 15
o
C), then the biofilm formation process may take several weeks longer.
9.6.4 Gravity filters
Typically, gravity filters are reinforced concrete structures that operate a water pressure drop
through the media of between 1.8 and 3.0 m. The hydrostatic pressure over the filter bed
provides the force needed to overcome the head loss in the media. Single-stage down-flow
gravity DMF filters are the predominant type of filtration pretreatment technology used in
desalination plants of capacity higher than 40,000 m3/day. Table 9.2 provides examples of
key design criteria for gravity filters at SWRO desalination plants of various size and water
quality. Some of the largest SWRO desalination plants in the world in operation today such
as the 325,000 m³/day Ashkelon SWRO plant are gravity single-stage DMF.
Seawater always contains a
measurable amount of algae,
with the concentration usually
increasing several times during
the summer period and possibly
increasing 10 times or more
during periods of algal blooms
(which may or may not exhibit
themselves as HABs). There are
thousands of algal species in the
seawater, as discussed in
Chapter 1, so generalizations
like this should be viewed with
caution.
Gravity filters (Figure 9.20) are
Figure 9.20. Gravity filters protected with plastic covers for control typically covered with light
of algal growth showing covers of gravity filters in Ashkelon, Israel. plastic covers that protect the
282
filter cells from direct sunlight to

prevent algal growth or are installed
in buildings (Figure 9.21).
An alternate method to aid control
of SWRO biofouling is the
installation of a granular-activated
carbon media layer (“activated
carbon cap”) on the surface of the
filters to enhance the removal of
some of the polysaccharides and
other organics in the source water.
This approach has been tested in
full-scale applications in the Middle
Figure 9.21. Single-stage dual media gravity filters at the Gold Coast East.
SWRO Desalination Plant
Table 9.2. Examples of large SWRO desalination plants with DMF gravity filters.
Desalination plant Pretreatment Average and

location and system maximum filter Notes
capacity configuration loading rates
Ashkelon SWRO 40 single-stage 10/12 m/h (avg./max) Open intake –
Plant, Israel – 1,000 m from shore
325,000 m³/day
Sydney SWRO Plant, 24 single-stage 8/12 m/h (avg./max) Open intake –

Australia – 300 m from shore
250,000 m³/day
Fujairah 1 SWRO 14 Single-stage Filtration Rate - Shallow offshore

Plant, UAE – 8.5 m/h (avg.) open intake. High
170,000 m³/day 9.5 m/h (max) bloom potential
Fujairah 2 SWRO 16 DAFs, 12 - DAF rate – Shallow offshore

Plant, UAE – Single-stage 21 m/h (avg.) open intake. High
136,000 m³/day 30 m/h (max) bloom potential
Filtration Rate – DAF operated only

10.5 m/h (avg.) if turbidity > 5 NTU
12.5 m/h (max)
Gold Coast SWRO 18 single-stage 8/10 m/h (avg./max) Open intake –

Plant, Australia – 1,500 m from shore
125,000 m³/day
283
9.6.5 Pressure filters

Compared to gravity media filters that operate under a maximum water level over the filter bed of
up to 3 m, pressure filters typically run at feed pressure equivalent to 15 to 30 m of water column,
which has the potential to damage HAB cells. Therefore, pressure filters may have the
disadvantage of causing accelerated biofouling when filtering source water with very high algal
content. This effect is likely to manifest itself mainly during algal blooms when the level of TOC
in the source water exceeds 2 mg/L.
Pressure filters are used in medium- and large-size desalination plants in Spain, Algeria, and
Australia; however, in most successful applications, the source water quality is very good (TOC
< 1 mg/L, SDI15 < 4 and turbidity < 4 NTU). In addition, the Spanish desalination plant intakes
are relatively deep and the algal content in the source water is commonly fairly low. Hence, the
ingress of algae is lower and biofouling caused by breakage and decay of algal cells may not be
as significant problem as it would be for shallow or near-shore open intakes (see Chapter 6).
Gravity media filters have a two- to three-times larger volume of filtration media and retention
time than pressure filters for the same water production capacity. This is a benefit for plants
exposed to algal blooms because source water turbidity and TSS could increase several times
during algal bloom events. The higher solids retention capacity allows the gravity filters to handle
such increases without decreasing the length of the filter cycle. Higher hydraulic retention
capacity of the gravity filters allow these filters to develop a more robust biofiltration layer near
the bottom of the sand media in the filters, which in turn results in more effective filtration.
Pressure filters usually do not handle solids/turbidity spikes as well because of their smaller
solids retention capacity (i.e. smaller volume of media pores that can store solids before the filter
needs to be backwashed). If the source water is likely to experience occasional spikes of high
turbidity (20 NTU or higher) due to rain events, HABs, shipping traffic, or ocean bottom
dredging operations in the vicinity of the intake, seasonal change in underwater current direction,
or spring upwelling of water from the bottom to the surface, then pressure filters will produce
effluent with inferior effluent quality (SDI15 and turbidity) during such events and, therefore, their
use would likely result in a more frequent RO cleaning.
Pressure filters have filter bed configurations similar to that of gravity filters, except that the
filter media is contained in steel pressure vessels. They have found application mainly for
small- and medium-sized seawater desalination plants – usually with production capacity of
less than 20,000 m3/d. There are, however, a number of installations worldwide where
pressure filters are used for pretreatment of significantly larger volumes of water (see Table
9.3).
In most cases for good source water quality (SDI15 < 4 and turbidity < 4 NTU), pressure
filters are designed as single stage, dual media (anthracite and sand) units. Some plants with
relatively poor water quality use two-stage pressure filtration systems. Pressure filters are
available in two vessel configurations – vertical and horizontal. Vertical pressure filters
(Figure 9.22) are customarily used in smaller plants and individual vessels have maximum
diameter of 3 m. Horizontal pressure filters (Figure 9.23) are used more frequently in
desalination plants and are more popular for medium and large-size plants. One example of a
desalination plant using horizontal pressure GMF for seawater pretreatment is the 140,000
m³/day Kwinana SWRO plant in Perth, Australia (Figure 9.23). Horizontal filters allow larger
filtration area per filter vessel compared to vertical units; however, usually vertical vessels
can be designed with deeper filter media, if deep filters are needed to handle spikes of source
seawater turbidity.
284
Table 9.3. Large seawater desalination plants with pressure granular media filters.
Desalination plant Pretreatment Average and
location and system maximum filter Notes
capacity configuration loading rates
Al Dur SWRO Plant, DAF followed by DAF surface loading Shallow offshore
Bahrain horizontal pressure Rate – 25 to open intake in algal
218,000 m3/day filters 30m³/m².h bloom prone area
Pressure filter rate –
18 to 24 m³/m².h
Barcelona SWRO DAF followed by DAF surface Loading Deep offshore open
Plant gravity filters and Rate – 25 to intake in industrial
200,000 m3/day horizontal pressure 30m³/m².h port and near river
filters Gravity filter rate – 8 estuary
to 10 m³/m².h
Pressure filter rate –
15 to 20 m³/m².h
Kwinana SWRO 24 single-stage dual 14.0/18 m³/m².h Shallow open intake

Plant, Perth, Australia media pressure (avg./max)
140,000 m³/day filters
Carboneras SWRO 40 single-stage dual 12.0/15.0 m³/m².h Offshore open

Plant, Spain media pressure (avg./max) intake
120,000 m³/day filters
El Coloso SWRO DAF followed by DAF surface loading Open intake in

Plant, Chile two-stage dual rate – 22 to industrial port with
45,400 m³/day media horizontal 33m³/m².h. Filter rate frequent red tides
pressure filters -25 m³/m².h
Since pressure filters are completely

enclosed, sunlight cannot reach the
filter weirs, distribution system and
media and induce green algal
growth that would have negative
impact on filter performance. The
visual indications that can be
observed for gravity filters are
difficult or impossible to perform
for pressure filters. Additionally,
flow distribution into individual
filters in a bank of filters may not
be easily controlled compared
with gravity filters, where equal
flow splitting can be achieved
Figure 9.22. Vertical pressure pretreatment filters capacity. through gravity.
Photo: Voutchkov 2013.
285
9.6.6 GMF filter

performance
The purpose of the pretreatment
filters for SWRO plants is not
only to remove over 99% of all
suspended solids in the source
water, but also to reduce the
content of the much finer silt
particles by several orders of
magnitude. Therefore, the
design of these pretreatment
facilities is usually governed by
the filter effluent SDI15 target
level rather than by target
Figure 9.23. Single-stage horizontal -pressure dual media filters at turbidity or pathogen removal
the Kwinana SWRO Plant. Photo: Water Corporation. rates. Full-scale experience at
many GMF plants indicates that
filters can consistently reduce source water turbidity to less than 0.1 NTU, and SDI15 less
than 2-4 depending on source water quality.
The rate of algal removal by the filters will depend mainly upon the size of the algae and the
size of the filter media. Most algae are typically retained on the surface of the top media
(anthracite/pumice) (Bar-Zeev et al. 2012). Depending upon the size of media and size of the
algae dominating in the source water, algal removal could typically vary between 20% and
90% or more. Based on recent research (Gustalli et al. 2013), UF membranes provide better
removal of algae than granular dual media filters (99%+ vs 74%).
Results of a comprehensive study of the effect of bloom intensity on GMF algal removal
efficiency indicates that the more intense the bloom, the lower the level of algal removal by
the filters (Plantier at al. 2013). After seven hours of filtration, the overall efficiency of the
tested GMF for a light algal bloom (30,000,000 algal cells/L) dominated by microalgae was
74%, while for a severe algal bloom (150,000,000 algal cells/L) this removal efficiency was
reduced to 49%. Another important observation of this study was that the first 30 cm of the
filter media retain more algae than the rest of the media, and that most of the algal retention
occurs over the first three hours of the filtration cycle.
Desalination pretreatment GMF would typically provide 99% (2 logs) of removal of
pathogens, but sometimes may have lower removal rates in terms of marine bacteria because
these bacteria are typically of small size and can pass through the filters. It is interesting to
note that while UF filters have two to three order of magnitude higher removal rates of
marine bacteria, such removal is inadequate to prevent heavy biofouling of the downstream
SWRO elements if sufficient amounts of bioavailable organics are present in the water.
Typical gravity and pressure dual media filters of conventional filter bed depth of 1.0 to
1.4 m have relatively low organic removal rates – 15 to 20% in mature filters. Algal
biopolymer removal of 18% was reported in the Barcelona DAF-DMF pilot study (see Case
Study 11.11) and TEP removals of 20 to 90% was reported in GMF (Bar-Zeev et al. 2012).
The removal rate, however, increases significantly with depth and could reach 25 to 35% for
filters with a total filter depth of 2.0 m or more. If a carbon cap is installed on the top of the
filter media (above the layer of anthracite), TOC removal rate could be increased to 40 to
50%, although the performance of this removal requires monitoring over time as the removal
capacity will become exhausted.
286
GMF with a total filter depth of 1.4 to 1.6 m typically removes 10 to 20% of the TOC, AOC
and AOM in the source seawater (Voutchkov 2013). Deep gravity media filters (filter depth
of 2.0 m or more) can remove over 40% percent of the TOC, AOC and AOM in the source
water. Such filters develop a biofiltration zone at a depth of 1.6 to 2.0 m, which significantly
enhances removal of dissolved organics contained in the source seawater.
9.7 MICROSCREENS FOR MEMBRANE PRETREATMENT
9.7.1 Overview
For SWRO desalination plants with membrane pretreatment (UF/MF), the seawater has to be
prescreened to remove very fine (50 to 500 µm) sharp particles (e.g. broken shells) which
could puncture the plastic membranes and compromise their integrity and performance.
Generally, there are two types of microscreens; micro-strainers or disk filters. These devices
can experience complete fouling of the screens that permanently builds differential pressure.
Additionally, fouling due to macroalgae and jellyfish can be experienced during
microscreening. The operation of microscreens during HABs is described in this section.
9.7.2 Types and configurations
Typically, microscreens, which could be the micro-strainer (Figure 9.24) or disk filter (Figure
9.25) type, are used in SWRO for
large particle removal (50 to 500
µm). SWRO desalination plants
with microscreens are also usually
equipped with conventional coarse
screens or a combination of coarse
and fine screens, which retain
debris >10mm upstream at the
seawater intake.
Most self-cleaning micro-strainers
consist of screens with small
openings located in a filtration
chamber. The source water enters
through the inner side of the
Figure 9.24. Self-cleaning micro-strainers. Photo: Voutchkov strainers, moves radially out
(2013).
through the screens and exits
through the outlet. The gradual
buildup of solids on the inner
Filter in Filter in
Filtration Mode Backwash Mode surface of the screens creates a
Inlet Inlet filter cake, which increases
differential pressure between
intake and filtered water overtime.
When the differential pressure
reaches a certain preset value, the
deposited solids are removed by
Drain jet of backwash water. The self-
cleaning process typically takes 30
Outlet to 40 seconds.
Figure 9.25. Disk filter – modes of operation. Source: Voutchkov Disk filters are equipped with
(2013). polypropylene disks, which are
diagonally grooved on both sides
287
to a specific micron size. A series of these disks are stacked and compressed on a specially
designed spine. The groove on the top of the disks runs opposite to the groove below,
creating a filtration element with series of valleys and traps for source water debris. The stack
is enclosed in corrosion and pressure-resistant housing.
During the filtration process, the filtration disks are tightly compressed together by the
spring's power and the differential pressure, thus providing high filtration efficiency.
Filtration occurs while water is percolating from the peripheral end to the core of the element.
Source water debris and aquatic organisms (mainly phytoplankton) with a size smaller than
the size of the microscreens (generally 80 to 120 µm) are retained and accumulated in the
cavity between the filter disks and the outer shell of the filters, thereby increasing the head
loss through the filters. Once the filter headloss reaches a preset maximum level (typically
0.35 bar or less) the filters enter backwash mode. Considering the aperture of the screens,
significant retention of algal cells could occur during a bloom. All debris (algal or otherwise)
retained on the outer side of the filters is then flushed by tangential water jets of filtered
seawater flow under 0.15 to 0.2 bar of pressure and the flush water is directed to a pipe,
which returns the debris and marine organisms retained on the filters back to the ocean
through the plant ocean outfall – usually along with the concentrate. In some cases, the debris
collected on the microscreens is disposed of through a separate pipe in the vicinity of the
intake area. Hence during a bloom, algal cell wall debris and suspended solids could be re-
entrained into the intake. Therefore, during a bloom, redirection of this waste stream could be
considered to prevent downstream impacts.
Because of the relatively low differential pressure at which these filters operate, they are
likely to minimize impingement of the marine organisms in the source water. Furthermore,
since the disk filtration system is equipped with an organism return pipe, the entrained
marine organisms are returned back to the source water body, thereby reducing their net
entrainment.
One of the key issues associated with using membrane pretreatment is that the membrane
fibers can be punctured by sharp objects contained in the source seawater, such as broken
shells or sharp sand particles. In addition, seawater can contain barnacles, which in their
embryonic phase of development are 130 to 150 µm in size and can pass through the screen
openings unless these openings are 120 µm or smaller. Experience at the Southern Seawater
Desalination Plant in Perth, Australia also shows that the source seawater could contain other
marine organisms such as sponges, polychaetes, and diatoms which have diameter of only 3
to 10 µm and could be sharp-enough to puncture membrane fibers (Ransome et al. 2015).
If larval shellfish and barnacles pass the screens, they could attach to the walls of
downstream pretreatment facilities, grow on these walls and ultimately interfere with
pretreatment system operations. Once barnacles establish colonies in the pretreatment
facilities and equipment, they are very difficult to remove and can withstand chlorination,
which is otherwise a very effective biocide for most other marine organisms. Therefore, the
use of fine microscreens or disk filters (80 to 120- µm size) is essential for reliable operation
of the entire seawater desalination plant using membrane pretreatment. Microscreens or disk
filters are not needed for pretreatment systems using granular media filtration because these
systems effectively remove fine particulates and barnacles in all phases of their development.
9.7.3 Algal bloom-related challenges
Full-scale experience shows that, depending upon the severity of the algal bloom and the size
and type of microscreens, these devices face two operational challenges: 1) complete
plugging of the screens which permanently builds differential pressure, triggering continuous
288
backwash; and 2) rapid increase in algal mass accumulation on the screens which causes very
frequent backwashes that in turn interrupt the normal operation of the pretreatment system
and the desalination plant. Such challenges are very frequent in the case of near-shore or
lagoon intakes of the Gulf (such as at Sohar) applying microscreens of size of 100 µm or
smaller during conditions of algal cell abundance in the water of over 100,000 cells/L (see
Case Study 11.3). In the case of the Jacobahaven demonstration plant, algal blooms resulted
in a significant increase in the clogging rate of the smaller aperture 50 µm strainers, reducing
the backwash interval to 5 minutes from 0.5 – 1.5 hour for non-bloom conditions at identical
turbidity (see Case Study 11.10). Similar impacts on microscreens are observed during
conditions of jellyfish outbreaks, strong winds or currents carrying sea grass or seaweeds
(macroalgae) in the vicinity of the intake.
One solution has been to select a larger size microscreen (i.e. screen with pore sizes of
300 µm or higher). While this solution could address the problems during certain algal
blooms and/or jellyfish outbreaks it creates potential problems with sharp particles entering
the membrane pretreatment system and creating micro-punctures on the membranes over
time. When adopting this solution, permanent MF/UF membrane integrity loss has usually
been observed after plant operation of 6 months or more.
Another solution is to use conventional relatively coarse granular gravity media filters instead
of microscreens in order to retain both the elevated content of algal mass and fine sharp
particles during blooms. This approach is being considered for the West Basin Desalination
Project in California, USA. As demonstrated by a comprehensive pilot study at the West
Basin Demonstration Plant (Figure 9.26), long-term side-by-side testing of disk microscreens
and GMF indicated comparable capability of these two pre-filters to remove harmful shell
fragments; however, the deep-bed high-rate GMF provided higher quality filtrate during
periods of poor raw ocean water quality (storm and algal bloom events). Testing showed that
the GMF allowed more sustainable MF permeability and affected an increased MF cleaning
interval compared to disc filtration during an algal bloom event (SPI Engineering 2010).
Figure 9.26. Configuration of West Basin SWRO Demonstration Plant. Figure: SPI Engineering (2010).
289
In order for microscreens to successfully handle conditions of high algal cell concentrations,
jellyfish outbreaks or sea grass plugging, they need to have an effective focused auto-
backwash system that maximizes the backwashing velocity across the strainer face with
moving nozzles that operate over the strainer surface (Figure 9.24) or a variable area mesh
that opens up on backwashing. For example, at the Southern Seawater Desalination Plant in
Perth, Australia the original conventional microscreens had to be replaced by screens with a
focused auto-backwash system to address operational challenges of intense screen plugging
due to large quantity of macroalgal sea grass fragments carried to the intake via underwater
currents (Ransome et al. 2015).
A microsreening technology that has potential to handle severe algal bloom conditions is
microfiber filtration (Eshel et al. 2013). Microfiber filters consist of cartridges with
multilayer textile threads, usually with pore sizes of 2 to 20 µm. While their ability to retain
algae, organic matter and mineral solids is comparable to cartridge filters, they have the
benefit that they can be backwashed automatically, similar to conventional microscreens.
Full-scale test experiments of this technology on freshwater during algal bloom conditions in
a lake have showed significant reductions in TEP (mean 47±21%), chlorophyll (mean
90±6%), total suspended solids (67±7%), turbidity (89±5%), and particles larger in size than
3 µm (93±4%). Such systems have not yet found full-scale implementation for seawater
screening.
9.7.4 Summary
Fouling of microscreens due to HABs cause two main operational challenges: 1) complete
plugging of the microscreens which permanently builds differential pressure triggering
continuous backwash; and 2) rapid increase in algal mass accumulation on the screens which
causes frequent backwashes that in turn interrupt the normal operation of the pretreatment
system and the desalination plant. Monitoring of the time between backwashes in
microscreen systems is therefore an important consideration during bloom periods.
9.8 MICROFILTRATION/ ULTRAFILTRATION
9.8.1 Overview
MF/UF membranes have been tested and applied as pretreatment for SWRO membranes for
more than a decade (Wolf et al. 2005; Halpern et al. 2005). MF/UF membranes offer several
advantages over GMF in SWRO pretreatment, namely, smaller footprint, consistently high
permeate quality in terms of SDI, higher retention of organic macromolecules (including
some AOM), lower overall chemical consumption (Wilf and Schierach 2001; Pearce 2010).
Although, MF and UF are both applied in SWRO, UF is often employed due to its smaller
pore size and better removal of particulate/colloidal organics, silt, and pathogens from
seawater (Voutchkov 2009).
9.8.2 MF/UF filtration modes
In contrast to RO membranes that operate in cross flow, MF/UF membranes are typically
operated in dead-end mode in two main configurations based on feed flow direction; PDI,
and pressure driven outside-in (PDO). MF/UF membranes in outside-in configuration may
also be operated in vacuum driven (submerged) mode, which are commonly polyvinyledene
fluoride (PVDF). PDI membranes are generally based on polyethersulphone (PES) and
blends thereof, although some PDI membranes based on cellulose triacetate and
polysulphone are also available in the market (Pearce 2007). In PDI filtration, feedwater
enters through the inside of the element and permeate is collected from the outside of the
capillaries. Hydraulic cleaning (or backwash) is performed by reversing the flow, whereby
product water flows through the membranes from the outside of the capillary, physically
290
lifting the accumulated material from the membrane surface and flushing it to the waste
stream. PDI membranes may be manufactured as single capillaries (hollow fibers) or multi-
channel tubes. Multi-channel tubes combine several individual capillaries in a robust fiber,
usually in a honeycomb arrangement. This configuration is considered to have higher
stability and breaking resistance as compared to single capillaries.
PDI UF membranes based on PES have seen a substantial growth in installed capacity due to
several large SWRO desalination projects, e.g., Ashdod (UF capacity 930,000 m3/d),
Shuwaikh (UF capacity 350,000 m3/d). Among the manufacturers of PDI membranes applied
in SWRO pretreatment are Pentair X-Flow, BASF inge, Hydranautics, Aquasource and 3M
Membrana, while the main manufacturers of PDO membranes are Hyflux, Dow, Pall, Toray
and GE Zenon. In general, PDO configuration can tolerate higher feed solids loadings (TSS >
300 mg/L; turbidity > 300 NTU) than PDI (TSS 100 mg/L; turbidity 100 NTU), and allows
the use of air scour, which can enhance hydraulic cleaning. PDI configuration membranes
tend to have higher permeability, due to the selection of PES rather than PVDF. However,
comparative data on operation of PDI vs. PDO membranes on algal-laden seawater is not
available. Table 9.4 summarizes the products of the main international suppliers in terms of
membrane material and pore size/molecular weight cut-off (MWCO). AOM compounds –
including particulate and colloidal algal biopolymers, TEPs, etc. – cover a wide size spectrum,
ranging from a few nanometers to more than 1 millimeter (Figure 2.2 in Chapter 2). Based
upon their MWCO (~ 80-150 kDa), most UF membranes are expected to remove only part of
the higher molecular weight fraction of AOM, while MF membranes will remove an even
smaller amount.
Table 9.4. Commercially available MF/UF membranes for SWRO pretreatment.
Pressure driven inside-out (PDI)
Manufacturer Product Material Nominal pore size Nominal MWCO
Pentair X-Flow Seaguard PES/PVP 0.02 µm 150 kDa
Seaflex
BASF inge Multibore® PES 0.02 µm 100-150 kDa
Hydranautics Hydracap® PES 0.02 µm 150 kDa
Aquasource ALTEON™ Hydrophilic 0.02 µm 150 kDa
PS
3M Membrana UltraPES™ PES n.a. 80 kDa
Pressure driven outside-in (PDO)
Manufacturer Product Material Nominal pore size Nominal MWCO
Hyflux Kristal® PES n.a. 120 kDa
Dow Integraflux™ PVDF 0.02 µm n.a.
Pall Aria™ PVDF 0.1 µm n.a.
Toray* TORAYFIL® PVDF n.a. 150 kDa
GE Zenon* Zeeweed® PVDF 0.02 µm n.a.
* Also available in vacuum driven (submerged) mode
MF/UF membranes are designed based on a certain feedwater quality composition. For
instance, Hydranautics operation guidelines for their PDI membranes are turbidity <
200 NTU, algae counts < 1,500,000 cells/L, SUVA < 4, while Dow specifies turbidity < 50
NTU, TOC < 10 mg C/L, and TSS < 50 mg/L to reduce the extent of fouling and ensure
longevity of their PDO membrane elements (Hydranautics 2015; Dow Water and Process
Solutions 2015).
291
9.8.3 Fouling of MF/UF during algal blooms

PDI and PDO membranes have reportedly exhibited some degree of fouling during bloom
periods due to high concentrations of algal cells and associated organics. The accumulation
of AOM and algal particulates on the surface and/or within the pores of MF/UF membranes
leads to a loss in membrane performance. As discussed in Chapter 2, high molecular weight
AOM such as biopolymers, particularly the very sticky TEP, have been identified as the main
cause of membrane fouling rather than the algal cells themselves.
Most MF/UF membrane plants are operated in constant flux mode to a flux set point to meet
RO feedwater requirements. In constant flux filtration, fouling is observed as an increase in
TMP through time, resulting in higher pumping requirements to maintain constant filtration
flow. This may impact the efficiency of membrane cleaning. In general, MF/UF membrane
fouling phenomena are a function of feedwater quality, membrane properties and operational
conditions (e.g., filtration flux) and may be classified as:
- Hydraulically reversible or back-washable fouling (permeability is restored with
hydraulic cleaning with air and water);
- Hydraulically irreversible or non-backwashable fouling (permeability cannot be
restored with hydraulic cleaning alone and CEB may be required); and
- Chemically irreversible fouling (permeability cannot be restored with CEB and/or
chemical cleaning in place (CIP))
AOM accumulation can cause a rapid increase in TMP and/or can increase non-
backwashable fouling in MF/UF. The impact of AOM on MF/UF membrane permeability
and backwashability can be explained using classic blocking and cake mechanisms as
described in Chapter 2 (see Section 2.4.4.2).
Control of fouling that may occur during a HAB can be achieved by optimizing operational
conditions and cleaning procedures (Murrer and Rosberg 1998; Brehant et al. 2002; Zhang et
al. 2006; Ma et al. 2007; Bu-Rashid and Czolkoss 2007; Di Profio et al. 2011; Schurer et al.
2012) or by feedwater conditioning. Operating MF/UF membranes at a lower flux can reduce
the extent of backwashable and non-backwashable fouling. Filtration flux affects the
characteristics of the AOM cake/gel layer formed on the membrane surface; layers formed at
low flux exhibit lower resistance to filtration and are less compressible than those formed at
higher flux values (Salinas Rodriguez et al. 2012; Tabatabai et al. 2014). The downside of
lowering filtration flux is that proportionally higher membrane surface area is required to
meet production capacity, resulting in a larger plant footprint and higher CAPEX. This can be
balanced in part by an increase in operational flexibility to allow backwash while maintaining
capacity.
During HAB events, cleaning procedures can be tailored to control fouling in membrane
pretreatment. Optimizing the frequency, duration and intensity of hydraulic cleaning regimes,
the type and concentration of cleaning chemicals, the sequence in which they are applied, and
the duration of soaking and rinsing steps can enhance permeability recovery and fouling
control. Optimum conditions for cleaning are site and event specific and membrane
manufacturer guidelines on cleaning regimes during HAB events are currently not available.
Alternatively, feedwater conditioning (e.g. with coagulation) can be undertaken to reduce the
extent of back-washable and non-back-washable fouling in MF/UF systems. Experience in
PDI UF operation in SWRO pretreatment has shown that other than for algal bloom events,
coagulation is generally not required to stabilize UF hydraulic performance (Schurer et al.
2013), even during storms with turbidity peaks as high as 50 NTU (see Case Study 11.10).
Most importantly, operating at lowered filtration flux of approximately 60 L/m2h on North
292
Sea water allowed for longer filtration cycles (as compared to operation at 90 L/m2h) and
resulted in enhanced permeability recovery by backwashing and improved overall
productivity during bloom conditions.
Coagulation may also further enhance UF operation during algal blooms when HAB cell
numbers and TSS concentrations are high by partially acting on the following mechanisms
(Figure 9.27):
- Reducing TMP development during filtration in PDI UF membranes;
- Reducing the extent of back-washable and non-back-washable fouling;
- Enhancing permeability recovery by backwashing;
- Improving permeate quality of PDI UF membranes.
CEB BW
TMP
BW
BW
BW
BW nBW
BW nBW
nBW
nBW
Rf
nBW Irr
nBW
Rm
Time
CEB
Raw water
TMP
Coagulated water
nBW
Rf
nBW
Rm
Time
Figure 9.27. Idealized schematic presentation of TMP development and fouling in dead-end MF/UF systems for
raw feedwater with algae (top panel) and coagulated feedwater (bottom panel). BW is backwash; nBW is non-
back-washable fouling, Irr is irreversible fouling, Rm is clean membrane resistance (where resistance is TMP
normalized for temperature variation) and Rf is fouled membrane resistance. Figures: Tabatabai (2014).
9.8.4 Removal of HAB cells using MF/UF
Removal of HAB cells typically exceeds 99% given UF/MF filters do not have any gross
integrity breaches. Some shear of the HAB cells may be experienced during pumping and
impingement against the membrane surface due to pressure, breaking a small number of cells
(~1-2%) under normal UF conditions. During this process, a small amount of AOM may be
released due to stress on the cells. Stress on the HAB cells without breakage may release
more AOM than if the cells were completely broken. Coagulation prior to the UF may aid the
process by encapsulating cells in floc and preventing some breakage (Chow et al. 1997;
Dixon et al. 2011a,b). While these experiments were at a laboratory scale using cyanobacteria
(both cultured and sampled from the field), the principles for marine species remain the same,
although breakage percentages may be higher for marine HAB cells that may be less robust
293
than cyanobacteria. Cell lysis may also be dependent on the species of HAB; however, this
phenomenon requires further research.
The type of HAB cell may bear less importance on UF/MF performance than the cell count,
as shown by Castaing et al. (2011) during a laboratory-scale trial for a variety of marine
species. When comparing MF performance (0.2 µm, polysulfone, submerged outside-in)
challenging the membrane with the
marine HAB species Heterocapsa
triquetra, Alexandrium minutum and
Prorocentrum lima (Figure 9.28)
showed that while small differences in
the loss of relative permeate flux were
identified between cell types, this was
not as significant as filtration of
1,000,000 cells/L vs 30,000,000
cells/L using A. minutum (Figure
9.29). The small differences between
cell type are most likely due to
differences in the production of AOM
from each cell type. The difference
Figure 9.28. Relative permeate flux at 20 oC vs time for each made by cell numbers in this
tested microfiltration in presence of the different micro-algae experiment are not surprising given
studied. the stark difference between the cell
numbers in each experiment.
Villacorte (2014) compared sixteen
common species representing
different major groups of algae,
showing potential differences in
MF/UF performance due to size and
shape of HAB cells. These algal
species cover a wide range of shapes
and sizes and are presented in Table
9.5. Typical cells are spherical or
ellipsoidal, but some are cylindrical
(e.g. most diatoms) or elongated (e.g.
Figure 9.29. Relative permeate flow at 20oC versus time Pseudo-nitzschia spp.). To normalize
for A. minutum microfiltration.
for such shape heterogeneities, the
equivalent spherical diameter was calculated based the estimated cell volume of the cells.
The equivalent cell diameters of the sixteen species range from 4 µm (Microcystis spp.) to
400µm (Noctiluca scintillans) and a median size of 12 µm. The maximum recorded
concentrations of each species are also shown in Table 9.5. The species with the highest and
lowest recorded concentration are Microcystis spp. (14,800,000,000 cells/L 1.4x1010) and
Noctiluca scintillans (1,900,000 cells/L), respectively. The calculated fouling potential (MFI-
UF) of algal cells considering the maximum recorded cell concentrations ranged from 0.4 to
70 s/L2.
When theoretically considered as spherical or partially spherical objects, the fouling potential
of algal cells themselves appear to be low (< 100 s/L2) even during severe bloom conditions
compared to what is reported for ambient surface water, which is typically higher than
1000 s/L2 (Salinas-Rodriguez 2011). The remarkable difference may be attributed to: 1)
presence of particles much smaller than algae (e.g. TEP–like colloidal or particulate
294
materials); and/or 2) algae that are surrounded by TEP–like materials that are blocking the
gaps (interstitial voids) between deposited algal cells.
Table 9.5. Cell characteristics, recorded severe bloom concentrations, and calculated
membrane fouling potentials of 16 species of common bloom-forming algae (Villacorte
2014).
Bloom-forming algae Cell shape Eq. diam. Sphericity Severe bloom situation
(µm) (µm) (a) (-) Cells/L (b) MFI-UF
(s/L3)
Dinoflagellates
Alexandrium RE 32 0.995 10,000,000 0.38
tamarense
Cochlodinium RE 33 0.985 27,000,000 1.07
polykrikoides
Karenia brevis RE 36 0.984 37,000,000 1.60
Noctiluca scintillans Sp 400 1.000 1,900,000 0.88
Prorocentrum micans FE 44 0.914 50,000,000 3.06
Diatoms
Chaetoceros affinis OC 15 0.968 900,000,000 16.76
Pseudo-nitzschia spp. 0.8*PP 7 0.391 19,000,000 1.01
Skeletonema costatum Cy 5 0.808 88,000,000 0.78
Thalassiosira spp. Cy 12 0.867 100,000,000 1.86
Cyanobacteria
Nodularia spp. Cy 21 0.543 605,200,000 50.77
Anabaena spp.* Sp 6 1.000 10,000,000,000 69.80
Microcytis spp.* Sp 4 1.000 14,800,000,000 68.87
Haptophytes
Emiliania huxleyi Sp 5 1.000 115,000,000 0.67
Phaneocystis globosa 0.9*Sp 6 0.933 52,000,000 0.42
Raphidophytes
Chattonella spp. Co+0.5*Sp 15 0.665 10,000,000 0.39
Heterosigma Sp 20 1.000 32,000,000 1.68
askashiwo
*Non-marine species of algae; (a) Equivalent diameter of sphere with similar volume as the cell; (b) Maximum
recorded concentrations reported in various literatures. RE = rotational ellipsoid; Sp = sphere; FE = flattened
ellipsoid; OC = oval cylinder; PP = parallelepiped; Cy = cylinder; Co = cone.
Plugging of PDI fibers by HAB cells and associated AOM may also cause issues during a
HAB bloom. To illustrate the potential effect of plugging, Villacorte (2014) calculated the
expected loss of active membrane area and the localized flux increase for different species of
algae at severe algal bloom situations based on cell size and abundance (Table 9.6). Based on
the calculations, some bloom-forming species may cause severe plugging problems in
capillary UF. Severe blooms caused by three species of algae (Noctiluca scintillans,
Prorocentrum micans, Nodularia spp.) may cause complete plugging of capillaries within 30
minutes of filtration (at a flux of 80 L/m2h). Two other species (Chaetoceros affinis,
Anabaena spp.) caused more than 50% loss in active membrane area and 2-5 times increase
in average flux of the remaining active membrane area. These findings indicate that large
algae can cause blooms that may have severe implications to the operation of PDI UF due to
plugging, as can high cell numbers.
These theoretical calculations show that plugging of capillaries by HAB cells during severe
blooms may substantially increase membrane fouling. Plugging can also cause/enhance non-
295
Table 9.6. Calculated loss in effective membrane area and localized increase in average flux due to
capillary plugging by 16 species of algae during severe bloom situations (Villacorte 2014).
Bloom-forming algae Diameter Cell conc. Active Ave. flux after
(µm) (cells/L)(a) membrane plugging
area lost (%)(b) (L/m2.h)(c)
Alexandrium tamarense 32 10,000,000 9 87.5
Cochlodinium 33 27,000,000 25 107.2
polykrikoides
Karenia brevis 36 37,000,000 45 146.0
Noctiluca scintillans 400 1,900,000 -----completely plugged------
Prorocentrum micans 44 50,000,000 -----completely plugged------
Chaetoceros affinis 15 900,000,000 80 390.7
Pseudo-nitzschia spp. 7 19,000,000 0.2 80.1
Skeletonema costatum 5 88,000,000 0.3 80.2
Thalassiosira spp. 12 100,000,000 5 83.8
Nodularia spp. 21 605,200,000 -----completely plugged------
Anabaena spp. 6 10,000,000,000 57 184.1
Microcystis spp. 4 14,800,000,000 25 106.4
Emailiana huxleyi 5 115,000,000 0.4 80.3
Phaeocystis globosa 6 52,000,000 0.3 80.2
Chattonella spp. 15 10,000,000 1 80.7
Heterosigma akashiwo 20 32,000,000 7 85.8
(a) Maximum recorded concentrations reported in various literature; (b) Membrane area lost as a percentage of
the initial clean membrane area after 30 seconds of filtration at flux 80 L/m2.h; and (c) is the average flux of the
remaining active membrane area after 30 minutes of filtration and keeping the permeate flow constant.
back-washable fouling if the accumulated HAB cells are not effectively removed from the
feed channel by hydraulic cleanings. Backwashing and chemical cleaning may not be 100%
effective in removing foulants in plugged capillaries and the plugged section of the
capillaries will continue to increase in the succeeding filtration cycles and cause severe non-
back-washable fouling. To minimize plugging problems in PDI UF membranes, the
following could be applied:
a) shortening the filtration cycle e.g., from 30 min to 15 mins;
b) reducing the average flux (Note: various PDI UF plants are operating at a more
conservative flux (50-60 L/m2h) in areas where HAB blooms are common).
9.8.5 Summary
In summary, during algal bloom events MF/UF operation can be controlled by lowering
filtration flux, applying amended backwash regimes, or through the addition of coagulant.
Under optimum process conditions, UF operation can be stabilized by coagulation at doses as
low as 0.5-1.5 mg Fe/L to yield less compressible AOM-iron hydroxide floc at this range. An
alternative mode of coagulant application (i.e. coating UF membranes with pre-formed flocs
of iron hydroxide) has proved efficient in controlling UF operation at very low dose (0.5 mg
Fe/L) during severe algal bloom conditions in laboratory studies. To remove algal
biopolymers from UF permeate by over 80%, doses as high as 5-10 mg Fe/L are required.
Coagulation at such high dose in MF/UF systems results in increased cost and complexities
associated with handling and treatment of coagulant-rich sludge and may also be detrimental
to the hydraulic performance (non-back-washable fouling and membrane plugging by iron
hydroxide flocs) of UF membranes. HAB cells are removed very well, greater than 99%,
although a small amount of damage to the cells through shear stress and pressure may cause a
release of AOM. Further, stress on live cells during the process may cause them to produce
296
more AOM. Some difference between removal of different cell types has been apparent,
particularly for larger cells such as Noctiluca, that can cause plugging of PDI UF fibers.
9.9 CARTRIDGE FILTERS FOR REVERSE OSMOSIS PRETREATMENT
9.9.1 Overview
Cartridge filters are a protection measure for SWRO membranes, pumps, and energy
recovery devices. The main purpose they serve is to capture particulates in the pretreated
source water that may have passed through the upstream pretreatment systems. During a
bloom, cellular material should have been removed prior to the cartridge filters so in most
cases an increase in differential pressure should not be due to cells. More likely, AOM can
build up on the cartridge filters and cause biofouling. Buildup of iron residuals can also occur
on cartridge filters. This section describes operation of cartridge filters during a bloom and
the subsequent challenges.
9.9.2 Types and configurations
Cartridge filters are fine micro-filters of nominal size of 1 to 25 µm made of thin plastic
fibers (typically polypropylene) which are wound around a central tube to form standard size
cartridges (Figure 9.30). Often
they are the only screening
device between the intake wells
and the SWRO system with well
intakes producing high quality
source water.
Although wound (spun)
polypropylene cartridges are
most commonly used for
seawater (and brackish water)
applications, other types, such as
melt-blown or pleated cartridges
of other materials have also
Figure 9.30. Cartridge filters installed in horizontal vessel.
found application. Standard
cartridge filters for RO desalination plants are typically 101.6 to 1,524 cm long and are
installed in horizontal or vertical pressure vessels (filter housings). Cartridges are rated for
removal of particle sizes of 1, 2, 5, 10, and 25 µm, with the most frequently used size being 5
µm.
Cartridge filters are typically installed downstream of the pretreatment to capture fine sand,
particles and silt that could be contained in the pretreated seawater following GMF. When the
source seawater is of very high quality (SDI15 < 2) and does not need particulate removal by
filtration prior to desalination, cartridge filters are used as the only pretreatment device,
which in this case serves as a barrier to capture fine silt and particulates that could
occasionally enter the source water during the start up on intake well pumps or due to
equipment/piping failure.
The main function of cartridge filters is to protect the downstream SWRO membranes, high-
pressure pumps and energy recovery devices from damage, not to provide removal of large
amount of particulate foulants from the source seawater. A typical indication of whether the
pretreatment system of a given desalination plant operates properly is the seawater SDI
reduction through the cartridge filters. If the pretreatment system performs well, then the SDI
of the source water upstream and downstream of the cartridge filters is approximately the
same.
297
If the cartridge filters consistently reduce the SDI of the filtered source water by over one
unit, this means that the upstream pretreatment system is not functioning properly.
Sometimes, SDI of the source water increases when it passes through the cartridge filters.
This condition almost always occurs when the cartridge filters are not designed properly or
are malfunctioning and provide conditions for growth of biofouling microorganisms on and
within the filters. The biofouling of the cartridge filters is usually lower if they are
chlorinated periodically. Therefore, it is recommended the point of dechlorination of
pretreated water to be downstream of the cartridge filters.
For UF or MF filtration systems that have a direct flow-through pattern where the
desalination plant feed pumps convey water directly through the membrane pretreatment
system without interim pumping, the pretreatment membranes are more likely to be exposed
to pressure surges. If the pretreatment membrane fiber material is weak and it easily breaks
under pressure surge conditions, such pretreatment systems are more likely to experience
fiber breaks. Broken membrane fibers can release small amounts of particles (or algal cells)
into the SWRO feed water, which can cause accelerated membrane fouling. Therefore, the
use of cartridge filters downstream of the membrane pretreatment system is a prudent
engineering practice.
Cartridge filters are operated under pressure and the differential pressure across these filters
is monitored to aid in determining when filter cartridges should be replaced. In addition,
valved sample ports should be installed immediately upstream and downstream of the
cartridge filter vessel(s) for water quality sampling and testing (including SDI field testing).
9.9.3 Algal-bloom related challenges
During moderate and heavy algal blooms, cartridge filters usually experience shortened
useful life due to accelerated bacterial growth on the
cartridge surface and core. Such growth is triggered as
a result of the elevated content of easily biodegradable
dissolved organic solids released from the algal
biomass, such as TEP. The accelerated biogrowth on
the cartridge filters peaks during the period of the algae
decay, which usually occurs several weeks after the
beginning of a typical algal bloom. In order to counter
the negative impact on the accelerated biogrowth on
the cartridge filters and their premature plugging, and
subsequent replacement, source water is usually
chlorinated more frequently and the cartridges are
exposed to chlorine at a higher dose (1 to 2 mg/L vs.
0.2 to 0.5 mg/L) for a longer period of time (6 hours vs.
2 to 3 hours). Figure 9.31 depicts a heavily fouled
cartridge filter during period of severe algal blooms
with cell concentrations exceeding 40,000,000 cells/L.
Figure 9.31. Heavily fouled cartridgeElectroadsorptive cartridge filtration is a novel
filter during severe algal bloom event
technology, which holds promise for pretreatment of
from the Pacific Ocean. Photo:
Voutchkov (2013). seawater exposed to frequent algal blooms. The
electroadsorptive filter media removes TEP through a
strong positive charge generated by nanofibers of the
mineral boehmite and the tortuous path created by the depth filter media itself. The filter
media has a mean flow pore of about 0.7 µm and very high nanofiber surface area that
produces a filter with low-pressure drop but a high filtration efficiency and high loading
298
capacity for TEP removal (Komlenic et al. 2013). Test results on Mediterranean seawater
indicate that the electoradsorptive cartridge filters can achieve TEP and chlorophyll-a
removal efficiencies of 40 to 75% and 60 to 80%, respectively.
9.10 REVERSE OSMOSIS
9.10.1 Overview
RO, the core of seawater desalination plant design, is discussed in this section. SWRO is very
susceptible to fouling and the risk of fouling is very high during algal blooms. If pre-
treatment is operated efficiently, the risk to the RO will be greatly reduced. In this section the
water quality required for the SWRO process unit is described, as well as the potential for
fouling the RO, the most common types of RO fouling associated with algal bloom events,
and diagnosis and mitigation of fouling events.
9.10.2 Raw water quality and pretreatment requirements for SWRO
Various membrane manufacturers have developed recommended pretreatment configurations
based on seawater intake and water quality parameters to characterize the particulate and
organic load of the raw source seawater to be desalinated, such as turbidity, total suspended
solids, silt density index and total organic carbon (Chapter 5). These parameters are
commonly used to indicate water quality of the raw water and across pretreatment. Of these
indicators, only turbidity can be measured continuously online. The others are conducted as
discrete measurements on water samples taken periodically. The configuration of the
pretreatment system has evolved around specific water sources and aims to produce
feedwater of acceptable quality to allow long-term operation.
In areas prone to HABs, the pretreatment system for open intakes requires augmentation by
additional treatment steps, as suggested in Table 9.7 for open intakes. Generally, plants with
seawater beach wells should attenuate the high organic and particulate load associated with
HABs, as the wells provide pretreatment prior to intake, removing a substantial amount of the
bacteria, algae, and algal organic matter (Chapter 6).
The majority of membrane manufacturers specify an upper limit of 5 for feedwater SDI15 in
their guarantees. Field results show that for stable, long-term performance of RO elements,
the SDI15 of feedwater should be consistently below 4. The SDI limit is dependent upon the
lead element flux. Consistently lower feedwater SDI15 allows for higher lead element flux.
For example, in RO systems with UF pretreatment, the lead element flux may be as high as
35 L/m2h, while for conventional pretreatment systems, this may only be as high as 32 L/m2h.
Field results have demonstrated that in the majority of cases water from deep beach wells has
very low SDI15, usually less than 1. RO systems, operating with good quality well water feed,
practically do not show any pressure drop increase across the membranes or flux decline.
Surface seawater, after a conventional pretreatment, usually has SDI15 in the 2 – 4 range. A
RO system processing feedwater with SDI15 in the 2 – 3 range will show stable membrane
performance. Membrane cleaning frequency for such feedwater does not exceed once or
twice per year. RO systems processing feedwater of higher SDI15, in the 3 - 4 range, usually
suffer from some degree of membrane fouling and somewhat higher membrane cleaning
frequency may be required. Long-term operation of RO systems with feedwater having SDI15
above 4 is not recommended, particularly when average system flux is above 15 L/m2h.
Many systems are being designed with these higher average and lead element fluxes and such
systems will be a major risk during HABs.
299
Table 9.7. Recommended configuration of pretreatment system according to raw water

quality (Hydranautics 2015).
Water quality Configuration of

Water source Comments
parameters pretreatment system
Seawater beach well Turbidity < 0.2 NTU Cartridge filtration If seawater is under
TSS < 2 mg/L influence of brackish
SDI15 < 1.0 water, acidification,
and scale inhibitor may
be required
Seawater beach well Turbidity > 0.2 NTU Sand filtration If seawater is under
TSS > 2 mg/L Cartridge filtration influence of brackish
SDI15 > 1.0 water, acidification,
and scale inhibitor may
be required
Seawater open intake Turbidity < 5 NTU Acidification Short excursion of
TSS < 5 mg/L Coagulation + turbidity up to 20 NTU
TOC < 2 mg/L flocculation is possible for few days
Single stage granular in year
Seawater open intake Turbidity < 5 NTU Membrane filtration Short excursion of
TSS < 5 mg/L turbidity up to 20 NTU
TOC < 2 mg/L is possible for few days
in year
Seawater open intake Turbidity 5 -20 NTU Acidification Short excursion of
TSS > 5 mg/L Coagulation + turbidity up to 30 NTU
TOC > 2 mg/L flocculation is possible for few days
Two stage granular in year
Seawater open intake Turbidity 5 -20 NTU Acidification Short excursion of
TSS > 5 mg/L Coagulation + turbidity up to 30 NTU
TOC > 2 mg/L flocculation is possible for few days
Membrane filtration in year
Seawater open intake Turbidity > 20 – Settling clarification Suspended solids
30 NTU Coagulation + mainly inorganic
TSS > 5 mg/L flocculation particles(silt)
TOC > 2 mg/L Single stage granular
Seawater open intake Turbidity > 20 – Settling clarification Suspended solids
30 NTU Coagulation + mainly inorganic
TSS > 5 mg/L flocculation particles(silt
TOC > 2 mg/L Membrane filtration
Seawater open intake Turbidity > 20 - DAF Suspended solids
30 NTU Coagulation + mainly organic and
TSS > 5 mg/L flocculation biological
TOC > 2 mg/L Single stage granular particles(algae)
Seawater open intake Turbidity > 20 - DAF Suspended solids
30 NTU Coagulation + mainly organic and
TSS > 5 mg/L flocculation biological
TOC > 2 mg/L Membrane filtration particles(algae)
300
Media filtration and membrane filtration are well understood. The prevention of biofilms was
misunderstood in the past as discussed in the chlorination-dechlorination section of this
Chapter. Recently most of the SWRO plants either eliminated the chlorination –
dechlorination process completely or apply shock chlorination.
Another alternative for AOM reduction is addition of activated carbon (Huang et al. 2015) to
absorb organics, although this is seldom practiced. A PAC system exists at the Adelaide
Desalination Plant in Australia, where PAC can be dosed before and after the disk filters and
PAC is removed by the UF, but no serious algal blooms have occurred at that plant to warrant
use of the PAC system. Its primary use is for hydrocarbon removal.
9.10.3 Effect of algal blooms on SWRO operation
Algal blooms may result in deterioration of raw water quality depending on the density of the
bloom and associated organics (intracellular or extracellular) or indirectly through a
reduction in oxygen as algae die off. This may be detected through raw water monitoring of
SDI, turbidity, TOC, and AOM such as TEP (see Chapter 5). An additional indicator that
may be specifically associated with an algal bloom event is an increase in the concentration
of chlorophyll-a, thus alerting RO operators of the potential for a fouling event. Chlorophyll
background levels are generally specific for different regions and might be in the range of 1-
5 µg/L. During a HAB bloom chlorophyll may increase significantly, reaching as high as 120
µg/L (Desormeaux et al. 2009; Franks et al. 2006). Some blooms will produce other algal
pigments and thus less chlorophyll-a, and thus may not be detected by chlorophyll sensors
(see Chapters 3 and 5).
During the HAB period, seawater drawn into the pretreatment system contains much higher
concentrations of suspended particulate matter, which significantly increases the solids load
in the influent to the filtration system (either media or membrane filtration). The result is a
need to increase frequency of backwashing, backwash volume, backwash time and air scour
time, which in turn reduces the yield of the pretreatment system and thus the flow to the RO
trains. Another problem can be an increased concentration of dissolved organic carbon (or
algal organic matter, AOM), which can easily pass the pre-treatment process. This can
contribute to an increase in biofouling of RO membrane elements, resulting in an increase in
differential pressure.
Alternatively, impacts may manifest in the RO due to fouling if pretreatment does not
mitigate the increases in these parameters. SWRO systems with poorly optimized
pretreatment may experience a rapid rate of differential pressure increase, due to particulate
fouling and biofouling during or following a bloom event (Franks et al. 2006; Unni et al.
2011).
9.10.4 Ferric coagulant SWRO fouling during algal bloom events
When coagulation/flocculation is employed as a part of a seawater feed pretreatment scheme,
it is important to dose the coagulant at a rate that will not result in filter breakthrough. While
aluminum and ferric salts will both foul the RO process by a similar mechanism, only ferric
is discussed as the most commonly used coagulant in SWRO (see Section 9.4.1). Ferric
coagulant that will pass the filtration step and reach the membrane elements will result in iron
hydroxide fouling of membrane elements (Figure 9.32). Ferric coagulant fouling can be
diagnosed in the plant through a rapid increase in differential pressure, a rapid increase in
normalized feed pressure and a rapid increase in normalized salt passage (normalization of
RO plant data is discussed in ASTM method D4516 (ASTM, 2010) and requires diligent
collection of plant data to be of use when diagnosing fouling). This trend is well described in
Hydranautics Technical Bulletin, TSB-107, Table 1 (Hydranuatics, 2015).
301
To demonstrate the localized

effect of ferric coagulant
foulant at the membrane level,
a system performance
evaluation was undertaken on
a full pressure vessel of
elements removed from a full-
scale plant (Hydranautics
2015). Each element was
individually wet tested and
differential pressure analyzed,
according to element position
in the vessel (Table 9.8).
At standard conditions, the
expected differential pressure
from a single RO element
should be close to 0.2 bar.
Figure 9.32. Surface of membrane in the element that was operated in
This differential pressure is
a SWRO system that was exposed to an excessive dose of ferric observed in tail elements
coagulant. seven and eight. The
differential pressure on lead
element one is approximately three times the nominal value. Differential pressure decreased
gradually along the pressure vessel. Observing the distribution, it is evident that the case of
increased pressure is result of poor quality of feedwater. In this particular SWRO system, the
problem was partially resolved by optimization of the coagulation process and an additional
filtration step, downstream of the media filters. During an algal bloom, selection of the
correct dose of coagulant is complicated, as the concentration of HAB cells in the feedwater
changes daily, and during exponential growth, even more frequently. The associated loading
of AOM will also change with the cell count. It is thus difficult to both properly coagulate
cells and avoid overdosing coagulant, unless regular jar tests are performed during a bloom
and dosing information is used to make on site changes to minimize any potential overdose.
Every SWRO system utilizing ferric coagulation will experience some degree of iron deposit
in RO membrane elements.
Table 9.8. Differential pressure of individual elements according to their position in pressure
vessel from a SWRO system experiencing excessive ferric coagulant dosing. Wet test for
element performance was conducted at standard test conditions, but with 10% recovery rate
rather than 8% standard.
Position 1 2 3 4 5 6 7 8
Differential 0.65 0.34 0.31 0.28 0.26 0.24 0.21 0.22
Pressure, bar
9.10.5 SWRO Biofouling during and following algal bloom events

All raw waters contain microorganisms such as bacteria and algae. The typical size of
bacteria is about 1 µm. Most microorganisms will be removed during pretreatment, but those
that remain can rapidly reproduce and form a biofilm under favorable conditions (Dow Water
and Process Solutions 2015). Microorganisms entering an SWRO system find a large
membrane surface where dissolved nutrients from the water are enriched due to concentration
302
polarization, thus creating an ideal environment for the formation of a biofilm. Biological
fouling of the membranes will seriously affect SWRO performance. Severe symptoms arising
from well-developed biofouling results in blockage of RO element feed channels as shown in
Figure 9.33 or even telescoping of elements in the direction of the tail end of the pressure
vessel, causing permanent mechanical damage to the elements. The blockage of feed
channels is expressed as an increase of differential pressure along the RO trains. An example
of differential pressure increase variations in a SWRO plant that experienced severe
biofouling, due to continuous chlorination – dechlorination is illustrated on Figure 9.34.
Figure 9.33. Left: Feed channels in clean RO membrane element; Right: Feed channels in an RO membrane
element with significant biofouling. Photos: D.H. Paul Training.
The potential for biological fouling

should be assessed during the design
or pilot phase (where a pilot phase
exists) so that the system can be
designed accordingly. Warm waters,
such as in the Gulf, generally have a
higher biofouling potential than cold
well waters.
When fouling of SWRO membranes
occurs during an algal bloom, this is
commonly biofouling. Biofouling will
present itself to the operator as a
marked increase in differential
Figure 9.34. Differential pressure increase (shown here as pressure in the first pass, first stage of
pressure drop) in a SWRO plant as a result of biofouling a SWRO plant. An increase will
Figure: Wilf et al. 2007. develop over 1-2 weeks during or
immediately after a HAB. To further
diagnose such an issue, the normalized feed pressure will also experience a marked increase
during this time and the normalized salt passage will remain normal or be slightly increased.
The change in salt passage will distinguish biofouling from ferric coagulant fouling, as ferric
coagulant fouling will have a much more rapid increase in normalized salt passage. This
trend is well described in Hydranautics Technical Bulletin, TSB-107, Table 1 (Hydranuatics
2015).
The regular assessment of the microbiological activity of the feed water should also be part
of the operating discipline of a plant so that any increase in microbiological activity can be
responded to at an early stage. AOC is one such method for undertaking this monitoring
303
along with TEP and LC-OCD (see Chapter 5). The frequency of sampling and analysis
depends upon the risk of biofouling.
Prevention of biofouling is an important consideration for operators as biofouling is very
difficult to remove and can permanently damage RO membranes. Once a RO membrane is
biofouled, subsequent biofouling will develop more rapidly. Figure 9.35 shows that during a
laboratory-scale membrane fouling simulator (MFS) RO experiment, the addition of further
AOM to the RO feedwater will cause an increase in differential pressure. Further to this, once
a biofilm has ‘pre-fouled’ a membrane during previous fouling events, the differential
pressure (feed channel pressure drop) will rise more rapidly than a brand new membrane
despite cleaning of the membrane (Villacorte 2014).
Figure 9.35. Differential pressure (feed channel pressure drop) across the feed channel of MFS cells with
clean membrane (MFS 1 & 2) and pre-fouled with AOM (3 & 4). Source: Villacorte 2014.
The most successful approach to minimize biofouling is the limitation or removal of nutrients
for microorganisms from the water in order to limit biological growth. This can be achieved
through the careful operation of the pretreatment as discussed previously (Section 9.4-9.9).
Coagulation is critically important in the pretreatment to remove biofouling. Without
coagulation, a high percentage of AOM from the feedwater will pass filtration/flotation steps
during blooms. If allowed to pass to the RO membranes, AOM will accelerate biofilm
formation. Effective coagulation will enable consistent RO operation at the designed output
capacity.
Additionally, the continuous addition of oxidation chemicals such as chlorine may increase
the nutrient level because organic substances may be broken down to smaller biodegradable
fragments (as discussed in Section 9.2). Dosing chemicals such as antiscalants or acids must
be carefully selected because they may also serve as nutrients. Preventive treatments are
much more effective than corrective treatments because single attached bacteria are easier to
deactivate and remove than a thick, aged biofilm (Vrouwenvelder 2009). Other physical
methods are targeted to remove microorganisms in the feedwater with microfiltration or
ultrafiltration to deactivate them with UV radiation.
One method for minimizing the attachment of bacteria to a membrane surface and their
growth is surface modification of the membrane (Dow Water and Process Solutions 2015).
304
This concept is available with fouling resistant elements. While popular for application in
brackish water applications, they have been less popular for seawater RO elements to date.
Other methods for biofouling reduction are based on chemicals that have a biocidal effect on
microorganisms. These sanitization chemicals are applied during the normal operation of the
plant either as a continuous dose to the feedwater stream or preferably as a discontinuous
(intermittent) dose in certain intervals. Typical treatment intervals are one to four per month
(Dow Water and Process Solutions 2015). Non-oxidizing biocides are used as chlorine is
oxidizing and cannot be used due to rapid damage of the polyamide layer in the RO.
One biocide chemical commonly used is sodium bisulfite that can be added into the feed
stream as a shock treatment during normal plant operation. In a typical application, 500–
1,000 mg/L sodium bisulfite is dosed for 30 minutes. Sodium bisulfite should be dosed as
sodium metabisulfite (food-grade) that is free of impurities and not cobalt-activated. The
treatment can be carried out periodically or when biofouling is suspected. The permeate
produced during dose will contain some bisulfite, depending on the feed concentration, the
membrane type and the operating conditions. Depending on the permeate quality
requirements, permeate can be used or discarded during shock treatment (Dow Water and
Process Solutions 2015).
As an alternative to sodium bisulfite, the non-oxidizing biocide DBNPA (2,2, dibromo-3-
nitrilo-proprionamide) can also be used to prevent biofouling. DBNPA has been found to be
cost effective, has acceptable transportation, storage, stability and handling characteristics,
has broad spectrum control (e.g. planktonic and sessile organisms) and is quickly
biodegraded in the environment. The standard method to apply DBNPA is shock
(intermittent) dosing. The amount of DBNPA used depends on the severity of the biological
fouling. With water that has reduced biofouling potential, dosing 10–30 mg/L of the active
ingredient for 30 minutes to 3 hours every 5 days can be effective. Because DBNPA is
deactivated by reducing agents (such as sodium bisulfite used for chlorine removal), a higher
concentration of DBNPA will be required if there is residual reducing agent in the feed water.
The concentration of DBNPA should be increased by 1 ppm of active ingredient for every
ppm of residual reducing agent in the RO feed water. To remove the dead biofilm, an alkaline
cleaning is also prudent. Biocides, their degradation products, and other ingredients in their
formulations are not always completely rejected by RO membranes. For this reason, during
shock dosing, it may be necessary to discharge the permeate because it may contain slightly
elevated levels of organics. Note that although DBNPA is non-oxidizing, it does produce an
oxidation/reduction potential (ORP) response in approximately the 400 mV range at
concentrations between 0.5 and 3 mg/L. For comparison, chlorine and bromine give a
response in the 700 mV range at 1 mg/L, which increases with increasing concentration. This
increase in ORP is normal when adding DBNPA and it is recommended the ORP set-point is
by-passed during DBNPA addition (Dow Water and Process Solutions 2015).
The optimal frequency for dosing sodium bisulfite or DBNPA will be site-specific and must
be determined by the operating characteristics of the RO system. In RO systems operating
with biologically active feed water, a biofilm can appear within 3–5 days after inoculation
with viable organisms (Villacorte 2014). Consequently, the most common frequency of
sanitization is every 3–5 days during peak biological activity (summer) and about every 7
days during low biological activity (winter). Dosing should be considered during a HAB
bloom to prevent the onset of biofouling (Dow Water and Process Solutions 2015).
Care should be taken when disposing of biocide-affected waste streams as these can be
detrimental in the environment and local permit conditions may prevent operators from
disposing of these in the outfall.
305
9.10.6 SWRO operational strategies during a bloom

At low algal concentrations, the RO system will most likely continue to operate, with no
system changes. Sometimes at moderate algae concentrations, the RO will operate at reduced
capacity, due to lower output of the feedwater from the pretreatment system. At severe algal
concentrations, operators may consider shutting down the RO to avoid severe (and somewhat
irreversible) fouling of membrane elements.
Optimization of pretreatment may be the best strategy for maximizing SWRO efficiency
during a bloom. Media filtration and membrane filtration optimization are discussed earlier in
this Chapter (Section 9.6 and 9.8). Coagulation should be considered, as without coagulation,
a high percentage of AOM from the feedwater will pass the pretreatment during the HAB
blooms (Section 9.5). If allowed to pass to the RO membranes, AOM will accelerate biofilm
formation. Acceleration of biofouling is also attributed to chlorination-dechlorination, as
discussed in the chlorination-dechlorination section of this Chapter (9.2 and 9.3). Most
SWRO plants either eliminate the chlorination – dechlorination process completely or apply
shock chlorination. Shock chlorination should be reconsidered or modified during a bloom.
Non-oxidizing biocide dosing to the RO could also be considered during a bloom period (e.g.,
DBNPA).
While RO operating conditions can be slightly changed during a bloom, the design envelope
to produce the required amount of permeate at an acceptable quality is usually set precisely
during the design process. When flux is lowered in an SWRO train, the permeate salinity
increases, therefore lowering flux too much may result in permeate water that is out of TDS
specification. In places where a boron specification for the permeate quality is low (e.g.
0.5mg/L), lowering the permeate flux may quickly produce an unacceptable permeate boron
concentration. While some plants may be able to deal with this issue through a second pass
design, others may not have this flexibility. Lowering recovery may slightly improve
permeate quality and at the same time decrease the lead element flux to alleviate some
fouling trends. This will result in higher pumping pressures and thus more energy to produce
the same amount of water; however, the minor impact on lead element flux experienced
when lowering recovery may not be sufficient to significantly alleviate the effects of fouling.
While flux and recovery could be lowered to avoid some fouling during an algal bloom, a
better result may be obtained by optimizing the pretreatment to yield better removal of AOM.
9.10.7 Membrane cleaning
Maintenance chemical cleaning for iron removal, applied periodically, will restore membrane
performance to clean membrane conditions. The common cleaning procedure for iron
removal (in the absence or organics) is recirculation of citric acid solution at a concentration
of about 2%.
The cleaning applied for biofouling consists of high pH flushing combined with prolonged
soaking. RO cleaning is discussed in greater detail in Appendix 5. A biofilm is difficult to
remove because it protects its microorganisms against the action of shear forces and biocide
chemicals. In addition, if not completely removed, remaining parts of a biofilm (in the form
of AOM) lead to accelerated biofouling (Villacorte 2014). Once an element has biofouled
once, it can be difficult to prevent further occurrences. Biological fouling prevention is
therefore a major objective of the pretreatment process. The common cleaning procedure for
biofouling removal is recirculation of an alkaline cleaning solution at pH 12. In extreme
cases, a biocide can be used (Appendix 5).
When iron fouling is combined with biofouling, the removal of foulant from the membrane is
difficult and restoration of performance much less effective. In such cases, the cleaning
306
procedure consists of alternate steps of high pH (organics removal) and low pH (iron
removal) cleaning. In extreme cases, a biocide can be used between the high pH and low pH
step. To adequately restore RO system performance, in the case of mixed iron – biofouling,
the cause of biofouling has to be addressed and eliminated (Appendix 5).
9.10.8 Summary
SWRO is very susceptible to fouling, particularly from overdosing ferric coagulants and
through biofouling. During HABs, this risk is very high as it is difficult to predict the precise
dose of ferric coagulant on any given day of a bloom, and AOM released from HAB cells can
pass the pretreatment process and promote biofouling; however, if pre-treatment is operated
efficiently, the risk to the RO will be greatly reduced in all but the most severe HABs.
Cleaning of the RO after ferric coagulant fouling is not difficult, but cleaning of biofouled
elements can be a challenge owing to the sticky nature of the biofilm. Other than
optimization of the pre-treatment, only a minor amount of adjustment can be made to the
lower flux and lower recovery of a SWRO process unit to prevent fouling from HAB AOM
due to tight design parameters to achieve acceptable permeate quality.
9.11 SUMMARY OF BIOMASS REMOVAL IN SWRO
This chapter presents the most common treatment methods used in SWRO systems and
discussed the impacts of a non-toxic HAB on plant operations. A summary of a broad set of
industry knowledge on the matter is summarized, giving guidance to designers and operators
alike. The following major concepts are discussed:
• Avoiding chlorination-dechlorination during a bloom will help to prevent downstream
fouling. While chlorination will destroy HAB cells, broken cells will release AOM,
causing downstream fouling (Section 9.2 and 9.3). The AOM is more important in
terms of negative impacts than the intact cells, which are more easily removed during
pretreatment.
• DAF is a good choice for cell removal in areas likely to experience heavy algal
blooms, as it lifts and removes cells from the seawater in a relatively gentle manner
(Section 9.5).
• GMF (Section 9.6) and UF (Section 9.8) can deal with cell removal in lighter blooms,
but using a DAF upstream for heavy blooms is prudent.
• Coagulation assists all three pretreatments (DAF, GMF, and UF) and acts to remove
AOM more effectively than the pretreatments alone. Less AOM in the pretreated
water is also an important objective to alleviate SWRO fouling (Section 9.4).
• Microstrainers can foul with algal material and may have shorter cycle times during a
bloom (Section 9.7).
• Cartridge filters are installed to protect the SWRO. These will most likely foul during
or following a bloom and require regular replacement during longer more intense
blooms (Section 9.9).
• There is little an operator can do to adjust the SWRO flux and recovery. A better
strategy may be to focus on getting the pretreatment working efficiently and,
consequently, the SWRO will operate more effectively. SWRO fouling from excess
ferric coagulant and biofouling may still be an issue during intense blooms and may
require focused cleaning after a bloom (Section 9.10).
307
9.12 REFERENCES
Agus, E., Voutchkov, N., and Sedlak, D. L. 2009. Disinfection by-products and their
potential impact on the quality of water produced by desalination systems: a literature
review. Desalination 237(1-3), 214-237.
Al-Rasheed, R., Althobity, E., Al-Sulami, S., and Al-Ruwished, A. 2009. Influence of
Different Disinfection Methods on the Presence of Bromate and Chlorate in the Product
Water of Seawater Desalination Plants. Proceedings of the International Desalination
Association World Congress, Dubai, UAE.
Amato, T., Nattress, J., and Mackay, D. 2012. Seawater v. Surface Water Treatment
Applying Enflo-DAFTM Technology. Proceeding of the American Water Works
Association Water Quality Technology Conference, Toronto, Ontario, Canada.
Appelgate, L., Erkembrecher Jr., C. and Winters, H. (1989) New chloramine process to
control aftergrowth and biofouling in Permasep B-10 RO surface seawater plants,
Bahamdan, T., Al Mobayed, A., and Ehsan, A. 1999. Prevention of biofouling of seawater
systems – a case study. Proceedings of the International Desalination Association World
Congress, San Diego, California.
Bar-Zeev, E., Belkin, N., Liberman, B., Berman, T., and Berman-Frank, I. 2012. Rapid sand
filtration pretreatment for SWRO: Microbial Maturation Dynamics and Filtration
Efficiency of Organic Matter, Desalination 286, 120-130.
Besson, A. and Guiraud, P. 2012. Characterization of bubbles produced by dissolved air
flotation in saline water. Proceedings of the International Water Association Flotation
Conference. Oct 29 - Nov 1. New York City, New York.
Bonnelye, V., Sanz, M. A., Durand, J-P., Plasse, L., Gueguen, F., and Mazounie, P. 2004.
Reverse osmosis on open intake seawater: pre-treatment strategy. Desalination 167, 191-
200.
Brehant, A., Bonnelye, V., and Perez, M. 2002. Comparison of MF/UF pretreatment with
conventional filtration prior to RO membranes for surface seawater desalination.
Bu-Rashid, K. A. and Czolkoss, W. 2007. Pilot tests of Multibore UF membrane at Addur
SWRO desalination plant, Bahrain. Desalination 203(1-3), 229-242.
Camp, T. R. and Stein, P. C. 1943. Velocity gradients and internal work in fluid motion.
Journal of the Boston Society of Civil Engineers 30, 219-237.
Castaing, J. B., Mass. A., Sechet, V., Sabiri, N.-E., Ponti, M., Haure, J., and Jaouen, P. 2011.
Immersed hollow fibres microfiltration (MF) for removing undesirable micro-algae and
protecting semi-closed aquaculture basins, Desalination 276, 386–396.
Huang, W., Chu, H., Dong, B., Hu, M., and Yu, Y. 2015. A membrane combined process to
cope with algae blooms in water. Desalination 355, 99–109.
Choi, K. Y. and Dempsey, B. A. 2004. Inline coagulation with low pressure membrane
filtration. Water Research 38, 4271-4281.
Chow, C. W. K., Panglish, S., House, J., Drikas, M., Burch, M. D., and Gimbel, R. 1997. A
study of membrane filtration for the removal of cyanobacterial cells. Journal of Water
Supply: Research and Technology 46(6), 324-334.
308
Crittenden, J. C., Trussell, R. R., Hand, D. W., Howe, K. J., Tchobanoglous, G., and
Borchardt, J. H. 2012. MWH’s water treatment principles and design. Hoboken (NJ),
Wiley: 1920 p.
Dennett, K. E., Amirtharajah, A., Moran, T. F., and Gould, J. P. 1996. Coagulation: its effect
on organic matter. Journal of the American Water Works Association 84(4), 129-142.
Desormeaux, E., Meyerhofer, P. F., and Luckenbach H. 2009. Results from Nine
Investigations Assessing Pacific Ocean Seawater Desalination in Santa Cruz, California,
Proceeding of the IDA World Congress, Atlantis, The Palm – Dubai, UAE.
Di Profio, G., Ji, X., Curcio, E., and Drioli, E. 2011. Submerged hollow fiber ultrafiltration as
seawater pretreatment in the logic of integrated membrane desalination systems.
Desalination 269 (1-3), 128-135.
Dixon, M. B., Richard, Y., Ho, L., Chow, C. W. K., O’Neill, B. K., and Newcombe, G.
2011a. Integrated membrane systems incorporating coagulation, activated carbon and
ultrafiltration for the removal of toxic cyanobacterial metabolites from Anabaena
circinalis. Water Science and Technology 63(7), 1405-1411.
Dixon, M. B., Richard, Y., Ho, L., Chow, C. W. K., O’Neill, B. K., and Newcombe, G.
2011b. A coagulation - powdered activated carbon - ultrafiltration multiple barrier
approach for removing toxins from two Australian cyanobacterial blooms. Journal of
Hazardous Materials 186(2-3), 1553-1559.
Dow Water and Process Solutions. 2015. Reverse Osmosis Technical Manual,
http://msdssearch.dow.com/PublishedLiteratureDOWCOM/dh_08db/0901b803808db77
d.pdf?filepath=liquidseps/pdfs/noreg/609-00071.pdf&fromPage=GetDoc, Accessed
online, November 2015.
Duan, J., Graham, N. J. D., and Wilson, F. 2002. Coagulation of humic acid by ferric
chloride in saline (marine) water conditions. Water Science and Technology 47(1), 41-48.
Duan, J. and Gregory, J. 2003. Coagulation by hydrolyzing metal salts. Advances in Colloid
and Interface Science 100-102, 475-502.
Edzwald, J. K., Tobiason, J. E., Amato, T., and Maggi, L. J. 1999. Integrating high rate
dissolved air flotation technology into plant design, Journal of American Water Works
Association 91(12), 41-53.
Edzwald, J. K. 2010. Dissolved air flotation and me. Water Research 44, 2077-2106.
Edzwald, J. K. and Haarhoff, J. 2011. Seawater pretreatment for reverse osmosis: chemistry,
contaminants, and coagulation, Water Research 45, 5428–5440.
Edzwald, J. K. and Haarhoff, J. 2012. Dissolved Air Flotation for Water Clarification.
American Water Works Association and McGraw-Hill: 352 p.
Eshel, G., Elifantz, H., Nuriel, S., Holenberg, M., and Berman T. 2013. Microfiber filtration
of lake water: impacts on TEP removal and biofouling development. Desalination and
Water Treatment 51, 1043-1049.
Ferguson, M., Towndrow, S., Broom, G., Jovanoski, M., Boardman, B., and Sibma, S. 2011.
Southern seawater microfiltration pilot plant trial – the impacts of shock chlorination.
Proceedings of the International Desalination Association World Congress, Perth,
Australia.
309
Franks, R., Wilf, M., Voutchkov, N., Murkute, P., and Kizer, J. 2006. A Pilot Study Using
Seawater Reverse Osmosis Membranes In Combination With Various Pretreatments To
Meet The Challenges Of Pacific Seawater Desalination, Proceeding of the AMTA
Biennial Conference.
Gabelich, C. J., Chen, W. R., Yun, T. I., Coffey, B. M., and Suffet, I. H. 2005. The role of
dissolved aluminum in silica chemistry for membrane processes. Desalination 180(1-3),
307-319.
Gabelich, C. J., Ishida, K. P., Gerringer, F. W., Evangelista, R., Kalyan, M., and Suffet, I. H.
2006. Control of residual aluminum from conventional treatment to improve reverse
osmosis performance. Desalination 190(1-3), 147-160.
Gregory, R., Zabel, T. F., and Edzwald, J. K. 1999. Sedimentation and Flotation. Chapter 7.
In: Water Quality & Treatment: A Handbook of Community Water Supply. (Letterman,
R.D., Ed). AWWA 5th Ed. McGraw-Hill. New York: 509 p.
Guigui, C., Rouch, J. C., Durand-Bourlier, L., Bonnelye, V., and Aptel, P. 2002. Impact of
coagulation conditions on the inline coagulation/UF process for drinking water
production. Desalination 147, 95-100.
Gustalli, A., Simon, F. X., Penru, Y., de Kirchove, A., Llorens, J., and Baig, S. 2013.
Comparison of DMF and UF pretreatments for particulate material and dissolved organic
matter removal in SWRO desalination. Desalination 322, 144-150.
Haarhoff, J. and Edzwald, J. K. 2013. Adapting dissolved air flotation for the clarification of
seawater. Desalination 311, 90–94.
Halpern, D.F., Mcardle, J., and Antrim, B. 2005. UF pretreatment for SWRO: pilot studies.
Henderson, R. K., Baker, A., Parsons, S. A., and Jefferson, B. 2008a. Characterisation of
algogenic organic matter extracted from cyanobacteria, green algae and diatoms. Water
Research 42, 3435-3445.
Henderson, R. K., Parsons, S. A., and Jefferson, B. 2008b. Successful removal of algae
through the control of zeta potential. Separation Science and Technolog 43(7), 1653-
1666.
Henderson, R. K., Parsons, S. A., and Jefferson, B. 2008c. The impact of algal properties and
pre-oxidation on solid-liquid separation of algae. Water Research 42, 1827-1845.
Henderson, R. K., Parsons, S. A., and Jefferson, B. 2010. The impact of differing cell and
algogenic organic matter (AOM) characteristics on the coagulation and flotation of algae.
Water Research 44(12), 3617-3624.
Heng, L., Yanling, Y., Weijia, G., Xing, L., and Guibai, L. 2008. Effect of pretreatment by
permanganate/chlorine on algae fouling control for ultrafiltration (UF) membrane system.
Desalination 222(1-3), 74-80.
Hung, M. T. and Liu, J. C. 2006. Microfiltration for separation of green algae from water.
Colloids and Surfaces B: Biointerfaces 51(2), 157-164.
Huang, W., Chu, H., Dong, B., Hu, M., and Yu, Y. 2015. A membrane combined process to
cope with algae blooms in water. Desalination 355, 99-109.
Hyde, R. A. 1975. Water Clarification by Flotation – 4 Technical Report No.TR13,
Medmenham, England Water Research Centre.
310
Hydranautics. 2015. www.membranes.com. Accessed 1 November 2015.

