Environment International 119 (2018) 220–231

Contents lists available at ScienceDirect

Environment International
journal homepage: www.elsevier.com/locate/envint

Review article

Synthesizing water quality indicators from standardized geospatial T

information to remedy water security challenges: A review
⁎
Vincent de Paul Obade , Richard Moore
The Ohio State University, School of Environment and Natural Resources, 2021 Coffey Road, Columbus, OH, United States of America

A R T I C LE I N FO A B S T R A C T

Handling Editor: Robert Letcher Water is vital not only for food, energy and sanitation but also for ecosystem functioning, human health, socio-
Keywords: economic progress and poverty reduction. Water security exists when all people have physical and economical
Fresh water access to sufficient, safe, and clean water that meets basic needs. However, water security is threatened by
Indicator growing human population, episodic environmental disasters, indiscriminate land management practices, con-
Water security taminants, and escalation in geopolitical conflicts. < 3% of the estimated 1.4 billion cubic kilometers of water on
Water quality earth is available for consumption. Although there exist a range of laboratory and field methods for measuring
the chemical, physical and biological properties of water, the information available to the public remains in-
consistent and patchy. To this end, we advance a new theory of a single-value objective water quality index
(WQI) that considers the interaction between the above properties, to provide concise information for source
water quality surveillance and monitoring. Although geospatial technologies such as remote sensing is credited
as a high frequency spatiotemporal mapping tool, exiguous information is available on its application for con-
structing single-value WQIs. Besides, no remote sensing device exists that directly measures water quality, which
must indirectly be inferred through modeling sensed remote sensing signals with measured water properties.
This review not only highlights the water security conundrum but also provides an overview of methods for
integrating geolocated qualitative (e.g., management data) with quantitative (i.e., measured water constituent
properties) into a WQI.

1. Introduction conclusions, poor regulatory enforcement, health risks, or even devalue

environmental concerns (EPA, 2016b; Reif, 2011). To avoid cata-
Other than climate change and the increasing population estimated strophic health ailments, diagnostic screening tools are required to
at 80 million per year, global challenges include indiscriminate land ensure drinking water quality satisfies international standards, speci-
use, rising economic disparities and anticipated deficiencies in food, fied by World Health Organization (WHO), American Public Health
energy and water (Esteban and Max, 2016; Gleick and Palaniappan, Association (APHA), or Environmental Protection Agency (EPA),
2010; Lal, 2015b; UNU, 2013). Although water is one of the crucial among others (EPA, 2016a; Kumar and Puri, 2012; WHO, 2011).
resources for life on earth, defining a generic water quality standard Water pollutants are categorized as point source (PS) and non-point
that satisfies all water uses is challenging. This is because standards source (NPS). PS pollutants are anthropogenic contaminants discharged
considered suitable for human consumption are different from those for via a discrete conveyance thus are traceable to single source (e.g., pipes
industrial or even agricultural use (Haji Gholizadeh et al., 2016; Ritchie discharging effluent of industrial or domestic wastewater). In contrast,
et al., 2003). Despite the diverse data processing platforms, water NPS have diffuse origins, and are facilitated by: (a) infrastructure
quality determination at field, watershed and even landscape scale is dysfunction (open/leaky sewer systems, landfill leakages, impervious
challenging because of spatiotemporal variation in the physical, che- urban surfaces), (b) land mismanagement (e.g., broadcasting fertilizer
mical and biological water constituents (Table 1). Moreover, gleaning on soil surface, flood irrigation, agricultural and livestock waste, soil
water quality information for routine management operations can be erosion), meteorological conditions (i.e., precipitation intensity, tem-
impeded by data scarcity and artifacts, nebulous baselines, and vali- perature and wind speed), (c) hydromodification and atmospheric de-
dation challenges (de Paul Obade et al., 2013, 2014). Indeed, in- position of industrial pollutants (Chipman et al., 2009; Kozlowski et al.,
coherence in water quality information can contribute to feigned 2016; Michalak et al., 2013; Selman et al., 2009; Swanson et al., 2015).

⁎
Corresponding author.
E-mail addresses: obadev@yahoo.co.uk (V. de Paul Obade), moore.11@osu.edu (R. Moore).

https://doi.org/10.1016/j.envint.2018.06.026
Received 11 May 2018; Received in revised form 18 June 2018; Accepted 20 June 2018
0160-4120/ © 2018 Elsevier Ltd. All rights reserved.
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

Table 1
Overview of parameters for water quality surveillance and monitoring.
Physical Chemical Biological

Temperature Chlorophyll Benthic macroinvertebrates:

-Water temperature determines the amount of oxygen -Chlorophyll pigment is useful for photosynthesis, -Macroinvertebrates are organisms large (macro) enough to
dissolved in the water, the photosynthesis rate of whereby green plants and algae convert sunlight be seen through the naked eye; and lack a backbone
aquatic plants and metabolism of aquatic organisms. energy into chemical energy by taking in CO2, H2O (invertebrate). Benthic refers to the bottom of a waterway.
and producing carbohydrates and O2. Chlorophyll Examples of benthic macroinvertebrates include insects in
-Causes of temperature change include weather, measures plant and algae pigments, and can be used as their larval or nymph form, crayfish, clams, snails, and
removal of shading streambank vegetation, proxy biomass estimate. Chlorophyll absorbs blue and worms. Most are attached to submerged rocks, logs, and
impoundments, discharge of cooling water, urban red light and reflects green (thus healthy plants appear vegetation.
storm water, and groundwater inflows to the stream. green).
-Some macro invertebrates are more sensitive to pollution
-Algal blooms make water unsuitable for swimming, than others.
toxic and unpalatable to the aquatic food chain.
Besides, unconsumed algae sink and decay, depleting
deeper water of oxygen.
pH Suspended solids/minerals
-pH indicates the alkalinity or acidity of a substance, -Suspended minerals or sediments move along in a
ranked on a scale from 1 to 14. pH < 7 is acidic stream, thus are dependent on water flow and rainfall.
whereas pH > 7 is alkaline. A pH of 7.0 is neutral. -sediments obstruct light and may harbor pathogens.
-suspended particles/algae pigments affect how
-Aquatic organisms differ as to the range of pH in ambient light is absorbed and reflected.
which they flourish.
Streamflow Colored Dissolved Organic Carbon contains fulvic Submerged Aquatic Vegetation: Submerged aquatic
-Streamflow, or discharge, is the volume of water that or humic acid, which makes water-bodies to have vegetation (SAV) provides invaluable benefits to aquatic
moves over a designated point over a fixed period of brownish tan color. ecosystems. It not only provides food and shelter to fish and
time. Turbidity estimates particulate matter suspended in invertebrates but also produces oxygen, traps sediment and
-It is affected by weather as it increases during water. Water with high turbidity is cloudy or opaque. absorbs nutrients such as nitrogen and phosphorus.
rainstorms and decreases during dry periods; and High turbidity increases water temperatures because Whereas SAV are dependent upon the transmission of
varies by season. suspended particles absorb heat, and reduce sunlight through the water, the location of individual
penetration of light into water. Murky water is unsafe species depends upon a variety of factors such as salinity,
for recreational purposes because hazardous depth and bottom sediment.
materials/objects are obscured.
Secchi Disk Transparency measures water clarity. -plankton include bacteria, archaea, algae, protozoa and
This measurement is done by lowering a black and drifting or floating animals that inhabit oceans, seas, lakes,
white disk into the water and recording the depth at ponds
which the disk is invisible. Clear water signifies
environmental health.
Dissolved Oxygen: Respiration of most aquatic
organisms requires Oxygen dissolved in water.
Rapidly moving water dissolves more Oxygen than
stagnant water. Colder water dissolves more oxygen
than warmer water.
-An oxygen-deficient aquatic environment results in
water bodies with excess organic material, which
cause death to aquatic life.

Phosphates which are largely lost through runoff and erosion are major locations; is rapid, non-destructive, reproducible, durable and provides
NPS pollutants; compared with nitrates which being soluble are leached both analog and digital data to support automated processing. How-
and recyclable depending on climate and bioactivity (Causse et al., ever, no remote sensing device exists that directly measures water
2015; de Paul Obade et al., 2013). In essence, high quality soils are quality, which can instead be indirectly inferred by modeling sensed
natural buffers against contaminants (Adhikari and Hartemink, 2016; remote sensing signals with measured water properties. Because, water
Lal, 2018). For brevity, soil organic Carbon (SOC), a proxy of soil management decisions determine ecosystem health and productivity,
quality plays a key role in: (i) water purification and retention, thereby other reviewed priorities include: (i) evaluating consequences of an-
preventing indiscriminately managed nutrients (e.g., nitrogen and thropogenic activities on water security, (ii) which strategies ensure
phosphorus) from contaminating surface or ground water, (ii) food, sustainable freshwater ecosystems? and (iii) what tools, models, and
fuel, and fibre production, (iii) biodiversity conservation, (iv) climate information are required for cost-effective monitoring of water quality?
regulation, and (v) nutrient cycling (de Paul Obade, 2017)
Conceptually, a water quality index (WQI) synthesizes complex 2. Risks and impacts of water insecurity
water characteristic properties into more interpretable and informative
format to support decision-making (Chaturvedi and Bassin, 2010; Khan Although 71% of the earth's total surface area (51 × 107 km2) is
et al., 2003; Liou et al., 2004; Mohebbi et al., 2013). In essence, WQIs water, only 2.5% (3.5 million cubic kilometers out of a total 1.4 bil-
should integrate water quality parameters using techniques that: (i) are lion cubic kilometers) is renewable freshwater (Lal, 2015a). Approxi-
simple to develop and use, (ii) correlate well with water constituent mately 70% of this freshwater is contained in ice caps, glaciers, per-
properties, and (iii) are replicable and accurate at variable scales manent snow, ground ice, permafrost, or ground water, and only 1.2%
(Cobbina et al., 2010; House, 1990; Mohebbi et al., 2013). Although is available for direct consumption by living organisms (Gleick and
guidelines exist for water quality determination (Ashbolt et al., 2001; Palaniappan, 2010). Consumptive water refers to water unavailable for
WHO, 2011), no universal comprehensive single-value WQI has been use after being evapotranspired, ground infiltrated or incorporated into
conclusively determined. Here, the feasibility of integrating field and plant or animal tissue, whereas, non-consumptive returns to surface
remote sensing data to generate a concise objective WQI is explored. runoff and is reusable after treatment. Blue water refers to fresh surface
Remote sensing technologies acquire data even from inaccessible field or ground water and includes precipitation, whereas green water is the

