PNEC Calculations

MARINE RISK ASSESSMENT
4 ENVIRONMENTAL RISK ASSESSMENT MARINE
4.1 INTRODUCTION
The extension of the existing risk assessment approaches to cover risks to the marine
environment is a logical and important development in the establishment of a comprehensive
risk assessment methodology. Both the Commission report on the operation of several pieces of
legislation in the area of chemicals (COMMISSION, SEC (1998) 1986 final) as well as the
OSPAR Hazardous Substances Strategy (OSPAR, 1998) recognise the need to extend the risk
assessment framework and methodology as developed under Directive 93/67 and Regulation EC
1488/94. This section, therefore, seeks to lay down the principles and concepts that should drive
an assessment of the impacts on the marine environment. In doing so, it also identifies the areas
where a similar approach can be adopted to that described elsewhere within the TGD, as well as
elaborating different methodologies where they are considered more appropriate.
The assessment approaches detailed within the TGD have been developed principally to address
risks, which might arise from emissions to the terrestrial and/or limnic aquatic environment.
These schemes can and must nevertheless act as a starting point for the development of a
comprehensive approach to risk assessment of substances in the marine environment, although
due recognition is given to the many differences both in technical detail and general approaches
which may be necessary. It is not the intention of this section, therefore, to repeat technical
descriptions or equations described elsewhere where the basic methodology for marine
assessment do not differ significantly to that applied to the freshwater environment. Such
technical detail will be appropriately referenced to ensure that clarity is maintained. Rather, the
section will focus on new approaches, which are considered necessary to cover the unique
features of the marine environment.
While the approaches to the assessment must conform to EC requirements for assessment under
Directive 67/548, Regulation 793/93 and Directive 98/8, they must also recognise the objectives
established by OSPAR policy. The approaches will be guided and implemented, therefore, in
accordance with the EU policy under the above legislation as well as taking into account the
OSPAR Strategy on Hazardous Substances. With respect to the OSPAR strategy the assessment
should specifically contribute to the identification of the sources of release for a chemical and
their relative significance in order to facilitate the eventual preparation of measures that
substantially, effectively and proportionately reduce the exposure.
The basic principles of the assessment have been derived in accordance with the experience
gathered by the procedure for chemicals in the frame of the original TGD (EC, 1996). In
attempting to extend current risk assessment methodology to cover the marine environment, it is
necessary to closely investigate the common concepts and protection goals of the available
methods. Where common protection goals were identified, an examination of the
appropriateness of the current methodologies to achieve them has been carried out.
Modifications have been made where necessary to enhance relevance to the marine environment.
Where environmental compartments were not adequately covered by the existing methodologies,
new approaches have been elaborated based on a sound scientific understanding of the problems
and taking account, where appropriate, of the precautionary principle.
The approaches of the original TGD for the inland environment and that required for assessment
of the marine environment share a number of common principles and objectives. Each must
attempt to address the concern for the potential impact of individual substances on the
134
environment by examining both the exposures resulting from discharges/releases of chemicals

and the effects of such emissions on the structure and function of the ecosystem. In the TGD for
the inland environment this is practically done by considering five environmental compartments,
namely the aquatic ecosystem, the terrestrial ecosystem, top predators, the functioning of
Sewage Treatment Plants (STP) and the atmosphere. The environmental compartments are
assessed at the local and the regional spatial scale by comparing the Predicted Environmental
Concentration (PEC) with the Predicted No Effect Concentration (PNEC) for the ecosystem
using data from representative species at different trophic levels for the particular environmental
compartment. Top predators are assessed by assuming an exposure through the food chain. The
assessment addresses the functioning of the ecosystem as determined by the survival and well-
being of all the species in the specific ecosystem. It is assumed that the protection of species
protects ecosystem structure and hence the ecosystem function. It addresses the survival and
well-being of species populations rather than an individual organism.
While this approach must clearly also apply to the marine environment, it must be recognised
that the concepts and methodologies for the inland environment have largely been developed
with the local and regional spatial scales in mind, rather than the potential for global impact.
There are, therefore, additional concerns for the risk assessment of the marine environment,
which may not be adequately addressed by the methodologies used for the inland environmental
risk assessment. These are:
a. the concern that hazardous substances may accumulate in parts of the marine environment
and that:
(i) the effects of such accumulation are unpredictable in the long-term;
(ii) that such accumulation would be practically difficult to reverse;
b. the concern that remote areas of the oceans should remain untouched by hazardous
substances resulting from human activity, and that the intrinsic value of pristine
environments should be protected.
Of these additional concerns (a) above may be seen as the main concern. This is characterised by
a spatial and temporal scale not covered by the inland risk assessment approach. It is a concern
that chemical substances which can be shown both to persist for long periods and bioaccumulate
in biota, can give rise to toxic effects after a greater time and at a greater distance than chemicals
without these properties. While this is also true for the freshwater environment, the additional
concern in the marine environment is that once the chemical has entered the open seas, any
cessation of emission will not necessarily result in a reduction in chemical concentration and
hence any effects become difficult to reverse. Equally, because of the long-term exposures and
long-life-cycle of many important marine species, effects may be difficult to detect at an early
stage.
To meet these concerns, which principally relate to substances that are considered as Persistent,
Bioaccumulative and Toxic (referred to as PBTs), or have other properties which give rise to a
similar level of concern, an assessment approach will be detailed that will give special
consideration to this new protection goal. In this context, the assessment of risk fulfils
specifically the purpose of determining what are the sources, routes and pathways to the marine
environment. This assessment will facilitate in the subsequent risk management decisions on
which measures are the most effective in order to reduce the levels.
The structure of this section on marine risk assessment basically follows the structure of the
inland environmental assessment. It starts with a section of exposure assessment where specific
issues are highlighted relating to marine partitioning processes and marine degradation and
135
where a description is given on how the predicted environmental concentration (PEC) for the
local and regional situation should be derived. In the next section on marine effects assessment
the specific procedures for the derivation of predicted no-effect concentrations (PNECs) for the
aquatic compartment and for sediment are described. This section also deals with the assessment
of possible effects through secondary poisoning via the foodchain in the marine environment.
The section ends with the section on PBT assessment that describes criteria for identification of
persistent, bioaccumulative and toxic substances and includes testing strategies to obtain the
necessary data for this identification. For the risk characterisation the reader is referred to Section 5.
4.2 MARINE EXPOSURE ASSESSMENT
4.2.1 Measured data

Guidance on the use of measured data in the environmental exposure assessment can be found in
Section 2.2 of the TGD for the inland environment. This section covers the selection of adequate
data, the allocation of these data to the regional or local scale and deals with the question
whether measured or estimated data (or both) should be used in the risk characterisation phase.
4.2.2 Partition coefficients

The distribution of a substance in the environment can be predicted from partition coefficients,
which describe the relative concentrations between environmental compartments at equilibrium.
Specific information on the derivation of the partitioning processes between air-aerosol, air-
water, and solids-water in the various compartments can be found in Section 2.3.5. This section
only highlights some specific issues related to the marine environmental conditions.
Measured partition coefficients between water and a second compartment, if available, are
usually derived from studies using non-saline water (freshwater or distilled/deionised water). In
the absence of measured data, the relevant partition coefficients must be extrapolated from the
primary data listed in Section 2.3.2. However, the techniques that allow such an extrapolation
are also largely based on freshwater data sets. Therefore, to assess the distribution of chemicals
in the marine environment, it is necessary to consider the extent to which partition coefficients
may differ between seawater and freshwater.
The ionic strength, composition, and pH of seawater, compared with freshwater, have potential
effects on the partitioning of a chemical with other compartments. To a large extent, these effects
are associated with differences in water solubility and/or speciation of the chemical, compared
with freshwater. The relatively high levels of dissolved inorganic salts in seawater generally
decrease the solubility of a chemical (referred to as salting-out), by about 10-50% for non-polar
organic compounds but by a smaller fraction for more polar compounds (Schwarzenbach et al.,
1993). A recent review found a typical reduction factor of 1.36 (Xie et al., 1997).
For non-ionisable organic substances, the decreased solubility in seawater, compared with
freshwater, is expected to result in proportional increases in the partition coefficients between
water and octanol, organic carbon and air. However, considering the uncertainty in measured
partition values and the uncertainty associated with the frequent need to predict some or all of
the partition coefficients, the differences attributable to the seawater environment (less than a
factor of 2) are unlikely to be significant in risk assessment. Thus, unless measured seawater data
136
of equal reliability are available, freshwater data can be used for non-ionisable organic
compounds without adjustment for the marine environment.
For ionisable organic compounds, as for freshwater, the pH of the environment will affect the
water solubility and partitioning of the substance. There is some evidence that the degree of
dissociation may also be directly affected by the ionic strength of seawater (Esser and Moser,
1982). However, the resulting shift in the dissociation curve is relatively small compared with
that which can occur due to pH for substances with dissociation constants close to the marine
water pH. It may, therefore, be preferable to obtain realistic measurements by use of seawater
instead of deionised water. Another option is to measure the log Kow dependency of the pH
directly (cf. the new draft OECD guideline 122 Log Kow pH-metric method for ionisable
substances (OECD, 2000g). Because the pH of seawater (approximately 8) tends to be more
constant than that of freshwater, the procedure to correct partition coefficients for ionisable
substances, as described in Appendix XI, may however be considered sufficiently reliable for
marine conditions.
For inorganic chemicals such as metals, the form or speciation of the substance can be directly
affected by the ionic composition of seawater, which may have a considerable influence on both
solubility and partitioning. On a case-by-case basis, there may be sufficient information
available to allow the relevant partition coefficient in seawater to be calculated from the
freshwater data; otherwise, measurements under marine conditions may be necessary.
4.2.3 Marine degradation
4.2.3.1 Abiotic degradation

Abiotic degradation (i.e. hydrolysis and photolysis) in marine environments should be assessed
in a similar manner to abiotic degradation in freshwater environments except that the different
physico-chemical conditions in marine environments should be taken into account. In particular
the stable pH of about 8 and the generally lower temperature of in average 9C (282 K) should
be considered.
4.2.3.2 Biotic degradation

The rate of biodegradation in the various marine environments depends primarily on the
presence of competent degraders, the concentration and the intrinsic properties of the chemical
in question, the concentration of nutrients and organic matter and the presence of molecular
oxygen. These factors vary significantly between various marine environments.
In estuarine environments, the supply of xenobiotics, nutrients and organic matter is much higher
than in more distant marine environments. These factors enhance the probability that
biodegradation of xenobiotics occurs with a greater rate in estuaries than is the case in more
distant marine environments. Furthermore, estuarine and coastal environments are often
turbulent and characterised by a constant sedimentation and re-suspension of sediment particles
including microorganisms and nutrients, which increase the biodegradation potential in these
environments compared to marine environments with a greater water depth. The presence of
suspended particles and surfaces for attachment may favour the degradation of xenobiotics in
estuarine environments.
137
ECETOC (1993) reviewed existing biodegradation data for the marine environments. They
showed that the biodegradation in estuaries was approximately a factor 4 lower than in
freshwater environments for a variety of substances: Linear Alkylbenzene-Sulfonates, Linear
Alkyl-Ethoxylates, m-cresol, chlorobenzenes, p-nitrophenol glutamate, hexadecane, and
methylparathion. However, for substances known to be very rapidly biodegradable (such as
sodium acetate, sodium benzoate and sodium dodecylsulphate), the rates were similar in
estuarine and freshwater environments. In this section the average degradation potential in
estuarine environments is assumed to be similar to the degradation potential in freshwater
environments.
Further away from the land-based sources of xenobiotics and allochthonous material the
conditions for microorganisms are less favourable than close to land. The adaptation pressure is
low due to much lower concentrations of xenobiotics as a result of degradation and dilution.
Moreover, the environment can in general be characterised as oligotrophic, and the
concentrations of nutrients and organic matter are lower than in marine environments closer to
land. Because of their low concentrations, the xenobiotics are hardly degraded as primary
substrates, and due to the relatively low microbial activity the degradation of xenobiotics as
secondary substrates is assumed to be limited. This implies that the degradation potential in
distant marine environments is anticipated to be much lower than the degradation potential in
estuaries.
A special case is areas with offshore-based sources as, e.g., oil platforms. It may be assumed that
the microorganisms associated with the sediment may be more or less adapted to degradation of
chemicals that are continuously emitted from these sources. However, several factors, like e.g.
nutrient limitation, may limit the biodegradation potential compared to the situation close to
land. Furthermore, microorganisms in the water column will to a large extent drift with the
currents and, therefore, a development of stable communities of competent degraders is
impeded.
Most marine sediments are anaerobic below the upper 0-5 mm. The assessment of the
biodegradation in marine sediments should ideally be based on results from investigations
simulating these conditions. If not available, other approaches may be used, e.g.:
an approach similar to the one used for freshwater sediments could be used, i.e. to use a
scenario consisting of a 30 mm thick sediment layer of which the upper 3 mm are
considered aerobic and the remaining part anaerobic. If separate degradation rates are
available for aerobic and anaerobic sediment, these could be used for estimating the half-
life. If only data on aerobic degradation in sediment (or soil) is available, no degradation in
the anaerobic compartment should be assumed and consequently, a 10 times longer half-life
than the half-life in aerobic sediment (or soil) should be used.
anaerobic screening tests may be performed using a sediment inoculum (Horowitz et al.,
1982; Madsen et al., 1995), and the observed biodegradability may then be used as an
indication of the potential biodegradability of the substance in anaerobic sediment.
Degradation rates should be derived by expert judgement.
if no degradation data from studies with sediment or soil are available, the use of data on
degradation in water could be considered. The degradation potential in the upper aerobic
sediment layer is generally assumed to be similar to the degradation potential in the
overlying water. However, the possible very low bioavailability in the sediment of highly
hydrophobic and/or poorly water-soluble substances should be taken into consideration as is
done also for freshwater sediments.
138
4.2.3.3 Marine biodegradation simulation tests