Jamaly, S., Darwish, N. N., Ahmed, I., and Hasan, S. W. 2014. A short review on reverse
osmosis pretreatment technologies. Desalination 354, 30–38.
Jansenns, J. and Buekens, A. 1993. Assessment of process selection for particle removal in
surface water treatment. Aqua 42(5), 279-288.
Junli, H., Li, W., Nenqi, R., Li, L. X., Fun, S. R., and Guanle, Y. 1997. Disinfection effect of
chlorine dioxide on viruses, algae and animal planktons in water. Water Research 31(3),
455-460.
Kim, S-H., Min, C.-S., and Lee, S. 2011. Application of dissolved air flotation as
pretreatment of seawater desalination. Desalination and Water Treatment 33, 261-266.
Komlenic, R., Berman, T., Brant, J. A., Dorr, B., El-Azizi, I., and Mowers, H. 2013. Removal
of polysaccharide foulants from reverse osmosis feedwater using electroadsorptive
cartridge filters. Desalination and Water Treatment 51, 1050-1056.
Konno, H. 1993. Settling and coagulation of slender type diatoms. Water Science and
Technology 27(11), 231–240.
Lattemann, S. 2010. Development of an environmental impact assessment and decision
support system for seawater desalination plants. Doctoral dissertation, UNESCO-
IHE/TUDelft, Delft.
Lattemann, S. and Höpner, T. 2008. Environmental impact and impact assessment of
seawater desalination. Desalination 220, 1-15.
Le Roux, J., Croue, J-P., and Nada, N. 2013. Tracing disinfection byproducts in desalination
plants. Proceedings of the International Desalination Association World Congress,
Tianjin, China.
Le Roux, J., Nihemaiti, M., and Croué, J-P. 2015. The role of aromatic precursors in the
formation of haloacetamides by chloramination of dissolved organic matter. Water
Research 88, 371-379.
Ma, M., Liu, R., Liu, H., and Qu, J. 2012. Chlorination of Microcystis aeruginosa
suspension: cell lysis, toxin release and degradation. Journal of Hazardous Materials,
217-218, 279-285.
Ma, W., Zhao, Y., and Wang, L. 2007. The pretreatment with enhanced coagulation and a UF
membrane for seawater desalination with reverse osmosis. Desalination 203(1-3), 256-
259.
Mir, J., Morato, J., and Ribas, F. 1997. Resistance to chlorine of freshwater bacterial strains.
Journal of Applied Microbiology 82, 7-18.
Murrer, J. and Rosberg, R. 1998. Desalting of seawater using UF and RO – results of a pilot
study. Desalination 118(1-3), 1-4.
Myklestad, S. 1995. Release of extracellular products by phytoplankton with special
Naidu, G., Jeong, S., Vigneswaran, S., and Rice, S. A. 2013. Microbial activity in biofilter
used as a pretreatment for seawater desalination. Desalination 309, 254–260.
Pearce, G. K. 2007. Introduction to membranes: manufacturers’ comparison: part 1.
Filtration+Separation, 44(8), 36-38.
311
Pearce, G. K. 2010. SWRO pre-treatment: markets and experience. Filtration+Separation,

47(4), 30-33.
Pivokonsky, M., Naceradska, J., Kopecka, I., Baresova, M., Jefferson, B., Li, X., and
Henderson, R. K. 2016. The impact of algogenic organic matter on water treatment plant
operation and water quality: a review. Critical Reviews in Environmental Science and
Technology 46(4), 291-335.
Plantier, S., Castaing, J. B., Sabiri, N. E., Masse, A., Jaouen, P., and Pontie, M. 2013.
Performance of sand filter in removal of algal bloom for SWRO pretreatment.
Desalination and Water Treatment 51, 1838-1846.
Ransome, T., Ferguson, M., de Miguel, A., and Sibma, S. 2015. Operation of Large Scale RO
plant with Ultrafiltration Pretreatment: Challenges in Early Operation. Proceedings of
the IDA World Congress on Desalination and Water Reuse, San Diego, CA.
Regula, C., Carretier, E., Wyart, Y., Sergent, M., Gésan-Guizioud, G., Ferry, D., Vincent, A.,
Boudot, D., and Moulin, P. 2013. Ageing of ultrafiltration membranes in contact with
sodium hypochlorite and commercial oxidant: experimental designs as a new ageing
protocol. Separation and Purification Technology 103, 119-138.
Resosudarmo, A., Nappa, L., Ye, Y., Le-Clech, P., and Chen, V. 2017. Effect of physical and
chemical stress on ultrafiltration membrane performance during marine algal
blooms. Separation Science and Technology 52, 364 – 373.
Ripperger, S., Gösele, W., and Alt, C. 2012. Filtration, 1. Fundamentals. In: Ullmann's
Encyclopedia of Industrial Chemistry. Wiley-VCH Verlag GmbH & Co. KGaA,
Weinheim, Germany: pp. 677–709.
Rovel, J. M. 2003. Why a SWRO in Taweelah? - Pilot plant results demonstrating feasibility
and performance of SWRO on Gulf water. Proceedings of International Desalination
Association World Congress, Nassau, Bahamas.
Salinas Rodriguez, S. G. 2011. Particulate and Organic Matter Fouling of Seawater Reverse
Osmosis Systems: Characterization, Modelling and Applications. Thesis dissertation,
UNESCO-IHE/TUDelft, ISBN 978-0-415-62092-5, CRC Press/Balkema, Leiden.
Salinas Rodriguez, S. G., Kennedy, M. D., Amy, G. L., and Schippers, J. C. 2012. Flux
dependency of particulate/colloidal fouling in seawater reverse osmosis systems.
Desalination and Water Treatment 42, 155-162.
Schurer, R., Janssen, A., Villacorte, L. O., and Kennedy, M. D. 2012. Performance of
ultrafiltration and coagulation in an UF-RO seawater desalination demonstration plant.
Schurer, R., Tabatabai, A., Schippers, J. C., and Kennedy, M. D. 2013. Three years
operational experience with ultrafiltration as SWRO pre-treatment during algal bloom.
Shen, Q., Zhu, J., Cheng, L., Zhang, J., Zhang, Z., and Xu, X. 2011. Enhanced algae removal
by drinking water treatment of chlorination coupled with coagulation. Desalination 271,
236-240.
Shutova, Y., Karna, B., Hambly, A. C., Lau, B., Henderson, R. K., and Le-Clech, P. 2016.
Enhancing organic matter removal in desalination pretreatment systems by application
of dissolved air flotation. Desalination 318, 12-21.
312
SPI Engineering. 2010. West Basin Municipal Water District, Ocean Water Desalination
Pilot Program, Final Comprehensive Report 2002 – 2009, September 2, 2010.
Tabatabai, S. A. A. 2014. Coagulation and Ultrafiltration in Seawater Reverse Osmosis
Pretreatment. Doctoral dissertation, UNESCO-IHE/TU Delft, ISBN 978-1-138-02686-5,
CRC Press/Balkema, Leiden.
Tabatabai, S. A. A., Schippers, J. C., and Kennedy, M. D. 2014. Effect of coagulation on
fouling potential and removal of algal organic matter in ultrafiltration pretreatment to
seawater reverse osmosis. Water Research 59, 283-294.
Unni, S., Gorenflo, A., Al Suwaidi, E., Al Mulla, M., and Scott K. 2011. Two Year
Operational Experience Of SWRO Membrane With Arabian Gulf And Gulf Of Oman
Water. Proceedings of the IDA World Congress, Perth, Western Australia.
Valade, M. T., Edzwald, J. K., Tobiason, J. E., Dahlquist, J., Hedberg, T., and Amato, T.
1996. Particle Removal by Flotation and Filtration: Pretreatment Effects. Journal of
American Water Works Association 88(12), 35-47.
Villacorte, L. O. 2014. Algal blooms and membrane-based desalination technology. Doctoral
dissertation. UNESCO-IHE/TUDelft, ISBN 978-1-138-02626-1. CRC Press/Balkema,
Leiden.
Villacorte, L. O., Ekowati, Y., Neu, T. R., Klein, J. M., Winters, H., Amy, G., Schippers, J.
C., and Kennedy, M. D. 2015. Characterisation of algal organic matter produced by
bloom-forming marine and freshwater algae. Water Research 73, 216-230.
Voutchkov, N. 2009. SWRO pretreatment systems: choosing between conventional and
membrane filtration. Filtration+Separation 46(1), 5-8.
Voutchkov, N. 2013. Desalination Engineering - Planning and Design. McGraw-Hill, New
York: 637 p.
Vrouwenvelder, J. S. 2009. Biofouling of spiral wound membrane systems. Doctoral
dissertation. Delft University of Technology, Delft, The Netherlands. ISBN: 978-90-
814743-1-3, Ipskamp Drukkers, Enschede, The Netherlands.
WHO. 2007. Desalination for Safe Water Supply, Guidance for the Health and
Environmental Aspects Applicable to Desalination. World Health Organization (WHO),
Geneva: 161 p.
Wilf, M. and Schierach, M. K. 2001. Improved performance and cost reduction of RO
seawater systems using UF pretreatment. Desalination 135, 61-68.
Wilf, M., Awerbuch, L., Bartels, C., Mickley, M., Pearce, G., and Voutchkov, N. 2007. The
Guidebook to Membrane Desalination Technology: Reverse Osmosis, Nanofiltration
And Hybrid Systems Process Design, Applications And Economics. Rehovot: Balaban
Publishers: 524 p.
Winters, H. 1995. Biofouling: Its history and how it affects today's Desalination industry.
Proceedings of the International Desalination Association World Congress, Abu Dhabi,
UAE.
Winters, H. and Isquith, I. 1995. A Critical Evaluation of Pretreatment to Control Fouling in
Open Seawater Reverse Osmosis - Has it been a success? Proceedings of the
International Desalination Association World Congress, Abu Dhabi, UAE.
Wolf, P. H., Siverns, S., and Monti, S. 2005. UF membranes for RO desalination
pretreatment. Desalination 182, 293-300.
313
Xu, H., Chen, W., Xiao, H., and Hu, X. 2014. Stability of an ultrafiltration system for
drinking water treatment, using chlorine for fouling control. Desalination 336(1), 187-
195.
Yap, R. K., Holmes, M., Peirson, W., Whittaker, M., Stuetz, R., Jefferson, B., and Henderson,
R. 2012. Optimising dissolved air flotation/filtration treatment of algae-laden lagoon
effluent using surface charge: a Bolivar treatment plant case study. Water Science and
Technology 66(8), 1684-1690.
Zhang, J. D., Liu, Y. W., Gao, S. M., Li, C. Z., Zhang, F., Zen, H. M., and Ye, C. S. 2006.
Pilot testing of outside-in UF pretreatment prior to RO for high turbidity seawater
desalination. Desalination 189(1-3), 269-277.
Zheng, X., Ernst, M., Huck, P. M., and Jekel, M. 2010. Biopolymer fouling in dead-end
ultrafiltration of treated domestic wastewater. Water Research 44, 5212-5221.
Zhu, I. X., Bates, B. J., and Anderson, D. M. 2014. Removal of Prorocentrum minimum from
Research 2(1), 1-10.
314
Removal of algal toxins and taste and odor compounds
10 REMOVAL OF ALGAL TOXINS AND TASTE AND ODOR

COMPOUNDS DURING DESALINATION
Mike B. Dixon1, Siobhan F.E. Boerlage2, Holly Churman3, Lisa Henthorne3, and Donald M.
Anderson4
1
2
3
Water Standard, Houston, Texas, USA
4
Woods Hole Oceanographic Institution,Woods Hole MA USA

10.1 Introduction .........................................................................................................................................315
10.2 Chlorination .........................................................................................................................................316
10.3 Dissolved air flotation (DAF) .............................................................................................................317
10.4 Granular media filters ..........................................................................................................................317
10.5 Ultrafiltration/microfiltration ..............................................................................................................318
10.6 Reverse osmosis ..................................................................................................................................320
10.7 Sludge treatment and backwash disposal ............................................................................................324
10.8 Toxin removal in thermal desalination plants .....................................................................................324
10.8.1 Chemical and physical properties of marine HAB toxins ...........................................................325
10.8.2 Toxin barriers in thermal desalination plants ..............................................................................325
10.9 Chlorination prior to the distribution system ......................................................................................327
10.10 Summary ...........................................................................................................................................328
10.11 References .........................................................................................................................................329
10.1 INTRODUCTION
A major challenge in desalination is the removal of harmful algal bloom (HAB) toxins and
taste and odor compounds (hereafter referred to as algal metabolites) using common
treatment techniques. Removal of other compounds such as polysaccharides, proteins or
transparent exopolymer particles (TEP) are discussed in Chapter 2. Taste and odor
compounds are materials produced during a HAB that are not detrimental to human health,
but cause customer dissatisfaction and often a misconception that the drinking water is not
suitable for consumption. Toxins are detrimental to human health and are discussed in
Chapter 2. Here the objective is to assess each process unit in a common desalination
treatment train, both for SWRO and thermal desalination, and address how each is best
optimized to act as a barrier to these specific algal metabolites. Where treatment techniques
in seawater applications exist, these have been referenced and used as examples. As little
documentation exists on removal of algal metabolites from seawater blooms, fresh water
algal species are referred to whenever needed. This information is relevant in understanding
the removal mechanisms that are possible. For clarity, these are denoted for each example.
Algal metabolites can be either intracellular or extracellular. Many algal species have high
percentages of intracellular metabolites, such as Microcystis (freshwater) in which the toxin
microcystin can be up to 98% intracellular (Chow et al. 1997). Lefebvre et al. (2008) showed
an approximate 81% intracellular saxitoxin (STX)-equivalent concentration for an
Alexandrium (seawater) bloom, although further data are needed to confirm this observation.
STX-eq (or STX-equivalents) is a measure of total toxicity due to all saxitoxin analogues in a
particular solution. In contrast, Smith et al. (2012) report that 60% of the okadaic acid
produced by Dinophysis cultures was extracellular, while Kudela (pers. comm.) reported total
and extracellular concentrations of 100 and 50 µg/L domoic acid respectively during a
massive bloom of Pseudo-nitzschia along the US west coast in 2014. Extracellular
315
metabolites are inefficiently removed by pretreatment processes, and this is discussed below
in more detail.
The nutritional status of HAB cells will affect the percentage of extracellular metabolites in a
bloom. At the outset of a bloom, HAB cells will be more robust than toward the end of the
bloom period when stresses from nutrient limitation, grazing, or other factors can lead to the
leakage of metabolites into the seawater. Smith et al. (2012) noted that, in general, the
concentration of extracellular toxin in a lab culture of Dinophysis acuminata (seawater)
significantly increased upon culture aging and decline; cells did not appear to be actively or
passively releasing toxin during the stationary phase (see Chapter 1, Figure 1.3), but rather
extracellular release was likely a result of cell death.
10.2 CHLORINATION
By applying intake chlorination during a HAB event, the cells are lysed (ruptured) and,
because of the breakdown of the cell wall, there is a release of cell organelles and dissolved
compounds (including metabolites, TEP, polysaccharides, and proteins) which can cause a
significant impact on downstream SWRO
treatment processes. These dissolved
substances are more poorly removed by
dissolved air flotation (DAF) and dual
media filtration (DMF) and associated
flocculation/coagulation processes than
the intact cells, and therefore a large
amount can pass through pretreatment
unit processes to the SWRO process unit,
causing biofouling of the elements
(Villacorte 2014; see Chapter 2). Daly et
al. (2007) showed that for the blue-green
algal (or cyanobacterium) species
Microcystis aeruginosa in fresh water, a
chlorine concentration of 7 mg/L
completely lysed a cell density of
54,000,000 cells/L in 30 min.
Additionally, even a small amount of
Figure 10.1. The effect of chlorination (0.5 mg/L) on chlorine (1mg/L) can cause leakage of
the marine dinoflagellate, Pyrodynium bahamense, toxins, as the HAB cell wall is damaged
Destruction of UV treated and chlorinated Pyrodinium
cells in seawater (a and b). Ruptured thecal plates of
(Daly et al. 2007). Azanza et al. (2001)
cells after 2 and 6 min UV exposure. (c) Cell enveloped showed the effect of chlorination (as well
in mucilaginous-like substance after 2 min of chlorine as UV) on the marine dinoflagellate
exposure. (d) Clumping of Pyrodinium cells at 8 min Pyrodynium bahamense, noting that the
chlorine exposure. cellulose thecal plates of the cell wall turn
into mucilage. Figure 10.1 shows the
effect of a chlorine dose of 0.5 mg/L on these cells through time. Resosudarmo et al. (2014)
also indicated that chlorination (1 - 40 mg/L) of the marine alga Tetraselmis suecica caused
significant cell lysis and release of cellular contents into the seawater. Thus, the use of
chlorine as a pretreatment step should be avoided where possible when a bloom is present in
the source water, as chlorine causes intact cells to lyse, releasing intracellular toxin into
solution, and removal becomes more difficult. While thermal and SWRO systems should
both remove toxins efficiently, as discussed in Chapter 8, maximizing the number of
effective barriers against toxins is a prudent removal strategy. Some soluble toxins can be
destroyed by chlorine (Laycock et al. 2012), although in untreated raw seawater, the pH and
316
chlorine demand may reduce the effectiveness of this reaction and therefore it cannot be
relied upon as a viable treatment option (Daly et al. 2007; Boerlage and Nada 2014). Laycock
et al. (2012) demonstrated that saxitoxin, domoic acid, and okadaic acid in synthetic seawater
were completely destroyed by exposure to 10 ppm hypochlorite at 37 °C for 10 min, whereas
brevetoxin was unaffected. While these operating conditions are extreme within desalination
plants, it demonstrates possibilities for further investigations. Equipment warranties may
need review to ensure that maximum allowable chlorine expose is not exceeded.
By avoiding the use of chlorination at the intake, cell lysis can be avoided and therefore algal
metabolites can be kept intracellular and more easily removed by downstream processes.
During the majority of the bloom, SWRO pretreatment processes should then remove
intracellular metabolites efficiently, but as a bloom dies and the metabolites become
extracellular, pretreatment will become less efficient for metabolite removal and downstream
processes (such as SWRO and product water chlorination) will become the relevant treatment
methods for toxin removal. Taste and odor compounds MIB and geosmin are not denatured
by chlorine.
10.3 DISSOLVED AIR FLOTATION (DAF)
As mentioned in Chapter 9, DAF will remove cells by floating them to the surface of the
DAF tank. Due to the very low shear forces and encapsulation of the HAB cells with
coagulant (if used to aid flotation), cells are not lysed and are floated, unharmed, to the
surface of the DAF tank (Zhu and Bates 2012; Zhu et al. 2014). As the unharmed cells are
skimmed from the surface of the tank, the intracellular metabolites will be removed from the
treated water. When the HAB species has a high percentage of intracellular metabolites, the
majority will be removed by this step (Teixeira and Rosa 2006, 2007; Teixeira et al. 2010).
This metabolite removal technique is practiced at the freshwater Myponga WTP in South
Australia, which is regularly challenged with concentrated cyanobacteria that are successfully
removed using dissolved air flotation and filtration (DAFF; Qian et al. 2014).
In the DAF process, seawater HAB cell removal is expected to be >75% (Zhu and Bates
2012; Zhu et al. 2014). DAF removal of cells has been studied by many groups previously as
detailed in Chapter 9, but cell counts were often either relatively low (less than 1,000,000
cells/L), or counts were not undertaken; however, several studies have incorporated cell
counting in their research programs. Wiley et al. (2014) showed that when a bloom consisted
of 100,000,000 cells/L of the marine green alga Tetraselmis, the removal by DAF was as
high as 97%. Zhu et al. (2014) showed that DAF could remove >90% of the marine
dinoflagellate Prorocentrum minimum.
SWRO desalination plants that incorporate DAF specifically designed for algal removal will
also be a barrier for intracellular metabolites, which may be as high as 97% of the total in a
HAB. Removal is clearly dependent on the intracellular toxin percentage, and that can vary
with the physiological status of the cells, as well as the pretreatment strategies (e.g.,
chlorination).
Brevetoxin is hydrophobic and accumulates in bubbles, therefore some of this toxin may be
removed during DAF (Boerlage and Nada 2014). Pierce et al. (2004) reported that adding a
slurry of natural clay at the rate of 0.25 g/L removed 97% of brevetoxins associated with live
marine Karenia brevis (intracellular toxins) from seawater.
10.4 GRANULAR MEDIA FILTERS
GMF, a conventional filtration method, will remove HAB cells by trapping them between the
sand granules while the clean water passes out the bottom of the filter. Gravity GMF can
achieve around 90% removal of algal cells (Chapter 9), and thus intracellular metabolites will
317
be removed to the same percentage (Desormeaux et al. 2009). HAB cell removal has been
shown to be between 75 and 97% for pressurized GMF, thus metabolite removal percentage
is similar when comparing the same cell types (Desormeaux et al. 2011).
Similar to the DAF discussion above, coagulation will aid removal of HAB cells and ensure
that those cells remain unharmed by the filtration process as they are encapsulated inside
flocs (Dixon et al. 2011b, c). Studies by Velzeboer et al. (1995), Chow et al. (1998, 1999),
Dixon et al. (2012) and Drikas et al. (2001) have shown that alum and ferric
coagulation/flocculation (with flash mixing at 200rpm) do not compromise the membrane
integrity of cyanobacterial cells; if this is borne out with marine HAB species, the process
would not cause extracellular metabolite release. As mentioned above, removal of toxin will
be maximized if the metabolite remains intracellular, as extracellular metabolites will be
poorly removed by both coagulation and GMF. To maximize the effectiveness of the multi-
barrier treatment approach, care should be taken to restart the GMF filter properly after a
backwash to ensure that a concentrated portion of extracellular metabolite is not sent
downstream (see Chapter 9).
In some cases, a biofilm forms in the GMF with bacteria specific for biodegradation of
extracellular toxins and taste and odor compounds. Studies at the Morgan Water Treatment
Plant (WTP) in South Australia show the biodegradation of geosmin, affording a 70%
removal (McDowall 2008) in a GMF.
10.5 ULTRAFILTRATION/MICROFILTRATION
Ultrafiltration/microfiltration (hereafter referred to as UF) will remove intracellular

metabolites reliably, but not extracellular metabolites. There are many mechanisms for
metabolite removal in UF processes such as those illustrated in Figure 10.2 (Schäfer et al.
2011), but the only ongoing reliable method is by size exclusion of the HAB cells and
subsequent intracellular toxin removal.
In some cases there may be some incidental absorption of extracellular metabolites onto the
UF membrane fibers, but they will become quickly saturated (Dixon et al. 2011a) and
therefore this removal method is not operationally reliable. Desormeaux et al. (2009) also
observed this phenomenon for extracellular toxin removal, seeing a removal of 9-28% of a
100% extracellular domoic acid surrogate using two different pilot UF systems. Additionally,
extracellular metabolites can be removed by foulant-toxin interaction (Figure 10.2); however,
a large amount of foulant is required for any efficiency of the process and there is an inherent
inefficiency during backwashing, when the foulant is mostly removed from the UF
membrane. In one example case in South Australia at the freshwater Cowirra UF Plant on the
Murray River, more algal metabolite was removed than expected for a PVDF UF membrane
due to absorption by the foulants on the surface of the UF membrane (Newcombe 2011).
While regular removal of the extracellular algal metabolites 2-methyl isoborneol (MIB) and
geosmin (GSM) is less than 20%, during a serious fouling episode, removal of GSM was 40-
60%. This was due to unusually high DOC concentrations in the raw water, a side effect of
the conclusion to the Australian drought and referred to as a ‘black water’ event. DOC was
10mg/L or greater. While this shows that removal mechanisms other than size exclusion of
cells are possible, they cannot be relied upon consistently as a treatment barrier for
extracellular metabolites.
Many UF operators have concerns that cells can be broken by shear or pressurization during
normal UF operation. With UF filtration, shear will only cause minor cell lysis in regular
operating conditions. In a study by Ladner (2009), shear was maximized by repeatedly
running water through a needle valve with a small aperture (power density was 4x1010 W/m3).
318
Figure 10.2. Mechanisms for algal metabolite removal using membranes (modified from
Schäfer et al. 2011).
Ladner (2009) showed that the needle valve lysed cells of the marine dinoflagellate
Heterocapsa pygmaea, but the pump used to circulate the cells did not cause major amounts
of damage. Given that SWRO plants use submerged UF systems, there would be little shear
in comparison to Ladner’s experiment, in which the shear was maximized to exacerbate the
phenomenon.
When considering the conditions experienced in full-scale pressurized UF systems,
Resosudarmo et al. (2014) observed minimal lysis by running experiments between 50 to 150
kPa (0.5 to 1.5bar) for marine Tetraselmis suecica; however, if transmembrane pressure
(TMP) becomes very high, as it might during a HAB event, it can cause a greater amount of
cell lysis, so TMP should be controlled to as low as possible by maintaining frequent
backwash during bloom periods. Dixon et al. (2011b) and Chow et al. (1997) showed less
than 1% cell lysis of freshwater cyanobacterial cells (Microcystis) in an outside-in
pressurized UF experiment undertaken at 1-3 bar. Campinas and Rosa (2010) found that a
small amount of cell lysis occurred throughout the entire algal cell life cycle for the
freshwater Microcystis; however, it was more pronounced with older cells from a lab culture.
319
Dixon et al. (2011c) undertook UF experiments using flocculation/coagulation and found that
encapsulation within floc protect the cell from damage.
Dixon et al. (2011b, c) showed excellent saxitoxin removal using a pressurized outside-in UF
system and by keeping toxin intracellular. In one experiment, Dixon et al. (2011c) found that
total saxitoxin concentrations (intracellular and extracellular combined) from the freshwater
cyanobacterium Anabaena circinalis were 2.2 – 2.7 µg/L STX-eq in the feed water to a UF
membrane laboratory system, of which 31–38% was extracellular (0.7–0.8 µg/L STX-eq).
Results showed that when using alum coagulant, up to 68% removal of total saxitoxin was
achieved in the membrane tank as intact A. circinalis cells were removed via coagulation
prior to contact with the UF membrane. The majority of saxitoxin that was not removed was
extracellular. Extracellular saxitoxin (STX-eq) removal by the UF membrane itself (without
the effect of coagulation) was less than 20%.
Thus in UF, if TMP is minimized, intracellular metabolite removal will be maximized. Use
of a coagulant can aid cellular removal and help keep toxin intracellular, while keeping TMP
lower than if the coagulant was not used.
10.6 REVERSE OSMOSIS
If the pretreatment process performs properly, then a very small number of HAB cells should
be present entering the RO treatment step. Therefore intracellular metabolite removal is no
longer relevant and the RO mechanism is used to remove extracellular metabolites.
Additionally, if the pretreatment process is optimized, there should be minimal extracellular
metabolite concentration at the entry to the RO.
RO is an excellent barrier for removing extracellular metabolites and the removal mechanism
is the same as for removal of ‘organic micropollutants’ such as personal care products and
pharmaceuticals, which has been well studied in Europe and North America (Bellona et al.
2004; Verliefde et al. 2007, 2009; Schoonenberg Kegel et al. 2010).
Metabolite removal is governed by the properties of the RO (or in some cases nanofiltration
(NF)) membrane and the properties of the specific metabolite itself. Bellona et al. (2004)
reported that in estimating the rejection of a solute by high pressure membranes (RO, NF),
properties such as molecular weight cut-off (MWCO), desalting degree, porosity, membrane
morphology, and hydrophobicity of the membrane, and the molecular weight, molecular size,
charge, and hydrophobicity of the solute as well as the feedwater chemistry must all be
considered. A complete understanding of the solute and membrane characteristics that
influence rejection could lay the foundation for a modeling approach capable of predicting
the fate of specific compounds during high pressure membrane applications.
Given these mechanisms, if a metabolite is larger in molecular weight than approximately
200-300 Da (as a guide), then there will be excellent removal of the metabolite using RO.
Molecules 50-200 Da are more difficult to remove by RO. While the MWCO of RO is
theoretically approximately 100 Da (Dixon et al. 2012), the charge of the molecule becomes
more important for the 50-200 Da molecular weight range. If the molecule is negatively
charged, then the molecule will be repelled from the negatively charged RO surface. If the
molecule is positively charged, then it will be attracted to the surface of the membrane and
might be sorbed into the polyamide and pass into the permeate (Bellona et al. 2004; Verliefde
et al. 2007, 2009; Schoonenberg Kegel et al. 2010).
Fortunately, the most common HAB toxins are above 200 Da as discussed in Chapter 2.
Common toxins such as domoic acid and brevetoxin are far larger than saxitoxin in MW and
molecular size and will be well removed by size exclusion. Desormeaux et al. (2009)
undertook a pilot study in Monterey Bay, California, and due to a lack of a natural HAB,
320
kainic acid was selected as a toxin surrogate to spike into the treatment system as it has a
similar chemical structure to domoic acid, but is non-toxic. Kainic acid is a natural marine
acid contained in some species of seaweed and is a commonly used surrogate for domoic acid.
Dissolved kainic acid was spiked at concentrations 100 - 1,000 times greater than observed
during blooms of domoic acid-producing algae. Removal of the toxin surrogate was greater
than 99.5% for two different RO pilot systems, with a detection limit of 0.6 µg/L in seawater
and 0.017 µg/L in the RO product water. Seubert et al. (2012) undertook bench-scale RO
experiments to explore the potential of extracellular algal toxins contaminating RO product
waters. Concentrations exceeding maximal values previously reported during natural blooms
were used in the laboratory experiments, with treatments comprised of 50 µg/L of domoic
acid, 2 µg/L of saxitoxin and 20 µg/L of brevetoxin. None of the algal toxins used in the
bench-scale RO experiments were detectable in the desalinated product water. In the same
study by Seubert et al. (2012) monitoring for intracellular and extracellular concentrations of
domoic acid and saxitoxin within the intake and RO treated water from a pilot RO
desalination plant in El Segundo, California was conducted from 2005 to 2009. During the
five-year monitoring period, domoic acid and saxitoxin were detected sporadically in the
intake waters but never in the RO treated water. Another relevant study is that of Laycock et
al. (2012) in which a small laboratory-scale RO device was used to study HAB toxin removal.
Starting with 10.3 µg/mL of saxitoxin, 17.2 µg/mL of domoic acid, and 0.4 and 0.9 µg/mL of
okadaic acid and brevetoxin respectively, removal was 99.4, 99.0, 99.7 and 99.9 %,
respectively. While only a single pass through an RO membrane, the results are consistent
with previous studies mentioned above.
Given that the only existing relevant water quality guidelines relating to algal toxins (Brazil
and New Zealand) are in the range of 0.2 to 3 µg/L for saxitoxin and a worst case scenario
bloom may contain up to 600 ug/L of extracellular toxin (Chapter 8, Table 8.3), 99%
membrane removal can be an adequate treatment barrier for HAB toxins, depending on local
guideline concentrations. With the relatively low molecular weights of saxitoxin and domoic
acid (299 and 311 Da, respectively) and their hydrophilic nature, they are the most likely of
the common HAB toxins to pass through RO, as their molecular weights are the closest of
any HAB toxin to the theoretical MWCO of a RO membrane (~100Da). Brevetoxin (895 Da)
and okadaic acid (805 Da) are approximately eight times the MWCO of a RO membrane and
will therefore be easily removed. Despite saxitoxin being a smaller molecule, a study by
Dixon (2014) showed that saxitoxin and its congeners were removed to 99% or greater by a
tight nanofiltration membrane with a MWCO of ~100 Da (Table 10.1). The saxitoxin
analogues STX, GTX 3 and 4, and C1 and 2 were removed to greater than 99% by both NF
membranes. In parallel studies by Dixon et al. (2010, 2011a), it was shown that both SWRO
and BWRO membranes always removed toxin more efficiently than nanofiltration
membranes for toxins similar to saxitoxin in charge and size, such as cylindrospermopsin.
One can thus expect RO to remove saxitoxin just as well as this particular NF membrane.
This saxitoxin removal information correlates well with the work undertaken by Seubert et al.
(2012).
Dixon (2014) also showed that the smaller molecular weight non-toxic taste and odor
compounds MIB and geosmin were removed less efficiently than for saxitoxin (71-94%)
owing to their smaller molecular weight and size (168 and 182 Da respectively) (Table 10.1).
Given a heavily concentrated HAB producing MIB or geosmin, if the pretreatment system
completely fails and all the MIB and geosmin is extracellular, then a small amount of
material may pass into the product water. Given the non-toxic nature of taste and odor
compounds, the worst-case scenario would be customer complaints, but no risk to public
health exists.
321
Table 10.1. Removal percentages for algal metabolites using two NF membranes (NF90 and
NF270, Dow Filmtec).
Table Cyanobacterial % removal
metabolite NF90 NF270
2-Methylisoborneol (MIB) 71 83
Geosmin 94 89
Microcystin (mLR equiv.) 93 100
Saxitoxin (STX) 100 100
Gonyautoxin 3 (GTX3) 100 100
Gonyautoxin 4 (GTX4) 100 100
C1 99 99
C2 99 99
Unlike removal of salts and inorganics by RO, there is a less pronounced effect of feedwater
concentration when calculating organic micropollutant removal (such as toxins and taste and
odors). When considering salts, while membrane rejection will be ~99.7%, system rejection
may be around 98-99% depending in part upon the feed salinity and membrane array design.
This may not be so with organic micropollutants, as Fujioka et al. (2014) showed that a larger
concentration of N-Nitrosodiethylamine (NDMA) (up to 800 ng/L) did not increase the
permeate concentration of NDMA. Therefore, removal of saxitoxin and other toxins should
be somewhat independent of the membrane array and feed concentration (at maximum feed
concentrations of 600 µg/L) and be maintained through the first pass at approximately 99%.
Given many SW plants are two pass, then the total removal of saxitoxin would be 99% of the
original first pass 99% removal. Any blending of 1st and 2nd pass permeate should be
considered, but the net result may still be very low residual concentrations of the toxin.
Salt passage increases (due to membrane ageing or oxidation) in the RO process unit would
occur far sooner than any increase in product water HAB toxin concentration. For this reason
a major increase in permeate TDS could be used to detect an integrity breach that could later
lead to an increase in permeate toxin concentration. In a hypothetical study by Dixon et al.
(2015) a set of theoretical RO projections were undertaken to understand the failure mode of
how damage to the RO membrane may affect the permeate saxitoxin concentration during a
typical bloom. LG NanoH2O’s Q+ RO projection software was used as it allows the user to
model membrane deterioration independently of simple ageing factors. A typical SWRO
system from the Gulf was modeled (38,000 mg/L TDS, 110 pressure vessels (PV) per train, 7
membranes (M) per pressure vessel, 35 oC feed water temperature, 42% recovery, 5 year old
membranes, with supporting full second pass appropriately sized). For this hypothetical
modeling study, a feed water saxitoxin concentration of 10 µg/L was used, and it was
assumed that the pretreatment experienced full failure for removal of saxitoxin, meaning the
RO inlet saxitoxin concentration was also 10 µg/L. To exceed a hypothetical local saxitoxin
guideline value after the first pass of 1 µg/L, the plant would need to experience a first pass
permeate TDS of 2000 ppm. This corresponds to a gross loss of rejection in the elements, for
example from 99.7 to 99.0% NaCl rejection (when measured using a standard wet test,
32,000 mg/L NaCl, 25 oC, 800 psi, 8% recovery, pH 8). This also assumes a gross loss of
saxitoxin rejection from 99% to 95% in the first pass, to give a very large safety factor,
322
particularly for the higher water temperature in this case study. Given a full two pass system,
the second pass permeate would be approximately 0.3ug/L, despite the damaged first pass
elements. Given a partial split system with only 25% of water sent to second pass, this would
still produce a combined permeate saxitoxin concentration under the 1 µg/L guideline limit
for the above hypothetical scenario. Figure 10.3 summarizes the above theoretical case study
by showing plant conditions, saxitoxin concentration, and probable conductivity alarms
throughout the plant.
Figure 10.3. A summary showing a hypothetical scenario for saxitoxin removal through a typical partial
two pass RO system. The figure illustrates that alarms will be generated for 1st pass and 2nd pass TDS before
saxitoxin reaches a hypothetical local guideline concentration of 1µg/L. Figure: Dixon et al. 2015.
Such rejection losses could occur in several ways: 1) chlorination and oxidation of the
membranes; 2) accidental overdose of acid to below pH 3 for an extended period of time; or
3) an abundance of rolled permeate seals in the pressure vessels. In each case, the allowable
permeate TDS would be exceeded, causing plant alarms for high conductivity in the first and
second pass permeate. Regular plant TDS monitoring would show any membrane damage in
the first pass and a plant with two passes with a permeate TDS of less than the typical
300 mg/L is very unlikely to have detectable saxitoxin in the product water in given this
hypothetical case. Plant designers and operators could use this modeling exercise to analyze
their plant performance at maximum toxin concentrations predicted for certain plant localities.
While this hypothetical study used a saxitoxin concentration of 10 µg/L, some localities may
predict higher toxin concentration from data previously collected in that specific seawater.
By using this tool they could assess for potential alarm limits for conductivity that indicate a
potential presence of toxin in the permeate.
Despite this, toxin analysis during a bloom is prudent, as any unforeseen errors during
treatment could have a major impact on local public health. Some details on HAB toxin
analysis methodology are given in Chapter 2, and relatively simple methods for toxin
screening using ELISA kits and other assays are found in Appendix 2.
It is important to note that during toxic bloom conditions, toxin will most likely remain in the
waste brine from the SWRO process. Algal toxins are unlikely to be destroyed by most pre-
treatment processes, unless chlorination is undertaken in one or more of the unit processes.
Chlorination appears to degrade saxitoxin (Zamyadi et al. 2010; Laycock et al. 2012), domoic
acid, and okadaic acid, but not brevetoxins (Laycock et al. 2012). Hypochlorite
concentrations of 4 ppm or higher were sufficient to react with all of the saxitoxins, domoic
acid and okadaic acid in the samples that contained initial toxin concentrations up to
1250 ng/mL. Brevetoxins appeared to be unaffected in experiments in which the toxins were
exposed to up to 30 ppm hypochlorite in seawater at 35 °C for 60 min (Laycock et al. 2012).
323
Given the complications associated with HABs and chlorination such as biofouling, it is
unlikely that chlorination will be performed during pre-treatment. Consequently, the
concentration at the outfall will be no more than double that of the intake water, given most
desalination plants operate at under 50% recovery. While the concentration of toxin in the
immediate vicinity of the outfall may be more than in the intake water, this will be diluted
quickly to background levels, especially in modern desalination plants where mixing is
designed to occur rapidly. Impacts greater than that already experienced naturally due to the
bloom, if any, would be in the immediate vicinity of the outfall.
10.7 SLUDGE TREATMENT AND BACKWASH DISPOSAL
At the time of publication, no literature was available discussing sludge concentrations of

toxins and taste and odors in SWRO plants; however, freshwater literature was available.
Cell lysis has been documented to occur in the clarifier sludge, releasing intracellular toxins
(Drikas et al. 2001). This becomes a problem during clarification processes in conventional
drinking water plants, especially if long sludge retention times are evident in sedimentation
tanks, or sludge blanket clarifiers, and in particular where recycling of the supernatant from
the sludge to the head of the WTP is practiced.
Once confined in sludge, fresh water cyanobacteria may lose viability, die, and release
metabolites into the surrounding water (Newcombe et al. 2010). This can occur within one
day of treatment for some cyanobacteria, and could potentially result in very high dissolved
concentrations of algal metabolites. Similarly, algal cells carried onto sand filters, in flocs or
individually, could rapidly lose viability. As a result, where cyanobacteria (or marine HABs)
are potentially toxic, all sludge and sludge supernatant should be isolated from the plant until
the toxins have degraded sufficiently, wherever this is possible. Microcystins are readily
biodegradable (Newcombe et al. 2010) so this process should take 1-4 weeks.
Cylindrospermopsin appears to be slower to degrade and the biological degradation of
saxitoxins has not yet been studied; however, the latter are known to be stable for prolonged
periods (greater than 4 weeks) in source water, so caution is recommended. Intracellular
geosmin and MIB may also be released in sedimentation tanks and sludge treatment facilities.
This could result in increased taste and odor levels through the plant, or in the sludge
supernatant which, if it is returned to the head of the plant, could contribute significantly to
the levels entering the treatment plant (Newcombe et al. 2010). The possibility of this
occurring in individual treatment plants should be the focus of regular in-plant sampling.
10.8 TOXIN REMOVAL IN THERMAL DESALINATION PLANTS
While HABs do not have major operational impacts on thermal desalination plants as
discussed in Chapter 2, some water supply authorities have expressed concern related to the
removal of marine toxins by thermal desalination during toxic blooms. The removal of algal
toxins by thermal desalination processes has not been well researched. The study by Laycock
et al. (2012) experimentally determined the removal of marine toxins in the dissolved form,
i.e. extracellular in synthetic seawater using a bench scale micro distillation system. Boerlage
and Nada 2014 reviewed this work and examined the physical and chemical properties of the
four major classes of marine toxins that might be present at plant intakes to determine their
fate in thermal (and SWRO) desalination plant processes and the potential (residual) risk in
desalinated drinking water. Barriers to remove intracellular toxins in intact algal cells and
extracellular toxins from ruptured cells were identified. The following section is a summary
of Boerlage and Nada (2014).
324
10.8.1 Chemical and physical properties of marine HAB toxins

Four of the most potent and well characterized groups of marine toxins which could appear at
desalination plant intakes include saxitoxin, domoic acid, okadaic acid and brevetoxin. As
discussed in Chapter 2, the toxins have been classified based upon the poisoning syndromes
the toxins elicit. Physical and chemical properties of these toxins are summarized in Table
10.2. Algal toxins are structurally and functionally diverse, with varying charge, polarity, and
size, and many being derived from unique synthetic pathways (Wang 2008). Most of the
marine toxins that have high molecular weights are acid stable and non-volatile. Brevetoxins,
for example, are reported to withstand heat up to 300 ºC.
Table 10.2. Physico-chemical properties of common marine toxins
(http://www.chemspider.com/ 2016; http://www.latoxan.com/ 2016).
Vapor
Molecular Melting/Boiling
Human poisoning Pressure
Toxin Solubility weight Point
syndrome (mmHg
(Da) (ºC)
at 25ºC)
Paralytic shellfish Water soluble at
Saxitoxin 299 BP 549-575 0
poisoning (PSP) pH <7; stable
Brevetoxin 1 867 BP 197-199 NA
Neurotoxic
Brevetoxin 2 Fat soluble 895 MP 265- 270
shellfish poisoning
Brevetoxin 3 (liposoluble) 897 BP 291 - 293
(NSP)
Brevetoxin 9 899 MP 289 - 293
Water soluble
Amnesic shellfish
Domoic acid at pH <7 311 BP 607 0
poisoning (ASP)
Diarrhetic shellfish Slightly water
Okadaic acid 805 BP 921.6 0
poisoning (DSP) soluble
10.8.2 Toxin barriers in thermal desalination plants

Thermal desalination systems are quite robust in terms of source water quality. Therefore,
pretreatment is limited prior to MSF and MED plants, and typically comprised of feedwater
chlorination, screening, and chemical addition to prevent scaling and foaming. Feedwater is
screened to remove coarse debris to prevent equipment erosion by suspended solids and
prevent equipment from becoming blocked. For MSF, the allowable particle size for seawater
entering the tubes varying between 5- 15 mm (Gille 2003). On the other hand, MED needs
finer filtration, with the allowable particle size for seawater going through the spray nozzles
being < 0.5 mm.
Open intake screening commonly consists of coarse bar screening (75 to 150 mm) to remove
large debris and flotsam followed by mechanical fine screening (6-9.5 mm), e.g. travelling
band screens and drum screens to remove finer material and protect downstream processes.
Alternatively, only wedge wire screen may be employed with apertures ranging between 0.5
to 10 mm. Dinoflagellate and diatom cells can easily pass through these screens; for example,
Alexandrium spp. (the potent saxitoxin producers) are typically 15 to 48 µm in size. Hence,
screening will not serve as a barrier for algal cells, unless the screen is blinded, nor for
extracellular toxins. Instead shear forces during intake pumping and screening may break
down algal cell walls, particularly unarmoured cells like Karenia brevis whose cell walls are
fragile, releasing toxins into the seawater. Brevetoxins produced by K. brevis could become
aerosolized around onshore screens and could pose a respiratory risk to plant personnel if not
enclosed.
325
Chemical conditioning is utilized in thermal desalination in two treatment streams: the

seawater cooling water component and the seawater makeup water (used within the
desalination process). An oxidizing agent (usually chlorine) or biocide is continuously added
to the cooling water to prevent marine fouling, while antiscalants are continuously dosed to
prevent scaling on the heat exchanger surfaces. In addition, antifoaming agents are
continuously added to thermal process to prevent foaming in the deaerator and flash
chambers. Neither the antifoaming chemicals (polypropylene/polyethylene oxide,
isopropanol) nor the antiscalant (commonly polyacrylactes, polycarboxylic acids) are
expected to assist in removal of algal cells or detoxification of extracellular toxins.
Antiscalants are designed to modify crystal formation and disperse scaling ions and not
oxidize organic matter. Antifoam agents may have an effect on organic compounds
associated with algal blooms, but are not expected to degrade the toxin itself.
Most thermal desalination plants practice continuous chlorination at the seawater intake to
provide a residual chlorine concentration of approximately 0.15 – 0.3 mg/L to prevent
fouling marine growth in piping and biofilm formation on heat exchange surfaces.
Chlorination has also been proven to detoxify some marine toxins, with domoic acid the most
sensitive to chlorine - requiring only 1 ppm hypochlorite. Exposure to ≥ 4 ppm hypochlorite
for 10 min at 37 ºC completely destroyed saxitoxin and okadic acid (Laycock et al. 2012);
however, brevetoxin (3 and 300 µg/L concentration) was unaffected by exposure to
hypochlorite up to 30 ppm for one hour. Hence, chlorination is not a barrier for all marine
toxins. The experiments of Laycock et al. (2012) were in synthetic seawater with toxins
isolated from laboratory cultures. In practice the higher organics present during a bloom will
exert a chlorine demand, thereby reducing the efficiency of toxin degradation by chlorination,
potentially rendering it impractical as a degradation strategy. It is unlikely that continuous
chlorination of intake seawater can be applied at 4 ppm hypochlorite in thermal desalination
plants. In addition to increasing chemical consumption costs, the higher concentration of
chlorine will have a deleterious effect on the venting system and plant corrosion and the
guarantee values of various equipment may be exceeded. Finally, as discussed earlier in this
chapter, chlorine can result in the lysis of algal cells, thereby releasing intracellular toxins
into solution. Hence, chlorination should be avoided during a HAB when possible.
In thermal desalination systems, volatile and semi-volatile organics with boiling points lower
than water’s boiling point may carry over in the steam to contaminate the distillate and
therefore are vented out in the process. It is often assumed that high molecular weight
organics with high boiling points will remain in the brine, but this can sometimes be
erroneous. This is because the evaporation of organics from seawater and their condensation
into distillates is governed by a multitude of factors such as the temperature and pressure of
the MSF stage or MED effect and the concentration, vapor pressure, latent heat of
condensation of the individual compounds (Kutty et al. 1994).
The four major toxins presented in Table 10.2 are all reported to be heat stable, have low
vapor pressures, and are non volatile. The boiling points of saxitoxins, domoic acid, and
okadaic acid are significantly higher than water (at atmospheric pressure). Similarly, the
boiling point of brevetoxin is higher than that of water. These factors would suggest that the
toxins will not carry over in thermal desalination systems or co-distill, but instead will remain
in the flashing brine.
The results of Laycock et al. (2012) support high removal of toxins in the MSF and MED
desalination processes. The maximum temperature in that study was 104 ºC which is
approaching the top brine temperature of MSF (90 to 112 °C), but above MED (60 and
64°C). Three of the toxins, at unusually high test concentrations for the marine environment,
326
saxitoxins (10,340 µg/L), domoic acid (17,150 µg/L), and okadaic acid (400 µg/L), produced
from laboratory cultures of toxin producing species with optimal nutrient conditions, were
combined in one test solution with a synthetic seawater base, salinity 37 . Algal cell walls
may be broken down under the varying temperature and pressure conditions of MSF and
MED (if not already damaged by shear forces of pumps at screens). Therefore the majority of
toxins are expected to be extracellular, justifying the approach of using dissolved toxins in
these laboratory studies. Distillation results from Laycock et al. (2012) showed 99.5 to 99.9%
removal of the three extracellular toxins. Removal of the fourth toxin, brevetoxin, was
conducted in a separate series of tests, with the removal of that toxin somewhat lower than
for the other toxins, but still high at 98.3% removal. Similar to the other toxins, the test
concentration of brevetoxin (900 µg/L) is considered unusually high for natural bloom
conditions in the marine environment. Laycock et al. (2012) suggested that due to the
aerosolization nature of brevetoxin, it may result in carry over in a MSF plant; however, this
is expected to be very unlikely in MSF (and MED) plants as the toxins are non-volatile and if
present in droplets, will be captured by the demisters. The work of Laycock et al. (2012)
demonstrated that thermal desalination is an effective barrier for the removal of these marine
toxins, assuming no leaks in the system. The fate of these non-volatile toxins is then to be
discharged with the brine, which is combined with power plant cooling water for co-located
plants or recirculated into thermal systems with brine recycling. Nonetheless, more research
on toxin removal is recommended whereby,temperature and pressure conditions in MSF and
MED plants are simulated in a laboratory study to provide a higher level of confidence in the
results.
10.9 CHLORINATION PRIOR TO THE DISTRIBUTION SYSTEM
Following remineralization of the distillate/permeate, the water may be chlorinated for

distribution to the consumer so that there is a chlorine residual at the customer tap. This
provides a further barrier for some toxin removal after RO. The study of Laycock et al.
(2012) showed chlorination (in seawater) was ineffective in degrading brevetoxin.
Above1 ppm chlorination was effective in degrading domoic acid while saxitoxin and okadic
acid required 4 ppm or more. Zamyadi et al. (2010) showed that at pH 6.8-8, saxitoxin at a
concentration of 1.5 µg/L was degraded to less than 0.1ug/L after a contact time (CT) value
of 15 – 20 mg.min/L (Figure 10.4). Thus when product water is chlorinated at 3 ppm,
saxitoxin is degraded in five minutes. As desalination plants regularly have transfer pipelines
to transport water to the distribution system, ample chlorination time is usually available for
saxitoxin degradation.
The STX congeners (e.g., GTX 2&3 and C1&2) behaved similarly to the parent compound
during chlorination (Zamyadi et al. 2010). The longest required CT value was 35 mg.min/L
or 11.7 min at 3 ppm of chlorine (Figure 10.4). Operators have the potential to maximize the
speed of degradation of saxitoxin by increasing chlorine dose, as long as no other related
factors are detrimentally affected (such as production of disinfection by-products in
distribution systems where water is blended with surface water).
While chlorine is not normally dosed in drinking water at 4 ppm due to the objectionable
taste, chlorination in the distribution system may be an effective final barrier for toxin
removal in the treatment train. While it may not be required in practice, chlorination could be
used to form another layer of treatment to provide confidence for operators and water
authorities during toxic blooms.
327
Taste and odor compounds like geosmin are not degraded by chlorination. Geosmin is very
well detected by smell and taste, at around 10 ng/L, but is not toxic. It is therefore regularly
detected by customers of river and reservoir water. It will be removed moderately well by RO
(80-95%) (Dixon et al. 2010, 2011a) and
more so by a full two pass RO. Typical
bloom concentration in river and reservoir
sources can be as high as 100 ng/L. The
instrument detection limit for GC/MS is 4
ng/L and a full two pass system will
approximately produce a permeate
concentration of this value given a worst-case
scenario, if geosmin is allowed to stay
intracellular for best removal during
pretreatment.
Operators may consider chlorination
experiments on remineralized
distillate/permeate to see whether
chlorination is effective in fresh water and
whether a lower dose than that used by
Laycock et al. (2012) might be effective in
degrading saxitoxin and okadic acid. As
degradation of toxins is pH dependent, this
would need to be investigated in these
experiments. Given the current status of
research on the effect of chlorination in
degrading marine toxins, further work is
clearly required for greater confidence in
chlorination as a final barrier to all common
HAB toxins.
10.10 SUMMARY
This chapter describes the process steps

within SWRO and thermal desalination that
specifically remove HAB toxin and taste and
odor compounds and the limitations of each
process step. The general principle for HAB
toxin and taste and odor compound removal
is to ensure toxin remains intracellular to
maximize the removal efficiency of each
pretreatment step. By avoiding chlorination
of the intake or any shock chlorinations
during a HAB bloom, cells will not be
ruptured and release toxin. In SWRO plants,
extracellular toxin is difficult for the
Figure 10.4. Toxin oxidation in Myponga Reservoir downstream processes to remove, apart from
water at pH 8 with 3 mg/L (solid line) chlorine dose the RO process. DAF will effectively remove
(a) STX, (b) GTX2 (c) GTX3 (d) C1 and (e) C2. intact cells to approximately greater than 90%,
(Dashed line represents chlorination at 2 mg/L) thus removing intracellular toxin along with
(modified from Zamyadi et al. 2010)
the cells. GMF will remove similar
percentages of intracellular toxins to DAF.
328
UF/MF will remove greater than 99% of HAB cells and associated intracellular toxins,
although some minor leakage of toxins from the cells may occur due to shear and pressure in
the unit process. RO is the major toxin removal step and removes greater than 99% of
extracellular toxin. A two pass system will remove another 99% of the remaining 1%
extracellular toxin, although operators should take care to assess a partial split two pass
system. Taste and odor compounds are difficult to remove when extracellular. Should off
taste and odor occur with the HAB bloom, pretreatment will remove intracellular compounds
to greater than 90% (similar to removal of toxins), while each full RO pass will remove 60-
80% of the extracellular taste and odor. Chlorination will not remove taste and odor
compounds. As common HAB tase and odor compounds can be detected at around 10ng/L,
single pass RO systems may experience customer complaints. Any sludge produced from
pretreatment will still contain toxin and care must be taken when considering any supernatant
return to the plant or disposal of the sludge. Removal of toxins from thermal systems should
be in the order of 99%, and toxin should exit the plant in the brine. As an additional barrier to
toxin removal, some toxins are denatured by chlorination, for example saxitoxin (STX) will
be removed with a CT of 15 µg.min/L. Thus the inherent multiple barrier approach to
seawater desalination systems creates an effective removal system for HAB toxin with built
in redundancy.
10.11 REFERENCES
Azanza, M. P. V., Azanza, R. V., Gedaria, A. I., Sententa, H. G., and Idjao, M. V. 2001.
Decimal reduction times of Pyrodinium bahamense var. compressum and Escherichia
coli in chlorine- and ultraviolet-treated seawater. Letters in Applied Microbiology 33(5),
371-376.
Bellona, C., Drewes, J. E., Xu, P., and Amy, G. 2004. Factor affecting the rejection of
during NF/RO treatment - a literature review. Water Research 38, 2795-2809.
Boerlage, S. F. E., and Nada, N. 2014. Algal toxin removal in seawater desalination
processes, In: Proceedings of European Desalination Society, Cyprus.
Campinas, M., and Rosa, M. J. 2010. Evaluation of cyanobacterial cells removal and lysis by
ultrafiltration. Separation and Purification Technology 70, 345-353.
Chemspider. 2016. www.chemspider.com. Royal Society of Chemistry, Accessed 13 Feb
2016.
Chow, C. W. K., Drikas, M., House, J., Burch, M. D., and Velzeboer, R. M. A. 1999. The
impact of conventional water treatment processes on cells of the cyanobacterium
Microcystis aeruginosa. Water Research 33(15), 3253–3262.
Chow, C.W.K., House, J., Velzeboer, R.M.A., 1998. The effect of ferric chloride flocculation
on cyanobacterial cells. Water Research 32, 808–814.
Chow, C. W. K., Panglish, S., House, J., Drikas, M., Burch, M. D., and Gimbel, R. 1997. A
study of membrane filtration for the removal of cyanobacterial cells. Aqua, 46(6), 324–
334.
Daly, R. I., Ho, L., and Brookes, J. D. 2007. Effect of chlorination on Microcystis aeruginosa
cell integrity and subsequent microcystin release and degradation. Environmental
Science and Technology 41(12), 4447-4453.
DesalData. 2015. Worldwide desalination inventory (MS Excel Format), downloaded from
DesalData.com (GWI/IDA) 2015.
329
Desormeaux, E. D., Meyerhofer, P. F., and Luckenbach, H. 2009. Results from nine
investigations assessing Pacific Ocean seawater desalination in Santa Cruz, California.
Proceedings of the IDA World Congress, Dubai, UAE, November 2009.
Desormeaux, E. D., Meyerhofer, P. F., Luckenbach, H., and Kudela, R. M. 2011. Pilot-
testing multiple pretreatment systems for seawater desalination. IDA Journal of
Desalination and Reuse 3(1), 42-52.
Dixon, M. B. 2014. Removal of toxin and taste & odor compounds using membranes and
associated processes. Proceedings of the Middle East Desalination Research Centre’s
HABs and Desalination workshop, Muscat, Oman, April 2014.
Dixon, M. B., Churman, H., and Henthorne, L. 2015. Harmful algae blooms and desalination:
a cells journey from sea to SWRO. Proceedings of the IDA World Congress, San Diego,
California, Sept 2015.
Dixon, M. B., Falconet, C., Ho, L., Chow, C. W. K., O’Neil, B., and Newcombe, G. 2010.
Nanofiltration for the removal of algal metabolites and the effects of fouling. Water
Science and Technology 61, 1189–1199.
Dixon, M. B., Falconet, C., Ho, L., Chow, C. W. K., O’Neil, B. and Newcombe, G. 2011a.
Removal of cyanobacterial metabolites by nanofiltration from two treated waters.
Journal of Hazardous Materials 188(1), 288–295.
Dixon M. B., Ho, L., Chow, C., Newcombe, G., Croue, J-P., Buisson, H., Cigana, J., and
Treuger, R. 2012. Water Research Foundation Report #4016: Evaluation of integrated
membranes for taste and odour and algal toxin control, Published by Water Research
Foundation, Denver, Colorado, USA.
Dixon, M. B., Richard, Y., Ho, L., Chow, C. W. K., O’Neil, B. and Newcombe, G. 2011b. A
coagulation-powdered activated carbon-ultrafiltration – Multiple barrier approach for
removing toxins from two Australian cyanobacterial blooms. Journal of Hazardous
Materials 186(2), 1553–1559.
Dixon, M. B., Richard, Y., Ho, L., Chow, C. W. K., O’Neil, B., and Newcombe, G. 2011c.
Integrated membrane systems incorporating coagulation, activated carbon and
ultrafiltration for the removal of toxic cyanobacterial metabolites from Anabaena
circinalis. Water Science and Technology 63(7), 1405-1411.
Drikas, M., Chow, C. W. K., House, J. and Burch, M. D. 2001. Using coagulation,
flocculation, and settling to remove toxic cyanobacteria. Journal American Water Works
Association 93, 100–111.
Fujioka, T., Tu, K. L., Khan, S. J., McDonald, J. A., Roux, A., Poussade, Y., Drewes, J. E.,
and Nghiem. L .D. 2014. Rejection of small solutes by reverse osmosis membranes for
water reuse applications: a pilot-scale study. Desalination 350, 28-34.
Gille, D. 2003. Seawater intakes for desalination plants. Desalination 156(1), 249-256.
Kutty P. C. M., Nomani, A. A. and Thankachan,T. S. 1994. Carry over of some high
molecular weight organics from seawater to MSF distillates SWCC RDC 25 Technical
Report.
Ladner, D. A. 2009. Effects of bloom-forming algae on fouling of integrated membrane
systems in seawater desalination. PhD dissertation, University of Illinois at Urbana-
Champaign.
330
Laycock, M. V., Anderson, D. M., Naar, J., Goodman, A., Easy, D. J., Donovan, M. A., Li,
A., Quilliam, M. A., Al Jamali, E., and Alshihi, R. 2012. Laboratory desalination
Lefebvre, K. A., Bill, B. D., Costa, P. R., Nance, S., Erickson, A., Trainer, V. L., and Baugh,
K. A. 2008. Characterization of intracellular and extracellular saxitoxin levels in both
field and cultured Alexandrium spp. samples from Sequim Bay, Washington. Marine
Drugs 6(2), 103-116.
McDowall, B. 2008. Removal of geosmin and 2-methylisoborneol from drinking water
through biologically active sand filters. PhD Thesis, The University of Adelaide.
Newcombe, G. 2011. Personal Communication.
Newcombe, G., House, J., Ho, L., Baker, P., and Burch, M. 2010. Management strategies for
cyanobacteria (blue-greenalgae): a guide for water utilities: cooperative research centre
for water quality and treatment research report #74.
Pierce, R. H., Henry, M. S., Higham, C. J., Blum, P., Sengco, M. R. and Anderson, D. M.
2004. Removal of harmful algal cells (Karenia brevis) and toxins from seawater culture
by clay flocculation. Harmful Algae 3(2), 141–148.
Qian, F., Dixon, D. R., Newcombe, G., Ho, L., Dreyfus, J. and Scales, P. J. 2014. The effect
of pH on the release of metabolites by cyanobacteria in conventional water treatment
processes. Harmful Algae 39, 253-258.
Resosudarmo, A. Ye, Y. Le-Clech, P.. amd Chen, V. 2014. UF pretreatment for desalination
during algal blooms: Negative effects of stress and benefits of hybridised systems.
Proceedings of the HABs and Desalination Worksop, Muscat, Oman April 2014.
Schaefer, A., Akanyeti, I., and Semiao, A. J. C. 2011. Micropollutant sorption to membrane
polymers: A review of mechanisms for estrogens. Advances in Colloid and Interface
Science 164(1–2), 100-117.
Schoonenberg Kegel, F., Rietman, B. M., and Verliefde, A. R. D. 2010. Reverse osmosis
followed by activated carbon filtration for efficient removal of organic micropollutants
from river bank filtrate. Water Science and Technology 61(10), 2603-2610.
Seubert, E. L., Trussell, S., Eagleton, J., Schnetzer, A., Cetinic, I., Lauri, P., Jones, B. H., and
Caron, D. A. 2012. Algal toxins and reverse osmosis desalination operations: laboratory
bench testing and field monitoring of domoic acid, saxitoxin, brevetoxin and okadaic
acid. Water Research 46(19), 6563-6573.
Smith, J. L., Tong, M., Fux, E., and Anderson D. M. 2012. Toxin production, retention, and
extracellular release by Dinophysis acuminate during extended stationary phase and
culture decline. Harmful Algae 19, 125–132.
Teixeira, M. R., Sousa, V., and Rosa, M. J. 2010. Investigating dissolved air flotation
performance with cyanobacterial cells and filaments. Water Research 44(11), 3337-3344.
Teixeira, M. R. and Rosa, M. J. 2006. Comparing dissolved air flotation and conventional
sedimentation to remove cyanobacterial cells of Microcystis aeruginosa. Part I: The key
operating conditions. Separation and Purification Technology 52(1), 84-94.
Teixeira, M. R. and Rosa, M. J. 2007. Comparing dissolved air flotation and conventional
sedimentation to remove cyanobacterial cells of Microcystis aeruginosa. Part II. The
effect of water background organics. Separation and Purification Technology 53(1),
126-134.
331
Verliefde, A. R. D., Cornelissen, E. R., Amy, G., Van der Bruggen, B., and Van Dijk, H.
2007. Priority organic micropollutants in water sources in Flanders and the Netherlands
and assessment of removal possibilities with nanofiltration. Environmental Pollution
146(1), 281-289.
Verliefde, A. R. D., Cornelissen, E. R., Heijman, S. G. J., Verberk, J. Q. J. C. and Amy, G. L.
2009. Construction and validation of a full-scale model for rejection of organic
micropollutants by NF membranes. Journal of Membrane Science 339(1), 10-20.
Velzeboer, R., Drikas, M., Donati, C., Burch, M., and Steffensen, D. 1995. Release of
geosmin by Anabaena circinalis following treatment with aluminium sulphate. Water
Science and Technology 31, 187–194.
Villacorte, L. O. Algal Blooms and Membrane Based Desalination Technology, UNESCO-
IHE PhD Thesis Dissertation, May 20 2014.
Wiley, R. 2014. Proceeding of the American Membrane Technology Membrane Technology
Conference, March 2014, Las Vegas, NV.
Wang D.-Z. 2008. Neurotoxins from marine dinoflagellates: a brief review. Marine Drugs
6(2): 349-371.
www.latoxan.com. Natural Active ingredients, Accessed 13 Feb 2016.
Zamyadi, A., Ho, L., Newcombe, G., Daly, R. I., Burch, M., and Baker, P. 2010. Release and
oxidation of cell-bound saxitoxins during chlorination of Anabaena circinalis cells.
Environmental Science and Technology 44(23), 9055-61.
Zhu, I. X. and Bates, B. 2012. Seawater desalination pretreatment for harmful algae blooms
using dissolved air flotation. IDA Journal of Desalination and Water Reuse 4(1), 34-37.
Zhu, I. X., Bates, B. J. and Anderson, D. M. 2014. Removal of Prorocentrum minimum from
Research 2(1), 47-56.
332
Case histories for HABs in desalination
11 CASE HISTORIES FOR HARMFUL ALGAL BLOOMS IN