221
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

soil water, which plants uptake and transpire. “Gray” or contaminated 2016).
water originates from domestic use, urban and industrial discharge; As realized from the extreme lead (Pb) levels in Flint, Michigan
whereas virtual water encompasses water consumed in production of (USA) in 2016 (Ingraham, 2016), and Toledo, Ohio (USA) in 2010 and
commodities (e.g., agricultural or industrial) and traded across inter- 2014 (Lindstrom, 2016; Stewart et al., 2014), no guarantees exist for
national boundaries (DIA, 2012; Lal, 2015b). clean water. Intake of water with elevated Pb levels can cause anemia,
Water scarcity occurs when human demand or consumption exceeds severe mental and physical impairment especially in children. Pb has
supply to the extent that per capita availability of renewable freshwater been used for ages in plumbing fittings and water distribution systems,
is < 1000 m3/person/year, whereas water stress or extreme lack of however, because of Pb poisoning, copper has replaced Pb as a safer
water to satisfy human or ecological demands occurs when water alternative. Alternately, chromium-6, a raw material for stainless steel
supply < 1700 m3/person/year (DIA, 2012; Lal, 2015b; Rijsberman, production, was found in approximately 75% of water samples in U.S.A.
2006). Globally, an estimated 2 billion people live under squalid con- between 2013 and 2015, and may have been consumed by an estimated
ditions in areas of high water stress with limited sanitary facilities, 200 million people (Zaremba, 2016). Chromium-6 pollution causes
insufficient clean water, and unreliable energy access (WHO, 2008; liver damage, reproductive health problems and cancer. Although
Water and United Nations, 2008; WHO/UNICEF, 2004). Producing permissible limit for Pb in treated drinking water is documented as 15
energy requires water, for instance, thermoelectric power expends parts per billion (ppb); that for Chromium-6 is unknown (Scott, 2016;
roughly 40% of freshwater in the U.S. (i.e., 160 billion cubic meters). Zaremba, 2016). Other adverse ramifications of pollution include soil
Yet, water purification and desalinization consumes energy and emits salinization, and outbreak of water-borne diseases (Rijsberman, 2006;
CO2(g) (FAO, 2008; Rijsberman, 2006; UNDESA, 2014). Reduction in Semenza et al., 2012; Vincent et al., 2004). In Kenya, almost half of the
soil-water storage or low soil-water availability at critical crop growth population (~16 million) resides in unhygienic sanitary conditions;
stages can result in pedological, or agronomical drought respectively. which has contributed to disease outbreaks, soaring infant mortality
Conversely, ecological drought refers to low water availability after rate, and lower graduation rates among children (http://water.org/
land-use conversion, whereas sociological drought occurs when human country/kenya/). Arsenic, detected in dangerous levels in some
water demand exceeds supply. Meteorological and hydrological drinking water supplies in India (i.e., Ganges delta) and rural China, is
droughts are long-term deficiencies in precipitation or water flow in carcinogenic and genotoxic in high concentration, and its intake can
reservoirs, respectively (Lal, 2013). Imperfections in climate models cause acute abdominal pain, vomiting, diarrhea, muscular pain, and
may explain the sometimes blurred and unreliable information on skin cancer (Komorowicz and Barałkiewicz, 2016; Kumar and Puri,
drought phenomenon (Awange et al., 2016; Cutter et al., 2012). Against 2012). Above all, the arsenic concentration in drinking water should
the backdrop of abrupt climate change and increasing water footprint, not exceed 10 μg L−1 (WHO, 2011).
over 50% of the global population will lack clean water, and about 700 Elevated nutrient concentration (i.e., Total Nitrogen (TN) or Total
million people risk water instigated displacement by 2050 (FAO, 2008; Phosphorus (TP)) in water reservoirs trigger blossoming of cyano-
Rijsberman, 2006; UNDESA, 2014). bacteria such as Anabaena sp., Planktothrix or Microcystis sp. that cause
Societal vulnerability to disasters depends on settlement patterns eutrophication. Eutrophication stresses aquatic ecosystems, reduces
and population density, economic status, precipitation intensity and water aesthetics impacting on tourism industry, threatens drinking
land use patterns (Gleick, 2014; Swanson et al., 2015). These disasters water supplies, produces a bad odor, and clogs reservoirs (de Paul
can have devastating consequences to human life, environment and Obade et al., 2013, 2014; Vincent et al., 2004). According to the World
even slow down economic progress (Vorosmarty et al., 2000). For in- Health Organization, the maximum threshold for phytoplankton mi-
stance, flooding destroy infrastructure, displaces people, submerge crocystin toxin in recreational waters is 20 μg/L (Michalak et al., 2013).
homes, inundate farmlands, and cause recurring humanitarian crisis in Water contaminated with toxic cyanobacterium can cause ailments
the states of Colorado, Missouri, Illinois, Tennessee, Arkansas, Mis- such as blue baby syndrome (methaemoglobinaemia), liver cancer,
sissippi, and Louisiana, USA (Coffman and Dobuzinskis, 2013). Not only nausea, vomiting, respiratory illnesses or even death (Chang et al.,
is the precise environmental footprint of the 2010 gulf of Mexico oil 2015; Wright, 2016). Extreme cases of “hypoxia” or “dead zones” at-
spill unknown, but also the efficacy of remedial measures; especially tributed to oxygen depletion have reportedly occurred in the Black Sea
after spending over US $ 12 billion on the restoration efforts (Haq, (Eastern Europe), Lake Taihu and Pearl River Delta (China), Lake
2010). In Africa, disastrous floods recur in the Limpopo basin of Mo- Winnipeg (Canada), Lake Erie, the Gulf of Mexico, and Chesapeake Bay
zambique (Spaliviero et al., 2014), and river Nyando in Kenya (IFRC, (U.S.) (Selman et al., 2009). Elsewhere, the water hyacinth (Eichhornia
2016). Minimizing disaster impacts requires proactive strategies that crassipes) invaded Lake Victoria, the largest lake in East Africa with over
include: (a) creating socio economic safety nets, (b) restoration plans 30 million people in its vicinity, blocking fishing access and providing
backed by scientifically credible data (Bouma and McBratney, 2013; breeding grounds for disease carrying mosquitoes and snails (le Roux
Lal, 2009a, 2009b; Power, 2010). Fig. 1 depicts the interconnection et al., 2016; UNEP, 2013).
between water quality, society and environment, based on the Driving Geopolitical tensions are simmering between Ethiopia and Egypt
Forces-Pressure-State-Impact-Response (DPSIR) framework (EEA, because of the grand renaissance dam, located in the Blue Nile river
1997). DPSIR gauges effectiveness of strategies for tackling environ- (Khaled et al., 2016). This dam built at an estimated cost of U.S. $4
mental challenges (Gari et al., 2015; Niemeijer and de Groot, 2008). billion has a capacity to retain about 70 billion cubic meters of water,
Demographic explosion and overcrowded settlements, resource over useful for irrigation, flood control, and generating 6000 MW of elec-
use, poor waste disposal, leaching of pollutants, flooding, dysfunction tricity. However, apart from Egypt risking the loss of 60% of its crop-
of infrastructure, and rapid groundwater depletion adversely affect lands due to water diversion, the environmental consequences could be
water quality and availability (FAO, 2008, 2013; Laurent and Ruelland, catastrophic (Chellaney, 2013; Khaled et al., 2016). Alternately, Tur-
2011; Mueller-Warrant et al., 2012; Ramadas and Samantaray, 2018; key's numerous dam construction along the Tigris and Euphrates river
Rijsberman, 2006). Due to water withdrawal for irrigation exceeding basin for military purposes, hydroelectric power and irrigation, has
average annual recharge, the groundwater levels at Ogallala aquifer reduced water flow and created rifts with neighboring Iraq and Syria
declined by 30 cm/year between 1996 and 2011 (i.e., totaling 4.3 m); (Gleick, 2014). Syria accuses Turkey of socio-economic sabotage and
which doubled to 60 cm/year after the 2012 drought (Kisekka et al., creating a humanitarian crisis by contaminating water with fecal
2017; Lal et al., 2012; NBC, 2016). Meanwhile, the risk of leakages from matter, a scenario that enhances risk of waterborne diseases, lowers
the $3.8 billion Midwestern U.S. oil pipeline traversing Missouri River yields of irrigated fruits and vegetables, and interferes with aquatic
not only threatens biodiversity and environmental health but also the ecosystems (Chellaney, 2013). Drought may have ignited the ongoing
sanctity of sacred sites belonging to Native American Indians (Levine, civil war in Syria, that has created a humanitarian crisis and displaced

222
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

.
State

.
Pressures

.
point (PS) & non-point source (NPS) pollutants

.
food, water, energy security

.
contaminated surface and ground water

.
ecosystem functions

.
peace and tranquility threatened
impaired habitats
infrastructure and industry issues

Responses (to sustain water quality)