As a general rule, degradation rates or half-lives determined in tests simulating the conditions in
the actual aquatic environment should be considered for use whenever available. Expert
judgement of the validity and quality of the test data is necessary. The origin (e.g. relevance of
sampling site) of the seawater/sediment inoculum shall always be evaluated in connection with
assessment and use of simulation test results. Biotransformation (identification of metabolisation
pathways and major metabolites) and mineralisation data may be derived from one of the
standardised simulation tests by using samples from the particular environment as inoculum.
Standardised simulation test methods for various marine compartments are:
Aquatic (pelagic) compartment: ISO/DIS 14592-1 Evaluation of the aerobic
biodegradability of organic compounds at low concentrations Part 1 (draft method 2001)
The ISO method has been the basis for a proposal for a new OECD guideline Simulation
test - Aerobic transformation in surface water (OECD, 2001d);
Turbid aquatic/sediment dispersed compartment: ISO/DIS 14592-2 Evaluation of the
aerobic biodegradability of organic compounds at low concentrations Part 2 (draft
method, 2001b) and OECD 308: Aerobic and anaerobic transformation in aquatic sediment
systems (aerobic test) (draft guideline, OECD, 2000c; draft Annex V C.24);
Anaerobic sediment compartment: OECD 308 Aerobic and anaerobic transformation in
aquatic sediment systems (strictly anaerobic test) (draft guideline, OECD, 2000c; draft
Annex V C.24). Data from anaerobic screening tests conducted with digested sewage sludge
(e.g. ISO 11734, 1994) cannot be used for predicting the degradation potential in sediments.
4.2.3.4 Use of biodegradation screening test data

For most chemicals, however, no test data from such simulation tests are yet available. For many
chemicals only data from screening tests are available. This may be data from marine
biodegradation screening tests or freshwater biodegradation screening tests. Marine screening
tests may be:
the OECD 306 Biodegradability in Seawater test (OECD, 1992e) comprises two methods,
the Shake Flask Method and the Closed Bottle Method. These tests are seawater variants of
the Modified OECD Screening Test (EU Annex V C.4-B and OECD 301E, 1992f) and
Closed Bottle Test (EU Annex V C.4-E and OECD 301D, 1992f), respectively, the main
difference being the use of a marine inoculum.
three additional screening tests were subjected for a ring test initiated by the OSPAR
Commission in 1995-96. The tests are the Marine CO2 Evolution Test, the Marine
BODIS Test and the Marine CO2 Headspace Test. The results of the ring test were
reported by Elf & IARE (1996).
When only results from marine or freshwater biodegradation screening tests are available, it is
recommended to use the default mineralisation half-lives for the pelagic compartment as
specified in Table 24.
Table 24 Recommended mineralisation half-lives (days) for use in marine risk assessment when only screening test data are
available
Freshwater 1) Estuaries 4) Other marine

environments 5)
Degradable in marine screening test N.a. 15 50
139
Readily degradable 2) 15 15 50
Readily degradable, but failing 10-d window 50 50 150
Inherently degradable 3) 150 150
Persistent
Notes to Table 24:

1) Half-lives from Table 7.
2) Pass level >70% DOC removal or > 60% ThOD in 28 days. Not applicable for freshwater.
3) A half-life of 150 days may be used only for those inherently degradable substances that are quickly mineralised in the MITI II or the
Zahn Wellens Test (cf. TGD Chapter 2.3.6). The half-life of 150 days is not fully scientifically justifiable (cf. TGD Chapter 2.3.6), but
reflects a guesstimate consensus between a number of experts.
4) Also including shallow marine water closest to the coastline
5) The half-lives mentioned under this heading are normally to be used in the regional assessment (coastal model) as described in
Section 4.2.5.
The half-lives for the marine environments that are described in Table 24 are provisional
recommendations, which should be reconsidered, when sufficient data for degradation of
different substances in screening tests and simulation tests have been evaluated. The basis for the
recommendation is the assumption that the degradation of xenobiotics in freshwater and
estuarine waters in general can be described by similar degradation rates, whereas the
degradation rates are lower in other marine environments more distant from the coastline (Here
the half-life is suggested to be increased by a factor of three relative to estuaries for readily
biodegradable substances and even more for more slowly degradable substances, see Table 24).
4.2.4 Local Assessment
4.2.4.1 Introduction
Usually releases to the environment stem from a point source leading to a locally high
environmental concentration of the substance. The highest risk resulting from discharges,
emissions and losses of a chemical into the environment is expected to be at this local scale close
to the point of emission. It should be recognised that this might not always be the case and that
other local high concentrations can arise some distance from the point of an emission due to
marine currents, transport and deposition of sediments etc. Where this is considered possible for
a local emission, specific modelling or measurements may be necessary. Since the aquatic
concentrations are highest at the point of emission, risks may be adequately assessed, at this
local scale, using the existing methodologies.
140
In addition to the inland sources of emission, there may also be direct discharges to the marine
environment. Thus, releases can occur from point sources:
to estuaries, either by direct discharges or from inland sources via riverine inputs (or both);
to coastal areas;
to harbour areas from port activity and shipping;
to open sea e.g. from offshore oil and gas installations and from ships;
atmospheric deposition.
4.2.4.2 Calculation of PEClocal for the aquatic compartment

In the current procedure of inland environmental risk assessment, the use of marine exposure
scenarios had become necessary whenever site-specific assessments were performed for a large
number of industrial sites, of which some actually discharge directly to the sea. A risk
assessment for the marine environment on a local scale was therefore only performed for specific
sites identified as releasing directly into the sea. In the context of a dedicated methodology for
marine risk assessment, a more generic exposure assessment for any given use is necessary.
While in some countries with long coastlines, the number of industrial sites discharging
wastewater to the sea is low compared with the overall number of sites (e.g. 5 10% in France;
IFEN, 1997), it can be very high in others (e.g. 58 % in Sweden; SCB, 2000). It is therefore
assumed that for all uses of a given chemical substance, potential local releases to the marine
environment can occur and, hence, it is necessary to perform a generic local exposure assessment
for the local marine environment.
As for inland risk assessment, the calculation of the PEClocal depends mainly on two
parameters: dilution and the presence (or absence) of a STP. Both of these parameters have large
influences on the local concentration (Clocalseawater).
Regarding the presence or absence of a STP, conflicting information is available. Experience
with the risk assessment of existing substances has shown that for chemical processing sites
located on the coast, the probability that the effluents are treated in a biological treatment plant is
much lower than for sites situated in land (see e.g., risk assessment reports for acrylonitrile,
cyclohexane or methylene dianiline). This is confirmed by a survey performed by HELCOM
(1998). While most industrial effluents from sites located on the Baltic Sea coast were treated
(up to 98 %), the report did not contain detailed information on the treatment used from all
contracting parties of HELCOM. However, from the data compiled in Sweden it appears that
less than 50% of the industrial wastewater discharged passes a biological treatment step. On the
other hand, statistics regarding treatment of municipal wastewater show that the treatment rate of
municipal wastewater from coastal municipalities is not different from overall treatment rates
(e.g. IFEN, 1997; HELCOM, 1998). On the other hand, four EU Member States have applied
Article 6 of Directive 91/271 allowing them to declare marine areas non sensitive to urban
wastewater meaning that they dont have to treat the wastewater biologically but only
mechanically.
It is therefore proposed, for a default assessment, that in a local setting, industrial effluents
(which may have been subject to some treatment on-site) are not treated in a municipal
biological STP. It is recognised though that the situation regarding the treatment of industrial
effluents is evolving rapidly and the present scenario could be revised in the near future. When
there is specific information available for a certain site that specific treatment facilities are
available this information needs to be assessed and can be used to override the default
141
assumption. In practice this information is often available for production and/or large processing
sites. It may also be possible to assume the presence of connection to an STP for certain industry
and/or use categories if appropriate justification about the general connection frequency to the
STP for that specific industry is provided. For releases to municipal wastewater of substances
that are used for private or public use (substances belonging to IC5 and IC6, Appendix I),
however, it can be assumed that the degree of treatment in a biological STP corresponds to the
inland scenario (see Section 2.3.7.1).
For discharges to a coastal zone, local dilution will be greater than in a freshwater river. First,
initial dilution may occur if the density between the effluent and the saline receiving medium
differs (Lewis, 1997). The initial dilution factor is usually around 10. Further dilution due to
currents can also be assumed, particularly if the point of release is subject to tidal influences. In
the Baltic or the Mediterranean sea, where there are almost no tidal influences compared to the
Atlantic Ocean or the North Sea, only initial dilution may occur on calm days, but normally,
further dilution due to currents is probable. Dilution factors of more than 500 have been
determined from model simulations (based on current measurements) in the North Sea, 200 m
away from the discharge point (e.g. Pedersen et al., 1994).
A dilution factor for discharges to a coastal zone of 100 may then tentatively be assumed, which
seems to be representative of a realistic worst case. The same estimation method as for inland
exposure assessment can then be used to obtain the local concentration in seawater (Clocalseawater,
see Section 2.3.8.3, equations 45-49).
In certain circumstances, it may be possible to identify specific emission points which would
allow the use of more precise information regarding the available distribution and fate processes.
Such site-specific assessments should only be used when it is known that all the emissions
emanating from the particular point in the life-cycle, e.g. manufacture, arise from a limited
number of specific and identifiable points. In these circumstances each specific point of release
will need to be assessed individually. If it is not possible to make this judgement, then the default
assumptions should be applied. In site-specific assessments, due account can be taken of the
true dilution available to the given emission as well as the impact of degradation, volatilisation,
etc. in the derivation of the PEC. Normally, only dilution and adsorption to suspended sediment
need be considered but site-specific conditions may indicate that valid local distribution models
can be used.
For estuaries, which are influenced by currents and tidal movements, it is assumed as a first
approach that they are covered by either the inland or the marine risk assessment. Thus, no
specific assessment is proposed.
Then, the local concentration in seawater can be obtained with:
Clocal eff
Clocal seawater =
( 1+ Kp susp SUSP water 10-6 ) DILUTION
(83)
142
Explanation of symbols
Clocaleff concentration of the substance in the STP effluent [mg.l-1] eq. (33)
Kpsusp solids-water partitioning coefficient of suspended matter [l.kg-1] eq. (24)
SUSPwater concentration of suspended matter in the seawater [mg.l-1] 15
DILUTION dilution factor [-] 100
Clocalseawater local concentration in seawater during emission episode [mg.l-1]
Kpsusp is derived as for inland risk assessment. For a specific estimation of the partitioning
behaviour of substances in saltwater environments see Section 4.2.2.
It is recognised that the dilution available to a discharge will also be related to the actual volume
of that discharge. In the freshwater scenario, this discharge volume is standardised to a volume
of 2,000 m3/day ie. the outflow from a standard STP. It is therefore proposed that the discharge
volume to the marine environment is also normalised at 2,000 m3/day such that the quantity of
the substance discharged (in kg/day) is assumed, for modelling purposes, to be diluted into this
volume prior to discharge.
For indirect human exposure and secondary poisoning, an annual average concentration in
surface water is calculated:
Temission
Clocal seawater,ann = Clocal seawater
365 (84)
Clocalseawater local concentration in seawater during emission episode [mg.l-1] eq. (83)
Temission number of days per year that the emission takes place [d.yr-1] App. IB
Clocalseawater,ann annual average local concentration in seawater [mg.l-1]
The concentration at the regional scale (PECregionalseawater) is used as background concentration

for the local scale. Therefore, these concentrations are summed:
PEClocal seawater = Clocal seawater + PECregional seawater (85)
PEClocal seawater,ann = Clocal seawater,ann + PECregional seawater (86)
Clocalseawater local concentration in seawater during episode [mg.l-1] eq. (83)