DESALINATION
Siobhan F.E. Boerlage1, Mike B. Dixon2 and Donald M. Anderson3

1
2
3

11.1 Introduction ...........................................................................................................................................333
11.2 Fujairah 2, United Arab Emirates – Effects of harmful algal blooms on plant operations in
2008 and 2013 .......................................................................................................................................347
11.3 Sohar, Oman – Harmful algal bloom impact on membrane pretreatment: challenges and solutions ...355
11.4 Barka 1, Oman - The impact of harmful algal blooms on the performance stability of UF
pretreatment ...........................................................................................................................................367
11.5 Shuwaikh, Kuwait – Harmful algal bloom cell removal using dissolved air flotation:
pilot and laboratory studies ...................................................................................................................383
11.6 La Chimba, Antofagasta, Chile – Oxygen depletion and hydrogen sulfide gas mitigation due to
harmful algal blooms .............................................................................................................................391
11.7 Mejillones, Chile – Operation of the ultrafiltration system during harmful algal blooms at
the Gas Atacama SWRO plant ..............................................................................................................399
11.8 Antofagasta, Chile - Abengoa water micro/ultrafiltration pretreatment pilot plant ..............................409
11.9 Tampa Bay, Florida (USA) – Non-toxic algal blooms and operation of the SWRO plant
detailing monitoring program for blooms .............................................................................................417
11.10 Jacobahaven, The Netherlands – Ultrafiltration for SWRO pretreatment: a demonstration plant ........425
11.11 Barcelona, Spain – SWRO demonstration plant: DAF/DMF versus DAF/UF .....................................439
11.12 Gold Coast, Queensland, Australia - Deep water intake limits Trichodesmium ingress .......................447
11.13 Berlin, Germany – akvola: An integrated DAF-UF pilot ......................................................................459
11.1 INTRODUCTION
Algae have long been an issue impacting desalination plant operation in areas prone to algal
blooms or where macroalgae (seaweeds) and detritus became dislodged from the seabed.
Previously and still today, operators and designers may elect to turn down production or shut
down SWRO plants, if contract obligations allow, when blooms are infrequent or of short
duration. Alternatively, in areas subject to frequent and prolonged blooms, additional
pretreatment such as conventional dissolved air flotation (DAF), hitherto designed for
brackish water applications, began to be employed as early as 1995.
The unprecedented 2008/2009 bloom of Cochlodinium polykrikoides in the Gulf of Oman
and the Gulf1, brought algal blooms to the fore in the desalination industry. SWRO plant
shutdowns were up to four months long as pretreatment processes struggled to remove the
increased biomass and produce the required RO feedwater quality. Apart from a few
exceptions, thermal desalination plants continued to operate without major issue throughout
the bloom, as phytoplankton blooms generally pass through intake screens and thermal
processes are very forgiving of source water quality. This was demonstrated at the Fujairah 1
hybrid desalination plant where the multi-stage flash (MSF) plant operated throughout the
bloom while the adjacent SWRO plant was shut down.
1
333
Globally, harmful algal blooms (HABs) similar to the 2008 bloom of Cochlodinium
polykrikoides are increasing in frequency and severity (Anderson et al. 2012). Coupled with
the increasing use of RO as the desalination technology of choice, HABs have become one of
the major challenges facing the industry as RO membranes are extremely vulnerable to
feedwater quality, making pre-treatment exceptionally important. Smooth operation is
contingent on the selection of appropriate pretreatment processes upstream to remove
organics, solids, colloids and other foulants from the RO feedwater. The 2008 Gulf HAB
highlighted the limitations of conventional pretreatment based on ferric chloride coagulation
and single stage dual media filtration (DMF) in removing algal biomass and organics.
Ongoing research efforts to identify the algal organic matter (AOM) constituents responsible
for membrane fouling and measurement of their removal in pretreatment intensified. To this
end, the spike in AOM occurring during a bloom was found to comprise mainly of high
molecular weight biopolymers (polysaccharides and proteins), which include sticky
transparent exopolymer particles (TEP) (Myklestad 1995; Villacorte 2014). TEP have been
shown to form microgels with a high hydraulic resistance and are increasingly recognized to
promote biofouling of RO membranes (Villacorte 2014; Berman and Holenberg 2005; Li et
al. 2015). With the increasing adoption of low pressure microfiltration (MF) and
ultrafiltration (UF) membrane pretreatment, questions were raised as to their performance
during algal bloom events and how they compared to conventional pretreatment in removal
of AOM.
In preparing the Manual and to address some of the above questions, operators, researchers,
and plant owners in the desalination industry were contacted as part of an informal survey
and invited to contribute
case studies related to their
experience with algal
blooms. As expected, it
became clear that algal
bloom issues were
predominantly encountered
in SWRO plants rather than
those using thermal
desalination. Twelve SWRO
plants (Figure 11.1.1) at
eleven different sites were in
a position to share their
experiences from a shortlist
Figure 11.1.1. Location of the 12 plants presented in the case studies
that have experienced algal blooms (two were at the same site).
of 30 sites that may have
experienced HAB issues.
Algal blooms, primarily phytoplankton, were reported in almost all geographic locations, in
cold and warm seas over a range of salinities affecting municipal and industrial desalination
plants. Notable areas affected include the warmer waters of the Gulf of Oman and the Gulf in
the Middle East. Case studies include Sohar and Barka 1 in Oman, Fujairah 2 in UAE and the
Shuwaikh plant located close to Kuwait’s most important commercial port in the upper
reaches of the Gulf where seawater quality is at its poorest. HABs are also commonly found
in the cooler waters off the coast of Antofagasta in Northern Chile supplying industry and
drinking water for towns in one of the driest areas of the world.
Key insights from the 12 case studies are summarized below in terms of impacts experienced,
if any, in both conventional and advanced MF/UF membrane pretreatment plants during algal
blooms. Commonly recommended measures implemented in the industry to combat algal
334
blooms are discussed in relation to the case studies mitigation strategies, and lessons learned.
This encompasses measures adopted during design and/or during plant operation, e.g. deep-
water intakes (Gold Coast), and DAF (Fujairah 2, Shuwaikh) and/or direct MF/UF filtration
(Jacobahaven, Sohar), or subsequently enacted in response to HAB events (La Chimba).
11.1.1 Algal-related impacts in desalination
HABs pose two main operational risks in desalination, namely safety of the drinking water
produced (which relates to the toxicity of the bloom) and security of supply (Boerlage and
Nada 2014). HABs are broadly classified as toxic or non-toxic. Toxic algal blooms produce
potent toxins, causing illness or mortality in humans, fish, marine mammals, and other
marine life through either the direct exposure to the toxin or ingestion of bioaccumulated
toxin in higher trophic levels e.g. shellfish consumption. In desalination, marine algal toxins
represent a potential health risk to the safety of desalinated drinking water if present in
sufficiently high concentrations in the seawater and they breakthrough through the
desalination process.
Non-toxic HABs can cause damage to ecosystems, fisheries resources, recreational areas, and
commercial facilities such as desalination plants, sometimes because of the biomass of the
accumulated algae, and in other cases due to the release of compounds that are not toxins
(i.e., reactive oxygen species, polyunsaturated fatty acids, organic matter, mucilage) that can
be lethal to marine animals or that can cause disruptions of other types. In desalination, these
blooms represent an operational risk to plants, threatening water supply security through
unplanned outages, and loss of production at any point in the process from blinding of intakes
(less common) through to failure of pretreatment unit processes and/or the RO system itself.
Such blooms can have massive financial impacts. Huge economic losses were reported, for
example, at Oman’s Sohar Industrial Port Area (SIPA), which supplies desalinated water and
seawater cooling water to industrial customers. The non-toxic bloom of Cochlodinium
polykrikoides in 2008/2009 resulted in an increase in the frequency of cleaning seawater
intake screens from every 12 hours to every 4 hours and an inability to maintain the required
free residual chlorine. The independent SWRO plant operated by the Sohar Refinery at SIPA
shut down for four months and required 100% membrane replacement due to severe
biofouling.
A myriad of HAB impacts in desalination were identified in the cases studies. Some are
easily identified with most posing a risk to the security of desalinated water supply as in
general, the causative species of the HAB, where identified, were non-toxic. Other impacts
were more obscure and relate more to the perception that desalinated drinking water may be
unsafe (e.g., malodorous) or even to the perceived environmental impacts of brine during an
algal bloom event.
Perceived impacts associated with safety of drinking water may follow the decay of an algal
bloom. In 2011, the La Chimba plant, supplying 60% of Antofagasta’s drinking water,
detected hydrogen sulfide at the seawater intake. The seawater, abstracted from a deep water
intake, had become hypoxic (DO<0.5 mg/l) due to the decay of an intense bloom of
Prorocentrum micans made worse by thermoclines that restricted mixing and movement of
water in the bay. Under these conditions, sulfate reducing bacteria (SRB) flourished and
hydrogen sulfide was generated. While the inhalation of hydrogen sulfide in air is well
known to be extremely toxic, there are no data on the human health effects of ingesting water
that contains hydrogen sulfide (NHMRC/NRMMC 2011). Instead, impacts are mainly related
to the foul smell of rotting eggs, with hydrogen sulfide having a very low taste and odor
threshold in water, estimated to be as low as 0.05 mg/l (WHO 2011). As a gas, hydrogen
sulfide can readily pass through RO membranes resulting in customer complaints and/or the
335
perception that water produced following a HAB is unsafe. Alternatively, oxidation of

hydrogen sulfide could lead to elemental sulfur which has a detrimental effect on membranes.
The plant owner therefore quickly enacted a raft of measures to remove H2S from the product
water.
Another atypical issue encountered during an algal bloom event was that a desalination plant
was thought to be the cause of a dark plume in the seawater in an area renowned for its
beaches and recreational fishing. A vast Trichodesmium bloom occurred in the area of the
brine outlet (and intake) during commissioning of the Gold Coast plant which led a member
of the public to complain to the local Environmental Protection Agency (EPA) of a dark
plume emanating from the brine outlet. The bloom was, in fact, demonstrated to frequently
occur naturally in the area (prior to the construction of the plant) and shown not to be
entrained in the plant intake. If it was entrained, the solids associated with the high
concentration of algal cells would be removed during residual treatment and not be returned
to sea. Hence, no plume would be evident from the brine outfall. Nonetheless, this
demonstrates potential community perception issues associated with algal bloom events in
desalination.
The more typical algal-related issues in SWRO desalination arising from the high biomass
and organics accompanying non-toxic HABs are well known in the desalination industry.
High biomass blooms can increase suspended solids beyond design thresholds, overloading
conventional granular media filters (GMF) and resulting in rapid filter clogging. If the
coagulation dose becomes prohibitively high in response to a bloom, surface clogging of the
media may occur, severely limiting filtration capacity. Downstream this may lead to more
frequent cartridge filter replacement and cleaning of the RO membranes due to a higher iron
residual and higher concentration of colloids and AOM remaining in the RO feedwater. At
worst, plant production is lost due to frequent media backwashing and short filter runs, and
conventional pretreatment may fail to produce water to meet RO guidelines. Plants are then
forced to void warranties and continue operating or shut down to avoid the risk of irreversible
RO membrane fouling. In the latter case, they may also incur cost penalties associated with
loss of production. Conventional pretreatment was employed in 5 of the 12 case studies
(Tampa Bay, Gold Coast, Barcelona, Atacama Plant 1, Fujairah 2). Shorter filter runs were
observed for the diatomaceous earth filters at Tampa Bay during algal blooms, along with
foaming. No negative impacts were observed in the gravity DMF at the Gold Coast plant as
the deep water intake limited ingress of the most frequently occurring algal bloom. The same
was true at Fujairah 2, which employs additional solids removal through an upstream DAF.
Similarly, HAB impacts are found in microfiltration or ultrafiltration which was the
predominant pretreatment choice in the case studies. This involved a variety of membrane
materials and pore sizes either in the pressurised inside-out or outside-in flow configuration.
Here impacts were as expected and included blocking of strainers due to higher solids
loading and increases in transmembrane pressure (TMP) in order to maintain a constant
permeate flux as pretreatment membranes fouled. Backwashing and chemically enhanced
backwashing (CEB) intervals were severely shortened in some cases and less effective than
in non-bloom conditions, with a marked membrane permeability decrease over time.
Additional cleaning-in-place (CIP) to recover initial permeability was required in some cases
(Sohar, Barka 1, Antofagasta, Jacobahaven).
Biofouling of the SWRO membranes was noted at the Sohar plant during algal bloom events
in 2013 (and earlier in 2008 at the Sohar Refinery SWRO plant as mentioned above), which
uses direct MF membrane pretreatment without coagulation. Additional RO membrane
cleaning was also required at the SWRO plant at Atacama using conventional pretreament
but the nature of fouling was not identified. Pretreatment was effective during algal bloom
336
events for most of the remaining case studies with no additional cleaning reported for the
SWRO or fouling was not directly attributed to algal blooms. RO fouling is typically
complex with more than one type of fouling occurring and may be the synergistic effect of
operating conditions prior, during or after a bloom e.g. overdosing of ferric coagulants
causing iron fouling along with biofouling during or following the termination of the bloom.
11.1.2 Algal bloom mitigation and lessons learned
11.1.2.1 Characterization of seawater quality and piloting
Ideally, prior to SWRO plant design, seawater quality is thoroughly characterized through a
long-term monitoring study (see section 11.1.2.1.1 for parameters) to assist in selecting an
intake site and pretreatment processes from which the plant footprint and layout can be
estimated. This should be coupled to a review of historical records to determine factors that
impact on water quality such as the frequency and severity of algal blooms. Hydrodynamic
conditions at the intake area should also be considered, such as the presence of thermoclines
or upwelling that may promote HABs. Monitoring methods and approaches are covered in
detail in Chapters 3 and 5.
It should be noted that water quality can vary in a region or be strongly influenced by intake
design. The importance of having data specific to a site cannot be understated, as highlighted
by the Sohar case study in Oman. Sohar’s SWRO plant based its design for a direct MF
membrane pretreatment system on blooms with a maximum duration of one month once a
year and on process data (e.g. MF flux) from a similar plant operating for 5 years located on
the Duqum Coast of Oman. Instead of one month, the plant encountered a prolonged bloom
of six months during plant commissioning and extremely poor water quality exacerbated by
its shallow lagoon intake system. The plant experienced a multitude of issues including
clogging of the intake screen, partially clogged self-cleaning strainers, high TMP on the MF
system, biofouling of the SWRO, and ultimately loss of production and supply for
downstream industrial users. Short-term mitigation measures to maintain supply included
lowering MF flux and renting containerized MF-RO plants to compensate for the applied
drop in production. A one-year feedwater characterization study for design was
recommended following their experience.
Pilot plant trials are often employed in addition to, or instead of, a seawater quality
assessment study and are a useful tool offering many benefits for process design
optimization, leading to successful long term operation of the full scale plant. Ideally, MF/UF
can be operated at high flux and without the addition of coagulant, thereby avoiding capital
and operating costs associated with chemical storage, residual handling and treatment for
coagulated solids. In practice, MF/UF cannot always achieve high fluxes during challenging
water quality conditions such as HABs or storms where the solids and/or organic feedwater
load increase. Piloting can therefore assist in determining whether solids removal by
clarification or flotation is required prior to conventional or membrane pretreatment. Should
no bloom occur during piloting, plants can be challenged through the addition of cultivated
algae to the raw water (see section 11.5). Process parameters can also be optimized such as
the fluxes that can be achieved on MF/UF with or without coagulation, efficacy of
mechanical (hydraulic backwash, air scour) cleaning frequency and duration of CEB or CIP
to maintain production targets (Sohar, Barka 1, Antofagasta, Jacobahaven). This information
is often critical in preventing over-capitalization of pretreatment facilities if coagulation is
incorporated for MF/UF pretreatment or upstream clarification or flotation. Moreover, these
facilities may only be required for a short period each year. Alternatively, pretreatment
requirements may be underestimated or key process parameters overly ambitious, potentially
leading to loss of production and severe disruptions to drinking or industrial water supply.
337
Four of the case studies directly relate to piloting pretreatment options and process
optimization during algal blooms; Abengoa, Jacobahaven, Barcelona, and the akvola pilots.
Pilot or lab studies were also used to support pretreatment plant design for many of the full-
scale plants reported in this paper. Sohar’s rental containerized plant, following significant
HAB events, acted as a pilot plant to determine long-term mitigation strategies so that the
existing plant could be redesigned to operate through such prolonged algal blooms.
11.1.2.1.1 Water quality parameters for HAB monitoring in design and operation
Methods to measure AOM, its fouling constituents, and their impact on membrane fouling
potential (organic, particulate and/or biofouling) are important to detect blooms and assess
pretreatment efficiency. Monitoring efforts also need to continue following the collapse of an
algal bloom. The succession of bacterial species which can thrive on decaying AOM may
release organic matter extracellularly including TEP which can contribute to fouling.
Conventional water quality parameters typically included in monitoring programs to provide
an indication of the increase in organics, solids, and fouling potential include: TOC, DOC,
TSS, turbidity, SDI, and DO. Monitoring of the same parameters for process control during
plant operation as during the design phase provides continuity so operational data can be
compared with baseline data.
Although, the aforementioned parameters are not specific to algal blooms, changes may
indicate their presence, e.g. increase in TSS (Abengoa pilot plant), or indirect impacts from
HABs such as low DO following decomposition of a dense bloom (La Chimba plant).
Despite the well-known limitations of the SDI (Schippers and Verdouw 1980; Kremen and
Tanner 1998; Boerlage et al. 2000; Boerlage 2008), it has proven useful in detecting algal
blooms at the intake compared to other parameters including turbidity and chlorophyll-a
(determined via fluorescence). Elevated SDI at the intake corresponded to algal bloom events
as seen at Fujairah 2, Barka 1, Sohar and Gas Atacama plants. Care should be taken however,
in interpreting results. Often the SDI was measured at intervals not recommended by the
ASTM standard, e.g. SDI3 and even SDI1. The SDI test was not designed to measure high
fouling feedwater such as algal-laden seawater nor for UF permeate where UF have smaller
pores than that of the SDI membrane. SDI results will underestimate the fouling potential of
feedwater during a bloom as it does not capture small particles responsible for fouling
including TEP precursors and the SDI is not linear with particle concentration. Moreover,
when assessing process performance, SDI cannot be directly compared for different filtration
intervals, e.g. SDI5 for raw water and SDI15 after pretreatment, or when measured at different
temperatures (Boerlage 2008).
Measuring TOC to detect AOM in the source seawater and for process control is generally
unreliable. TOC (and DOC) measure bulk organic matter and therefore provide no
information as to the composition or concentration of potential AOM foulants produced
during an algal bloom. While TOC increases in the source seawater were found at Fujairah 2,
Sohar, and Tampa Bay, this is not always true. Measuring TOC removal to assess
pretreatment efficiency processes and as a process trigger is also inaccurate due to the
difficulties in measuring low-level TOC residuals in saline process streams (as discussed in
the Fujairah 2 case study).
SDI and TOC are often interpreted in conjunction with other parameters which directly
identify the presence of algal species in the source water through algal cell identification and
enumeration or an increase in algal productivity or advance warning through remote sensing.
Algal counts were reported at 10 of the plants, but the dominant species were not always
identified. Although cell counting can be automated using new biosensors if conducted on
338
discrete samples at external laboratories, this can result in prohibitively long turnaround
times and therefore cannot be used as an alert to trigger process adjustment, as discussed for
Fujairah 2. Identification of the algal species, size, and toxicity is important for plant
operation and process control, but must be done on discrete samples by experienced
personnel. Published and online taxonomic guides listed in the Manual can be of great value,
but trained personnel are needed to insure consistency through time.
Chlorophyll-a measurements were reported at 6 plants with La Chimba obtaining
chlorophyll-a data from satellite images provided by NPOES and MODIS satellites. Results
from chlorophyll-a data using fluorescence measurements at the seawater intake are not
always a reliable indicator of the severity of a bloom nor are the instruments easy to
maintain. The relationship between chlorophyll-a fluorescence and cell biomass is not
constant across all phytoplankton species, nutritional conditions, and times of sampling.
Some species have a low content of chlorophyll despite their size e.g. Noctiluca scintillans
which ranges up to 2 mm in size. Other factors influencing chlorophyll-a include nutrients
and light history, with limitations often resulting in lower chlorophyll-a content than the
same cells under more favorable conditions. This may explain why chlorophyll-a readings
were very low at Barka 1 (species not identified), yet operational issues were observed, while
Tampa Bay (Ceratium furca and Phaocystis) and the Jacobahaven plant (Phaeocystes,
Chaetoceros) reported much higher chlorophyll-a values at times of operational difficulties.
Similarly, Fujairah 2 reported no particular trends in chlorophyll-a concentrations during a
bloom.
Satellite-derived chlorophyll-a in conjunction with information on prevailing currents has
been used as an early warning system by desalination plants to track algal blooms
approaching plant intakes. La Chimba in Chile used the Bricker et al. (Bricker et al. 2003)
eutrophication status classification based on chlorophyll-a levels from satellite images to
assess the risk for algal blooms during the summer months. For the 2013 bloom, water in
Antofagasta Bay was classed as hyper-eutrophication (chlorophyll-a > 60 µg/L).
Recently, more sophisticated tests have been developed to determine constituents of AOM
which may better indicate the biofouling and particulate fouling potential of seawater and
process streams during a bloom. Villacorte (2014) demonstrated that biopolymers and TEP
can promote fouling of both pretreatment and SWRO membranes. Monitoring of biopolymer
and TEP concentrations during a bloom would therefore be informative. Biopolymers can be
determined by liquid chromatography - organic carbon detection (LC-OCD). LC-OCD
fractionates natural organic matter (NOM), primarily by size and also by ion interaction and
hydrophobic interaction. Fractions vary from larger biopolymers (> 20,000 Da) of which TEP
is a component, to low molecular weight compounds (< 350 Da). In a LC-OCD
chromatogram, a spike in biopolymers is observed during an algal bloom. Two methods have
been developed to determine fractions of TEP in seawater (Villacorte 2014). TEP0.4µm
measures larger TEP (>0.4 µm) while TEP10kDa includes both TEP and most of the smaller
TEP precursors. TEP precursors dominate AOM during a bloom, therefore, applying both
tests allow the differences in pretreatment removal efficiency for these algal-derived foulants
to be distinguished (Villacorte 2014); however, the degree of difficulty and cost in
determining them is correspondingly higher. As samples need to be sent to specialized
laboratories, delays in obtaining the results mean these parameters cannot be employed to
alert a plant of a bloom or to adjust process parameters during plant operation. Consequently,
only a few plants used these tests. The Barcelona and Jacobahaven plants both used LC-OCD
to examine the performance of pretreatment steps to remove biopolymers. TEP0.4µm was also
monitored for three years at the Jacobahaven plant and a correlation between TEP and
fouling rates in the UF pretreatment system was observed.
339
A method to determine assimilable organic carbon (AOC) based on luminescence using

Vibrio harveyi has also been developed to indicate the biofouling potential of the source
seawater and process streams (Weinrich et al. 2011). The saline AOC test was trialed at the
Tampa Bay plant to monitor pretreatment. High levels of AOC were measured in the source
seawater during a bloom that further increased following chlorine dioxide prechlorination
which may be due to oxidation of organic matter. The AOC was greatly reduced by
bioactivity in the subsequent sand filtration step (LeChevallier 2014; Tampa Bay Case
Study).
Finally, there are improved methods to measure particulate fouling which can be used during
a bloom. The Modified Fouling Indices (MFI), developed to address the limitations of the
SDI, comprise the MFI-0.45 (an ASTM standard) and the MFI-UF using ultrafiltration
membranes (Schippers and Verdouw 1980; Boerlage et al. 2000; Salinas-Rodriguez 2011).
The MFI-UF with smaller pore size membranes can measure UF permeate unlike the SDI and
MFI-0.45. Both MFI tests are normalized to reference temperature, area and pressure values
and can measure high fouling feedwater. Therefore, MFI can be compared and pretreatment
efficiencies can be determined. When using a 10 kDa membrane in the MFI-UF test, a high
correlation was found between MFI-UF and TEP10kDa, measurements (Villacorte 2014). This
indicates the MFI-UF could be used to investigate the fouling potential of feedwater
containing the smaller high fouling TEP precursors across a plant during a bloom. Of the
twelve case studies, only the Jacobahaven plant employed the MFI-UF where it was used to
measure pretreatment efficiency.
11.1.2.2 Intake depth and location
Careful selection of intake type, depth, and location, is often considered the first defense in
preventing the entrainment of algal blooms into a plant coupled to a comprehensive
investigation of seawater quality and site conditions as discussed above.
Subsurface intakes are one option to reduce the ingress of algal cells and associated AOM
into a plant; however, subsurface intakes are not feasible for all sites and were not covered in
the case studies, though chapter 6 covers intake designs in detail. Most large scale SWRO
plants use an open ocean (or surface) intake. Abstracting water at depth is often promoted as
a means to prevent entrainment of algae into desalination plants with open intakes. Indeed,
the deep water (20 m water depth) intake option selected for the Gold Coast plant appears to
be successful in providing good water quality and preventing the ingress of dense floating
mats of Trichodesmium erythraeum, the most frequent algal bloom observed in that region.
Limited ingress of Colpomenia, a macroalgae typically attached to rocks, has occurred. This
is most likely not as a result of a bloom but through dislodgment with wind and waves
breaking it into flakes which may then have become entrained into the intake. The success of
deep water intakes depends on the bloom-forming species. Some are motile or display diel
vertical migration so that they move within the water column and are not found solely within
the surface, mixed layer. Indeed, they can be at the surface during the daytime, and at 10 or
even 20 m depths at night. Moreover, the distribution of AOM may not reflect the
distribution of algal cells in the water column, as AOM can be extracellular and detrital in
form, sinking to the seabed or rising to the surface. Hence, a deep-water intake can assist in
limiting ingress of some algal blooms, but is no guarantee in preventing all algae and AOM
as seen in various SWRO desalination plants in Chile.
The La Chimba and Gas Atacama plants, located 50 km apart, abstract seawater from the
sheltered bays of Moreno and Mejillones, respectively, using a deep (25 m) and shallow (5m)
intake. Both plants entrain algae into the intake. Abengoa’s pilot plant employed the La
Chimba intake and measured a significant increase in algal concentration, suspended solids
340
and organics during bloom events. TEP was found to increase five-fold to 3500 µg x-eq/L
and suspended solids increased to 40 mg/L on average and up to 76 mg/L. The deep-water
intake did not therefore prevent algal bloom and AOM from being entrained into the intake.
Moreover, thermoclines in the Moreno Bay led to SRB and the generation of H2S issues at
the La Chimba plant as discussed above. Further up the coast, the shallow intake of the Gas
Atacama is less likely to suffer from hypoxic events as at La Chimba, but directly from algae
and AOM.
11.1.2.3 Chlorination
Chlorine lyses algal and bacterial cells and oxidizes high molecular weight organics into
biodegradable low molecular weight compounds which are more easily metabolized by
biofouling bacteria. Some of the organic compounds released or formed may adsorb onto
MF/UF pretreatment membranes or pass through to the RO system, increasing fouling and
biofouling of the RO membranes. Hence, the SWRO industry moved from continuous to
shock dosing of chlorine. During an algal bloom, chlorination at the intake could be
suspended to limit lysis of algal cells and release of fouling AOM and toxins (if it is a toxin-
producing bloom).
The small Jacobahaven demonstration plant (360 m3/d) avoided chlorination to prevent the
formation of chlorination byproducts by pigging their intake lines (once every 2 weeks),
which would be infeasible for large full scale plants. Yet, it still suffered from AOM fouling
of the UF membranes. Although, the intake at the Fujairah 2 hybrid plant was switched from
continuous to shock dosing during studies of HABs at the SWRO plant, it is difficult to show
direct data from their case study to demonstrate that the plant reduced algal lysis. Shuwaikh,
Gas Atacama and Gold Cost practice shock chlorination.
Some plants continue to practice continuous chlorination such as Barka 1 and Sohar. Sohar
intends to change to shock chlorination, but has issues due to other co-located industries that
require chlorinated seawater. La Chimba can continuously chlorinate, but doesn’t use the
system. At Tampa Bay, continuous chlorination is practiced at two locations; chlorine
dioxide at the intake and hypochlorite prior to the coagulation mixing basins. A 65% increase
in the AOC during algal blooms was found following disinfection with chlorine dioxide
compared to the raw water at the plant intake. Increases in AOC have been observed
following disinfection with chlorine dioxide (Haas et al. 2015). Increases in AOC may lead to
increased biofouling potential and may be directly linked to HABs.
11.1.2.4 Dissolved air flotation
DAF is generally thought to be one of the leading approaches to mitigate algal-related issues
in downstream processes. However, DAF can be an expensive addition to a plant if it is only
required to remove algae and blooms are not severe or frequent. A disadvantage of operating
DAF in bypass mode is that it needs to be brought online prior to the bloom, which is often
difficult to time correctly. The advantage of DAF is that the algal cells will be gently lifted by
air bubbles, minimizing damage to cells or release of AOM. Successful performance of DAF
is contingent on optimized coagulation flocculation to promote attachment of algal cells to
the micro-bubbles generated in DAF. In general, cell removal is higher for more dense
blooms.
DAF was utilized in five of the twelve case studies, Shuwaikh, Fujairah 2, Gas Atacama,
Barcelona and akvola. Two of these cases (Shuwaikh and Fujairah 2) require DAF for other
problems besides HABs, such as very high turbidity, suspended solids, and hydrocarbons.
At Gas Atacama, a conventional DAF+DMF system ran side-by-side with a UF system. The
DAF+DMF encountered significant operational issues: high coagulant dose (25 ppm),
341
increased SWRO membrane cleaning and cartridge filter replacement, and eventually
premature SWRO membrane replacement. The alternative system that employed UF did not
require cartridge filter replacement or SWRO membrane cleaning even when challenged by
HAB events; however, it should be noted that the DAF+DMF plant is approximately 20 years
old and employs older designs for DAF and SWRO.
The Shuwaikh plant uses high rate Leopold Clari- DAF before autostrainers and UF
(pressurized inside out). The DAF uses ferric chloride and sulfuric acid. The system
maintained full capacity during a bloom after optimization of coagulant and acid dose. DAF
pilot studies conducted on Antofagasta Bay reported in the Shuwaikh case study showed
greater than 95% removal for varying species of algae.
At Fujairah 2, Spidflow® DAF is utilized in line with DMF. Ferric chloride average dose rate
during blooms in 2011-2012 was < 20 ppm and in 2013 <15 ppm. Fujairah 2 has provision in
the DAF for acid dosing, but this was only used in commissioning and has provision for
polymer dosing, but this has never been used. A DAF pilot plant operated throughout the
2008/2009 Cochlodinium bloom and the full-scale plant operated through a 2013 bloom.
The Barcelona pilot operated a parallel trial of DAF followed by (i) conventional DMF or (ii)
pressurized outside-in UF. The combination of DAF followed by UF gave superior removal
of biopolymers (41%) compared to only 18% for DAF and conventional DMF pretreatment.
The ferric chloride coagulant dose for the Barcelona DAF pilot study was low being 0-
6 mg/L and acid and coagulant aid were not found to be crucial in optimizing DAF
performance. Algal cell removal averaged at 75% and up to 87%.
11.1.2.5 Microfiltration/Ultrafiltration
Microfiltration and ultrafiltration were commonly used for treatment of algal rich water, with
eight of the twelve plants employing advanced membrane pretreatment: Sohar, Barka 1,
Shuwaikh, Gas Atacama, Abengoa pilot, Jacobahaven, Barcelona and akvola.
Due to the pore size of MF/UF membranes, cell removal is generally 99-100% and cells only
pass the MF/UF through broken fibers (Dixon et al. 2011). Most of the plants employed
ultrafiltration which will achieve a higher removal of AOM and particulates due to their
smaller membrane pores (or MWCO rating) than microfiltration with larger pores. As fouling
biopolymers and TEP precursors range upwards of 20,000 Da in molecular weight and a few
nm in size, respectively, ultrafiltration membranes (with a typical MWCO of 100,000 –
150,000 Da) will not remove them completely. Coagulation, albeit at a lower dose than for
DMF, is therefore practiced to remove dissolved AOM, improve removal of biopolymers and
TEP, and the filterability of the cake deposited on UF membranes. Several of these plants
used DAF in combination with UF in order to reduce the suspended solids and organics
loading on the membranes: Shuwaikh, Barcelona and akvola. The akvola pilot plant is a
single unit combined DAF+UF system employing ceramic membranes.
Several operating techniques to prevent hydraulically irreversible fouling are available to
allow UF plants to continue operations during an algal bloom. The following were used in the
case studies: reducing flux (4 plants), recirculating flow (1 plant), use of coagulation
involving optimizing the dose or switching on coagulation in response to blooms (3 plants),
decreasing backwashing intervals (2 plants), increased CEB frequency (5 plants) and CIP (3
plants). The specific actions taken at each plant to maintain membrane performance are
described in more detail in the following paragraphs.
342
At the Sohar plant which employs direct MF filtration, immediate short-term mitigation
measures included reducing the MF flux from 100 L/m2h maximum to 60-73 L/m2h. The
operators also reduced backwash cycle time, increased CEB frequency, and optimized
cleaning strategies using chlorine and caustic (twice/day) and acid cleans (every 3 days).
Coagulation will be incorporated into pretreatment as part of long term bloom mitigation
measures to improve removal of AOM.
At Barka 1, the plant experienced significant UF fouling yielding a permeability drop of
100 L/m2hbar during light to moderate blooms. The plant maintained flux without
coagulation by reducing the CEB design interval from 18 hours to 12-15 hours to delay the
CIP to greater than once in 180 days. In the case of a more significant bloom, the plant could
reduce flux or turn on coagulation.
The Gas Atacama UF plant maintained operation during blooms as coagulation is employed
at the plant (maximum 2 ppm).
The Abengoa pilot plant experienced a rapid increase in TMP during their bloom and
responded first by reducing the CEB interval from 30 to 7-8 hours, using hypochlorite and
caustic for CEB cleaning, with the pH increased from 9 (normal pH) to 12 during blooms.
Later, mitigation included reducing flux (16%), which allowed CEB to be conducted every
26 hours. Unlike most other UF plants, no coagulation was utilized. The UF membranes were
operated with increased recirculation flow allowing scouring of the membrane. While the UF
plant usually achieves 96% recovery, during the bloom, recovery dropped to 80% due to the
reduction in flux while maintaining backwash frequency.
Jacobahaven (a demonstration plant) allowed for maximum flexibility in operation. Rapid
fouling of UF coincided with high chlorophyll-a, TEP and algal counts. CEB decreased to
every 6-12 hours while in non-bloom conditions, CEB was typically every 24 hours, and as
infrequent as once every 2 weeks. Modification of the backwash interval gave limited
improvement and increased backwash volume to 20%. Similarly, modification of CEB
conditions and frequency was ineffective. Flux reduction did not yield a satisfactory CEB
interval. Coagulation was effective but CEB remained ineffective during coagulation and
HABs, requiring a CIP.
At Barcelona, no impacts were noted with algal cell counts up to 1.2 million/L. Both
conventional and membrane pretreatment performed well; however, UF gave better
performance with respect to SDI15, turbidity, algal cell, and biopolymer removal.
11.1.2.6 Dual media filtration/Granular media filtration
While DMF is a common pretreatment practice, only four of the case studies employed DMF
as part of their pretreatment. These were Fujairah 2, Gas Atacama, Barcelona and the Gold
Coast. While DMF has been proven as an effective HAB removal method, giving between 20
and 90% removal in several cases (Gustalli et al. 2013), in this study little cell removal data
was corroborated.
No negative impacts were observed in the gravity DMF at the Gold Coast plant as the deep
water intake limited ingress of the most frequently occurring algal bloom. Hence, no cell
removal data were recorded. Anecdotally, no increased DMF clogging was experienced
during blooms with ferric sulfate (13 mg/L) coagulation and DMF. Similarly, no impacts
were reported at Fujairah 2 and the Barcelona pilot plant, which employ additional solids
removal through an upstream DAF.
At Barcelona cell removal with DMF was 74% and removal of biopolymers was 18% for
DAF and DMF. As Barcelona was a side-by-side pilot plant, it showed DAF+UF cell
343
removal (100%) was superior to DAF+DMF. Likewise, biopolymer removal was 41% versus
18%.
11.1.3 Summary
The twelve case studies present an array of mitigation strategies for operating in algal-rich
seawater from around the globe including deep water intakes, DAF, conventional and
advanced pretreatment with or without coagulation. A myriad of impacts, algal water quality
monitoring and detection techniques and management practices are described in the case
studies. In many cases, the observations collected from plant operations as part of the Manual
development validate some of the key perceptions held by our industry about how to manage
operations during HABs. For example:
• Careful, informed site analysis with a focus on the potential for different types of
algal blooms can be a cost-effective strategy that minimizes future pretreatment
needs;
• Long-term monitoring programs to characterize seawater quality should be conducted
at the plant intake site. This should include a review of historical records examining
the frequency and severity of algal blooms (if available);
• Piloting at site is recommended with challenge testing using cultivated algae if no
blooms occur;
• Conventional parameters such as the SDI have proven useful in detecting blooms at
an intake. Care needs to be taken with the SDI as it will underestimate the fouling
potential of feedwater, especially during algal blooms;
• Parameters have been developed to monitor AOM constituents and the increase in
(bio)fouling potential of seawater during a bloom e.g. biopolymer concentration by
LC-OCD, TEP10kDa, MFI-UF using a 10 kDa test membrane and the AOC
luminescence test using Vibrio harveyi. These tests can be used to assess and optimize
pretreatment during a bloom. While promising, the degree of difficulty and cost in
determining them is correspondingly higher. As yet, these parameters cannot be
employed to alert a plant to a bloom or to adjust process parameters during plant
operation;
• Deep-water open ocean intakes may be successful in avoiding some bloom-forming
species, but some species are motile or display diel vertical migration so that they
move 10 m or more daily within the water column. Moreover, the distribution of
AOM may not reflect the distribution of algal cells in the water column. Deep water
intakes are thus not guaranteed to prevent the ingress of algae and AOM into a plant;
• Avoiding chlorination-dechlorination during a bloom will help to prevent downstream
fouling as fouling AOM will be retained intercellularly along with toxins and less
assimilable organic carbon is generated through oxidation of organic matter;
• Coagulation is important for all three pretreatments (DAF, UF, and DMF) and acts to
remove AOM more effectively than these pretreatment processes alone. Less AOM in
the pretreated water will help to alleviate SWRO fouling;
• DAF is a good choice for cell removal in areas predicted to experience heavy blooms,
as it removes cells by floatation, minimizing cell rupture;
• DMF or UF alone can accomplish cell removal and some dissolved AOM (if
coagulation is used for UF) for light algal blooms, but using a DAF upstream is
prudent for heavy blooms; and
• The combination of DAF and UF is increasingly being applied in the industry, when
applied in bypass mode, care needs to be taken to bring the DAF online in time.
344
11.1.4. References
Anderson, D. M., Cembella, A. D., and Hallegraeff, G. M. 2012. Progress in understanding
harmful algal blooms: paradigm shifts and new technologies for research, monitoring,
and management. Annual Review of Marine Science 4, 143-176.
Berman, T. and Holenberg, M. 2005. Don't fall foul of biofilm through high TEP levels.
Filtration and Separation 42(4), 30-32.
Boerlage, S. F. E. 2008. Understanding the Silt Density Index and Modified Fouling Indices
(MFI0.45 and MFI-UF). Desalination and Water Reuse Quarterly May –June, 12-21.
Boerlage, S. F. E, Kennedy, M., Aniye, M., Abogrean, E., El-Hodali, D., Tarawneh, Z., and
Schippers, J. 2000. Modified Fouling Index - ultrafiltration to compare pretreatment
processes of reverse osmosis feedwater. Desalination 131, 201-214.
Boerlage, S. F. E. and Nada N. 2014. Algal toxin removal in seawater desalination processes.
In: Proceedings of European Desalination Society, Cyprus.
Bricker, S. B., Ferreira, J. G., and Simas, T. 2003. An integrated methodology for
assessment of estuarine trophic status. Ecological Modelling 169(1), 39–60.
Dixon, M. B., Richard, Y., Ho, L., Chow, C. W. K., O’Neill, B. K., and Newcombe, G. 2011.
A coagulation - powdered activated carbon - ultrafiltration multiple barrier approach for
removing toxins from two Australian cyanobacterial blooms. Journal of Hazardous
Materials 186(2-3), 1553-1559.
Gustalli, A., Simon, F. X., Penru, Y., de Kirchove, A, Llorens, J., and Baig, S. 2013.
Comparison of DMF and UF pretreatments for particulate material and dissolved
organic matter removal in SWRO desalination. Desalination 322, 144-150.
Haas, C. N., LeChevallier, M. W., and Weinrich, L. A. 2015. Application of the
Bioluminescent Saltwater Assimilable Organic Carbon Test as a Tool for Identifying
and Reducing Reverse-Osmosis Membrane Fouling in Desalination. Alexandria: Water
Environment & Reuse Foundation. Project 11-07.
Kremen, S. S., and Tanner, M. 1998. Silt density indices (SDI), percent plugging factor (%
PF): their relation to actual foulant deposition. Desalination 119(1), 259-262.
LeChevallier M. W. 2014. Measurement of biostability and impacts on water treatment in the
US. In: Microbial Growth in Drinking-Water Supplies Problems, Causes, Control and
Research Needs. IWA Publishing.
Li, S., Winters, H., Villacorte, L. O., Ekowati Y., Abdul-Hamid. E., Kennedy, M. D., and
Amy, G. L. 2015. Compositional similarities and differences between Transparent
Exopolymer Particles (TEP) from two marine bacteria and two marine algae:
significance to surface biofouling. Marine Chemistry 174, 131–140.
Myklestad, S. M. 1995. Release of extracellular products by phytoplankton with special
NHMRC/NRMMC. 2011. Australian Drinking Water Guidelines Paper 6: National Water
Quality Management Strategy. Canberra: National Health and Medical Research
Council / National Resource Management Ministerial Council.
Salinas-Rodriguez, S. G. 2011. Particulate and organic matter fouling of SWRO systems:
characterization, modelling and applications. Doctoral dissertation, UNESCO-
IHE/TUDelft, Delft.
345
Schippers, J. C. and Verdouw, J. 1980. The Modified Fouling Index, a method of

determining the fouling characteristics of water. Desalination 32, 137-148.
Villacorte, L. O. 2014. Algal blooms and membrane-based desalination technology. CRC
Press/Balkema, Leiden: 200 p.
Weinrich, L. A., Schneider, O. D., and LeChevallier, M. W. 2011. Bioluminescence-based
method for measuring assimilable organic carbon in pretreatment water for reverse
osmosis membrane desalination. Applied and Environmental Microbiology 77, 1148–
1150.
WHO. 2011. Guidelines for Drinking Water Quality, Fouth Edition. Geneva: World Health
Organization, 564 p.
346
11.2 FUJAIRAH 2, UNITED ARAB EMIRATES – EFFECTS OF

HARMFUL ALGAL BLOOMS ON PLANT OPERATIONS IN 2008
AND 2013
Herve Faujour1, Cyril de Vomecourt1, and Jérôme Leparc 2

1
2
Veolia Recherche & Innovation, Maisons Laffitte, France
Figure 11.2.1. Location and aerial view of the Fujairah 2 hybrid plant in UAE. Photos: Veolia.
347
Table 11.2.1 Overview of Fujairah 2 SWRO plant.

Plant/Project Name
Location Fujairah (Qidfa)
Primary product water use Municipal
Desalination Technology Hybrid SWRO/MED, case study only covers SWRO as HAB did
not affect MED plant
Total Production Capacity 136,000 m3/d for the SWRO (23%)
(m3/d) 453,000m3/d for the MED (77%)
SWRO recovery (%) First pass approximately 41% Overall plant recovery 39%
Commissioning date Official start of operation: Oct 2010
Intake
Feedwater source Oman Sea
Intake type Open submerged intake – 3 risers
Intake description 500 – 600 m off shore, 10 m depth
Intake screening Static and rotating drum screens
(shock) chlorination Hypochlorite:
Strategy, dose rate continuous dosing from 10/2010 to 12/2012 with additional shock
dosing a few times per day from 01/2013 onwards, only shock
dosing at intake risers
Online raw water monitoring SDI, conductivity, temperature, pH, turbidity, and hydrocarbons
Discrete raw water analysis SDI, TOC, algal counts during presumed HABs, chlorophyll
relevant to HABs
Pretreatment
Process description Spidflow® DAF, Dual Media Gravity Filtration,
5 to 10 µm cartridge filtration
Chemical dosing FeCl3 Average dose rate in 2011-2012 was <20 ppm
and in 2013 <15 ppm
Provision in the DAF for acid dosing (but only used in
commissioning) and for polymer dosing (never used)
Feedwater design parameters Feedwater during blooms

Temperature range (˚C) 22 to 33 22 to 33
Salinity range (TDS mg/L) < 40,500
Total Suspended Solids (mg/L) 4 to 6 10 to 30
SDI (%/min) SDI3 = 18 to 22 SDI3 = 26 to 30
Turbidity (NTU) 0.2 to 0.3 0.5 to 1.0
Organic Matter TOC < 1 TOC = 1.0 to 1.5
Algal cell count (cells/L) 2 to 4x105 Up to 106
Algal species (if known) N/A N/A
Chlorophyll-a or total N/A N/A
chlorophyll (µg/L)
Additional water quality N/A N/A
parameters for design or
observed during algal bloom
Desalination Design During bloom conditions
Filter rates DAF up to 30 Same as normal design
(DAF/DMF etc.) m/h
UF flux (L/m2h) N/A N/A
RO flux (L/m2h) Confidential Same as normal design
N/A - not available
348
11.2.1 Introduction
High demographic growth in the United Arab Emirates (UAE) in recent years has resulted in
a higher demand for drinking water and power. In order to satisfy the growing demand, the
Fujairah 2 Independent Water and Power Plant (IWPP) was launched by the Emirate of Abu
Dhabi for completion during 2010. The location of this new plant was immediately next to
the existing Fujairah 1 IWPP in the Emirate of Fujairah on the Gulf of Oman (Figure 11.2.1).
Until recently, the Fujairah 2 IWPP was the largest hybrid desalination plant in the world
with a combined net design capacity of 590,000 m3/d linked to a power plant with a
2,000 MW/d net design capacity. Desalinated water is produced by a Seawater Reverse
Osmosis (SWRO) plant and a Multi-Effect Distillation plant (MED).
In recent years, the unwelcome phenomenon of harmful algal blooms (HABs) have been
experienced in the Gulf of Oman along the coasts of the UAE and Oman. Impacts had been
devastating, particularly during a 2008/2009 HAB event which caused the shutdown of many
SWRO plants due to the inability of conventional pretreatment processes to deal with the
high concentrations of algae, and an associated increase in suspended solids and dissolved
organics. The Fujairah pilot plant, which incorporated dissolved air flotation (DAF) as a
pretreatment strategy, continued to operate during that bloom.
Effective pretreatment process performance is paramount for any SWRO plant, as feedwater
quality needs to meet the generally stringent O&M contractual specification. Veolia meets
the very stringent SWRO plant availability requirement of Fujairah 2 of 98%. In order to
achieve such availability with challenging raw water quality, the Veolia SPIDFLOW®
Dissolved Air Flotation process has been included in the pretreatment with the aim of
mitigating algal blooms/red tide events in particular and ensuring a superior and highly
consistent water quality.
11.2.2 Plant overview
11.2.2.1 SWRO desalination plant
A shared seawater intake supplies seawater to the power plant as well as the MED thermal
and SWRO desalination plants with a common remineralization section where MED distillate
and SWRO permeate are blended and remineralized. The Fujairah 2 RO plant process
configuration comprises the SPIDFLOW® DAF units (16) with coagulation/flocculation,
rapid gravity dual media filters (12), cartridge filtration (16), first pass Reverse Osmosis (10)
and partial second pass Reverse Osmosis (2) with the second pass to achieve boron removal
(Figure 11.2.2).
During the first five years of operation, two distinctly different intake chlorination strategies
were employed. These had potential effects on general treatment process performance and
water quality. During the first 27 months of operation, continuous intake chlorination was
practiced plus an extra “shock-chlorination” every 6 – 8 hours, which could reach up to
2.5 mg/L. Since January 2013, shock or intermediate intake chlorination has been practiced at
various frequencies and concentrations. Residual Cl2 was monitored upstream of the RO,
and when detected, was neutralized with sodium bisulfite (SBS) injection in the suction of
the HP pump.
Provision was made for polymer and acid dosing to reach an optimum pH for coagulation,
but experience from the five previous years showed that the efficiency of FeCl3 alone is
sufficient.
349
Figure 11.2.2. Process flow schematic for the SWRO desalination plant at Fujairah 2.
11.2.3 Inlet Water Monitoring

11.2.3.1 Summary of main feedwater quality excursion events since 2008
Poor seawater quality events during pilot plant operation (from 2008) and full scale plant
operation (from 2010) are summarized in Table 11.2.2, highlighting the main parameters that
characterized the event, along with the corresponding origin of the degradation. Two of these
events were due to algal blooms (2008 and March 2011). The 2008 event was caused by
Cochlodinium polykrikoides; the causative organism for the 2013 event is not known.
Despite degraded inlet seawater quality, the DAF was always able to maintain good water
quality downstream of the dual media filter (DMF). The major red tide of 2008 was treated
by the pilot plant only, without affecting the SDI15 after DMF (with higher FeCl3 doses).
Once the full-scale plant was operational, the DAF pilot was kept in operation in parallel to
the main plant to investigate adjusting operating parameters on the pilot before applying them
to the main plant. This approach confirmed that the results on the DAF were highly
representative of the large-scale plant, which gives confidence that even very high intensity
algal blooms similar to that of 2008 will not affect the RO performance when the inlet
pretreatment includes a Spidflow® DAF.
The nature of the feedwater quality excursion is difficult to characterize in detail, and work is
underway in collaboration with local research institutes such as MEDRC in Muscat to have
more parameters analyzed when degradation is observed. This analysis will also be applied to
normal operation to create background data for comparison to the excursions.
The nature of the poor water quality events was also interpreted in conjunction with remote
satellite observations and on-site photographs (Figure 11.2.3), and through the description of
the nature of the effects on the coastal industry. For instance, on the 18th of February 2009,
the Khaleej Times quoted the following: “Red Tide Plaguing Fujairah, will continue to
affect area: The red tide that has been plaguing Fujairah’s coastline since November last
year will continue to affect the area for some more months”.
350
Table 11.2.2. Inventory of main feedwater quality excursion events at Fujairah 2 intake.
Ref Value 2008 03/11 03/12 03/13 07/13 09/13 02/14 10/14
SDI3 (inlet) 18-22 26-30 26.4 27.3 26-28 22 24.4 30 24-29
SDI15 (before 3-3.5 5 4.8 5 5 No 4.3 4.5 < 3.5
RO) Impact
Turbidity 0.2-0.3 0.5-0.8 0.6 ave 0.3 ave 0.2ave 2.9max 0.5 ave 0.3 ave
(NTU) 6.5 max 29 max 1.5 peak 13 max
TSS (ppm) 4-6 10-30 N/A N/A N/A +10% in N/A 20 1.5-9
HC
Temp (oC) +/- 2C N/A normal max more No Start to Normal Normal
variability around Impact increase
(D/N) 16/3
DMF clogging N/A N/A x2 x3 N/A No x2 N/A N/A
rate Impact
DMF BW freq 40-54 40 40 40 ave 26 No 55 ave N/A N/A
(h) 17h min Impact 42 min
Shore obs. N/A N/A red jellyfish brown oil patch jellyfish N/A N/A
patches shore
Satellite N/A N/A N/A N/A visible N/A visible N/A N/A
picture
Algae count 2-4x105 N/A N/A 3-4x105 105-106 N/A N/A N/A N/A
(cells/L)
TEP 20- N/A N/A N/A N/A N/A N/A N/A N/A
(>0.45µm) 40µg/L
Chl-a (µg/L) 0.02- N/A N/A N/A N/A N/A N/A N/A N/A
0.06
TOC (ppm) ~1 N/A N/A N/A 1-1.5 N/A N/A N/A N/A
Interpretation algal algal jelly fish not oil spill jelly fish high not
bloom bloom known turbidity known
Note: ave, max and min refers to average, maximum and minimum values, respectively
On the 8th of March 2013, the Khaleej

Times again reported another HAB in
the same area (Kalba, 30 km south of
Fujairah). “The directorate [of SEWA,
the Sharjah Electricity and Water
Authority] had decided to stop the
operation of the desalination plant for
the safety of the residents in the region.
He explained that SEWA regularly
tests desalinated water to ensure that it
meets required specifications. The
operation will resume immediately
after the red tides recede in the sea”.
During the 2013 bloom period, a strong
Figure 11.2.3. Red tide of Cochlodinium polykrikoides on
the coast of Fujairah during 2008/2009. Photo: D.smell was present both outside and
Anderson. inside the plant. Water discoloration
was clearly visible at the sea surface,
especially close to the coast where the waves were yellow-orange. Fujairah 2 SWRO was not
affected, as reported in the Water Desalination Report Volume 49 – Number 10, dated from
the 11th of March 2013 in an interview with Erich Koenig, plant manager: “I am not sure if
we can call it red tide, because the visual signs were more yellowish green, although satellite
351
images did also show red tide patches close to Oman”, he said, adding: “Our process was
stable with significant adjustment to our ferric chloride dose in order to satisfy coagulant
demand and to ensure the effectiveness of our signature high-speed DAF system.”
11.2.3.2 Overview of SDI and turbidity of seawater
The trends in online SDI3 and turbidity from January 2012 are shown in Figure 11.2.4. Spikes
in both parameters were observed during the March 2013 algal bloom. SDI3 and turbidity are
not sufficient to entirely characterize all potential pollution events, but they remain the easiest
parameters to monitor. Operational feedback for the monitoring of other parameters can be
summarized as:
- TOC: TOC was not a reliable indicator to trigger adjustment of operational/process
parameters. Interpretation of trends in TOC were also difficult due to the issues
associated with accurately measuring TOC at low residual levels of 1 ppm with high
salinity seawater.
- Temperature: This was an easy parameter to monitor, but a sharp change in
temperature may not always signal that operational disturbance is expected.
- Chlorophyll: The laboratory analyzer did not show any particular trends, plus the
low level of algal pigment makes local monitoring difficult. Furthermore, a HAB is
not the only water quality excursion that needs a robust pretreatment, so adjustment of
operating parameters may not be related only to chlorophyll.
- Algal count: Samples were sent to the USA several times for more accurate
characterization, but the turn-around time was long, so this was used only for
diagnosis rather than as an alert to trigger operational changes. Increase in the algal
count was not sufficient to be relevant in March 2013.
- Satellite data: Imagery was used as a way to help the interpretation of the event
rather than to adjust operating parameters;
- Shore observation: Some observations from shore may not be relevant to the plant
intake due to the submergence of the intake, but it remains a pragmatic monitoring
approach.
- TEP: Transparent Exopolymer Particles, associated with algal blooms could be
monitored. As lab protocols are complex these were not set up locally. Long turn-
around times for European laboratories make decision making based on laboratory
data impractical.
11.2.4 Plant Performance
11.2.4.1 SDI monitoring
In combination, the pretreatment processes ensured the consistent delivery of exceptional
water quality characterized by a SDI15 between 2 and 4 for the filtered water, even during the
March 2013 algal bloom (Figure 11.2.5). A slight degradation of the RO feedwater SDI
observed in 2013. It may be related to the modification of the pre-chlorination regime or to
the optimization of the FeCl3 dose, but the increase of SDI remains quite minor, in the range
of 2.5 to 3.0, which is quite comparable to UF filtered water, showing that a robust
combination of DAF with DMF with the injection of coagulant can achieve as good a result
as UF, with the probable additional benefit of ability to take highly degraded seawater
quality.
352
6.00 30.00
5.00 25.00
Seawater Turbidity (NTU)
4.00 20.00
SDI3 Seawater
3.00 15.00
2.00 10.00
1.00 5.00
0.00 0.00
Jan-‐‑12 Jan-‐‑13 Jan-‐‑14 Jan-‐‑15 Jan-‐‑16
SEA WATER INTAKE TURBIDITY SEA WATER INTAKE SDI³
Figure 11.2.4. Online monitoring of inlet seawater turbidity and SDI3 in January, 2012.
Figure 11.2.5. SDI3 of the SWRO intake and SDI15 of the first pass RO feed in 2013.
Cartridge filtration, although not considered a treatment process, has also contributed to
achieving well-polished feedwater for the RO membranes. In terms of SDI, the cartridge
filters have generally reduced the SDI by 0.5 units.
11.2.4.2 Plant production and availability
As shown in Figure 11.2.6, the capacity of the plant was maintained at its highest
requirement regardless of the quality of the raw feedwater during six poor water quality
events experienced between 2008 and 2014. For example, during October 2014, SDI3 at the
inlet increased to 24 - 29, the cause of which was unknown, but the produced flow remained
high and even exceeded the flow corresponding to when all skids including the standby (9
duty + 1 standby) were in operation.
A 98% availability (monitored as a cumulative value in the month of October every year) is a
very stringent requirement, but over 5 years of operation, despite multiple HABs and other
events, this has always been achieved.
353
Figure 11.2.6. Water production and plant availability since 2011.
11.2.5 Conclusions
The pretreatment process combination of the Fujairah 2 SWRO Plant and optimized
operations have allowed excellent performance of the SWRO plant. The plant has
consistently met availability targets during challenging water quality events. This can be
attributed to the success of the SPIDFLOW® DAF system that has proven to be a reliable and
flexible treatment process under different water quality conditions, including moderate
HABs.
In terms of additional cost of operation, the kWh needed for the Spidflow® DAF is in the
range of 2% of the total plant consumption (excluding plant inlet and remineralization which
are not in the scope of the Veolia O&M contract). While the FeCl3 dose rate was kept high as
a precaution during the first years of operation, it has been reduced to values that are
comparable with the industry standard of SWRO plants that do not include a DAF.
RO membranes perform at their best when they are fed with excellent and consistent water
quality. The overall result led to a contract performance far above the expectation of Veolia
and the client.
354
11.3 SOHAR, OMAN – HARMFUL ALGAL BLOOM IMPACT ON

MEMBRANE PRETREATMENT: CHALLENGES AND
SOLUTIONS
Abdullah Said Al-Sadi1 and Khurram Shahid2

1
Majis Industrial Services, Sohar, Oman
2
Water Solutions International Ltd, Gatwick, England
Figure 11.3.1. Map of the location of the Sohar desalination

plant (top) and aerial view showing the combined intake channel
and outlet for power and desalination plant (bottom).
355
Table 11.3.1. Overview of the Sohar SWRO desalination plant.

Maijs SWRO Plant
Location Sohar, Oman
Primary product water use Process and potable
Desalination technology MSF and SWRO
Total production capacity 150,000 MSF and 20,000 SWRO
(m3/d)
SWRO recovery (%)
Commissioning date September 2013
Intake
Feedwater source Gulf of Oman
Intake type Shallow intake channel
Intake description Lagoon of 4 m depth
Intake screening Chain raked bar screen (30 mm spacing); Travelling band
screens (3 mm mesh size)
(Shock) chlorination Hypochlorite continuous dosing to provide a free chlorine
strategy, dose rate residual between 0.4 to 0.8 mg/L
Online raw water Conductivity, temperature, pH, turbidity, hydrocarbon
monitoring analyzer; residual chlorine, dissolved oxygen (DO)
Discrete raw water analysis TOC, algal counts
relevant to HABs
Pretreatment
Process description 100-µm self-cleaning strainers MF pressurized outside in
Feedwater design Feedwater during bloom
parameters conditions
Temperature range (˚C) N/A N/A
Salinity range N/A N/A
(TDS mg/L)
Conductivity (mS/cm) N/A N/A
Total Suspended Solids N/A N/A
(mg/L)
SDI (%/min) 5-15 SDI5 SDI5 unmeasurable;
SDI2.5 >35 100% of the time
Turbidity (NTU) N/A N/A
Organic Matter TOC <3 5.8
TOC and COD mg/L COD <20 56
Algal cell count (cells/L) N/A 5.6 million
Algal species N/A Gonyaulax polygramma
Chlorophyll-a or total N/A N/A
chlorophyll (µg/L)
2
MF flux (L/m h) <100 <76
2
RO flux (L/m h) 16.6 (SWRO) and 36.5
(BWRO second pass)
356
11.3.1 Introduction
The Oman Sohar Industrial Port Area (SIPA) is home to major petrochemical, iron and steel,
methanol, and fertilizer plants, as well as three power stations. Since its inception in 2006 it
has drawn some US$ 15 billion in investment and holds a strategic place in the future
development of Oman. As the sole water services provider at SIPA, Majis Industrial Services
(Majis) has a significant role in the economy of the Sultanate.
Majis provides seawater for cooling as well as high-quality desalinated water for process
water applications, as well as potable water and wastewater collection, treatment, and re-use
for irrigation. After the completion of its second seawater pumping station, the cooling water
pumping facility with a capacity of almost 700,000 m3/h will be one of the largest in the
Middle East. Some of the industries at SIPA have their own treatment plants to produce
process and boiler water using this seawater.
The plant comprises a 585MW combined-cycle, gas-fired power plant and a 170,000 m3/d
desalination plant located in the Sohar Industrial Port area in the North Al Batinah region of
the Sultanate of Oman (Figure 11.3.1). The desalination plant comprises four conventional
Multi Stage Flash (MSF) units and a SWRO system. The 20,000 m3/d SWRO plant is unique
in nature by using direct membrane filtration, not conventional pretreatment. The raw water
originating from an open intake is pumped directly through the microfiltration (MF)
membrane filtration system, then directly to the RO units without a break tank. The
characteristics of the SWRO plant are given in Table 11.3.1.
HABs have been observed in the Sohar Region since 1970s, but presumably occurred long
before that. Majis started its operations at SIPA in 2006-2007, and since 2008 has kept a
record of algal bloom events. In 2008-2009 a severe red tide of Cochlodinium polykrikoides
disrupted operations at SIPA resulting in huge economic loss. Sohar Refinery shut down its
own SWRO plant and intake screens required cleaning every 4 hours instead of every 12. In
2011 and 2013 green algal blooms also impacted SIPA operations.
In September 2013, when the RO plant was ready for commissioning, the seawater
conditions were very challenging with very high levels of SDI2.5 >35, and high algal bloom
cell concentrations > 5.8 million cells/L. This prolonged HAB event lasted almost six months
during which the installations had to be commissioned, the feedwater consistently having the
same high level of SDI.
11.3.2 System description
The screened and continuously chlorinated seawater is supplied from the Seawater Intake
Pumping Station (SWIPS I). The SWIPS I has an open intake and the water comes from a
shallow lagoon of 4 m depth which provides an ideal environment for algal blooms (Figure
11.3.1). The SWRO plant consists of MF pretreatment (with no coagulation or chemical
injection prior) followed by five SWRO trains, each having a capacity of 4,000 m3/d. Two
trains are dedicated for process water and the other three for potable water. A diagram of the
20,000 m3/d MF plant is provided in Figure 11.3.2. The microfiltration (MF) system uses Pall
Microza UNA-620A, high crystalline PVDF Membranes, with < 0.1 µm pore size, and an
outside-inside filtration mode. The first-pass SWRO system consists of a combination
Hydranautics membrane, SWC5max/SWC6max at train 1 and 2 and SWC4max/SWC5max
at train 3, 4 and 5. The second pass BWRO trains consist of ESPA2-LD/Hydranautics. The
average membrane flux for the SWRO and BWRO systems is 16.6 L/m2h and 36.5 L/m2h,
respectively.
The 20,000 m3/d plant, was designed for standard seawater quality conditions where the
duration of any red tide or algal bloom events were expected to last for two to a maximum of
357
four weeks, once per year. The MF plant was designed to operate at a maximum flux of
100 L/m2h. This flux has been demonstrated on a similar direct feed seawater installation that
has been operating for 5 years in Ducum, Oman. While pre-commissioning the Majis
20,000 m3/d plant, SDI5 could not be measured as the SDI membrane filter completely
blocked. These high levels continued and a few weeks later, a significant number of jellyfish
affected the intake. These were up to 20 cm in size but due to their sheer numbers, organic
matter penetrated some of the bar screens and this was not fully removed by the band
screens. This high load partially clogged the downstream 100 µm self-cleaning strainers. The
jellyfish attack was then rapidly followed by a HAB event consisting of unidentified fine
green algae with an excessively large presence of the dinoflagellate Gonyaulax polygramma
(Figure 11.3.3, Table 11.3.2).
Figure 11.3.2. Process schematic for the Sohar SWRO desalination plant – the 100 µm screens first
installed were replaced by a 300 µm strainers.
Figure 11.3.3. Clogged self-cleaning strainer during jellyfish and HAB event (left) and Gonyaulax
polygramma, the dominant species during the 2013 HAB event (right).
358
Table 11.3.2 summarizes the biological analysis for samples taken from MF feed, filtrate, and
backwash waste during the event. The direct impact of the massive changes in feedwater
conditions are summarized below:
- Partially clogged self-cleaning strainers;
- High transmembrane pressure (TMP) on the MF units due to unexpected
feedwater conditions; and
- Biofouling on the downstream SWRO units, due to the high level of dissolved
organic carbon (DOC), that was not removed by the MF pretreatment.
Table 11.3.2 Algal cell counting and species identification of the MF feedwater, filtrate, and
backwash water during a 2013 algal bloom event.
MF Feed
Phytoplankton Species MF Backwash MF Filtrate
seawater
cells/L cells/L cells/L
Cylindrotheca closterium < 25,000 250,000 Absent
Pleurosigma normanii < 25,000 250,000 Absent
Pleurosigma sp. < 25,000 250,000 Absent
Ceratium furca < 25,000 250,000 Absent
Dinophysis caudata < 25,000 250,000 Absent
Gonyaulax polygramma 5,637,500 53,125,000 Absent
Prorocentrum sigmoides 37,500 375,000 Absent
Protoperidinium steinii < 25,000 250,000 Absent
Total 5,825,000 55,000,000 Absent
11.3.3 Short-term mitigation measures

11.3.3.1 Intake
Short-term mitigation measures were urgently required to supply water during the 2013 event.
Thus the backwash frequency of the intake screens was increased to minimize the chances of
intake screen blockage, along with an increase in the chlorine dose. Change over to shock
dose chlorination was considered as a long-term option.
11.3.3.2 Self-cleaning strainers
Due to repeated partial-to-complete clogging of the existing 100 µm self-cleaning strainers,
the pressure drop across the strainers lead to insufficient pressure at the MF feed pump inlet
and caused automatic shutdown of the pretreatment plant. Moreover, there was insufficient
feed to the MF and the SWRO plant. The sticky nature of the jellyfish made it difficult-to-
impossible to clean the strainers. Due to time limitations and the long procurement time
needed for installing an alternative type of strainer, it was decided to take advantage of
having robust MF modules downstream of the self-cleaning strainers, and temporarily replace
the 100 µm mesh by a 300 µm mesh. In addition, the screen mesh material was changed to
non-metallic instead of metallic, thereby facilitating the removal of sticky materials from the
screen surface. Although, these measures helped in addressing the pressure drop issue on the
359
strainers and allowed for continuous operation for the self-cleaning strainers, it transferred a
greater solids load to the downstream MF units.
11.3.3.3 Microfiltration system
The high-crystalline outside-in PVDF membranes are known for their ability to handle
difficult feedwater without the need for upstream conventional pretreatment. Although the
outside-in filtration concept makes these modules more suitable for difficult water
applications, the unexpected feed conditions in 2013 presented a challenge for the MF
membranes, as it was originally sized at a flux of 100 L/m2h, similar to another 5 year old
installation on the Duqum coast of Oman. This plant operated for the majority of the time in
normal seawater conditions. Majis therefore, invited the MF membrane supplier to revisit
their design based on the new feedwater conditions and to come up with short-term and long-
term mitigations to improve and sustain the MF performance.
Due to the high organic load, and the other parameters outside of the design envelope, it was
decided to operate the MF plant at reduced capacity to allow for stable MF operation at
reduced flux. Table 11.3.3 shows the original MF design flux and the applied reduced flux
for the actual feed conditions.
Furthermore, Majis rented and installed additional 3,000 m3/d containerized MF-RO units
near the main plant to compensate for the applied drop in production and to allow the Pall
team to pilot the best operational settings for the main plant under the new feedwater
conditions using the same membranes.
Unlike the main plant, the new rental MF units could be sized based on actual water
conditions. They thus offered Majis an opportunity to demonstrate the new design settings
planned in the main plant for long-term operation. The rental MF containers were operated
on the same setting and the same cleaning regime as adapted for the main plant during the
short-term mitigations.
Table 11.3.3. Original MF design flux versus reduced flux.
Short-term
Original MF
Description Unit mitigation
unit design
conditions
Daily feed to MF m3/d 51,600 42,000
3
Daily net filtrate to RO m /d 48,347 38,737
Number of MF racks Racks 7 7
Number of modules/rack Modules/rack 84 84
2
Total installed membrane surface m 29,400 29,400
area
Flux condition when all racks in 73 60
filtration (N) L/m2h
Flux condition at N-1 85 70
L/m2h
Flux condition at N-2 L/m2h 102 84
Operation at a reduced flux allowed for a lower permeability drop during the filtration cycle,
and a lower frequency of cleans per day. It was necessary to revisit the cleaning regime
360
considering the high level of organics and the high SDI values in the feedwater. The high
crystalline outside-in PVDF membranes are known for their high chemical stability for
chlorine and caustic. Given the poor nature of the feedwater (i.e., the high organic load)
mixed with high concentrations of feedwater iron, it was important to apply skilled use of
alkaline and acidic chemical cleaning regimes to achieve the right balance and to restore MF
permeability. Table 11.3.4 summarizes the original cleaning plan and the changes made
based on the feedwater conditions.
Table 11.3.4. Original MF cleaning versus modified cleaning based on feedwater conditions.
Short-term
Description Original design
mitigation
Filtration cycle time / backwash Every 30-40 minutes Every 20-30 minutes
frequency
Number of EFM1 cleanings per Once/day Twice/day

day
Alkaline EFM regular cleaning Chlorine Chlorine + Caustic
Acid EFM cleaning frequency Once every 5 days Every 3 days
Acidic EFM cleaning HCl HCl/Citric
EFM cleaning process From outside surface of From outside/inside

the fibers surface of the fibers
1. EFM = enhanced flux maintenance
11.3.3.4 Results of short-term mitigation measures

Figure 11.3.4 shows the MF performance before and after applying the short-term mitigation
measures. Permeability started to improve once the plant’s operational settings had been
adjusted to suit the algal bloom conditions with the modified cleaning regime starting in early
February. This change led to an immediate improvement in the permeability trend, with the
modules recovering their normal permeability within two weeks. The acid EFMs used
hydrochloric acid (HCl) to counter the effect of hardness scaling from seawater and were
alternated with citric acid cleans to counter the impact of the iron fouling. Alkaline EFMs
also used NaOH to counter the effects of the high algal concentration in the seawater feed.
The number of EFM cycles and the backwash frequency were increased during this period to
control the permeability decline. Considering the long duration of the disruptions that were
experienced, it was recommended to keep the flux < 85 L/m2h to reduce chemical and energy
consumption.
As per the normalized data, the differential pressure (DP) (Figure 11.3.5) started to increase
in the SWRO trains in the middle of December when there was a significant change in
seawater quality due to a green algal bloom. Such events are usually associated with
increased amounts of organics in RO feedwater, and this was confirmed by TOC
measurements. Since at least 50% of the TOC was in a dissolved state, it passed through the
361
120
MF Permeability (L/m2/hr.barg)
100
80
60
40
MF Performance MF Performance Stable MF
20 before applying after applying short Operation
short-term -term mitigations under algal
mitigations bloom event
0
05/Feb/14
10/Feb/14
15/Feb/14
20/Feb/14
25/Feb/14
02/Mar/14
26/Jan/14
31/Jan/14
Figure 11.3.4. MF Performance before and after applying the short-term mitigation measures.
Figure 11.3.5. Normalized differential pressure during the period that short-term mitigation measures were
adopted.
MF and reached the RO stage. Furthermore, the seawater intake is continuously chlorinated
with a high chlorine concentration. The high dose of SMBS to de-chlorinate also has a
negative impact on the RO membranes.
Chlorine, a strong oxidizing biocide, breaks down larger organic chains into smaller more
assimilable organics, which are much easier for microorganisms to utilize as food. The
combination of all these factors triggered biological fouling of the SWRO membranes. A
362
cleaning regime with high pH and warm water flushing yielded positive results and DP were
brought under control. This resulted in frequent RO cleaning cycles, which again
detrimentally impacted the daily production rate of the plant. Overall the plant operator
adapted a strategy to achieve optimal plant availability and to keep plant productivity up to
88%. In addition the SWRO was operated in a manner to control DP increases using frequent
short cleaning followed by high pH solution and then hot water flushing. The short cleanings
helped reduce downtime periods of the SWRO system to once per month.
11.3.3.5 Results of demonstration study
The short-term mitigation measures successfully led to stable operation for the self-cleaning
strainers and the MF units. But, these measures were not sufficient to confirm long-term plant
operation. When considering the seasonally, unstable seawater conditions, it was important to
demonstrate the reduced MF flux for a longer time on actual pilot operation. It was also
necessary to find a permanent solution for the DOC that passed through pretreatment and
affected the downstream RO.
Therefore, a demonstration study was carried out on site using the rented containerized MF-
RO units, with net RO capacity of 3,000 m3/d near the main plant and sharing the same feed
source. The objective was to demonstrate and confirm the reduced MF flux, to confirm the
fiber status, and to select the optimum cleaning solutions that could fully recover the modules
and achieve sustainable operation in the long term. Table 11.3.5 summarizes the MF
container sizing parameters.
Table 11.3.5. MF-RO demonstration study sizing parameters.
Description Unit Original MF Units Design
Daily feed to MF container m3/day/container 4,225
Daily net filtrate from MF container m3/day/container 3,950
Number of MF racks/container racks 2
Number of modules/rack modules/rack 30
Total installed membrane surface area m2 / container 3000
Instantaneous feed flow/container m3/h 230
Constant operation flux L/m2h 76
Figure 11.3.6 shows the MF stable performance during 8 months of continuous operation.
The containerized MF unit operation was stable and the permeability continuously
Figure 11.3.6. MF rack 1 and MF rack 2 Performance in the containerized MF unit.
363
maintained above 100 L/m2h.bar, except for a few days during an algal bloom event that
repeated again in September 2014.
11.3.4 Long-term mitigation measures
With a strategic view on the desalination business of Majis, it has been decided to replace
continuous chlorination with shock chlorination. This will be challenging, as the Majis intake
system serves various types of industries that are dependent on the chlorine level at their heat
exchangers.
For the removal of DOC, it is important to use coagulation upstream of the MF system to
remove dissolved organics that are currently fouling the RO membranes. Either a direct
coagulation before the MF units will be applied in the existing plant, or during phase 2 of
plant expansion, a common gravity filter stage with coagulation dose could be installed
preceding the new RO system. This would be located upstream of the MF system to cover
both phase 1 and phase 2.
The self-cleaning filters will be expanded by adding an extra two units, and the drum screens’
pore size will be reduced back to 100 µm from 300 µm. This increase in the screen area will
allow for continuous supply and the use of non-metallic parts instead of the metallic screens
that have enhanced the removal of sticky materials from the screen surface.
Based on the research study, operation under a relaxed flux range of < 76 L/m2h showed very
stable operation for the MF containerised units. The same flux was used as the maximum MF
flux for long term mitigation of poor feedwater in the main plant. Extra membrane surface
area will be added to allow full production at the reduced flux. Table 11.3.6 summarizes
design changes for long-term mitigation of algal blooms.
The frequency of HCl cleaning will be increased as will the use of citric acid in addition to
HCl. Alkaline cleaning will use a mixture of chlorine and caustic soda (Table 11.3.7). The
use of direct coagulation is one of the options proposed for removing dissolved organics.
Table 11.3.6. Planned design changes at Majis.
Description Unit Original MF Long-term
unit design mitigation
Daily feed to MF m3/d 51,600 58,500
Daily net filtrate to RO m3/d 48,347 54,800
Number of MF racks Racks 7 9
Number of modules/rack modules/rack 84 96
Total installed membrane m2 29,400 43,200
surface area
Flux condition when all 73 56
racks in filtration (N) L/m2h
Flux condition with N-1 85 63
L/m2h
Flux condition with N-2 102 72.5
L/m2h
Since the pretreatment system has been modified, the RO system performance is expected to
improve. There is also a plan for reducing the flux rate on the SWRO from 16.6 L/m2h to
14 L/m2h by adding extra RO membranes. With the use of coagulation in front of the MF
units, the majority of DOC should be removed, and this will further achieve better results in
the RO system.
364
Table 11.3.7. Long-term operational settings for the MF system.