.
1. Research:
quantitatively demystify enigmas (e.g., sources and fate of contaminants, land

.
management vis–à–vis water quality)
new technology (e.g., integration of sensor technology) and metrics (WQI)
2. tighter legislation (e.g., on water quality trading), riparian/wetland management
3. comprehensive solutions (moral and technical)

.
Impacts Driving Forces (i.e., influencing water quality)

. .
pathogens, epidemics Natural

. .
environmental & socio-economic costs weather (precipitation (dry/wet), temperature)

.
elemental water and energy cycles soil type/texture, topography
ecosystem processes (e.g., leaching)

.
Anthropogenic (land use and management conflicts)
agricultural practices (i.e., excess irrigation,
pesticide and fertilizer use, monoculture, intensive

.
grazing, subsurface tile drains, runoff)
urban sources (impervious surface, sewage

.
overflows/leaks, open landfills/waste dumps)
atmospheric deposition (e.g., urban traffic air
pollution deposited by rainfall)

Fig. 1. Driving Forces-Pressures-State-Impact-Responses (DPSIR) water quality infrastructure. Abbreviation WQI stands for Water Quality Indicator.

over 2 million people (Gleick, 2014). not objectively integrate qualitative and quantitative water quality
constituent variables, and thus is non-comprehensive and unreliable. In
this regard, the question remains how to blend data on water properties
3. Remediation options overview into a credible and robust water quality metric? Otherwise, WQT is a
market strategy designed to reduce water pollution through issuing
Despite existing legislation, the complexity of natural systems and permits and providing economic incentives to support ecosystem
their interactions with anthropogenic activities complicate prioritiza- management and restoration initiatives.
tion of remediation strategies to control pollution. Curtailing water Approximately half of water reservoirs within USA do not meet
pollution requires understanding of key driving forces influencing minimum water quality standards (EPA, 2000). Purifying polluted
water quality, strict enforcement mechanisms, and proactive manage- water resources in USA is estimated to cost millions of dollars each year
ment strategies. Examples of feasible management options include the (de Paul Obade et al., 2013; Ritchie et al., 2003). The Clean Water Act
Proper Functioning Condition (PFC) assessment method and water (CWA) was enacted in 1972 to: (i) control water effluent standards by
quality trading (WQT). PFC method evaluates the hydrology, vegeta- establishing a pollution cap based on mandated Total Maximum Daily
tion and erosion/deposition within riparian areas thereby providing Loads (TMDLs), (ii) ensure the water reservoirs are fishable and
critical information for understanding the drivers of ecosystem function swimmable (EPA, 2016b; Freeman, 2000; Ghosh et al., 2011). Based on
(Kozlowski et al., 2016; Swanson et al., 2015). However, PFC is a the sum of all pollutants, TMDLs provide permissible pollution limits
subjective approach because the assessment criteria relies on expert allowable in water reservoirs. Under section 303 (d) of CWA, all im-
opinion and visual interpretation of field sites either directly, or by paired waters within respective watersheds must be listed (EPA,
using raw aerial photographs and satellite imagery. Besides, PFC does

223
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

2016b). Although CWA is credited for significantly controlling PS water time-consuming, labor intensive process, and thus expensive (Ramesh
pollutants, NPS pollution remains a problem. et al., 2010; Ritchie et al., 2003).
As a strategy to control pollution, the U.S. federal CWA requires all From a practical standpoint, laboratory analyzed field data and
facilities that discharge wastewater, or storm water within specific geospatial technology (e.g., remote sensing and Global Positioning
watersheds to purchase a WQT permit applicable within the defined Systems (GPS)) are integrated using a Geographical Information
jurisdiction. WQT allows point polluters (e.g., industries) to opt out Systems (GIS) (Fig. 3) to provide spatially/temporal continuous data
from fully implementing pollution reduction technologies but instead that can be validated and complemented from the sampled field points
purchase “pollution credits” from NPS polluters (e.g., farmers) (de Paul (Gitelson et al., 2008; Haji Gholizadeh et al., 2016; Ritchie et al., 2003).
Obade et al., 2014; Moore, 2014). However, prior to participating in GIS are computer-based tools that integrate database operations and
WQT, U.S. Environmental Protection Agency (EPA) requires these NPS can query, statistically analyze, overlay, visualize, manage and store
polluters to incorporate best management practices (BMPs). Examples geographically referenced data. The GIS-based Soil and Water Assess-
of BMPs include crop rotation, water conservation, allocating a con- ment Tool (SWAT) driven largely by weather data approximates surface
servation easement, prescribed burning, grazing management plan, runoff, discharge, sediment and nutrient loads and can thus be used to
conservation agriculture, recycling waste, integrated nutrient and pest predict agricultural management impacts on water quality (Ha et al.,
management (Kozlowski et al., 2016; Lal, 2013; Lal et al., 2012; 2018; Michalak et al., 2013). Other hydrologic and water quality
Swanson et al., 2015). models include the ADAPT, ANNAGNPS, APEX, COUPMODEL,
WQT permits are issued by the National Pollutant Discharge CREAMS/GLEAMS, DRAINMOD, EPIC, HYDRUS, HSPF, InVEST, KI-
Elimination System (NPDES) a subsidiary of EPA, which enforces the NEROS2/AGWA, MACRO, MIKE SHE, MT3DMS, RZWQM2, SHAW,
CWA mandate (Ghosh et al., 2011; Moore, 2014; Stephenson and SWIM3, STANMOD, TOUGH, WARMF etc. (Adhikari and Hartemink,
Shabman, 2011). Approximately 1000 permits have been issued costing 2016; Kisekka et al., 2017; Yuan et al., 2015). Among the parameters
between US $ 400 to 2000. However, NPDES continues to impose integrated and analyzed by these models include hydraulic con-
stringent measures for WQT permits, for instance, by reducing TP limits ductivity, TN, TP, pesticide concentration, bacteria loading rate, eva-
to 1 mg/L, with considerations of lowering further to between 0.5 and potranspiration (ET), soil porosity and erodibility, soil bulk density (ρb),
0.75 mg/L, which is arduous to monitor (de Paul Obade et al., 2014; fecal coliform concentration, soil organic carbon (SOC), runoff curve
Moore, 2013). No wander, specifying spatial and temporal scales for number, rooting depth, crop yield, subsurface drainage flow etc.,
which WQT can be most effective remains a challenge. (Adhikari and Hartemink, 2016; Kisekka et al., 2017; Yuan et al., 2015).
To recapitulate, WQT credibility depends on: (i) enforcement of a However, none of these models explicitly generate “single-value” WQI.
pollution cap, (ii) BMP verification system, and (iii) restrictions cur- Table 2 reviews standardized WQIs that have been utilized with mixed
tailing pollution (de Paul Obade et al., 2014; EPA, 2016b; Savard, 2000; results (Mohebbi et al., 2013; Ramesh et al., 2010).
Shrestha et al., 2008). WQT can be a risky venture because legal lia-
bility is not transferrable for credits purchased, implying that the credit 4.1. Water quality surveillance and monitoring using remote sensing
buyer may be fined for exceeding TMDL cap stated in original permit
should the contract with the non-point polluter lapse (Selman et al., Determination of the spatial extent of water resources is not a dif-
2009). The Alpine and Muskingum WQT programs (Moore, 2014); and ficult task at a specific site; the problem is monitoring water quality.
the electrical power research institute's (EPRI), exemplify WQT in- This knowledge gap is a major impediment in the process of planning or
itiatives operational in Ohio, Kentucky and Indiana (USA) (EPRI, 2016), even informing water resource managers and decision makers on the
that can be emulated in other localities. socio-economic impacts of development projects. Thus, the potential of
technologies such as remote sensing which provide continuous spatial
4. Monitoring water quality strategies and temporal data needs to be fully explored to understand water
quality dynamics. Optical, thermal, active and passive remote sensing
Sustainable water quality management entails: (i) developing robust based systems that are either hand-held, or operated from boats, air-
water quality assessment models and tools, and (ii) utilizing technolo- craft, and satellites have transformed the paradigm of mapping by
gies that can promote water use efficiency, for instance, by increasing sensing beyond the visible electromagnetic spectrum and providing
green water, purifying gray water and minimizing virtual water (de repetitive, spatially continuous data in real-time that can be upscaled or
Paul Obade et al., 2013, 2014; Lal, 2015a; Maruthi Sridhar and Vincent, downscaled (de Paul Obade and Lal, 2013; Oliver and Webster, 2014;
2007; Mohammed, 2002). Besides, new monitoring capabilities pro- Ouma, 2016; Pérez Hoyos et al., 2016; Roy et al., 2008, 2014). Active
viding real time information are required to assess the risks and im- systems emit and detect own energy to and from the target, whereas
plications of anthropogenic land management practices on water re- passive systems rely on energy from the sun. Unlike passive sensors,
servoirs (Bushaw-Newton and Seller, 1999; Vincent et al., 2004). Many active sensors generate pulses that penetrate clouds and smoke. Al-
different statistical methods, automated or semi-automated technolo- though remote sensing systems provide frequent synoptic coverage, and
gies utilizing field data, and interpretations from aerial or satellite are non-destructive, its caveats include: (a) bidirectional reflectance
imagery, or both to assay water quality exist (Hajigholizadeh and distribution function (BRDF) effects and spectral mixing problem
Melesse, 2017). A conceptual flow chart portraying a holistic yet syn- caused by adjacency effects and uneven illumination, (b) data-gaps
thesized WQI is depicted in Fig. 2. attributed to cloud cover, smoke, sun-glint, aerosols, atmospheric in-
Water quality information is gleaned by a range of techniques terference; all of which lower the signal-to-noise ratio (SNR), (c) rig-
varying in complexity and sophistication. The most commonly used orous and continuously evolving data processing and calibration re-
methods regardless of limitations include: (i) in situ (local/point/field) quirements (e.g., changes in data quality and formats following the
measurements (e.g., Secchi disk depth (SDD) which evaluates the depth launching of newer satellites), (d) mismatches between spatial, spectral
of water transparency/clarity in reservoirs but may be limited for and temporal resolution, (e) high costs of data acquisition, archiving
flowing river systems), (ii) laboratory analyzed samples, and (iii) em- and absence of long-term data (Chang et al., 2015; de Paul Obade et al.,
pirical/analytical modeling of remotely sensed data from scanned lo- 2013). Haze in imagery attributed to variation in sensor slope angles or
cations or samples (Haji Gholizadeh et al., 2016; Maruthi Sridhar and directions can be minimized by ratio-ing sensor bands. Notwith-
Vincent, 2007; Olmanson et al., 2013; Reif, 2011). Although lacking in standing, the explosion of digital geo-information products create re-
spatial coverage and temporal extent, adequately collected field data dundant information (Ouma, 2016; Vitharana et al., 2008).
for a point is accurate, and thus used to validate data from other Although an overview is provided here on water quality sensors
sources, such as remote sensing. However, collecting field data is a (Tables 3, 4, & 5); the sensor specifications, other sundry details such as