Clocalseawater,ann annual average concentration in seawater [mg.l-1] eq. (84)
PECregionalseawater regional concentration in seawater [mg.l-1] 4.2.5
PEClocalseawater predicted environmental concentration during episode [mg.l-1]
PEClocalseawater,ann annual average predicted environmental concentration [mg.l-1]
If relevant site-specific information is available, it can be used to improve the assessment. Some
significantly different exposure situations need to be reviewed though:
143
substances released from offshore platforms. A harmonised mandatory control system for
the use and reduction of the discharge of offshore chemicals is already agreed within
OSPAR (OSPAR, 2000a;2000b). For this specific exposure situation within the EU
legislation, the methodology proposed by OSPAR can be taken into consideration5;
substances released from harbours, marinas, fish farms and dry-docks. Specific scenarios
will have to be developed for these situations, which are most relevant for biocides.
4.2.4.3 Calculation of PEClocal for the sediment compartment.

The concentration in freshly deposited sediment is taken as the PEC for sediment; therefore the
properties of suspended matter are used. The concentration in bulk sediment can be derived from
the corresponding water body concentration, assuming a thermo-dynamic partitioning
equilibrium (Di Toro et al., 1991):
K susp water
PEClocal sed = PEClocal seawater 1000
RHOsusp
(87)
PEClocalseawater concentration in seawater during emission episode [mg.l-1]

Ksusp-water suspended matter-water partitioning coefficient [m3.m-3] eq. (24)
RHOsusp bulk density of suspended matter [kg.m-3] eq. (18)
PEClocalsed predicted environmental concentration in sediment [mg.kg-1]
Highly adsorptive substances may not be considered adequately with the approach described
above, as they are often not in equilibrium distribution between water and suspended matter
because of their cohesion to suspended matter; however they may be desorbed after ingestion by
benthic organisms.
Suspended matter exposed to local releases can subsequently be transported over long distances
and deposited to sediment in distant areas. Therefore, it is possible that areas unrelated to local
settings are exposed to the same sediment concentrations as would be expected only in the
immediate vicinity of the releases. This has especially to be taken into account when comparing
measured concentrations to estimated concentrations.
4.2.5 Regional assessment

For the release estimation of substances, a distinction is usually made between substances that
are emitted through point sources to which specific locations can be assigned, and substances
that enter the environment through diffuse releases.
Point source releases may have a major impact on the environmental concentration on a local
scale (PEClocal) and contribute to the environmental concentrations on a larger scale
(PECregional). Like with the freshwater environment for the marine situation it is necessary to
5
The methodology for assessing releases from platforms (e.g. CHARM-model) that has been developed in the
context of these OSPAR decisions was not re-discussed in the context of the development of the present
guidance document for marine risk assessment.
144
evaluate the impact of substances that are released from point and diffuse sources over a wider
area. The PECregional is supposed to take into account the further distribution and fate of a
chemical upon release. The resulting PECregional is assumed to be a steady-state concentration
of the substance.
The regional system for the freshwater environment is a relatively large area of 200 by 200 km
which consists of 97% of soil and 3% of water. This system is surrounded by a larger area of the
size of Europe, called the continent (see Sections 2.1.2 and 2.3.8.7). If for the marine region an
area of similar size would be chosen where the water of the freshwater region would enter into,
the resulting concentrations would be around 0.1% of the freshwater concentrations, mainly due
to the dilution of the freshwater in the much larger seawater region.
To assess the potential impacts of multiple point and diffuse sources of substances on the marine
environment a river plume in coastal sea water is considered as a marine regional generic
environment as follows:
An area of coastal sea that receives all
the water from the rivers from the
regional system. This seawater
exchange with global scales
compartment is exchanging chemical
with the continental seawater continent
compartment by dispersion and
advection (a current of seawater
flowing in a certain direction). The
size of the coastal compartment is region
40 km long, 10 km wide and 10 m
deep. In addition to the input from the dispersion reg. rivers
regional river water it receives 1% of
the direct emissions from the inland
sources which is supposed to
represent a relevant fraction of the advection
sources that are located near the sea cont. rivers
and also have direct emissions into
the sea compartment. Most of the
relevant characteristics of the coastal
compartment are similar to the Figure 15 Coastal sea scenario.
freshwater compartment apart from
the suspended matter concentration that is set to 5 mg/l. In the absence of specific information
(e.g. from marine simulation tests) it is assumed that the biodegradation rate in the water column
is approximately three times lower than in freshwater as described in Section 7.3. This scenario
is shown in Figure 15.
This scenario can be modelled with the multi-media fate model that is used for the freshwater
PEC calculations, modified to allow dispersive exchange between the coastal zone to the
continental sea water. By default, mixing of river water into the coastal sea gives a dilution
factor of approximately 10. As a result concentrations in coastal seawater are expected to be a
factor of 10 (for conservative chemicals) or more (for chemicals that react, volatilize or
sediment) lower than in river water. The extent of degradation, volatilization, etc. in this coastal
sea scenario is adequately modeled using the multi-media model.
145
More details on the features of these models can be found in the section on calculation of
PECregional for the freshwater environment (Section 2.3.8.7.)6.
The calculation of PECregional according to this scenario provides the results for the generic
risk assessment that is necessary for the risk evaluation for new and existing substances and
biocides. Sufficient information on sources and emissions and site-specific information on the
suspended matter concentration, the flow rate and the dispersion velocity may be available so the
generic assessment can be made more site-specific by overriding some of the default parameters
or even can be replaced by site-specific models. The dispersion velocity greatly affects all
calculated concentrations, while in addition the suspended matter content further affects the
dissolved concentration in seawater for chemicals with high log Kow. For the marine
environment, models are available that can be used to assess the concentrations in certain
specific compartments (bays, estuaries, regions) of the marine environment to which specific
industrial sites discharge wastewater.
4.3 MARINE EFFECTS ASSESSMENT
4.3.1 Effects Assessment for the aquatic compartment
Historically, the patterns of chemical production and usage resulting from urban and industrial
development have led to the freshwater environment being considered to be the hydrosphere
most at risk from these substances. Consequently, most regulatory schemes for evaluating the
hazards and risks posed by new and existing substances have focussed primarily on the
protection of freshwater communities. As a result there is a considerable body of data on the
ecotoxicity of chemical substances to freshwater organisms (ECETOC, 1994a)7.
Where there is a need to assess the potential impact of substances entering estuarine and marine
waters, any hazard or risk assessment should ideally be based upon data generated using a range
of ecologically relevant saltwater species (for example algae, invertebrates and fish). This is
particularly important given the greater diversity of species (particularly invertebrates) present in
6
A default length:width ratio of the coastal marine compartment has been set at 4:1. Assuming that this reflects the
plume shape in the generic assessment situation, this implies a ratio between the advective sea current along the
coast and the dispersive transport velocity perpendicular to that. If, in addition to the compartment dimensions, a
value is chosen for the sea current, the value of the lateral dispersion coefficient follows, or vice versa. If then a
value for the freshwater discharge into the coastal marine compartment is set too, mixing of freshwater with
coastal seawater is determined completely. In the generic regional model the river discharges approximately
1000 m3/s into the continental model. With the dimensions of the sea compartment set to 40,000 m.10,000 m.10 m,
and a suggested default value for the sea current of 0.03 m/s, taking into account the necessary dispersion coefficient
of 50 m2/s, the freshwater content of the sea water inside the selected box would become approximately 10%.
It should be noted that river water plumes in coastal waters vary greatly with local conditions (river flow, sea
current, tide, depth, etc.). Prediction of site-specific dilution of river water into coastal seawater requires site-
specific knowledge of flows and salinity distributions. Rhine and Meuse water (2,000 m3/s) are known to mix with
a sea current of 0.035 m/s in the southern North Sea, yielding a very long-streched plume with approximately 20%
river water in the first 10 km of the coast. A dispersion coefficient of 20 m2/s adequately describes this situation.
The Amazon river is known for its great plume.
7
The ECETOC database consists of 2,203 entries on 361 chemicals, covering 121 species. Data on freshwater
species accounted for 1862 entries (84.5%) while data for saltwater (estuarine/marine) species accounted for
341 entries (15.5%).
146
marine waters, relative to freshwaters (cf. Appendix XIV). There are also circumstances,
however, where the special conditions existing in a particular environment such as that existing
in the Baltic Sea, give rise to a reduced or limited species diversity and/or specific stresses such
as low or variable salinity. In such circumstances of low species diversity, adverse impacts in
individual species can have devastating impacts on the specialised ecosystem. Thus, while high
species diversity may lead to a wide sensitivity distribution, but also considerable functional
overlap, low species diversity may result in a lower sensitivity distribution but increase the
ecosystem function dependency on individual keystone species.
In both cases, the effects assessment must use, where possible, data relevant to the
environmental compartment that is considered. However, compared to the situation for
freshwaters, there are relatively few data on the effects of chemical substances on estuarine and
marine organisms. Therefore, in practice there will be situations where saltwater toxicity data are
needed for hazard/risk assessments, but may not be available. In these situations it may be
necessary to use freshwater data in lieu of data for estuarine/marine species (Schobben et al., 1994;
Karman et al., 1998). In using data on freshwater species to characterise the risk in the marine
waters, a clear understanding of the comparability of effects data generated on both types of species
is necessary. Furthermore, there is some evidence, e.g. for some metals, that species living in
brackish water are more susceptible because of the salinity (osmotic) stress they have to endure in
contrast to those of the same species living in truly marine conditions. Under these circumstances
the applicability of the toxicity data needs to be considered on a case-by-case basis.
4.3.1.2 Evaluation of data