Description Original design Long-term mitigation
Filtration cycle time / backwash frequency Every 30-40 Every 20-30
minutes minutes
Number of EFM cleaning per day Once/day Twice/day
Alkaline EFM regular cleaning Chlorine Chlorine + Caustic
Number of EFM cleaning per day Once/day Twice/day
Acidic EFM cleaning HCl HCl/Citric
Acid EFM cleaning frequency Once Every 5 days Every 3 days
EFM Cleaning Process From outside From outside/inside
surface of the fibers surface of the fibers
11.3.5 Conclusions
Although the raw water during commissioning was significantly poorer in quality than
expected, it was possible with some modifications to start successfully and operate the
20,000 m3/d plant at the Majis port of Sohar. Lessons learned from the various challenges
faced during this phase of operation include:
• The open seawater intake with shallow depth, high silt concentration, and continuous
chlorination is challenging enough, but when jellyfish and algal blooms also appear,
raw water quality can deteriorate to the extent that normal operation becomes difficult.
In those instances, a special operating regime needs to be adopted to enable the plant
to operate with maximum output.
• Accurate and detailed feedwater analysis over a period of at least one year is required
for the design, especially in the case of open intake installations liable to be affected
by jellyfish and HABs.
• It is possible to apply direct membrane filtration in open intake installations
challenged by high SDIs and prolonged algal bloom events without the need of
conventional pretreatment if accurate feedwater analysis over a period of time is
provided, and a combination of direct coagulation and membrane filtration are used as
pretreatment.
• The membrane installation should always be designed to allow for the addition of
extra modules, and space should be allocated for an additional rack to mitigate against
a significant deterioration in the feedwater quality.
• The skilled use of alkaline and acid cleaning regimes can restore the permeability of
the microfiltration membranes, even when challenged with unexpected feedwater
conditions.
• Highly crystalline PVDF modules and the outside-in configuration are extremely
robust and able to effectively operate in variable and unexpected seawater conditions.
365
366
11.4 BARKA 1, OMAN - THE IMPACT OF HARMFUL ALGAL

BLOOMS ON THE PERFORMANCE STABILITY OF UF
PRETREATMENT
Graeme K. Pearce1
1
Membrane Consultancy, Reading, England
Figure 11.4.1. Location and aerial view of Barka 1 desalination plant in

Oman.
367
Table 11.4.1. Overview of the Barka 1 desalination plant.

Plant/Project Name
Location Barka, Oman
Desalination Technology UF-SWRO
Total Production Capacity 45,500
(m3/d)
SWRO recovery (%) 43%
Commissioning date February 2014
Intake
Feedwater source Gulf of Oman
Intake type Shore intake via MSF cooling system of neighboring
plant
Intake description Intake designed for MSF (1200m from shore, 500 m min
5 m depth)
Intake screening Intake screen (3 mm) designed for MSF
Chlorination Hypochlorite continuous dosing, at 0.1-0.2 mg/L residual
Online raw water Conductivity, temperature, pH, turbidity, residual
monitoring chlorine
Discrete raw water analysis TOC, Total N,
relevant to HAB chlorophyll-a
Pretreatment
Process description Pressurized UF, inside-out with 300 µm pre-strainers
Option for in-line Fe³+ coagulant dosing, but not used
during 9-month operational period under review
Feedwater design parameters Feedwater during bloom
conditions
Temperature range (˚C) 40 - 45
Salinity range Approx. 40,000
(TDS mg/L)
Conductivity (mS/cm) Approx 55
Total Suspended Solids <5 Unknown
(mg/L)
SDI₂.5 (2.5 minute) 20 - 35 Light bloom: 35 – 40
Significant bloom: > 40
Turbidity (NTU) Approx. 0.2
Organic matter (mg/L) Approx. 1 TOC
UF flux (L/m2h) 65 6565
368
11.4.1 Introduction
The original desalination and power plant at the industrial port city of Barka was based
around a 91,000 m3/d multi-stage flash (MSF) thermal desalination unit. Barka is located 60
km north east of Oman’s capital, Muscat, and draws its feedwater from the Gulf of Oman as
shown in Figure 11.4.1. In 2012, the plant operator, ACWA Power Barka won a contract
from the government agency Oman Power and Water Procurement (OPWP) for expansion of
the facility with a 45,500 m3/d seawater RO (SWRO) unit. The contract to build this first
phase expansion plant, subsequently referred to as Barka 1, was awarded to Abeinsa. The
contractor, a wholly owned subsidiary of the Spanish EPC Abengoa, is experienced in UF-
SWRO design and operation including with feedwater that are prone to algal bloom
(Bernaola 2012). Commissioning for Barka 1 started in September 2013 with continuous
operation in February 2014. In 2015, a second phase expansion for a further 57,000 m3/d,
known as Barka 2, was also brought online.
The Barka 1 plant uses an open intake seawater feed without any pretreatment by
clarification or flotation. Before arriving at the RO desalination plant, the feedwater passes
through the heat exchangers of a condensing system providing cooling water for the MSF
facility. The plant has a provision for in-line coagulant dosing which was intended for
periods of poor feedwater quality, but for the period of review in this case study, this dosing
facility has not been used.
Details of the algal species at Barka during the period of this investigation were not
determined directly. The HAB and red tide causing species commonly found in these areas
are: Noctiluca scintillans, Dinophysis spp., Gonyaulax sp., Trichodesmium sp., and Ceratium
spp. of which the dinoflagellate, N. scintillans is a major species that produces red tide
blooms yearly (Thangaraja et al. 2007). More recently, 24 algal species have been identified
in these waters (Al-Azri 2014) with relatively abundant concentrations of the dinoflagellates
Prorocentrum minimum and Cochlodinium polykrikoides being reported for the first time.
The pretreatment system at Barka 1 utilizes Pentair-Xflow UF. The Xflow membranes are
made from polyethersulfone (PES) and have a nominal 0.02 µm pore size rating. The
membranes operate with an inside-out feed flow configuration, with four 1.5 m elements
mounted in a 6 m horizontal pressure vessel.
This case study focuses on the performance stability of the UF system during a 9-month
period which took place 5 months after continuous operation commenced in February 2014.
This ensured that initial conditioning of the membrane was completed and effective chemical
cleaning procedures were established. The period under review started in July 2014 and
finished in April 2015 and was assessed through the monitoring of permeability trends for
three months at the start of the review period and comparing these trends with the three
months at the end of the period.
11.4.2 Plant overview
A simplified block flow diagram for the Barka 1 desalination plant is shown in Figure 11.4.2.
The seawater feed is taken from an open intake 1200m from shore at a depth of 5 m below
the lowest tide level. Following 3-mm screens, the seawater is used as cooling water for the
neighboring MSF desalination plant, which increases the temperature by 9 ºC. The feedwater
is chlorinated with a 2 ppm free chlorine dose prior to the MSF, but by the time it reaches the
membrane units, the chlorine residual has reduced to around 0.2 ppm. The desalination plant
feedwater is filtered by 300 µm strainers before passing through the eight UF units. There is
an option for in-line coagulant dosing, but this was not employed during the nine-month
period of evaluation described in this case study. The UF is connected directly to the RO
369
without a balance tank. The RO high-pressure pumps feed three RO trains that normally
operate continuously at a constant flux.
Figure 11.4.2. Simplified block flow diagram showing feed and product flows for the Barka I SWRO
desalination plant.
Essential details for the Barka 1 desalination plant are summarized in Table 11.4.1 with
details of location, intake, and pretreatment design as well as a description of the feedwater
quality. Note that the feedwater from the MSF cooling system is continuously chlorinated
with de-chlorination prior to the SWRO. Algae is frequently experienced in this feedwater,
but the blooms have not been too severe to date.
11.4.3 UF System Design
11.4.3.1 Membranes
The Barka UF pretreatment system utilizes the Pentair-Xflow Xiga 55, which has a
membrane area of 55 m² for each element. The plant is arranged in eight units. Each unit has
46 housings that hold 4 elements per housing, thus there are 184 elements per unit with a
total membrane area of 10,120 m². The total membrane area for the whole plant is 80,960 m².
There are two vacant places on each of the units, so that a total of 48 housings per unit could
be used if production capacity needed to be expanded allowing a 4.3% increase in membrane
area.
The housing for the Xiga 55 element has a 9 inch (225 mm) nominal bore. Prior to 2012,
most Pentair-Xflow systems were based on the 8 inch (200 mm) Xiga 40 element. The Xiga
55 still uses a central filtrate collector, but has bypass channels on the outer circumference
rather than around the central collection pipe. The additional diameter combined with the
different filtrate collection arrangement has enabled a significantly higher membrane area to
be obtained from a single element, even though the dimensions of the membrane fiber are
identical (internal diameter, 0.8 mm, external diameter, 1.3 mm).
370
11.4.3.2 In-line design

The UF is connected directly to the RO in a direct-coupled design that eliminates the break
tank between UF and RO systems. Pentair-Xflow is a strong proponent of the in-line UF-RO
system design, especially for large scale seawater duties. The control system provides a
constant inlet pressure to the RO feed pump at approximately 2 bar. Eliminating the RO
feedwater tank reduces space and cost for the system and saves energy. Of arguably greater
significance, the intermediate tank can promote biofouling, and so by eliminating it from the
flowsheet, the operational stability of the RO can be improved and chemical cleaning
frequency reduced (Pearce 2013; Knops 2014).
In-line designs need to employ a significant number of units to ensure that flow can be kept
reasonably constant. In order to minimize fouling it is preferable to keep flux variation within
10% of the average if possible (+/- 15% maximum), or alternatively the average flux should
be reduced by a few percent (Pearce 2015). In the Barka 1 design, the flux selected is
moderate and the flux variation for the in-line design has been optimized such that the
variation falls within the 10-15% guideline. Accelerated fouling would therefore not be
expected, and as will be seen from the results review in section 11.3.4, has not been
observed.
11.4.3.3 Flux
For a plant without flux variation, it would be expected that 70 L/m2h would be a suitable
design flux for the Pentair-Xflow membrane on seawater from the Gulf based on comparison
with other plants. The Barka 1 plant has been designed to operate at an average UF flux of
around 65 L/m2h, but with a modest degree of flux variation. Flux varies during the filtration
cycle (time between backwash) between 60 and 70 L/m2h, with occasional short-lived
excursions down to 55 L/m2h and up to 75 L/m2h. Fine-tuning the process cycle has
minimized the flux variation. At the design flux, the filtration cycle time is approximately 50
minutes, but this cycle time may extend at lower flux.
When a unit is taken off-line for backwash, the remaining units increase their flow in
proportion causing an increase of flux over the average for when all 8 units are available, that
is, from around 62 L/m2h with 8 units in filtration, to about 70 L/m2h when one unit is in
backwash. The filtration cycle takes about one minute. At the same time, some filtrate is
taken from the manifold to replenish the backwash tank. The replenishment rate has been
optimized to ensure that there is no adverse effect on flux variation.
11.4.3.4 Temperature correction
Warm feedwater temperatures increase clean water permeability. For example, an elevation
of 9 oC above ambient should increase permeability by 30%. Furthermore, the elevated
temperatures would reduce lumen length pressure drop as well as pressure drop in the shell,
resulting in improved hydrodynamic efficiency of the module. The benefit at 70 L/m2h
would, however, be minimal since hydrodynamic stress at this moderate flux is relatively
low.
Furthermore, the benefit of elevated temperature for a real feedwater is observed to be
significantly less than the 30% permeability improvement indicated above. For seawater from
the Gulf, fluxes are determined as much by the concentration of particles in the lumen as by
the permeability across the membrane. Any increase in flux would increase the particle
concentration in the lumen and at the membrane surface thereby increasing the fouling rate.
Thus in these circumstances, elevated temperature would have no significant benefit. This
assumption is reinforced by the practical observation that there does not appear to be a
371
benefit from high temperature operation for other power plant cooling water desalination
systems.
11.4.4 UF system performance
11.4.4.1 Expected permeability
A typical Pentair-Xflow membrane has a permeability of 400-500 L/m2h.bar during early
operation on a surface water feed at 20oC (for example during take-over testing). Generally,
the membrane would stabilize after the first month at 300-400 L/m2h.bar, depending on
feedwater quality, and this would be expected to represent an initial baseline.
Seawater feeds may have lower permeabilities than for surface waters at the same
temperature since the viscosity and density are both higher. Permeability is also strongly
affected by foulants and the adsorption of organics, especially algae and algal derivatives. A
higher temperature would decrease viscosity, which would give an increase in permeability.
For the Gulf water feedwater at Barka, taking account of the effects described above, it
would be expected that the baseline permeability after the initial operating period should be
around 350-400 L/m2h.bar, depending on local feedwater quality characteristics. Permeability
reported in subsequent sections shows that the Barka 1 plant has been able to achieve the
expected permeability level.
11.4.4.2 Use of chemical cleaning to maintain permeability
Performance stability is maintained by conducting a chemical enhanced backwash (CEB) on
a regular basis. Under normal circumstances with good quality feedwater, the CEB would be
scheduled once every 23 or 24 backwash cycles, equivalent to an interval of around 18 hours
between consecutive CEB’s (which is the period referred to in this review as a CEB cycle). If
feedwater quality deteriorates, for example due to an algal bloom, or permeability drops for
some other reason, the CEB is performed more frequently, perhaps after 12 or 15 hours. The
change of CEB frequency is initiated manually.
If there is a long-term deterioration in CEB, a Clean-In-Place (CIP) is performed. The CIP is
more effective than CEB since the contact time is longer and the chemicals may be
recirculated. The longer CIP downtime would mean that recovery and hence output would be
reduced if conducted too frequently, so it is beneficial to make the interval between CIP as
long as possible with a target of at least 6 months. It is therefore important for the CEB to
recover a high proportion of the permeability lost during the course of a CEB cycle.
360 CEB cycles have been performed in the period of this review. The following time
periods have been evaluated which used the full design flux at a consistent level:
1. Period 1: CEB cycles 8 to 119 during 10 July to 7 October 2014; and

2. Period 2: CEB cycles 201 to 360 during 1 January to 27 April 2015
Prior to cycle 8, flux varied. No data were evaluated for cycles 120 to 200 in the last quarter
of 2014, but the plant appeared to operate in a stable fashion at this time. The starting
permeability at the beginning of Period 2 was fully satisfactory and in line with expectations
for this type of operational plant with membranes in good condition.
The analysis focuses on the performance of unit 7 since this unit was found to have average
performance, although, variation between units was found to be almost negligible.
372
11.4.4.3 Permeability trends during period 1

Figure 11.4.3 shows the permeability trend data for Period 1, CEB cycles 8 to 119, together
with the CEB cycle time on the right-hand axis.
Figure 11.4.3. Permeability at start and end of CEB cycle for cycles 8-119
The chart shows the loss of permeability during the course of a CEB cycle was significant at
around 100-150 L/m2h.bar, and that the CEB was reasonably effective at recovering the
permeability (Figure 11.4.3). The CEB was not fully effective, and there was a gradual loss
of permeability with time. The loss reduced slightly in the later cycles and showed a response
to external factors.
At the beginning of CEB cycle 8, the initial permeability was fully satisfactory. The chart
shows the trend lines for start and end permeability. The two trend lines gradually converge,
indicating that the rate of fouling was falling, and this issue will be discussed in more detail
below. The slope can be used to predict the requirement for CIP and quantify the CIP interval
and this will be discussed below.
For a given flux, permeability is affected by the following control variables:

• Filtration cycle time
• CEB interval
• Feedwater quality
The loss of permeability in a single filtration cycle was approximately 5 L/m2h.bar per cycle.
A series of 24 cycles would therefore lead to a permeability loss of approximately
120 L/m2h.bar. In reviewing the data, there has not been an opportunity to examine the effect
of changing the filtration cycle time, since this value is linked to the flux, and only increased
if the flux reduced. It is possible that the rate of fouling could be reduced slightly by reducing
the filtration cycle time from the standard interval of 50 minutes. This action would not be
recommended, since recovery would fall significantly. The first of the three parameters is
therefore not a control variable as far as the operator is concerned.
373
The second factor listed is the interval between CEB. Although normally set at one CEB
every 23 or 24 backwash cycles (equivalent to a time interval of 18 hours), the interval can be
reduced as shown by the green triangles in Figure 11.4.3.
The final operational factor affecting performance is the feedwater quality. The main
parameter that appears to correlate with permeability is the chlorophyll concentration.
Typical concentrations vary between 0.02 and 0.04 µg/L, but minor excursions to 0.1 µg/L
do not appear to cause the UF system any problems and appear to indicate a normal non-
problematic seawater quality. A moderate algal bloom event between cycles 40 and 60 (mid
to late August 2014) had a significant effect on permeability, but the plant managed to
maintain flux and recovered without any intervention other than a reduction in the CEB cycle
time. For a more significant algal bloom, it might be necessary to reduce flux or possibly
consider coagulant dosing.
Figure 11.4.3 shows that increasing CEB frequency from once in 18 hours to once in 15
hours or 12 hours during a period of high algae between cycles 40 and 60 enabled the plant to
cope with the challenging conditions. Although permeability dropped during this period by
around 100 L/m2h.bar, the plant recovered without having to reduce the flux. Permeability
then recovered to a level only marginally below the original level once feedwater quality
returned to normal conditions.
Note that the CEB cycle time remained slightly below the original level between cycles 60
and 80, which appeared to arrest the rate of decline slightly. The situation was made more
complicated by the fact that there was a sticking valve on unit 7 at around cycle 75, that is,
during the 2nd week of September. The result of this was that CEB was not fully effective
during this period and permeability temporarily declined. After the problem was resolved, the
plant returned to previous levels without any further intervention, again underlining the fact
that the operational regime was sufficiently robust for performance to recover after a minor
upset.
11.4.4.4 Permeability Trends after One Year of Operation
Figure 11.4.4 shows permeability trends for cycles 201 to 360. The figure shows that the
starting permeability in cycle 201 was similar to the starting permeability at cycle 119, thus
the gradual decline discussed in the previous section had been arrested. The difference
between the start and end permeability from cycle 201 onwards is normally around
100 L/m2h.bar or just over which is similar to or slightly less than in cycle 119.
The initial permeability did not decline further between cycles 119 and 201 because of the
use of a slightly shorter CEB cycle time, i.e., CEB frequency increased. Shortening CEB
cycle time reduces the loss of permeability in the course of a CEB cycle. For cycles from 201
onwards, CEB cycle time was 12 - 15 hours as opposed to 18 hours previously. The higher
frequency was implemented in response to the poor feedwater quality from an algal bloom
that began in December and continued for much of January through April. Increasing CEB
frequency made the start permeability more stable. This has the effect of lengthening the CIP
interval and will be quantified in section 11.4.4.7.
374
Figure 11.4.4. Permeability at start and end of CEB cycle for cycles 201-360.
11.4.4.5 Analysis of CEB Cycle Time
Figure 11.4.5 shows the effectiveness of the CEB at recovering permeability. Between cycles
8 and 119 there was a steady drop in the recovery obtained, and that drop can be used to
predict the required CIP interval. It would be possible to improve recovery by changing the
procedure or the frequency of the CEB. Procedural changes would include the following:
• increasing the concentration of the chemicals used;
• increasing the contact time;
• changing the chemicals used; and
• changing the order in which the CEB’s were performed and chemicals utilized in the
CEB procedures
The other changes that could be considered for improving the effectiveness of chemical
cleaning and might be practical for a CIP procedure are:
• introducing flow-through or recirculation to improve mass transfer (due to
replenishment of spent chemical at the foulant location); and
• warming the chemicals (only applicable in cold climates and not relevant at Barka)
These changes would depend on whether they could be accommodated by the system design.
A simpler alternative to changing the CEB procedure is to increase the CEB frequency, and
the effectiveness of this approach is demonstrated in Figure 11.4.6 for CEB cycles 201 to
360. Whereas the average CEB cycle time was 18 hours in Figure 11.4.5, the cycle time in
Figure 11.4.6 was reduced to 12 or 15 hours up until cycle 320. The consequence of this was
that permeability recovery steadily improved despite the fact that feedwater quality was
375
Figure 11.4.5. Permeability recovery by CEB; cycles 8-119 (CEB cycle mainly 18 hours)
Figure 11.4.6 Permeability recovery by CEB; Cycles 201-360 (CEB cycle mainly 12-15 hours)
poorer. Early in April, the CEB interval was increased to the original level of 18 hours,
resulting in a sharp drop off in recovery. The problem was exacerbated due to a second
sticking valve incident which made CEB completely ineffective for 3 or 4 cycles between
cycle 356-358 on 22nd to 24th April 2015. The strong downward trend was already well
376
established by this time, so the sticking valve did not appear to initiate the problem of
permeability drop off.
11.4.4.6 Fouling Rate Trends
Figure 11.4.7 shows the fouling rate trend for cycles 8 to 119. The fouling rate is calculated
by dividing the permeability loss in units of L/m2h.bar in the course of a CEB cycle by the
CEB cycle time in hours. Data are also shown for the first 7 cycles in which the lower and
variable flux gave rise to a lower fouling rate. After an effective CIP at the end of June,
fouling rates after cycle 8 between July and October were steady and slightly decreasing,
apart from the period of poor feedwater quality (high algae) in the third week of August
(corresponding to CEB cycles 40 to 60).
The fouling rate during CEB cycles 201 to 360 is shown in Figure 11.4.8. In general, the
fouling rates during this period were higher than those previously. Whereas in period 1, one
algal bloom happened in the third week of August between cycles 40 and 60, Figure 11.4.8
shows that there were multiple occasions in period 2 when chlorophyll spiked higher. There
was one extended event from 13 January to 9 February (cycles 220 to 260) and there were
four other shorter episodes between late February and early April. The chart shows that the
fouling rate increased during the first episode, and stayed at a relatively high level through
the beginning of April. This sustained poor feedwater quality did not cause operational
stability problems however, since the CEB cycle time was shortened throughout this period.
Figure 11.4.7. Fouling rate vs CEB cycle number for Cycles 1 to 119.
As of October, regular measurements of the SDI were taken and these are shown on the chart
as well. Note that the values used by the operator were not the conventional 15-minute SDI,
and for simplicity the reported chart values in Figure 11.4.8 have been adjusted to show the
relative trend rather than the absolute values of SDI2.5. For very high fouling feeds, the
operator used an SDI1 measure at one minute. The operator’s SDI was 20 times the value on
the chart axis, that is, normally falling between 30 and 40 units for SDI2.5 and up to 90 for
SDI1.
377
High SDI was observed sporadically after 13 January (cycle 220) and for a sustained period
in early February when the algal bloom was at its most intense (cycles 250-260). After that,
high SDI was observed occasionally at times corresponding to the high chlorophyll
measurements. Many high SDI measurements were made during mid-late March (cycles 310-
340), but these were interspersed with lower SDI measurements, suggesting that the algal
bloom was light, or at least not sustained. SDI is useful as an indicator of high algal
concentrations, but the correlation is only partial, since other factors influence SDI values.
The fouling rate fell during April after cycle 340 due to the improvement of feedwater quality
and extension of the CEB cycle. Notwithstanding, as shown in this review, the frequency of
CEB was not sufficient to maintain the permeability during this period, which led to a
significant drop off in permeability (Figure 11.4.4).
Figure 11.4.8. Fouling rate, SDI and CEB cycle time for cycles 201 to 360 (Note – SDI values indicate a
relative trend of SDI2.5 values and are not SDI15 values).
11.4.4.7 Projected CIP Interval
Figure 11.4.3 shows that after a CIP, the permeability at the start of a CEB cycle can be
expected to be 350-400 L/m2h.bar. With an 18-hour interval between CEB, as was applied
for most of the time between cycles 8 and 119, the slope of the line of best fit through the
data shows a decline in permeability of around 30 L/m2hbar over 100 cycles or 1800 hours
(74 days). This is a rate of decline of 0.4 L/m2hbar/day.
Figure 11.4.3 also shows that the loss of permeability in the course of a CEB cycle is around
100 to 150 L/m2hbar between the start and end of each CEB cycle (the difference between
the two lines of best fit). The difference is slightly greater when the membrane is cleaner and
less towards the end of the cycle.
378
If we take it that just prior to a CIP, we should expect the permeability loss to be towards the
lower end of this range, say 100-125 L/m2h.bar, we can use this value to predict the
allowable decline in the starting permeability. An idealized chart is reproduced in Figure
11.4.9 using the typical permeabilities discussed. The chart shows that for an 18-hour CEB
cycle, the projected CIP interval would be 6 months. This would mean that after 6 months of
operation and prior to a CIP, the worst-case permeability would be 150-200 L/m2hbar
(assuming the permeability is measured at the end of the CEB cycle).
It should be noted that this analysis assumes that swift action is taken during a light to
moderate algal bloom to reduce the CEB interval first to 15 hours, and then to 12 hours if
necessary, and this shortened interval in maintained until the starting permeability has fully
recovered to its pre-bloom level.
400
350
300
If CIP performed
Permeability, lmh.bar
250
Permeability Start
200
Permeability End
Permeability After CIP
150 Permeability After CIP
100
50
0
0 2 4 6 8
Number of Months of Operation
Figure 11.4.9 Projected CIP interval for an 18 hour CEB cycle.

The permeability trends for cycles 201-360 show that shortening the CEB cycle to 15 hours
for extended periods would extend the interval between CIP beyond 12 months, especially
without feedwater quality excursions or equipment failure. Figure 11.4.4 showed that the
long-term permeability could be maintained at this frequency even during light algal events
and the system could easily recover itself during moderate algal events without flux
reduction.
Plant experience that has shown a 6 month CIP interval was achieved for unit 7 with the June
CIP being followed by a CIP in early February. Allowing for short shutdowns, unit 7 has
achieved just over 6 months of operation at full flux in line with this expectation. The
experience of chemical use on unit 7 has shown that it is better to manage permeability
expectations by slightly increasing CEB frequency and reducing the use of CIP.
The analysis has not determined the effect of a severe algal bloom, since none occurred
during the period of evaluation. Projecting the expected performance based on behavior
during moderate algal bloom events indicates that a stepwise reduction of flux may be
379
necessary during very poor feedwater quality. For example, an initial reduction of flux to
80% of the design level (to 52 L/m2h), accompanied by an immediate reduction of the CEB
interval to 12 hours would probably accommodate most situations. A further set of reductions
may be necessary if the algal bloom is dense and sustained.
Another option would be to use coagulant during significantly poor feedwater quality
episodes including high algal cell concentrations. This has not been found to be necessary
with the feedwater quality experienced to date.
11.4.5 Conclusions
• The permeability of Pentair-Xflow membranes for the seawater duty at Barka 1 was
expected to be at least 350-400 L/m2h.bar after cleaning by CIP; actual permeabilities
observed after the CIP in June 2014 were found to be in the acceptable range;
• The average flux since July has been consistently around 65 L/m2h with an acceptable
level of variation through the cycle (normally +/- 5 L/m2h and nearly always within
+/- 10 L/m2h), i.e. a maximum variation of around 15%;
• The fouling rate is significant but it is well controlled by CEB to ensure long term
stability of the permeability;
• Occasional high fouling periods have been observed and these can be linked with the
occurrence of moderate to high chlorophyll in the feedwater due to algal blooms,
which is also shown by an increase in SDI2.5 (>35 units);
• Plant operation without coagulation is stable and maintainable in the long term with a
CIP interval of ≥ 180 days and a CEB interval of 18 hours, provided that the CEB
interval is reduced to 15 or 12 hours during light or moderate algal blooms; the plant
has also been able to cope with algal blooms at this level without intervention (apart
from CEB cycle reduction) or reduction of flux;
• Under normal feedwater conditions and using an 18 hour CEB interval, the required
CIP frequency would be ≥ 180 days to ensure that the average end permeability was
≥ 175 L/m2h.bar; this should enable a CIP to work effectively to restore a high
starting permeability (note that CIP intervals at the plant have been > 180 days);
• A more severe and sustained algal bloom may require a reduction of flux (for
example by 20%); use of coagulant during such an episode could be an alternative;
• If the CEB interval were to be reduced from 18 to 15 hours for normal conditions, this
would have the effect of significantly extending the CIP interval well beyond 6
months (e.g. ≥ 12 months), and would ensure that permeabilities were steady and
stable with a significantly higher average permeability and hence lower pressure
requirement;
• With a 15 hour CEB interval, the plant would have much better resilience in handling
algal blooms or occasional equipment failure
11.4.6 References
Al-Azri, A. R. 2014. Expansion of harmful algal blooms in the coastal water of Oman, and
potential threat to desalination plants. In: Proceedings of the Harmful Algal Blooms
Conference, Mucat, Oman.
Bernaola, F. 2012. Red Tide Event Prevention and Influence on Desalinated Water
Production Cost, Expert Workshop Red Tide and HABs: Impact on Desalination Plants,
Muscat –Sultanate of Oman.
380
Knops, F. N. M. 2014. Reducing SWRO life cycle costs by direct coupling of UF and SWRO
Systems. In: Proceedings European Desalination Society Conference, Limmasol,
Cyprus.
Pearce, G. K. 2013. The Potential Advantages of Direct Coupled UF System Designs for Bio-
Fouling Control in Pre-treatment of Seawater RO Systems. In: Proceedings
International Desalination Association Conference, Tianjin, China.
Pearce, G. K. 2015. How Much Flux Variation Can We Allow in UF System Design;
Striking the Right Balance between Performance Stability and Cost Minimization. In:
Proceedings American Membrane Technology Association Membrane Technology
Conference, Orlando, Florida, USA.
Thangaraja, M., Al-Aisry, A., and Al-Kharusi, L. 2007. Harmful algal blooms and their
impacts in the middle and outer ROPME sea area. International Journal of Oceans and
381
382
11.5 SHUWAIKH, KUWAIT – HARMFUL ALGAL BLOOM CELL

REMOVAL USING DISSOLVED AIR FLOTATION: PILOT AND
LABORATORY STUDIES
Robert Wiley1, Mike Dixon2 and Siobhan F. E. Boerlage3

1
Leopold, a Xylem Brand, Zelienople, PA, USA
2
MDD Consulting, Kensington, Calagry, Alberta, Canada
3
Figure 11.5.1. Location of the Shuwaikh, SWRO Plant in

Kuwait (top). Algal (species unidentified) bloom near intake
(middle). DAF Float during the bloom (bottom).
383
Table 11.5.1. Overview of Shuwaikh plant.
Plant/Project Name
Location Shuwaikh, Kuwait
Desalination Technology SWRO
Total Production Capacity (m3/d) SWRO 135,000
Commissioning date October 2011
Intake
Feedwater source The Gulf
Intake type Open near shore intake
Intake description Intake depth 3 m, 200 m from shore, 5 mm coarse
screening grill aperture
Intake screening Travelling band screens
Shock chlorination Shock chlorination utilized. Dose unknown.
Strategy, dose rate
Online raw water monitoring Conductivity, temperature, pH, turbidity, dissolved
oxygen hydrocarbon analyzer.
Discrete raw water analysis DOC, TOC
relevant to HABs
Pretreatment
Process description DAF, autostrainers, UF (pressurized inside out), UF
direct coupling to RO (no cartridge filtration)
Chemical dosing Ferric chloride, Sulfuric acid
conditions
Temperature range (˚C) 20-35
Salinity range (TDS mg/L) 42,000
Conductivity (mS/cm)
Total Suspended Solids (mg/L) 5-43 (average 14.5) 17
SDI
Turbidity (NTU) 1-55 (average 5.2) >30
Organic Matter
TOC/DOC (mg/L), TEP,
biopolymers etc.
Algal cell count (cells/L) Unknown
Algal species Unknown
Chlorophyll–a (µg/L)
DAF Loading Rate (m/h) 22
UF flux (L/m2h) 80
RO flux (L/m2h) 15
384
11.5.1 Shuwaikh plant

DAF has been used by the majority of countries in the Gulf Region for all new SWRO
projects that utilize open intake systems. The application of the DAF process on seawater,
however, is still in its infancy. The Shuwaikh Seawater Reverse Osmosis (SWRO)
Desalination Plant, which was built for the Kuwait Ministry of Electricity and Water (MEW)
incorporates the Leopold Clari-DAF® System followed by pressurized inside-out
ultrafiltration (UF) for pretreatment. The pretreatment capacity is 350,000 m3/day with a
SWRO output of 136,000 m3/d. Figure 11.5.2 shows a model of a typical Clari-DAF System
as used at the Shuwaikh, Kuwait SWRO Plant.
Figure 11.5.2. Model of typical Clari-DAF system as used at the Shuwaikh, Kuwait SWRO plant.
The plant is located directly next to the Port of Shuwaikh, which accommodates the majority
of Kuwaiti shipping. Seawater in the region is warm, highly saline, laden with organics and
polluted due to port activities with occasional harmful algal blooms (HABs) known to occur.
Therefore, a robust pretreatment system was designed (Figure 11.5.3) with the intention of
protecting the downstream UF and RO from potential contamination (Figure 11.5.4). The
small footprint of the site necessitated a high-rate pretreatment system capable of dealing
with any contamination associated with the shipping traffic, such as oil, grease, and
suspended solids removal due to the shallow, near-shore intake.
The DAF was designed with in-line static mixing for the ferric chloride and sulfuric acid
dosing. It incorporates two stages of flocculation with total flocculation time of 13.2 minutes
at design flow. To generate the bubbles, a vertical packed tower saturator is utilized with
10% of the design flow being recycled and saturated at pressure of 5.5 bar within the
saturator. The DAF basin hydraulic loading rate is designed at 22 m/h and utilizes hydraulic
desludging for removal of the generated solids.
The plant was commissioned in October of 2011 and at that time was the world’s largest
SWRO plant using dissolved air flotation and pressurized inside-out ultrafiltration (UF) as
pretreatment.
While, this area of the Gulf does not have a history of severe, recurring HAB events,
occasional blooms have occurred. In the context of this case study, during plant operation in
May, 2012 and once again in April, 2014 a localized bloom near the port occurred (Figure
11.5.5). Turbidities of >30 NTU were seen in the source water, and total suspended solids
(TSS) increased up to 17 mg/L in May 2012 indicating a rather substantial bloom, as seen in
Figure 11.5.6. DAF reduced the turbidity and TSS to less than 3 NTU and 11 mg/L,
respectively (Im et al. 2012). These blooms came on without warning, but most likely
385
exhausted the nutrients needed to support the algal growth quite quickly, and lasted for
approximately 10 days.
Minor changes were made to the plant’s chemistry to accommodate this increase in organic
loading. This included increasing the coagulant feed from an average of 4 mg/L of ferric
chloride to 12 mg/L as well as decreasing the coagulated pH with a sulfuric acid dose of
40 mg/L during the bloom. In addition, the chemically enhanced backwash (CEB) frequency
of the UF system was increased. With these changes the DAF and UF pretreatment system
provided high-quality seawater feed with the SDI15 consistently less than 3.0 at all times to
the SWRO system. The Shuwaikh plant was therefore, able to maintain capacity throughout
the duration of the bloom allowing for the uninterrupted supply of potable water to the
residents.
Figure 11.5.3. Process flow diagram for Shuwaikh. (HP = High Pressure, LP = Low Pressure).
Figure 11.5.4. Shipping traffic passing by the Shuwaikh intake (pictured at left), heading
into the Shuwaikh Port (left and out of frame).
386
Figure 11.5.5. Seawater conditions during the algal bloom (top) showing the condition of the seawater
close to shore (intake pipe at right of image); bottom shows a close-up of the green-brown hue
produced by the algae.
Figure 11.5.6. DAF Performance in HAB conditions from May 5 to May 10 2012 (data from Im et al. 2012).
387
The thick floating blanket formed in the DAF cell during a bloom is shown in Figure 11.5.7.
The DAF process performed well during the blooms in Shuwaikh, but the severity of the
bloom, type and concentration of algae, and the percent algal removal from the system were
not determined.
Figure 11.5.7. Floc blanket on the surface of the DAF unit.

As an example of the UF performance, the operating data from one UF skid is shown before,
during and after the algal bloom event (May 5-10, 2012) in Figure 11.5.8
(http://xflow.pentair.com/en/case-studies/shuwaikh, 2016). The top line shows in membrane
permeability over time. The bottom line shows development of transmembrane pressure
(TMP) of the UF unit. The UF performance data confirms that the UF skids operated during
the bloom event, which allowed continuous operation of the SWRO desalination plant.
Figure 11.5.8. UF permeability (top line) and transmembrane pressure (bottom line) during a bloom in May
2012. Figure modified from: http://xflow.pentair.com/en/case-studies/shuwaikh.
388
11.5.2 Laboratory and Pilot Scale DAF Testing

To confirm the performance of the full-scale DAF system during a HAB event, multiple
HAB species were cultivated at Woods Hole Oceanographic Institution for DAF bench scale
testing. These species included Cochlodinium polykrikoides, the algal species that caused the
2008 - 2009 bloom in the Gulf. Different coagulation techniques were investigated to help
lower the high charge-neutralization demand of the algal cells. The most promising
techniques found to date are oxidation with sodium hypochlorite (typically less than 1 mg/L)
and adjustment of the pH prior to coagulation. These techniques lowered the coagulant
demand while improving cell removal by a few percentage points. Following laboratory
bench tests, the process was tested on a naturally occurring Cochlodinium bloom in Buzzards
Bay, MA in 2011 (14,000 cells/mL), with over 99% of the algae removed with the DAF
process.
The most interesting pilot study conducted to date was a simulated algal challenge test
conducted for the Fujairah I expansion in Fujairah, UAE. To demonstrate the effectiveness of
the DAF design which now feeds the existing Fujairah I filtration system and the ongoing
expansion of the plant, a pilot study with up to 100,000 cells/mL in the source water was
required as part of the contract. To accomplish this test with an algal concentration
approximately 4-5 times that present in the 2008- 2009 Gulf bloom, a cultivated alga was
utilized. The DAF pilot consistently removed over 95% of the algal cells and at times
exceeded 99% removal, as seen in Figure 11.5.9. The DAF pilot was followed by a gravity
media filtration pilot which utilized the current Fujairah I plant’s media profile. Interestingly
when the algae was introduced to the DAF system, the downstream filtration system Silt
Density Index (SDI) actually improved and filter head loss accumulation decreased. This is
most likely due to the algae acting as a sweep floc to incorporate colloidal material in the
source water.
Figure 11.5.9. Samples of feedwater (left) and DAF-treated water (right) from the challenge test at the
Fujairah I pilot plant using cultured algae.
389
Several pilots have also been conducted on naturally occurring blooms including a pilot at the
University of Texas Marine Science Institute at Port Aransas, Texas during a bloom of
Karenia brevis. Also a pilot study was conducted in Antofagasta Bay in Chile which
experiences a prolonged HAB event yearly, lasting at times up to 3 to 4 months due to the
bay’s currents which serve to feed nutrients to the algae continually over the summer months.
This prolonged bloom tends to support the growth of many algal species. The results from
each of the pilot tests on the different species of algae all showed greater than 95% removal,
suggesting that DAF would be equally effective on differing strains of algae and should be
considered a robust process for pretreatment around the world when an uninterrupted supply
from open intake systems needs to be assured.
11.5.3 Conclusions
DAF effectively removed cells during a HAB bloom in Kuwait, producing a stable and dense
floc blanket and effectively removed elevated levels of TSS and turbidity. The UF performed
well during the bloom, with no TMP spikes during this period. After optimization of the
ferric chloride dose, the DAF and UF pretreatment system provided high-quality seawater
feed with the SDI15 consistently less than 3.0 at all times to the SWRO system.
The DAF-UF pretreatment system at the Shuwaikh Desalination Plant provided a very stable
and effective system for providing RO feedwater during a HAB bloom. The DAF system has
also been tested and its effectiveness demonstrated through pilot studies on multiple other HAB
species in the laboratory and the field.
11.5.4 References
Im, J., Park, K., and Yim, W. 2012. Doosan as Pioneer in Supplying and Operating Large-
scale RO Desalination Plant at Shuwaikh on Coast of Kuwait. In: Proceedings of the
International Water Association World Water Congress & Exhibition, Busan, South
Korea.
390
11.6 LA CHIMBA, ANTOFAGASTA, CHILE – OXYGEN DEPLETION

AND HYDROGEN SULFIDE GAS MITIGATION DUE TO
HARMFUL ALGAL BLOOMS
Walter Cerda Acuña1, Carlos Jorquera Gonzalez1, and Victor Gutierrez Aqueveque1
1
Figure 11.6.1. Left and top right - Geographic location of the La Chimba desalination plant. Photo:
maphill.com Prepared by F.Knops. Bottom right – Aerial view of the La Chimba desalination plant. Photo:
Google Earth.
391
Table 11.6.1. Overview of La Chimba desalination plant.
Plant/Project Name
Location La Chimba, Antofagasta, Chile
Total Production Capacity (m3/d) 52,000
SWRO recovery (%) 52
Commissioning date 2002-2003
Intake
Feedwater source Pacific Ocean, Moreno Bay
Intake type Deep water intake
Intake description Intake depth: 25 m, distance from shore: 400 m
Intake screening Offshore open intake with 8 screens of 1 m2
protected with trellised boxes of fiber-reinforced
plastic
Chlorination Continuous chlorination installed but not used
Strategy, dose rate
Online raw water monitoring Conductivity, temperature, pH, turbidity, dissolved
oxygen (DO), ORP
Discrete raw water analysis Chlorophyll-a, algal counts, DO, pH and ORP
relevant to HABs
Pretreatment
Process description Pressure multimedia filtration, 5 µm cartridge
filtration
Chemical dosing 1ppm antiscalant; sulfuric acid to pH 5 for sulfate

reducing bacteria control
conditions
Temperature range (˚C) 14 – 19 14
Salinity range (TDS mg/L) 36,000 36,000
Conductivity (mS/cm) 53.0 52.8
Total Suspended Solids (mg/L) 6.0 unknown
SDI (15 minute interval)(%/min) 5.0 unknown
Turbidity (NTU) <1 unknown
Organic Matter
TOC/DOC (mg/L), TEP,
biopolymers etc.
Algal cell count (cells/L) unknown 11.3 Million
Algal species Prorocentrum micans
Chlorophyll-a (µg/L) 23 65-109
Additional relevant water quality Low DO (0.5 - 2mg/L),
parameters H2S 50mg/L max
DMF Filter rate m/h 7.6 7.6
RO flux (L/m2h) 11.6 11.6
392
11.6.1 Introduction
“Red tide” or harmful algal bloom (HAB) phenomena are a recurrent issue off the coasts of
Antofagasta, Chile, due to blooms of phytoplankton that, depending upon environmental
conditions, can vary in severity. The presence of phytoplankton in coastal waters can
originate from fluctuations in essential nutrients (nitrate, phosphate, and silicate) as a result
of surface runoff and/or coastal hydrographic phenomena and wind conditions.
Depending upon the species present in the algal bloom, these effects can be quite varied,
from innocuous to lethal. On some occasions, the high rates of nutrients and phytoplankton
biomass produced during these blooms greatly exceeds the natural capacity of the marine
environment to assimilate the phytoplankton growth, generating conditions of eutrophication.
Under this condition, the degradation of the organic material causes a reduction in dissolved
oxygen levels, resulting in the generation of noxious gases, such as methane and hydrogen
sulfide.
Given the geographic conditions of the location of the La Chimba Desalination Plant, it is
regularly affected by the presence of HABs. From 1964 to 1999, 41 algal bloom events were
recorded in the area known as Moreno Bay. On 23 occasions, the cause was the
dinoflagellate Prorocentrum micans, the same species that was dominant in a major bloom in
March 2011. Figure 11.6.2 presents a historical series of cell counts in the Bay of
Antofagasta from 1970. Notable among these events is the bloom of March of 2011 with a
peak cell density of over 11 million cells/L, five times greater than the maximum historical
level recorded in prior years.
Figure 11.6.2. Algal blooms between 1970 and 2012 in the Bay of Antofagasta.
11.6.2 Presence of hydrogen sulfide gas in a desalination plant

The presence of hydrogen sulfide (H2S) gas in a seawater desalination plant is uncommon.
More typically, information in the literature related to hydrogen sulfide generation in SWRO
plants is as a result of overdosing sodium metabisulfite to neutralize chlorine added in the
pretreatment. Sodium metabisulfite not only neutralizes the chlorine, but also, due to its
excess, reduces oxygen levels and therefore generates ideal conditions for the appearance of
sulfate-reducing bacteria (SRB).
393
There is little information related to desalination plants where hydrogen sulfide gas is
naturally present, as during very intense algal blooms; some plants may elect to cease
operations until the algal bloom activity is reduced.
11.6.2.1 Generation of hydrogen sulfide gas in the seawater
In the presence of very intense algal blooms, with water retained in the bay and the presence
of a thermocline that impedes mixing and oxygenation of seawater, the appearance of sulfate
reducing bacteria is quite likely. Oxygenation of the water is limited and the death and
decomposition of organic matter leads to hypoxic (<0.5 mg/l DO) and even anoxic (absence
of DO) conditions in the deepest zones, which favors the proliferation of SRB, a group of
totally anaerobic organisms that use the sulfate ion (SO42-) as an electron receptor in their
energy metabolism. They are distributed in sulfate-rich anoxic sediments, such as marine
sediments. In marine ecosystems, a good part of the metabolic activity occurs in the
oxic/anoxic interface and also in the deeper and anoxic zones of the sediment. This process is
illustrated in Figure 11.6.3.
Figure 11.6.3. Diagram of sulfate reducing bacteria metabolism.
11.6.2.2 Hydrogen sulfide gas in La Chimba desalination plant

Figure 11.6.4 shows the counts of SRB in all unit processes of La Chimba during the algal
bloom in March 2011. The seawater intake had the highest concentration of SRB
(700,000 CFU/mL), which strongly contrasts with the 5000 CFU/mL at the sand filters and
no presence of SRB in the RO stage. The difference between internal (inside the intake
tower) and external intake (outside the intake tower, or surrounding ocean) is noted in Figure
11.6.4 to demonstrate that the bacterial concentration originates in the ocean as opposed to
inside the intake tower due to poor maintenance. This indicates that there was no operational
problem, but rather the problem originated from external sources in the Bay of Antofagasta.
11.6.3 Operation of La Chimba Plant in the presence of hydrogen sulfide gas

At first it was thought that the appearance of hydrogen sulfide gas in the plant was caused by
an operating problem or by local contamination. The count of SRB demonstrated that the
origin of the problem was the source seawater (seawater intake). Notwithstanding the origin
of the phenomenon, and given that this desalination plant supplies 60% of the potable water
for the city of Antofagasta, Aguas Antofagasta, who owns and operates the plant, had to
implement a quick and effective solution that would eliminate the hydrogen sulfide gas
present in the potable water.
394
Figure 11.6.4. Presence of SRB in the La Chimba intake and desalination plant processes,
March 2011. CFU: Colony Forming Unit.
Potential strategies to eliminate hydrogen sulfide gas present in the source seawater can be
classed as aerobic and anaerobic solutions. There are three aerobic solutions:
• Elimination of H2S by degassing at the entry to the plant;
• Precipitation of sulfides to sulfates and decantation; and
• Oxidation of sulfur derivatives and decantation
The aerobic alternatives are risky due to the probability that elemental sulfur is generated by
exposure to oxidizing agents, such as dissolved oxygen or chlorine. This elemental sulfur can
then reach the membranes to cause damage and thus a loss of salt rejection. Therefore, the
alternative anaerobic solution was pursued whereby pretreatment was maintained under
anaerobic conditions, with no contact with oxygen or other oxidants (Fethi 2003). A system
was implemented that enabled maintenance of the presence of sulfur as dissolved sulfuric
acid through the SWRO process and eliminating it in the final stage with a stripping system
and through the controlled addition of chlorine. This solution involved the following:
1. Installation of dissolved oxygen sensors in the seawater intake pit for early
detection of conditions that could favor the presence of H2S in the seawater;
2. Addition of H2SO4 to reduce the pH of the seawater, to stabilize the H2S present,
and eliminate the probability of generation of H2S in the plant pretreatment;
3. Installation of H2S sensors on the outlet from the dolomite post-treatment filters,
product water tank access hatch and distribution of potable water;
4. Implementation of stripping with air in the product water tank, so as to displace
the H2S present in the product water;
5. Direct dose of sodium hypochlorite in the product water tank to eliminate traces
of H2S by oxidation to sulfate; and
6. Contingency procedure to operate in seawater conditions with a low level of
dissolved oxygen.
In the following diagram (Figure 11.6.5), existing water quality variables measured in the
designed plant line are; ORP, CE, Cl, pH and TS, while new variables incorporated are
labeled as *H2S and *O2. The intervention points for eliminating the sulfuric acid were the
hypochlorite (hypo) and air added to the potable water storage tank.
395
Figure 11.6.5. Diagram of operation during H2S events in La Chimba desalination plant (CE = Conductivity;
ORP = oxidation and reduction potential; Cl = Free Chlorine; TS = total dissolved solids).
11.6.3.1 Analysis and discussion of the application of the anaerobic method
The application of the anaerobic method to eliminate hydrogen sulfide gas is examined in the
following sections. The measures adopted, from the implementation of the dissolved oxygen
measurement systems, portable hydrogen sulfide gas sensors and the installation of aeration
in the product water tank, were very effective. The traces of hydrogen sulfide gas in the
potable water were almost completely eliminated. Nonetheless, despite the very low
concentration of hydrogen sulfide (< 0.3 mg/L) it produced an unpleasant smell.
11.6.3.2 Water quality during 2011 bloom
Dissolved oxygen levels during the 2011 bloom are presented in Figure 11.6.6 where it was
observed that the seawater tank was hypoxic and even very close to anoxic. During this
period, the levels of hydrogen sulfide gas concentrations eliminated from the product water
tank reached a maximum level of 50 mg/L (Figure 11.6.7). Nonetheless, the anaerobic
alternative was highly efficient and the presence of hydrogen sulfide gas was not detected at
any time in the transmission of potable water to the tanks in the city of Antofagasta.
Figure 11.6.6. Concentration of dissolved oxygen (ppm, y-axis) in the seawater tank, March 2011.
396
Figure 11.6.7. Concentration of H2S (ppm, y-axis) extracted from product water tank.
11.6.3.3 Prevention strategy
The strategy implemented to prevent and/or mitigate the effects of atypical phenomena seen in
the coastal area of the city of Antofagasta in March 2011 has two complementary aspects:
• Early prevention (discussed below) and
• Mitigation
Early prevention aspects can indicate the presence of conditions that generate the presence of
sulfate-reducing bacteria in seawater and/or minimize the probability of their appearance
during operations. The following were implemented:
• Seawater dissolved oxygen sensors;
• Monitoring of SRB bacteria during the summer months (December to March);
• Satellite monitoring of chlorophyll concentration and eutrophication status based on
chlorophyll (Bricker et al. 2003) during summer months (Figure 11.6.8);
• Monitoring of pH, temperature, dissolved oxygen and redox potential of the water
column in the seawater abstraction zone during summer;
• Dredging of the seawater suction pond twice per year;
• Dredging the seawater intake tower every two years; and
• Continuous inline dosing of sulfuric acid in the pretreatment during summer to reduce
the pH to levels where SRB cannot survive.
Figure 11.6.8. Satellite image of chlorophyll (left) and the risk of algal blooms based on eutrophication status
(right), 1 February 2013. Low eutrophication (Bajo): ≤ 5 ppb; Medium eutrophication (Medio): > 5 ppb and ≤ 20
ppb; High eutrophication (Alto):> 20 ppb and ≤ 60 ppb; Hyper eutrophication (Hipereutrof):> 60 ppb.
397
In terms of mitigation, when dissolved oxygen concentrations <2 mg/L are detected,
operational procedures are activated based on the anaerobic solution described above, with
additional actions that minimize the production of H2S gas. These actions are as follows:
• Dosing sulfuric acid prior to the multimedia filters to lower pH to 4 - 5. Growth of
SRB bacteria is prevented in multimedia and cartridge filters and/or RO membranes;
• Activate the aeration system in the product water tank to strip out H2S; and
• Dosing sodium hypochlorite directly to the product water tank, transforming the
traces of H2S into sulfates.
Additional actions:
• Permanently washing the multimedia filters (even when the pressure differential is
low) to remove biological matter and reduce the probability of its decomposition;
• Change out of filter cartridges when differential pressure is >0.7 kg/cm2;
• Maintain a minimum level of water in the product water tank, equivalent to 90% of its
total capacity, so the reaction of the H2S with air is effective;
• Monitoring of pH, odor and dissolved oxygen in all stages of the process, to detect,
the sources of generation of H2S gas (if possible); and
• Hourly monitoring of the product water quality for pH, residual chlorine, turbidity,
odor (by observation), taste and H2S.
11.6.4 Conclusions
• Red tide is a recurrent phenomenon on the coast of Antofagasta; however, the episode
in March 2011 was the most intense recorded to date, leading to a major problem with
hypoxia and H2S gas.
• The anaerobic solution described above was successful. During the bloom, DO in the
seawater was at hypoxic levels until mid-March 2011, with the presence of H2S, but
La Chimba was able to operate with no impacts on potable water quality.
• The measures adopted, from the implementation of DO measurement systems,
portable H2S gas sensors, and the installation of aeration in the product water tank
were effective. Other actions included dosing sulfuric acid prior to the multimedia
filters to lower pH, and dosing sodium hypochlorite directly to the product water tank.
• The presence of H2S in seawater is an exceptional event; nevertheless, desalination
plants that supply potable water to cities can have effective operating systems under
these conditions.
• As it is very likely that this phenomenon will occur again, Aguas Antofagasta
desalination plants have been equipped with instrumentation in anticipation of the
presence of H2S in the source intake water.
• Monitoring of phytoplankton communities, bacteria, and the determination of
chlorophyll in the seawater, especially in periods when red tides appear, are good
approaches to detect the appearance of conditions that could favor the proliferation of
sulfate reducing bacteria and thereby provide an early warning of this phenomenon.
11.6.5 References
Bricker, S. B., Ferreira, J. G., and Simas, T. 2003. An integrated methodology for assessment
of estuarine trophic status. Ecological Modelling 169(1), 39–60.
Fethi, K. 2003. The 15,000 m3/d RO Djerba Plant: Desalination of Brackish Water with a
High Hydrogen Sulfide Concentration. In: Proceedings of the International
Desalination Association World Congress, The Bahamas.
398
11.7 MEJILLONES, CHILE – OPERATION OF THE

ULTRAFILTRATION SYSTEM DURING HARMFUL ALGAL
BLOOMS AT THE GAS ATACAMA SWRO PLANT
Frans Knops1 and Alejandro Sturnilio2

1
X-FlowBV/Pentair Water Process Technology BV, Enschede, the Netherlands
2
RWL Water Unitek, Mar del Plata, Argentina
Figure 11.7.1. Intake to the Gas Atacama SWRO Plant (top left and bottom) and the algal bloom discussed in
this case study. Bloom photo: Aquacien Consultoria Maritima.
399
Table 11.7.1. Overview of Gas Atacama desalination plant.

Plant/Project Name
Location Mejillones, Chile
Primary product water use Industrial (boiler feed to power plant, NOx reduction)
Desalination technology SWRO first pass, BWRO second pass, CEDI (polishing)
Total production capacity 1,200 m3/d (plant #1), DAF and media filters
(m3/d) 2,600m3/d (plant #2), UF pretreatment
Commissioning date 1995 (plant #1), 2010 (plant #2)
Intake
Feedwater source Pacific Ocean
Intake type Open intake (existing, shared with intake of power plant)
Intake description Intake 5 m below sea level, 3 m above ocean floor
Intake screening unknown
(shock) chlorination Hypochlorite shock dosing 3 mg/l (frequency unknown)
Online raw water TDS (measured as conductivity), turbidity, temperature
monitoring
Discrete raw water analysis pH, dissolved oxygen
Pretreatment (plant #2)
Process description Automatic backwashable screens (200 µm), UF
(pressurized, inside-out filtration mode), cartridge
filtration (5 µm)
Chemical dosing UF feed coagulant – FeCl3 (max 2 ppm as Fe), UF
cleaning chemicals (NaOCl, H2SO4, NaOH), SWRO
antiscalant (5 ppm), SWRO Na2H2SO5 (5 mg/L)
Feedwater design parameters Feedwater during
bloom conditions
Temperature range (˚C) 11 – 22 18 – 20
Salinity range (TDS mg/L) 33,000 No change
pH 7.6 – 8.8 No change
DO (at 20 m depth) 1.8 – 5.4 mg/L (24 – 71%) No change
DO (at 5 m depth) 6.7 – 10.4 mg/L (92 – No change
148%)
Total suspended solids - -
(mg/l)
SDI (%/min) - SDI5 > 18
Turbidity (NTU) 10 35
Organic Matter - -
TOC/DOC (mg/L)
Algal cell count (cells/L) ~25,000 400,000 – 1,300,000
Algal species Leptocylindrus danicus Leptocylindrus danicus
(Diatom) Prorocentrum graciles
Chlorophyll-a (µg/L) < 20 40 – 120
Additional relevant ORP 140 – 220 mV ORP 20 – 75 mV
parameters
Desalination Design Bloom conditions
Filter rates N/A N/A
UF flux (L/m2h) 75 75
2
RO flux (L/m h) - -
400
11.7.1 Introduction
Gas Atacama operates a combined-cycle thermal power plant in Mejillones, Chile (see
location map Figure 11.7.1), with an installed capacity of 780 MW. The plant uses seawater
as its only source of water. Two SWRO plants, with their corresponding pre- and post-
treatment, are installed in parallel and are operated simultaneously. Each of the two plants
uses different pretreatment technologies, so a comparison can be made between them.
The existing cooling water intake for the power plant is an open intake, without any
treatment. The flow of water used for cooling is much higher than the flow needed to feed the
desalination plants. As the intake was capable of handling this additional flow, using the
existing cooling water intake was therefore considered the most viable option.
The source water is characterized by seasonal variations in turbidity; during autumn and
winter it ranges from 1 to 5 NTU, while during spring and summer, the range increases from
3 to 35 NTU. Water temperature ranges widely from 11 to 22°C, resulting in a significant
variation in SWRO operating pressure between winter and summer. It also affects the
microbiological conditions of seawater. Another important issue is the occasional presence of
algal blooms, responsible for SWRO biofouling on the membrane surface that can require
frequent chemical cleaning. Such algal blooms occur one to three times per year with
durations of up to a week. As the power plant uses SWRO desalination as the sole source of
water, a robust pretreatment process is required that can reliably operate through algal bloom
events with minimal down time to ensure high plant availability.
Plant 1
Installed in 1995, Plant 1 produces 50 m3/h of demineralized water reaching a conductivity of
less than 0.1 µS/cm. The main stages are:
• Dissolved Air Flotation (DAF) with upfront coagulation;
• Pressurized depth filters;
• Seawater Reverse Osmosis with Pelton Turbine energy recovery;
• Cation ion exchange;
• Forced draft degasifier;
• Anion ion exchange; and
• Mix Bed ion exchange
Plant 2
Installed in 2010 by Unitek (now RWL Water), Plant 2 produces 108 m3/h (0.7 MGD) of
demineralized water reaching a conductivity of less than 0.1 µS/cm. The main steps are:
• In line coagulation with FeCl3;
• Ultrafiltration (UF) (Figure 11.7.2);
• SWRO shown in (Figure 11.7.2) with ERI pressure exchanger (not visible);
• Brackish Water Reverse Osmosis (BWRO); and
• Continuous Electrodeionization (CEDI)
Note: UF backwash waste and SWRO brine are blended and discharged to the ocean without
further treatment. BWRO and CEDI concentrate streams are recycled to SWRO inlet.
401
Figure 11.7.2. Ultrafiltration (left) and SWRO (right) trains at the Gas Atacama Plant.

11.7.2 Water quality
Only limited water quality information (turbidity, temperature, and TDS) is available, as the
plant operators do not monitor raw water on a regular basis. A nearby facility (Aguas de
Antofagasta) at Antofagasta conducted an extensive study of seawater quality in Antofagasta
during a HAB event in January/February 2013, measuring chlorophyll-a, dissolved oxygen,
pH, algal cell counts, Secchi depth, and ORP at 0, 5, 10, 15 and 20 m depths, (Atacama Agua
y Tecnología Ltda. 5o, 9o and 11o, 2013). The monitoring program covered the coast line
approximately 40 km to the north and to the south of Antofagasta. The distance from
Antofagasta to Mejillones is 50 km, and both Mejillones and Antofagasta are located in
sheltered bays. Therefore the results obtained are considered representative for the conditions
in Mejillones.
Although, the algal bloom event occurred in the Southern Hemisphere summer season, with
elevated seawater temperatures (18–20 oC), there was no direct relationship between seawater
temperature and algal counts. The level of dissolved oxygen varied between 1.8 and 5.4 mg/L
(24–71%) at 20 m depth and 6.7 and 10.4 mg/L (92–148%) at 5 m water depth.
No anoxic conditions were observed, but the level of eutrophication and the presence of
Prorocentrum graciles suggests that anoxic conditions might occur during other HAB events
as this species has been known to deplete nutrients and cause anoxia (Cassis et al. 2012). pH
varied from 8.0 to 8.9 at the seawater surface and from 7.6 to 8.0 at 20 m depth. Typically it
was 0.5 to 1 unit lower at 20 m depth, compared to the surface. One observation (on 27
December 2012) showed no variation (pH 8.0 for the entire water column).
Satellite images of chlorophyll-a data were provided by NPOES and MODIS satellites
processed with SeaDAS 6.3 software (Figure 11.7.3). Four images of chlorophyll-a were
collected every week for 2 ½ months to monitor chlorophyll-a. The following classification
based on chlorophyll-a concentration was used (based on Bricker et al. 2003) to assess the
risk for algal blooms:
• Low eutrophication: ≤ 5 ppb
• Medium eutrophication: > 5 ppb and ≤ 20 ppb
• High eutrophication: > 20 ppb and ≤ 60 ppb
• Hyper eutrophication: > 60 ppb
402
Figure 11.7.3. Satellite image with chlorophyll-a concentration on 17th of January 2013
adjacent to La Chimba.
Figure 11.7.4 gives a time series of the chlorophyll-a concentration during this period. Three
individual readings were averaged to a single value. Background chlorophyll-a
concentrations preceding and following the chlorophyll-a spike (up to 120 µg/l) during the
bloom were typically around 10 - 40 µg/L, which is higher than the initial design envelope
(See Table 11.7.1). From January 14 to around the 4 February, chlorophyll-a levels were in
the high- to hyper-eutrophic zones.
Seawater samples were taken during this study and phytoplankton counted on three separate
occasions: 14 January 2013, 14 February 2013 and 28 February 2013 (Figure 11.7.5). On 14
January, the algae count was 1.3 million cells/L with the highest count close to the sea
surface (at 1 m depth) and only 11,000 cells/L at 15 m depth. One dinoflagellate species
accounted for 99.8% of the bloom at that time (Prorocentrum graciles). On 14 February, the
total algal count was 417,000 cells/L at 1 m depth and 160,000 cells/L at 15 m depth. One
diatom species (Leptocylindrus danicus) accounted for 99% of the population at that time.
The dinoflagellate that had previously been observed had almost completely disappeared,
accounting for only 0.2% of total count. Characteristics of these two species are summarized
in Appendix 1. Both are not known to produce toxins and are relatively large, especially
Leptocylindrus danicus, which forms long chains. Hence, intact cells should be well removed
by both pretreatment systems. The pore size of the UF membranes (25 nm) is two to three
orders of magnitude smaller than the typical size of the algae. The equivalent pore size of the
403
media filters would be less than one order of magnitude smaller than the size of the algal
cells, but still adequate.