224
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

GIS

.
field (insitu) data (e.g.,)
Remote Sensing (passive, active;

.
pigment samples extracted in hot ethanol data fusion
passive + active), Google Earth

.
phytoplankton chlorophyll quantified fluorometrically
Engine, GPS

.
turbidity determined turbidity sensor, secchi disc
dissolved/particulate organic carbon determined by high
temperature combustion

.
synthesize (e.g., single-value) water quality index

.
algorithm development
empirical, analytical & hybrid (e.g., structural equation models)

.
uncertainty / sensitivity analyses (e.g.,)

.
2
R , Root Mean Square Error (RMSE),
coefficient of variation = (standard deviation/mean)×100

Fig. 2. Framework for monitoring water quality. Passive Sensors rely on naturally occurring external energy source (e.g., sun for illumination), whereas Active
sensors provide own energy source to scan target.
Abbreviations: GIS, Geographical Information Systems; GPS, Global Positioning Systems.

Fig. 3. The system-based approach to Water Quality determination.

Abbreviations: GIS: Geographic Information Systems; GPS: Global Positioning Systems.

software's, digital processing techniques are beyond the scope of this quality that accurately represents measured surface signals. Atmo-
work, but are accessible online or from the following references (Chang spheric signal attenuation and geometric distortions are assumed min-
et al., 2015; de Paul Obade and Lal, 2013; Haji Gholizadeh et al., 2016; iscule in proximal sensing given the reduced atmospheric path length
Hajigholizadeh, 2016; Ouma, 2016; Pérez Hoyos et al., 2016; Revilla- and proximity to target of measurement. However, data acquired from
Romero et al., 2016). Prior knowledge of spatial (pixel size, or size of aerial/space platforms have substantial path length, necessitating geo-
smallest feature distinguishable by sensor), spectral (band ranges, or metric and radiometric corrections. Radiometric correction minimizes
region of electromagnetic spectrum sensed), temporal (frequency of atmospheric distortions thereby enhancing clarity of imagery, whereas
imagery acquisitions), radiometric (color depth, or number of digital geometric correction compensates for systematic and random errors
levels representing data) resolution, and swath (area of coverage) are necessary for accurate extraction of distance, area and direction in-
important considerations in sensor selection (Ouma, 2016). Notably, formation from imagery (Chang et al., 2015).
remotely sensed data should be calibrated to ensure highest data Spaceborne sensors may be of high spatial resolution (e.g., Compact

225
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

Airborne Spectrographic Imager (CASI) and HyMap hyperspectral,

-complex and time consuming calculation,

quality parameters and evaluation criteria

-insensitive score system i.e., equal effect

of different water quality parameters in
-lack of flexibility in selection of water
QuickBird, IKONOS, and SPOT-5/HRG), medium (e.g., Landsat), or

-non-comprehensive input parameters,

coarse resolution with limited spectral range (e.g., National Oceanic
and Atmospheric Administration (NOAA) which is a consortium of
several coarse resolution sensors) (Table 3) (Chang et al., 2015; Haji

-subjective rating curves,

-subjective rating curves
Gholizadeh et al., 2016; Hajigholizadeh, 2016; Pérez Hoyos et al., 2016;
Reif, 2011; Revilla-Romero et al., 2016). Aerial sensors operating from
-inflexible structure,

final DWQI score

relatively lower platforms (altitude) have greater spectral and spatial
Disadvantages

resolution but less swath width (Table 4) (Olmanson et al., 2013). Be-
cause hyperspectral sensors have numerous, narrower, spectral bands,
they contain more precise and detailed information than multispectral
and panchromatic respectively; but require greater storage and pro-
cessing requirements (Chang et al., 2015; Haji Gholizadeh et al., 2016).

water quality benchmarks and

Data from coarse spatial resolution satellite sensors are generally lower
-flexibility in selection of
in cost and sometimes free, compared with high spatial resolution sa-
estimating water quality
estimates water quality

Simple, and stable for

tellite or aerial sensors (Chipman et al., 2009; Vinciková et al., 2015).

-simple calculations,

Besides, airborne surveys require more flight planning, and adjustment

considerations to cater for air traffic, solar conditions, weather and
Advantages

parameters

flight dynamics (i.e., pitch, roll, yaw) (Chang et al., 2015; Chipman
et al., 2009; Reif, 2011). Data acquired using ground based sensors such
as portable hand-held spectroradiometers (e.g., Fieldspec,1 PSR+
35002) can be used to calibrate or even synergize data from space or
unique parameters i.e., dissolved oxygen, fecal coliforms, pH,

Turbidity, Copper (Cu), Fecal coliforms, Fluoride (F−), Lead

Acceptability parameters: Aluminum (Al), Ammonia nitrogen,
Calcium (Ca2+), Chloride (Cl−), Hardness (total), Iron (Fe),
Magnesium (Mg2+), pH, Sodium (Na+), Sulphate (SO42−),

aerial platforms (de Paul Obade et al., 2013; Maruthi Sridhar and
(Pb), Manganese (Mn), Mercury (Hg), Nitrate (NO3−) and
Health – based parameters: Arsenic (As), Cadmium (Cd),

Vincent, 2007).
biochemical oxygen demand, nitrate, total phosphate,

Oftentimes relating remotely sensed spectra to specific water

properties can be challenging. In general, the criteria used to decipher
information from remotely sensed data include color, pattern, texture
temperature change, turbidity, total solids

and tone. Gleaning water quality information from remotely sensed

Total dissolved solids (TDS), Zinc (Zn).

data initially requires correlating sensed signals (e.g., reflectance,

brightness temperature because water has low emissivity, backscatter)
directly recorded or available in spectral libraries to surrogate in situ
measured variables such as total sediments (or turbidity), algae con-
centrations (~chlorophyll-a), particulate and/or dissolved organic
22 parameters i.e.,

matter (PDOM), temperature, dissolved oxygen, salinity, total phos-

phorus (TP), total nitrogen (TN), chlorides, fecal coli bacteria, trans-
Nitrite (NO2−)

parency (e.g., Secchi disk depth), or pH (Chipman et al., 2009; Ramadas

Parameters

and Samantaray, 2018; Ritchie et al., 2003). Basically, clear water

flexible

generates higher reflectance within the blue region of the visible elec-
tromagnetic spectrum than in the red, and vice versa for turbid water.
-uses rating curves ranging from 0 to 100 with zero denoting

Compares observed water quality parameters to standard or

Moderate Resolution Imaging Spectrometer (MODIS) sensor on the

-Selects and aggregates product of the respective weighted

standardizing observations using subjective rating curves

Some standardized water quality index (WQI) (adapted from Mohebbi et al., 2013).

Aqua satellite, Landsat, and Earth Observing-1 (EO-1) satellite data

guideline values considered benchmarks rather than
-development of sub-index with regression statistics

have been used to model the relationship between surface reflectance,

-final aggregation of DWQI with Min-Max operator

chlorophyll-a, and water hyacinth presence (le Roux et al., 2016).

-assignment of weight factors to parameters,

Awange et al. (2008a, 2008b) demonstrated using Gravity Recovery

and Climate Experiment (GRACE) and auxiliary data that abrupt cli-
poor quality and 100 high water quality

-parameter selection and categorization,

mate change and not the expanded hydroelectric power station in

Uganda significantly contributed to the receding Lake Victoria water
level. At a coarse spatial resolution (i.e., 300–400 km), GRACE satellite
maps temporal variations of the Earth's gravitational field which highly
water quality parameters,

correlates with terrestrial water storage after screening-out atmospheric

and oceanic effects (Pérez Hoyos et al., 2016). Aircraft-mounted
thermal sensors, Advanced Very High Resolution Radiometer (AVHRR)
4 definite steps:

or other satellite platforms, not only permit mapping of thermal plumes

Method

but may also provide new insights on the role oceans play in regulating
weather (Ritchie et al., 2003). Alternately, thermal pollution arising
from extreme water temperature variability attributed to thermal re-
(2) Drinking Water Quality Index

leases from electrical power plants, adversely affect aquatic ecosystem.