It has been recognised for many years that there is a wider diversity of taxonomic groups
(particularly invertebrates) in saltwaters compared to freshwaters and that many groups are only
found in marine waters (see Russell and Yonge, 1928; Tait, 1978). Moss (1988) stated that
56 phyla were present in marine waters compared to 41 in freshwaters. No phyla are confined to
freshwaters only while 15 phyla are found only in marine waters. These differences are partly
due to the fact that multicellular animals originated in the seas and they have been well
populated since the earliest fossil records.
Nevertheless, an important part of any evaluation of data must involve an assessment of the
usefulness of the main body of freshwater ecotoxicity data in predicting effects in the marine
environment. Where such data can be used, the focus of further investigation can concentrate on
additional factors which specifically characterise the marine conditions. Studies conducted on
the comparability of sensitivity of freshwater and marine species have been hampered by the low
level of substances for which a comparable dataset has been available. Nevertheless where such
data are available, it has tended to show that there is no systematic bias in sensitivity where
comparable tests and endpoints are paired. A recent report which collated much of the available
data confirmed these findings (ECETOC, 2000). Based on the currently available data, it can be
concluded that:
overall, the data reviewed and current marine risk assessment practice suggest a reasonable
correlation between the ecotoxicological responses of freshwater and saltwater biota - at
least for the usual aquatic taxa (i.e., fish, crustacea, algae). No marked difference in
sensitivity between freshwater and saltwater biota appears that systematically applies across
all three trophic levels considered;
147
where evaluated, differences between trophic levels within each medium were generally as
significant or even more marked than between media. Such variation is implicitly assumed
in the use of assessment factors in current risk assessment practice;
where differences in the apparent sensitivity of freshwater and marine biota were observed
for individual compounds, such differences were consistently within a factor of 10 (<1 log
unit) and usually somewhat less;
average differences in sensitivity for such paired species comparisons were typically within
a factor of ~2;
however, within trophic levels differences larger than a factor of 10 were shown for several
metals and pesticides indicating that for these substances fresh water and saltwater data
should not be pooled for effects assessment and PNEC derivation.
The use of freshwater acute effects data in lieu of or in addition to saltwater effects data for risk
assessment purposes is not contra-indicated by the empirical data reviewed. Use of pooled data
is therefore recommended. Under such circumstances, PNEC values should be derived from the
most sensitive endpoint regardless of the medium.
No comparison of long-term effects data has been made due to the lack of suitable data but again
there are no reasons to believe that a systematic bias to freshwater or marine species would exist.
Therefore it is proposed that data on freshwater or marine fish, crustacea and algae be used
interchangeably for evaluation of the risks to either compartment.
4.3.1.3 Derivation of PNEC

The greater species diversity in the marine environment, compared to freshwaters (see
Appendix XIV), including the presence of a number of taxa that occur only in that environment,
may mean that the distribution of sensitivities of species is broader. It is necessary to consider,
therefore, whether the three-taxa model offers sufficient certainty that sensitive species will be
covered using the assessment factors developed for the freshwater systems. Since it is not
possible to make a clear judgement on the basis of available data, it is considered prudent to
assume that this greater diversity of taxa will produce a broader distribution of species
sensitivity. Thus, where only data for freshwater or saltwater algae, crustaceans and fish is
available a higher assessment factor than that for the derivation of PNECwater for freshwaters
should be applied, to reflect the greater uncertainty in the extrapolation. Where data is available
for additional taxonomic groups, for example rotifers, echinoderms or molluscs the uncertainties
in the extrapolation are reduced and the magnitude of the assessment factor applied to a dataset
can be lowered. Test protocols for these groups are available from organisations such as the
American Society for Testing and Materials, the International Council for the Exploration of the
Seas and the United States Environmental Protection Agency (OECD, 1998a). The assessment
factors given are based on current scientific understanding on the species comparability of toxicity
between freshwater and saltwater species and the issue of differences in diversity in freshwaters
and saltwaters. These may need to be revisited as additional information becomes available.
It is recognised that the assumption of a greater species sensitivity distribution covering the
additional marine taxa is based on limited data and is precautionary. The generation of additional
toxicity data on marine species may allow this assumption to be further refined such that lower
or higher assessment factors may be considered following a systematic review of accumulating
evidence.
148
The additional assessment factor is also considered sufficient to cover the situations noted above
where low species diversity may result in high ecosystem dependency on individual species.
The assessment factors decrease in magnitude from higher values for short-term acute studies
from which L(E)C50 values have been derived to lower values for long-term chronic studies
from which NOECs have been derived. For long-term studies the magnitude of the assessment
factors also decreases as information on a wider range of species becomes available. The
assessment factors described in Table 25 are those that would normally be applied to the
datasets available. There are some circumstances, however, where expert judgement may be
applied to the interpretation of a dataset which may allow a pragmatic approach to the
application of the factors and the generation of new data. In each case where expert judgement is
so applied, a full justification must be provided.
Table 25 Assessment factors proposed for deriving PNECwater for saltwater for different data sets
Data set Assessment factor

Lowest short-term L(E)C50 from freshwater or saltwater representatives of three taxonomic 10,000 a)
groups (algae, crustaceans and fish) of three trophic levels
Lowest short-term L(E)C50 from freshwater or saltwater representatives of three taxonomic 1000 b)
groups (algae, crustaceans and fish) of three trophic levels, + two additional marine
taxonomic groups (e.g. echinoderms, molluscs)
One long-term NOEC (from freshwater or saltwater crustacean reproduction or fish growth 1000 b)
studies)
Two long-term NOECs from freshwater or saltwater species representing two trophic levels 500 c)
(algae and/or crustaceans and/or fish)
Lowest long-term NOECs from three freshwater or saltwater species (normally algae and/or 100 d)
crustaceans and/or fish) representing three trophic levels
Two long-term NOECs from freshwater or saltwater species representing two trophic levels 50
(algae and/or crustaceans and/or fish) + one long-term NOEC from an additional marine
taxonomic group (e.g. echinoderms, molluscs)
Lowest long-term NOECs from three freshwater or saltwater species (normally algae and/or 10
crustaceans and/or fish) representing three trophic levels + two long-term NOECs from
additional marine taxonomic groups (e.g. echinoderms, molluscs)
Notes to Table 25:

Evidence for varying the assessment factor should in general include a consideration of the availability of data from a wider selection of species
covering additional feeding strategies/ life forms/ taxonomic groups other than those represented by the algal, crustacean and fish species
(such as echinoderms or molluscs). This is especially the case, where data are available for additional taxonomic groups representative of
marine species. More specific recommendations as with regard to issues to consider in relation to the data available and the size and variation
of the assessment factor are indicated below.
When substantiated evidence exists that the substances may be disrupting the endocrine system of mammals, birds, aquatic or other wildlife
species, it should be considered whether the assessment factor would also be sufficient to protect against effects caused by such a mode of
action, or whether an increase of the factor would be appropriate.
a)
The use of a factor of 10,000 on short-term toxicity data is a conservative and protective factor and is designed to ensure that substances with
the potential to cause adverse effects are identified in the effects assessment. It assumes that each of the identified uncertainties described
above makes a significant contribution to the overall uncertainty.
For any given substance there may be evidence that this is not so, or that one particular component of the uncertainty is more important than
any other. In these circumstances it may be necessary to vary this factor. This variation may lead to a raised or lowered assessment factor
depending on the evidence available. Except for substances with intermittent release, as defined in Section 2.3.3.4, under no circumstances
should a factor lower than 1000 be used in deriving a PNECwater for saltwater from short-term toxicity data.
Evidence for varying the assessment factor could include one or more of the following:
evidence from structurally similar compounds which may demonstrate that a higher or lower factor may be appropriate.
149
knowledge of the mode of action as some substances by virtue of their structure may be known to act in a non-specific manner. A lower
factor may therefore be considered. Equally a known specific mode of action may lead to a higher factor.
the availability of data from a variety of species covering the taxonomic groups of the base set species across at least three trophic
levels. In such a case the assessment factors may only be lowered if multiple data points are available for the most sensitive taxonomic
group (i.e. the group showing acute toxicity more than 10 times lower than for the other groups).
There are cases where a complete short-term dataset even for freshwater algal, crustacean and fish species will not be available, for example
for substances which are produced at < 1 t/a (notifications according to Annex VII B of Directive 92/32). In these situations, the only data may
be short-term L(E)C50 data for Daphnia. In these exceptional cases, the PNEC should be calculated with a factor of 10,000.
Variation from an assessment factor of 10000 should be fully reported with accompanying evidence.
b)
An assessment factor of 1000 applies where data from a wider selection of species are available covering additional taxonomic groups (such
as echinoderms or molluscs) other than those represented by algal, crustacean and fish species; if at least data are available for two additional
taxonomic groups representative of marine species.
An assessment factor of 1000 applies to a single long-term NOEC (freshwater or saltwater crustacean or fish) if this NOEC was generated for
the taxonomic group showing the lowest L(E)C50 in the short-term algal, crustacean or fish tests.
If the only available long-term NOEC is from a species which does not have the lowest L(E)C50 in the short-term tests, it cannot be regarded
as protective of other more sensitive species using the assessment factors available. Thus, the effects assessment is based on the short-term
data with an assessment factor of 10,000. However, normally the lowest PNEC should prevail.
An assessment factor of 1000 applies also to the lowest of the two long-term NOECs covering two trophic levels (freshwater or saltwater algae
and/or crustacean and/or fish) when such NOECs have not been generated for the species showing the lowest L(E)C50 of the short-term tests.
This should not apply in cases where the acutely most sensitive species has an L(E)C50-value lower than the lowest NOEC value. In such
cases the PNEC might be derived by applying an assessment factor of 1000 to the lowest L(E)C50 of the short-term tests.
c)
An assessment factor of 500 applies to the lowest of two NOECs covering two trophic levels (freshwater or saltwater algae and/or crustacean
and/or fish) when such NOECs have been generated covering those trophic levels showing the lowest L(E)C50 in the short-term tests with
these species. Consideration can be given to lowering this factor in the following circumstances:
It may sometimes be possible to determine with a high probability that the most sensitive species covering fish, crustacea and algae has
been examined, that is that a further longer-term NOEC from a third taxonomic group would not be lower than the data already available.
In such circumstances an assessment factor of 100 would be justified;
a reduced assessment factor (to 100 if only one short-term test, to 50 if two short-term tests on marine species are available) applied to
the lowest NOEC from only two species may be appropriate where:
short-term tests for additional species representing marine taxonomic groups (for example echinoderms or molluscs) have been
carried out and indicate that these are not the most sensitive group, and;
it has been determined with a high probability that long-term NOECs generated for these marine groups would not be lower than
that already obtained. This is particularly important if the substance does not have the potential to bioaccumulate.
An assessment factor of 500 also applies to the lowest of three NOECs covering three trophic levels, when such NOECs have not been
generated from the taxonomic group showing the lowest L(E)C50 in short-term tests. This should, however, not apply in the case where the
acutely most sensitive species has an L(E)C50 value lower than the lowest NOEC value. In such cases the PNEC might be derived by
applying an assessment factor of 1000 to the lowest L(E)C50 in the short-term tests.
d)
An assessment factor of 100 will be applied when longer-term toxicity NOECs are available from three freshwater or saltwater species (algae,
crustaceans and fish) across three trophic levels.
The assessment factor may be reduced to a minimum of 10 in the following situations:
where short-term tests for additional species representing marine taxonomic groups (for example echinoderms or molluscs) have been
carried out and indicate that these are not the most sensitive group, and it has been determined with a high probability that long-term
NOECs generated for these species would not be lower than that already obtained;
where short-term tests for additional taxonomic groups (for example echinoderms or molluscs) have indicated that one of these is the
most sensitive group acutely and a long-term test has been carried out for that species. This will only apply when it has been determined
with a high probability that additional NOECs generated from other taxa will not be lower than the NOECs already available.
A factor of 10 cannot be decreased on the basis of laboratory studies only.
150
Statistical extrapolation methods for calculation of PNEC for marine organisms could be used
when sufficient data are available. More information on these methods and the prerequisites to
apply them for risk assessment purposes can be found in Section 3.3.1.2.
4.3.2 Effects assessment for the sediment compartment
Substances that are highly hydrophobic may be assessed as of low risk for pelagic fauna but can
accumulate in sediments to concentrations at which they might exert significant toxic effects
(SETAC, 1993). This may be of concern particular in the marine environment, where the
sediment may act as a permanent sink for highly hydrophobic substances that can be
accumulated to a large extent. Because marine sediment constitutes an important compartment of
marine ecosystems it may be important to perform an effects assessment for the marine sediment
compartment for those substances.
In principle the same strategy as applied to freshwater sediment is recommended (see
Section 3.5) for the effects assessment of marine sediment). Several test methods on sediment
are developed and used in Member States of the European Union. Most of the tests are used for
sediment management purposes; only a few tests are conducted for risk assessment of
substances. An inventory of tests with marine organisms for the evaluation of dredged material
and sediments has been compiled by the Federal Environment Agency of Germany, UBA
(Herbst and Nendza, 2000). It comprises of biotests with various species of marine organisms of
different trophic levels on whole sediment, pore water or sediment extracts. In addition OECD
has prepared a detailed review paper on aquatic ecotoxicity tests including marine sediment test
methods (OECD, 1998a). Only whole sediment tests with infaunal and epibentic organisms are
considered suitable for being used in a risk assessment of the marine sediment compartment.
From examination of the UBA and OECD inventories it is clear that no fully internationally
accepted, standardised test methods for whole sediment are currently available.
Most of the existing whole sediment tests measure acute toxicity; only a few measure long-term,
sub-lethal endpoints. Only the latter tests are considered applicable to marine risk assessment
because of the long-term exposure of benthic organisms to sediment-bound substances that occur
under field conditions.
In Section 4.3.1 freshwater toxicity data are compared to marine and estuarine data. It is
concluded that the use of freshwater acute effects data in lieu or together with saltwater effects
data is acceptable for risk assessment purposes. Although it is not sure that this also applies to
marine and freshwater sediment data, it is nevertheless recommended to use pooled marine and
freshwater sediment toxicity data for effect assessment for the sediment compartment. However,
when sufficient data for ecologically relevant saltwater species are available lower assessment
factors can be applied.
4.3.2.2 Strategy for effects assessment for sediment organisms