Figure 11.7.4. Average chlorophyll-a concentration measured from satellite data measured at La Chimba during
the summer bloom in January – February 2013 and associated eutrophication classification.
Figure 11.7.5. Algal cell counts at various depths in January and February 2013.
On 28 February the total algal count was 27,000 cells/L at 1 m depth and 12,000 cells/L at 15 m.
The most abundant species was again Leptocylindrus danicus, but the population was more
diverse, with the most abundant species only accounting for 80% of the total count. Figure 11.7.5
provides a graphic overview of sampling results.
The chlorophyll-a measurements suggest that an algal bloom started on or around 21 January and
ended on or around 4 February. Seawater sampling showed elevated algae counts immediately
prior to (14 January) and after the bloom event (14 February). Unfortunately no samples were
taken during the period of highest chlorophyll-a. Extrapolating the data, it can be estimated that
algal counts may have reached a maximum of 2 to 5 million cells/L at 1 m depth during the peak
of the bloom.
404
During the same period, transparency and oxidation-reduction potential (ORP) were monitored as
well. Transparency was monitored by means of a Secchi disk. ORP was averaged from 5
individual samples: 0, 5, 10 15 and 20 m depth. Both data are plotted in Figure 11.7.6 along with
chlorophyll-a. Transparency is a good indicator for seawater turbidity. Bricker et al. (2003)
employs the following classification for turbidity:
• High turbidity Secchi disk depth < 1 m

• Medium turbidity Secchi disk depth ≥ 1 m and ≤ 3 m
• Low turbidity Secchi disk depth > 3 m
Turbidity (as measured by Secchi depth) showed a good correlation with chlorophyll-a
concentration at the start of the HAB event (around 15 January; Figure 11.7.6). After the
decrease in chlorophyll-a (10 February), however, it took another 7 – 10 days for the Secchi
transparency to reach a value of more than 2 m. This suggests that the end of the HAB event
resulted in a further water deterioration due to release of organics from decaying algae.
The ORP value monitored on 4 January was 29 mV (Figure 11.7.6). This value appeared to
be an anomaly, but was confirmed by multiple samples. Apart from this data point, there is a
good relationship between transparency and ORP. A drop in ORP indicates that oxygen
concentration is being reduced (assuming no other oxidant or reducing agent is present in the
water). Hence, it was concluded that ORP may be a good indicator for the occurrence of
HAB events.
Figure 11.7.6. ORP, Chlorophyll-a and seawater transparency (Secchi disk depth) measured at
La Chimba during the summer bloom in January – February 2013.
11.7.3 Operational data

The following data (Table 11.7.2) were provided by the end user of the plants (Knops et al.
2012; Vasini et al. 2013). No detailed plant operating settings were provided. Both plants
405
were operating at design capacity during all periods. HAB events in the older plant (SWRO
#1) were managed by additional SWRO cleanings and cartridge filter replacements. It should
be noted that SWRO Plant 1 is approximately 20 years old, and employs older designs for
DAF and SWRO. Up to 25 ppm of coagulant was required to keep the DAF system
operational.
In contrast, SWRO plant #2 did not require additional cleaning or cartridge filter
replacements. The ultrafiltration system in that plant provided sufficient pretreatment during
the algal blooms to enable trouble free operation of the SWRO system. A maximum of 2 ppm
coagulant was used directly in front of the UF system.
Table 11.7.2. SWRO plant design and operating parameters.
SWRO Plant 1 SWRO Plant 2
Pretreatment DAF, media filters and 5 µm 200 µm strainers, pressurized UF and
cartridge filters 5 µm cartridge filters
Coagulant dose 25 ppm FeCl3 in front of DAF 1 – 2 ppm FeCl3 in front of UF
Cartridge filter 15 – 30 days during winter No filters replaced in 2 years

replacement Once every 4 days during summer
SWRO replacement 165% over 5 years of operation, 0% over 2 year’s operation

frequency approximately 30% per year.
SWRO cleaning Once per month during winter No cleaning performed in 2 years
frequency 15 – 30 days during summer
11.7.4 Conclusions
• Satellite data providing chlorophyll-a concentrations were very useful, and indicated
that an algal bloom event occurred over approximately two weeks, from 21 January
2013 to the 5 February 2013.
• This bloom was confirmed by seawater sampling that documented elevated cell
counts immediately prior to (14 January) and after the event (14 February).
Unfortunately no samples were taken during the period of highest chlorophyll-a
concentration. Total algal counts are estimated to have reached a maximum of 2 –
5,000,000 cells/L during the peak of the bloom.
• Secchi disctransparency and ORP measurements showed a good correlation with
chlorophyll-a and cell count levels, and confirmed the occurrence of an algal bloom
event and a deterioration in water quality after the bloom due to decay of the algae.
• The bloom event consisted of two separate periods in which different species of algae
dominated. Initially Prorocentrum graciles (a dinoflagelate) dominated. Within four
weeks this species was completely replaced by Leptocylindrus danicus (a diatom).
• During the blooms, a single species prevailed whereas, during periods of low and
medium eutrophication, the algal population was more diverse.
• Two SWRO systems were operating on a common intake system. These two systems
had different pretreatment designs (DAF plus media filtration versus ultrafiltration)
and allowed for a comparison of pretreatment under identical operating conditions.
• Both SWRO systems operated at design capacity during the algal bloom events.
• The SWRO system that employs conventional pretreatment (DAF plus media filters)
did encounter significant operational issues: high coagulant dose, increased SWRO
406
membrane cleaning and cartridge filter replacement and eventually premature SWRO
membrane replacement. 12% of SWRO membranes were replaced within 18 months
of operation.
• The alternative system that employed UF did not require cartridge filter replacement
or SWRO membrane cleaning even when challenged by HAB events. During the first
two years of operation, no SWRO membranes had to be replaced. The lower
mechanical stress and chemical exposure suggest that extended SWRO lifetime can
be achieved.
11.7.5 References
Atacama Agua y Tecnología Ltda. 2013. Programa de vigilancia ambiental preventivo
Comunidades fitoplanctónicas, Período estival 2012 – 2013, sector aducción de agua de
mar Planta Desaladora la Chimba, Antofagasta, 5º , 9º, 11º Monitoreo semanal.
Bricker, S. B., Ferreira, J. G., and Simas, T. 2003. An integrated methodology for assessment
of estuarine trophic status. Ecological Modelling 169(1), 39-60.
Cassis, D. et al. Phyto'pedia., 2012.
http://www.eos.ubc.ca/research/phytoplankton/index.html accessed 1 September 2015.
Knops, F., Kahne, E., Garcia de la Mata, M., and Mendoza Fajardo, C. 2012. Seawater
Desalination of the Chilean Coast for Water Supply to the Mining Industry. In:
Proceedings of the Water in Mining Conference, Santiago, Chile.
Vasini, V., Garcia de la Mata, M., and Sturniolo, A. 2013. Two full-scale desal plants
operating side-by-side: results and comparison of conventional vs new technology. In:
Proceedings of the International Desalination Association World Congress on
Desalination and Water Reuse, Tianjin, China.
407
408
11.8 ANTOFAGASTA, CHILE - ABENGOA WATER

MICRO/ULTRAFILTRATION PRETREATMENT PILOT PLANT
Francisco Javier Bernaola1, Israel Amores1, Miguel Ramón1, Raquel Serrano1,

and Juan Arévalo1
1
Abengoa, Spain
Figure 11.8.1. Aerial photo of pilot plant area at La Chimba

Antofagasta (top). Photo: Google Earth). Intake area of the pilot
plant (bottom). Photo: Abengoa.
409
Table 11.8.1. Overview of the Abengoa Water multimembrane pilot plant.
Plant/Project Name Multimembrane - Antofagasta

Location Antofagasta (Chile)
Desalination Technology MF / UF pretreatment
Total Production Capacity 336
(m3/d)
Commissioning date May 2013
Intake
Feedwater source Pacific Ocean
Intake type Shore intake-submerged bar screen
Intake description Intake depth: 25 m, distance from shore: 400 m
Treatment
Process description Prefilter strainer plus two parallel membrane systems one MF
and UF both pressurized with outside-in configuration
Chemical dosing No chemicals added
Feedwater design parameters
Temperature range (˚C) 17-20 ºC
pH 7.5-8.5
Turbidity (NTU) 0.4-4
Organic Matter 3500 Measured once during the
TEP µg xanthan-eq/L algal bloom
Algal cell count (cells/L) ~up to 1.2 x106
Algal species Pseudo-nitzschia delicatissima
Ceratium fusus
Chlorophyll-a (µg/L) ≤ 50
Treatment Design (*) During normal conditions During bloom conditions
UF flux (L/m2h) 76 64
MF flux (L/m2h) 83 70
(*) Different fluxes as both MF and UF systems were designed to operate maintaining the same pressure.
11.8.1 Introduction
Low-pressure membranes are an emerging technology as a pretreatment option in seawater
desalination processes using reverse osmosis. Many advantages can be obtained when
microfiltration (MF) or ultrafiltration (UF) membranes are used instead of conventional
pretreatment such as:
• Positive barrier to particulates and pathogens;
• Lower operating costs due to significantly reduced RO membrane fouling and
cleaning frequency;
• Lower chemical requirement and extended RO membrane life;
• Reliable production of high quality RO feedwater regardless of raw water turbidity;
• Higher RO membrane flux; and
• Smaller footprint for both RO and pretreatment systems resulting in lower capital
costs.
410
It remains necessary to evaluate the performance of membrane pretreatment when faced with
challenging feedwater conditions such as algal bloom events, when the seawater composition
changes drastically.
The Chilean coast suffers red tides frequently and at the same time desalination plants are
increasingly being installed along the coast to solve water scarcity problems, providing a
water source to the cities and mines, mainly in the north, one of the driest deserts in the
world. Therefore, Abengoa Water conducted a pilot study in the Bay of Antofagasta, in
northern Chile, in which various MF and UF membrane technologies, processes and
configurations were tested as pretreatment for the seawater reverse osmosis (SWRO)
installation.
The MF and UF membranes both operated under pressure from a feed pump mainly in the
outside-in configuration whereby, feedwater is pushed through the membrane pores from the
shell side into the lumen of the hollow fibers. In the first stage of the pilot plant, one module
was tested in the inside-out configuration. During mechanical cleaning, reverse flow was
used, and air could also be introduced into the membrane modules in order to create
turbulence along the membrane surface to increase the efficiency of cleaning if required.
Rising air bubbles scour and clean the surface of the membrane fibers, maximizing
membrane cleaning to restore flux.
To assess the quality and capacity of membrane fouling with this feedwater, Abengoa
developed new measurement indexes, the AMFI1 and AMI2, and measured new parameters,
including, among others, the content of TEP (Transparent Exopolymer Particles) - gelatinous
particles composed mainly of negatively charged polysaccharides that are very sticky and can
dramatically clog filters and membranes (see Chapter 2 and Appendix 3).
The multi-membrane Abengoa Water MF/UF pilot plant is shown in Figure 11.8.2 with a
block diagram of the pilot plant given in Figure 11.8.3. The configuration allowed collection
of operational as well as
analytical data to evaluate
the performance of
different membranes tested
at the same time and
treating the same water.
This is a very robust
system as it was designed
to treat seawater of varying
quality with different types
of membranes. The pilot
plant is designed to work
at the same time with two
different modules, but both
modules must be working
either in outside-in or
inside-out filtration mode.
Figure 11.8.2. Multi-membrane pilot plant.
1
AMFI is based on the MFI procedure but simplified to allow automatic measurement.
2
AMI is a procedure using a hollow fiber membrane at lab scale to evaluate clogging capacity.
411
Figure 11.8.3. Process schematic for the Abengoa multi-membrane pilot plant operating in Antofagasta.
The objectives of the Antofagasta pilot study were to demonstrate the feasibility of UF/MF
for SWRO pretreatment during normal feedwater conditions and algal bloom events and to
optimise the following operational parameters:
• UF/MF flux;
• UF/MF backwash frequency and parameters;
• chemical cleaning frequency and duration; and
• cleaning efficiency and procedures
The trials were carried out in two stages. In the first, five different membrane modules were
tested in the rig with various pore sizes and material (summarized in Table 11.8.2) in order to
identify membranes for further long term testing in the second stage. In Figure 11.8.4, the
behavior of different modules is shown along with AFF, a clogging index, showing the
fouling kinetic of each membrane. No direct relationship was found between membrane pore
size and clogging capacity. Therefore, clogging capacity is more likely to be related to
membrane manufacture and the affinity of the membrane for the contaminant, for example
membrane hydrophilicity. Membranes P0046 and P0035 were selected based on their lower
AFF value (fouling capacity at different flux) as shown in Figure 11.8.4 and a AMFI below a
maximum value that was considered acceptable in an RO pretreatment plant).
Table 11.8.2. Characteristics of MF and UF membranes used in the first stage of the pilot
study.
Pressure
Module Pore Size Flow mode decay test
Material
Number (µm) (PDT)
(mbar/min)
PVDF
F1223 0.02 Outside-In 1.4
reinforced
P0033 PVDF 0.03 Outside-In 0.9
P0047 PESM 0.02 Inside-Out 1.1
412
0.5
0.45
0.4
0.35
AFF (mbar/s)
0.3 F1223
0.25 P0046
0.2 P0035
P0033
0.15
P0047
0.1
0.05
0
60 80 100 120 140 160 180 200
J (LMH)
Figure 11.8.4. AFF Index obtained for the MF and UF membranes used in the first stage of the pilot plant
study. P0046 and P0035 membranes were selected for the second stage of the study.
In the second stage of the study, the pilot plant, was equipped with two commercial modules
and operated in parallel at the same AFF value, so that each module had its own specific
operating flux.
11.8.2 Results
The pilot plant achieved excellent results testing different MF/UF technologies for normal
seawater quality through comprehensive monitoring of the operating conditions of the
system, the frequency of chemical cleaning, resulting in high filtrate quality, high flow and
good membrane fouling management. A recovery of 96% was achieved.
The pilot plant also faced challenging situations where algal bloom counts reached 1.2 x106
cells/L. During this time the rate of membrane fouling grew exponentially compared to the
usual rate, necessitating modification of plant operating conditions. The frequency of
cleaning was also increased. It was possible to maintain stable control of the process during
the bloom without plant shutdowns and without the addition of chemical reagents such as
coagulants, so that chemical consumption was minimized. Disposal of coagulant residuals
was avoided, decreasing the environmental impact of the process while achieving high
quality feedwater for SWRO.
The pilot plant faced some algal bloom issues in the warm periods of the year (December -
March), when the cell counts increased from a background level of ~ 6 x 103 to > 1.2 x 106
cells/L measured at the intake basin (Figure 11.8.5). TEP concentrations were found to
increase from ~ 700 µg xeq/L to 3500 µg x-eq/L during this period.
Based on the concentration of chlorophyll-a, the eutrophication status and the risk of algal
blooms was classified as medium/high (Bricker et al. (2003); Gas Atacama case study). The
two most abundant species in the blooms were Pseudo-nitzschia delicatissima (diatom) and
Ceratium fusus (dinoflagellate), both are non-toxic, but can foul membranes.
413
1,400,000
1,200,000
1,000,000
Cells/L
800,000
600,000
400,000
200,000
0
1-‐‑Jan-‐‑14
12-‐‑Feb-‐‑14
19-‐‑Feb-‐‑14
15-‐‑Jan-‐‑14
21-‐‑Jan-‐‑14
29-‐‑Jan-‐‑14
5-‐‑Feb-‐‑14
Figure 11.8.5. Total cell counts in the multi-membrane pilot plant feedwater showing an increase during an
algal bloom.
The presence of these microscopic algae caused problems in filtration performance as the
suspended solids concentration increased due to the high cell numbers and organic
compounds (TEP) associated with the bloom. This resulted in a rapid two-fold increase in
transmembrane pressure. TSS measured in this period increased from 6 – 8 ppm during non-
bloom periods up to an average value of 40 ppm, with a peak of 76 ppm during the bloom,
measured downstream of the pre-filter.
Figure 11.8.6 shows how the membrane fouling kinetics significantly increased during the
time that algal blooms were more frequent, that is during the summer months with higher
temperature in the southern hemisphere.
4 4.5
3.5 4
3.5
3
3
2.5
Flow (m3/h)
2.5
Kinetic
2
2
1.5
1.5
1
1
0.5 0.5
0 0
0 5 10 15 20 25 30 35
Data Point
Kinetic Flow (m3/h)
Figure 11.8.6. Membrane fouling kinetic during the 2014 algal bloom event. Data points were taken at
uneven times and 34 data points were taken between early November 2013 to mid-March 2014.
This increase in transmembrane pressure made it impossible to operate the installation under
the normal operating conditions. Therefore it was necessary to adapt the process to the
feedwater quality changes during the bloom. To allow for continued membrane operation
414
during the algal bloom it was necessary to adjust the flux, filtration mode and cleaning
frequency as described further below.
It was necessary to reduce the flux by 16% on average to return the fouling kinetics to its
initial value, where the fouling was controlled and the permeability restored using
mechanical cleaning or simple backwashing.
The increase in suspended solids concentration measured in bloom periods also made it
necessary to adapt the filtration process. In normal conditions the UF operated in dead-end
mode with all the seawater pumped across the membrane. During bloom events the water
composition was more complex and the higher suspended solids made it more attractive to
operate in cross flow, where part of the water pumped scours the outside membrane surface
and removes the excess of solids off the membrane. In this case, a 10% recirculation ratio
was used. While operating in cross flow mode may require more energy, the average pressure
was less and any increase in energy consumption would be more related to the loss of
recovery than from recirculation.
Finally, the frequency of chemical enhanced backwash (CEB) was greatly increased during
algal blooms. The increase in TSS and a much more complex aqueous seawater matrix made
more frequent cleaning necessary. The initial plant configuration was fixed with a CEB
period of 30 hours, but during the bloom event and prior to changing flux the CEB period
was reduced down to 7-8 hours (Figure 11.8.7). Following the optimization of flux during
bloom conditions the CEB was set to 26 hours to achieve stable operation.
The CEB and clean in place (CIP) were carried out with the same chemicals and
concentration, but the duration for each procedure was different. CIP comprised of a chlor-
alkali phase followed by an acid phase.
50
Hours Between Chemical Cleanings
45
40
35
30
25
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Data Point
Figure 11.8.7. The duration between chemical cleaning (CEB) during normal feedwater conditions and during
the algal bloom. Data points were taken between early November 2013 and late February 2014 on 23 occasions.
From data point 21, both hypochlorite and sodium hydroxide were used during the CEB.
The increased cleaning frequency resulted in much higher operational costs, due to the higher
chemical consumption. To lower the chemical consumption during algal blooms, the pH
value of the chemical cleaning solution was increased to values around 12 compared to 9
during normal operation. This resulted in chemical cleaning restoring UF recovery to 94%.
415
11.8.3 Conclusions
• Based on the concentration of chlorophyll-a, the phytoplankton bloom experienced at
the pilot plant can be classified as medium/high, using the classifications of Bricker et
al. (2003).
• The pilot plant initially tested inside-out UF membranes but thereafter tested outside-
in. Conclusions therefore relate to outside-in membranes.
• The pilot plant achieved an average recovery of 96% in normal conditions when no
algal blooms were present. During events, recovery decreased to 85% due to the
reduction in flux when keeping the same mechanical cleaning configuration
(backflush flow rate and time and frequency used). Therefore, more water required
disposal and less filtered water was produced.
• To decrease the kinetics of membrane fouling, the optimal choice was to reduce flux
before shortening the cycle time and switching to recirculation mode. When the plant
was operated in recirculation mode, membrane fouling did not steadily increase until
a cleaning was required. Instead, fouling increased asymptotically until it reached a
value at which it stabilized and remained constant, or increased very slowly. This
allowed membrane chemical maintenance to extend to every 20 - 24 hours, enough
for stable operating conditions.
• The most effective cleaning was with an elevated pH of 12 using chlor-alkali in the
chemical cleaning stage. This achieved up to 94% flux restoration for these types of
membranes versus 40% obtained by cleaning at pH 9.
• The pilot plant was operated in recirculation mode during the algal bloom; however,
no chemicals (for example, coagulant) were added to the feedwater to improve solids
and organics removal, thereby reducing operating costs.
11.8.4 References
Bricker, S. B., Ferreira, J. G. and Simas, T. 2003. An integrated methodology for assessment
of estuarine trophic status. Ecological Modelling 169, 39-60.
416
11.9 TAMPA BAY, FLORIDA (USA) – NON-TOXIC ALGAL BLOOMS

AND OPERATION OF THE SWRO PLANT DETAILING
MONITORING PROGRAM FOR BLOOMS
Lauren Weinrich1
1
American Water, Voorhees, NJ, USA
Figure 11.9.1. Location of the Tampa Bay Seawater

Desalination Plant (top) and aerial view (bottom; Accessed on 11
Nov. 2016 http://www.acciona.us/projects/water/desalination-
plants/tampa-bay-desalination-plant/).
417
Table 11.9.1. Overview of Tampa Bay seawater desalination plant.
Plant/Project Name Tampa Bay Seawater Desalination Plant

Location Gibsonton, Florida, U.S.A.
(m3/d)
SWRO recovery (%) 60% (additional recovery up to 73-75% is achieved for the
plant from recirculating pretreatment supernatant)
Max. Feed RO Temp. (°C) 40
Commissioning date 2008 (Final acceptance test passed 2010)
Intake
Feedwater source Tampa Bay
Intake type Cooling outflow from co-located power plant, mixed with
bay water when needed to meet temperature requirements.
Intake description Outfall canal from power plant (power plant uses open
intake canal system).
Intake screening Plant intake uses 0.5 mm fine mesh for reducing
impingement and entrainment.
(shock) chlorination Chlorine dioxide is dosed at the intake continuously
Strategy, dose rate between 0.5 – 1.1 mg/L. Sodium hypochlorite is fed
continuously prior to the coagulation mixing basins.
Online raw water Conductivity, temperature, pH, turbidity, dissolved oxygen
monitoring
Discrete raw water analysis
relevant to HAB
Pretreatment
Process description Upflow, deep bed sand filters, diatomaceous earth
(precoat) filtration, cartridge filtration (5µm)
Chemical dosing Chlorine dioxide at the intake. Ferric chloride, sodium

hypochlorite, and sulfuric acid (for pH adjustment) are
dosed prior to the rapid mix basins. Sodium bisulfite is
added after the cartridge filters.
Feedwater during bloom
Feedwater design parameters
conditions
Temperature range (˚C) 15.5 – 40 36 - 38
Salinity range 25,000 – 31,000 29,000
(TDS mg/L)
Conductivity (mS/cm) 37 – 43 N/A
Total Suspended Solids 10 - 30 37
(mg/L)
SDI15 6 – 6.5 SF effluent N/A
2.8 – 4.0 DE effluent
2.5 – 3.5 CF effluent
Turbidity (NTU) 1 – 3 Seawater 2-5
Organic Matter TOC 1 – 12 mg/L TOC 5 – 7.6 mg/L
AOC 60 – 260 µg acetate C AOC 230 – 490 µg acetate
per L C per L
418
Table 11.9.1 (Continued)

conditions
Algal cell count (cells/L)
Algal species Ceratium furca (red-tide)

Phaeocystis (foaming)
Chlorophyll-a (µg/L) 9.6
Additional relevant water DO 4 – 9 mg/L
quality parameters for Tot. N 0.4 mg/L Tot. N 2 mg/L
design or observed spikes
during algal bloom
Sand filter 2.4 – 12.2 m/h N/A
Diatomaceous earth 24.4 – 97.7 m/h N/A
RO flux (L/m2h) 23.3 N/A
Terms:
AOC assimilable organic carbon
CF cartridge filter
DE diatomaceous Earth (Precoat)
DO dissolved oxygen
N nitrogen
SF sand filter
TOC total organic carbon
11.9.1 Introduction
The Tampa Bay Seawater Desalination Plant (TBSDP) is located in Gibsonton, Florida on
the south shore of Apollo Beach on Hillsborough Bay (Figure 11.9.1). TBSDP can produce
up to 94,635 m3/d and supplies 10% of the drinking water to the region. Tampa Bay Water
owns the plant and the joint venture American Water – Acciona Agua is responsible for
water production. TBSDP is co-located with a coal-fired power plant and receives feedwater
from either the plant’s cooling loop or, if the cooling water exceeds the temperature limits of
the membranes, bay water can be mixed with the cooling water. Pretreatment comprises
upflow deep bed sand filters, diatomaceous earth (DE) (precoat) filtration and cartridge
filtration (Figure 11.9.2). There are seven reverse osmosis treatment trains. Since the plant
went online, TBSDP has not experienced a toxic harmful algal bloom (HAB). Non-toxic
micro-algae contributed to periodic operational challenges within the pretreatment process,
described in the following case study.
11.9.2 Periodic performance issues
Plant personnel observed operational problems and recorded adverse water quality events
during late 2009. In order to track the severity of intermittent yet recurring issues, staff
assigned a grade to both the water event and the loss of membrane performance on a scale of
0 – 3 (3 being the most severe). From September to December 2009, there were seven events
classified as level 3 that lasted between one and nine days. Other problems associated with
the level 3 events included reduced production capacity, foaming in the pretreatment basins
(Figures 11.9.3 and 11.9.4), and shortened diatomaceous earth filter run times.
419
Figure 11.9.2. Process flow diagram of Tampa Bay Seawater Desalination Plant. Photo:
http://www.tampabaywater.org/Portals/0/desal-fact-sheet.pdf. Accessed May 01, 2017).
Figure 11.9.3. Normal pretreatment basin appearance (Voutchkov 2009).
420
Figure 11.9.4. Foaming observed in the rapid mix basins during pretreatment at TBSDP. Source:
Voutchkov (2009).
After a particular episode of shortened DE filter run time and foaming in the rapid mix basins
in late September 2009, microscopic analysis of source seawater, sand filter and DE filter
backwash was conducted. The algal species Ceratium furca and Phaeocystis spp. were
present in the source seawater in October, between the initial event in September and another
level 3 episode in November (Voutchkov 2009). The presence of algae in the sand and DE
filter backwash waters suggest that algae were not retained by the sand filters and passed
through to the DE filters, thereby blinding the DE filters and causing shorter DE filter run
times. Other causes of the adverse event were considered, such as an increase in alluvial
organic matter from the nearby Alafia River, changes in salinity and calcium concentration,
or other intermittent issues from the co-located TECO power plant. Total organic carbon
(TOC) measured in the raw seawater during the performance issue on September 30, 2009
was 7.6 mg/L (Owen 2015, pers.comm., 21 July), which was in the upper range of the
fluctuations that TBSDP experiences and commensurate with increased algal activity.
Averages determined from a study in 2015 found that TOC ranged from 5.5 to 6 mg/L (Haas
et al. 2015).
During the November event, samples were collected for assimilable organic carbon (AOC)
and TOC from the raw seawater, after initial chemical pretreatment, after the settling basins
and before sand filters, after sand filters, after diatomaceous earth filters and the RO feed.
AOC ranged from 230 – 490 µg acetate C per L; average AOC in the raw water was 360 ±
180 µg acetate C per L over three days (Figure 11.9.5). The pretreatment sample was
collected after chlorine dioxide dosing and ranged from 190 – 760 µg acetate C per L over
the three days of monitoring, and averaged 440 ± 290 µg acetate C per L. AOC was 65%
higher in the pretreatment location (following disinfection with chlorine dioxide) compared
to the raw water. Oxidation of organic matter or sheared algal cells from turbulent discharge
at the power plant may have caused the increase in AOC during the pretreatment stage.
Increases in AOC have been observed following disinfection with chlorine dioxide (Haas et
421
al. 2015; see also Chapter 9).

While the actual cause of the
operational challenges was
not confirmed to have
exclusively resulted from an
algal bloom, increases in
AOC may lead to increased
biofouling potential which is
an important consideration
for seawater RO plants
worldwide.
Biological and organic RO
fouling are the most common
Figure 11.9.5. AOC and TOC at TBSDP during periodic operational fouling issues at TBSDP.
issues in November 2009 thought to result from a non-toxic HAB. Recent studies have shown
Pretreatment refers to post-chlorine dioxide addition. SF = sand filter that although the levels of
(after FeCl3 coagulation), DE = diatomaceous earth filter. Error bars
organic carbon at the intake
represent standard deviations of samples collected over three days.
are generally high and
variable, the treatment
process impacts the biodegradable fraction. Intake AOC levels ranged from 60 – 260 µg
acetate C per L in recent studies (Schneider et al. 2012; Haas et al. 2015) and were often
reduced after filtration. AOC levels increased in the RO feed following chemical addition of
sodium bisulfite for reducing ORP. The presence of AOC greater than 30 µg acetate C per L
has been correlated to increased RO biofouling potential ( ich et al. 2016; Haas et al. 2015).
11.9.3 Algal bloom monitoring and HAB strategy
Monitoring for HAB in Tampa Bay is conducted by the Florida Fish and Wildlife
Conservation Commission for protection of human health, ecosystems, recreation, tourism,
and aquaculture. The most common HAB species in the region, Karenia brevis, produces
brevetoxin, with outbreaks occurring nearly every year along the coast and is monitored at
numerous stations around Florida, with results reported through the website
http://myfwc.com/REDTIDESTATUS. TBSDP personnel regularly monitor the website for
updates on occurrence of algal blooms in the area (Martorell Cebrian 2015). HABs and
biotoxin occurrence affect recreation and potentially desalination treatment processes.
Studies in southern California have suggested potential desalination plant concern associated
with marine algal biotoxins that include domoic acid, saxitoxin, brevetoxin, okadaic acid, and
yessotoxin (Caron et al. 2010; see Chapter 2). Tampa Bay, however, has not had a major
event since 2005. In 2014 a K. brevis HAB occurred north of Tampa Bay between Franklin
and Citrus counties in the Gulf of Mexico. Algal blooms caused by other species have
occurred in Old Tampa Bay (the northwest arm of Tampa Bay) since 2008 but did not impact
TBSDP, which is between Tampa Bay and Hillsborough Bay (Figure 11.9.6).
Watershed and water quality management are the focus of numerous projects to improve
water quality and reduce nutrient loading in Old Tampa Bay. Despite the lack of HAB
occurrence at TBSDP, monitoring for increased algal growth in water near the intake is
important for maintaining effective treatment, avoiding adverse effects, and maintaining
historical records for establishing patterns of water quality changes. There would be tangible
benefits if a monitoring program were also in place for organic carbon, chlorophyll-a and
other algal constituents, nutrients, particle size and abundance for evaluating water quality
trends and to track changes and make adjustments to preempt adverse effects of algal blooms.
422
If toxic HABs occur, the plant would opt to shut down (Owen 2015). The water portfolio for
the region currently ensures that the majority of the drinking water supply is from other
freshwater sources. The
desalination plant can provide
10% of the Tampa Bay Water
supply. Operations staff would
monitor intake online water
quality parameters such as
turbidity and pH and work with
state and local authorities to
conduct additional analyses to
decide when the plant could be
started and brought back online.
11.9.4 Conclusions
With a diverse water supply
portfolio, Tampa Bay Water has
been able to meet supply
demands from other resources
when adverse water quality
events (e.g., non-toxic algal
blooms) affected production
Figure 11.9.6. Satellite photo of Tampa Bay (A); Tampa Bay capacity. In the events described
Seawater Desalination Plant (TBSDP) (B); Hillsborough Bay (C); above that caused foaming in the
and Old Tampa Bay (D). rapid mix basins and shortened
filter runs, TBSDP staff worked
to define the severity of events and track their occurrence over time. When the events
occurred, staff had to adapt treatment to the limitations caused by the event and the result was
a decline in plant production. To date, there have been no toxic HABs and TBSDP continues
to monitor for the occurrence of HABs using information provided by the Florida Fish and
Wildlife Conservation Commission. If a toxic HAB occurred and threatened the water quality
of the intake supply, then TBSDP would stop processing seawater and shut down the plant.
Taking these necessary precautions is critical for avoiding health risks, as Tampa Bay Water
would supply safe drinking water using other sources in their water supply portfolio.
11.9.5 References
Cetinić, I., Filteau, G., Lauri, P., Jones, B., and Trussell, S. 2010. Harmful algae and
their potential impacts on desalination operations off southern California. Water
Research 44(2), 385-416.
Haas, C. N., LeChevallier, M. W., and Weinrich, L. A. 2015. Application of the
Bioluminescent Saltwater Assimilable Organic Carbon Test as a Tool for Identifying
and Reducing Reverse-Osmosis Membrane Fouling in Desalination. Water Environment
& Reuse Foundation. Project 11-07.
Martorell Cebrián, A. 2015. Personal Communication, July 2015.
Owen, C. 2015. Personal Communication, June 2015.
423
Schneider, O. D., Weinrich, L. A., Giraldo, E., Kennedy, M., and Salinas, S. 2012.
Investigation of Organic Matter Removal in Saline Waters by Pretreatment. Water
Research Foundation. Project 4280.
Voutchkov, N. 2009. Initial biofouling analysis. Water Globe Consulting Tech. Memo. Proj.
No. 0007-2009-0001 on Nov. 5, 2009.
Weinrich, L., LeChevallier, M. and Haas, C. N. 2016. Contribution of assimilable organic
carbon to biological fouling in seawater reverse osmosis membrane treatment. Water
Research 101, 203-213.
424
11.10 JACOBAHAVEN, THE NETHERLANDS – ULTRAFILTRATION

FOR SWRO PRETREATMENT: A DEMONSTRATION PLANT
Rinnert Schurer1,2, Loreen O. Villacorte2,3, Jan C. Schippers2 and Maria D. Kennedy2

1
Evides Water Company, Rotterdam, the Netherlands
2
UNESCO-IHE Institute for Water Education, Delft, the Netherlands
3
Figure 11.10.1. Location of the UF RO demonstration site in the

Netherlands (top). Algal foam at demonstration plant intake site
(bottom); inset – Phaeocystis colony. Photos: D. J. Patterson
and R. A. Andersen; A. Al-Hadidi.
425
Table 11.10.1. Overview of Jacobahaven UF-SWRO demonstration plant.

Plant/Project Name Evides Desalination Demoplant Jacobahaven
Location Jacobahaven, Oosterschelde Estuary, the Netherlands
Primary product water use Pilot for drinking water and/or industrial water production
Desalination technology SWRO first pass and BWRO second pass
Total production capacity 360 (output)
(m3/d)
Operational period January 2009 to July 2012 (decommissioned)
Intake
Feedwater source North Sea / Oosterschelde. Subject to tidal currents and
sediments of the nearby tidal flats
Intake type Open intake
Intake description 2 parallel open pipes DN150 mm, submerged 4 m below
water surface into a rapid-flowing tidal current; coarse
screening (aperture unknown); 10 mm-perforated
secondary screen in intake pump suction
UF pre-screening 50-µm microstrainer, automatic backwash (purpose:
retention of mussel seed)
(Shock) Chlorination Not practiced, intake pipe was pigged regularly in summer
strategy, dose rate
Online raw water monitoring Conductivity, temperature, pH, turbidity
Discrete raw water analysis TOC, algal cell counts and identification, chlorophyll-a,
relevant to HABs TEP (Transparent Exopolymer Particles)
Pretreatment
Process description In-line coagulation (seasonally optional); pressurized
inside-out dead-end UF; cartridge filter 10 µm; SWRO;
BWRO (seasonally optional for boron removal);
remineralization
Chemical dosing UF in-line coagulation (seasonally optional, initially PAC,
later superseded by ferric, 0.5 – 1.0 mg/L);
SWRO pH control (HCl, to pH 7.2, later omitted, no
antiscalant dose); BWRO pH control (NaOH, to pH > 9.5);
Remineralization CO2 (~ 100 mg/L) and CaCO3
(100 mg/L)
Feedwater design Conditions outside algal Conditions during algal
parameters bloom bloom
Temperature range (˚C) 2 – 25 ~ 9 – 15 (Spring)
Salinity range (TDS mg/L) 29,000 – 35,000 (unaltered)
Conductivity (mS/cm) 45 – 50 at 25 °C (unaltered)
Total suspended solids <1 - ~ 100 (avg: 35) (unaltered)
(mg/L)
SDI (5, 10 or 15 minute No data available SDI15 >5 (see Al-Hadidi,
interval)(%/min) 2012)
Turbidity (FTU) 5 – avg. 15 – 50 (shut (unaltered)
down, peaks up to 100)
Organic Matter DOC 1.2 – 1.8; DOC: unaltered;
TOC/DOC (mg/L), TEP, TOC 1.5 – 2.2 TOC: 3.0
biopolymers
426
Feedwater design Conditions outside algal Conditions during algal

parameters bloom bloom
Algal cell count (cells/L) 100 – 300 0.5 x 106 – 12 x 106 (peak)
Algal species Centric diatoms Phaeocystis, Chaetoceros
Chlorophyll-a (µg/L) 1–5 ~ 15 -20 (peak)
Additional relevant water pH: 7.9 – 8.1 pH: 8.5
quality parameters ortho-P: 0.04 mg/L ortho-P: < 0.01 mg/L
DO: 8 – 11 (avg.) – 18 (same, possibly due to
mg/L intake turbulence)
TEP0.4µm: <0.05 mg Xeq/L TEP0.4µm: 0.05 - 0.74 mg
Xeq/L
Desalination design Conditions outside algal Conditions during algal
Bloom Bloom
UF type 150 kDa MWCO, Pentair (unaltered)
UF flux (L/m2h) 55 – 90, without 55, with coagulation
coagulation activated
UF backwash Backwash interval 45 – 90 Backwash interval 45 (30)
min. min.
Backwash flux 250 L/m2h Other settings unaltered
for 45 seconds
UF feedwater pH 6.6 if pH correction ~7.6 for Fe-coagulation,
deployed, otherwise 7.4 to ~6.6 for Al-coagulation
8.1 - 8.4 (ambient)
UF coagulation No coagulant dosed 0.5 – 1.5 mg/L Fe inline
coagulation for UF
UF CEB CEB initiated when TMP Omitted during coagulation,
(at 20 °C) had increased to as not effective
0.30 bar
NaOH pH 9.4, NaClO 200- Same protocol, but less
400 ppm, HCl pH 2, effective (as described in
soaking time 10 – 20 min. the text)
UF CIP Only practiced after or Solution of 1% ascorbic
during coagulation, acid and 0.4 % oxalic acid,
otherwise no need soaking time 12 – 48 h. No
temperature control
UF wastewater handling Direct discharge to sea, as Buffering, secondary
no UF-coagulation was coagulation and
practiced outside algal clarification, clarified water
bloom discharged to sea
other wastewater handling Discharge to sea, combined Discharge to sea, combined
with strainer backwash with strainer backwash
water, SWRO and BWRO water, SWRO and BWRO
brine and UF waste water brine and clarified UF waste
water
SWRO flux (L/m2h) 13.4 13.4
SWRO recovery (%) 40 40
BWRO recovery (%) 90 90
427
11.10.1 Introduction
Evides Waterbedrijf, the South-Western Netherlands water supply company providing
municipal drinking water, industrial water, as well as wastewater treatment services, owned
and operated a UF-RO (Ultrafiltration – Reverse Osmosis) desalination plant on the
Oosterschelde for demonstration and research purposes between 2009 and 2012. The overall
aim of the project was to study desalination as one of the alternatives to the water resources
already used by Evides (i.e., fresh surface water, ground water and infiltrated water), in the
framework of the company’s efforts to safeguard water supply in the long term with respect
to climate change. The desalination research was conducted under locally relevant conditions
regarding water quality (seasonally variable) and temperature (seasonally low). The
characteristics of the plant are given in Table 11.10.1.
As the site featured an open intake, and raw water quality was variable, conditions were
challenging for the downstream treatment processes. A major characteristic in this respect
was the occurrence of seasonal algal blooms, which impacted the UF pretreatment in terms of
fouling, necessitating the temporary application of in-line coagulation. To guide plant
operation and design, a major objective was to gain an understanding of the following topics
relative to the ultrafiltration pretreatment:
• Seasonal variation in raw water quality, especially of parameters associated with algal
blooms, expected to be relevant for the strainer and UF treatment steps such as algal
presence, associated algogenic organic components and nutrients;
• Impact of algal bloom-related raw water quality variations on UF permeability
performance i.e., fouling, and its maintenance by means of hydraulic backwash,
chemically enhanced backwash (CEB), and application of coagulant;
• Resultant UF permeate quality and SWRO membrane permeability.
During the plant’s operation, ample data were collected; several scientific institutes
conducted or participated in research at the site. A valuable dataset, experience, and lessons
learned have therefore been acquired1. The dissemination of this material by means of the
current Case Study aims to contribute to the ongoing improvement of operating standards in
the desalination industry.
11.10.2 Feedwater intake
The demonstration plant feedwater was drawn from the Oosterschelde water body, close to
the southern landfall of the identically-named storm surge barrier (Figure 11.10.2). This feed
can be regarded as North Sea water in terms of salinity, temperature, and organic and
inorganic constituents. Due to its geographical position, seasonal temperature variations were
applicable.
The intake was located at a site featuring highly turbulent conditions caused by tidal currents,
hence stratification was absent. During stormy weather, severe turbidity peaks occurred, due
to mobilization of sediments from the tidal flats in the Oosterschelde water body. If turbidity
exceeded 50-75 FTU for more than several hours, the limited capacity of the strainer forced a
plant shut down, which occurred two to four times a year.
A specific characteristic of the feedwater was the recurrence of seasonal algal blooms in
Spring (April-May) every year, as described in Section 11.10.5. For further feedwater data,
see Table 11.10.1.
1
For further literature specific for the current Case Study, the reader is also referred to Al-Hadidi et al. (2012);
Villacorte (2014); Tabatabai (2014) and Schurer et al. (2012, 2013).
428
No intake chlorination was

conducted, contrary to common
practice. The main reasons were to
preclude the formation of
chlorination by-products
(trihalomethanes) and to safeguard
SWRO membranes from chlorine
degradation. Although clogging
by barnacles and shells did occur
in the (multiple) intake structures
during summer, this could be
controlled satisfactorily by regular
(2-weekly) pigging due to the
Figure 11.10.2. Intake location at arrow, Oosterschelde storm relatively small-scale system.
surge barrier (structure in center) Photo: Pinterest. 11.10.3 Process line-up
The plant comprised process stages depicted by the block diagram in Figure 11.10.3 and
further specified in Table 11.10.1. Figure 11.10.4 presents images of the main UF and RO
equipment, respectively.
Figure 11.10.3. Treatment process block diagram for the SWRO Jacobahaven demonstration plant.
11.10.4 Occurrence of algal

blooms
During the four years of site
operation, algal blooms recurred
every spring, from mid-April
onwards. Such events are common
in this section of the North Sea
(Janse et al. 1996). As the blooms
always commenced at the same time
of the year, it is speculated that
favorable daylight length and
Figure 11.10.4. Demonstration plant ultrafiltration (left) and
seasonal temperatures were
SWRO (the 5 pressure vessels on the left of the right hand contributing factors in bloom
image) and BWRO (the 3 pressure vessels on the right side of development. Typically, blooms
the skid). developed in a short period (2-4
weeks) in spring and subsequently
declined after several months in summer. The severity of the blooms varied, ranging from
429
peak levels of 12 million cells/L to a relatively subdued bloom in 2012 at 2 million cells/L
(which could possibly be attributed to the cold and stormy weather conditions in that
particular year).
Algal speciation varied from year to year. Diatoms were generally present during each bloom
(Chaetoceros, Thalassiosira), whereas Phaeocystis or Chrysochromulina prevailed only in
some years (2010 and 2010 and 2011, respectively) during the four years of monitoring
(Evides data). Example images of these species are given in Appendix 1. The prevalence of
Phaeocystis caused noticeable foam formation as shown in Figures 11.10.1 and 11.10.5.
11.10.4.1 Associated raw water
quality parameters
During algal blooms, levels of
chlorophyll-a (Figure 11.10.6) and
algal counts (Table 11.10.1)
correspondingly spiked up to 20 µg/L
and 10,000/mL, respectively.
Interestingly, a measurable elevation
of the pH (~0.5 pH above average)
occurred during the algal bloom
periods. In contrast, no deviation in
dissolved oxygen was noted, possibly
because of the highly turbulent
Figure 11.10.5 Foam formation at the intake site during
Phaeocystis bloom occurrence. conditions at the raw water intake site.
Also, in associated research work, the
presence of TEP0.4µm was measured. TEP0.4µm levels fluctuated in rather close unison with
chlorophyll-a levels, increasing 20 to 60-fold at the algal bloom peak (Villacorte 2014;
Schurer et al. 2012), as shown in Section 11.10.6 – (Impact on UF). More data on TEP
measurements are presented in Chapters 2 and 5 and the methods for TEP measurement are
outlined in Appendix 4.
Figure 11.10.6. Seasonal patterns of raw water temperature and algal-boom related parameters (chlorophyll,
TEP0.4µm).
430
11.10.5 Impact of harmful algal blooms on microstraining

During algal blooms, a significant increase in the clogging rate of the 50-µm strainer was
observed, manifested as a backwash interval reduction to 5 minutes from 0.5 – 1.5 hours for
non-bloom conditions at identical turbidity. The overall plant production capacity remained
unaffected as the required strained water output could still be met.
11.10.6 Impact on Ultrafiltration (UF)
11.10.6.1 General overview of seasonal fouling pattern
A set of UF transmembrane pressure (TMP) profiles as typically encountered over the
seasons is depicted by Figure 11.10.7. Here, the TMP is normalized to 20 °C by inclusion of
a correction factor according to the temperature-dependency of the viscosity of water.
Figure 11.10.7. Typical seasonal UF fouling patterns. TMP is normalized to 20°C and at 55 L/m2h.
Typically, a CEB (Chemically Enhanced Backwash) was initiated once the TMP had
increased to 0.30 bar (temperature-corrected and at the default flux of 55 L/m2h), instead of
fixed time or volume intervals. As the applied membranes were of a high-permeability type,
equivalent permeability ranged from 180 – 200 L/m2hbar just before CEB, and 300 –
500 L/m2h bar after CEB.
As is evident from Figure 11.10.7, a remarkable seasonal variation occurred in the UF
fouling pattern, ranging from a comparatively slow increase and eventual stabilization of the
TMP in Winter and Autumn (curves D, respectively C) to a steep increase in Spring (curve
A).
The fouling behavior of the UF throughout the research period is depicted in Figure 11.10.8.
In this graph, the UF fouling rate is expressed as the (temperature-corrected) TMP increment
incurred between two successive CEB’s, divided by the elapsed time. TMP just after CEB
ranged generally between 0.12 - 0.15 bar. The fouling rates are shown for both UF operation
with coagulation, which was typically around April-July, and without coagulation for the
remainder of the year, as is further elaborated in the next paragraphs. A fouling rate of 0.25 –
0.3 bar/day was considered to be an upper practical limit in terms of CEB interval (< 12 h)
and hence UF water loss for the operating conditions applicable in this case study.
431
Figure 11.10.8. Seasonal pattern in UF fouling rate, for periods without respectively with coagulation deployed.
Also, CIP’s are indicated.
11.10.6.2 UF operation during algal bloom: fouling

As highlighted by a comparison of Figures 11.10.7 and 11.10.8, algal blooms had a
significant impact on the fouling rate of the ultrafiltration system, as the occasions of high
fouling rate coincided with the elevated levels of chlorophyll-a, TEP0.4µm and algal counts
(the latter is not shown). Typically, the UF fouling rate accelerated in a 2 - 4 week period in
April to such an extent that CEB intervals declined to less than 12 or even 6 hours. In this
fouling regime, unaltered operation was no longer practical due to UF downtime and water
loss which eventually led to a shortfall in the net permeate output, as no redundancy in UF
production capacity was installed in the plant.
In the first instance, operators attempted to control the aforementioned bloom-related UF
fouling by adaptation of the hydraulic backwash regime, that is, extension of backwash
duration and shortening of backwash intervals from 45 to 25 minutes. Only a marginal
reduction in UF fouling was achieved which was outweighed by the associated increase in
backwash water losses (20%). Furthermore, intensified CEB (prolonged soaking, increased
concentrations) proved unable to compensate for the rapid fouling, whereas a frequent CEB
execution (e.g., more than 1-2 times daily) implied excessive UF production loss due to
downtime and CEB water losses. Hence, extended regimes of hydraulic backwash or CEB
were not capable of countering the increase in UF fouling. Similarly, lowering of the flux to
30 L/m2h still did not yield a feasible CEB interval (and would have meant an
uneconomically oversized UF system for the non-bloom periods).
11.10.6.3 UF operation during algal bloom: coagulation
As modifications of backwash, CEB and flux proved unable to adequately control UF fouling
during algal blooms, coagulation was seasonally implemented. In the first year of operation
(2009), aluminum (~0.5 mg/L Al) coagulant was applied at pH 6.6, but although effective to
counter UF fouling, it proved detrimental to RO operation due to aluminum silicate scaling,
and hence was abandoned in favor of ferric coagulant.
Ferric doses ranging from 0.5 to 1.5 mg Fe/L proved generally capable of controlling UF
fouling during the algal bloom, as shown by curves B1 and B2 in Figure 11.10.7. During Fe
432
coagulation, a pH of 7.6 was maintained. In some cases, the coagulant became less effective
for UF fouling control, for which no clear reason was found. At such times, the dose was
increased. Eventually, cleaning-in-place (CIP) was required to restore permeability, as further
elaborated in Section 11.10.6.4.
Coagulation efficacy was improved by relocating the dosing point, which was originally sited
in the UF feed tank, to the UF feed pump suction side, thereby creating a better defined
hydrodynamic regime for the rapid mixing of the coagulant. This resulted in longer periods
of stable UF operation. During coagulation, the UF was always operated at the default flux of
60 L/m2h. Hence, it is not known whether a higher flux would have been feasible.
Although it was relatively clear when to commence the UF coagulation (that is, at UF fouling
rate exceeding ~ 0.25 – 0.30 bar/day), it was less obvious when to actually terminate the
coagulation. In general, cessation of Fe-dosing after the algal bloom had apparently passed
(as indicated by chlorophyll-a levels) still resulted in a rapid permeability decline. Only after
conducting a CIP (see Section 11.10.6.4), a situation with only moderate fouling without
coagulation dosing was re-established.
11.10.6.4 UF operation during algal bloom: CEB and CIP
An important observation was that the regular CEB was no longer effective during periods of
Fe-coagulation. UF permeability was only restored by conducting a CIP, where a mixture of
ascorbic acid (1%) and oxalic acid (0.3 %) proved adequate. Other chemicals (for example,
citric acid), or omission of either ascorbic or oxalic acid, were not effective in recuperation of
permeability. Although not explicitly tested, the CIP appeared to be more effective at higher
temperatures for example in mid-summer when compared to during May. Each algal bloom
period necessitated 1 or 2 such CIPs. For one such CIP, the liquid was analyzed for metallic
composition, as shown in Table 11.10.2. Unfortunately, quantification of organic foulant
release by a CIP was not feasible due to the high carbon content of the CIP medium itself.
Table 11.10.2. Metallic compounds released by CIP, as g/m2 membrane area.
Metal Membrane area Comments
Fe 0.3 g/m2 Likely originating from the
coagulant
Al 0.2 g/m2 Possibly immobilized clay
particles
Mn not detectable Potential UF foulant, originating
as trace compound in Fe-
coagulant
11.10.6.5 UF operation outside algal blooms

Outside of algal blooms, from July till April, the UF fouling rate was low, allowing CEB
intervals ranging from 24 hours up to a week or even or even a fortnight. No coagulation was
deployed, as there was no need since fouling rates were low. Increasing the filtration cycle
duration from 45 min (default value) to 60 and even 90 min proved to be possible without
incurrence of noticeable acceleration in fouling (see Figure 11.10.9), thereby reducing UF
backwash water loss from 10 % to < 5 %. The backwash duration utilized (hence backwash
volume) proved to be sufficient, as a more prolonged backwash did not establish further
permeability recuperation. Furthermore, filtration fluxes could be increased up to 90 L/m2 h
433
without excessively raising the fouling rate (Figure 11.10.10), but this was only utilized for
intermittent trials as there was no use for the surplus UF permeate.
UF feedwater pH was initially
kept at 6.6 outside application of
coagulant, but this pH correction
was discontinued after 2009 as
no obvious need was apparent,
the RO being in a non-scaling
regime due to the relatively low
recovery of 40%. In the
subsequent period, pH was
varied between 6.6, 7.4 and
ambient (~8.1 – 8.4), but no
significant effect on UF fouling
was observed. Turbidity, even at
Figure 11.10.9. UF fouling rate outside algal bloom, and without
elevated levels during storm
coagulation, for various backwash intervals at the default 55 L/m2h
flux. events, did not affect UF
permeability.
11.10.6.6 UF wastewater
handling
Outside of algal bloom periods,
the spent UF backwash water
was discharged directly into the
Oosterschelde as it did not
contain additional chemicals.
During periods of coagulation,
the spent wastewater was
conveyed to the plants integrated
Figure 11.10.10. UF fouling rate outside algal bloom, and without wastewater treatment, which
coagulation, for 55 and 90 L/m2h flux. consisted of secondary
coagulation (ferric, 5–10 mg/L
Fe dose) and a combined flocculator / lamellae clarifier. The produced sludge was thickened
and stored in a tank, and eventually disposed of by truck, due to the relatively small sludge
quantities involved. The clarified wastewater was generally discharged back into the
Oosterschelde, as the risk of additional UF fouling by inadvertent reintroduction of algal
biopolymers and/or excess unsettled coagulated material was considered not to outweigh the
water intake savings, the latter being quite limited in absolute sense. In a few short-term trials
runs where clarified water was purposely recirculated to the UF feed, no immediate effect on
UF fouling was noticed.
11.10.6.7 Alternative UF backwashing by SWRO concentrate
As the SWRO concentrate did not have scaling properties due to the relatively low applied
recovery (40%), it was theoretically possible to use this for UF backwash water. Several trial
runs were achieved, where indeed no increase in the UF fouling rate was encountered
compared to the regular backwash with UF permeate. Hence, UF backwash with RO
concentrate appeared feasible. This was not tested extensively during the more critical period
of algal bloom, where accumulation of algogenic organic compounds in the SWRO
concentrate is a potential issue.
434
11.10.6.8 UF permeate quality and membrane integrity

Selected UF permeate quality data are presented by Table 11.10.3. Appropriate RO feedwater
quality standards were met. Furthermore, UF integrity testing revealed that UF fiber breakage
amounted to < 0.06 % over a period of 3.5 years of nearly continuous operation under
periodically challenging conditions, attesting to the resilience of UF pretreatment in the
retention of particulate and biomass matter in conjunction with the upstream-positioned
microstrainer.
Table 11.10.3. UF permeate quality data and reductions.
Parameter Result
Turbidity < 0.05 FTU
Particle count reduction > 3.6 log for particles ≥ 0.5 µm
SDI15 < 2 (Al-Hadidi et al. 2012)
Fe during coagulation < 0.05 µg/L (detection limit)
Biopolymer (LC-OCD) ~50 % reduction (Salinas Rodríguez 2011;
Villacorte 2014)
TEP0.4µm < 0.01 mg Xeq/L (detection limit), equivalent
reduction > 90% (Appendix 4; Villacorte 2014).
11.10.6.9 Further data on UF foulant levels

During an algal bloom period in May 2013, various parameters relevant to HABs were
monitored at the same Oosterschelde site. Since operation of the demo-site UF unit was
discontinued in July 2012, samples for the UF feed and UF permeate were collected from a
smaller pilot UF unit (1” module) with similar membrane properties and operational settings
as in the previous demo-site unit. On this occasion, the raw water recorded algal density was
6.25 x 106 cells/L and the chlorophyll-a concentration was 10.6 µg/L. Table 11.10.4 shows
the comparison of the different parameters and their reduction over the pretreatment
processes.
Table 11.10.4. Measured parameters for raw water, strained water and UF permeate during
the 2013 algal bloom. Data from Villacorte et al. 2014.
MFI- Bio-
UF10kDa TEP0.4 µm TEP10 kDa DOC polymers SUVA
Sampling
point s/L2 (mg Xeq/L) (mg Xeq/L) mg C/L mg C/L L/(mg.m)
Raw water 16000 0.079 1.49 1.61 0.34 2.38
After strainer 9200 0.075 0.78 1.55 0.32 2.52
UF permeate* 1200 0.001 0.44 1.25 0.10 3.02
* The sample for UF permeate was collected from a 1” module size pilot system operated with the same feedwater,
membrane type and hydraulic settings as the previous demonstration plant.
The strainer showed significant reduction of MFI-UF10kDa and TEP10kDa but no significant
reduction of other parameters was observed. The UF demonstrated remarkable reduction in
MFI-UF10kDa, TEP10kDa, TEP0.4µm and biopolymers. DOC was only slightly reduced mainly
because the low molecular weight (<1 kDa) organic substances such as humic substances,
building blocks, organic acids and organic neutrals, comprising about 80% of the DOC, are
435
poorly removed by UF by nature of the UF MWCO (150 kDa). The increase in SUVA
(UV254nm/DOC) was due to the reductions in DOC and relatively unchanged UV absorbance
over the pretreatment processes. UV absorbance was attributed to the aromatic compounds
such as humics, which were poorly removed in the pretreatment processes.
11.10.7 Operational characteristics of downstream Reverse Osmosis (RO)
During the experimental period, SWRO fouling was assessed by relative MTC (Mass
Transfer Coefficient) as a measure of permeability and NPD (Normalized Pressure Drop) to
indicate spacer fouling. Figure 11.10.11 presents their respective behavior throughout the
research period.
At the first spring season of SWRO operation, an initial sudden permeability decline was
associated with the original application of aluminum coagulant (Gallego et al. 2007; Gabelich
et al. 2005), which was thereafter abandoned. In the subsequent three years of SWRO
operation, RO permeability declined by approximately 20%. No clear relation (e.g. a sudden
concomitant permeability decline) with algal blooms was observed, although autopsies
indicated the presence of some biopolymeric organic deposition. CIP’s by caustic and acid
restored permeability only marginally. Iron deposits present (see Figure 11.10.12) could be
removed by CIP. Mineral scaling (e.g. calcium carbonate) was absent, as attested by
autopsies and as expected with the practiced regime of (low) recovery and pH control.
Figure 11.10.11. SWRO performance throughout the research period as feed pressure and relative MTC (initial =
100%), NPD. Periods where coagulation was deployed are indicated. Between December 2009 and March 2010,
the RO was out of operation for mechanical reasons.
Normalized spacer head loss was continuously low throughout the trial period, indicative of
the absence of biofouling and particulate fouling and thereby confirming the beneficial
potential of UF as RO pretreatment in this respect. This was corroborated with the results of
membrane autopsies (see Figure 11.10.12) whereby the maximum ATP (Adenosine Tri-
Phosphate) concentration (< 50 pg/cm2) measured at the lead element was substantially lower
than the biofouling threshold (> 1000 pg ATP/cm2) reported by Vrouwenvelder et al. (2008).
436
Biopolymers and TEPs were also

observed mainly in the lead element.
Salt passage remained continuously <
0.5 % (measured as conductivity), being
in line with the SWRO specifications
(not shown further).
11.10.8 Conclusions
Seasonal algal blooms encountered at
the North Sea test site posed
challenging conditions to the UF
Figure 11.10.12. Concentrations of iron, ATP and some pretreatment i.e., severe UF fouling.
organic constituents per surface area of autopsied RO This could be controlled by in-line
membranes. The autopsies were conducted in December coagulation. Nevertheless, algal bloom
2009 for the lead and last RO elements of the SWRO train events essentially governed the overall
(adapted from Villacorte et al. 2014).
UF design in terms of size and auxiliary
equipment, as the algal-related UF fouling capped the UF flux and necessitated the inclusion
of additional coagulation and wastewater treatment facilities.
Multiple issues were identified that merit further research based on the experience gained
from the UF-RO demo-plant study in Jacobahaven. These pertain to the use of UF for SWRO
pretreatment for feedwater prone to algal blooms.
• During the plant design phase, scenario studies should be included regarding algal-
related UF fouling, taking into account the estimated UF fluxes attainable and
equipment cost and space required for coagulant storage and dosing, wastewater
treatment and sludge handling facilities. Since feedwater properties may fluctuate
seasonally and can significantly impact UF layout and operation, as highlighted above,
local advance piloting is recommended to establish an optimal plant lay-out.
• For a relatively large UF plant that experiences challenging feedwater properties,
(semi-) continuous parallel operation of a small-scale UF unit for optimization
purposes may be a worthwhile consideration. This especially pertains to optimizing
coagulation: duration of coagulation required, dose and applied filtration fluxes in
relation to algae-related fouling without coagulation.
• Understanding of local and seasonal presence of algal-bloom related compounds
(biopolymers, polysaccharides, TEP) is important for UF operation, as these impact
UF fouling.
• For plants where recirculation of treated UF backwash wastewater into the seawater
intake is considered, verification of presence or rather absence of residual algogenic
foulants in the returned (clarified) UF wastewater water is a worthwhile provision, in
order to assess risk for accelerated UF fouling by these recirculated compounds.
• Further optimization in establishing a more effective CEB during coagulation is
worthwhile to further enhance process robustness and possibly limit the need for the
more cumbersome CIP’s. The cause for the observed decline in CEB efficacy was not
studied in detail in the current study, though a relation with iron compounds seems
plausible. Conservation of UF permeability may be attempted by e.g. dose of citric
acid during backwash rather than as CEB or CIP, and a further improvement of the
exact configuration (hydrodynamic regime, contact time) of the coagulant
introduction into the main water flow.
437
• Although it was not directly studied, UF CIP efficacy appeared to be enhanced by

elevated ambient temperature, suggesting that a temperature-controlled CIP system
could be worthwhile to shorten system downtime.
• Further study of the exact cause of the decline in SWRO MTC is recommended, by
conducting more extensive membrane autopsies specifically aimed at determining the
presence and identification of organic compounds.
11.10.10 References
Al-Hadidi, A. M. M., Kemperman, A. J. B., Schurer, R., Schippers, J. C., Wessling, M., and
van der Meer, W. G. J. 2012. Using SDI, SDI+ and MFI to evaluate fouling in a UF/RO
desalination plant. Desalination 285, 153-162.
Gabelich, C. J., Chen, W. R., Yun, T. I., and Coffey, B. M. 2005. The role of dissolved
aluminum in silica chemistry for membrane processes. Desalination 180(1-3), 307-319.
Gallego, M. S. and Darton, M. E. G. 2007. Simple laboratory techniques improve the
operation of RO pretreatment systems. In: Proceedings of the International Desalination
Association World Congress, Maspalomas, Gran Canaria, Spain.
Janse, I., Van Rijssel, M., Gottschal, J. C., Lancelot, C., and Gieskes, W. W. 1996.
Carbohydrates in the North Sea during spring blooms of Phaeocystis: a specific
fingerprint. Aquatic Microbial Ecology 10(1), 97-103.
Salinas Rodriguez, S. G. 2011. Particulate and Organic matter fouling of seawater reverse
osmosis systems. Thesis Dissertation, UNESCO-IHE / TU Delft PhD Thesis, CRC
Press.
Schurer, R., Janssen, A., Villacorte, L., and Kennedy, M. 2012. Performance of ultrafiltration
and coagulation in an UF-RO seawater desalination demonstration plant. Desalination
and Water Treatment 42(1-3), 57-64.
Schurer, R., Tabatabai, A., Villacorte, L., Schippers, J. C., and Kennedy, M. D. 2013. Three
years operational experience with ultrafiltration as SWRO pre-treatment during algal
bloom. Desalination and Water Treatment 51(4-6), 1034-1042.
Tabatabai, S. A. A. 2014. Coagulation and Ultrafiltration in Seawater Reverse Osmosis
Pretreatment. Dissertation, Unesco-IHE / TU Delft.
Villacorte, L. O. 2014. Algal Blooms and Membrane Based Desalination Technology. Thesis
Dissertation, Unesco-IHE / TU Delft.
Vrouwenvelder, J. S., Manolarakis, S. A., Van der Hoek, J. P., Van Paassen, J. A. M., Van
der Meer, W. G. J., Van Agtmaal, J. M. C., Prummel, H. D. M., Kruithof, J. C., and Van
Loosdrecht, M. C. M. 2008. Quantitative biofouling diagnosis in full scale nanofiltration
and reverse osmosis installations. Water Research 42(19), 4856-4868.
438
11.11 BARCELONA, SPAIN – SWRO DEMONSTRATION PLANT:

DAF/DMF VERSUS DAF/UF
Joan Llorens1, Andrea R. Guastalli1 and Sylvie Baig2
1
2
Degrémont SA, Rueil-Malmaison Cedex, France
Barcelona
SWRO Plant
Barcelona Airport
Llobregat
River
Pilot Plant discharge
Pilot Open
intake
Figure 11.11.1. Location of the Barcelona seawater desalination pilot plant, the open intake and the Llobregat
River discharge. Photos: Google Earth and Wikimedia Commons).
439
Table 11.11.1. Overview of Barcelona SWRO demonstration plant.
Plant/Project Name Barcelona Pilot Plant

Location Southern shore of Barcelona, Spain
Primary product water use Municipal drinking water
Desalination technology SWRO
Total Production Capacity (m3/d) 200 of desalinated water (with a
pretreatment capacity of 750)
Recovery (%) 45
Operational period May 2007 to April 2010
Intake
Feedwater source Mediterranean Sea
Intake type Open intake
Intake description 1.2 km offshore at a depth of 12 m
Online raw water monitoring Conductivity, turbidity, pH, temperature
Discrete raw water analysis relevant to TOC, Algal count, SDI
HAB
Pretreatment
Process description Flotation (DAF) + Filtration (DMF) or
DAF + 400 µm pre-screens +
Ultrafiltration (UF)
Chemical dosing (as FeCl3) DAF pilot is a rapid flotation using ferric
chloride as a coagulant (0-6 mg/L)
Feedwater design parameters Mean ± SD (maximum value due to
degradation or “green” algal blooms)
Temperature (oC) 19 ± 4 (27.8)
pH 8.2 ± 0.1 (8.4)
Conductivity (mS/cm) 56 ± 1 (60)
TDS (g/L) 36 ± 2 (-)
Turbidity (NTU) 1.7 ± 0.8 (63)
TSS (mg/L) 6 ± 3 (117)
SDI 75% (%/min) 20 ± 10 (> 50)
TBC (cells/mL) 7 x 105 ± 3 x 105 (1 x 106)
Algae count (cells/L) 332,000 ± 302,000 (1,254,000)
Absorbance 254nm (1/m) 0.7 ± 0.2 (1.6)
DOC (mg C/L) 0.9 ± 0.1 (1.1)
A SWRO desalination pilot plant was operated to provide important information for the
development of a large-scale desalination plant in the coastal area of Barcelona. An
assessment of raw water quality considered all parameters that could have a capital or
operational impact on pretreatment and/or the SWRO membrane process such as turbidity,
suspended solids, algal blooms, and organic carbon.
Two pretreatment strategies, dual-media filtration (DMF) and ultrafiltration (UF), running in
parallel to reduce the particulate material and natural organic matter (NOM) from
440
Mediterranean seawater were studied and compared.1 Figure 11.11.1 above shows an aerial
view of the Demonstration Pilot Plant location at Barcelona.
11.11.1.1 Feedwater
Barcelona’s desalination pilot plant had an open intake, located 1.2 km offshore at a depth of
12 m in the Barcelona area of El Prat de Llobregat, Spain. The raw water quality was
monitored on line and extensively analyzed for 21 months from June 2007 to March 2009.
Annual analysis of the raw water demonstrated that the open-intake water was generally of an
excellent quality, with turbidity lower than 4 NTU for 90% of the year. Occasional
divergences in quality were mostly related to severe weather events in which the open intake
was simultaneously affected by the adjacent Llobregat River discharge and wave turbulence
at sea which caused the high TSS (refer Table 11.11.1).
The algal count in seawater averaged 130,000 cells/L for 65 % of the year and increased
almost tenfold, reaching peaks of more than 1,200,000 cells/L during spring and summer
blooms.
11.11.1.2 Process Line-up
The first level of pre-treatment was a dissolved air flotation unit (DAF). The DAF (25 m/h
average) had three main zones: coagulation, flocculation (11 minutes average) and flotation
with ferric chloride used as coagulant. The water then flowed along two different treatment
pathways, as shown in Figure 11.11.2. In the first pathway, additional coagulant could be
added to the seawater prior to a dual-media filter (DMF) unit with a 0.55 m layer of
anthracite (0.95mm size) and 0.45 m layer of sand (0.28 mm size), which operated in down-
flow mode at 14 m/h. DMF was designed to reduce turbidity and the presence of colloids in
the water by physical straining. In the second parallel pathway, seawater was passed through
400 µm pre-screens and then a pressurized outside-in UF hollow fibre membrane (PVDF;
0.02 µm nominal pore size). The UF permeate was then passed through 5 µm security
cartridge filters and fed through a RO module (thin film composite membrane operating at
14 L/m2h and 45% recovery).
Figure 11.11.2. Schematic representation of the Barcelona pilot plant pretreatment scheme.
1
Further literature regarding the current case study can be obtained from Guastalli et al. (2013); Simon et al.
(2013); Sanz et al. (2007).
441
11.11.2 Occurrence of algal blooms

The initiation of blooms in the pilot plant feedwater and subsequent death of algal cells were
found to correlate with the increase and decrease of temperature and sunlight in the seawater,
starting in March and ending in July (peaking in late June). This period is referred to as the
“green season”. During the “green season”, the high phytoplankton productivity maintains
algal concentrations around 600,000 cells/L with occasional short blooms being observed,
leading to concentration peaks greater than 1,200,000 cells/L (see Table 11.11.2). Between
August and February, algal cells have a low activity and the average concentration remained
around 130,000 cells/L.
Table 11.11.2. Algal count (cells/L) measured in the raw seawater seasonally for the
Barcelona SWRO pilot plant from June 2007 to March 2009.
Minimum Maximum Average Standard dev.