Sanitation Foundation Water

Indeed, high water temperatures kill aquatic life because essential gases
(3) Canadian Drinking Water
(1) NSFWQI (U.S. National

Quality Index (DWQI)

(e.g., oxygen) are released from the water into the atmosphere (Kumar
and Puri, 2012). Although temperatures of surface water bodies sig-
nificantly fluctuate on a seasonal or even daily basis, that for
Quality Index)

(DWQI)
Table 2

1
http://www.asdi.com/products-and-services/fieldspec-spectroradiometers/fieldspec-
Index

4-hi-res
2
http://www.spectralevolution.com/spectroradiometer_PSR_plus.html

226
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

Table 3
Some current spaceborne water quality systems (Modified from Hajigholizadeh, 2016; Chang et al., 2015).
Satellite sensor Spectral bands (wavelength, nm) Spatial resolution Approximate revisit interval Category
(m) (days)

OrbView-2 SeaWiFS 8 (402–885) 1130 16 Coarse (regional−global)

Envisat-1 MERIS 15 (390–1040) 300–1200 daily
Terra MODIS 2 (620–876)-5 (459–2155)-29 (405–877 & thermal band) 250, 500, 1000 1 to 2
EO-1 Hyperion 242 (350–2570) 30 16 Moderate resolution
Landsat-8 OLI/TIRS 5 (430–880)-1 Pan (500–680)-2SWIR (1570–2290)-1cirrus cloud 30, 15, 100 16
detection (1360–1380)-2TIRS (10600–12,510)
HICO 128 (350–1080) 100 10
Digital Globe 4 (430–918)-1 Pan (450–900) 2.62–0.65 2 and half High resolution
Quickbird
SPOT-5 HRG 3 (500–890)-1 Pan (480–710)-1SWIR (1580–1750) 2.5; 5–10–20 2 to 3
CARTOSAT Pan (500–850) 2.5 5

groundwater is relatively stable (Pérez Hoyos et al., 2016). Active Table 5

sensor Radar data such as the synthetic aperture radar (SARs) are useful Some current microwave radiometers for water quality studies (Hajigholizadeh,
for mapping oil spills in water bodies, ocean topography and for re- 2016).
gional ice monitoring (Chang et al., 2015; Haji Gholizadeh et al., 2016; Satellite Spatial resolution (km) Swath Frequency (GHz)
Minvielle et al., 2015). width (km)
Different inversion methods, for instance, empirical (e.g., statis-
Nimbus-5 25 3000 19.4
tical), analytical (bio-optical) and hybrid (i.e., empirical and analytical
TRMM 8 × 6 at 85.5 GHz to 760 10.7,19.4,21.3,37.0, & 85.5
combined) are frequently used to extract water quality information 72 × 43 at 10.7 GHz
from remotely sensed data; sensed from specific bands, and transformed SEASAT 22 at 37.1 GHz to 100 600 6.6, 10.7, 18.0, 21.0, and
into indices (Chang et al., 2015; Gitelson et al., 2008; Ramadas and at 6.6 GHz 37.1
Samantaray, 2018). Remote sensing indices are numerical indicators
derived from ratios of spectral bands that highly correlate with specific
object of interest. Empirical methods are site specific and non-trans- WQI can be constructed from measured water quality constituents (e.g.,
ferable, whereas analytical methods based on the correlation between Table 1) sensed over a broad electromagnetic spectrum, for instance, by
optical information and the radiative transfer equation are transferable screening key variables using factor analyses in tandem with multiple
(Reif, 2011; Ritchie et al., 2003). However, generic analytical models regression methods to analyze the interrelationship between measured
are yet to be constructed with data from hyperspectral bands (Chang and latent critical variables. Eq. (1) exemplifies WQI parameters, which
et al., 2015). Mapping using remotely sensed data entails preprocessing can subsequently be graphed as WQI (y-axis) over time (x-axis) to de-
and classification techniques such as supervised and unsupervised, pict spatial-temporal trends:
which utilize statistical algorithms to discriminate among groups.
Major preprocessing steps especially critical in change detection and WQI ≡ {weather (e. g., precipitation, temperature, wind speed)
time-series analyses include georeferencing, radiometric calibration, + land use and management (e. g., fertilizer
atmospheric corrections and cloud removal (Pérez Hoyos et al., 2016). /manure application and tillage practiced)
Classifying digital imagery can be useful for making predictions and
interpolation. Common classification techniques include: (i) cluster + remotely sensed data (spectral library acquired at different time
analyses, which group data based on similarities within classes, or /location) + water constituent properties (i. e.
dissimilarities of different classes, (ii) structural equation models (e.g., , clarity, pesticides, bacteria, nutrient loading)
multivariate regression, least squares optimization, factor analyses, and
discriminant analysis or supervised pattern recognition technique that + others (e. g.
uses linear combinations of several variables to construct statistical , assumed baseline, reservoir circulation
classification of sample into categorical-dependent values, and (iv) , residence time for water constituents) } (1)
geostatistics (de Paul Obade and Lal, 2013; Larson and Pierce, 1994).
It is desirable to develop a standardized objective single-value WQI WQI represents water quality index
that rates water quality based on specific use on a % scale, ranging from Applicable parametric multivariate regression methods that syn-
zero to 100 with for instance, 100% denoting high/excellent quality thesize complex models include the: (i) forward selection which selects
and 0% poor quality water (Abtahi et al., 2015). Ascertaining high and predictor variables sequentially until the model fit (R2) cannot be im-
low quality water thresholds a priori is critical for generating a stan- proved, (ii) backward elimination or reverse of forward selection, or
dardized framework to ease comparability between WQIs values. Hy- (iii) stepwise method which deletes one non-significant predictor
pothetically, utilization of electromagnetic radiation to monitor water variables upon each iteration. However, parametric statistics must sa-
quality is feasible because backscattering characteristics of water is tisfy the following assumptions: (i) independence of observations, (ii)
dependent on the type and concentration of substances in water. Thus, linearity, (iii) homoscedasticity, and (iv) normal distribution of errors

Table 4
Some current Airborne Water Quality Sensors (modified from Hajigholizadeh, 2016).
Sensor Scan system/sensor type Number of bands, & spectral range (μm) Resolution (m) Imaging swath

Airborne Visible Infrared Imaging Spectrometer Whiskbroom/Hyperspectral 224 (0.40–2.50 μm) 17 12 km (614 pixels/scanline)
(AVIRIS)
Daedalus Multispectral Scanner (MSS) Pushbroom/Multispectral 12 (0.42–14.00 μm) 25 714 pixels/scanline
PROBE-1 in USA (HyMap) Whiskbroom/Hyperspectral 128 (0.40–2.50 μm) 3 to 10 512 pixels

227
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

(Chong and Jun, 2005; Hajigholizadeh and Melesse, 2017; Mehmood assessment technique whereby each measured observation is sequen-
et al., 2012). In contrast, non-parametric methods such as artificial tially estimated and error determined (de Paul Obade and Lal, 2013;
neural networks (ANNs), support vector machines (SVMs), genetic al- Mehmood et al., 2011, 2012). Confusion matrix or classification table is
gorithms (GAs), decision tree techniques, principal component analyses normally used to evaluate accuracy of remotely sensed data
(PCA), partial least squares regression (PLSR) are parsimonious (Hajigholizadeh and Melesse, 2017). This table has rows representing
(Hajigholizadeh and Melesse, 2017; Liou et al., 2004). Multivariate observed categories of the dependents, the columns having predicted
statistics not only screen significant model variables but can also extract categories for each dependent, whereas the diagonal representing per-
regression coefficients even from quantitative and qualitative (e.g., fect predictions. Thus, the percentage of correct classifications is com-
management) data simultaneously, which can conjoined as exemplified puted based on relating the diagonal values with the totals in the row or
in de Paul Obade and Lal, (2016). In principle, the construction of a Soil columns, using Kappa. Kappa (KHAT statistic) is a metric that quantifies
Quality Index (SQI) and WQI are similar, because both constitute che- the difference between actual and chance (Congalton, 1991;
mical, physical and biologic attributes. Eq. (2) is a prototype water Hajigholizadeh and Melesse, 2017). Kappa values range between ne-
quality model subsequently aggregated, standardized and transformed gative to positive one, with positive one value indicating a perfect
into WQI (%) (Eq. (3)). For monitoring purposes or to conduct relative classification significantly better than a random result, whereas nega-
comparison of water quality which is the gist of WQIs, the units cancel tive values represent a poor classification (Congalton, 1991; de Paul
out so long as “oranges are compared with oranges”, that is, same input Obade et al., 2014; Foody, 2002, 2010).
variables (e.g., physical, chemical and biologic properties) are modelled
into WQI per site.
6. Conclusions and future perspectives
Attribute Index (AI) α {water properties (e. g.
, clarity, nutrient concentration…..)} (2) This review highlights water security issues and explicates potential
requirements for constructing a synthesized WQI linking biogeochem-
(WQI = AI MAXAI − MINAI ) × 100 (3)
ical water properties to refined remotely sensed data. Indeed, water
quality is determined by quantifying concentration or existence of
AI: Attribute Index computed from derived remote sensing band specific biological, chemical and physical water constituent properties.
ratios. However, constituent elements that can threaten water quality do
MAXAI: Maximum Aggregate Index remotely sensed for good change over time (i.e., sometimes high concentration of Arsenic occur).
quality water (e.g., good drinking water per WHO standards). Thus, this review proposes a WQI generated from fused field and re-
MINAI: Minimum Aggregate Index remotely sensed for poor quality motely sensed data that can integrate multivariate qualitative and
water (e.g., waste-water). quantitative water constituent properties into a “single-value” WQI, so
Other techniques include the parsimonious decision tree which re- insights can be gained to guide future model development and appli-
cursively splits data into mutually exclusive subsets using tree like cations. WQI information is pertinent for (i) decision making and
partitions, and classifies data either continuously or categorically (dis- management especially regarding control and remediation of con-
cretely) (de Paul Obade et al., 2014; Saghebian et al., 2014). Yet, the tamination hotspots, (ii) reporting and understanding various water
challenge remains identifying specific spectral resolution sensitive to quality threats to human health, and (iii) surveillance of water quality
water constituents to rank and map critical parameters influencing life cycle, for instance, from the reservoir to tap.
water quality. Nonetheless, the multivariate regression models and Some future remote sensing systems equal to this task are also listed
geostatistics may facilitate interpolation and mapping of point data in Table 6. It is important to note that small payload sensors are be-
over a continuous surface, especially through fusing field data with coming fashionable because their electronics and detection sensors
remotely sensed data. have reduced size and mass (Ouma, 2016). For instance, unmanned
aerial vehicles (drones), or nanosatellites (e.g., from Terra Bella3 and
5. Calibration and validation Planet Labs4) packed with high-powered optics and sensors are not only
affordable in comparison to current aerial and space borne sensors, but
Although uncertainty and sensitivity analyses (SA) are often carried their agility allows them to achieve attitude-change maneuvers rapidly
out in tandem, they serve different purposes; because uncertainty fo- allowing them to operate from any orbit; thereby providing high tem-
cuses on error propagation, whereas SA explores the strength of re- poral resolution products (Olmanson et al., 2013; Ouma, 2016; Rango
lationships between model inputs and outputs (Yuan et al., 2015). SA is et al., 2009).
applied in parameter fixation, screening-out of key model parameters, From a practical standpoint, calibrated remote sensors/devices sy-
resource allocation for parameter and data measurement, model or al- nonymous to “no contact thermometers” should be constructed that
gorithm corroboration, and to justify scientific based decisions directly measure water quality (e.g., in %, with 100% denoting ex-
(Hajigholizadeh and Melesse, 2017; Yuan et al., 2015). However, SA cellent/high quality and 0% low quality). Such devices will not only
results of models are derived for specific scenarios thus are site and provide an insight on the deviation from pristine conditions for natural
condition dependent. For all applications, the reliability of geo-in- water bodies (e.g., to inform recreational water users); but also enhance
formation products depends on sampling strategy; accuracy between knowledge on the interrelationship between water quality dynamics,
model inputs, outputs and actual field conditions. Among the common abrupt climatic change and anthropogenic land use/management.
sampling techniques for field data acquisition include; random, cluster, Thus, robust algorithms accurately synthesizing water quality in-
stratified, systematic, and Latin hypercube among others. Visual in- formation and transferable to other regions are required. Finally, de-
spection of statistical input versus output plots (e.g., box and whisker velopment of high fidelity and concise WQI will be useful for time series
plots) can rudimentarily inform on model reliability. analyses so as to enhance the understanding of socio-economic dy-
Frequently used “goodness-of-fit-measures” for quantifying accu- namics vis–à–vis water quality status.
racy include: (i) coefficient of determination (R2), which describes the
closeness of data to the fitted regression line, (ii) standard deviation,
and (ii) root-mean square error (RMSE). For instance, a lower RMSE
signifies higher accuracy, and conversely for R2. R2 > 0.7 is considered
as the limit of applicability for linear regression-based sensitivity 3
https://terrabella.google.com/?s=in-action&c=case-mongolia
4
techniques (Yuan et al., 2015). Cross validation is another accuracy https://www.planet.com/