Substances that are potentially capable of depositing on or sorbing to sediments to a significant
extent have to be assessed for toxicity to sediment-dwelling organisms. In addition, marine
sediment effects assessment is necessary for substances that are known to be persistent in marine
151
waters, and may accumulate in sediments over time. In general substances with a Koc < 500
1000 L/kg are not likely sorbed to sediment (SETAC, 1993). To avoid extensive testing of
chemicals a log Koc or log Kow of 3 can be used as a trigger value for sediment effects
assessment.
For most existing chemicals the number of toxicity data on infaunal and epibenthic organisms
will be limited. As a screening approach the equilibrium method can be used to compensate for
the lack of toxicity data if a PECmarine sediment can be determined on the basis of a measured
concentration of the substance in water that is independent of the value of the Koc. If the
PEC/PNEC determined using this method is > 1 then the need for testing with benthic organisms
using spiked sediment should be considered.
It is not necessary to apply the equilibrium partitioning method to predicted environmental
concentrations obtained from application of an exposure model when such a model will have
used the same Koc or log Kow value as that used to predict the PNECsediment. The reason is that
the resulting PEC/PNEC ratio for sediment will have the same value as for the water
compartment. In this case no quantitative risk characterisation for marine sediment should be
performed. Under these circumstances the assessment conducted for the aquatic compartment
will also cover the sediment compartment for chemicals with a log Kow up to 5. For substances
with a log Kow > 5 (or with a corresponding Koc), however, the PEC/PNEC ratio for the aquatic
compartment is increased by a factor of 10. The increased factor is justified by the fact that the
equilibrium partitioning method considers mainly the exposure via the water phase and does not
include that potential additional accumulation via sediment ingestion may occur for certain types
of sediment dwelling invertebrates (see Section 8.2.3).
Four situations can be distinguished for deriving a PNECsediment:
1. where only results from acute tests with benthic freshwater organisms are available (at least
one) the risk assessment is performed both on basis of the tests and on the basis of the
equilibrium partitioning method. The lowest PNECmarine sediment is then used for the risk
characterisation.
2. where, in addition to the tests with freshwater benthic organisms, an acute toxicity test is
performed with a marine benthic organism that is preferably representative of the same
taxon that is judged to be the most sensitive in the freshwater tests. Under these
circumstances an assessment factor of 1000 is applicable. A reduction of the assessment
factor is only justified if sufficient long-term tests with sediment-dwelling organisms are
available, and, if possible, where other evidence indicates that these tests include sensitive
taxonomic groups. Also in this case a comparison with the screening approach has to be
made and the lowest PNECsediment should be used for the risk characterisation.
3. where long-term toxicity data are available for benthic freshwater organisms. Under this
circumstance the PNECmarine sediment is calculated using assessment factors for long-term tests.
This approach is explained in Section 4.3.2.4.
4. where long-term toxicity data are available for benthic freshwater and a minimum of two
marine organisms. Under these circumstances a PNECmarine sediment is calculated using the
lower assessment factors that are associated with data obtained from long-term tests. A
PNECmarine sediment obtained from such data is preferred for risk assessment. This approach is
explained in Section 4.3.2.4.
Table 18 in Section 3.5.2 presents an overview of different data configurations and explains how
to use them for the risk characterisation for sediment. Attention should be paid to the fact that
very often contaminants are not analysed in whole sediment but in a certain fraction of the
152
sediment, for example in the sediment fraction of particles < 63 m. The organic carbon content
of this fraction is typically 15-30% for marine sediment while for whole marine sediments it is
generally less than 2%. It is important, for reasons of comparability of PEC and PNEC values,
that the organic carbon content of sediment used for toxicity tests are comparable with those of
actual marine sediments. If not there are likely to be concerns regarding the relative
bioavailability of a substance in the different sediments.
4.3.2.3 Calculations of PNEC for marine sediment using the equilibrium method
In the absence of any ecotoxicological data for sediment-dwelling organisms, but with measured
data to predict the PECmarine sediment, the PNECmarine sediment may provisionally be calculated using
the equilibrium partitioning method. This method uses the PNECsaltwater for aquatic organisms
and the marine suspended matter/water partitioning coefficient. The assumptions that are made
in this method are described in Section 3.5.3. Based on the equilibrium partitioning the following
equation is applied:
K susp water
PNECmarine se dim ent = PNECsaltwater 1000 (88)
RHOsusp
PNECsaltwater Predicted No Effect Concentration in saltwater [mg.l-1]

RHOsusp bulk density of suspended matter [kg.m-3] eq. (18)
Ksusp water partition coefficient suspended matter water [m3.m-3] eq. (24)
PNECmarine sediment Predicted No Effect Concentration in marine sediment [mg.kg-1]
In Section 3.5.2 a remark is made with respect to the calculation of PNECmarine sediment using the
equilibrium partitioning method. The equilibrium partitioning method considers uptake via the
water phase, while uptake may also occur via other exposure pathways such as ingestion of
sediment or direct contact with sediment. This may be important, especially for chemicals that
have a tendency to adsorb to sediment organic matter, for example those with a log Kow greater
than 3. Direct uptake from marine sediment is also observed in studies with marine benthic
organisms and may significantly contribute to the uptake of organic contaminants such as PAHs
(Kaag, 1998). There is also however evidence from studies in soil and in marine sediment that
the proportion of the total dose taken up through intake of sediment particles remains low for
chemicals with a log Kow up to 5. From other studies it is obvious that feeding mode also
influences uptake of substances (via water or ingestion of sediment). Furthermore the absorption
of contaminants in the gastrointestinal tract has been found to be increased compared with
absorption from the surrounding water (Mayer et al., 1996; Voparil and Mayer, 2000). However,
no quantitative conclusions can be drawn from these studies regarding uptake of substances from
sediment.
For substances with a log Kow greater than 5 (or with a corresponding Kpsed) the equilibrium
partitioning method is used in a modified way in order to take account of possible uptake via
ingestion of sediment. Thus the resulting PEC/PNEC ratio is increased by a factor of 10 for these
compounds. It should be borne in mind that this approach is considered as a screening level
assessment of the risk to sediment dwelling organisms. If with this method a PEC/PNEC ratio >
1 is derived then tests, preferably long-term, with benthic organisms using spiked sediment have
153
to be conducted in order for a realistic risk assessment appropriate to the sediment compartment
to be carried out.
4.3.2.4 Calculation of PNEC for marine sediment using assessment factors

If results from whole-sediment tests with benthic organisms are available the PNECmarine sediment
has to be derived using assessment factors. In establishing the size of the assessment factors, a
number of uncertainties have to be addressed (cf. Section 3.2). Due to the generally long-term
exposure of benthic organisms to sediment-bound substances, long-term tests with sub-lethal
endpoints like reproduction, growth, emergence, sediment avoidance and burrowing activity are
regarded as most relevant.
In contrast to the concept applied to the pelagic marine compartment, it is only necessary to have
results from one acute sediment test for the assessment factor of 10000 to apply. Furthermore if only
results from short-term tests with freshwater sediment-dwelling organisms are available (at least
one) an assessment factor of 10,000 is also applied to the lowest value. The PNECmarine sediment should
also be calculated from the PNECsaltwater using the equilibrium-partitioning method.
If, in addition to the results of tests with freshwater benthic organisms, a result from an acute
toxicity test with a marine benthic organism (preferably representative of the same taxa that is
most sensitive in aquatic freshwater or saltwater tests) is available then an assessment factor of
1000 is applicable. Once again a PNECmarine sediment should also be calculated from the
PNECsaltwater using the equilibrium partitioning method. A reduction of the assessment factor is
only permitted if results from long-term tests with sediment-dwelling organisms are available.
A PNECmarine sediment is derived by application of the following assessment factors to the lowest
LC50 value from acute tests:
Table 26 Assessment factors for derivation of PNECmarine sediment from short-term sediment toxicity tests
Available test results Assessment factor PNECmarine sediment

One acute freshwater or marine test 10,000 Lowest of LC50 /10,000 and equilibrium-
partitioning method
Two acute tests including a minimum of one marine 1000 Lowest of LC50 /1000 and equilibrium-
test with an organism of a sensitive taxa partitioning method
A PNECmarine sediment is derived by application of the following assessment factors to the lowest
NOEC/EC10 value from long-term tests:
154
Table 27 Assessment factors for derivation of PNECmarine sediment from long-term sediment toxicity tests
Available test results Assessment factor a)
One long-term freshwater sediment test 1000

Two long-term freshwater sediment tests with species representing different living and feeding conditions 500
One long-term freshwater and one saltwater sediment test representing different living and feeding 100
conditions
Three long-term sediment tests with species representing different living and feeding conditions 50
Three long-term tests with species representing different living and feeding conditions including a minimum 10
of two tests with marine species
a) The general principles of notes (c) and (d) as applied to data on aquatic organisms (Section 4.3.1.3) shall also apply to sediment data.
Additionally, where there is convincing evidence that the sensitivity of marine organisms is adequately covered by that available from
freshwater species, the assessment factors used for freshwater sediment data may be applied. Such evidence may include data from long-
term testing of freshwater and marine aquatic organisms, and must include data on specific marine taxa.
If no results from long-term tests with sediment organisms are available and the PEC/PNEC ratio
derived from the results of short-term sediment tests or via the equilibrium partitioning method
is a cause for concern then the need for long-term testing with sediment organisms should be
considered.
Since there are no chronic marine sediment test methods that are internationally accepted the
results from any tests should always be carefully evaluated. Several factors can contribute to
variability in test results. Of major importance to sediment tests are the effects of grain size and
organic carbon content of the sediment on the bioavailability of a substance. Sediment grain size
can also be an important factor in tests for other reasons. For example, the extent to which
bacteria can be adsorbed onto the sediment varies with particle size. Likewise, different species
of amphipods prefer sediments with different particle size distributions. No satisfactory solution
to the question which reference sediment should be considered appropriate is therefore currently
available. One should thus consider the tolerance of a given species with regard to the grain size
distribution of the sediments in question. Also spiking techniques have to be optimised because
often water is spiked after spiking the sediment. In addition, more insight is needed in the uptake
route of sediment bound contaminants in the organisms (exposure assessment).
Next to standardisation and test guidelines, it is necessary to further investigate the sensitivity,
reproducibility and inter-laboratory variability of the tests. It must be mentioned that most
available data on these facts concern the tests applied on field sediments, and not on spiked
sediments.
Examples of sub-chronic and chronic toxicity tests with whole sediment are given in Table 28.
Most of the tests have been developed for amphipods and polychaetes and some of them are
recommended by the OECD (1998a). There is a need for chronic tests to be developed for
Mollusca. Early life-stage tests with mussels and oysters are available for testing aqueous phases
but no standardised test is available for testing whole marine sediment samples. Chronic tests
that measure effects on community structure are also available but these tests seem to be very
insensitive. Functional endpoints tests, e.g. nutrient release rates, have been used to assess the
effects of contaminated sediments (Dahllff et al., 1999).
A final point that should be borne in mind is that single-species toxicity tests do not take account
of the interactions between the sediment inhabiting fauna and the fate or behaviour of chemical
substances, caused by e.g. bioturbation (Ciarelli et al., 1999; 2000). No procedures are currently
155
available for assessing the significance of such interactions but it is clear that they could be of
potential significance, particularly in respect of the bioavailability of a sediment contaminant.
Table 28 Acute and chronic whole sediment toxicity tests
Test organism Acute or Duration Endpoints Reference

chronic test
AMPHIPODS
Corophium sp. (C. Chronic 28d survival, ASTM (1993), Degrader. Organisms can be field
volutator or C. growth and Environment Canada collected. Cultivation causes
arenarium) reproduction (Burton, 1992), intermediate to high expenses
(OECD, 1998a Organism does not like coarse
recommended) sediment.
Low concern with regard to animal
welfare
Ecologically important organisms
relevance for exposed ecosystems
high.
SOP 1) available with field-collected
organisms.
Ringtested
Leptocheirus chronic 28 d survival, ASTM (1993), Degrader
plumulosus Environment Canada
growth and grain size has a significant effect on
(Burton, 1992),
reproduction survival, growth and reproduction.
US EPA (1996)
Survival is highest between 25%
clay and 75% sand.
Low concern with regard to animal
welfare Ecologically important
organisms relevance for exposed
ecosystems very high
SOP 1) available with field-collected
organisms.
Ringtested
POLYCHAETES
Nereis/Neanthes sp subacute/ 12 d - 28 d survival - ASTM (1994) Distributed widely throughout the