Bloom (March-July) 215,000 1,254,000 575,000 290,000
No bloom (August-February) 13,000 282,000 126,000 53,000
11.11.3 HAB-associated water quality parameters

11.11.3.1 Temperature
Temperature variations can affect salt passage in RO membranes and hence, product water
quality. It is therefore an important design parameter for desalination facilities. There is also
a close relationship between temperature and microorganism reproduction rates and seawater
temperature has also been related to the occurrence of algal blooms at the pilot plant. Raw
seawater temperature ranged from 11.7 to 27.8 oC with an average of 19.2 oC. As can be seen
in Figure 11.11.3, the maximum temperature occurs in July-August (summer) whereas the
minimum occurs in January-February (winter).
11.11.3.2 Algal count
The algal counts found, regardless of the period, are low compared to the concentrations
detected in other seas. Nonetheless, the pretreatment should consider effects due to the
possible maximum load of algae. Measurement of algal count over the extended period of 21
months allowed the detection of a sudden algal bloom followed by a five month “green
season” when the algal count was high and stable between March and July.
Algal counts above 200,000 cells/L (defined as an algal bloom for the purposes of this study)
corresponded to 30% of the time. As can be seen in Figure 11.11.3, an algal bloom started in
mid-February when minimum temperatures were reached and continued during increasing
seawater temperatures. Algal blooms remained until the temperature reached average
summer values in July (approximately 22 oC). The rest of the year, corresponding to 70% of
the samples taken, the value observed was less than 200,000 cells/L.
11.11.3.3 Turbidity and SDI
Storm and rain events, prevalent in winter, had the greatest impact in deteriorating water
quality and gave rise to elevated SDI75%. Generally, apart from these events, water quality
was good, with raw seawater turbidity showing little variability with values below 4 NTU for
over the 90% of the samples including algal bloom events. SDI was measured as SDI75%
following the method in Mosset et al. (2008).
442
10000 30
Algae Temperature 28
26
Algae concentration (cell/mL)
24
1000
Temperature (ºC)
22
20
18
100
16
14
12
10 10
01 07
31 07
30 07
29 07
29 07
28 07
28 07
27 07
26 08
27 08
26 08
26 08
25 08
25 08
24 08
23 08
23 08
22 08
22 08
21 08
20 09
09
6/
7/
7/
8/
9/
0/
1/
2/
1/
2/
3/
4/
5/
6/
7/
8/
9/
0/
1/
2/
1/
2/
/0
/0
/0
/0
/0
/1
/1
/1
/0
/0
/0
/0
/0
/0
/0
/0
/0
/1
/1
/1
/0
/0
01

Figure 11.11.3. Algal count (points) and measured temperature (line) in pilot plant raw seawater.
Dashed squares show algal bloom events.
Almost 60% of the SDI75% measured in summer samples were below 21%/min. The
occurrence of high SDI75% values in summer that raised the SDI75% over 20%/min occurred
simultaneously with the completion of the algal bloom and occasional turbidity spikes (see
Figure 11.11.4). Hence, degradation of algae in water could be responsible for the high
SDI75% values observed in summer.
Algae Turbidity SDI(75%)

10000 50
45
40
Algae concentration (cells/mL)
Turbidity (NTU); SDI (%/min)

35
30
1000 25
20
15
10
100 0
Figure 11.11.4. Algal count (points), turbidity (line) and SDI75% (dashed line) in raw seawater
during a HAB between March and April 2008 at the Barcelona pilot plant intake.
443
11.11.4 Impact on Pretreatment Processes

Results from pilot operation demonstrate the specific performance of each pretreatment
process technology (see Table 11.11.3 for comparison). For more information on the
operation of each pilot, see Guastalli et al. (2013).
Table 11.11.3. Quality of seawater treated by DAF followed by either DMF or UF.
Removals are based on the raw seawater (RSW) quality.
DAF DMF UF
% % %
Units Average SD Average SD Average SD
Removal Removal Removal
Temperature °C 18 4 - 18 4 - 18 3 -
pH - 8.0 0.1 - 8.2 0.1 - 8.2 0.1 -
Conductivity mS /cm 56 1 - 57.6 0.4 - 57.5 0.3 -
0.0 0.0
Turbidity NTU 1.5 0.9 38% 0.1 96% 0.06 98%
2 2
a a
TSS mg/L 3 2 58% N.D. - - N.D. - -
SDIb % /min 12 4 42% 2 1 90% 1 1 95%
3 2 4
Cells/
TBC 6 x105 x1 16% 3 x105 x1 57% 7 x102 x10 99.9%
mL
05 05 2
cells
Algal Count 81 89 75% 87 72 74% 3 1 99%
/mL
a
N.D. Not Detected
b
The SDI method used was SDI75% for FW, and SDI15min for DMF and UF waters. TBC = total
bacterial counts, SD = standard deviation.
11.11.4.1 DAF
Optimization of the coagulant dose (0 to 6 mg/L as FeCl3) was specific to the feedwater
quality. The use of acid and coagulant aid was not found to be crucial for the optimal
operation of DAF. During algal blooms, DAF fulfilled its purpose by successfully removing
up to 87% (and 75% on average) of the algal content confirming the benefits of using this
technology when algal blooms events are present.
11.11.4.2 UF
Excellent water quality was obtained by ultrafiltration with no coagulant addition, with high
removal of algae and bacteria. Algal removal by UF increased to 99% on both DAF treated
water and raw water, SDI15 was low (1%/min) and turbidity averaged 0.06 NTU (see Table
11.11.3). Pressurized UF maintained high and stable permeability in direct seawater filtration
at high fluxes (approximately 100 L/m2h at 20 ºC). Filtration cycles were up to 35 minutes.
Successful UF performance was due to the high efficiency of the chemical cleaning protocols
that involved tangential circulation and alternating hypochlorite and citric acid cleaning in
place. Backwash conditions were both air/water and only water.
11.11.4.3 DMF
The dual media filter (sand and anthracite) incorporated all additional equipment for air and
water filter backwash, in the same conditions as full scale. Similar filtered water quality as
UF was observed over a wide range of filtration feed velocities (11-19 m/h), post-coagulant
doses (1-3 ppm as FeCl3) and seawater conditions (normal turbidity, high turbidity, and
adverse climatic events), such that the average SDI15 values were typically between 2.0 and
444
2.8 %/min, turbidity was generally less than 0.13 NTU, algal rejection was 74% on average,
and no residual iron was detected.
11.11.5 Removal of natural organic matter by LC-OCD
An important aspect of the pilot plant project was characterization of Natural Organic Matter
(NOM) and understanding of the performance of different pretreatment options on NOM
removal prior to the SWRO membranes. Liquid chromatography with organic carbon
detection (LC-OCD) was used for the fractionation of the organic carbon into biopolymers
(> 10 kDa), humic substances (humics and building blocks, 0.3-10 kDa) and low molecular
weight compounds (LMW acids and neutrals, < 0.3 kDa).
The program was implemented over a period of 8 months mainly in winter (from July 2009
to March 2010) and
organic carbon was
monitored across the
SWRO treatment line
periodically. LC-
OCD chromatograms
are shown in Figure
11.11.5. Raw sea-
water was mainly
composed of 47% of
low molecular weight
organics (LMW),
47% of humic
substances and 6% of
biopolymers.
The main fraction
rejected after DAF +
UF was biopolymers
(41% removal).
Removal of this
Figure 11.11.5. LC-OCD chromatograms of raw and seawater treated by DAF fraction is relevant
followed by DMF or UF at the Barcelona pilot plant. since biopolymers
contain transparent
exopolymer particles (TEP). TEP are highly sticky and accumulate on membranes that
potentially induce fouling in membranes by providing favorable conditions for biofilm
formation. Humic substances and LMW were only removed in UF permeate to a small
extent, with an 8% and 6% rejection, respectively.
The combined effect of DAF + DMF eliminates 18% of biopolymers; the other fractions
were practically not removed.
11.11.6 Conclusion
Algal counts increased in March and declined in July. Maximum algal cell counts of
approximately 1,200,000 cells/L were observed.
Compared to conventional DMF, the UF showed excellent performance with good quality
and stable permeate. During dense algal blooms, UF yielded almost 100% removal of algal
cells. This value and other results highlight an important advantage of the UF with respect to
the DMF.
445
Acknowledgment: The authors are grateful to CDTI (Spain) through the project CENIT-
Sostaqua (CEN20071039).
11.11.7 References
Guastalli, A. R., Simon, F. X., Penru, Y., de Kerchove, A., Llorens, J., and Baig, S. 2013.
Comparison of DMF and UF pretreatments for particulate material and dissolved organic
matter removal in SWRO desalination, Desalination 322, 144-150.
Mosset, A., Bonnelye, V., Petry, M., and Sanz, M. A. 2008. The sensitivity of SDI analysis:
from RO feedwater to raw water. Desalination 222, 17-23.
Sanz, M. A., Cremer, G., Beltran, F., Bonnelye, V., and Del Campo, I. 2007. Barcelona:
200,000 m3/day of Potable Water coming from Mediterranean Sea. In: Proceedings of
the International Desalination Association World Congress, Maspalomas, Gran Canaria,
Spain.
Simon, F. X., Penru, Y., Guastalli, A. R., Esplugas, S., Llorens, J., and Baig, S. 2013. NOM
characterization by LC-OCD in a SWRO desalination line. Desalination and Water
Treatment 51, 1776-1780.
446
11.12 GOLD COAST, QUEENSLAND, AUSTRALIA - DEEP WATER

INTAKE LIMITS TRICHODESMIUM INGRESS
Dianne L. Turner1, Siobhan F. E. Boerlage2 and Scott Murphy1
1
2
Boerlage Consulting, Gold Coast, Australia
Figure 11.12.1. Location of Gold Coast Desalination Plant in

Queensland (top). Trichodesmium erythraeum bloom off shore at
intake (bottom). Insert - colony of Trichodesmium erythraeum
cells – photo: WHOI.
447
Table 11.12.1. Overview of the Gold Coast SWRO desalination plant.
Plant/Project Name Gold Coast Desalination Plant (GCDP)

Location Tugun, South East Queensland, Australia
Desalination Technology SWRO first pass (SWC5) and BWRO second pass ESPA2+
(m3/d)
RO recovery SWRO 45%, BWRO membrane 85%
Commissioning date February 2009
Intake
Feedwater source South Pacific ocean – Coral Sea
Intake type Deep water intake
Intake description Tunneled intake 1.5 km offshore at 20 m depth. Coarse
screen (2 m high) on intake riser with 140 mm maximum
aperture bars.
Intake screening 3 mm variable speed drum screens.
(shock) chlorination Sodium hypochlorite is dosed at 3-5 mg/L for 2 hours daily
Strategy, dose rate
Online raw water Conductivity, temperature, pH, turbidity, hydrocarbon
monitoring analyzer, total chlorine, oxygen reduction potential(ORP)
Discrete raw water analysis TOC, Total N,
relevant to HAB Total P, Ammonia,
TSS, SDI3
Pretreatment
Process description Ferric sulfate coagulant is injected at second static mixer
(Including UF prescreens, followed by flocculation and gravity DMF consisting of both
cartridge filters etc.) sand and Australian filter coal. 5 µm cartridge filtration prior
to SWRO.
Chemical dosing (acid, 13 mg/L Ferric sulfate for coagulation. Antiscalant is dosed at
coagulants, polymers) 1.7mg/L for first pass and 2 mg/L for second pass RO to
prevent membrane scaling. NaOH is dosed at the second pass
for pH control of 10 for boron removal.
Feedwater design parameters Feedwater during bloom conditions
(Event 2) (Event 3)
Temperature range (˚C) 17 to 28 20.8-23.6 21.5
Salinity range 34,000 to 39,000 Unaltered Unaltered
(TDS mg/L)
Conductivity (mS/cm) 50.0 to 53.0 Unaltered Unaltered
Total Suspended Solids <10 0.5-7 0.2
(mg/L)
SDI3 N/A 8.3 - 29.1 26.4
Turbidity (NTU) 8 0.1 – 1.8 0.7
Organic Matter
TOC (mg/L) <4 N/A N/A
Algal cell count (cells/L) N/A 20 x 106 to 24 x 107 N/A
Algal species Trichodesmium erythraeum
Chlorophyll-a (µg/L) N/A N/A N/A
448
Feedwater design parameters Feedwater during bloom conditions

(Event 2) (Event 3)
TN mg/L N/A N/A N/A
TP (mg/L) N/A N/A N/A
Ammonia (mg/L) 0.2 N/A N/A
DMF Filter rates <8 Unaltered Unaltered
(m/h)
SWRO flux (L/m2h) <16 N/A 12.9
In response to an unprecedented drought, the Queensland Government developed a strategy
for South East Queensland (SEQ), to ensure reliable water supply for the next 50 years. The
strategy comprised a regional Water Grid which linked dams, water treatment plants and
water storages, allowing the transport of water from surplus areas to water deficit areas
(Cannesson 2009). In addition, a key component of that strategy was the construction of the
133,000 m3/d Gold Coast Desalination Plant (GCDP). Due to the drought risk posed to the
reliability of regional water supplies, the GCDP plant was critical to maintaining supplies of
potable water and design and construction were fast tracked.
The GCDP was the first large-scale seawater desalination plant on Australia’s eastern
seaboard (commissioned in February 2009) and is located in Tugun adjacent to an
international airport, approximately 20 km south of the iconic Surfers Paradise and close to
the border with New South Wales (Figure 11.12.1). The plant abstracts seawater from the
pristine embayment of Kirra-Tugun in the Coral Sea that is renowned for fishing, swimming
and surfing. The plant was designed and constructed by the GCD Alliance comprising
Seqwater, Veolia Water Australia, John Holland, with sub Alliance partners; Sinclair Knight
Merz and Cardno and is now operated and maintained by Veolia Water Australia.
Prior to construction, a Seawater Quality Assessment Study was commissioned in November
2005 at the three proposed desalination plant intake sites for the GCDP extending from
Tugun (the selected site) at the southern end of the Gold Coast to 40 km further in the north
(Boerlage 2006). The study provided water quality data for the GCDP design envelope and
information on factors that would impact seawater quality to assist in selecting and designing
pretreatment. Marine algal blooms identified in the report that could occur in the SEQ region
included Trichodesmium, Lyngbya majuscula, Colopomenia sinuosa (cornflake weed),
various brown macroalgae and Anaulus autralis. A marine cyanobacterium Trichodesmium
was identified as the most frequently occurring species throughout the Gold Coast, typically
commencing from late spring to early summer and lasting from four to ten days followed by
Colopomenia sinuosa which occurs sporadically every few years.
Trichodesmium, first described by the English explorer, Captain James Cook in Australian
waters (Beaglehole 1955), is commonly found along the Queensland coast, particularly in the
warmer North tropical regions, the sub-tropical seawater of the Gold Coast and also along the
NSW coastline. It is a ubiquitous genus with blooms also found in the Gulf and other oceans.
Surface blooms of Trichodesmium can be extensive with some covering on the order of
100,000 m2 (Sudek et al. 2006; Mckinna 2015). Under stagnant conditions, T. erythraeum
can release a clear organic compound that changes the bloom’s color from rust brown to
449
green and hence the blooms are commonly mistaken as oil slicks. In addition, the bloom can
release a purple photosynthetic pigment so that the water appears pink; the Red Sea is so
named due to the presence of Trichodesmium blooms and coloration of the water. Decaying
blooms of Trichodesmium spp. may lead to anoxic conditions and mortality and/or an
unpleasant fishy smell may be associated with the blooms.
To avoid operational issues at the GCDP, design measures were incorporated to minimize the
ingress of algae including Trichodesmium into the seawater intake and growth of algae
during pretreatment as discussed herein.
11.12.2 Seawater intake
Design of the seawater intake (and outfall) was challenging as the plant site is located within
the highly developed Gold Coast tourist coastal strip with white sandy beaches and is highly
visible from landing aircraft at the adjacent airport. Various intake options were therefore
considered which would limit community and environmental impacts while delivering high-
quality seawater. The Neodren technology for subsurface intakes was investigated as it was
recognized that this may reduce pretreatment capital and operating costs due to “prefiltration”
of the intake water through sand at a rate similar to slow sand filtration, reducing algal issues,
turbidity, SDI and TSS. Following a Multi Criteria Assessment, a deep-water intake option
comprising an on-shore shaft, tunnel and riser at sea was found to be the most feasible and
preferred solution when considering factors such as cost, risk, scheduled delivery window,
environmental impact, community disruption, visual amenity, and water quality. An
additional factor was that the design had to be capable of meeting the 100 year design life.
This was the first time that tunneling had been used for the intake (and outfall) for a large-
scale SWRO desalination plant anywhere in the world (Burch and Murphy 2008).
The tunnel intake length was finalized to ensure that seabed infrastructure was clear of the
active beach zone and at a depth to provide water with an acceptable quality. Twin tunnels
(2.8m internal diameter) were constructed for the intake and outfall, approximately 1.5 km
off shore and 2.2 km from the plant (see Figure 11.12.2). Design aspects were incorporated to
minimize the ingress of sand, macroalgae (seaweed), algae and marine fauna into the intake.
The intake riser, at 22 m depth was fitted with a vertical coarse bar screen which abstracts
seawater 6 m above the seabed to limit the entrainment of sand and benthic organisms
through the screens (see Figure 11.12.3). The riser was also located in an area devoid of
seaweed, based on marine surveys prior to construction. At the top of the intake riser, a dome
was fitted to convert vertical intake flows to horizontal flows to reduce marine entrainment.
Similarly, the intake flow velocity was designed to operate at 0.05 m/s at the bar screen to
reduce the impingement and entrainment of marine flora (including algae) and fauna
(Cannesson et al. 2009).
The intake riser was located approximately 250 m away from the outfall diffuser. Modelling
was conducted to ensure the brine diffuser achieved high brine exit velocities and dilution
and to determine the footprint of the mixing zone. The intake depth and location took into
account modelling results to prevent recirculation of the brine and any seawater constituents
back into the intake. including dissolved algal organic matter concentrated in the brine.
Diffuser performance was later validated during plant commissioning and operation
(Boerlage and Gordon 2011).
11.12.3 Treatment process overview
The GCDP is required to produce 45.6 million m3/year at 94% availability. It has a modular
design, allowing it to operate at 33, 66 or 100% of its maximum capacity to enable the plant
to respond to dam levels and demand. An overview of the treatment process is presented in
450
GCDP Surfers Paradise
Figure 11.12.2. Self-elevating platform over intake during construction with the GCDP in the background (left)
and Surfers Paradise (right).
Figure 11.12.3. Intake tunnel (left), intake shaft with intake pumps (middle) and the intake coarse screen prior
to installation (right).
Figure 11.12.4. The seawater flows by gravity through the intake tunnel to the plant that is
intermittently chlorinated, typically once a day to limit marine growth. Following fine
screening (3 mm rotating screens), the seawater undergoes conventional pretreatment with
optional pH correction with sulfuric acid, coagulation by the addition of ferric sulfate and a
cationic polymer, followed by flocculation and gravity dual media filtration (DMF) which
will remove intact algal cells, some organics and suspended solids. The DMF and the filtered
seawater tank were enclosed in buildings to prevent sunlight stimulation of algal growth.
Antiscalant is added to the filtered seawater before final filtration through 5 µm cartridge
filters prior to the RO system, which comprises two passes. The first pass is operated with
split permeate extraction. The high-quality front-end permeate is directly transferred to
remineralization while the rear permeate is desalinated in a second pass. The blended
permeate is remineralized by addition of carbon dioxide and lime water, dosed with sodium
hypochlorite, and fluoridated.
Backwash from the DMF is treated in the residuals treatment section of the plant by lamella
thickeners and centrifugation. Solids are transferred to a landfill and the treated water
returned to sea with the RO brine through a purpose-built diffuser. Hence, particulate and
solid algal matter that may be entrained during a bloom will not be returned to sea.
451
Figure 11.12.4. GDCP process schematic for desalination and residual treatment (Burch and Murphy 2008).
11.12.4 Algal bloom events

11.12.4.1 Event 1 (January 2008)
During plant construction, a dense algal bloom was observed at the surface of the seawater in
late January (22nd – 23rd, 2008) during monthly boat surveys. The bloom was extensive,
covering both the desalination plant intake and outlet areas. Similar, more extensive blooms
were observed a year earlier in January 2007, extending up to Surfers Paradise during marine
monitoring surveys. On both occasions, a sample of the dense brown algae was taken which
colored the seawater a profuse pink color following storage. This indicated the bloom was
caused by Trichodesmium, as a quick test for this species is to shake a sample of the bloom
and let stand for several hours after which the water will turn pink/purple due to pigments
dissolving (pers. comm. A. Negri). In the case of the 2008 event, the sample was sent for
analysis and the dominant species was indeed confirmed as the filamentous blue green
cyanobacteria Trichodesmium erythraeum, which was expected from the seawater quality
assessment study.
The prior weather conditions experienced in the Kirra-Tugun embayment were favorable for
an algal bloom, with heavy rainfall (4th, 5th and 11th of January) and flooding of the Tweed
River (5km to the south) bringing high nutrients with it, followed by clear weather, calm seas
and high summer temperatures (Boerlage 2008). The nutrient input due to high rainfall was
not likely to be the driver for the bloom as Trichodesmium erythraeum is well known as a
nitrogen fixing organism (e.g., capable of taking dissolved nitrogen gas from the water)
thriving in nutrient poor tropical- sub tropical waters with temperatures > 20 ˚C. Instead the
combination of rough marine conditions followed by calm winds and warm water
temperatures most likely stimulated the bloom, as such conditions are known to increase the
growth and accumulation of Trichodesmium. Typically, Trichodesmium algal cells are barely
visible to the naked eye and spread throughout the water column. Under such favorable
452
conditions the single filamentous algal cells can aggregate in strings and clumps to form
colonies of up to 200 cells (0.5 to 3 mm in size). As the cells age, they become positively
buoyant and begin to float like sawdust on the surface, forming extensive blooms described
as mats or rafts (Negri et al. 2004). This is what was observed on the Gold Coast in 2007 and
2008 (Figure 11.12.5).
As the GCDP was still being constructed, no impacts on desalination plant operation could be
investigated.
Figure 11.12.5. Algal blooms observed in January 2007 (left) and 2008 at the GCDP seawater intake and outfall
(right). Inset - coloration of seawater sample (2008 bloom) of the photosynthetic pigment – phycoerythrin
produced by Trichodesmium (Boerlage 2008).
11.12.4.2 Event 2 (January to February 2009)
Another bloom similar to that previously observed in 2007 and 2008 was seen the following
summer in late January to mid-February 2009 during monthly boat surveys when the GCDP
was being commissioned. Rafts of the bloom were estimated to cover several hundred
meters. A member of the public suggested it emanated from the plant brine outlet. Surface
water samples collected on two occasions were again an intense pink color and the bloom
was confirmed by laboratory analysis to be a natural bloom event caused by Trichodesmium
erythraeum. The bloom was dense with viable cell counts ranging from 238 million cells/L in
January reducing ten-fold to 20 million cells/L in mid-February. If indeed an algal bloom
were significantly entrained in the deep-water intake, the solids associated with the high
concentration of algal cells would be removed during residual treatment and not be returned
to sea. The brine diffuser was designed and proven to achieve a high dilution with a
minimum footprint (Boerlage and Gordon 2011). Hence, no plume would be evident from the
brine outfall.
As the intake is 1.5 km off shore, visible inspection of the intake area is not possible from the
plant and therefore the exact date for the dissipation of the bloom is not known. As a result,
all water quality data for February 2009 was examined. At the time of this bloom event, the
plant was mainly operating at 33% capacity with no significant impacts observed on raw
seawater quality, operation of the DMF filters, cartridge filters or RO feedwater quality. No
spikes in SDI3 (Figure 11.12.6), or TSS were observed for late January – early February and
all parameters were within the design envelope with the exception of TOC that was 1mg/L
above design on one occasion (shown in Table 11.12.2). The maximum SDI15 value observed
in the DMF filtrate occurred in early February as a one-off measurement and was still within
guarantee values. This may be attributable to limited ingress of the algal bloom due to the
deep-water intake. While Trichodesmium species are generally regarded as “floaters”
453
literature suggests they may sink during die-off. Some algal cells might then be entrained into
the intake resulting in the higher SDI3 values observed for late February (>25). Nevertheless,
no impacts on plant operation were observed.
Figure 11.12.6. Raw water SDI3 measured at the seawater intake during plant commissioning in January –
February 2009.
Table 11.12.2. Intake and filtered seawater quality (following dual media filtration) during a
February 2009 bloom event.
Algal Bloom Data Intake seawater Filtered seawater
Event 2
SDI3 Temperature Turbidity TSS TOC SDI15
°C NTU mg/L mg/L
1st February – Minimum 8.3 20.8 0.1 0.5 0.5 1.1
28th 2009 Maximum 29.1 26.9 1.8 7 5.0 4.0
Average 19.0 23.6 0.7 2.53 1.65 2.8
11.12.4.3 Event 3 (November 2012)

The City of Gold Coast reported that algal blooms were observed in coastal areas on
Monday, 5th November 2012, identified as Trichodesmium from water quality testing by the
City Council. No cell counts were available. The bloom was reported to extend south of
Surfers Paradise down to Coolangatta, an area that incorporates the seawater intake
abstraction site.
Raw seawater quality data for October, the month preceding the bloom, and for November
when the bloom occurred are presented in Table 11.12.3. SDI3 and turbidity are plotted in
Figure 11.12.7. Higher SDI3 and turbidity values were generally found in October. No
noticeable increase in SDI3 (21.6), turbidity (0.639 NTU) and TSS (1.6 mg/L) were observed
on November 5th, the date corresponding to the bloom report by the City Council. Nor was
there any impact on plant operation. The plant was operating on the 5th of November at the
454
desired 33% capacity. The clean-bed head loss was stable for all filters, just under 2.5kPa on
average for the month of November 2012 indicating no fouling/clogging of the DMF.
Similarly to Event 2, all seawater and RO feedwater quality parameters were within the
design envelope and RO membrane guarantee values (Table 11.12.4), respectively. While
limited nutrient data were available, no increases in total nitrogen were observed from that
observed in marine environmental monitoring studies.
Raw water SDI Turbidity NTU

35 2.5
30
2
25
Turbidity (NTU)
SDI3
1.5
20
15
1
10
0.5
5
0 0
5-Nov
1-Oct
6-Oct
10-Nov
15-Nov
20-Nov
25-Nov
30-Nov
11-Oct
16-Oct
21-Oct
26-Oct
31-Oct
Figure 11.12.7. Raw seawater quality in October and November 2012 during which a Trichodesmium bloom
was reported by the Gold Coast City Council on November 5th.
Table 11.12.3. Intake seawater quality data for October and November 2012.
Algal Date Data SDI3 Temp Turbidity TSS Total Ammonia

Event 3 statistics °C NTU mg/ Nitrogen mg/L
L mg/L
Oct. Min 17.2 19.4 0.225 0.2 N/A N/A
2012 Average 23.0 21.6 0.698 1.84 N/A N/A
5 Nov. Max 28.8 24.0 2.21 5.2 0.09 0.079
2012 Nov. Min 15.2 20.9 0.249 0.2 <0.05 N/A
2012 Average 20.4 22.9 0.411 1.16 <0.05 N/A
max 26.6 24.5 0.795 3 <0.05 0.015
Table 11.12.4. Monthly RO feedwater quality before and after algal event in November
2012.
Algal event Water quality Data SDI15 Turbidity TOC
Dates month date statistics NTU mg/L
October Min 2.21 0.01 1

2012 Average 2.95 0.045 1
5 November
Max 3.69 0.135 1
2012
November Min 2.13 0.01 1
2012 Average 2.57 0.025 1
Max 3.51 0.05 1
455
11.12.5 Long-term plant operation

The GCDP plant has been in operation for more than seven years, with anecdotal reports of
Trichodesmium blooms similar to algal events 1-3 described above occurring during the
summer months when conditions are expected to promote blooms. In the absence of any
observed impacts at the plant, algal blooms are likely to be unreported as the intake is 2.5 km
from the plant and algae are not specifically monitored for by the GCDP.
Blooms of another marine algal species, Colpomenia sinuosa, have been detected
sporadically, once in 2010 and again in 2014, through an increase in the mass of intake
screenings. In the latter case, this resulted in the unusually high mass of screenings
(approximately 400 kg). Colpomenia sinuosa, a globe-shaped brown macroalga, is much
larger than Trichodesmium, ranging in size from 1 to 3 mm and up to 10 mm. It is found in
intertidal areas growing on rocks or other algae. Colpomenia sinuosa is typically in the
attached form, it becomes dislodged with wind and waves breaking into flakes and may then
have become entrained into the intake.
In general, the depth of the intake and pretreatment design is believed to have largely
mitigated algal bloom impacts, especially surface Trichodesmium blooms with limited
ingress of Colpomenia. Water quality associated with algal blooms does not appear to have
deteriorated - SDI3, TSS and TOC were within the range observed throughout the year. No
operational issues were observed at the plant such as increased DMF clogging. No sudden
increases in RO operating pressures due to biofilm or adsorptive organic fouling have been
observed associated with the potential increase in organics from an algal bloom entrained at
the seawater intake.
Membrane autopsies and the low cleaning frequency of the RO trains further support the
contention that no organic fouling or significant biofouling has occurred during plant
operation. Autopsies of a cartridge filter, first and second pass RO membranes in May of
2009 following algal event 2 showed no evidence of organic deposits or fouling with only
minor inorganic deposits. Similar autopsies were conducted in August 2009, March 2010,
August 2010 and April 2013. In general, visual inspection of lead membranes showed that
the level of fouling was very light with only minor biofilm deposits present. The loss of
ignition results indicated that the organics on the membrane were low.
Only limited CIP cleans of RO membranes have been carried out, which is attributed to the
effectiveness of pretreatment coupled to good raw water quality. Clean in place have not
been conducted in response to any fouling issues but rather as a trial run on a limited number
of RO trains or as preventative cleans.
11.12.6 Conclusions
Trichodesmium, the most frequent bloom-forming species in the region, occurred prior to
operation of the GCDP and continues to occur during the summer months as a result of
marine and meteorological conditions that promote the growth of this genus. Dense surface
blooms have been extensive, covering the seawater intake and brine discharge area with a
density of up to 238 million cells/L. Occasional blooms of Colpomenia have also occurred.
The deep water (20 m water depth) intake option selected for the GCDP appears to be
successful in providing good water quality and preventing the ingress of Trichodesmium,
which typically floats at the surface with limited ingress of Colpomenia. No increase in raw
water TOC, TSS, SDI3 have been observed as a result of blooms. Membrane autopsies also
show no organic or biofilm fouling due to the potential increase in organics associated with
such dense blooms.
456
11.12.7 References
Beaglehole, J. C. 1955. The Journals of Captain James Cook on His Voyages of Discovery. I.
The Voyage of the Endeavour 1768–1771. Cambridge: Cambridge University Press.
Boerlage, S. F. E. 2006. Seawater quality investigation at proposed desalination intake sites
Consultancy Report.
Boerlage, S. F. E. 2008. Identification of marine algal bloom, Gold Coast Desalination
Alliance, Internal Memo.
Boerlage, S. and Gordon, H. 2011. Assessing diffuser performance and discharge footprint
for the Gold Coast Desalination Plant. In: Proceedings of IDA World Congress, Perth,
Western Australia.
Burch, R. and Murphy, S. 2008. Twin intake/outfall tunnels save space and environment.
IDA Journal of Desalination and Water Reuse 18(2), 22-28.
Cannesson, N., Johnstone, P. M., Mitchell, M. A., and Boerlage, S. F. 2009. Minimizing
community, environmental, and marine impacts at the Gold Coast Desalination Plant.
IDA Journal of Desalination and Water Reuse 1(1), 74-87.
Negri, A. P., Bunter, O., Jones, B., and Llewellyn, L. 2004. Effects of the bloom-forming
alga Trichodesmium erythraeum on the pearl oyster Pinctada maxima. Aquaculture
232(1), 91-102.
McKinna, L .I. 2015. Three decades of ocean-color remote-sensing Trichodesmium spp. in
the World’s oceans: a review. Progress in Oceanography 131, 177-199.
Sudek, S., Haygood, M. G., Youssef, D. T., and Schmidt, E. W. 2006. Structure of
trichamide, a cyclic peptide from the bloom-forming cyanobacterium Trichodesmium
erythraeum, predicted from the genome sequence. Applied and Environmental
Microbiology 72(6), 4382-4387.
457
11.13 BERLIN, GERMANY – AKVOLA: AN INTEGRATED DAF-UF

PILOT
1
Johanna Ludwig1 and Matan Beery
1
Figure 11.13.1. Pilot plant location at Landwehrkanal Berlin.
Figure 11.13.2. Pilot plant at Landwehrkanal Berlin.
459
Table 11.13.1. Overview of the akvola Technologies’ pilot plant.
Plant/Project Name akvoFloat pilot at Berlin City River

Location City River (Landwehrkanal), Berlin, Germany
Total Production
Capacity (m3/d) 15-17
Commissioning date January 2013
Intake
Feedwater source Surface Water (River: Landwehrkanal), Berlin
Intake type Surface intake
Intake description Intake depth 0.5-1 m, distance from shore 2 m
Intake screening 15 mm mesh
Discrete raw water TOC, turbidity, algae cell count, temperature, pH
analysis relevant to
HABs
Pretreatment
Process description 300 µm automatic strainer, coagulation-flocculation, flotation and
MF/UF (submerged out-in)
Chemical dosing Coagulant FeCl3 8-12 mg/l as Fe; Polyacrylamide 0.1 mg/l
Feedwater design parameters Feedwater during Annual avg.
bloom conditions feedwater
conditions
Temperature range (˚C) 5-25 14-21 0.2-25
Salinity range 490-550 490-505 410-526
(TDS mg/L)
Conductivity (mS/cm) 660-830* 732-753 612-785
Total suspended solids 1.7-12* 5.5-12 1.7-12
(mg/L)
SDI
Turbidity (NTU) 1.5-7.5 2.1-7.5
Organic matter 7.8-10.8 (TOC) 8.6-10.2 (TOC)
Algal cell count (cells/L) 6.4 - 56 million* 30 - 56 million 6.4 - 56 million
Algal species Blue-green algae (60- Blue-green algae (82- Blue-green algae
95%), Diatoms (3- 95%), Diatoms (3- (60-95%),
37%) 8%), Diatoms (3-37%)
Main species: Main species:
Pseudanabaena Pseudanabaena
limnetica, Anabaena limnetica,
planctonica, Anabaena
Aphanizomenon flos- planctonica,
aquae, Aphanizomenon
Aphanizomenon flos-aquae,
issatschenkoi* Aphanizomenon
issatschenkoi*
460

Feedwater design parameters Feedwater during Annual avg.
bloom conditions feedwater
conditions
Chlorophyll-a (µg/L) 2-18* 7-18 < 2 - 18
Additional relevant water 3-11 mg/L DO*, 5-7 mg/L DO, 3-15 mg/L DO,
quality parameters for 0.09-0.27 mg/L TP* 0.2-0.27 mg/L TP 0.06-0.27 mg/L
design or observed spikes TP
during algal bloom e.g.
low DO at intake, H2S,
ammonia
Filter rates 0.8-10 m/h (including 0.8 m/h (not optimized)
(DAF) m/h both flotation and
filtration area)
2
UF flux (L/m h) 107-167 146-158
*
Data measured by third party: Berlin’s Senate Department for Urban Development and the Environment
(Department VIII E 24 Integrative Ecology)
Originally developed for SWRO pretreatment as a reliable and energy efficient technology,
akvoFloat was first piloted at a freshwater river. The pilot results are valid for seawater as
well, as laboratory studies have shown flotation bubbles to be 5-10% smaller (and thus more
efficient) in saltwater than in freshwater. Seawater application requires corrosion resistant
materials for all parts, but the overall performance of akvoFloat in terms of separation and
efficiency will be similar or better. (See Chapter 9 for a detailed comparison on DAF
conditions in fresh water vs seawater).
Specializing in the treatment of hard-to-treat waters in the industrial, commercial and
municipal sectors, akvola Technologies is a water technology company with a focus in oil-
water separation and suspended solids removal for applications such as SWRO pretreatment.
In 2013, pilot tests of the akvoFloat technology, comprising a hybrid flotation-filtration
process using ceramic membranes in both steps, were conducted using Berlin’s city river
water at 0.5-0.7 m³/h. The goal was to determine the optimal operating conditions giving a
constantly low transmembrane pressure of the MF/UF membrane. An algal bloom occurs
every summer at the river. More information can be found in Hög et al. (2015) and Beery et
al. (2014).
11.13.1.1 Feedwater
In the pilot plant, water from the Berlin city river, Landwehrkanal (a branch of the river
Spree), was treated during summer to autumn 2013. The period was characterized by
fluctuating levels of total organic carbon (TOC), turbidity and temperature (Figure 11.13.3).
The slow flow (10 cm/s) of the river and low depth (2 m) promotes occasional algal blooms
in summer. Furthermore, the canal is used to relieve the sewage system during heavy rain
events. Phytoplankton counts close to the water inlet are given in Figure 11.13.4, using data
provided by Berlin Senate Department for Urban Development and the Environment
(Department VIII E 24 Integrative Ecology). An algal bloom was declared when the
phytoplankton concentration was higher than 30 million cells/L (Aug-Oct.), see Figure
11.13.4 and the image in Figure 11.13.5.
461
temperature [°C], pH 25
20
TOC [mg/l],
15
10
5
0
03/13 05/13 06/13 08/13 10/13 11/13 01/14
TOC Temp pH
Figure 11.13.3. Average TOC, pH and temperature during pilot plant testing.
100% 60,000
Phytoplankton
Percentage of Total
Phytoplankton
Cells/ml
40,000
50%
20,000
0% 0
May June July Aug Sept Okt
May June July Aug Sept Oct
Cyanophyceae Bacillariophyceae Centrales Total Phytoplankton

Figure 11.13.4. Algal cell count close to the water inlet, data provided by the Berlin Senate Department for
Urban Development and the Environment (Department VIII E 24 Integrative Ecology).
11.13.2 Process Description

The akvoFloat process is a hybrid of
flotation and membrane filtration with
inline coagulation/flocculation (iron (III)
chloride). The production of micro-
bubbles for flotation is achieved by a
bubble generator, composed of ceramic
diffusor disks, giving 50-100 µm bubble
size. A ceramic MF/UF membrane is
submerged in the flotation tank operating
at constant flux. The membrane is
backwashed with permeate which
simultaneously removes the float layer
hydraulically over a weir. Backwashes
typically had a two-minute duration and a
volume flow of 1200 L/h reaching 2.1 bar
absolute (shorter backwash intervals <30s
yield the same result, as shown in later
experiments).
Figure 11.13.5. Water inlet during an algal bloom.
Figure 11.13.6 shows the process scheme
of the pilot plant, Figure 11.13.7 shows a photo of the pilot plant and the resulting
sludge/float layer during algal bloom operation.
462
Figure 11.13.6. Process flow schematic of the pilot plant.
Figure 11.13.7. Side view of the akvoFloat pilot (left) and resulting float layer during operation (right).
11.13.2.1 Process Results

The process results are shown in Figure 11.13.8. Turbidity of the feed was constantly reduced
by up to 95% with typical effluent turbidity < 0.1 NTU. TOC was reduced by 30-40%, with
practically 100% of the microalgae removed within the test accuracy. No difference in plant
performance was observed during algal bloom conditions – in permeate quality or plant
performance (fouling, backwash effectiveness). The transmembrane pressure was always 0.1-
0.2 bar. A combination of a CEB/CIP with sodium hypochlorite and a subsequent CIP with
citric acid, 30 minutes each, allowed the complete recovery of the membrane performance at
any given time.
11.13.3 Conclusions
The high efficiency of the akvoFloat process for surface water treatment during algal bloom
conditions was demonstrated. A periodic backwash (2h) and a flux of up to 150 L/m2h was
shown to be sustainable, yielding low overall fouling, maintaining up-time and high quality
permeate during an algal bloom. A combination of a CEB with sodium hypochlorite and a
CIP with citric acid allows the full recovery of the membrane performance. akvola is
currently collaborating with a large international EPC running trials for seawater
pretreatment but the results are thus far confidential.
463
Figure 11.13.8. Water quality of feed and treated water.
11.13.4 References
Beery, M., Ludwig, J., and Hög, A. 2014. River Water Treatment using Ceramic Membranes
in Berlin: Pilot Testing. In: Proceedings of the American Membrane Technology
Association Membrane Technology Conference, Las Vegas, Nevada, USA.
Berline Senate, Department for Urban Development and the Environment (Senatsverwaltung
für Stadtentwicklung und Umwelt Berlin), Abt. Integrativer Umweltschutz, Monitoring
Wasserrahmenrichtlinie. 2014. “Ökologisches Monitoring und Bewertung von
Oberflächengewässern VIII E 24”, Berlin.
Hög, A., Ludwig, J., and Beery, M. 2015. The use of integrated flotation and ceramic
membrane filtration for surface water treatment with high loads of suspended and
dissolved organic matter. Journal of Water Process Engineering 6, 129-135.
464
Appendix 1 – Harmful algal species

Appendix 1
ALGAL SPECIES POTENTIALLY HARMFUL TO DESALINATION

OPERATIONS
David G. Borkman1 and Donald M. Anderson2
1
Pausacaco Plankton, Saunderstown, RI USA
2
Woods Hole Oceanographic Institution, Woods Hole, MA 02543 USA
1 INTRODUCTION
It is now well established that harmful algal blooms (HABs) represent a serious and growing
threat to seawater reverse osmosis (SWRO) desalination plants worldwide. In many plants,
these threats are indirectly monitored using parameters such as the Silt Density Index (SDI)
or chlorophyll-a (see Chapter 5), but these only provide a general indication of the particulate
fouling propensity of the water or the abundance of phytoplankton, respectively. Although it
is often a challenge to obtain data on the phytoplankton species composition and abundance
in the raw seawater, such information can be of great value in the long-term operation of
desalination plants. Individual algal species vary dramatically in their properties and
therefore in the extent to which they can disrupt plant operations (e. g., through the
production of toxins that represent a potential threat to the safety of the drinking water
produced, or organic matter that can clog filters and foul membranes). As a result, it is
important for a desalination plant to make (and record) species identifications, and the
concentrations of those species that are in the source seawater, particularly those that have
disrupted normal plant operations. As described in Chapter 3, monitoring programs for
seawater outside a plant and process monitoring at the plant can provide this type of
information.
Identification of the algal species in seawater samples can be a challenge however. In Chapter
3, methods for sample collection, fixation, and identification are presented. Section 3.6.1.1
lists books that provide useful taxonomic information on marine HAB species, while section
3.6.1.2 lists websites where taxonomic information on algal species can be found. To
augment this information and to provide a quick resource for operators or managers who
need identification assistance, this appendix presents brief descriptions and a photograph of
some algal species that either have caused problems at desalination plants, that produce
potent toxins, or that are known to produce sufficient organic matter or biomass to be
problematic. The list of species covered here is not comprehensive, as this is not intended to
be an operator’s sole source of taxonomic information. Instead, it is offered as a quick
reference guide. For example, there are more than 30 species in the Alexandrium genus, and
about half of those are toxic, but only three are described here. Readers are urged to refer to
the many other resources that provide more detailed descriptions and photographs.
In this manual, we define toxic algae as those that produce potent toxins (i.e., poisonous
substances produced within living cells or organisms), e.g., saxitoxin. These can cause
illness or mortality in humans as well as marine life through either direct exposure to the
toxin or ingestion of bioaccumulated toxin in higher trophic levels e.g. shellfish. Confusion
arises, however, because non-toxic HABs can also result in mass mortalities of fish and other
marine life. In this instance, the mortality results from the indirect effect of compounds
produced by the algae - compounds that do not have specific targets or receptors, but instead
are more general in their mechanism of damage, sometimes requiring chemical modifications
by other compounds to become lethal. Examples of “harmful” but not “toxic” substances are
reactive oxygen species that, when combined with polyunsaturated fatty acids, can rapidly
465

kill fish and other animals. Another example is a proteinaceous compound produced by
Akashiwo sanguinea that accumulates on bird feathers, causing a loss in natural water
repellency and widespread mortality of affected animals.
In this appendix, species that do not produce toxins but that do cause marine mortalities are
termed “harmful”.
The following glossary defines some of the taxonomic terminology used here.
Antapex – posterior-most (bottom) part of the cell body, excluding spines, lists and
similar structures
Apex – anterior-most (top) part of the cell body
Areolate – ornamentation of the thecal plates that consists of depressions of variable
depth and form
Cingulum – a furrow encircling the cell that contains the rotary flagellum
Cyst – the diploid zygotic dormant state in the sexual life cycle; usually morphologically
dissimilar from the haploid motile stage; also called the ‘dinocyst’ or
‘hypnozygote’
Dorsal Side – back side of the cell, opposite of the front or ventral side where the sulcus
is located
Epicone – the part of a dinoflagellate cell above the cingulum; usually refers to an
‘unarmored’ (lacking cellulose plates) cell; may also be known as the epitheca or
episome
Epitheca – the part of a dinoflagellate cell above the cingulum; usually refers to a thecate
(with cellulose plates) cell; may also be known as the epicone or episome
Flagellar Area – the vicinity of the origin of the two flagella
Hypocone – the part of a dinoflagellate cell below the cingulum; usually refers to an
‘unarmored’ (lacking cellulose plates) cell; may also be referred to as the
hypotheca or hyposome
Hypotheca – the part of a dinoflagellate cell below the cingulum; usually refers to a
thecate (with cellulose plates) cell; may also be referred to as the hypocone or
hyposome
Lists – membranous extensions of the cingulum and/or sulcus that extend beyond the cell
wall boundary
Pores – openings in the theca that can be involved in the extrusion of certain structures
from the cell
Sulcus – a longitudinal furrow, often partially enclosing the propulsive flagellum
Thecate – having a cell wall of cellulose plates which have special designations and
symbols according to their location on the cell
466

Akashiwo sanguinea
Dinoflagellate
A medium- to large-sized dinoflagellate

with cell divided into approximately equal-
sized conical epitheca and bilobed
hypotheca, strongly cleaved by the sulcus.
Cells are 40-80 µm long and 40-60 µm
wide. Chloroplasts are numerous and
golden; nucleus is large and centrally
located.
Harmful. Produces surfactants that can be

associated with bird and fish kills. Photo:
http://planktonnet.awi.de/repository/rawdata-
PlanktonNet2/viewable/alexandra_akashiwo_080115_
1_20150110174310_small.jpg
Alexandrium (genus description)
Alexandrium is a genus of planktonic, thecate (i.e., with rigid, cellulose cell walls)
dinoflagellates. Many Alexandrium spp. are toxic and are the causative organisms for
paralytic shellfish poisoning (PSP) outbreaks. All Alexandrium spp. in the descriptions that
follow share these features: spherical to sub-spherical shape, a descending cingulum (with
the right end of the cingulum displaced approximately one cingulum width), numerous
golden-brown chloroplasts, and the presence of a comma-shaped aperture in the apical pore
plate (Po plate). Species identification is based on thecal plate tabulation that requires
specialized staining and microscopy. The descriptions below will aid in identification of
Alexandrium spp. based on general morphology that is visible using light microscopy.
Several common Alexandrium species are described below.
Alexandrium catenella
Dinoflagellate
Cells spherical to sub-spherical in shape and

approximately 28-50 µm long by 25-45 µm
wide; typically slightly longer than wide.
Can be solitary, or in chains with 2, 4, and 8
cells. Note that there are five species
recognized in what was once considered the
A. tamarense complex that includes A.
catenella (see John et al., 2014).
Toxic. Produces multiple toxins within the

saxitoxin family, causes paralytic shellfish Photo: D.M. Anderson, Woods Hole Oceanographic
poisoning (PSP). Institution
467

Alexandrium monilatum
Dinoflagellate
Cells flattened anteriorly-posteriorly and are

20-30 µm long by 30-45 µm wide; often
forms chains of cells up to 32 cells or
longer. The cingulum (e.g., groove halfway
between the top and bottom of the cell) is
deep, relatively wide and easily seen in light
microscopy. The nucleus is located in the
center of the cell, and the many chloroplasts
radiate outward from the nucleus
Toxic. Produces goniodomin, linked to fish

and shellfish kills. Has not been associated Photo: W.M. Jones, III, Virginia Institute of Marine
with human toxicity although shellfish can Science.
be impaired by reduced filtration.
Alexandrium ostenfeldii
Dinoflagellate
Cells are relatively large (40-60 µm long by

40-50 µm wide) and have a nearly spherical
shape. The epitheca has a slightly more
pointed or conical shape than the hypotheca.
The hypotheca often has a central flattened
area. The cingulum is relatively shallow and
has no lists along it and the sulcus is very
shallow and difficult to see in light
microscopy. This species has a
characteristic large ventral pore on the first
apical plate.
Toxic. Produces saxitoxins and spirolides, Photo:

has been associated with paralytic shellfish Canadian Register of Marine Species (CARMS)
poisoning (PSP). http://images.vliz.be/resized/40248_alexandrium-
ostenfeldii.jpg
468

Aureococcus anophagefferens
Pelagophyte
Small (2 µm diameter) solitary spherical

cells with dark brown color. Due to the
extremely small cell size and non-descript
features, electron microscopy or
immunofluorescent (antibody-based)
methods are needed for positive
identification.
Harmful. Not known to produce toxins, but

can form “brown tides” or dense blooms
(109 cells L-1) that cause water discoloration.
Associated effects include shellfish
mortality (shellfish may not consume Photo: Koonja Yang, State University of New York at
Aureococcus cells and can thus lose Stony Brook.
viability) and seagrass mortality due to light
attenuation during brown tide blooms.
Azadinium/Amphidoma species
Dinoflagellate
Some species of Amphidomataceae are

producers of azaspiracids, a group of
lipophilic toxins responsible for azaspiracid
shellfish toxicity. Up to now four of the ~20
species of the family have been found to
produce toxins, Azadinium spinosum (A),
Az. poporum (B), Az. dexteroporum (C),
and Amphidoma languida (D). All are
small, solitary thecate dinoflagellates;
photosynthetic, with single chloroplast that
is branched into both episome and
hyposome; one or more pyrenoids (arrows)
present, large nucleus usually visible in the
hyposome. Note that there are other non-
toxic species that are very similar at the
light microscopy level, e.g., Heterocapsa.
Cells: Depending upon the species,

approximately 7-15 mm long and 5-12 mm
wide.
Photos: U. Tillmann,
Alfred Wegener Institute.
Toxic. Some species produce azaspiracid
toxins, responsible for azaspiracid shellfish
toxicity.
469

Chaetoceros socialis
Centric diatom
A colonial Chaetoceros that forms

spherical, mucous–bound colonies
comprised of thousands of cells; colonies
may reach millimeter-scale diameter. C.
socialis is in the subgenus Hyalochaete, so
its setae (spines) are thin and do not contain
chloroplasts. Individual cells are 3 – 15 µm
in diameter and contain a single chloroplast.
A distinctive feature is that one setae is
much longer than the others, and this fourth,
elongate setae extends towards the center of
the spherical colony.
Cells: 3 – 15 µm diameter
Colonies up to several mm in diameter
Photo: http://planktonnet.awi.de/
Harmful. Non-toxic, but heavy mucilage
producer.
Cochlodinium polykrikoides
Dinoflagellate
An athecate (fragile cell walled)

dinoflagellate that may occur as single cells
or as long chains of cells. Cells are variable
in shape, ranging from elongate to sub-
spherical; brown rod-shaped chloroplasts
present. Epitheca is slightly conical while
hypotheca is bi-lobate. The key feature is a
transverse cingulum that wraps
approximately two times around the cell
and the placement of the sulcus. 35-50 µm
in length, 35-50 µm in width. Note,
Cochlodinium fluvescens is toxic, and
Photo:
morphologically similar. http://www.serc.si.edu/labs/phytoplankton
/guide/addtl_collections/Cape%20Cod/
Harmful. Non-toxic, but is associated with Cochlopoly.aspx
fish kills, larval mortality; also produces
large amounts of mucilage.
470

Coscinodiscus wailesii
Girdle view
Centric diatom
A large (to 500 µm diameter), usually

solitary centric diatom with numerous,
irregular-shaped chloroplasts. In girdle view
the mantle is tall and meets the valve at a
right angle. The valve face has a clear
central hyaline area and has rows of striae
radiating from the central area to the valve
edge. Two rings of marginal processes may
be seen with light microscopy in some
preparations.
Valve view
300-500 µm in diameter
200-400 µm in height
Harmful. Not toxic, but considered a

nuisance taxon due to formation of dense
mucilage-producing blooms.
Photo: http://planktonnet.awi.de/
Cylindrotheca closterium
Pennate diatom
A slender, spindle-shaped pennate diatom

approximately 50 to 400 µm long and <10
µm in width at the widest central region.
Spindle-shaped central region with two thin,
tapering extensions of the valve. Wide,
central region of the cell has two
chloroplasts. Nitzschia longissima is similar
in size and shape, but is more heavily
silicified.
50 - 400 µm in length
3 – 10 µm in width

nuisance taxa due to release of mucilage and Photo: http://planktonnet.awi.de/
alleopathic chemicals when present in dense
blooms.
471

Dinophysis spp.
Dinoflagellate
Dinophysis spp. are thecate dinoflagellates characterized by a distinctive dinophysoid body

shape. This form features a small epitheca with a cingular list developed into a “collar” and
a much larger hypotheca with a large left sulcal list. Dinophysis spp. cells range from small
(20 µm long) to large (100 µm long). Species are distinguished by size, shape in lateral
view, and presence of distinctive hypothecal horns in some species.
Most Dinophysis spp. are toxic, producing okadaic acid and other toxins responsible for
diarrhetic shellfish poisoning (DSP); these other toxins include dinophysistoxins (DTXs).
Several common Dinophysis species are described below.
Dinophysis acuminata
Dinoflagellate
Medium sized, thecate cells with a small

collar-like epitheca (~1/5 of cell) and much
larger (~4/5 of cell) hypotheca. Compressed
laterally and usually viewed on microscope
slide in lateral view. Cells ovoid to round in
overall profile, with left sulcal list that
extends from epitheca past the midline of
the cell. Cell surface has visible pores and
sometimes at the antapex (bottom of the
hypotheca), a series of bumps or
protrusions. Key features distinguishing D.
acuminata from other Dinophysis spp. are
smaller size, ovoid shape, and protrusions
on the antapical end of the cell
Size: 40 – 60 µm long by 30 – 40 µm wide.

Photo:
http://planktonnet.awi.de/index.php?contenttype=ima
Toxic. Produces okadaic acid and other ge details&itemid=60229#content
toxins causing diarrhetic shellfish poisoning
(DSP).
472

Dinophysis fortii
Dinoflagellate
Large (60 – 80 µm long by 40 – 60 µm

wide) thecate dinoflagellate with
dinophysoid body shape: small epitheca
with collar and much larger hypotheca and
large left sulcal list. Distinguished in lateral
view by the large, rounded hypotheca, with
widest point of cell near the bottom of the
cell. Cell surface covered with depressions
and pores that are visible with a light
microscope.
Size: 60 - 80 µm long by 40 – 60 µm wide.
Toxic. Produces okadaic acid and

dinophysistoxins (DTXs) that are
responsible for diarrhetic shellfish poisoning
Photo:
(DSP). http://planktonnet.awi.de/repository/rawdata-
PlanktonNet2/viewable/vera_veloso_dinophysis_forti
i c01_xxw_20070621124535_small.jpg
Dinophysis tripos
Dinoflagellate
Large cells (90-120 µm long by 50-60 µm

wide) with dinophysoid body shape, with
the hypotheca formed into two posterior
projections or horns.
Non-toxic.
Photo:
http://planktonnet.awi.de/repository/rawdataPlankton
Net2/viewable/alexandra_dino_tripos_a1_ps95_stn2_
20m_20160224104606_small.jpg

473

Gambierdiscus spp.
Dinoflagellate
Multiple toxic species in this genus. As an

example, Gambierdiscus toxicus is a
solitary, benthic thecate dinoflagellate.
Relatively large (40-150 µm in diameter,
50-150 µm in dorso-ventral length), sub-
spherical shaped (slightly flattened anterio-
posteriorly) so that the cells appear in apical
or antapical views (i.e., from the top or
bottom) in the counting slide. The epitheca
and hypotheca are approximately the same
size. Numerous chloroplasts visible. Cell
surface with visible pores. This is an
epiphytic species, meaning that it lives on
surfaces like dead coral, seaweeds, sand, Photo:
etc. https://www.whoi.edu/redtide/photos/cfp?tid=542&ci
d=83388&c=3&idx=2&slideshow=30812
Toxic. Gambierdiscus species can produce
ciguatoxins, gambiertoxins and maitotoxin

Gonyaulax fragilis, Gonyaulax hyalina
Dinoflagellates
Small thecate dinoflagellates with a shallow

sulcus and thin thecal plates. Numerous
chloroplasts present. Epitheca is slightly
pointed and an elongate apical pore complex
is moderately visible in light microscopy at
the apex. Hypotheca is more broad and
rounded than the epitheca.
Gonyaulax fragilis and G. hyalina are

similar in size and shape and can be
differentiated by displacement of the
cingulum. The cingulum of G. hyalina is
displaced by one cingular width while that
of G. fragilis is displaced (offset
longitudinally) by greater than one width. Gonyaulax fragilis. Photo: F. Guerrini, University of
30 - 45 µm in length; 20 – 30 µm in width. Bologna.

harmful taxa because blooms can release
large amounts of mucilage.

474

Gymnodinium catenatum
Dinoflagellate
An athecate marine dinoflagellate that often

forms multi-cell chains. Individual cells are
30-45 µm in diameter by 35-65 µm in
height. Cells have a distinct cingulum
separating a slightly smaller, conical
epitheca from the slightly larger and
squared-off hypotheca. The cingulum is
displaced 1.5 to 2-times it width. The sulcus
divides the hypotheca into two lobes.
Numerous chloroplasts are visible.
Toxic. Produces PSP toxins (saxitoxins).

Photo:

http://planktonnet.awi.de/ - content

Heterocapsa circularisquama
Dinoflagellate
Small photosynthetic dinoflagellate

approximately 15-20 µm wide by 20-30 µm
in length. Cells have a conical epitheca that
is more pointed than the round,
hemispherical hypotheca. H.
circularisquama is differentiated from other
species by the presence of circular organic
scales with radial ridges, general body
shape, shape and position of the nucleus.
Study of scales requires advanced
microscopy. Photo:
http://feis.fra.affrc.go.jp/hcaphp/biology_e.h
Toxic. Blooms have been associated with tm
shellfish mortalities.
475

Heterocapsa triquetra
Dinoflagellate
Small (20 - 30 µm long by 15-20 µm wide

at sulcus) thecate dinoflagellate with bi-
conical shape. Has distinct transverse
cingulum that divides cell into two
approximately equal halves. The hypotheca
is conical and formed into a point; the
epitheca is also conical, but more rounded
than the hypotheca. Chloroplasts are brown.
Harmful. Not toxic, but a bloom former that

can cause high-biomass, water-discoloring
blooms.
Photo:
http://nordicmicroalgae.org/taxon/Heterocapsa%20tr
iquetra?media_id=Heterocapsa%20triquetra_4.jpg

Heterosigma akashiwo
Raphidiophyte
Small (15-20 µm diameter) spherical to

ovoid shaped cells with yellow-green to
brown colored chloroplasts that give the cell
a ‘lumpy’ appearance. Two flagella visible
in live material.
Harmful. Associated with fish kills, but

mode of toxicity and fish mortality not well
understood. Mortality appears to be caused
by the production and release of free fatty
acids that are chemically altered and made
toxic by reactive oxygen species.
Photo: http://www.marinespecies.org/hab/
476

Karenia brevis
Dinoflagellate
Cells are flattened dorso-ventrally resulting
in an oblong or square shape of 20 – 45 µm
wide (along cingulum) and 15 – 40 µm
long. An apical bump is usually visible on
the epicone and the cingulum and sulcus are
deep and prominent, with the sulcus
extending well into the epitheca. Nucleus is
located on the left side of the hypocone;
numerous yellowish green chloroplasts are
visible as is a long trailing flagellum.
Toxic. Produces brevetoxins responsible for
neurotoxic shellfish poisoning (NSP),
marine vertebrate deaths; aerosolized toxin Photo:
www.nmfs.noaa.gov/pr/pdfs/health
causes respiratory irritation in humans. /brevetoxin.pdf
Karenia mikimotoi
Dinoflagellate
Cells are approximately ovoid and are 20-30
µm wide by 15-30 µm long with numerous
yellow-brown chloroplasts. The relatively
deep cingulum and sulcus appear to divide
the cell into four lobes The cingulum is
descending, displaced approximately twice
the cingulum width. The sulcus extends into
the epitheca. The nucleus is often at the
periphery of the cell on the left side.
Toxic. Produces gymnocins. Blooms have
been associated with marine invertebrate Photo:
http://nordicmicroalgae.org/taxon/
mortalities and fish kills. Karenia%20mikimotoi
Karlodinium veneficum
Dinoflagellate
Small ovoid dinoflagellate with single, wide,
distinct cingulum that is displaced 2+
cingular widths along the cell length. Cells
are 15-20 µm long by 10-15 µm wide, have
a centrally located nucleus and several
chloroplasts. Difficult to distinguish from
other small gymnodinoid species.
Toxic. Produces karlotoxins and is Photo:
associated with fish kills. https://microbewiki.kenyon.edu/index.php/Karlodini
um_veneficum
477

Noctiluca scintillans
Dinoflagellate
A distinctive, large dinoflagellate with a

sub-spherical shape, a single flagellum and a
large, curving tentacle with visible
transverse striations. Cells pinched slightly
at the area of tentacle emergence giving a
sub-spherical shape. This species is
heterotrophic and uses the tentacle in
feeding. Although it lacks its own
chloroplasts, photosynthetic symbionts or oil
and pigment accumulations typically give
the cytoplasm (and blooms) a green or pink
color. Bioluminescent.

300 µm to 2 mm in diameter
Harmful. Non- toxic, but forms dense, Photo:

http://planktonnet.awi.de/
water-discoloring blooms. Ammonia present
in the cytoplasm may accumulate to toxic
levels and has been linked to fish and
invertebrate kills. Very common bloom
former in the Arabian Sea and Gulf region.
Ostreopsis siamensis
Dinoflagellate
A benthic, thecate dinoflagellate that is

anterio-posteriorly flattened and typically is
presented on microscope slides in apical or
antapical view. Cells are oval to tear-drop
shaped and are large (100-135 µm long by
75-95 µm wide) in apical/antapical view.
Cell surface covered with visible pores.
Identification to species based on
measurement ratios, pores, and cingulum
orientation.
Photo:
http://www.whoi.edu/science/B/redtide
Toxic. Produces palytoxins. /species/cfp_images.html
478

Phaeocystis spp.
Phaeocystis spp. have a complex life cycle that alternates between a small (<10 µm
diameter), motile, bi-flagellate single cell form and a large (cm-scale) colonial form
comprised of up to thousands of cells. Single cells require specialized microscopy to
identify, but colonial forms can be identified based on colony morphology. Some bloom-
forming species are described below.
Phaeocystis antarctica
Colony forming Prymnesiophyte
Phaeocystis antarctica forms spherical

colonies of up to 2 mm in diameter. Cells
are distributed evenly throughout the
periphery of the colony. Found in Southern
Hemisphere in the Southern Ocean.
Colonies 20 µm to 2 mm in diameter
Cells 2-6 µm in diameter
Harmful. A nuisance alga if dense blooms

are agitated by wave action to create beach
foaming. Abundant mucilage also a
Photo:
potential problem with desalination plants. http://www.bco-dmo.org/project/546758
Phaeocystis globosa
Phaeocystis globosa forms spherical

colonies that are much larger (up to 3 cm in
diameter) than those of P. antarctica or P.
pouchetii. Cells are distributed evenly
throughout the periphery of the spherical
colonies. Found in the North Atlantic and
Indian Oceans.
Colonies 25 µm to 3 cm in diameter

are agitated by wave action to create beach Photo:
potential problem with desalination plants.
479

Phaeocystis pouchetii
Phaeocystis pouchetii is the only

Phaeocystis spp. to form lobed colonies.
The colonies are generally smaller (up to 2
mm in diameter) than those of P. antarctica
or P. globosa. Unlike P. globosa and P.
antarctica, cells within a colony have a
clumped distribution with more cells in the
edges of lobes, and often cells are in groups
of four. Found in the Northern Hemisphere
in polar to boreal waters of the Atlantic and
Photo:
Pacific Oceans. http://planktonnet.awi.de/
Colonies 25 µm to 2 mm in diameter
are agitated by wave action to create beach
potential problem with desalination plants.
Prorocentrum spp.
Prorocentrum spp. are laterally flattened thecate dinoflagellates that are generally ovoid to
almond-shaped and range from small (10-15 µm wide P. minimum) to large in size. Have
two anterior flagella and are photosynthetic; cell surface often covered with visible pores.
Species preliminarily distinguished by size, shape, pore and poroid patterns, and presence or
absence of apical processes. Some bloom forming species are described below.

Prorocentrum lima
Dinoflagellate
Medium-sized cells approximately 35-40

µm long by 25-30 µm wide with ovate to
ellipsoid shape being, typically, widest at
posterior of cell. The anterior end of the left
valve is straight while the anterior end of the
right valve has a triangular concavity.
Surface of cell covered with pores except
for flattened central area. Photo:
Toxic. Produces okadaic acid, implicated in Net2/viewable/maria_antnia_sampayo_prorocentrum_
lima_culture_2cells1theca_
DSP and ciguatera poisoning.
480

Prorocentrum micans
Dinoflagellate
Prorocentrum micans is a medium to large

size (30 – 70 µm long by 25 – 50 µm wide)
thecate dinoflagellate that generally has a
teardrop shape. Cells are flattened and are
pointed at the posterior end, more rounded
at the anterior end and are widest across the
middle of the cell. Cells have variable
lengths and widths, but usually have a
length-to-width ratio of 2:1. A distinctive
anterior spine (~ 10 µm long) is generally
visible in light microscopy.
Harmful. Toxicity is not confirmed for this Photo:

species, but it can form dense, water- http://planktonnet.awi.de/repository/rawdataPlankton
discoloring blooms that have been linked to Net2/viewable/alexandra_pro_mic
shellfish mortality and reduced water he461_e1_180416_20160509114831_small.jpg
column oxygen.
Prorocentrum minimum
Dinoflagellate
Small cells (15-20 µm long by 10-15 µm

wide) with oval to triangular shape.
Flattened on one side, often giving a “D”
shape. A single apical spine often visible.
Species also known as P. cordatum, a

synonym.
Toxic. Some strains may produce toxins;

some blooms associated with shellfish
poisoning and fish kills.
Photo:
Net2/viewable/fjouenne_sbrbill0165w_20080304121
251_small.jpg
481

Pseudo-nitzschia spp.
Pennate diatom
There are multiple toxic Pseudo-nitzschia

species in this large genus. The example
shown here, P australis, is a toxic form. It is
very difficult to distinguish between
Pseudo-nitzschia species using conventional
light microscopy. When that level of
identification is required, scanning electron
microscopy or molecular approaches are
typically used.

Photo:
Toxic. Some Pseudo-nitzschia species K. Hubbard, Florida Fish and Wildlife Conservation
produce domoic acid. Commission

Scrippsiella trochoidea
Dinoflagellate
Relatively small (15 – 35 µm long by 20 –

25 µm wide) thecate dinoflagellate having a
pointed or conical epitheca and a rounded
hypotheca. The epitheca has a short, blunt
apical process while the hypotheca is
smooth and lacks any spines or processes.
Photosynthetic, with numerous gold-brown
chloroplasts visible.
Harmful. Not toxic, but can cause high-

biomass, water-discoloring blooms.

Photo:
J. Lewis, University of Westminster

482

Trichodesmium erythraeum
Colonial cyanobacterium
Planktonic colonial cyanobacteria forming

bundles or fascicles of cells that are visible
with the naked eye. Cells are 7-15 µm wide
and are joined into chains (trichomes) that
may be 50 to 800 µm long. Numerous
trichomes are joined by mucilage to form
bundles or tufts that can form water-
discoloring blooms.
Cells: 7 -15 µm in diameter

Photo:
Trichomes: 30 - 800 µm in length and 7 – http://www.whoi.edu/oceanus/v2/article/
15 µm in width images.do?id=62206
Toxic. Bundles form dense, water-

discoloring blooms and are reported to
contain toxins including microcystins and
palytoxins.