228
 V. de Paul Obade, R. Moore

Table 6
Future water quality observation systems.
Sensor Swath width Spatial resolution (m) Spectral channels & resolution (nm) Temporal Application Launch year Reference
resolution (days)

Sentinel-2A (European 290 km 10 m (4 visible and near-infrared Multispectral imager (MSI) 13 spectral bands 5 Chlorophyll, 23rd June 2015 1
Space Agency) bands), 20 m (6 red-edge/shortwave- (443 nm–2190 nm) Harmful algal blooms (HABs), turbidity, Operational
infrared bands (SWIR)) & 60 m (3 water content indices 7 years
atmospheric correction bands)
Sentinel-2B/2C • monitor ocean conditions, topography, pollution, currents. 2017/2021 1,4
Sentinel-3A • monitor impact of climate changes on melting ice, ocean temperature 2016 2,4
Sentinel-3B, 3C, 3D, 4 • improve meteorological forecasts Before 2021 2
Operate for
7 years
Environmental Mapping 30 km 30 × 30 m spectral range from 420 nm to 1000 nm (VNIR) 4 -assess spatial and temporal water scarcity 2018 and operate 4
and Analysis and from 900 nm to 2450 nm (SWIR) and water quality problems? for 5 years
Programme (EnMAP) Hyperspectral; Visible to SWIR -how climate change, intensive agriculture,

229
water demanding industries and high
population density affect water availability?
-how water quality is impacted by land use
change and climate?
Hyperspectral InfraRed instruments on 60 m @nadir visible to short wave infrared (VSWIR: VSWIR = 19 -water quality not assigned ~ 3
Imager (HyspIRI) low orbit 380 nm–2500 nm) in 10 nm bands and a TIR = 5 -surface temperature maps 2016
multispectral 3 to 12 μm in the mid and -Evapotranspiration (ET).
thermal infrared (TIR) -drought
GeoEye-2 0.25 2016 4
Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) http://pace.gsfc.nasa.gov
Web-based decision support systems simple and streamlined to enhance data accessibility to end-users and decision makers 4
Google Earth Engine can be used to create cloud free mosaics https://earthengine.google.com/

1. http://www.esa.int/Our_Activities/Observing_the_Earth/Copernicus/Sentinel-2/Facts_and_figures
2. http://space.skyrocket.de/doc_sdat/sentinel-3.htm
3. https://hyspiri.jpl.nasa.gov/.
4. Additional sensor information available also in Chang et al., 2015; Haji Gholizadeh et al., 2016; Ouma, 2016.
Environment International 119 (2018) 220–231
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