Neanthes chronic survival/growth world.
arenaceodentatakan
Can be cultivated on the laboratory
cultivated
degrader Low concern with regard
to animal welfare relevance for
exposed ecosystems very high.
SOP 1) available, equipment and test
species commercially available.
Ringtested.
Table 28 continued overleaf
156
Table 28 continued Acute and chronic whole sediment toxicity tests
Test organism Acute or Duration Endpoints Reference

chronic test
POLYCHAETES (continued)
Arenicola marina chronic 28 d Survival ASTM (1994) Degrader, wide tolerance of

sediment grain size. Organism is
(OECD, 1998a
found extensively over the OSPAR
recommended)
and Helsinki conventions area;
cultivation is difficult Low concern
with regard to animal welfare
relevance for exposed ecosystems
very high.
Ringtested.
Arenicola marina subacute 10 Casting rate Thain and Bifield (2001) see above row .
Changes in feeding rate have
consequences for sediment
communities.
OSPAR ringtested
ECHINODERMES
Echinocardium acute/ 14 d Survival Stronkhorst, in press Degrader, SOP 1) available with
cordatum subchronic field-collected organisms
(OECD, 1998a
recommended) Ringtested
MICROCOSM
Nematodes chronic 60 d community (Austen and
structure Somerfield, 1997)
1) Standard operating procedure
4.3.3 Assessment of secondary poisoning
The assessment of the potential impact of substances on top predators in the marine environment
can be based, in principle, on the same methodology as that used for a freshwater scenario. As
with freshwater ecosystems the accumulation of hydrophobic chemicals through the marine food
chains may follow many different pathways along different trophic levels. This accumulation
may result in toxic concentrations in predatory birds or mammals ingesting aquatic biota
containing the chemical. This effect is called secondary poisoning and should in principle be
assessed by comparing the measured or estimated concentrations in the tissues and organs of the
top predators with the no-effect concentrations for these predators expressed as the internal dose.
In practice, however, data on internal concentrations in wildlife animals are hardly ever available
157
and most no-effect levels are expressed in term of concentrations of the food that the organisms
consume (i.e. in mg.kg-1 food). Therefore, the actual assessment is normally based on a
comparison of the (predicted) concentration in the food of the top predator and the (predicted)
no-effect concentration which is based on studies with laboratory animals. A distinction is made
between the methodology used to assess the effects of substances whose effects can be related
directly to bioconcentration (direct uptake via water) and those where also indirect uptake via the
food may contribute significantly to the bioaccumulation.
Highly bioaccumulative substances have both a very high bioconcentration potential (log Kow
typically >4.5 or BCF > 500) and are also resistant to biotransformation in animals.
Biomagnification of such chemicals (increased food chain accumulation) is a major risk to the
top predators of food webs, as the consumption of contaminated food is a major source of
contaminants in predatory marine birds and mammals. In contrast the direct uptake of substances
from the environment (that is from water and sediment) is only of minor relevance (Biddinger
and Gloss, 1984; Opperhuizen, 1991). Factors that make these very hydrophobic substances of
particular concern to the marine environment include longer food chains, migratory and
reproductive aspects that may cause especially high exposure of progeny of marine species
likely, long-life of many marine predators, and a higher fat content. However, whilst steady state
levels in birds may be reached within weeks depending on the biological half-life of the
chemical (Pearce et al., 1989), contamination levels in mammals may continually increase with
age, with a plateau only being evident after several years (Thompson, 1990; Teigen et al., 1993).
No distinction can effectively be made between the spatial scales in the approach to the
assessment since the predators will take food from sources spread across local and regional
marine scenarios, as well as from the open sea. In the assessment it is therefore proposed to use a
PECsaltwater based on the mean of the local and regional concentrations for the assessment of the
local situation, and for the regional situation to apply a spatially broader scale. Given that marine
predators may have a wider range of foraging and that the regional sea concentrations will
normally be lower, this is considered as a reasonable worst-case assumption.
Bioaccumulation of metallic species is not considered explicitly in this section.
4.3.3.2 Assessment of bioaccumulation and secondary poisoning
The assessment scheme

The principal endpoints for the secondary poisoning assessment are the predators and top
predators that prey on organisms that are in direct contact with the marine aqueous phase and
receive the substances from this source. A relatively simple food chain is modelled which
consists of the marine water phase, marine food, marine fish and two separate levels of
predators. This food chain is visualised in Figure 16. As can be seen from this scheme risks for
three different trophic levels need to be assessed:
1. risks to marine fish: No specific calculation needs to be performed for estimating the risk to
marine fish as this is covered by the risk assessment for aquatic organisms.
2. risks to marine predators: The risk to marine predators is calculated as the ratio between the
concentration in their food (marine fish) and the no-effect concentration for oral intake
(PNECoralpredator). The concentration in the marine fish (Cfish) is obtained from
bioconcentration of the substance from the aqueous phase and (for very hydrophobic
substances) as a result of bioaccumulation from the food the fish consumes (which consists
158
of different types of aquatic organisms). Therefore, both a bioconcentration factor (BCF)

and a biomagnification factor (BMF1) are used to calculate Cfish. Note that for the BCFfish
also information for other organisms such as mussels may be considered.
3. risks to marine top predators: The risk to marine top-predators is calculated as the ratio
between the concentration in their food (marine predators) and the no-effect concentration
for oral intake (PNECoraltop predator). Since very hydrophobic substances may biomagnify in
the tissue and organs of the predator, for the calculation of the internal concentration of the
predator an additional biomagnification factor (BMF2) must be applied. Note that no
additional BMF factor for the top predator itself is required since the comparison between
PECoral and PNECoral is not based on internal concentrations but on intake rates.
Marine water Fish Predator

Cfish from Cpredator from
Top-predator
Marine food BCF and BMF 1 Cfish and BMF 2
Figure 16: Secondary poisoning food chain
It is realised that food chains of the marine environment can be very long and complex and may
consist of 5 or more trophic levels. The possible extent of bioaccumulation in marine food chains
with more than the above three to four trophic levels should be evaluated case by case if
necessary input data for such an evaluation is available, using the principles for the shorter food
chain. Also if further data are available it may be possible to refine the assessment of secondary
poisoning via marine food chains by employing more advanced modelling that takes the
differences in for instance uptake and metabolic rates into account for the different trophic
levels.
In the relatively simple food chain given above the concentration in the fish (i.e. the food for the
fish-eater) ideally should take account of all possible exposure routes, but in most instances this
will not be possible because it is not clear what contribution each potential exposure route makes
to the overall body burden of a contaminant in fish species. Therefore for very hydrophobic
substances a simple correction factor for potential biomagnification on top of the biocon-
centration through the water phase is applied.
Calculation of PEC in food of predators

The actual calculation of the concentration of a chemical in the food of the predators and top
predators will include the following steps:
PEC oral , predator = PEC seawater BCF fish BMF1

(89)
PEC oral ,toppredator = PEC oral , predator BMF2 = PEC water BCF fish BMF1 BMF2
(90)
159
PECoralpredator concentration in the food of the predator [mg.kg-1]

PECoraltop predator concentration in the food of the top predator [mg.kg-1]
PECseawater concentration in seawater [mg.l-1]
BCFfish bioconcentration factor [l.kg-1] eq. (73)
BMF1 biomagnification factor in fish [-] Table 29
BMF2 biomagnification factor in the predator [-] Table 29
The biomagnification factors used should, ideally, be based on measured values. However, the
limited availability of such data means that in most instances the default values described below
may have to be used. The use of a default value represents a screening approach designed to
identify substances for which it may be necessary to obtain more detailed information on the
biomagnification factor.
Although there may be relationships between the magnitude of the BMF and the log Kow of the
substance under defined conditions, the available data are not conclusive. Other more complex
intrinsic properties of substances than the lipophilicity (log Kow) seems to be important as well
as the species under consideration (e.g. its biology in relation to uptake, metabolism etc.). As a
simple screening approach, however, it seems reasonable to assume that for organic substances
with a log Kow up to 4.5 biomagnification seems generally to be low and thus BMF = 1. For
higher log Kow the biomagnification increases up to around log Kow 7 and then it decreases
again to be low around log Kow 9 (Fisk et al., 1998). Based on data published by Rasmussen et
al. (1990), Clark and Mackay (1991), Evans et al. (1991) and Fisk et al. (1998), the default BMF
values in Table 29 are suggested. If a BCF for fish is available, it is possible to use that as a
trigger instead of log Kow. The BCF triggers recommended are less conservative than the log
Kow triggers because they more realistically take the potential for metabolism in biota (i.e. fish)
into account. Due to this increased relevance, the use of BCF as a trigger would take precedence
over a trigger based on log Kow.
Table 29 Default BMF values for organic substances with different log Kow or BCF in fish
log Kow BCF (fish) BMF1 BMF2

<4.5 < 2,000 1 1
4.5 - < 5 2,000-5,000 2 2
58 > 5,000 10 10
>8 9 2,000-5,000 3 3
>9 < 2,000 1 1
The derivation of appropriate default BMFs can only, at this stage, be considered as preliminary
for use in screening of chemicals for the purposes of identifying those that need further scrutiny.
In reviewing the appropriateness of the BMF applied in any particular assessment, it should be
recognised that factors other than the log Kow and BCF should also be taken into account. Such
factors should include the available evidence that may indicate a potential for the substance to
metabolise or other evidence indicating a low potential for biomagnification. Evidence of a
potential for significant metabolism may include:
data from in vitro metabolism studies;
data from mammalian metabolism studies;
evidence of metabolism from structurally similar compounds;
160
a measured BCF significantly lower than predicted from the log Kow, indicating possible
metabolism.
Where evidence exists suggesting that such metabolism may occur, the BMF detailed above may
be reduced. Where such reductions are proposed, a detailed justification must be provided.
Application of different spatial scales

Apart from the fact that for the assessment of the risks to the top predator an additional
biomagnification factor is used the assessment also differs in terms of the input values that are
used for the seawater concentrations that lead to the concentrations in the food of the different
predators. For the first tier (or trophic level) of predators a worst-case assumption is that they
obtain their prey equally from the local and regional area, respectively. This is in line with the
assessment for freshwater and terrestrial organisms where a similar choice is made. For the
calculation of the PECoral for the predators this implies the following:
PEC seawater = 0.5 (PEClocal seawater ,ann + PECregional seawater )