Tripos furca (=Ceratium furca)
Dinoflagellate
Occurs as solitary cells with 2 flagella; often

forms large mono-specific blooms. Cells
contain numerous yellow-brown
chloroplasts. The cells are straight, slightly
dorso-ventrally flattened and widest at the
cingulum area. The epitheca tapers
gradually into the anterior horn. Hypotheca
is subtrapezoid, extending into a long left
and a short right antapical horn, which are
usually straight in line with the cell and may
be slightly toothed along the sides. The left
horn is twice as long as the right. Thecal
plates are ornamented with a reticulum of
ridges and pores. The nucleus is situated in
the epitheca. 150-230 µm in length, 30-35
µm in width. Note: All marine Ceratium
species were reassigned to the genus Tripos
by Gómez et al. (2013). (CICIMAR
Oceanides 28:1-22). Photo:
Harmful. Non toxic, but can cause dense

blooms with ecosystem impacts, often due
to oxygen depletion.
483

Tripos fusus
Dinoflagellate
Occurs as solitary cells with 2 flagella;

contain numerous yellow-brown
chloroplasts. The cells are elongate, straight
to slightly curved and spindle-shaped in
overall appearance. Only one long antapical,
left horn. Right horn rudimentary or only
slightly developed. Cell is 15-35 µm wide at
center (near transverse cingulum) tapering
to several um wide at tips. Length 125 to
300 µm, 15-35 µm in width.
Harmful. Non toxic, but can cause dense

blooms with ecosystem impacts, often due
to oxygen depletion.
Photo:
484
Appendix 2 – Rapid screening methods for HAB toxins
Appendix 2
RAPID SCREENING METHODS FOR HARMFUL ALGAL BLOOM

TOXINS
Donald M. Anderson1, Maurice Laycock2, and Fernando Rubio3

1
2
Scotia Rapid Testing Ltd, Chester Basin, Nova Scotia Canada
3
Abraxis LLC, Warminster, PA USA
1 Introduction .................................................................................................................................................485
2 ELISA methodology ...................................................................................................................................486
2.1 Abraxis ELISA kit for saxitoxin .........................................................................................................487
2.1.1 Principle .......................................................................................................................................487
2.2 Abraxis ADDA-DM ELISA kit for microcystins/nodularins .............................................................489
2.2.1 Principle .......................................................................................................................................489
2.3 Abraxis ELISA kit for Anatoxin-a (ATX) ..........................................................................................490
2.3.1 Principle .......................................................................................................................................490
2.4 Biosense ELISA Kit for Domoic acid (DA) .......................................................................................491
2.4.1 Principle .......................................................................................................................................491
3 Imunochromatographic or lateral flow format ............................................................................................492
3.1 The Scotia tests....................................................................................................................................492
3.1.1 Principle .......................................................................................................................................492
3.2 Abraxis microcystins finished drinking water immunochromatographic tests ...................................494
3.2.2 Sample preparation methods .......................................................................................................494
3.2.2.1 Toxins from phytoplankton .................................................................................................495
3.2.2.2 Preparation of shellfish samples ..........................................................................................496
3.3 Obtaining quantitative data with Abraxis ELISA reader ....................................................................497
3.3.1 Essential steps to obtain data with the Abraxis + reader .............................................................497
4 Summary and conclusions ..........................................................................................................................499
5 References ...................................................................................................................................................499
1 INTRODUCTION
At the core of all national harmful algal bloom (HAB) programs are the monitoring
programs needed to detect HAB toxins in shellfish, fish, water, or other resources
sufficiently early to take management actions (Anderson et al. 2001). These programs
measure toxins produced by multiple species of algae, with the methods used varying
dramatically in scope and complexity due to the types of toxins that need to be detected, the
nature of the affected resource, and regulatory requirements.
Some of the methods developed for analysis of shellfish tissues and algal blooms can be of
direct use in desalination plants for analysis of toxins in water – both the raw, untreated
water before desalination, and the treated, fresh water. A major concern, however, are the
detection limits of the assays. All analytical methods have limits of detection (LODs) and
the choice of a method should be consistent with potential bloom concentrations and
possible toxin levels. With desalination plants, toxins need to be measured at exceedingly
low levels in water, whereas shellfish concentrate toxins to much higher levels. A recent
study summarized the epidemiological data for four common algal toxins (Laycock et al.
2010) and estimated the potential contamination of water that might enter a desalination
plant during major blooms. The assessment was based on a hypothetical (and dense) bloom
of toxic algae consisting of 107 cells/L with a toxin cell quota of 40 pg toxin/cell. If all of
that toxin were released from the cells into the water, that would give a concentration in
485
seawater of 400 µg/L. An alternative approach to estimating the total amount of toxin
present in a bloom is given in Chapter 1 (Table 1.4), where the amounts of toxin contained
in hypothetical blooms of various common HAB species are presented. The values range
from a few hundred to 1,000 µg/L. Given that 99% or more of a toxin is likely to be
removed by thermal or reverse osmosis desalination (Chapter 10), the sensitivity of an
analytical method must therefore be at least 0.1 – 1.0 µg/L or 0.1 – 1.0 ng/mL. Therefore,
analysis of water samples for dissolved or particulate toxins (i.e., inside algal cells) will
require high sensitivity methods, such as enzyme-linked immunosorbent assays (ELISAs).
For example, the LOD for saxitoxin (STX) using the Abraxis STX ELISA kit is 0.02 ng/mL
and there is similar sensitivity for domoic acid.
This appendix presents details on simple screening methods for HAB toxins. More complex
analytical methods are described or cited in Chapter 2. The screening methods are presented
here as a guide to desalination plant staff who wish to conduct on-site analyses. These
analyses could be of raw intake water, treated water, or algal cell extracts from monitoring
programs (Chapter 3).
The example assays are restricted to four HAB toxins i.e., saxitoxins, domoic acid,
microcystins/nodularins, and anatoxin-a. Although sample preparation procedures may
differ for the other HAB toxins not include here, the commercial ELISA kit protocols are
similar to each other. Sample preparation procedures, however, vary depending on solubility
of the toxins, source (e.g., phytoplankton, shellfish, or cyanobacteria) and method of
analysis. Sample preparation methods will be described in detail, as will procedures used to
obtain samples. Methods of analysis other than ELISA are also presented.
Lateral flow tests (such as the Scotia tests) are described as simpler alternatives to the
ELISA kits. The advantages and disadvantages of both tests will be discussed.
2 ELISA METHODOLOGY
For the purpose of demonstrating antibody methods that are potentially useful for
monitoring HAB toxins in water, four commercial ELISA kits are discussed here: saxitoxins
(STXs), microcystins/nodularins (MCT/NOD), anatoxin-a (ATX) manufactured by Abraxis,
and the domoic acid (DA), manufactured by Biosense Laboratories. Other kits are
commercially available – these specific kits are presented as examples, not as endorsements.
Antibodies are especially useful for analysis of HAB toxins because of their high sensitivity
and specificity. Since the introduction of the ELISA format in 1971 by Engvall and
Perlmann (1971), this microtitre plate format has been used for thousands of clinical
applications. For a review of ELISAs for shellfish toxins, see Usleber et al. (2001).
Antibodies to the common HAB toxins are relatively easy to prepare using purified toxins
conjugated to carrier proteins that are injected into rabbits or other animals from which
crude serum can be used. Most other components for this technology are commercially
available. Nevertheless, the basic ELISA requires a well-equipped laboratory and skill in
performing the numerous steps involved. Equipment required includes a plate reader (Figure
1) and multi-channel pipettes.
There are several ELISA plate readers available commercially at a cost of approximately
$10,000 (Figure 1). A plate reader also requires a computer to record and plot data. A
simple reader is available from Abraxis for ~ $1,200 and is described below.
486
Aside from being very sensitive, antibody methods such as ELISA are also very specific to a
compound or group of closely related compounds that occur as mixtures. For the STXs, of
which there are more than 20 different derivatives or congeners (Chapter 2), anti-STX
antibodies have different affinities for the individual congeners, i.e., different cross-
reactivities. The congeners also have different specific toxicities that vary over a wide range.
For example, the Abraxis antibodies have poor cross-reactivities for GTX1,4. The limitation
of antibody methods for saxitoxins is thus that quantitative test results can be misleading for
both concentration and toxicity. The net result is that antibody methods for naturally
occurring mixtures of saxitoxins tend to underestimate toxicity.
Figure 1. An ELISA plate reader (left) and a multi-channel pipette.

The standard ELISA is an indirect method and is therefore a more lengthy procedure than
that used in the commercial kits described below where a direct method is used. Fewer steps
are involved in using kits because a solution of toxin-enzyme conjugate or antibody-enzyme
conjugate (e.g., STX-HRP (horse radish peroxidase) or anti-domoic acid-HRP conjugate)
are provided and competition occurs directly with free toxin. In an indirect method, the
enzyme is linked to a second antibody (sheep anti-rabbit) that becomes attached to wells that
have bound primary (rabbit) antibody on their surfaces. Wells coated with a toxin conjugate
(as in the DA kit) would not bind the primary antibody if there were sufficient toxin in the
sample to outcompete the toxin conjugate. The net result is less color with free toxin and
more color with less toxin.
2.1 Abraxis ELISA kit for saxitoxin
2.1.1 Principle
Competition occurs between STX in samples and a STX-HRP conjugate. With no STX in
the sample, all STX-HRP conjugate binds to the rabbit anti-STX antibody that becomes
bound to the sheep anti-rabbit antibody well coating. With excess STX in the sample, no
STX-HRP conjugate binds to the rabbit anti-STX antibody. There is thus no HRP attached
to the well coating, resulting in no color when the enzyme substrate (TMB) is added. The
steps of the procedure can be seen in Figure 2.
This kit has been developed for testing shellfish, as well as drinking and fresh water. To
obtain accurate results when analyzing seawater samples using the Abraxis Saxitoxin
ELISA Kit, Seawater Matrix Standards and an alternate testing procedure are necessary.
Seawater samples which exceed the calibration range of the assay must be diluted using the
Seawater Matrix Sample Diluent and re-analyzed.
487
Figure 2. Schematic of the Abraxis saxitoxin ELISA procedure.
Twelve strips of eight micro-titer wells pre-coated with sheep anti-rabbit antibody are
provided with the kit. The coating procedure is equivalent to steps 1-5 in the standard
ELISA protocol given above.
Well coating Sheep anti rabbit antibody
Well solution STX solution 50𝜇L +STX-HRP 50𝜇L + STX antibody 50𝜇L
Incubate 30 min
Wash 4x 300𝜇L
TMB solution 100𝜇L
Incubate 30 min
Stop 100𝜇L
Read 450 nm
Limit of detection 0.02 ng/mL
488
2.2 Abraxis ADDA-DM ELISA kit for microcystins/nodularins

2.2.1 Principle
There are over 140 microcystins/nodularins congeners known. The antibody used in ADDA-
DM ELISA has been designed for congener-independent detection of both toxin families.
The antibody binds the ADDA part of the molecule which is common to all toxic congeners.
This kit is an example of a direct ELISA in which competition occurs between microcystins/
nodularins (MCT/NOD) in samples and a MCT-HRP conjugate. With no MCT/NOD in the
sample, all MCT-HRP conjugate binds to the mouse anti-ADDA antibody that becomes
bound to the goat anti-mouse antibody well coating. With excess MCT/NOD in the sample,
no MCT-HRP conjugate binds to the mouse anti-ADDA antibody so there is no HRP
attached to the well coating resulting in no color when the enzyme substrate (TMB) is added.
Twelve strips of eight micro-titer wells pre-coated with goat anti-mouse antibody are
ELISA protocol given above. The steps of the procedure can be seen in Figure 3.
Figure 3. Schematic of the Abraxis microcystins ELISA procedure.
489
This kit has been developed for testing drinking and fresh water. The manufacturer provides
a technical bulletin for testing brackish and sea water which involves the addition of one
reagent to the sample to remove salt.
Well coating Goat anti mouse antibody
Well solution MCT solution 100𝜇L +MCT-HRP 50𝜇L + MCT antibody 50𝜇L
Incubate 90 min
Wash 4x 300𝜇L
TMB solution 150𝜇L
Incubate 20-30 min
Stop 100𝜇L
Read 450 nm
2.3 Abraxis ELISA kit for Anatoxin-a (ATX)
2.3.1 Principle
This is a direct ELISA. Competition occurs between ATX in samples and ATX-HRP
conjugate. With no ATX in the sample, all ATX-HRP conjugate binds to the mouse anti-
ATX antibody that becomes bound to the goat anti-mouse antibody well coating. With
excess ATX in the sample, no ATX-HRP conjugate binds to the mouse anti-ATX antibody
so there is no HRP attached to the well coating resulting in no color when the enzyme
substrate (TMB) is added.
Twelve strips of eight micro-titer wells pre-coated with goat anti-mouse antibody are
ELISA protocol given above. The steps of the procedure can be seen in Figure 4.
This kit has been developed for testing drinking, fresh water and sea water. Brackish and sea
water can be analyzed without any sample treatment.
Well coating Goat anti mouse antibody

Well solution ATX solution 50 𝜇L +ATX-HRP 50𝜇L + ATX antibody 50𝜇L
Incubate 60 min
Wash 4x 300 𝜇L
TMB solution 100 𝜇L
Incubate 20-30 min
Stop 100 𝜇L
Read 450 nm
490
Figure 4. Schematic of the Abraxis anatoxin-a ELISA procedure.
2.4 Biosense ELISA Kit for Domoic acid (DA)

2.4.1 Principle
This is an example of an indirect ELISA. Competition occurs between DA in samples and
the well coating in the presence of an antibody-HRP conjugate. With no DA in solution, all
antibody-HRP binds to the well giving strong color. With excess DA in solution no
antibody-HRP binds to the well and no color develops.
Well coating DA-protein conjugate
Well solution Sample 50 𝜇L + Antibody-HRP conjugate 50𝜇L
Incubate 60 min
Wash 4X 300 𝜇L
TMB solution 100 𝜇L
Incubate 15 min room temperature
Stop with 0.3M H2SO4 100 𝜇L
Read 450nm after 2-5 min
491
3 IMUNOCHROMATOGRAPHIC OR LATERAL FLOW FORMAT
These are simple devices intended to detect the presence/absence of a target analyte in a
sample without the need for specialized or costly equipment. The lateral flow format for
these tests is simple and fast compared to ELISA, and makes them suitable for field-work
without laboratory equipment. They are useful when it is only necessary to show the
presence of a toxin and not suitable to try to measure toxin concentrations. They are less
sensitive than ELISA and only semi-quantitative in that toxin concentration can only be
estimated from the sensitivity of the test. They are not deemed suitable for testing of low
levels of toxins in water samples, but can be easily used to test for toxins in concentrated
plankton net tow extracts (Chapter 3).
The benefits of immunochromatographic tests include:
• User-friendly formats
• Short time to get results
• Long-term stability over a wide range of climates
• No need for specialized or costly equipment
• No need for experienced lab personnel
• Can be run on-site or in a laboratory setting
3.1 The Scotia tests
Scotia Rapid Testing Ltd. manufactures lateral flow immunochromatographic (LFI) tests for
three HAB toxins: the saxitoxins, domoic acid, and okadaic acid. These kits were formerly
known as Jellett Rapid Test Kits, but are now called Scotia Rapid Test Kits.
3.1.1 Principle
In a similar manner to the ELISA kits, immnunochromatographic tests are competition
assays. Free toxin competes with bound toxin for specific antibodies. Anti-toxin antibodies
are attached to colloidal gold particles (diameter 40 nm) that are dried during manufacture
onto a fiberglass pad on the test strip. When a sample is added, the red colored gold particles
are carried by capillary action along a nitrocellulose membrane where they interact with a
line consisting of a toxin-protein conjugate known as the test or T-line. If there is no toxin in
the sample, the antibodies interact with bound toxin forming a red line of colloidal gold
particles. If, on the other hand, toxin is present in the sample, it competes with bound toxin
on the test line and fewer gold
particles stick to the line resulting
in no red color or a fainter line
depending on the concentration of
toxin in the sample. Further along
the membrane, the particles cross
another line that consists of a
protein recognizable by other
antibodies on the particles. This is
a control line that should bind
particles when the test is working
properly.
The Scotia tests for shellfish toxins
can be used as spot tests on a single
Figure 5. The Scotia rapid test kits. sample with minimal and
unsophisticated equipment as
492
shown in Figure 5. Figure 6 shows a schematic of the lateral flow process, and Figure 7 an
example of the results.
Figure 6. Instructions for using the Scotia STX lateral flow assay. Source: Scotia Rapid Testing.
Figure 7. Lateral flow immnochromatography (LFI).
493
3.2 Abraxis microcystins finished drinking water immunochromatographic tests

Abraxis offers several immunochromatographic tests for microcystins/nodularins.The
antibody used is the same as in the ELISA test kit, and therefore offers a broad reactivity for
congener detection. Available test kits are for source waters, recreational waters, and
finished drinking water. The differences are detection limits and the option to lyse the
cyanobacterial cells. The drinking water, source water, and recreational water test kit are not
recommended for estuarine or sea water because of salt interference with the test. The
drinking water test can be used to analyze finished drinking water from desalination plants,
and it has a sensitivity of 0.5 ng/mL. Chlorinated finished drinking water must be quenched
immediately with sodium thiosulfate.
3.2.1 Principle
The test is based on the recognition of microcystins, nodularins, and their congeners by
specific antibodies. The toxin conjugate competes for antibody binding sites with
microcystins/nodularins that may be present in the water sample. The test device consists of
a vial with specific antibodies for microcystins and nodularins labeled with a gold colloid
and a membrane strip to which a conjugate of the toxin is attached. A control line, produced
by a different antibody/antigen reaction, is also present on the membrane strip. The control
line is not influenced by the presence or absence of microcystins in the water sample, and
therefore, should be present in all reactions. In the absence of toxin in the water sample, the
colloidal gold labeled antibody complex moves with the water sample by capillary action to
contact the immobilized microcystins conjugate. An antibody-antigen reaction occurs
forming a visible line in the ‘test’ area.
The formation of two visible lines of similar intensity indicates a negative test result,
meaning the test did not detect the toxin at or above the cut-off point established for the
assay.
If microcystins are present in the water sample, they compete with the immobilized toxin
conjugate in the test area for the antibody binding sites on the colloidal gold labeled
complex. If a sufficient amount of toxin is present, it will fill all of the available binding
sites, thus preventing attachment of the labeled antibody to the toxin conjugate, therefore
preventing the development of a colored line. If a colored line is not visible in the Test Line
Region, or if the Test Line is lighter than the negative Control Line, microcystins are present
at the levels of concern (>1 ppb). Semi-quantitative results can be obtained by comparing
the test line intensity to those produced by solutions of known microcystin concentrations
(control solutions). The steps of the procedure can be seen in Figure 8.
3.2.2 Sample preparation methods
There is no general method for sample preparation for all HAB toxins from a variety of
sources. The same methods can be used for both Abraxis and Scotia tests except for
solutions used to dilute samples. These solutions are provided with the kits. Methods depend
on the following considerations.
Type of organism. Shellfish have fairly uniform extractability, whereas some algal
cells with rigid cell walls are difficult to extract in plankton samples. These cells can
be most easily disrupted by sonication or freeze-thaw cycles.
Solubility. Molecular characteristics such as charge distribution determine the
degree of water solubility. Saxitoxins and domoic acid are readily soluble in aqueous
solutions. For less polar compounds like okadaic acid, methanol is a useful solvent
as it can be easily removed by evaporation, and at low concentrations (<10%) does
494
Figure 8. Schematic of the Abraxis Microcystins Finished Drinking Water Lateral Flow procedure.
not interfere with antibody reactions. An additional problem when working with
hydrophobic compounds (e.g., brevetoxins) is that they are readily adsorbed onto
surfaces; plastic containers should therefore be avoided.
Method of analysis. The final steps in sample preparation provide a solution that is
compatible with the analytical procedure. Antibodies are susceptible to organic
solvents, pH, and salt concentrations that can interfere with immuno-reactions or
alter their protein folding.
3.2.2.1 Toxins from phytoplankton
Initial water sampling can be done with a plankton net (see Chapter 3). Because cells occur
at various depths, the net can be raised through a column of water to obtain an integrated
sample. Alternatively, if a bloom is visible and at the surface, a known volume of water can
be filtered through a 20-cm diameter Nitex sieve and further concentrated onto a 1-cm
diameter filter using a syringe. (Figure 9). For saxitoxins, the filter can be transferred to a 4-
mL glass tube for extraction with 0.1M hydrochloric acid. The extract is heated at 100 ℃ for
495
10 min. This step is used to convert C toxins

to gonyautoxins which are more readily
detected.
Cells from a plankton net and Nitex filters
can be concentrated on a 1-cm diameter
fiberglass filter pad which is transferred to a
4-mL glass tube for extraction with 0.5 mL
dilute hydrochloric or acetic acid.
The Biosense method for extracting Pseudo-
nitzschia cells taken from Fehling et al.
(2004) involves sonicating in sample buffer
(PBS: phosphate buffered saline). If an
ultrasonic apparatus is not available,
Figure 9. Tools needed to filter and concentrate a freezing and thawing several times should
water sample for analysis.
be effective.
The Biosense protocol recommends extracting 4 g homogenate with 10 mL 50% methanol
followed by centrifuging. There is no heating step. For more details see the Biosense
website http://www.biosense.com.
The following is an example modified from the Biosense manual. The original example was
for shellfish tissue, but similar procedures would apply to concentrated plankton net tow
samples. Because ELISA kits are very sensitive and quantitative, a series of dilutions of
both calibration standards and extracts is necessary to obtain data in the dynamic range (i.e.,
on the slope of the calibration graph). For Scotia tests, the ASP buffer provided with the
tests is used for diluting samples.
3.2.2.2 Preparation of shellfish samples
Extraction of DA from plankton samples
1)
Accurately weigh a known mass of sample into a 50 mL centrifuge tube.
2)
Add 16 mL of Extraction solution (50% methanol).
3)
Mix well by vigorous shaking for 1 min.
4)
Centrifuge at 3000 x g for 10 minutes at room temperature.
5)
Retain the supernatant for further dilution prior to analysis. The extracts can be stored
at -20˚C for up to 14 days, although with a possible reduction in DA content.
Dilution of extracts
1) Prepare dilutions of the shellfish extract in Standard/Sample buffer (10% methanol in
PBS-T) as follows:
1:20 dilution: 50 µL shellfish extract + 950 µL buffer
1:200 dilution: 50 µL of the 1:20 dilution + 450 µL buffer
1:2000 dilution: 50 µL of the 1:200 dilution + 450 µL buffer
1:20 000 dilution: 50 µL of the 1:2000 dilution + 450 µL buffer
1:200 000 dilution: 50 µL of the 1:2000 dilution + 450 µL buffer
Cap and vortex each dilution before proceeding to the next dilution step.
2) Analyze the sample dilutions according to the DA concentration range of interest (see
Table 1), to give absorbance values within the calibration curve working range.

496
Table 1. Shellfish extract dilution for quantification of domoic acid.

DA concentration range of interest Corresponding sample extract dilution
(mg/kg) to be analyzed
0.01 - 0.25 1:200 (minimum dilution)
0.1 - 2.5 1:2,000 dilution
1.0 - 25 1:20,000 dilution
10 - 250 1:200 000 dilution
3.3 Obtaining quantitative data with Abraxis ELISA reader

Results obtained using ELISA can be measured using a laboratory ELISA reader. There are
several such instruments available which cost between $5,000 and $25,000. They are not
designed to be portable, weighing around 10 kg. Abraxis has, however, developed a
relatively simple portable instrument for field use, such as on survey vessels.
3.3.1 Essential steps to obtain data with the Abraxis + reader
Detailed instructions and software are provided with the Abraxis reader (Figure 10). A
simplified step-by-step guide to using the instrument is provided below.
1. Load the software from the disc to a computer.
3. Connect the reader and computer.
3. Plug the 12 volt DC supply into the reader.
4. Click on the M6+ icon to launch the computer software.
5. Set up a COM port (see operation manual).
6. Make sure the M6+ is showing "Abraxis LLC" before selecting the COM port for the
connection between M6+ and the
computer. If not it may be necessary to
restart the computer.
7. In the "add a parameter" function
(Green button A) choose a name,
concentration units, concentration range,
filter wavelength and curve type. (The
default parameter is for microcystins). If
this has already been done for saxitoxin
and domoic acid, click on the one for the
current analysis.
Note: 1. The displayed data will be lost
message is OK. The loss is temporary.
Note: 2. The parameter window is
'greyed out'. This is normal.
8. Click 'Set up' (top row middle on
screen). Select OK in comport setting
window if the correct comport is Figure 10. Portable Abraxis plate reader.
showing (COM5). This connects the
M6+ and PC (showing "on line").
497
9. Click on the third green button (capture data) to take readings from a strip of wells. In the
capture data window on the PC select "Clear sample data and get real time reading then OK.
A blank must be taken at this stage.
10. The M6+ window should show: Abraxis LLC. If not, scroll round the options by
pressing SEL/ESC several times until it appears.
11. Insert a strip with water (blank) in a well until it clicks into position in the light path then
press READ/ENTER on the M6+.
12. The reading will be 0.000 ABS., then repeat before going to the next level
13. On M6+ press SEL/ESC twice to get to level [2] PARAMETER. Here the number of
standards is shown, e.g. 6, but should correspond with those listed in the parameter with
their respective concentrations. Press READ/ENTER to select the number of replicates.
14. In the screen on the M6+ the number of replicates can be selected. Unlike the number
of standards which corresponds to the pre-determined PARAMETER any number of
replicates can be chosen for the standards at this stage up to a maximum of 5 by pressing
SEL/ESC repeatedly then press READ/ENTER.
15. Now the number of sample replicates can be chosen up to a maximum of 5 by pressing
SEL/ESC repeatedly (usually 2).
16. Press SEL/ESC to get to level [3] STANDARD. Here the standards can be read. Press
READ/ENTER to read the first Standard. The M6+ screen shows S1-1 Standard. Read
Standard? Press READ/ENTER. The screen will show S1-1 and the absorbance (eg 0.528A).
After S1-2 the PC screen will show the two values and the average. To move to standard
number two press SEL/ESC. Move the strip to the next well and press READ/ENTER. Press
SEL/ESC. The M6+ screen will ask if you want to continue. Press READ/ENTER to
continue.
17. Move the strip of wells ahead after the replicates have been read for the standard. The
readings and averages will appear on the PC screen under the Data tab. If they do not, repeat
steps 9-17. After the second replicate of standard 6 the M6+ screen will show END.
18. To read the samples press SEL/ESC to go to level [4] SAMPLE.
19. Press READ/ENTER. M6+ screen shows T01-1 Sample and Read Sample? Press
READ/ENTER repeatedly to read the replicates of the first sample. The PC screen will
show the absorption values for each replicate and their average.
18. Continue through all samples, then click on the green button to stop reading.
19. To see the calibration graph and sample concentrations click on the quantitative tab to
show the calibration graph and sample concentrations.
20. To print the quantitative results click File on the top line then print. If no printer is
attached to the PC save as a pdf to an external memory (flash drive).
20. To save the data so that this existing calibration graph can be used for new samples
without generating another graph each time samples are analyzed (and using up the limited
number of wells) scroll up to [6] RS232 to PC in the M6+ window followed by
READ/ENTER to M6 -OK->PC. Then capture data (third green button) by selecting upload
batch data in the capture data window ("not clear sample data and get real-time reading" as
before). Click OK in message window "Please load STD data".
21. Samples should be read using a calibration graph prepared at the same time. However,
for approximate values calibration data in the memory can be used.
498
22. Select [6] RS232 to PC then enter to show M6->PC in the M6 window. Then Capture
Data in the PC screen.
23. Select Upload batch data in the PC screen to show the data for calibration.
24. To read new samples using the uploaded calibration click on Capture Data button and
select "Clear sample data and get real time reading, then OK. The sample data will be erased
from the table under data in the PC screen.
25. Select [4] SAMPLE. The M6 screen will show "Rewrite Memory-are you sure?
26. Press read/enter on M6 and read the first new sample. If two replicates were chosen for
samples - read again.
27. The new sample data will be shown in the Table in the Quantitative tab on the PC.
28. To save data after obtaining sample concentrations
a. Under File select Load data
b. Select the protocol, eg Saxitoxin
c. Click on the data capture icon
d. Select Real Time Reading
e. Select Read Sample on M6+
f. Stop reading
g. Save and give a new name
4 SUMMARY AND CONCLUSIONS
Desalination plants should establish a capability for screening samples for HAB toxins. This
would not be a routine analysis, but instead an opportunistic approach triggered by
identification of toxic cells in plankton samples, or other indications that toxic species are
entering the plant. It is advisable for operators in areas subject to toxic HABs to have some
lateral flow or ELISA kits on site, as it can take days or even weeks to get new kits during a
crisis. ELISA kits are suitable for use when an estimate of toxin concentration is necessary.
Lateral flow tests are more suitable than ELISA for routine monitoring of phytoplankton
and shellfish extracts, but are not sensitive enough for water analysis, nor are they
quantitative.
If a positive result is indicated by the ELISA or the lateral flow tests, samples should be
collected and evaluated with more sophisticated analytical instruments. A fully equipped
analytical laboratory capable of measuring algal toxins is unrealistic at a desalination plant,
but capabilities for HAB toxin analysis exist within many countries or regions at
government agencies or universities. Operators should have contact details and sampling
protocols prepared as part of their Alert Level Framework actions (Chapter 8) in the event
that this type of external or outsourced analysis is required. Local or regional HAB experts
should also be identified and contacted for guidance during potentially dangerous or
damaging bloom events.
5 REFERENCES
Anderson, D. M., Andersen, P., Bricelj, V. M., Cullen, J. J., and Rensel, J. E. 2001.
Monitoring and Management Strategies for Harmful Algal Blooms in Coastal Waters.
Asia Pacific Economic Program, Singapore, and Intergovernmental Oceanographic
Commission, Paris. 268 pp.
Engvall, E. and Perlmann, P. 1971. Enzyme-linked immunosorbent assay (ELISA)
quantitative assay of immunoglobulin G. Immunochemistry 8(9), 871-874.
499
Fehling, J., Green, D. H., Davidson, K., Bolch, C. J., and Bates, S. S. 2004. Domoic acid
production by Pseudo-nitzschia seriata (Bacillariophyceae) in Scottish waters. Journal
of Phycology 40, 622-630.
Laycock, M. V., Donovan, M. A., and Easy, D. J. 2010. Sensitivity of lateral flow tests to
mixtures of saxitoxins and applications to shellfish and phytoplankton monitoring.
Toxicon 55, 597-605.
Usleber, E., Dietrich, R., Burk, C., Schneider, E., and Martlbauer, E. 2001. Immunoassay
methods for paralytic shellfish poisoning toxins. Journal of AOAC International 84,
1649-1656.
500
Appendix 3 - Methods for measuring TEP and their precursors in seawater
Appendix 3
METHODS FOR MEASURING TRANSPARENT EXOPOLYMER

PARTICLES AND THEIR PRECURSORS IN SEAWATER
Loreen O. Villacorte1,2, Jan C. Schippers1 and Maria D. Kennedy1
1
2
1! Introduction ................................................................................................................................................... 501!

2! TEP0.4µm method ............................................................................................................................................ 501!
3! TEP10KDA method .......................................................................................................................................... 504!
4! Storing water samples ................................................................................................................................... 507!
5! References ..................................................................................................................................................... 507!
1! INTRODUCTION
Transparent exopolymer particles (TEP) and their precursors produced by phyto-/bacterio-

plankton in fresh and marine aquatic environments are increasingly considered as a major
cause of organic/particulate fouling in MF/UF membranes and organic/particulate and
biological fouling in SWRO membranes. The following sections comprise detailed
descriptions of two methods for measuring transparent exopolymer particles in seawater,
namely TEP0.4µm and TEP10kDa. The TEP0.4µm method measures transparent exopolymer
particles retained by membrane filters having pores of 0.4 µm and conventionally known as
TEP (Passow and Alldredge, 1995). The TEP10kDa method covers transparent exopolymer
particles retained by membrane filters with molecular weight cut-off of 10 kDa.
Consequently, this method covers both TEP and most (if not all) of their colloidal precursors.
TEP0.4µm is a more rapid method than TEP10kDa and is recommended for routine TEP
monitoring in untreated seawater. The TEP10kDa method is more time consuming, however, it
gives much more information because it covers both TEP and their colloidal precursors.
2! TEP0.4µM METHOD
The most widely used method for TEP, referred to here as TEP0.4µm, was developed in the
mid-90s (Passow and Alldredge 1995). Recently, it was found that this method may
overestimate TEP in seawater due to the reaction of Alcian blue (AB) with residual saline
moisture on the filter (Villacorte 2014). To minimize this effect, Villacorte et al. (2015)
proposed that the membrane with the retained TEPs be pre-rinsed with ultrapure water
(UPW) before staining with AB. The modified procedure for measuring TEP0.4µm is
illustrated in Figure 1 and is briefly described as follows:
Preparation of Alcian blue dye solution
1.! The stock dye solution is prepared by dissolving 250 mg/L of Alcian blue in ultrapure
water spiked with acetic acid to lower pH to 2.5. The prepared stock solution is stirred
until most of the Alcian blue powder is dissolved (up to 18 hours). The working solution
is prepared on the same day of analysis by filtering 20 mL of the stock solution through
0.05 µm polycarbonate (PC) membranes before staining. The remaining stock solution is
stored in the dark at 4oC and a new stock solution should be prepared after 4 weeks.
501
Figure 1. Procedural diagram for measuring TEP0.4µm (adapted from Villacorte et al. 2015)
Appendix 3 – Methods for measuring TEP and their precursors in seawater
Sample filtration and staining

2.! First, the water sample (typically >20 mL) is filtered through a 47 mm diameter
polycarbonate filter (0.4 µm pore size) by applying a vacuum of 0.2 bar. Then filter
(<0.2 bar vacuum) 2 mL of UPW through the filter-retained TEP to wash the
remaining sample moisture through the filter.
3.! Subsequently, 1 mL of the working AB dye solution is applied over the filter, allowed
to react with TEP for 10 seconds, and then the unreacted dye is flushed through by
vacuum filtration (<0.2 bar). To remove the remaining unreacted dye, a rinsing step is
performed by filtering 2 mL of UPW.
4.! The rinsed filter is transferred to a 50 mL glass beaker and soaked in 6 mL of 80%
sulfuric acid solution. The beaker is covered with parafilm and gently mixed on a
shaker for 2 hours.
5.! The acid solution is then transferred to a 1cm cuvette and the absorbance (At) is
measured at 787 nm wavelength - the wavelength of maximum absorbance of Alcian
blue when dissolved in sulfuric acid.
Blank measurement
6.! A filter blank (Af) is measured in the same way (steps 2-5), but filtering TEP-free
blank samples (e.g., synthetic water with similar ion concentration as the water
sample) instead of actual water samples. For sample correction (As), water samples
are filtered in the same way as for determination of the total absorbance, but skipping
the AB staining procedure.
Concentration calculation (without calibration)
7.! The concentration of TEP0.4µm in terms of abs/cm/L is calculated as follows in
equation 1:
A+ − A- − A.
TEP$.&'( =
V-
(1)
where (At) is the total absorbance of the dye that reacted with TEP and the filter
(abs/cm); (Af) is the absorbance of the dye adsorbed to the filter (abs/cm); (As) is the
absorbance of unstained sample (abs/cm) and Vf is the volume of sample filtered (L).
Concentration calculation (with calibration)
8.! TEP0.4µm can be further calibrated and expressed in terms of equivalent weight of
standard acid polysaccharide - Xanthan gum – as mg Xeq/L:
A+ − A- − A.
TEP$.&'( =
m121 V-
(2)
where m787 is the slope of the calibration line [(abs/cm)/mg Xeq] which is determined by
plotting the mass of the standard (Xanthan gum) against the corresponding absorbance of AB
that reacted to it (see Passow and Alldredge 1995).
Note: The standard calibration procedure in step 8 involves dry weight measurements to
determine the mass of Xanthan gum retained on polycarbonate filters. This is prone to
several inaccuracies during drying (dust contamination) and weighing (electrostatic force
503
Appendix 3 – Methods for measuring TEP and their precursors in seawater
interference) at very low quantities (5-50 µg) of Xanthan. It is also very challenging to
prepare a homogeneous and artifact-free solution of Xanthan for the calibration experiment.
An option to avoid such difficulties is to skip the calibration step entirely (follow until step 6
only), whereby concentrations of TEP are expressed in terms of abs/cm/L. It is only feasible,
however, when comparing results using the same batch of AB staining solutions. Moreover,
absorbance results are not directly comparable with TEP10kDa due to differences in
wavelength at which the absorbance were measured. For standardized results, an alternative
calibration procedure was introduced by Villacorte (2014) as described in steps 9-13.
Alternative calibration method
9.! Homogenized standard solutions (4 mL) containing different concentrations (0, 1, 2, 3,
4 and 5 mg/L) of Xanthan gum are prepared from a stock solution (40 mg/L). The pH
is adjusted to pH 2.5 by adding 0.05 mL acetic acid to each solution and then briefly
agitated.
10.!The solution is then stained by adding 1 mL of pre-filtered AB staining solution,
mixed for 10 s and incubated for 10 min.
11.!Filter 4 mL of the resulting solution through a 0.1 µm PC membrane by vacuum
filtration (0.2 bar).
12.!The PC membrane used to filter AB-stained standard solution is carefully transferred
to a 50 mL beaker. Six mL of 80% sulfuric acid solution is added, covered with
Parafilm and mixed on an auto-shaker for 2 hours. The acid solution is transferred to a
1cm cuvette and absorbance was measured at 787 nm.
13.!To determine the calibration slope (m787), the mass of Xanthan gum retained on the
PC membrane is calculated by multiplying the volume filtered (4 mL) by the
concentration of Xanthan in the stained standard solution. The calculated mass is then
plotted against the corresponding AB absorbance measured at 787 nm wavelength,
whereby the average linear slope is the m787. TEP0.4µm is then calculated using
equation 2.
3! TEP10KDA METHOD
The TEP10kDa is a method intended to measure TEP and their precursors (down to 10 kDa).
This method was developed by Villacorte et al. (2015), which is partially based on the
principles developed by Thornton et al. (2007) for measuring acidic polysaccharides. Figure
2 illustrates the procedure for measuring TEP10kDa and is described as follows:
Sample filtration and TEP extraction
1.! The water sample is filtered constant flux (60 L/m2/h) through a 10 kDa MWCO
regenerated cellulose membrane (25 mm diameter) using a syringe pump.
2.! After filtering 10-100 mL of sample, the syringe is replaced with a clean syringe
containing about 10 mL of air. Air is then injected to the filter holder (60 L/m2/h)
until all the remaining water in the feed side of the membrane holder has passed
through the membrane. The total filtered sample volume is then measured after
collecting the filtrate
3.! To rinse out residual saline moisture, 5 mL of UPW is injected to the filter holder at
60 L/m2/h. Air is again injected until all the rinse water on the feed side of the
membrane holder has passed through the membrane.
4.! The membrane is then carefully removed from the filter holder and placed feed side
down, in a clean disposable plastic container filled with 10 mL of UPW. The sample
is covered, vortexed for 10 s and sonicated for 60 min.
504
Figure 2. Procedural diagram for measuring TEP10kDa (adapted from Villacorte et al. 2015).
Staining and absorbance measurement

5.! Transfer 4 mL of the re-suspended TEP solution to a clean 20 mL disposable plastic
container. To adjust the sample pH to 2.5, 0.05 mL of acetic acid solution is added to
the solution.
6.! Add 1 mL of the working AB dye solution (for dye preparation see step 1 of TEP0.4µm
method) to the sample, mix vigorously and leave to react for 10 min.
7.! Filter 4 mL sample of the TEP-AB solution through a 0.1 µm PC filter by vacuum
filtration (<0.2 bar).
8.! The filtrate is collected in a plastic container (10 mL), transferred to a 1-cm cuvette
and absorbance (Ae) at 610 nm wavelength - the wavelength of maximum absorbance
(visible light range) of AB when dissolved in acetic acid solution - is measured with a
spectrophotometer.
Blank measurement
9.! The blank absorbance (Ab) is measured to correct for the amount of stain adsorbed by
the polycarbonate filter (0.1µm). This is performed following steps 5-8 but replacing
the sample with UPW.
Concentration calculation (without calibration)
10.!The TEP10kDa concentration in absorbance per cm per liter of filter water (abs/cm/L) is
calculated as follows:
A+ − A-
TEP$%&'( =
V/
(3)
where Ab is the absorbance of filtered blank (abs/cm), Ae is the absorbance of the
excess or un-reacted dye (abs/cm) and Vf is the volume of filtered sample (mL).
Concentration calculation (with calibration)
11.!Alternatively, the TEP10kDa concentration can be calibrated and expressed in terms of
mg Xanthan equivalent per litre (mg Xeq/L):
1 V3
TEP$%&'( = A − A+
12$% V/ -
(4)
where m610 is the slope of the calibration curve [(abs/cm)/(mg Xeq/L)] and Vr is the total
volume of the re-suspended TEP sample solution (i.e., 10 mL).
12.!For the calibration, standard solutions (4 mL) containing different concentrations (0, 1,
2, 3, 4 and 5 mg/L) of Xanthan gum are prepared from a homogenized stock solution.
To adjust sample pH to 2.5, 0.05 mL acetic acid is added to each solution and then
briefly agitated. The solution is then stained by adding 1 mL of pre-filtered AB
staining solution, mixed for 10 seconds and incubated for 10 min.
13.!Filter 4 mL of the resulting solution through a 0.1 µm PC membrane by vacuum
filtration (0.2 bar). The filtrate is collected, transferred to 1cm cuvette and absorbance
is measured at 610 nm.
14.!The Xanthan concentration of the stained dye is plotted against the measured
absorbance (excess dye absorbance) and the average linear slope is the m610. Since
506
concentration is inversely proportional to the excess dye absorbance, the calibration

slope (m610) is always a negative value.
Note: The TEP10kDa concentration calculated using Eqn. 4 can be directly compared with
TEP0.4µm calculated based on Eqn. 2 as both are calibrated with Xanthan gum standard.
However, results for TEP10kDa calculated based on Eqn. 3 and TEP0.4µm calculated based on
Eqn. 1 are not comparable as the absorbance values were measured at different
wavelengths - 610nm and 787nm, respectively.
4! STORING WATER SAMPLES
Storing samples for a period of time before analysis may lead to significant disparity between
the measured and in-situ TEP concentration as a consequence of coagulation, bacterial
release, bacterial degradation, adsorption on walls of sample bottles or a combination thereof.
Bottle tests at 4oC temperature revealed that TEP0.4um concentration may increase over a
period of time either due to coagulation of TEP precursors or through bacterial TEP release
(Villacorte 2014). On the other hand, TEP10kDa concentration can rapidly decrease by up to
45% within the 3 days of storage resulting from either bacterial degradation or adsorption to
walls of sample bottles (Villacorte 2014). Further investigations are still necessary to fully
understand the mechanisms involved in the TEP loss or increase as well as to develop
reliable measures to preserve TEP samples (e.g., sample bottle, freezing, preservative
addition). It is therefore important that samples be analyzed immediately (within 24 hours)
after sampling to obtain reliable TEP concentrations. For TEP10kDa, it is advisable to filter and
then flush the sample immediately after sampling so the membrane can be kept at 4oC until
analysis.
5! REFERENCES
Passow, U. and Alldredge, A. L. 1995. A dye-binding assay for the spectrophotometric

measurement of transparent exopolymer particles (TEP). Limnology and Oceanography
40(7), 1326-1335.
Thornton, D. C. O., Fejes, E. M., Dimarco, S. F. and Clancy, K. M. 2007. Measurement of
acid polysaccharides (APS) in marine and freshwater samples using alcian blue.
Limnology and Oceanography: Methods 5, 73–87.
Villacorte, L. O. 2014. Algal blooms and membrane-based desalination technology. ISBN
978-1-138-02626-1, CRC Press/Balkema, Leiden.
Villacorte, L. O., Ekowati, Y., Calix-Ponce, H. N., Amy, G. L., Schippers, J. C., AND
Kennedy, M. D. 2015. Improved method for measuring transparent exopolymer
particles (TEP) and their precursors in fresh and saline water. Water Research 70 (1),
300–312.
507
Appendix 4 – Preservatives and methods for algal cell enumeration
Appendix 4
PRESERVATIVES AND METHODS FOR ALGAL CELL

ENUMERATION
Donald M. Anderson1 and Bengt Karlson2
1
Woods Hole Oceanographic Institution, Woods Hole MA USA
2
1 Introduction ...................................................................................................................................................509
2 Miscellaneous supplies, recipes, and protocols ............................................................................................509
2.1 Preservatives..........................................................................................................................................509
2.1.1 Advantages of Lugol’s solution.....................................................................................................510
2.1.2 Advantages of formaldehyde .........................................................................................................510
2.2 Sample storage ......................................................................................................................................510
2.3 Settling chambers ..................................................................................................................................510
2.4 Inverted microscope equipped with phase contrast and fluorescence...................................................511
2.5 Software that aids counting and data processing ..................................................................................511
2.6 Manual cell counter ...............................................................................................................................511
2.7 Qualification requirements for staff ......................................................................................................512
2.8 Assistance with species identification ...................................................................................................512
3 Enumeration of phytoplankton using inverted microscopy (the utermöhl technique) ..................................512
3.1 Equipment and reagents ........................................................................................................................512
3.2 Procedure ...............................................................................................................................................512
3.3 Calculations ...........................................................................................................................................515
4 Phytoplankton counting using a sedgewick rafter slide ................................................................................515
4.1 Procedure ...............................................................................................................................................515
5 References .....................................................................................................................................................516
1 INTRODUCTION
There are multiple ways to preserve phytoplankton samples and determine the algal species
composition and abundance. This Appendix provides details on some of the most common
methods. Additional relevant publications are included in Section 5, References. One of the
most useful is Karlson et al. (2010), which can be downloaded at http://hab.ioc-
unesco.org/index.php?option=com_oe&task=viewDocumentRecord&docID=5440.
2 MISCELLANEOUS SUPPLIES, RECIPES, AND PROTOCOLS
2.1 Preservatives
Lugol’s solution has proved to be a suitable preservative for many phytoplankton samples.
The solution preserves the shape of the algal cells; for flagellated species, the flagella remain
attached to the cells and the iodine staining of the algal cells gives a good contrast between
the organisms and the background when using phase-contrast microscopy. For seawater
samples, neutral or alkaline Lugol’s solution is most appropriate because calcareous
flagellates (coccolithophorids) can comprise a considerable proportion of the algal biomass.
If acidic Lugol’s solution is used, a parallel sample fixed in alkaline Lugol’s or with
hexamine buffered formalin should be taken because the acid dissolves the calcareous
casings (coccoliths) which are the basis for species determination of coccolithophorids.
Diatoms, which have silicate cell walls, preserve well in acidic Lugols. In samples fixed with
alkaline Lugol’s solution, fungal infections may occur, and for long-term storage, such
samples should be post-fixed by adding a few drops of concentrated formalin solution. If the
samples are to be stained with Calcofluor to enhance the pattern of cellulose plates on
509
armored dinoflagellates, it is an advantage to use neutral or basic Lugol’s solution or a

buffered formaldehyde solution, because acid prevents staining of the cellulose.
If cells are rugged and do not distort, a 2-5% formaldehyde solution is generally used,
working from a stock formalin sample (37% formaldehyde ) that is often acidified with acetic
acid. Recipes for these fixatives follow:
Acidic Lugol’s solution: Potassium iodide (KI) 100 g, Distilled water 1000 mL, Iodine
(I2) 50 g, Acetic acid glacial (100 % CH3COOH) 100 g, Neutral Lugol’s solution,
Potassium iodide (KI) 100 g, Distilled water 1000 mL, Iodine (I2) 50 g.
Alkaline (basic) Lugol’s solution: Potassium iodide (KI) 100 g, Distilled water 1000
mL, Iodine (I2) 50 g, Sodium acetate (CH3COONa) 100 g.
Neutral, weakly basic formalin solution: Analytical quality formalin is diluted to 20 %
(1 part concentrated solution to 1 part distilled water). 100 g hexamethylenetetramine
(hexamine) C6H12N4 is added to 1 L of the 20 % formaldehyde solution.
Acidic formalin: Equal amounts of analytical quality formalin (formaldehyde 37 %) and
concentrated acetic acid are mixed.
2.1.1 Advantages of Lugol’s solution

Lugol’s enters the cell more quickly than formaldehyde, leaving shock-sensitive organisms
better preserved in the sample. The specific weight of organisms is increased by iodine,
resulting in faster settling times. Detection of phytoplankton cells is made easier by the
enhanced contrast between organisms and the surrounding fluid. The iodine stains starch,
aiding recognition of those groups of algae (e.g. Chlorophyta) which use this as a storage
compound.
2.1.2 Advantages of formaldehyde

Formaldehyde is a good fixing and preserving agent for cells that have a relatively rigid cell
wall, as the cell wall structures and other characteristics such as eye spots remain visible.
When stored properly in appropriate bottles, samples will stay in good condition for many
years without attention. Autofluorescence of chlorophyll-a, though decaying, remains intact
for at least several days if the samples are stored in constant dark.
2.2 Sample storage
Preserved phytoplankton samples should be stored in the dark at room temperature to prevent
photo-oxidation. The maximum storage time for iodine (Lugols)-fixed samples that have
been stored in a dark and cool place (1 to 5 °C) is 12 months. During this period, the samples
should be checked every second month and topped up with Lugol’s solution if necessary (the
sample should be golden brown in color, like tea). Reference material should be collected and
stored in suitable separate sample bottles. For all phytoplankton samples to be stored for
more than 3 to 4 months, dark glass bottles are the preferred container.
2.3 Settling chambers
If an inverted microscope is to be used (which is highly recommended), consider purchasing
the Hydrobios Combined Plate Chamber (Utermöhl-Chamber) set or the equivalent, with one
cylinder each of 10, 50, and 100 mLs (Figure 1). Three cover plates are also needed. There
510
are two versions of the chambers. The sliding version is desirable since the main part of the
chamber is removed during analysis. This makes it possible to have better illumination and to
work at higher magnifications. The non-sliding version can be rotated, which is useful when
analyzing several transects (diameters).

Figure 1. Settling chambers for the Utermöhl technique. Left: chambers and chimneys for different volumes.
Right: a stand for the settling chambers that will collect the water sample when the chimney is moved
sideways, away from the observation chamber. Square cover slip for the observation chamber is at left. Photos:
Hydrobios. http://www.hydrobios.de/product/combined-plate-chamber/
2.4 Inverted microscope equipped with phase contrast and fluorescence

Inverted microscopes are very useful for phytoplankton counting, since samples can be
settled and counted directly in the settling chamber. If an inverted microscope is not available,
then it is also possible to use a standard microscope,
but with a Sedgewick Rafter slide (1 mL capacity;
Figure 2) for counting. In this instance, the water
sample needs to be settled in centrifuge tubes or
graduated cylinders, the overlying water removed
using aspiration, and then a subsample taken from
the residue (after mixing) for counting, taking
careful consideration of the initial and final
volumes. More details on filling the Sedgewick
Rafter slide are given below. It is useful to have a
Figure 2. Sedgewick Rafter counting slide. microscope fitted with a digital camera to take
Photo: D. Kulis. pictures. If the camera can take pictures under
fluorescence illumination, this is preferred.
2.5 Software that aids counting and data processing
There are free (e.g. Plankton Toolbox; Karlson et. al. 2015) and commercial software that can
aid in the process of counting phytoplankton samples. The Plankton Toolbox provides a
counting module that may be used on a computer placed next to the microscope during
analysis of samples. The software facilitates the counting of many species concurrently, and
also helps in calculations of the cell volumes of the organisms. Cell
volumes are used to calculate the biomass of the plankton.
2.6 Manual cell counter
Manual counters are very useful, but surprisingly expensive. A
Figure 3. Single-position multi position counter is the most useful – one on which the worker
counter.
511
can keep count of 5 or more species at the same time, but smaller, single counters are also
useful (Figure 3).
2.7 Qualification requirements for staff
Technical staff should have competence in identification of marine phytoplankton, preferably
through formal training courses in marine botany (algal flora) at the university level or
through practical experience in phytoplankton identification. Short courses are offered, such
as those given by the IOC HAB Programme:
http://hab.ioc-unesco.org/index.php?option=com_content&view=article&id=32&Itemid=0
A database of pictures of the species present in the samples is useful and can clarify doubts in
identification for analysts with less experience. A limited HAB species image database with
some species description details is found in Appendix 1.
2.8 Assistance with species identification
The best advice to plant operators seeking to mitigate the effects of a specific algal bloom is
to collect samples and identify the causative organism, hopefully to the species level, but at
least to genus. With some training and modest microscope facilities, this can be done on site
(Chapter 3). There are also outside experts and services that will do this type of work on
demand, but this will cost time. The Intergovernmental Oceanographic Commission (IOC)
Science and Communication Centre on Harmful Algae, University of Copenhagen, Denmark
http://hab.ioc-unesco.org/ can offer assistance in identification of microalgae. Small samples
with a limited number of species will be examined free of charge while analysis of a larger
number of samples may incur a charge. There are also several other institutes, universities
and companies that have the capability to analyze phytoplankton samples.
3 ENUMERATION OF PHYTOPLANKTON USING INVERTED MICROSCOPY (THE UTERMÖHL
TECHNIQUE)
The most common method for counting phytoplankton uses an inverted microscope (i.e., one
in which the sample sits above the microscope optics) and the Utermöhl method for settling
and concentrating the sample. The following sections describe this procedure in detail. For
facilities that only have a standard microscope, another cell counting method using the
Sedgewick Rafter slide is described in section 4 below.
3.1 Equipment and reagents
Sample containers must be clean, not permeable, watertight, wide mouth, and typically are
100-250 mL in volume. Sample containers must not be completely filled, as some air at 20%
of the total capacity will facilitate the homogenization of the sample before settling.
Settling chambers consist of a cylinder and a thin base (no more than 0.17 mm) allowing
observation from below using the inverted microscope. The chambers must be cleaned
between samples using a thin paintbrush, detergent and distilled water. The settling chambers
must be calibrated.
The inverted microscope should be equipped with phase contrast and/or Nomarski
(Differential Interference Contrast = DIC), and if possible a digital camera. The oculars must
have a micrometer and a reticule. The microscope must be calibrated for each objective.
Fluorescence is a useful capability on the microscope.
3.2 Procedure
1. For monitoring of phytoplankton it is important to analyze the samples rapidly, so it is
better to maintain the samples close to room temperature during transport by keeping the
512
samples inside a cooler with cooling blocks. On arrival to the laboratory, the samples
should be placed on the bench. A portion of the sample is saved for live examination and
another portion preserved (if this was not done in the field) and either stored or settled.
To demonstrate that the temperature in the room and the temperature in the sample are
close, a thermometer for the air temperature in the laboratory and other thermometer to
measure temperature in the water sample are used. In a notebook, records are kept of the
code of the sample, the complete reference of the sample, the date and hour of settling,
the number and volume of the settling chamber, temperature in the laboratory and in the
sample, and the analyst who did the settling.
2. During sample storage, suspended particles settle out and (small) algae become
indiscernible by incorporation in detritus aggregates or by adhesion to other large algal
cells. Resuspension and separation of particles before settling them in the Utermöhl
chamber can be achieved by shaking the sample as gently as possible. The method used
for manual shaking should be described clearly in order to minimize differences between
operators. A combination of alternating horizontally rolling and vertical turning upside
down of the sample bottle for a specific number of times provides better mixing than
straightforward shaking. The homogenization of the samples must be gentle, combining
circular horizontal movements with vertical movements.
3. One subsample is maintained alive for in vivo examination and another is preserved in
Lugols. Samples must be preserved as soon as possible after collection. Live samples
must be observed within 36 h. Preserved samples must be stored in the dark; room
temperature is appropriate if the analysis is done within 3 weeks. Otherwise, refrigerate
at 4oC. Lugol’s- fixed samples can be stored for 1 year in the dark and at temperatures
between 1-5ºC; for longer periods of storage it is necessary to add formaldehyde.
4. The volume of the settling chamber must be adapted to the type of waters to be
analyzed: 100 mL for oligotrophic waters, 50 mL for mesotrophic, 5-25 mL for
eutrophic. For nanoplankton and picoplankton the settling chambers of 100 mL are not
adequate and require preconcentration. The settling chambers must be placed on a
horizontal surface, all the chamber must be filled with the sample without air spaces, and
the chamber must be closed using a glass cover. The bench must be free of vibrations.
The recommended settling time depends on the volume (height) of the chamber, 48 h for
100 mL, 24 h for 50 mL, 12 h for 25 mL, 8 h for 10 mL, 3 h for 2 mL. After settling, the
cylinder is displaced and a glass cover is placed on top of the chamber.
5. Preserved samples must be acclimated to the ambient temperature, homogenized and
transferred directly from the container to the settling chamber. Temperature differences
between sedimentation chamber and medium may produce convection currents that have
differential effects on the settling of phytoplankton species, depending on their physical
properties. Furthermore, bubbles may develop in relatively cold samples as the solubility
of gases declines with the gradual rise of the temperature of the sample. When the
temperature of the chamber is higher than that of the sub-sample, the larger and heavier
particles tend to concentrate towards the chamber wall and the smallest particles more
towards the center of the chamber. If the subsample has a higher temperature, the
smallest particles collect near the chamber wall.
6. There are three counting strategies; the strategy to choose depends on the abundance and
size of each species.
513
Quantification by fields: 5 fields randomly distributed in the chamber; more fields can be
counted to increase the total number of cells counted.
Quantification using transects, 1 or 2 across the middle of the chamber one on the side of
the other.
Quantification using the full chamber: counting all cells present in the total volume settled.
514
For the 3 strategies, it is necessary to define a normal procedure for the quantification of cells
at the boundary of the field; for example, do not include cells lying across the top line and the
left line but include the cells crossing the lower and the right lines of the reticule.
As a general rule, if a precision of 5% in the estimation of the mean number of cells per
sample is required, then 400 objects should be counted. It makes no difference whether 10
fields are counted with 40 objects per field or 80 fields with just 5 objects. It is, however,
sometimes difficult to count 400 cells, so in practice, the number of cells counted should be
at least 50-100 cells in total for each dominant species. The area to be counted should be
adjusted to reach this number of cells. For some species this will not be possible because they
are in low density. In that case, the full chamber should be examined.
3.3 Calculations
The following equation related the number of cells per unt volume to the area examined:
Ad
N= X
av
N is the number of cells per unit volume (e.g., cells/mL)
X is the average number of cells per field, transect or full chamber
A is the area of the chamber
v is the volume of the chamber. If the result is in mL then we use mL, if the result is in L we
will use L
a is the area of the total field counted
d is the dilution factor (when the sample is diluted)
Example of calculations for the case of counting 10 cells in a full chamber (area 529.13 mm2)
after settling 50 mL.
N = 10 * 529.13 * 1 = 0.2 cells/mL = 200 cells/L
529.13 * 50 mL
4 PHYTOPLANKTON COUNTING USING A SEDGEWICK RAFTER SLIDE
For laboratories that do not have an inverted microscope, an alternative to the Utermohl’s
technique is to use a Sedgewick Rafter slide (SRS) and a regular compound microscope (see
sction 2.4). The Sedgewick Rafter slide holds exactly one mL. The thickness of the slide
generally restricts magnification to 200X or lower.
4.1 Procedure
The SRS without grid with cover glasses is available from Wildlife Supply Company, 95
Botsford Place, Buffalo NY 14216 (www.wildco.com ). Two styles are available – one
without, and one with grids. The latter is useful if only a portion of the slide is to be
counted.
1. The preserved sample can be counted directly if the cell concentrations are high enough,
or the sample can be concentrated by settling a known volume in a graduated cylinder
for several days, with the settled material careful removed from the bottom with a
pipette, and its volume measured.
2. Swirl the sample manually or with a mechanical agitator or vortexer to randomize the
organisms.
515
3. Withdraw a 1.0 mL sample and dispense it into the SRS. This is best accomplished by
rotating the cover slip so that it sits diagonally across the SRS such that small openings
are in each diagonal corner. The sample is pipetted into one open corner, and air exits
from the other opening. When the slide is full, straighten the cover slip so there are no
more openings. Placement of the cover glass limits the volume to 1 mL.
4. The gridded SRS has 20 rows, 50 columns, and 1000 squares. Using a microscope
magnification that allows the cells of interest to be easily identified, move the
microscope stage laterally, counting cells within a gridded row. Repeat with a return
transect across the slide, either directly below the first transect, or randomly – the latter
if only a portion of the slide is to be scanned. Have a consistent policy for cells that fall
on a line. For example, do not include cells lying across the top line, but do count those
lying across the bottom line. Try to count 100 - 400 algal cells, and calculate cells/mL
from that. For example, if you counted 150 cells in 8 rows, you have 150 * 20/8 = 375
cells/mL. This should be multiplied by the concentration factor determined from the
settling step in # 1 above, if that was done.
5. Note that if you have a non-gridded SRS, you can still count less than the entire slide.
Just move the microscope stage horizontally, counting all cells that are visible in the
field of view, and then when the edge of the chamber is reached, slide the field of view
vertically to a new, random position, and then scan horizontally to the opposite edge.
Repeat with another random scan. Keep track of the number of horizontal scans until
you have counted 100 – 400 cells of interest. You can then count how many horizontal
scans there are along the vertical edge of the SRS, and use the same approach described
above for the gridded SRS to calculate a cell concentration.
5 REFERENCES
Edler, L. and Elbrächter, M. 2010. The Utermöhl method for quantitative phytoplankton
analysis. In: Microscopic and molecular methods for quantitative phytoplankton
analysis, Eds. Karlson, B., Cusack, C. and Bresnan, E. Intergovernmental
Oceanographic Commission Manual and guides, pp 13-20.
EN 15204. 2006. Water quality - Guidance standard on the enumeration of phytoplankton
using inverted microscopy (Utermöhl technique).
Hallegraeff, G. M., Anderson D. M., and Cembella, A. D. 2003. Manual on Harmful Marine
Microalgae UNESCO Publishing.Hasle, G. R. 1978. The inverted-microscope method, p.
88–96. In A. Sournia [ed.], Phytoplankton Manual. UNESCO.
Karlson, B., Andreasson, A., Johansen, M., Mohlin, M., Skjevik, A-T., and Strömberg, P.
2015. Plankton Toolbox – open source software making it easier to work with plankton
data, Proceedings of the 16th International Conference on Harmful Algal Blooms, 27 Oct.
to 1 November, 2014, Wellington, New Zealand.
Software available at http://nordicmicroalgae.org/tools
Karlson, B., Cusack, C., and Bresnan, E. (Eds.) 2010. Microscopic and molecular methods
for quantitative phytoplankton analysis. Paris, UNESCO. (IOC Manuals and Guides, no.
55.). 110 pages.
http://hab.ioc-
unesco.org/index.php?option=com_oe&task=viewDocumentRecord&docID=5440
516
LeGresley, M., and McDermott, G. 2010. Counting chamber methods for quantitative
phytoplankton analysis - haemocytometer, Palmer-Maloney cell and Sedgewick-Rafter
cell. in: Microscopic and molecular methods for quantitative phytoplankton analysis,
Eds. Karlson, B., Cusack, C. and Bresnan, E. Intergovernmental Oceanographic
Commission Manual and guides, pp 25-30.
Moestrup, Ø., Akselman, R., Cronberg, G., Elbraechter, M., Fraga, S., Halim, Y., Hansen, G.,
Hoppenrath, M., Larsen, J., Lundholm, N., Nguyen, L. N., and Zingone, A. (Eds) 2016.
IOC-UNESCO Taxonomic Reference List of Harmful Micro Algae. Available online at
http://www.marinespecies.org/hab/
Utermöhl, H. 1958. Zur vervollkommnung der quantitativen phytoplankton-methodik. Mitt.
int. Ver. theor. angew. Limnology 9, 1–38.
517
Appendix 5 - Autopsy and cleaning of RO elements affected by HABs
Appendix 5
AUTOPSY AND CLEANING OF REVERSE OSMOSIS ELEMENTS

AFFECTED BY HARMFUL ALGAL BLOOM-CONTAMINATED
SEAWATER
Nuria Peña1, Steve Chesters2, Mike Dixon3, and Siobhan F. E. Boerlage4
1
Genesys Membrane Products, S.L. Spain
2
Genesys International, Cheshire, United Kingdom
3
4
1 Fouling diagnosis ..........................................................................................................................................519

2 Pre-autopsy ....................................................................................................................................................519
3 Autopsy .........................................................................................................................................................520
4 Cleaning protocol for HAB-affected RO elements .......................................................................................522
5 Additional notes on cleaning .........................................................................................................................525
6 References .....................................................................................................................................................525
1 FOULING DIAGNOSIS
Following an algal bloom, if a change is observed in the reverse osmosis (RO) performance,
an initial visual plant inspection should be carried out, including looking at and removing
cartridge filters and membrane elements from different positions in the pressure vessel.
Fouling may be a combination of organic, biofouling, particulate, and metal hydroxide.
Biofouling is often slimy or gelatinous which may be accompanied by a bad smell while iron
fouling is a reddish brown deposit. Figure 1 shows evidence of biological fouling on a
cartridge filter and inside a pressure vessel following an algal bloom event.
Figure 1. Severe fouling due to algal bloom. Algal fouling on a cartridge filter (L), remnants of algal
fouling left inside a pressure vessel after removing the lead element (R). Note: Elements are from a
brackish water reverse osmosis system, fouling may appear as different colors in a SWRO system.
If it is obvious that there is severe fouling, membrane autopsy would be the most appropriate
tool for identifying the nature of the foulant, and the best cleaning protocol for removal of the
foulant. Autopsy results would be interpreted in conjunction with an analysis of plant
performance data.
2 PRE-AUTOPSY
The following procedure would be recommended prior to membrane autopsy:
519
• The first element from the SWRO first pass should be chosen for autopsy, although
best characterization of plant status would be obtained by autopsying elements from
both the first and last positions of the pressure vessel. Note that for two-pass seawater
desalination plants, it is unlikely that harmful algal bloom (HAB) fouling will be
evident in the second pass, as low fouling permeate from the SWRO is the feedwater.
• Membrane elements should be wrapped in dark plastic and taped securely to ensure
that water is retained within the element. This is necessary for accurate
microbiological analysis and deposit characterization. Dry membrane elements yield
unrepresentative samples that can give false results.
• Do not use any chemicals for membrane preservation before sample delivery.
Otherwise, microbiological count results will not be representative.
• Membrane elements should be sent for autopsy the day of removal. If membrane(s)
cannot be shipped immediately, store horizontally in a cool dark place. It is important
to avoid light and high temperatures.
In addition to membrane elements, further samples could be analyzed for complementary
information about the algal bloom:
• Water samples from different points in the pretreatment process;
• Silt density index (SDI) membrane discs from different points in the pretreatment
process; and
• Cartridge filters
3 AUTOPSY
Autopsy involves different tests and analyses that should be carried out in order to obtain a
final diagnosis. This process includes the following steps that should be followed during algal
bloom events:
External and internal inspection of membrane element
During the inspection, failure due to a significant presence of fouling would be verified.
Fouling after an algal bloom would be expected to:
• Appear organic, with development of a biofilm on the RO membrane surface;
• Be a sticky deposit with a high content of water that would probably also adhere to
the spacer material; and
• Exhibit different coloration depending upon the type of algal bloom and fouling
content.
Analyses of foulant
• The LOI (loss on ignition) test is a common and widely used method to estimate the
organic content of foulants. A very high organic matter content would be expected
(>60%), although the high sodium chloride content in seawater systems could
interfere with the measurement.
• Optical microscopy analysis for algae identification (see Chapter 3, Appendix 3).
Analyses of foulant samples and membrane surface samples
• Scanning electronic microscopy with Energy Dispersive X-ray detection (SEM-
EDX): with this analytical technique it is possible to get an elemental analysis of
520
fouling and inspection at very high magnifications. This method can assist in
identifying colloidal, scaling, and metal hydroxide fouling, but cannot identify
biofouling.
• Infrared spectroscopy (IR): This technique provides information related to the
presence of specific functional groups and it is a very good tool for identification of
organic components. For algal bloom fouling it would verify the presence of
biopolymeric substances such as polysaccharides and proteins produced during an
algal bloom.
• Liquid Chromatography - Organic Carbon - Organic Nitrogen Detection (LC-OCD-
OND): a very accurate tool for the characterization of natural organic matter
including biopolymers (see Chapter 5).
Analyses of foulant samples and membrane surface samples for Algal Organic Matter
At present, there is no standard test protocol for determining Algal Organic Matter (AOM)
deposits on RO membranes or spacers. HAB research studies have examined the
accumulation of organic deposits using LC-OCD and staining of the membrane/spacer using
Alcian blue in membrane autopsies (Villacorte 2014).
• TEP deposits can be identified by staining a sample of RO membranes and/or spacer
with a solution of Alcian blue (see Appendix 3 for preparation) followed by rinsing
with ultrapure water.
• The stained samples are then viewed by an optical microscope. The presence of a
blue/green stained deposit is indicative of the presence of TEP. It cannot distinguish
between TEP which may have originated from algae and/or bacteria in the feedwater
or generated by biofilm growth in the membrane.
• For more quantitative information on the deposit on RO membranes/spacers
following a HAB bloom, Villacorte (2014) submerged samples of membrane and
spacer in a known quantity of ultrapure water and sonicated for 1 hour to extract the
organic matter accumulated on the membrane.
• The extracted solutions can then be analyzed for TEP0.4µm and TEP10kDa (see
Appendix 3) and natural organic matter such as biopolymer concentrations using LC-
OCD as described above.
The identification of NOM constituents such as TEP could assist in determining the
appropriate cleaning agent/chemical.
Microbiological counts
• A very high presence of microorganisms would be expected in the biofilm on the
membrane surface after an algal bloom.
• Characteristic microorganisms would be: bacteria (aerobic and anaerobic), moulds,
and yeasts, etc. Results would be recorded as UFC/cm2.
Biocide tests
Based on microbiological counts, biocide tests would be carried out to determine the most
suitable biocide for biofilm removal.
Sampling for membrane performance parameters and cleaning tests
Once the nature of fouling has been identified, it is important to know the effect of the
foulant on membrane performance. By working with membrane coupons on a flat sheet test
521
rig it is possible to examine the permeate flux and salt rejection conditions of a membrane
after an algal bloom. If the membrane was not damaged during algal bloom operation and
there is fouling present, a decrease in both permeate flux at constant applied pressure and salt
rejection would be expected.
Different cleaning tests should be carried out in order to choose the most suitable cleaner and
optimal cleaning conditions (e.g., concentration, contact time, temperature) as the incorrect
cleaning agent may cause a further loss in membrane performance. Depending upon the
results obtained during the biocide tests, an additional biocide step should be tested and
included in the final cleaning protocol if appropriate.
Integrity tests
An important impact of fouling on reverse osmosis membranes is the damage that can be
caused on the polyamide layer and on their rejection capabilities (Peña et al. 2013).
Therefore, integrity tests should be carried out in order to check if the polyamide layer has
suffered irreversible damage during operation after fouling due to algal bloom. Considering
the nature of fouling after an algal bloom, it is very common to observe increases in
differential pressure that can produce physical damage on the membrane surface.
The methylene blue test would detect this kind of damage, staining damaged areas blue. Care
should be taken not to scratch or damage the membrane during the autopsy process.
4 CLEANING PROTOCOL FOR HAB-AFFECTED RO ELEMENTS
Before cleaning RO elements, operators are advised to review the technical manual, technical
bulletins, or specification sheets for each element type used. Common manufacturers’
literature is below:
• Hydranautics: Foulants and Cleaning Procedures, TSB-107,
http://www.membranes.com/docs/tsb/TSB107.pdf
• Dow Filmtec: Filmtec Reverse Osmosis Membranes Technical Manual, Section 6,
page 121, http://www.dow.com/en-us/water-and-process-
solutions/resources/resource-finder
• Toray: Operation, Maintenance and Handling Manual, TMM-300, 310, 320, 330, 340,
350, http://www.toraywater.com/knowledge/pdf/HandlingManual.pdf
While general cleaning solutions are presented below, specially formulated cleaners are
available from RO chemical suppliers that are more effective than standard commodity
chemicals. Globally there are many common RO cleaning chemical suppliers such as (but not
limited to); Genesys, Avista, PWT Chemicals, Nalco, Alkema Solutions, GE, and BASF.
It is very important always to follow membrane manufacturers’ guidelines. While most
SWRO elements will tolerate pH 2 - 12, care should be taken as some RO membrane
elements will only withstand high pH values for short periods of time (e.g. 30min) or may
only withstand lower alkaline pH (e.g. pH 11.0). Cleaning fluid volume should be on the
order of 35 - 40 liters per 8” element and made with permeate water (rather than feedwater).
Table 1 shows a common cleaning procedure for HAB-affected RO elements where organic,
biological plus iron fouling, may occur. Ideally the clean should consist of a high pH clean to
soften and remove organics, biofilm and algae (Step 1), followed by a biocide in extreme
cases to kill these organisms (Step 2), Step 1 is repeated to remove dead organisms and the
remainder of the organics, and finally a low pH acid clean (Step 4) to remove carbonates and
iron and generally restore the membranes. In order to restore membrane plant performance as
quickly as possible after an algal bloom event has fouled the membranes, proprietary cleaners
522
are recommended rather than simple commodities such as caustics and acid. For instance, the
alkaline cleaner in Step 1 may comprise of a mixture of detergents, penetrants, sequestrants,
and other targeted components, and preferably with in-built biocidal properties. Each
cleaning step is further described below.
Table 1. Common cleaning procedures for SWRO elements following a HAB event.
Stage Cleaner Dose rate pH Temperature