Acknowledgements Foody, G.M., 2002. Status of land cover classification accuracy assessment. Remote Sens.
Environ. 80, 185–201.
Foody, G.M., 2010. Assessing the accuracy of land cover change with imperfect ground
The authors acknowledge the valuable feedback from the reviewers. reference data. Remote Sens. Environ. 114, 2271–2285.
Freeman, M.I.A., 2000. Water pollution policy. In: Portney, Paul R., Stavins, R.N. (Eds.),
References Public Policies for Environmental Protection, Washington, D.C.
Gari, S.R., Newton, A., Icely, J.D., 2015. A review of the application and evolution of the
DPSIR framework with an emphasis on coastal social-ecological systems. Ocean
Abtahi, M., Golchinpour, N., Yaghmaeian, K., Rafiee, M., Jahangiri-Rad, M., Keyani, A., Coast. Manag. 103, 63–77.
Saeedi, R., 2015. A modified drinking water quality index (DWQI) for assessing Ghosh, G., Ribaudo, M., Shortle, J., 2011. Baseline requirements can hinder trades in
drinking source water quality in rural communities of Khuzestan Province, Iran. Ecol. water quality trading programs: evidence from the Conestoga watershed. J. Environ.
Indic. 53, 283–291. Manag. 92, 2076–2084.
Adhikari, K., Hartemink, A.E., 2016. Linking soils to ecosystem services — a global re- Gitelson, A.A., Dall'Olmo, G., Moses, W., Rundquist, D.C., Barrow, T., Fisher, T.R., Gurlin,
view. Geoderma 262, 101–111. D., Holz, J., 2008. A simple semi-analytical model for remote estimation of chlor-
Ashbolt, N., Grabow, W., Snozzi, M., 2001. Water Quality: Guidelines, Standards and ophyll-a in turbid waters: validation. Remote Sens. Environ. 112, 3582–3593.
Health. IWA Publishing, London, UK (ed^eds). Gleick, P.H., 2014. Water, drought, climate change, and conflict in Syria. Weather
Awange, J.L., Ogalo, L., Bae, K.-H., Were, P., Omondi, P., Omute, P., Omullo, M., 2008a. Climate Soc. 6, 331–340.
Falling Lake Victoria water levels: is climate a contributing factor? Clim. Chang. 89, Gleick, P.H., Palaniappan, M., 2010. Peak water limits to freshwater withdrawal and use.
281–297. Proc. Natl. Acad. Sci. 107, 11155–11162.
Awange, J.L., Sharifi, M.A., Ogonda, G., Wickert, J., Grafarend, E.W., Omulo, M.A., Ha, L., Bastiaanssen, W., van Griensven, A., van Dijk, A., Senay, G., 2018. Calibration of
2008b. The falling Lake Victoria water level: GRACE, TRIMM and CHAMP satellite spatially distributed hydrological processes and model parameters in SWAT using
analysis of the Lake Basin. Water Resour. Manag. 22, 775–796. remote sensing data and an auto-calibration procedure: a case study in a Vietnamese
Awange, J.L., Mpelasoka, F., Goncalves, R.M., 2016. When every drop counts: analysis of River Basin. Water 10, 212.
droughts in Brazil for the 1901–2013 period. Sci. Total Environ. 566–567, Haji Gholizadeh, M., Melesse, A.M., Reddi, L., 2016. Spaceborne and airborne sensors in
1472–1488. water quality assessment. Int. J. Remote Sens. 37, 3143–3180.
Bouma, J., McBratney, A., 2013. Framing soils as an actor when dealing with wicked Hajigholizadeh, M., 2016. Water Quality Modelling Using Multivariate Statistical
environmental problems. Geoderma 200–201, 130–139. Analysis and Remote Sensing in South Florida. FIU Electronic Theses and
Bushaw-Newton, K.L., Seller, K.G., 1999. Harmful Algal Blooms. NOAA's State of the Dissertations Florida International University (2992).
Coast Report. National Oceanic and Atmospheric Administration, Silver Spring, MD. Hajigholizadeh, M., Melesse, A.M., 2017. Assortment and spatiotemporal analysis of
Causse, J., Baurès, E., Mery, Y., Jung, A.-V., Thomas, O., 2015. Variability of N export in surface water quality using cluster and discriminant analyses. Catena 151, 247–258.
water: a review. Crit. Rev. Environ. Sci. Technol. 45, 2245–2281. Haq, H., 2010. In: Monitor, T.C.S. (Ed.), BP Oil Spill 2010: How Much Will It Cost?.
Chang, N.-B., Imen, S., Vannah, B., 2015. Remote sensing for monitoring surface water House, M.A., 1990. Water quality indices as indicators of ecosystem change. Environ.
quality status and ecosystem state in relation to the nutrient cycle: a 40-year per- Monit. Assess. 15, 255–263.
spective. Crit. Rev. Environ. Sci. Technol. 45, 101–166. IFRC, 2016. Emergency Plan of Action (EPoA) Kenya: Floods. Ministry of Health, WFP,
Chaturvedi, M.K., Bassin, J.K., 2010. Assessing the water quality index of water treatment UNICEF, UNOCHA, Ministry of Water and Irrigation, Child Fund, PSK, County
plant and bore wells, in Delhi, India. Environ. Monit. Assess. 163, 449–453. Governments.
Chellaney, B., 2013. Water, Peace, and War: Confronting the Global Water Crisis. Ingraham, C., 2016. This Is How Toxic Flint's Water Really Is. The Washington Post.
Rowman & Littlefield Publishers (ed^eds). Khaled, A., Omar, F., Asma, A., John, S., 2016. Regional Powers Back Studies on Impact of
Chipman, J.W., Olmanson, L.G., Gitelson, A.A., 2009. Remote Sensing Methods for Ethiopia's Nile Dam.
LakeManagement: A Guide for Resource Managers and Decision-Makers. Khan, F., Husain, T., Lumb, A., 2003. Water quality evaluation and trend analysis in
NorthAmerican Lake Management Society and University of Minnesota, Madison, WI selected watersheds of the Atlantic region of Canada. Environ. Monit. Assess. 88,
(ed^eds). 221–248.
Chong, I.G., Jun, C.H., 2005. Performance of some variable selection methods when Kisekka, I., Schlegel, A., Ma, L., Gowda, P.H., Prasad, P.V.V., 2017. Optimizing preplant
multicollinearity is present. Chemom. Intell. Lab. Syst. 78, 103–112. irrigation for maize under limited water in the High Plains. Agric. Water Manag. 187,
Cobbina, S.J., Anyidoho, L.Y., Nyame, F., Hodgson, I.O.A., 2010. Water quality status of 154–163.
dugouts from five districts in Northern Ghana: implications for sustainable water Komorowicz, I., Barałkiewicz, D., 2016. Determination of total arsenic and arsenic species
resources management in a water stressed tropical savannah environment. Environ. in drinking water, surface water, wastewater, and snow from Wielkopolska, Kujawy-
Monit. Assess. 167, 405–416. Pomerania, and Lower Silesia provinces, Poland. Environ. Monit. Assess. 188, 504.
Coffman, K., Dobuzinskis, A., 2013. In: Johnston, C., Osterman, C., Chang, R. (Eds.), Kozlowski, D., Hall, R., Swanson, S., Heggem, D., 2016. Linking management and riparian
Likely Death Toll in Colorado Floods Rises to at Least 10. Reuters. physical functions to water quality and aquatic habitat. J. Water Resour. Prot. 8,
Congalton, R.G., 1991. A review of assessing the accuracy of classifications of remotely 797–815.
sensed data. Remote Sens. Environ. 37, 35–46. Kumar, M., Puri, A., 2012. A review of permissible limits of drinking water. Indian J.
Cutter, S., Osman-Elasha, B., Campbell, J., Cheong, S.M., McCormick, S., Pulwarty, R., Occup. Environ. Med. 16, 40–44.
Supratid, S., Ziervogel, G., 2012. Managing the risks from climate extremes at the Lal, R., 2009a. Technology Without Wisdom. (ed^eds).
local level. In: Field, C.B., Barros, V., Stocker, T.F., Qin, D., Dokken, D.J., Ebi, K.L., Lal, R., 2009b. Tragedy of the Global Commons: Soil, Water and Air. (ed^eds).
Mastrandrea, M.D., Mach, K.J., Plattner, G.-K., Allen, S.K., Tignor, M., Midgley, P.M. Lal, R., 2013. Climate-strategic agriculture and the water-soil-waste nexus. J. Plant Nutr.
(Eds.), Managing the Risks of Extreme Events and Disasters to Advance Climate Soil Sci. 176, 479–493.
Change Adaptation. IPCC Working Group II, Technical Support Unit (TSU), 260 Lal, R., 2015a. Research and development priorities in water security. Agron. J. 107,
Panama Street, Stanford, CA 94305, USA. 1567–1572.
de Paul Obade, V., 2017. Quantifying the interactions of land management practices and Lal, R., 2015b. World water resources and achieving water security. Agron. J. 107,
agricultural productivity using a soil quality index. Soil Use Manag. 33, 639–652. 1526–1532.
de Paul Obade, V., Lal, R., 2013. Assessing land cover and soil quality by remote sensing Lal, R., 2018. Sustainable intensification of China's agroecosystems by conservation
and geographical information systems (GIS). Catena 104, 77–92. agriculture. Int. Soil Water Conserv. Res. 6, 1–12.
de Paul Obade, V., Lal, R., 2016. Towards a standard technique for soil quality assess- Lal, R., Delgado, J.A., Gulliford, J., Nielsen, D., Rice, C.W., Van Pelt, R.S., 2012. Adapting
ment. Geoderma 265, 96–102. agriculture to drought and extreme events. J. Soil Water Conserv. 67, 162A–166A.
de Paul Obade, V., Lal, R., Chen, J., 2013. Remote sensing of soil and water quality in Larson, W.E., Pierce, F.J., 1994. The dynamics of soil quality as a measure of sustainable
agroecosystems. Water Air Soil Pollut. 224, 1–27. management. In: Doran, J.W., Coleman, D.C., Bezdicek, D.F., Stewart, B.A. (Eds.),
de Paul Obade, V., Lal, R., Moore, R., 2014. Assessing the accuracy of soil and water Defining Soil Quality for a Sustainable Environment.
quality characterization using remote sensing. Water Resour. Manag. 28, 5091–5109. Laurent, F., Ruelland, D., 2011. Assessing impacts of alternative land use and agricultural
DIA, 2012. Intelligence Community Assessment. Defense Intelligence Agency (in: (DIA) practices on nitrate pollution at the catchment scale. J. Hydrol. 409, 440–450.
D.I.A., ed. Intelligence Community Assessment). le Roux, Jeanné, Winstead, Daryl Ann, Vacek, Austin, Fischer, Christina, Amirazodi, Sara,
EEA, 1997. The DPSIR Framework. European Environmental Agency. Luvall, Jeffrey, Griffin, D.R., 2016. Mapping Water Quality. Lake Victoria's Water
EPA, 2000. Water Quality Conditions in the United States. A Profile from the 1998 Quality: Past, Present, and Future. NASA Marshall Space Flight Center, Huntsville,
National Water Quality Inventory Report to Congress. National Water Quality Alabama.
Inventory: 1998 Report to Congress (EPA841-R-00-001) United States Environmental Levine, S., 2016. Judge Rejects Riot Charges for Journalist Amy Goodman after Oil
Protection Agency. Pipeline Protest the Guardian.
EPA, 2016a. Drinking Water Contaminants – Standards and Regulations. Lindstrom, L., 2016. Toledo Looks at Requiring Landlords to Remove Lead, Pre-1978
EPA, 2016b. Implementing Clean Water Act Section 303(d): Impaired Waters and Total Properties would Be Inspected the Blade.
Maximum Daily Loads (TMDLs). US Environmental Protection Agency. Liou, S.-M., Lo, S.-L., Wang, S.-H., 2004. A generalized water quality index for Taiwan.
EPRI, 2016. Ohio River Basin Trading Project. Electric Power Research Institute. Environ. Monit. Assess. 96, 35–52.
Esteban, O.-O., Max, R., 2016. World Population Growth. OurWorldInData.org. Maruthi Sridhar, B.B., Vincent, R.K., 2007. Spectral reflectance measurements of a mi-
FAO, 2008. Coping With Water Scarcity. An Action Framework for Agriculture and Food crocystis bloom in upper Klamath Lake, Oregon. J. Great Lakes Res. 33, 279–284.
Security. Food and Agriculture Organization of the United Nations, Viale delle Terme Mehmood, T., Martens, H., Saebo, S., Warringer, J., Snipen, L., 2011. A partial least
di Caracalla, 00100 Rome, Italy. squares based algorithm for parsimonious variable selection. Algorithms Mol. Biol. 6.
FAO, 2013. The State of Food Insecurity in the World. The Multiple Dimensions of Food Mehmood, T., Liland, K.H., Snipen, L., Saebo, S., 2012. A review of variable selection
Security. Food and Agricultural Organization of the United Nations, Rome. methods in partial least squares regression. Chemom. Intell. Lab. Syst. 118, 62–69.