(91)
When PECseawater is substituted in equation 89 this results in the following equation:
PEC oral , predator = (PEClocal seawater ,ann + PECregional seawater ) 0.5 BCF fish BMF1
(92)
PECoralpredator concentration in the food of the predator [mg.kg-1]

PECseawater concentration in seawater [mg.l-1]
BCFfish bioconcentration factor [l.kg-1] eq. (73)
BMF1 biomagnification factor in fish [-] Table 29
PECregionalseawater predicted environmental concentration in the region [mg.l-1]
PEClocalseawater,ann annual average predicted environmental concentration [mg.l-1]
For the second tier of organisms, the top predators, it can be assumed that they obtain their prey
mainly from the larger-scale regional marine environment which is to a lesser extent influenced
by point source discharges. However, since it cannot be ruled out that certain top predators prey
on organisms that receive their food from relatively small areas it is proposed to assume, as a
realistic worst case, a 90/10 ratio between regional and local food intake. For the calculation of
the oral intake rate for the top predator (PECoraltop predator) this implies:
PEC water = 0.1 PEClocal seawater ,ann + 0.9 PECregional seawater

(93)
When PECwater is substituted in equation 90 this results in the following equation:
PEC oral ,top predator = (0.1 PEClocal seawater ,ann + 0.9 PECregional seawater ) BCF fish BMF1 BMF2
(94)
161
Derivation of the PNECoral values

In the derivation of the PNECoral values only toxicity studies reporting on dietary and oral
exposure are relevant as the pathway for secondary poisoning refers exclusively to the uptake of
chemicals through the food chain. However, reliable toxicity data for predatory marine birds
(such as gulls and penguins) and mammals (such as seals, dolphins, whales and polar bears) are
extremely limited (Nendza et al., 1997). Furthermore, testing of such species would be ethically
unsound and contrary to animal welfare concerns. Therefore, it is necessary to extrapolate
threshold levels for marine species from terrestrial species assuming there are interspecies
correlations between laboratory bird species and marine predatory bird species, and between
laboratory mammals (e.g. rats) and the considerably larger marine predatory mammals. This
procedure is identical to that applicable for other media (see Section 3.8.3.5).
4.3.3.3 Testing strategy

If the PEC/PNEC ratio based on use of default BMF values indicates potential problems at any
trophic level it should first be considered whether a refinement of the PEC-assessment is
possible, i.e. the release and exposure assessment, including the fate related parameters such as
determination of log Kow or BCF. In special cases it may even be considered to start with
bioaccumulation studies in fish to determine the assimilation coefficient and the biological half-
life of the substance (i.e. to determine BMF1) prior to estimating or determining the
bioconcentration factor (BCF). Also a refinement of the PNECoral could be considered, i.e. to
require a long-term feeding study with laboratory mammals or birds to derive a more realistic
NOECoral value. In conducting such a study according to current test methods, it may in special
cases be considered whether to extend such studies to include satellite groups for determination
of the concentration of the substance in the animals during exposure (i.e. to measure
BMF2 values). Alternatively or supplementary to actual testing can be monitoring of biota for
which it is clear that they have lived in the environment that is covered in the assessment. Of
course no active sampling of (top)predators should be performed, but for instance animals that
are found dead can be used to get an indication about possible biomagnification factors in
wildlife. Useful information might also be obtained from eggs or from biopsies of skin or
blubber of marine birds or mammals.
4.4 PBT ASSESSMENT
4.4.1 Introduction
The PBT assessment is considered to be different from the local and regional assessment
approaches, as it seeks to protect ecosystems where the risks are more difficult to estimate.
These additional concerns for the marine environment, which may not be adequately addressed
by the traditional risk assessment methodologies, can be summarised as:
a. the concern that hazardous substances may accumulate in parts of the marine environment
and that:
(i) the effects of such accumulation are unpredictable in the long-term;
(ii) that such accumulation would be practically difficult to reverse;
162
b. the concern that remote areas of the oceans should remain untouched by hazardous
substances resulting from human activity, and that the intrinsic value of pristine
environments should be protected.
These concerns particularly occur with substances that can be shown both to persist for long
periods and bioaccumulate in biota, and can give rise to toxic effects after a greater time and at a
greater distance than chemicals without these properties. While this is also true for the
freshwater environment, the additional concern in the marine environment is that once the
chemical has entered the open seas, any cessation of emission will not necessarily result in a
reduction in chemical concentration and hence any effects become difficult to reverse. Equally,
because of the long-term exposures and long-life-cycle of many important marine species,
effects may be difficult to detect at an early stage.
For PBT substances a safe concentration in the environment cannot be established with sufficient
reliability. The PBT assessment is particularly developed to take into account the unacceptable high
uncertainty in predicting reliable exposure and/or effect concentrations hampering quantitative risk
assessment. The PBT assessment basically consists of two different steps:
identification of PBT substances using specific criteria for the inherent properties; and
an evaluation of the sources, major emissions and pathways to the marine environment to
sufficiently establish the most appropriate and effective measures to reduce the releases to
the marine environment.
The urgency and stringency of possible measures may, however, be dependent on the potential
of the substance to be transported to the open sea. This can be assessed qualitatively by
considering the use pattern, volumes and emissions or by using measured data. Open
applications and wide dispersive uses of the substance are regarded particularly relevant as well
as non-minimised direct discharges from production, formulation and industrial use.
4.4.2 PBT criteria

The criteria to be used to decide if a substance must be regarded as a PBT substance are
summarised in Table 30 below. The testing strategies to obtain the data that are necessary to
decide whether a substance fulfils these criteria are given in separate sections on persistence,
bioaccumulation and toxicity. The table contains two sets of criteria, one for PBT substances and
a second category for so-called very persistent and very bioaccumulating substances (vPvB).
This second category is developed under the recognition that for substances that are very
persistent and bioaccumulate significantly in the food chain, high but unpredictable levels may
be reached in wildlife or man over extended time periods. For such substances it is not necessary
to demonstrate toxicity in laboratory testing as long-term effects can be anticipated anyway.
For most substances the available data will not allow to come to a definitive answer to the
question if the substance must be considered under the PBT assessment. Hence screening data
that identify whether the substance has a potential to be a PBT have to be made use of. The
testing strategies in the following paragraphs should be followed and further information should
be asked for accordingly. In deciding which information is requested (on P, B or T) care must be
taken to avoid animal testing where possible. This implies that when for several properties
further information is needed the assessment should be focussed on clarifying the potential for
persistence first. When it is clear that the P criterion is fulfilled a stepwise approach should be
followed to elucidate the B criterion, eventually followed by toxicity testing to clarify the T
criterion.
163
Table 30 Criteria for identification of PBT and vPvB substances
Criterion PBT criteria vPvB-criteria

P Half-life > 60 d in marine water or > 40 d in freshwater* Half-life > 60 d in marine- or freshwater or >180 d in
or half-life > 180 d in marine sediment or > 120 d in marine or freshwater sediment
freshwater sediment*
B BCF > 2,000 BCF > 5,000
T Chronic NOEC < 0.01 mg/l or CMR or endocrine Not applicable
disrupting effects
* For the purpose of marine environmental risk assessment half-life data in freshwater and freshwater sediment can be overruled by data
obtained under marine conditions.
In principle, substances are selected when they fulfil the criteria for all three inherent properties
P, B and T. However, certain flexibility is required in their application for instance in cases
where one criterion is marginally not fulfilled but the others are exceeded considerably. This
may include for example substances that do not fulfil the persistence criteria but bioaccumulate
significantly and are measured in marine biota distant from anthropogenic sources.
It is realized that the individual trigger values may be scientifically disputable when considered
in isolation. However, by applying the combined set of criteria it is expected that substances will
be selected for which quantification of the risk by using the PEC/PNEC approach is considered
too uncertain.
The PBT assessment has links to similar concepts discussed in other fora (e.g. the UNEP
Stockholm Convention on Persistent Organic Pollutants, the OSPAR Hazardous Substances
Strategy (OSPAR, 1998)). The discussions from the other fora have been carefully considered.
4.4.3 Testing strategy for the P criterion
The persistence of a substance reflects the potential for long-term exposure of organisms but also
the potential for the substance to reach the marine environment and to be transported to remote
areas. The assessment of the (potential for) persistency in the marine environment should in
principle be based on actual half-life data determined under marine environmental conditions.
Depending on whether a substance has a half-life smaller or greater than the cut-off criterion it is
decided if a substance fulfils the P criterion. When these key data are not available other types of
available information on the degradability of a substance can be used to decide if further testing
is needed to assess the potential persistence. In this approach three different levels of information
are defined according to their perceived relevance to the criteria:
experimental data on persistence in the marine environment;
other experimental data;
data from biodegradation estimation models.
An explanation on what type of information is relevant within these levels and the relevant cut-
off values is given below. It must be noted that this approach reflects existing knowledge on
biodegradation and should be considered as a pragmatic approach to make optimal use of the
available data and methods. Clearly, more research is needed to better estimate the persistence in
the marine environment from existing biodegradation tests. Moreover, other degradation
164
mechanisms such as hydrolysis and photolysis should be taken into account where they can be
shown to be relevant.
4.4.3.2 Experimental data on persistence in the marine environment

In principle the persistence in the marine environment should be assessed in simulation test
systems that determine the half-life under relevant environmental conditions. The determination
of the half-life should include assessment of metabolites with PBT-characteristics. The half-life
should be used as the first and main criterion in order to determine whether substances should be
regarded as persistent. Hence appropriate half-life data from valid simulation tests override data
from the other levels of information. Substances with a half-life in marine water of > 60 days or
a half-life of >180 days in marine sediment are considered as being persistent. Substances are
considered very persistent (vP) when the half-life in marine or freshwater is > 60 days or the
half-life in marine or freshwater sediments is > 180 days.
Recommended simulation test methods for water and sediment are described in Section 4.2.3.3.
Tests performed under marine conditions should use media from marine areas not directly
influenced by freshwater outlets or runoffs. This means that samples taken for inoculation and
conduction of marine biodegradation simulation tests must not contain significant amounts of
freshwater microorganisms as these to a larger extent could already have been pre-exposed or
adapted to the substance. It is not possible to establish specific criteria and each test must be
evaluated case-by-case. However, the content of freshwater in the sample should be low (i.e. a
large dilution as e.g. determined by salinity), the sample should be taken from the water column
(and not the surface), the content of microorganisms should be low (compared to freshwater) and
cross-contamination during handling, transport and testing should be avoided.
4.4.3.3 Other experimental data

In case no half-life data are available for marine water or sediment the decision whether a
substance is potentially persistent needs to be based on other experimental data. If available, use
can be made of the half-life values from simulation tests of degradation in freshwater. Since the
degradation in marine waters other than estuaries is expected to be slower than in freshwater a
criterion of >40 days for freshwater and >120 days for freshwater sediments has to be used.
Where no data are available which allow the assignment of a degradation half-life in the
environment based on simulation test data, other experimental data may be considered. The
available information relating to biodegradability is dominated by test results on Ready
Biodegradability (EU Annex V C.4 A, C, E and F; OECD guidelines 301A-D, 1992f, 306, 1992e
or equivalent) and to a lesser extent by data on the Inherent Biodegradability (EU Annex V C.12
and C.9; OECD TG A-C, 1981d or equivalent). The conditions for degradation in the marine
environment are far from the conditions applied in these standard tests (cf. Section 4.2.3). Hence,
extrapolation of the existing biodegradation information (either measured data from ready and
inherent tests or results from QSAR modelling) to degradation rates in the marine environment is
very difficult and care should be taken not to over-interpret the outcome of such tests. However,
in order to use the available information to select potentially persistent substances, this
information should be used in the following way:
readily biodegradable substances (fulfilling or not fulfilling the 10-day window criterion)
are considered as not persistent in the PBT assessment;
165
when a substance does not fulfil the criteria for ready biodegradability as defined in sections
on biodegradation and for the marine ready test (see Sections 2.3.6 and 4.2.3.4), it is
considered as being potentially persistent. The 10-day window criterion should not be used
here as an additional criterion. If the substance fulfils the criteria for B and T, further testing
is needed. It must be noted that in this case it is not considered appropriate to perform
inherent biodegradability tests but rather to go directly to simulation testing;
when results are available showing that a substance does not fulfil the criteria for inherent
biodegradability as defined within the Annex V method or the OECD guideline this is a
clear indication that the substance will not biodegrade in the marine environment either. The
substance will be regarded as potentially persistent. When the (screening) criteria for B and
T are also fulfilled, further testing is needed in order to determine the half-life in the
environment.
when a substance passes the criteria for inherent biodegradability tests this does not
necessarily indicate that it will not be persistent under environmental conditions. However,
in order to make the best use of available information it is accepted to use the results of two
specific tests when they fulfil certain criteria as an indication that the substance is not
persistent. These test are:
- Zahn-Wellens Test (EU Annex V C.9, OECD 302B, 1992g): Pass level (70%
mineralisation) must be reached within 7 days, log-phase should be no longer than 3
days, percentage removal in the test before degradation occurs should be below 15%,
not tested with pre-adapted microorganisms;
- MITI II -test (OECD 302C, 1981d): Pass level must be reached within 14 days; log-
phase should be no longer than 3 days, not tested with pre-adapted microorganisms.
In case a range of biodegradation data, including conflicting data, is available, a case-by-case
assessment is needed, using a weight of evidence approach, in order to decide whether a
substance has the potential to be persistent (see also Section 2.3.6).
4.4.3.4 Data from biodegradation estimation models