Flush Permeate Ambient
Step 1 Alkaline ~1-3% w/vol As high as permitted 35-40 oC
detergent (pH 11.8 – 12.3)
Step 2 Non-oxidizing ~400 ppm 5.5 – 7.0 Ambient
biocide (e.g. (natural)
DBNPA)
Step 3 Alkaline ~ 1-3% w/vol As high as permitted 35-40 oC
detergent (pH 11.8 – 12.3)
Step 4 Low pH, ~4% vol/vol pH 2-4 25-30 oC
organic acid,
Step 1: Alkaline detergent procedure to remove organics

- Ideally in order to prepare membranes for most efficient cleaning, heat the cleaning
tank with only permeate (no chemical) to 35oC and circulate around membranes for
10-15 minutes maintaining temperature. This will raise the temperature of the
membranes allowing for a more efficient cleaning.
- Prepare cleaning solution as per Table 1 with chemical (strength dependent on
severity of fouling), at pH 11.8 – 12.3, and at 35-40oC.
- Low-flow pumping - Pump the mixed, preheated cleaning solution to the pressure
vessel at low flow (3 - 4.5 m³/h per pressure vessel) and low pressure (2 - 2.5 bar to
prevent permeate being produced) to displace the process water already in the
membranes. Low pressure minimizes redeposition of foulants on the membrane.
Reject the concentrate to prevent dilution of the cleaning solution.
- Measure pH and adjust if necessary, maintain temperature, and recycle around the
system for 20 - 30 minutes. If the solution shows any sign of significant discoloration
then discard and prepare fresh as per Table 1 to prevent possible membrane abrasion.
Recirculating dirty cleaning solutions will lead to poor cleaning results and possible
membrane damage.
- Once the pH, temperature, and color have stabilized, allow the elements to soak for as
long as possible, ideally 4 - 6 hours. For difficult fouling, an extended soaking period
may be required, overnight for example. Contact time is critical for organic removal.
523
- High-flow pumping. Feed the cleaning solution at high flow rates (8 - 10.2 m3/h at 1.5
- 4.0 bar) for 30 - 60 minutes. The high flow rate flushes out the foulants removed
from the membrane surface by the cleaning.
- Flush with good quality permeate to return to natural pH levels.
Step 2: Non-oxiding biocide to remove biofouling
- Prepare cleaning solution with non-oxidizing biocide.
- Low-flow pumping - Pump the mixed, preheated cleaning solution to the pressure
vessel at low flow (3-4.5 m³/h per pressure vessel) and low pressure (2 - 2.5 bar to
prevent permeate being produced) to displace the process water already in the
membranes. Low pressure minimizes redeposition of foulants on the membrane.
Reject the concentrate to prevent dilution of the cleaning solution.
- Measure pH and adjust if necessary, and recycle around the system for 20 minutes. If
the solution shows any sign of significant discoloration then discard and prepare fresh
as per above to prevent possible membrane abrasion. Recirculating dirty cleaning
solutions will lead to poor cleaning results and possible membrane damage.
- Once the pH, temperature, and color have stabilized, allow the elements to soak for as
long as possible. One hour should be sufficient unless biological fouling is very
severe. For difficult fouling, an extended soaking period (several hours) may be
required followed by further recirculation.
- 4.0 bar) for 30 - 60 minutes. The high flow rate flushes out the foulants removed
from the membrane surface by the cleaning.
- Flush with good quality permeate to natural pH levels.
Step 3: Alkaline procedure (Repeat of Step 1)
Biocide application works best if applied between alkaline cleanings. The first alkaline
application “softens” the fouling, allowing the biocide to penetrate and kill the
microorganisms, and the second alkaline cleaning step optimizes their removal.
Step 4: Low pH, organic acid procedure to remove iron hydroxide fouling
If iron hydroxide fouling is present in addition to biofouling:
- Ideally, in order to prepare membranes for most efficient cleaning, heat the cleaning
tank with permeate only to 30oC and circulate around membranes for 10 - 15 minutes
maintaining temperature, this will allow for more efficient cleaning.
- Prepare cleaning solution as per Table 1 with low pH, organic acid (such as citric
acid), at around pH 2 - 4, and heat to 25-30oC
- Low-flow pumping. Pump the mixed, preheated cleaning solution to the vessel at low
flow (3-4.5 m³/h per pressure vessel) and low pressure to displace the process water.
- Measure pH and adjust if necessary, maintain temperature and recycle around the
system for 20-30 minutes. If the solution shows any sign of significant discoloration
then discard and prepare fresh as per above to prevent possible membrane abrasion.
- Once the pH, temperature and color has stabilized allow the elements to soak for as
long as possible ideally 2-4 hours. For difficult fouling an extended soaking period
may be required, overnight for example.
524
- 4.0 bar) for 30-60 minutes. The high flow rate flushes out the foulants removed from
the membrane surface by the cleaning.
- Flush with good quality permeate to return to natural pH levels.
5 ADDITIONAL NOTES ON CLEANING
Always measure the pH during cleaning. If the pH increases more than 0.5 pH units during
acid cleaning, more acid needs to be added. If the pH decreases more than 0.5 pH units
during alkaline cleaning, more caustic needs to be added.
Long soak times should be broken up with short circulation periods. It is possible for the
cleaning solution to become fully saturated and the foulants can deposit back onto the
membrane surface. In addition, the temperature will decrease during this period, therefore the
soaking becomes less effective. It is recommended to circulate the solution regularly in order
to maintain the temperature (temperature should not drop more than 5°C) and add chemicals
if the pH needs to be adjusted. Soaking times may be reduced if it is felt that no more
deposits are being removed.
Turbid or strong-colored cleaning solutions should be replaced with a freshly prepared
solution to avoid recirculating foulants around the system causing membrane damage.
6 REFERENCES
Peña, N., Gallego, S., del Vigo, F. and Chesters, S. P. 2013. Evaluating impact of fouling on
reverse osmosis membranes performance, Desalination and Water Treatment 51, 958-
961.
Villacorte, L. O. 2014. Algal blooms and membrane-based desalination technology. Ph.D.
thesis UNESCO-IHE/TUDelft, ISBN 978-1-138-02626-1, CRC Press/Balkema, Leiden.
525
INDEX
A Amphidoma species, 469

Abengoa pilot plant, 338, 343, 410 amyotrophic lateral sclerosis, 64
Abraxis ELISA Kit for Anatoxin-a, 485, 490 Anabaena circinalis, 64
Abraxis ELISA Kit for Saxitoxin, 485, 487 anatoxin-a, 64-65, 87, 485-486, 490-491
Abraxis ELISA Reader, 485, 497 angle wells, 169, 188
acidic formalin, 510 anoxic events, 135
acidic Lugol's solution, 509-510; see Lugol's antapex, 466, 472
acidic polysaccharides, 504 Antofagasta desalination plant (Chile),
acoustic Doppler current profiler, 92; see 409-416
ADCP ANZFA, 227; see Australian and New
ADCP, 92; see acoustic Doppler current Zealand Food Authorities
profiler AOC, 147-152, 163-164, 340-341, 344,
ADDA-DM ELISA kit for 418-419, 421-422; see assimilable organic
Microcystins/Nodularins, 485, 489 carbon
adenosine tri-phosphate, 436; see ATP AOM, 134, 136-138, 140-142, 144-148,
Akashiwo sanguinea, 466-467 150, 152, 155, 159-160, 162-164,
akvoFloat process, 462-463 252-255, 257, 259-260, 266-273, 275,
akvola pilot plant, 342, 460 277-280, 282, 287, 290-297, 301-302,
alcian blue, 141, 145, 166, 501, 503, 507 304, 306-307, 310, 334, 336, 338-344,
Alert Level, 223, 240-247, 499 521; see algal organic matter
Alexandrium catenella, 467 AOM adsorption, 270-271
Alexandrium fundyense, 26, 42 AOM components, 137, 142, 160, 252, 291
Alexandrium minutum, 26, 40-41, 43 AOM deposition, 53, 58
Alexandrium monilatum, 468 apex, 466, 474
Alexandrium ostenfeldii, 468 Arabian Sea region, 22-23
Alexandrium species, 19, 21, 24, 26, 32, areolate, 466
40-44, 46, 465, 467-468 ArvorC, 100
Alexandrium tamarense, 41, 44, 46 ASP, 18-19, 21, 32; see amnesic shellfish
algae identification, 465-483, 520 poisoning
algal counting, 338, 343, 348, 352, 356, 392, ASP toxins, 32
402-404, 406, 430, 432, 440-445, 509,517; assimilable organic carbon, 133, 147-148,
see cell counts; phytoplankton counting 165-168, 340, 344-346, 419, 421, 423-424;
algal metabolites, 53-75, 315, 317-318, 322, see AOC
324, 330; see metabolites ASTM International, 136, 153, 158, 164
algal organic matter, 53-62, 133-134, 137, atmospheric correction, 119-120, 124-125
142, 162, 167, 252, 299, 301, 313, 334, ATP, 436-437; see adenosine tri-phosphate
450, 521; see AOM ATP methods, 148
algal-derived foulants, 146, 163 Aureococcus anophagefferens, 25, 35, 48,
alkaline cleaning, 364 469
alkaline Lugol's, 509; see Lugol's Australian and New Zealand Food
alongshore transport, 17, 38-39, 44 Authorities, 227; see ANZFA
alum coagulant, 320 Australian Drinking Water Guidelines, 223,
amnesic shellfish poisoning, 18, 47; see ASP 225, 227-228, 241, 249
amorphous substances, 56 auto-backwash system, 290
527
INDEX
Autonomous Microbial Genosensor, 108 bloom development, 36

autonomous underwater vehicles, 102; see bloom dynamics, 17, 36, 41-42, 51
AUVs bloom initiation, 17, 31, 36
AUVs, 102; see autonomous underwater bloom mechanisms, 17, 34
vehicles bloom mitigation, 205-221, 337, 343
Azadinium, 469 bloom transport, 17, 36, 108-109, 117
AZAs, 63-64, 69; see azaspiracid BMAA, 63-65, 80, 84; see
azaspiracid, 53, 65, 69, 75, 78-79, 81, 84, 86; beta-methyl-amino alanine
see AZAs books for identification of cells, 89, 103; see
azaspiracid shellfish poisoning, 18; see AZP taxonomic reference lists
AZP, 18-19, 22; see azaspiracid shellfish brackish water applications, 333
poisoning brackish water reverse osmosis, 401; see
BWRO
B brevetoxin, 53, 63, 65, 69, 75-76, 78-81, 83,
backwashing, 58, 60-62, 343 85, 248, 317, 320-321, 323, 325-327, 331;
bacterial decomposition, 135 see BTX
bacterial degradation, 507 BTX, 53, 63-65, 69-70, 75-76, 78-81, 83,
bacterial release, 507 85, 227, 248, 317, 320-321, 323, 325-327,
band-subtraction method, 126 331; see brevetoxin
Barcelona SWRO demonstration plant, bubble size, 276
439-446 BWRO, 356-357, 400-401, 426-427, 429,
Barka 1 desalination plant, 367-382 448; see brackish water reverse osmosis
barley straw, 205, 209, 213-218, 221-222
beach gallery systems, 169, 182, 189 C
BEAM Software, 119, 124, 130 Calcofluor, 509
Berlin akvola pilot plant, 459-464 calculated residual risk, 234
beta-methyl-amino alanine, 64; see BMAA Carman-Kozeny relationship, 158
biochemical screening, 72 cartridge filters, 252, 263, 290, 297-299,
biocide tests, 521-522 307, 311
biofilms, 149, 256, 301 case histories, 333-463
biofiltration, 278-279, 281-282, 284, 287 CCL, 224, 226; see contaminant candidate
biofouling, 53, 57-59, 62, 81; see fouling list
issues CEB, 62, 336-337, 342-343, 372-380, 386,
biogenic substances, 143 415, 427-428, 431-433, 437, 463; see
biological control, 221 chemical enhanced backwash
bioluminescence, 149, 166, 168 CEB cycle time, 373-375, 377-378
biomass, 91-93, 96-98, 101-102, 114; see CEB interval, 343, 373, 376, 379-380,
phytoplankton biomass 431-433
biomass removal, 252, 307 CEDI, 400-401; see Continuous
biopolymer concentration, 145, 147, 161 Electrodeionization
biopolymer fraction, 145 cell biomass, 155
biopolymer removal, 271-272, 286 cell counts, 225, 240-242, 509-518; see
biopolymers, 137, 140, 142-150, 152, algal counting; phytoplankton counts
160-163, 256-257, 266-269, 271-272, cell growth, 17, 35
275, 278-279, 281, 291-292, 296, 334, cell lysis, 254-257, 273, 278, 280, 294, 311
339, 342-343, 384, 392, 426, 434-435, cell size, 17, 22, 25, 41
437, 445 cellular release of organic matter, 53
Biosense ELISA kit for Domoic Acid, 485, Ceratium furca, 339, 359, 419, 421, 483;
491 see Tripos furca
bloom control, 205-214 Ceratium fusus, 410, 413
528
INDEX
CFP, 18-19, 70; see ciguatera fish poisoning 296, 301-303, 307
cfu, 256; see colony forming units coagulant addition, 258-259, 265-266
Chaetoceros, 54, 57, 267-271, 295-296, 470 coagulant dose, 257, 259-260, 262, 264-265,
Chaetoceros affinis, 54, 57, 267-271, 267-272
295-296 coagulation, 251-252, 254, 257-262,
chemical additions, 205, 211 264-272, 274-275, 277-278, 282,
chemical cleaning, 361, 369, 371-372, 375, 292-293, 296, 300-302, 304, 306-313,
401, 412-413, 415-416, 444 334, 336-337, 341-345, 349, 357,
chemical composition of AOM, 54 364-365, 380, 389, 401, 418, 422,
chemical enhanced backwash, 372, 415; see 426-428, 431-438, 441, 448, 451, 460, 462
CEB coagulation feed systems, 251, 260
chl a, 98-99, 338-339, 343, 348, 352, 356, coastal oceanography, 17, 36
368, 374, 377-378, 380, 384, 392, coccolithophorids, 509
397-398, 400, 402-406, 410, 413, 416, Cochlodinium polykrikoides, 23-24, 29,
419, 422, 426-427, 430, 432-433, 435, 31-32, 35, 39-41, 44, 48-50, 333-335,
448, 461; see chlorophyll-a; chlorophyll 350-351,357, 369, 389, 470
chlorination, 172-173, 177-178, 200, 205, colloidal biopolymers, 59
210-212, 215, 217-218, 251-257, 275, colloidal Fe-biopolymer complexes, 269
288, 301, 303, 306-307, 309, 311-312, colloidal hydrogels, 56
315-317, 323-329, 332, 341, 344, colloidal suspension, 258
348-349, 352, 356, 359, 364-365, 368, colony forming units, 256; see cfu
370, 384, 392, 400, 418, 426, 429, 448 Colpomenia sinuosa, 456
chlorination - dechlorination process, 301, commercial ELISA kit protocols, 486
306 competitive interactions, 213-214
chlorine dose, 316, 327-328 comprehensive monitoring strategy, 230
chlorophyll, 98-99, 338-339, 343, 348, 352, conductivity, temperature, depth, 2, 94,
356, 368, 374, 377-378, 380, 384, 392, 97-98, 113, 135; see CTD
397-398, 400, 402-406, 410, 413, 416, conservative flux, 296
419, 422, 426-427, 430, 432-433, 435, contaminant candidate list, 226; see CCL
448, 461; see chlorophyll-a; chl-a Continuous Electrodeionization, 401; see
chlorophyll fluorescence, 89, 92-95, 98, 101, CEDI
105-106, 111, 114, 117 control variables, 373
chlorophyll-a, 89, 98-99, 114, 338-339, 343, conventional plant water quality parameters,
348, 352, 356, 368, 374, 377-380, 384, 134,139, 163
392, 397-406, 410, 413, 416, 419, 422, copper sulfate, 205, 211, 215, 220; see
426-427, 430, 432-433, 435, 448, 461; see CuSO4
chlorophyll; chl-a Coscinodiscus wailesii, 30, 48, 471
chlorophytes, 17, 90 counting strategies, 338, 343, 348, 352, 356,
ciguatera fish poisoning, 18, 70, 87; see CFP 392, 402-404, 406, 430, 432, 440-445,
ciguatoxin, 53, 63, 65, 70, 75; see CTX 509, 517; see algal counting methods
cingulum, 466-468, 470, 474-478, 483-484 criteria for siting, 198
CIP, 336-337, 342-343, 372-375, 377-380, critical control points, 235, 238-239
415, 427, 432-433, 436-438, 456, 463; see cryptophytes, 17, 90
Clean-In-Place CTD, 92, 94, 97-98, 113, 135; see
circulation models, 89, 109 conductivity, temperature, depth
Clean-In-Place, 372; see CIP CTX, 63-64, 70; see ciguatoxin
cleaning solutions, 522-525 CuSO4, 211, 216, 219; see copper sulfate
closed-loop stripping analysis, 73; see CLSA cyanobacteria, 90, 101, 105
CLSA, 73; see closed-loop stripping analysis cyanobacterial toxins, 63-64, 80, 86,
coagulant, 252, 257-272, 275, 278-279, 282, 224-225, 228
529
INDEX
cyclic imine, 53, 63-65, 67, 75-76 disk filters, 253, 287-288, 301
Cylindrospermopsin, 53, 73-74, 78, 82 dissolved air flotation, 58, 134, 252, 259,
Cylindrospermopsis, 227 273, 308-312, 314-317, 331-333, 341,
Cylindrotheca closterium, 23, 29, 471 349, 383, 385, 401, 441; see DAF
cysts, 31-33, 36, 42, 44, 50, 466 dissolved organic carbon, 136, 139,
CytoSense, 106 142-143, 267, 301, 359; see DOC
dissolved oxygen, 133-135
D
distribution of AOM, 179, 200
DA, 63-64, 66-67, 72, 227, 485-487, 491,
diurnal changes in pH, 135; see pH
496-497; see domoic acid
DMF, 316, 333-334, 336, 341-345, 348,
DAF, 58, 134, 155, 162, 243, 251-252, 254,
350-352, 392, 439-441, 444-446,
257, 259, 261-262, 264-266, 272-278,
448-449, 451, 453, 455-456
283, 285-286, 300, 307, 315-318, 328,
DO depletion, 135
333, 335-336, 341-344, 348-350, 352,
DOC, 136-137, 139-140, 142-144, 165, 267,
354, 383-390, 400-401, 406, 439-441,
272, 338, 359, 363-364, 384, 392, 400,
444-445, 459, 461; see dissolved air
426, 435-436, 440; see dissolved organic
flotation
carbon
DAFF, 273, 276; see flotation and filtration
domoic acid, 53, 63, 65-67, 73-74, 77-80,
data storage and distribution, 89, 112
82, 84, 87, 227, 231-232, 234, 248-249,
data subscription services, 119, 123
315, 317-318, 320-321, 323, 325-327,
DBPs, 253-254, 311; see disinfection
331, 485-487, 491-492, 494, 497, 500; see
byproducts
DA
dechlorination, 173, 251-253, 257, 275, 298,
domoic acid quantification, 497
301, 303, 306-307
dorsal side, 466
deep water intakes, 178-180, 340-341, 344,
downwelling, 38-39, 108
450, 453
DSP, 18-19, 35, 48, 227, 472-473, 480; see
definition of the levels, 223, 242
diarrhetic shellfish poisoning
destratification, 205, 208
DSP-producing species, 35
detection level, 223, 241-242
DTxs, 68-69; see dinophysistoxins
detection techniques, 53, 71; see toxin
dual flow, 174
detection
dual media filter, 349-350, 444
deterioration in feedwater quality, 134, 163
diarrhetic shellfish poisoning, 18, 227, E
472-473; see DSP Ekman transport, 38
diatoms, 17, 21, 25, 30, 35, 43, 90, 107, 224, electrolysis, 211-212, 214-215, 218
232, 325, 400, 403, 406, 413, 427, 430, electron microscopy, 89, 103, 105
460 ELISA format, 486
differential pressure, 361-362, 398 ELISA methodology, 485-486
dinoflagellates, 17, 21-22, 25, 29-30, 33, 35, ELISAs, 64, 72, 74-75, 77, 80, 84, 485-492,
39-40, 42, 44, 46, 49-50, 90, 96, 98, 105, 494, 496-497, 499; see Enzyme-Linked
224, 232, 358, 369, 393, 403, 413 Immunosorbent Assays
Dinophysis, 26-27, 35, 37, 39, 315-316, 331 empirical models, 108
Dinophysis acuminata, 26, 37, 39, 41, 48, Environmental Protection Agency, 336; see
472 EPA
Dinophysis acuta, 40-41, 48 Environmental Sample Processor, 108, 115,
Dinophysis blooms, 108 117; see ESP
Dinophysis fortii, 473 enzyme substrate, 487, 489-490; see TMB
Dinophysis tripos, 473 Enzyme-Linked Immunosorbent Assays, 64,
dinophysistoxins, 53, 68, 75; see DTxs 72, 74-75, 77, 80, 84, 485-492, 494,
disinfection byproducts, 252-253, 311; see 496-497, 499; see ELISAs
DBPs
530
INDEX
EOM, 54-55 449, 461-463

EOM substances, 54 forecast, 89, 108-111, 113-115, 117; see
EPA, 336; see Environmental Protection HAB forecast systems
Agency formaldehyde solution, 510
epicone, 466, 477 foulant components of AOM, 137
epidemiological data, 485 fouling, 334, 336-346, 361-362, 364, 371,
epitheca, 466-468, 470, 472-477, 482-483 373, 377-378, 380, 410-416, 422-424,
ERDDAP site, 121 428, 431-434, 436-438, 445, 455-456, 463;
ESA, 121, 123-125, 127, 130; see European see biofouling; fouling issues
Space Agency fouling diagnosis, 519
ESP, 108, 117; see Environmental Sample fouling indices, 133, 147, 153-154, 158, 165
Processor fouling issues, 53, 57-62, 79, 81, 83, 87; see
euglenophytes, 17, 90 biofouling
European Space Agency, 121, 130; see ESA fouling organics, 273, 278
eutrophication, 214-217, 221 fouling rate trends, 377
excessive production of AOM, 54 fouling reversibility, 270
extracellular levels, 232 fresh- and brackish-water toxins, 223
extracellular metabolites, 316, 318, 320; see freshwater bioluminescent-AOC test, 149
metabolites freshwater biotoxins, 65
extracellular polysaccharides, 54 fronts, 17, 36-37, 44-45
extracellular saxitoxin, 320, 331 FTAs, 67
extraction of DA, 496 Fujairah pilot plant, 347-354
F G
FAI, 126-127; see floating algal index gallery types, 182
ferric coagulant, 252, 260, 263, 267, 272, Gambierdiscus spp, 474
301-303, 307 Gas Atacama desalination plant, 400
Ferrybox systems, 89, 98-99, 101-102, 114 Gas Atacama UF plant, 343
filter grains, 279-280 GCDP, 336, 340, 343, 447-458; see Gold
filter media backwash, 279 Coast Desalination Plant
filter media type, 280 Geographic Information System, 199; see
filtration cycle time, 361, 365, 371, 373 GIS
filtration cycles, 444 geosmin, 53, 63, 65, 71-74, 83, 85, 317-318,
filtration flux, 271, 292, 296 321-322, 324, 328, 331-332; see GSM
first stage filter, 281 germination, 31, 36
fixatives, 509-517; see preservatives Giovanni site, 122, 130
flagellar area, 466 GIS, 199; see Geographic Information
Floating Algal Index, 126-127; see FAI System
floating blanket, 388 GMF, 55, 58, 62, 134, 251-252, 254, 257,
flocculation, 205, 209-210, 212, 214-215, 259, 261-262, 265-266, 272-273,
217, 219-222, 251, 257-261, 264-268, 277-281, 284, 286-287, 289-290, 297,
272-275, 277-278, 300-301 307, 317-318, 328, 336; see granular
flotation and filtration, 273, 313; see DAFF media filters
FlowCam, 100, 106 GMF filter performance, 251, 286
fluorescence, 509, 511-512 GMF filter rates, 257
fluorescence microscopy, 89, 103, 105 GoFlo bottles, 95
fluorometers, 98 Gold Coast Desalination Plant, 336, 340,
flux, 336-337, 342-343, 348, 356-358, 343, 447-458; see GCDP
360-361, 363-364, 368, 370-374, 377, Gonyaulax fragilis, 29, 49, 474
379-381, 384, 392, 400, 410-413, Gonyaulax hyaline, 29
415-416, 419, 427, 431-434, 437, 444, granular media filters, 55, 58, 62, 134,
531
INDEX
251-252, 254, 257, 259, 261-262, highly crystalline PVDF modules, 365
265-266, 272-273, 277-281, 284, HOBr, 254; see hypobromus acid
286-287, 289-290, 297, 307, 315, HOCl, 254; see hypochlorus acid
317-318, 328, 336; see GMF horizontal wells, 169, 182, 188-189, 196,
granular media filtration, 55, 134, 251-252, 201; see HDD
265, 278, 288; see GMF HPLC, 72-74, 78, 98; see High Performance
gravity DMF, 336, 343, 448 Liquid Chromatography
gravity media filters, 251, 276, 282-285 human illnesses, 19
grazers and trophic cascades, 205, 214 humic substances, 140, 143-144
green tide, 20, 30 hybrid flotation-filtration, 461
GSM, 53, 63, 65, 71-74, 83, 85, 317-318, Hydrobios Combined Plate Chamber, 510;
321-322, 324, 328, 331-332; see geosmin see Utermöhl-Chamber
guideline values, 223-225, 227-228, 241, 243 hydrochloric acid, 361; see HCl
Gulf of Maine (USA), 39 hydrogen peroxide, 205, 209, 211-212,
GYM, 63, 67-68; see gymnodimines 215-216, 219-220
gymnodimines, 63, 67-68; see GYM hydrogen sulfide, 333, 335-336, 341,
Gymnodinium breve, 69, 85 391-398, 461; see H2S
Gymnodinium catenatum, 21, 31-33, 41, 43, hydrophobic compounds, 143
47, 475 hypobromus acid, 254; see HOBr
hypochlorus acid, 254; see HOCl
H hypocone, 466, 477
H2S, 333, 335-336, 341, 391, 398, 461; see hypotheca, 466-468, 470, 472-476, 482-483
hydrogen sulfide hypoxic conditions, 135
HAB buoys, 93
HAB control, 205, 207, 210, 212 I
HAB forecast systems, 15, 89, 108-111, identification and enumeration of HAB
113-115, 117; see forecast organisms, 89, 102, 465-483, 509-517; see
HAB observing system, 89-117 algal counting methods
HAB prevention, 205-222 IFCB, 100, 106-107, 116; see Imaging
HACCP, 223, 229, 235-240, 247-248; see FlowCytobot
Hazard Analysis and Critical Control Imaging FlowCytobot, 100, 106-107, 116;
Point see IFCB
haptophytes, 17, 90 immunochromatographic tests, 485, 492, 494
hardware requirements, 119, 124 in-line design, 371
Hazard Analysis and Critical Control Point, Independent Water and Power Plant, 349;
223, 229, 235-240, 247-248; see HACCP see IWPP
HCl, 361, 364-365, 426-427; see Indian Ocean, 29
hydrochloric acid infrared spectroscopy, 521; see IR
HDD, 169, 182, 188-189; see horizontal inlet water monitoring, 350
wells intake, 122-123, 133-137, 146, 148,
HDF, 123; see Hierarchical Data Format 150-153, 155-158, 161, 163, 169-199,
Heterocapsa circularisquama, 475 333, 335-341, 343-344, 348-349,
Heterocapsa triquetra, 476 351-353, 355-359, 362, 364-365,
Heterosigma akashiwo, 22, 31, 40-41, 45, 368-370, 383-387, 390, 392, 394-395,
48, 51, 476 397-401, 406, 409-410, 413, 418,
Hierarchical Data Format, 123; see HDF 422-423, 425-430, 434, 437, 439-441,
high biomass algorithms, 119, 126 443, 447-457, 460-461
high molecular weight AOM, 55, 59 intake design, 169-199
High Performance Liquid Chromatography, intake feedwater quality, 133-134
72-74, 78, 98; see HPLC intake siting, 91, 198
high transmembrane pressure, 359 integrated hose sampling, 95
532
INDEX
integrity tests, 522 lipopolysaccharides, 227; see LPS

Intergovernmental Oceanographic liquid chromatography/mass spectroscopy,
Commission, 21-22, 44, 47, 51, 512, 230; see LC/MS
516-517; see IOC LMW, 256; see low molecular weight
international guideline values, 224 LOD, 64, 73-75, 485-486; see limit of
internet access to imagery, 119, 121 detection
intracellular metabolites, 315, 317-318; see LOI, 520; see loss on ignition
metabolites loss on ignition, 520; see LOI
intracellular organic matter, 54-55, 254-256; low molecular weight, 256; see LMW
see IOM low-nutrient stress, 54
intracellular toxins, 231-232, 238-239 LPS, 227; see lipopolysaccharides
inverted microscope, 511 LPT, 109; see Lagrangian particle transport
IOC, 21-22, 44, 47, 51, 512, 516-517; see Lugol's, 509-510, 513; see acidic Lugol's
Intergovernmental Oceanographic solution; alkaline Lugol's; Lugol's solution
Commission Lugol's solution, 509-510; see Lugol's
IOM, 54-55, 254-256; see intracellular
organic matter M
IOM classification, 55 MAC, 224-225; see maximum acceptable
IR, 521; see infrared spectroscopy concentrations
iron hydroxide flocs, 269-271, 296 macroalgae, 17-18, 30; see seaweed
IWPP, 349; see Independent Water and maitotoxin, 65, 70; see MTX
Power Plant manual cell counter, 511-512
marine snow, 54, 62
J mass spectrometry, 72, 77-79, 82, 84-88
Jacobahaven UF-SWRO demonstration
mass transfer coefficient, 436; see MTC
plant, 339-340, 425-438
maximum acceptable concentrations,
224-225; see MAC
K maximum contaminant levels, 224, 226; see
Karenia, 19, 23-24, 27, 36-37, 41, 44 MCLs
Karenia brevis, 19, 27, 36-37, 41, 44, 477 MCLR, 53, 63, 65, 68, 72-74, 76-80, 82-84,
Karenia mikimotoi, 477 86-87, 224-225, 231; see microcystin
Karlodinium veneficum , 477 MCLs, 224, 226; see maximum contaminant
Korea, 40, 48 levels
measuring biofouling potential, 133, 147
L MED, 63, 170, 173-174, 325, 327, 348-349
La Chimba desalination plant (Chile), 335, media information dissemination, 223, 245
338,341, 391-397 Mediterranean Sea, 29, 44, 49
Lagrangian particle transport, 109; see Mejillones desalination plant (Chile),
LPT 399-408; see Gas Atacama Plant
lateral flow format, 485, 492 membrane autopsy, 519-525
LC-OCD analysis, 145-146 membrane cake and pore constriction, 60
LC/MS, 230; see liquid membrane cleaning, 252, 292, 299, 306
schromatography/mass spectroscopy membrane fouling potential, 134, 161
levels of an alert framework, 223, 241-242 Membrane Fouling Simulator, 133, 148,
life history, 17, 31 152, 168, 304
light microscopy, 89, 103-105, 107 membrane integrity, 435
limit of detection, 74, 488, 490; see LOD MERIS, 119-121, 124, 126, 131
Lingulodinium polyedrum, 21, 40-42 metabolite removal, 317-320
lipophilic toxins, 75, 77, 86 metabolites, 53-75, 315-318, 320, 322, 324,
330-331; see algal metabolites;
533
INDEX
extracellular metabolites; intracellular MTC, 436, 438; see mass transfer coefficient
metabolites MTX, 70; see maitotoxin
meteorological conditions, 36 mucilage, 18, 29-30, 44, 47, 49
methods for measuring TEP, 501, 503-504, mucilage events, 54
506-507 multi effect distillation, 63; see MED
methylisoborneol, 53, 71-72, 85; see MIB Multi Stage Flash, 63, 170, 173-174,
MEW, 385; see Ministry of Electricity and 325-327, 330, 333, 356-357, 368-370; see
Water MSF
MF design flux versus reduced flux, 360 multi-spectral scanners, 119
MF/UF pretreatment, 251-252, 266, 290, multiple barrier concept, 223, 228, 230
308, 501
MFI, 340, 344-345, 411, 435, 438
MFI-UF, 133, 159, 251-252, 266, 290, 308, N
501; see Modified Fouling Index-UF N-DBPs, 253; see nitrogenous DBPs
MFs, 53, 55, 58-62, 81, 87, 148, 152, 304, N-Nitrosodiethylamine, 322; see NDMA
334-337, 341-343, 356-365, 410-413, nanofiltration, 55, 320-321, 330, 332; see
460-462; see microfiltration NF; NF membranes
MIB, 63, 71-73; see methylisoborneol nanoplankton, 96, 106, 513
microbes, 205, 212 NASA OBPG system, 123
microbiological counts, 521 National Oceanic and Atmospheric
microcystin, 53, 63, 65, 68, 72-74, 76-80, Administration, 119, 121, 130; see NOAA
82-84, 86-87, 224-225, 231; see MCLR natural organic matter, 169, 171, 178,
Microcystis, 207-209, 217-219, 222, 180-181, 183-187, 189, 191-194,
294, 296, 315-316, 319, 329, 331 200-201, 203, 339, 440, 445-446; see
microfiltration, 53, 55, 58-62, 80-81, 87, NOM
148, 152, 304, 334-337, 341-343, NDMA, 322; see N-Nitrosodiethylamine
356-365, 410-413, 460-462; see MFs neurotoxic shellfish poisoning, 18, 477; see
microfiltration system, 360 NSP
microfiltration/ultrafiltration, 342 New Zealand, 29, 43, 47, 49
microplankton, 96, 116 NF, 55, 320-321, 330, 332; see nanofiltration
microscopy, 89, 94, 103-105, 107, 114-115 NF membranes, 55, 320-321, 330, 332; see
migration, 35-36, 39-40, 48, 179, 200-201; nanofiltration
see vertical migration Niskin bottle, 95
Ministry of Electricity and Water, 385; see nitrogenous DBPs, 253; see N-DBPs
MEW NMR techniques, 66
mitigation, 205-214 no-observed adverse effect levels, 224; see
mixing, 205, 208 NOAEL
mixing intensity, 258, 261, 267-268 NOAA, 119, 121, 130; see National Oceanic
Modified Fouling Index, 133, 153-154, and Atmospheric Administration
158-159, 164-167, 340, 344-345, 411, NOAEL, 224-225, 234; see no-observed
435, 438; see MFI adverse effect levels
Modified Fouling Index-UF, 133, 159, Noctiluca, 23-25, 30-31, 43, 45
251-252, 266, 290, 308, 501; see MFI-UF Noctiluca algorithms, 119, 126
MODIS, 119-121, 123-124, 126-129, 131 Noctiluca scintillans, 23-25, 30, 43, 45,
molecular weight cut-off, 320 294-296, 339, 369, 478
monitoring of phytoplankton, 512 nodularin, 53, 63, 65, 68, 73-74, 77-78, 84
monitoring program design, 89-118 NOM, 169, 171, 178, 180-181, 183-187,
MSF, 63, 170, 173-174, 325-327, 330, 333, 189, 191-194, 339, 440, 445-446; see
356-357, 368-370; see Multi Stage Flash natural organic matter
non-oxidizing biocide, 254, 305, 524
non-toxic HABs, 335-336, 419
534
INDEX
non-UV absorbing compounds, 144 467; see PSP

non-volatile toxins, 327 particulate fouling potential, 133-134, 153,
normalized pressure drop, 436; see NPD 155, 162-164
NPD, 436; see normalized pressure drop particulate organic carbon, 134, 139; see
NSP, 18-19, 22, 33, 477; see neurotoxic POC
shellfish poisoning passive screen intake, 175
numerical models, 89, 108-111 PDA, 73-74; see photo diode array detection
nutrient enrichment, 30, 33; see pollution PDI, 268-271, 290-293, 295-297; see
nutrient source, 255 pressure driven inside-out
nutrients, 24, 31, 33, 37-39 PDI membranes, 290-291
PDI UF membranes, 269-271, 291, 293, 296
O PDO, 290-292; see pressure driven
OA, 63-64, 68-69, 72, 75, 227 outside-in
OBPG, 121-124, 130; see Ocean Biology pelagophytes, 17, 90
Processing Group permeate quality, 266, 271-272, 290, 293,
observing system for blooms, 89-114 305-307
OCD, 133, 137-138, 140, 142-147, 150, 166; PES, 369; see polyethersulfone
see organic carbon detection pH, 132-135, 141, 145, 150, 163; see diurnal
Ocean Biology Processing Group, 121-124, changes in pH
130; see OBPG Phaeocystis antarctica, 479
okadaic acid, 53, 63, 65, 68, 72, 75, 77-78, Phaeocystis globosa, 479
83, 85-86, 227, 231-232, 234, 248, 315, Phaeocystis pouchetii, 480
317, 321, 323, 325-327, 331; see OA Phaeocystis spp, 421, 479-480
OLCI sensor, 119 phase contrast, 509, 511
Oman Power and Water Procurement, 369; photic zone, 135
see OPWP photo diode array detection, 73; see PDA
ON, 138, 140, 143, 145, 166; see organic photosynthetic pigments, 98, 105
nitrogen phycocyanin, 101
OPWP, 369; see Oman Power and Water phytoplankton, 17-18, 20, 29-30, 33, 44, 48,
Procurement 50, 89-96, 98-99, 101-107, 109, 112,
organic carbon detection, 133, 137, 140, 114-117
142, 166; see OCD phytoplankton analysis, 95, 103, 106,
organic nitrogen, 138, 140, 143, 145, 166; 115-116, 465-484, 509-518; see cell
see ON counts; algal counting
ORP, 392, 395-396, 400, 402, 405-406, 422, phytoplankton biomass, 92-93, 98, 101-102,
448; see oxidation-reduction potential 114; see biomass
Ostreopsis ovata, 71, 77 phytoplankton counting, 509-518; see cell
Ostreopsis siamensis, 478 counts
ovatoxins, 63-64, 70; see OvTX phytoplankton population dynamics, 93; see
OvTX, 63-64, 70; see ovatoxins population dynamics
oxidation-reduction potential; see ORP phytoplankton samples, 509-512
oxygen depletion, 333, 391 phytoplankton taxonomy, 93; see taxonomy
ozonation, 211, 215, 220 picoplankton, 513
P pilot Scale DAF Testing, 389
PAC, 256, 301; see powdered activated pilot studies, 342, 347-354, 390, 409-416,
carbon 425-446, 459-464
PACls, 258; see polyaluminum chlorides plankton net, 96, 113, 115
palytoxin, 63, 65, 70-71, 77, 80, 83, 85-86; Plankton Toolbox, 105, 116, 511, 516
see PLTX planktonic bacteria, 59
paralytic shellfish poisoning, 18, 50, 227,
535
INDEX
PLTX, 63-64, 70-71; see palytoxin red Noctiluca, 30, 126

POC, 134, 139; see particulate organic Red Sea, 180-181, 185, 195-196, 201-202
carbon red tides, 18, 25, 30, 33-34, 37, 40, 42,
pollution, 19, 30, 33-34, 42, 46, 50, 52, 134, 44-50, 52
136, 163, 205-207, 214-215, 222; see region-specific algorithms, 120
eutrophication remote sensing, 89, 108, 110, 112, 114-115,
polyaluminum chlorides, 258; see PACls 117, 119-132; see satellite remote sensing
polyethersulfone, 369; see PES removal of marine algae, 251, 277
polysaccharide compounds, 55 residual handling, 134
population dynamics, 93; see phytoplankton reverse osmosis, 33-337, 341, 348-354,
population dynamics 356-357, 360-364, 369-371, 381,
potassium permanganate, 205, 212 384-385, 390, 392, 394, 398, 400, 410,
powdered activated carbon, 256, 309; see 412, 418-419, 421-422, 425, 446, 448,
PAC 451, 453, 455-456, 519-525; see RO
Practical Salinity Scale, 135 risk assessment
precautionary impact prevention, 205 maximum risk, 230, 237
preservatives, 509-517; see fixatives residual risk, 230, 234, 237-239, 247
pressure driven inside-out, 268, 291; see PDI risk evaluation
pressure driven outside-in, 290-291; see detection and quantification methods, 231
PDO exposure assessment, 231
pressure filters, 251, 279, 284-285 hazard characterization, 231
pretreatment, 54, 58, 87, 133-134, 136-137, hazard identification, 231, 249
145-152, 155-158, 160-163, 165-168, management and monitoring, 231
252-307, 333-334, 337, 344-345, 349, public call for data, 231
352, 435-436, 444 risk characterization, 231
process line-up, 429, 441 toxicological and epidemiological data,
projected CIP interval, 378-379 231
prokaryotes, 17, 90 RO, 333-337, 341, 348-354, 356-357,
Prorocentrum, 254, 259, 264, 294-296, 314, 360-364, 369-371, 381, 384-385, 390,
400, 402-403, 406, 410, 413, 480-481 392, 394, 398, 400, 410, 412, 418-419,
Prorocentrum graciles, 400, 402-403, 406 421-422, 425, 428-429, 432, 434-438,
Prorocentrum lima, 480 441-442, 446, 448, 451, 453, 455-456,
Prorocentrum micans, 40-41, 481 519-525; see reverse osmosis
Prorocentrum minimum, 259, 264, 314, 481 RO feedwater, 136, 146, 148, 151-152, 155,
Pseudo-nitzschia, 19, 21, 23-24, 27-28, 162
42-43, 45, 47-49, 315, 410, 413, 482, 500 RO membranes, 147, 151-153, 162
Pseudo-nitzschia delicatissima, 410, 413 rotating drum screens, 169, 174
PSP, 18-19, 21, 32, 35, 39, 43-44, 227, Ruttner sampler, 95
467-468, 475; see paralytic shellfish
poisoning S
PSP-toxin, 39 salinity, 133, 135, 141, 144-146, 149-150,
Ptychodiscus brevis, 69; see Karenia brevis 160, 163, 165
sampling methods, 89, 114
satellite remote sensing, 89, 108, 110, 112,
Q 114-115, 117, 119-132; see remote
quantification by fields, 514 sensing
quantification using transects, 514 saxitoxin, 64, 66-67, 74, 76, 224, 232, 234,
486-488, 493; see STX
R SBS, 349; see sodium bisulfite
Ranney Wells, 169, 182, 187 scale of blooms, 17, 34
raphidophytes, 17, 22, 90 scanning electron microscopy, 105; see SEM
536
INDEX
scattered light, 136 Sohar Industrial Port Area, 335, 357; see
Scotia Rapid Test, 85-486, 492-494, 496 SIPA
Scrippsiella trochoidea, 482 Sohar SWRO desalination plant, 355-356,
SDI, 134, 136-137, 152-153, 155-159, 358
163-164, 166, 262-266, 272, 278-280, solid phase microextraction, 73; see SPME
290, 297-299, 301, 520; see Silt Density species identification, 102-108, 465-484,
Index 509, 512
Sea of Oman, 20, 22-24, 29, 32, 38-39 species identification assistance, 509, 512
sea sawdust, 30 specific ultraviolet absorbance, 136; see
sea surface temperature, 120-121, 126, 128, SUVA
131; see SST SPME, 73; see solid phase microextraction
seabed gallery systems, 169, 182, 190, 192, SRB, 335, 341, 393-395, 397-398; see
196 sulfate-reducing bacteria
SeaDAS, 119, 121, 123-124, 128, 130 SRS, 515-516; see Sedgewick Rafter slide
seasonal fouling pattern, 431 SST, 120, 123, 129-130; see sea surface
seawater quality, 334, 337, 340, 344, 350, temperature
352, 357, 361, 374, 402, 413, 449, standard algorithms, 119, 125, 130
452-455, 457 STX, 64, 66-67, 74, 76, 224, 232, 234,
seaweed, 17-20, 38, 50; see macroalgae 486-488, 493; see saxitoxin
SeaWiFS, 119, 124, 127, 132 STX analogues, 66
Secchi disc, 89, 97 STX-equivalents, 315
Sedgewick Rafter slide, 509, 511-512, 515; subsurface intakes, 169, 171-172, 176,
see SRS 181-185, 194, 196, 200-203
sedimentation chamber, 513 sulcus, 466-468, 470, 474-477
self-cleaning strainers, 337, 356, 358-360, sulfate-reducing bacteria, 136, 393, 397
363 sulfuric acid, 205, 212
SEM, 105, 114; see scanning electron surface charge, 17, 25, 29
microscopy surface intake, 169, 172, 176, 182, 199
semi-volatile organics, 326 SUVA, 136-137, 139, 143; see specific
settling chambers, 509-513 ultraviolet absorbance
seven principles of HACCP, 236 SWRO brine, 401
Sharjah Electricity, 351 SWRO elements, 522-523
shear, 37
ship-based sampling, 119 T
shock chlorination, 243-244 Tampa Bay Seawater Desalination Plant,
shore observation, 352 417-424; see TBSDP
short-term mitigation measures, 337, 343, taste and odor compounds, 53, 55, 63,
359, 361-363 71-72, 74
Shuwaikh Desalination Plant, 334, 342, taxonomic reference lists, 104; see books for
383-390 identification
Silt Density Index, 133-134, 153-154, taxonomic terminology, 466
164-166, 262, 299, 520; see SDI taxonomy, 93, 104, 112; see phytoplankton
single media filters, 280 taxonomy; algae identification
SIPA, 335, 357; see Sohar Industrial Port TBSDP, 417, 424; see Tampa Bay Seawater
Area Desalination Plant
site selection, 177, 197 TDS, 135, 149, 165; see total dissolved
sludge blanket clarifiers, 324 solids
sludge treatment, 315, 324 temperature, 120-121, 126, 128, 131,
sodium bisulfite, 349, 418, 422; see SBS 133-135, 137, 139, 148, 150, 154,
157-158, 163-164, 338, 340, 348, 352,
356, 368-369, 371-372, 384, 392, 397,
537
INDEX
400-402, 410, 414, 418-419, 426-431, 426, 445, 501, 507; see TEPs
433, 438, 440, 442-444, 448, 452, 454, Trichodesmium, 120, 125, 127-128, 131-132,
460-462 135, 333, 336, 340, 369, 447-450, 452-457
TEP0.4µm, 501-504, 506-507 Trichodesmium algorithms, 119, 127
TEP10kDa, 501, 504-507 Trichodesmium erythraeum, 23, 30, 46, 340,
TEP methods, 142, 146-147 447-448, 452-453, 457, 483
TEP precursors, 57, 338-340, 342 Trichodesmium thiebauthi, 67
TEPs, 53-63, 76-77, 81, 83, 87, 133, trichotoxin, 63-65, 67, 84
136-138, 140-142, 145-147, 153-155, Tripos furca, 483; see Ceratium furca
160-166, 168, 255, 257, 267, 270, 279, Tripos fusus, 484
286, 290-292, 294-295, 298-299, 301, TSS, 153, 156, 163, 293; see total suspended
304, 309, 315-316, 334, 338-343, 345, solids
351-352, 384, 392, 410-411, 413-414, TTX, 66; see tetrodotoxins
426, 430, 437, 445, 501, 503-507; see turbidity, 133-134, 136-137, 139, 153,
Transparent Exopolymer Particles 155-156, 159, 163, 338, 341, 343, 348,
Tetraselmis suecica, 254-256 351-353, 356, 368, 384-385, 390, 392,
tetrodotoxins, 53, 63, 65-66, 73, 78, 86-87; 398, 400-402, 405, 410, 418, 423, 426,
see TTX 428, 431, 434-435, 440-445, 448, 450,
thecate, 466-467, 469, 472-474, 476, 478, 454-455, 460-461, 463
480-482 turbulence, 37
thermal desalination, 169-170, 172-173, two-stage filtration, 251, 278, 281
202, 315, 324-328
TMB, 487-491; see enzyme substrate U
TMP, 268-269, 271, 292-293, 336-337, 343, UAE, 334, 347, 349, 389; see United Arab
359, 388, 390, 427, 431; see Emirates
transmembrane pressure UF, 53, 55, 58-62, 81, 83, 87, 133, 137,
TOC, 54, 136-137, 139, 142, 147, 150-151, 146-147, 154-155, 158-165, 167-168,
163, 338, 348, 351-352, 356, 361, 368, 318-320, 329, 331, 333-335, 337-345,
384, 392, 400, 418-419, 421-422, 426, 348, 352, 367-372, 374, 381, 384-386,
440, 448, 453-456, 460-463; see total 388, 390, 400-401, 403, 406-407,
organic carbon 410-413, 415-416, 425-441, 444-446,
TOC measurements, 137 448, 459-462; see ultrafiltration
total dissolved solids, 135; see TDS UF backwash waste, 401
total organic carbon, 54, 136-137, 139, 142, UF backwashing by SWRO concentrate, 434
147, 150-151, 163, 338, 348, 351-352, UF foulant levels, 435
356, 361, 368, 384, 392, 400, 418-419, UF system design, 370, 381
421-422, 426, 440, 448; see TOC UF system performance, 372
total suspended solids, 137, 153; see TSS UF systems, 260-261, 266-268, 272,
toxin detection, 32; see detection techniques 292-293, 296, 318-319
toxin guidelines, 223-247 UF wastewater handling, 427, 434
toxin measurement methods, 71-75, 485-500 ultrafiltration, 55, 58, 79, 81, 85, 137,
toxin oxidation, 328 145-146, 154, 159, 162, 165-167, 315,
toxin removal, 315-328 318, 329-330, 333-334, 336, 340, 342,
toxin screening methods, 485-500 345, 385, 399, 401-402, 406, 409-410,
transmembrane pressure, 159, 255, 268, 319, 425, 428-429, 431-432, 438, 440, 444; see
336, 359, 388, 414, 431, 461, 463; see UF
TMP ultrasound, 211, 215, 219, 222
transmission electron microscopy, 105 ultraviolet absorption, 136, 139
Transparent Exopolymer Particles, 53-54, United Arab Emirates, 333, 347, 349; see
56, 76, 81, 83, 87, 133, 136, 138, 145, 164, UAE
166, 168, 255, 315, 334, 345, 352, 411,
538
INDEX
UPW, 501, 503-504, 506

upwelling, 17, 33, 37-39, 44, 51, 116
upwelling relaxation, 38
Utermöhl method, 104-105, 116
Utermöhl-Chamber, 510; see Hydrobios
Combined Plate Chamber
UV radiation, 137
V
velocity caps, 169, 175, 177, 199
vertical cell distributions, 17, 39, 93
vertical migration, 36, 39-40, 48, 179,
200-201; see migration
Vibrio, 149, 152, 163-164
Vulcanodinium rugosum, 68
W
water quality monitoring, 133-162
water quality parameters, 338, 423, 430, 442
water sampling devices, 93, 95, 99, 101
water supply system, 53
water treatment plant, 317-318, 324; see
WTP
WHO, 223-226, 230-232, 234-235, 241,
247, 250; see World Health Organization
windrows, 37-38
Wirewalker, 100
World Health Organization, 223, 226, 235,
247, 250; see WHO
WTP, 317-318, 324; see water treatment
plant
X
xanthan gum, 503-504, 506-507
Z
zooanthids, 70
539
IOC Manuals and Guides
No. Title
1 rev. 2 Guide to IGOSS Data Archives and Exchange (BATHY and TESAC). 1993. 27 pp. (English, French,
Spanish, Russian)
2 International Catalogue of Ocean Data Station. 1976. (Out of stock)
3 rev. 3 Guide to Operational Procedures for the Collection and Exchange of JCOMM Oceanographic Data.
Third Revised Edition, 1999. 38 pp. (English, French, Spanish, Russian)
4 Guide to Oceanographic and Marine Meteorological Instruments and Observing Practices. 1975.
54 pp. (English)
5 rev. 2 Guide for Establishing a National Oceanographic Data Centre. Second Revised Edition, 2008. 27 pp.
(English) (Electronic only)
6 rev. Wave Reporting Procedures for Tide Observers in the Tsunami Warning System. 1968. 30 pp. (English)
7 Guide to Operational Procedures for the IGOSS Pilot Project on Marine Pollution (Petroleum)
Monitoring. 1976. 50 pp. (French, Spanish)
8 (Superseded by IOC Manuals and Guides No. 16)

9 rev. Manual on International Oceanographic Data Exchange. (Fifth Edition). 1991. 82 pp. (French, Spanish,
Russian)
9 Annex I (Superseded by IOC Manuals and Guides No. 17)
9 Annex II Guide for Responsible National Oceanographic Data Centres. 1982. 29 pp. (English, French, Spanish,
Russian)
11 The Determination of Petroleum Hydrocarbons in Sediments. 1982. 38 pp. (French, Spanish, Russian)
12 Chemical Methods for Use in Marine Environment Monitoring. 1983. 53 pp. (English)
13 Manual for Monitoring Oil and Dissolved/Dispersed Petroleum Hydrocarbons in Marine Waters and on
Beaches. 1984. 35 pp. (English, French, Spanish, Russian)
14 Manual on Sea-Level Measurements and Interpretation. (English, French, Spanish, Russian)
Vol. I: Basic Procedure. 1985. 83 pp. (English)
Vol. II: Emerging Technologies. 1994. 72 pp. (English)
Vol. III: Reappraisals and Recommendations as of the year 2000. 2002. 55 pp. (English)
Vol. IV: An Update to 2006. 2006. 78 pp. (English)
Vol. V: Radar Gauges. 2016. 100 pp. and Supplement: Practical Experiences. 100 pp. (English,
French,, Spanish)
15 Operational Procedures for Sampling the Sea-Surface Microlayer. 1985. 15 pp. (English)
16 Marine Environmental Data Information Referral Catalogue. Third Edition. 1993. 157 pp. (Composite
English/French/Spanish/Russian)
17 GF3: A General Formatting System for Geo-referenced Data
Vol. 1: Introductory Guide to the GF3 Formatting System. 1993. 35 pp. (English, French, Spanish,
Russian)
Vol. 2: Technical Description of the GF3 Format and Code Tables. 1987. 111 pp. (English, French,
Spanish, Russian)
Vol. 3: Standard Subsets of GF3. 1996. 67 pp. (English)
Vol. 4: User Guide to the GF3-Proc Software. 1989. 23 pp. (English, French, Spanish, Russian)
No. Title
Vol. 5: Reference Manual for the GF3-Proc Software. 1992. 67 pp. (English, French, Spanish, Russian)
Vol. 6: Quick Reference Sheets for GF3 and GF3-Proc. 1989. 22 pp. (English, French, Spanish, Russian)
18 User Guide for the Exchange of Measured Wave Data. 1987. 81 pp. (English, French, Spanish,
Russian)
19 Guide to IGOSS Specialized Oceanographic Centres (SOCs). 1988. 17 pp. (English, French, Spanish,
Russian)
20 Guide to Drifting Data Buoys. 1988. 71 pp. (English, French, Spanish, Russian)
22 rev. GTSPP Real-time Quality Control Manual, First revised edition. 2010. 145 pp. (English)
23 Marine Information Centre Development: An Introductory Manual. 1991. 32 pp. (English, French,
Spanish, Russian)
24 Guide to Satellite Remote Sensing of the Marine Environment. 1992. 178 pp. (English)
25 Standard and Reference Materials for Marine Science. Revised Edition. 1993. 577 pp. (English)
26 Manual of Quality Control Procedures for Validation of Oceanographic Data. 1993. 436 pp. (English)
27 Chlorinated Biphenyls in Open Ocean Waters: Sampling, Extraction, Clean-up and Instrumental
Determination. 1993. 36 pp. (English)
28 Nutrient Analysis in Tropical Marine Waters. 1993. 24 pp. (English)
29 Protocols for the Joint Global Ocean Flux Study (JGOFS) Core Measurements. 1994. 178 pp . (English)
30 MIM Publication Series:
Vol. 1: Report on Diagnostic Procedures and a Definition of Minimum Requirements for Providing
Information Services on a National and/or Regional Level. 1994. 6 pp. (English)
Vol. 2: Information Networking: The Development of National or Regional Scientific Information
Exchange. 1994. 22 pp. (English)
Vol. 3: Standard Directory Record Structure for Organizations, Individuals and their Research Interests.
1994. 33 pp. (English)
31 HAB Publication Series:
Vol. 1: Amnesic Shellfish Poisoning. 1995. 18 pp. (English)
32 Oceanographic Survey Techniques and Living Resources Assessment Methods. 1996. 34 pp. (English)
33 Manual on Harmful Marine Microalgae. 1995. (English) [superseded by a sale publication in 2003, 92-3-
103871-0. UNESCO Publishing]
34 Environmental Design and Analysis in Marine Environmental Sampling. 1996. 86 pp. (English)
35 IUGG/IOC Time Project. Numerical Method of Tsunami Simulation with the Leap-Frog Scheme. 1997.
122 pp. (English)
36 Methodological Guide to Integrated Coastal Zone Management. 1997. 47 pp. (French, English)
nd
37 International Tsunami Survey Team (ITST) Post-Tsunami Survey Field Guide. 2 Edition. 2014. 120 pp.
(English)
38 Guidelines for Vulnerability Mapping of Coastal Zones in the Indian Ocean. 2000. 40 pp. (French,
English)
39 Manual on Aquatic Cyanobacteria – A photo guide and a synopsis of their toxicology. 2006. 106 pp.
(English)
40 Guidelines for the Study of Shoreline Change in the Western Indian Ocean Region. 2000. 73 pp.
(English)
No. Title
41 Potentially Harmful Marine Microalgae of the Western Indian Ocean

Microalgues potentiellement nuisibles de l'océan Indien occidental. 2001. 104 pp. (English/French)
42 Des outils et des hommes pour une gestion intégrée des zones côtières - Guide méthodologique, vol.II/
Steps and Tools Towards Integrated Coastal Area Management – Methodological Guide, Vol. II. 2001.
64 pp. (French, English;; Spanish)
43 Black Sea Data Management Guide (Cancelled)

44 Submarine Groundwater Discharge in Coastal Areas – Management implications, measurements and
effects. 2004. 35 pp. (English)
45 A Reference Guide on the Use of Indicators for Integrated Coastal Management. 2003. 127 pp.
(English). ICAM Dossier No. 1
46 A Handbook for Measuring the Progress and Outcomes of Integrated Coastal and Ocean Management.
2006. iv + 215 pp. (English). ICAM Dossier No. 2
47 TsunamiTeacher – An information and resource toolkit building capacity to respond to tsunamis and
mitigate their effects. 2006. DVD (English, Bahasa Indonesia, Bangladesh Bangla, French, Spanish,
and Thai)
48 Visions for a Sea Change. Report of the first international workshop on marine spatial planning. 2007.
83 pp. (English). ICAM Dossier No. 4
49 Tsunami preparedness. Information guide for disaster planners. 2008. (English, French, Spanish)
50 Hazard Awareness and Risk Mitigation in Integrated Coastal Area Management. 2009. 141 pp.
(English). ICAM Dossier No. 5
51 IOC Strategic Plan for Oceanographic Data and Information Management (2008–2011). 2008. 46 pp.
(English)
52 Tsunami risk assessment and mitigation for the Indian Ocean;; knowing your tsunami risk – and what to
do about it. 2009. 82 pp. (English)
53 Marine Spatial Planning. A Step-by-step Approach. 2009. 96 pp. (English;; Spanish). ICAM Dossier No.
6
54 Ocean Data Standards Series:

Vol. 1: Recommendation to Adopt ISO 3166-1 and 3166-3 Country Codes as the Standard for
Identifying Countries in Oceanographic Data Exchange. 2010. 13 pp. (English)
Vol. 2: Recommendation to adopt ISO 8601:2004 as the standard for the representation of date and
time in oceanographic data exchange. 2011. 17 pp. (English)
55 Microscopic and Molecular Methods for Quantitative Phytoplankton Analysis. 2010. 114 pp. (English)
56 The International Thermodynamic Equation of Seawater—2010: Calculation and Use of
Thermodynamic Properties. 2010. 190 pp. (English)
57 Reducing and managing the risk of tsunamis. Guidance for National Civil Protection Agencies and
Disaster Management Offices as Part of the Tsunami Early Warning and Mitigation System in the North-
eastern Atlantic, the Mediterranean and Connected Seas Region – NEAMTWS. 2011. 74 pp. (English)
58 How to Plan, Conduct, and Evaluate Tsunami Exercises / Directrices para planificar, realizar y evaluar
ejercicios sobre tsunamis. 2012. 88 pp. (English, Spanish)
59 Guide for designing and implementing a plan to monitor toxin-producing microalgae. Second Edition.
2016. 63 pp. (English, Spanish)
st
60 Global Temperature and Salinity Profile Programme (GTSPP) — Data user's manual, 1 Edition 2012.
61 Coastal Management Approaches for Sea-level related Hazards: Case-studies and Good Practices.
No. Title
62 Guide sur les options d’adaptation en zone côtières à l’attention des décideurs locaux – Aide à la prise
de décision pour faire face aux changements côtiers en Afrique de l’Ouest / A Guide on adaptation
options for local decision-makers: guidance for decision making to cope with coastal changes in West
Africa / Guia de opções de adaptação a atenção dos decisores locais: guia para tomada de decisões de
forma a lidar com as mudanças costeiras na Africa Ocidental. 2012. 52 pp. (French, English,
Portuguese). ICAM Dossier No. 7.
63 The IHO-IOC General Bathymetric Chart of the Oceans (GEBCO) Cook Book. 2012. 221 pp. (English).
Also IHO Publication B-11
64 Ocean Data Publication Cookbook. 2013. 41 pp. (English)

65 Tsunami Preparedness Civil Protection: Good Practices Guide. 2013. 57 pp. (English)
66 IOC Strategic Plan for Oceanographic data and Information Management (2013-2016). 2013. 54 pp.
(English/French/Spanish/Russian)
67 IODE Quality Management Framework for National Oceanographic Data Centres (in preparation)
68 An Inventory of Toxic and Harmful Microalgae of the World Ocean (in preparation)
69 A Guide to Tsunamis for Hotels: Tsunami Evacuation Procedures (in preparation)
70 A guide to evaluating marine spatial plans. 2014. 96 pp. (English)
71 IOC Communication Strategy for Marine Information Management (2015-2017). 2015
72 How to reduce coastal hazard risk in your community – A step-by-step approach. 2016
73 Guidelines for a Data Management Plan. 2016
74 Tsunami Ready Guidelines for the Caribbean and Adjacent Regions. 2016. (English/French/Spanish)
75 ICAN (International Coastal Atlas Network) - best practice guide to engage your CWA (Coastal Web
Atlas) user community. 2016
76 Plans and Procedures for Tsunami Warning and Emergency Management – Guidance for countries in
strengthening tsunami warning and emergency response through the development of Plans and
Standard Operating Procedures for their warning and emergency management authorities. 2017
77 IOC Strategic Plan for Data and Information Management (2017-2021). 2017
78 Harmful Algal Blooms (HABs) and Desalination: A Guide to Impacts, Monitoring and Management. 2017

Arid countries throughout the world are heavily reliant on seawater desalina-
tion for drinking and industrial process water. With nearly 20,000 plants oper-
ating globally and capacity projected to rise 12% per year, the industry is
large and growing. A major operational challenge facing the industry is also
expanding globally – the phenomena termed harmful algal blooms or HABs.
Algal “blooms” are cell proliferations caused by the growth and accumulation
of individual species; they occur in virtually all bodies of water. Cells can
reach concentrations sufficient to make the seawater appear red (hence the
common term “red tide”), though other colors are also observed, and many
HABs are invisible.
Of the thousands of algal species, most are beneficial to humans and the
environment, but some cause harm due to either their potent toxins or the
copious quantities of dissolved and particulate organic matter they generate.
Marine toxins pose a risk to both thermal and seawater reverse osmosis
(SWRO) plants while the organics typically impacts only SWRO plants.
Impacts of HABs are a significant issue in desalination, exacerbated by the
lack of knowledge of these phenomena among plant operators, engineers,
and others in the industry, including regulatory agencies. This Manual was
prepared to provide practical information about HABs, their toxins, biomass,
and extracellular products, monitoring approaches inside and outside plants,
treatment technologies, and risk assessment strategies. Case studies are
presented describing HAB events at desalination plants throughout the
world, detailing impacts and the strategies used to combat them. Their expe-
riences and lessons learned can be beneficial to others encountering similar
challenges. This Manual, with its practical areas of guidance and multidisci-
plinary approach, should be of great value to many in the industry.
medrc.org usaid.gov ioc.unesco.org

Harmful Algal Blooms (Habs) and Desalination: A Guide To Impacts, Monitoring, and Management

Uploaded by

Copyright:

Available Formats

Harmful Algal Blooms (Habs) and Desalination: A Guide To Impacts, Monitoring, and Management

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Harmful Algal Blooms (Habs) and Desalination: A Guide To Impacts, Monitoring, and Management

Uploaded by

Copyright:

Available Formats

Manuals and Guides 78

Harmful Algal Blooms (HABs)

Harmful Algal Blooms (HABs) and Desalination: A

*Corresponding Author’s email: danderson@whoi.edu

First published in 2017 by the United Nations Educational,

Appendix 1. Algal species potentially harmful to desalination operations

Appendix 2. Rapid screening methods for harmful algal bloom toxins

Appendix 3. Methods for measuring transparent exopolymer particles and

Appendix 4. Preservatives and methods for algal cell enumeration

Appendix 5. Autopsy and cleaning of reverse osmosis elements affected by harmful

Abdullah Said Al-Sadi

1! HARMFUL ALGAL BLOOMS

1.1! ALGAL BLOOMS

Table 1.1. Human illnesses associated with HABs.

Paralytic Alexandrium species, Saxitoxins Eating shellfish Acute paresthesias and

Neurotoxic Karenia brevis and Brevetoxins Eating shellfish Gastrointestinal and

Diarrhetic Dinophysis species; Okadaic acid Eating shellfish Acute gastroenteritis

Azaspiracid Azadinium spinosum Azaspiracids Eating shellfish Neurotoxic effects with

Amnesic Pseudo-nitzschia Domoic acid Eating shellfish (or, Gastroenteritis,

some of these plants and nearshore

Table 1.2. (Continued)

Microalgae Toxin(s) Poisoning References

Raphidophytes Brevetoxins Neurotoxic Shellfish Loeblich and Fine

Table 1.3. (Continued)

Dinophysis caudata Okadaic acid,

Dinophysis tripos Okadaic acid,

Gymnodinium Saxitoxin 101 10.1 50.5 505 Band-Schmidt et al. 2006

Karenia brevis Brevetoxin 49 4.9 24.5 245 Corcoran et al. 2014

Karlodinium veneficum Karlotoxins 1.34 0.134 0.67 6.7 Fu et al. 2010

Ostreopsis siamensis Palytoxins 16 1.6 8 80 Suzuki et al. 2012

Pseudo-nitzschia Domoic acid 37 3.7 18.5 185 Garrison et al. 1992

Pyrodinium bahamense Saxitoxin 120 12 60 600 Usup et. al. 1994

areas (e.g., Thangaraja et al. 2007;

1.5.1! Scale of blooms

1.5.5! Bloom transport

Band-Schmidt, C., Bustillos-Guzmán, J., Morquecho, L., Gárate-Lizárraga, I., Alonso

Corcoran, A. A., Richardson, B., and Flewelling, L. J. 2014. Effects of nutrient-limiting

2 ALGAL ISSUES IN SEAWATER DESALINATION

Philipp Hess1, Loreen O. Villacorte2,3, Mike B. Dixon4, Siobhan F.E. Boerlage5,

Natural organic matter in seawater is comprised of a diverse mixture of particulate, colloidal

2.2.3 Transparent Exopolymer Particles (TEP)

Pressure-driven capillary UF membranes have reportedly exhibited some degree of fouling

Figure 2.9. Homotropanes: anatoxin-a (ATX-a) and its

anatoxin-a FW bicyclic amine C10H15NO 165.1 hydrophilic fast acting Na channel

geosmin FW bicyclic alcohol C12H22O 182.3 lipophilic odor olfactive

saxitoxin FW + M alkaloid C10H17N7O4 299.1 hydrophilic fast acting Na channel

trichotoxin M chlorinated phenyl- C20H27CLO 318.2 lipophilic neurotoxin unknown

tetrodotoxin M alkaloid C11H17N3O8 319.1 hydrophilic neurotoxin Na channel

gymnodimine M cyclic imine, C32H45NO4 507.3 lipophilic fast acting Na channel

13desmethyl- M cyclic imine, C41H61NO7 691.4 lipophilic fast acting Na channel

pinnatoxin G M cyclic imine, C42H63NO7 693.5 lipophilic fast acting Na channel

okadaic acid M polyether C44H68O13 804.5 lipophilic diarrhetic PP2a

nodularin FW pentapeptide C41H60N8O10 824.4 lipophilic liver- PP2a

azaspiracid M polyether C47H71NO12 841.5 lipophilic diarrhetic unknown

brevetoxin-B M polyether C50H70O14 894.5 lipophilic diarrhetic Na channel

microcystin-LR FW + M heptapeptide C49H74N10O12 994.5 lipophilic liver- PP2a

ciguatoxin M polyether C60H85O16 1061.6 lipophilic diarrhetic Na

maitotoxin M polyether C164H258O68S2 3380 amphiphilic neurotoxin Ca-

Although the algal