230
V. de Paul Obade, R. Moore Environment International 119 (2018) 220–231

Michalak, A.M., Anderson, E.J., Beletsky, D., Boland, S., Bosch, N.S., Bridgeman, T.B., Helder, D., Irons, J.R., Johnson, D.M., Kennedy, R., Scambos, T.A., Schaaf, C.B.,
Chaffin, J.D., Cho, K., Confesor, R., Daloğlu, I., Depinto, J.V., Evans, M.A., Schott, J.R., Sheng, Y., Vermote, E.F., Belward, A.S., Bindschadler, R., Cohen, W.B.,
Fahnenstiel, G.L., He, L., Ho, J.C., Jenkins, L., Johengen, T.H., Kuo, K.C., Laporte, E., Gao, F., Hipple, J.D., Hostert, P., Huntington, J., Justice, C.O., Kilic, A., Kovalskyy,
Liu, X., Mcwilliams, M.R., Moore, M.R., Posselt, D.J., Richards, R.P., Scavia, D., V., Lee, Z.P., Lymburner, L., Masek, J.G., McCorkel, J., Shuai, Y., Trezza, R.,
Steiner, A.L., Verhamme, E., Wright, D.M., Zagorski, M.A., 2013. Record-setting algal Vogelmann, J., Wynne, R.H., Zhu, Z., 2014. Landsat-8: science and product vision for
bloom in Lake Erie caused by agricultural and meteorological trends consistent with terrestrial global change research. Remote Sens. Environ. 145, 154–172.
expected future conditions. Proc. Natl. Acad. Sci. 110, 6448–6452. Saghebian, S.M., Sattari, M.T., Mirabbasi, R., Pal, M., 2014. Ground water quality clas-
Minvielle, P., Massaloux, P., Giovannelli, J.F., Minvielle, P., 2015. Fast 3D Synthetic sification by decision tree method in Ardebil region, Iran. Arab. J. Geosci. 7,
Aperture Radar Imaging from Polarization-Diverse Measurements. 4767–4777.
Mohammed, S.M., 2002. A review of water quality and pollution studies in Tanzania. Savard, M., 2000. Modelling risk, trade, agricultural and environmental policies to assess
AMBIO J. Hum. Environ. 31, 617–620. trade-offs between water quality and welfare in the hog industry. Ecol. Model. 125,
Mohebbi, M.R., Saeedi, R., Montazeri, A., Azam Vaghefi, K., Labbafi, S., Oktaie, S., 51–66.
Abtahi, M., Mohagheghian, A., 2013. Assessment of water quality in groundwater Scott, F., 2016. Cancer-Causing Metal Found in Drinking Water in Bergen, Passaic
resources of Iran using a modified drinking water quality index (DWQI). Ecol. Indic. Counties.
30, 28–34. Selman, M., Greenhalgh, S., Branosky, E., Jones, C., Guiling, J., 2009. Water Quality
Moore, R., 2013. School of Environment and Natural Resources (SENR), The Ohio State Trading Programs: An International Overview. (in: Institute W.R., ed. 10 G Street, NE
University, Personal Communication. Washington DC 20002).
Moore, R., 2014. The Role of Trading in Achieving Water Quality Objectives. Semenza, J.C., Höser, C., Herbst, S., Rechenburg, A., Suk, J.E., Frechen, T., Kistemann, T.,
Mueller-Warrant, G.W., Griffith, S.M., Whittaker, G.W., Banowetz, G.M., Pfender, W.F., 2012. Knowledge mapping for climate change and food- and waterborne diseases.
Garcia, T.S., Giannico, G.R., 2012. Impact of land use patterns and agricultural Crit. Rev. Environ. Sci. Technol. 42, 378–411.
practices on water quality in the Calapooia River Basin of western Oregon. J. Soil Shrestha, S., Kazama, F., Newham, L.T.H., 2008. A framework for estimating pollutant
Water Conserv. 67, 183–201. export coefficients from long-term in-stream water quality monitoring data. Environ.
NBC, 2016. Sustainability: Water. NBC Universal Media, 30 Rockefeller Plaza, New York, Model Softw. 23, 182–194.
NY, pp. 10112. Spaliviero, M., De Dapper, M., Maló, S., 2014. Flood analysis of the Limpopo River basin
Niemeijer, D., de Groot, R.S., 2008. Framing environmental indicators: moving from through past evolution reconstruction and a geomorphological approach. Nat.
causal chains to causal networks. Environ. Dev. Sustain. 10, 89–106. Hazards Earth Syst. Sci. 14, 2027–2039.
Oliver, M.A., Webster, R., 2014. A tutorial guide to geostatistics: computing and mod- Stephenson, K., Shabman, L., 2011. Rhetoric and reality of water quality trading and the
elling variograms and kriging. Catena 113, 56–69. potential for market-like reform. JAWRA 47, 15–28.
Olmanson, L.G., Brezonik, P.L., Bauer, M.E., 2013. Airborne hyperspectral remote sensing Stewart, L.R., Farver, J.R., Gorsevski, P.V., Miner, J.G., 2014. Spatial prediction of blood
to assess spatial distribution of water quality characteristics in large rivers: the lead levels in children in Toledo, OH using fuzzy sets and the site-specific IEUBK
Mississippi River and its tributaries in Minnesota. Remote Sens. Environ. 130, model. Appl. Geochem. 45, 120–129.
254–265. Swanson, S., Wyman, S., Evans, C., 2015. Practical grazing management to maintain or
Ouma, Y.O., 2016. Advancements in medium and high resolution Earth observation for restore riparian functions and values. JRA 2, 1–28 (University of Idaho, Rangeland
land-surface imaging: evolutions, future trends and contributions to sustainable de- Center).
velopment. Adv. Space Res. 57, 110–126. UNDESA, 2014. International Decade for Action “Water for Life” 2002–2015.
Pérez Hoyos, I., Krakauer, N., Khanbilvardi, R., Armstrong, R., 2016. A review of ad- UNEP, 2013. Taking the Pulse of the Planet; Connecting Science with Policy. UNEP Global
vances in the identification and characterization of groundwater dependent ecosys- Environmental Alert Service (GEAS).
tems using geospatial technologies. Geosciences 6, 17. UNU, 2013. “Water security”: Experts Propose a UN Definition on Which Much Depends.
Power, A.G., 2010. Ecosystem services and agriculture: tradeoffs and synergies. Philos. United Nations Univ., Dresden, Germany.
Trans. R. Soc. B 365, 2959–2971. Vincent, R.K., Qin, X., McKay, R.M.L., Miner, J., Czajkowski, K., Savino, J., Bridgeman, T.,
Ramadas, M., Samantaray, A.K., 2018. Applications of remote sensing and GIS in water 2004. Phycocyanin detection from LANDSAT TM data for mapping cyanobacterial
quality monitoring and remediation: a state-of-the-art review. In: Bhattacharya, S., blooms in Lake Erie. Remote Sens. Environ. 89, 381–392.
Gupta, A.B., Gupta, A., Pandey, A. (Eds.), Water Remediation. Springer Singapore, Vinciková, H., Hanuš, J., Pechar, L., 2015. Spectral reflectance is a reliable water-quality
Singapore. estimator for small, highly turbid wetlands. Wetl. Ecol. Manag. 23, 933–946.
Ramesh, S., Sukumaran, N., Murugesan, A.G., Rajan, M.P., 2010. An innovative approach Vitharana, U.W.A., Van Meirvenne, M., Simpson, D., Cockx, L., De Baerdemaeker, J.,
of drinking water quality index—a case study from southern Tamil Nadu, India. Ecol. 2008. Key soil and topographic properties to delineate potential management classes
Indic. 10, 857–868. for precision agriculture in the European loess area. Geoderma 143, 206–215.
Rango, A., Laliberte, A., Herrick, J.E., Winters, C., Havstad, K., Steele, C., Browning, D., Vorosmarty, C.J., Green, P., Salisbury, J., Lammers, R.B., 2000. Global water resources:
2009. Unmanned Aerial Vehicle-Based Remote Sensing for Rangeland Assessment, vulnerability from climate change and population growth. Science (New York, N.Y.)
Monitoring, and Management. APPRES (3:033542-033542-033515). 289, 284–288.
Reif, M., 2011. Remote Sensing for Inland Water Quality Monitoring: A US Army Corps of Water, U., 2008. Tackling a global crisis: international year of sanitation. In: United
Engineers Perspective. U.S. Army Corps of Engineers, Washington D.C. Nations (Ed.), Nations U. United Nations, Paris, France.
Revilla-Romero, B., Wanders, N., Burek, P., Salamon, P., de Roo, A., 2016. Integrating WHO, 2008. Safer Water, Better Health: Cost, Benefits and Sustainability of Interventions
remotely sensed surface water extent into continental scale hydrology. J. Hydrol. to Protect and Promote Health. World Health Organization, Geneva, Switzerland.
543, 659–670. WHO (Ed.), 2011. Guidelines for Drinking Water Quality. World Health Organization (in:
Rijsberman, F.R., 2006. Water scarcity: fact or fiction? Agric. Water Manag. 80 (1–3), WHO, ed).
5–22. WHO/UNICEF, 2004. Joint Monitoring Programme for Water Supply and Sanitation;
Ritchie, J.C., Zimba, P.V., Everitt, J.H., 2003. Remote sensing techniques to assess water Meeting the MDG Drinking Water and Sanitation Target: A Mid-Term Assessment of
quality. Photogramm. Eng. Remote Sens. 69, 695–704. Progress. (in: Fund W.H.O. a.U.N.C.s., ed).
Roy, D.P., Ju, J., Lewis, P., Schaaf, C., Gao, F., Hansen, M., Lindquist, E., 2008. Multi- Wright, P., 2016. 10 things to know about Florida's harmful algae blooms. weather.com.
temporal MODIS–Landsat data fusion for relative radiometric normalization, gap Yuan, Y., Khare, Y., Wang, X., Parajuli, P., Kisekka, I., Finsterle, S., 2015. Hydrologic and
filling, and prediction of Landsat data. Remote Sens. Environ. 112, 3112–3130. Water Quality Models: Sensitivity.
Roy, D.P., Wulder, M.A., Loveland, T.R., Woodcock, C.E., Allen, R.G., Anderson, M.C., Zaremba, J., 2016. Cancer-Causing Metal Found in Drinking Water Across New Jersey.

231