For new substances, priority existing substances and biocides, information from a ready
biodegradability test is normally available and therefore an initial decision whether the substance
is potentially persistent can be taken. However, for many other substances no data will be
available or the available information is difficult to interpret. For these substances it can be
helpful to apply models that estimate the potential for biodegradation in the environment.
In a preliminary assessment whether a substance has a potential for persistence in the marine
environment and hence for asking for actual test data it is proposed to consider use of the
BIOWIN program. This program estimates aerobic biodegradability of organic chemicals using
6 different models:
1. linear model,
2. non-linear model,
3. ultimate biodegradability timeframe model,
4. primary biodegradability timeframe model,
5. MITI linear model,
6. MITI non-linear model;
166
The use of the results of these programs in a conservative way may fulfil the needs for
evaluating the potential for persistency. The use of three out of the six models is suggested as
follows:
non-linear model prediction: does not biodegrade fast (<0.5) or
MITI non-linear model prediction: not readily degradable (<0.5) and
ultimate biodegradation timeframe prediction: > months (<2.2)
When predictions of these three models are combined relatively few not readily biodegradable
substances will not be identified, without in the same time causing a significant increase in the
number of falsely included readily biodegradable substances.
The preliminary character of this method to identify potentially persistent substances in the
marine environment is emphasised, and further possible development of a suitable methodology
is recommended. The BIOWIN program is available from the US EPA's internet site
(http://www.epa.gov/oppt/exposure/docs/episuitedl.htm).
4.4.3.5 Summary of the P assessment

A summary of the assessment of biodegradation data in the context of the P criterion is given in
Table 31 below.
167
Table 31 Overview of P-assignment for different types of biodegradation data
Type of data Criterion Definitive assignment Screening assignment 1)

DT50 marine water > 60 d VP -
DT50 freshwater 2) > 40 d P 3) -
> 60 d VP -
DT50 marine sediment > 180 d VP -
DT50 freshwater sediment 2) > 120 d P 3) -
> 180 d VP -
Readily biodegradable 4) Yes Not P -
No - P or vP
Inherently degradable Yes Not P 5) -
No - P or vP
QSAR Non-linear model prediction < - P or vP
0.5 or MITI non-linear model
prediction < 0.5 and Ultimate
biodegradation timeframe
prediction < 2.2
1) These screening methods give an open-ended categorisation of the substance as either being potentially P or vP, which cannot easily
be related to a half-life for biodegradation.
2) Data for estuaries should also be considered in this category.
3) Half-life data in freshwater and freshwater sediment can be overruled by data obtained under marine conditions.
4) Regardless of whether the 10-d window criterion is fulfilled.
5) This only applies to cases where the specific criteria as mentioned in Section 4.4.3.3 are fulfilled.
4.4.4 Testing strategy for the B criterion
Substances can accumulate in aquatic organisms directly from the water, i.e. bioconcentration, or
via uptake through the foodchain, i.e. biomagnification. A high bioaccumulation potential of a
substance is of particular concern for the marine environment due to the possible accumulation
in the foodchains and the potential long-term effects that may occur in organisms at the top of
these foodchains. Whereas different models and parameters are available to evaluate
bioconcentration for organic chemicals, suitable parameters to evaluate accumulation in marine
foodchains are not available. The bioconcentration factor (BCF) in aquatic organisms is
traditionally used as a first indicator for bioaccumulation (see Section 3.8.2).
In principle, the assessment of the (potential for) bioaccumulation in the context of the PBT
assessment makes use of measured bioconcentration factors in marine or freshwater organisms.
Where these are not available BCF values may be estimated from the octanol/water partition
coefficient (Kow) using QSAR models. In addition, Kow values, either experimentally
determined or estimated can be used directly to assess the potential for bioaccumulation.
168
4.4.4.2 Assessment of measured BCF data

The decision whether or not a substance fulfils the B criterion should in principle be based on
measured data on bioconcentration in aquatic species. Data from freshwater as well as marine
species can be used. Extensive guidance on the quality assessment of such data can be found
elsewhere (e.g. OECD, 2001c). A substance is considered to fulfil the B criterion when the
measured BCF on a wet weight basis exceeds a value of 2,000. A substance is considered very
bioaccumulative (vB) when the BCF exceeds a value of 5,000.
4.4.4.3 Assessment of the potential for bioaccumulation

When measured BCF values are not available the Kow or the BCF based on modelling can be
used to indicate the liability to bioaccumulate from water. For substances with log Kow < 6
assessment on the basis of Kow or estimated BCF does not make a real difference since all
available BCF models are linear (see Section 3.8.3.2). The B criterion for log Kow is therefore
directly derived from this linear relationship. A substance is considered to potentially fulfil the B
criterion when log Kow exceeds a value of 4.5.
For highly hydrophobic substances, e.g. with log Kow > 6, experimentally derived BCF values
tend to decrease with increasing log Kow. Several conceptual explanations as well as
explanations referring to experimental artefacts can be given for this decline. For these
substances the available BCF models can lead to very different results. As a consequence the
potential for bioaccumulation is assessed by expert judgement on the basis of the log Kow value
and the estimated BCF using the available BCF models. Such an assessment must be done on a
case-by-case basis taking into account what is known about the BCF QSAR-models and the
specific properties of the substance, in particular what is known to affect uptake and the potential
for metabolism in aquatic organisms. Care must be taken that substances with high log Kow
values are not deleted from selection processes without applying expert judgement to them.
It must be noted that for priority existing substances, new substances and biocides a measured
octanol/water partition coefficient is usually available. Additionally a range of QSAR models
can be used to estimate this parameter (see e.g. Chapter 4, Meylan et al., 1999; OECD, 2001c).
4.4.4.4 Other information relevant for assessment of the B criterion

In addition to the above-mentioned data on bioconcentration or bioaccumulation in aquatic
species evidence that a substance shows high bioaccumulation in other species may also be used
to decide whether the B criterion is fulfilled. Such evidence may be based on information from
specific laboratory tests or from field studies. Specific attention needs to be paid to measured
data in biota. Measured data in biota are a clear indicator that a substance is taken up by an
organism. However, they are not an indicator that significant bioconcentration or
bioaccumulation has occurred. The interpretation of such data in terms of actual bioaccumulation
or biomagnification factors can especially be difficult when the sources and levels of the
exposure (through water as well as through food) are not known or cannot be estimated
reasonably.
169
4.4.5 Testing strategy for the T criterion
For persistent and bioaccumulative substances long-term exposure can be anticipated and
expected to cover the whole life-time of an organism and even multiple generations. Therefore
chronic or long-term ecotoxicity data, ideally covering the reproductive stages should in
principle be used for the assessment of the T criterion. In practice, however, the principal data
available for most chemicals will be for short-term effects, and this must, in the first instance, be
used to drive initial selection. Mammalian toxicity data must also be considered in the selection
due to the fact that toxic effects on top predators, including man may occur through long-term
exposure via the food-chain. The selection criteria should therefore consider two types of effect
data, either of which will trigger selection.
4.4.5.2 Chronic effects data

A substance is considered to fulfil the toxicity criterion when:
the long-term NOEC for marine or freshwater organisms is less than 0.01 mg/l. When other
information is available such as data on sediment toxicity or data from feeding studies this
needs to be assessed on case-by-case basis. For biocides and pesticides results from sub-
chronic, chronic or reproduction avian toxicity tests may be available. A chronic NOEC of
less than 30 mg/kg/food can be used as a criterion.
or
when the substance is classified as Carcinogenic, (category 1 & 2), Mutagenic (category 1 &
2) or Toxic for Reproduction (category 1, 2 & 3) or when there is evidence of chronic
toxicity, as identified by the classifications T, R45, R46, R48, R60 and R61 or Xn, R48,
R62, R63 and R64 8.
when a substance is classified as Carcinogenic category 3 or Mutagenic category 3 a case-
by-case assessment must be carried out to decide whether the evidence is sufficient for the
substance to be considered as toxic, in the context of this PBT assessment, or whether
further information is needed to clarify this potential concern.
or
when there is substantiated evidence of long-term toxicity (e.g. endocrine disrupting
effects). Such evidence needs to be considered on a case-by-case basis.
4.4.5.3 Acute effects data (screening level)

Where data on chronic effects are not available short-term toxicity data for marine or freshwater
organisms can be used to determine whether a substance is a potential PBT provided the
screening criteria for P and B are fulfilled. In the context of the PBT assessment a substance is
considered to be potentially toxic when the L(E)C50 to aquatic organisms is less than 0.1 mg/l.
If a substance is confirmed to fulfil the ultimate P and B criteria chronic toxicity data are
8 In relation to the use of R64 in the context of the PBT assessment care should be taken that the actual assignment
of the R-phrase is a result of results of one or two generation studies in animals which indicate the presence of
adverse effects on the offspring due to transfer in the milk (see Annex VI to Directive 67/548).
170
required to deselect this substance from being considered as a PBT. In principle chronic toxicity
data, when obtained for the same species, should override the results from the acute tests.
In the context of the PBT assessment acute mammalian toxicity tests are normally not considered to
provide an appropriate indication of chronic effects. However, it should be noted that when a
substance is classified as Very Toxic or Toxic after oral dosing (LD50 < 200 mg/kg bw/d) and the
toxicity is expected to be the result of systemic effects, the probability that the chronic NOAEL
after repeated dosing (e.g. 28 d or 90 d) will be less than the trigger value for R48 ( 150 or
50 mg/kg bw/d, respectively) will be high. The substance would therefore be classified and
considered as fulfilling the T criterion. In that case verification of the actual chronic toxicity by
performing animal testing is not recommended. When the P and B screening criteria are also
fulfilled the substance can be considered as a PBT unless additional information indicates
otherwise.
4.4.5.4 Estimated effects data

In case where no acute or chronic toxicity data are available the assessment of the T criterion at a
screening level can be performed using data obtained from quantitative structure activity
relationships (QSARs). Guidance on the use of QSARs for specific groups of substances can be
found in Chapter 4.
It must be noted that since long-term effects can be anticipated for very bioaccumulative
substances (vPvB), further animal testing for such substances is deemed unnecessary.
171

PNEC Calculations

Uploaded by

Copyright:

Available Formats

PNEC Calculations

Uploaded by

Document Information

Original Description:

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

PNEC Calculations

Uploaded by

Copyright:

Available Formats

MARINE RISK ASSESSMENT

4 ENVIRONMENTAL RISK ASSESSMENT MARINE

environment by examining both the exposures resulting from discharges/releases of chemicals

4.2 MARINE EXPOSURE ASSESSMENT

4.2.1 Measured data

4.2.2 Partition coefficients

4.2.3 Marine degradation

4.2.3.1 Abiotic degradation

4.2.3.2 Biotic degradation

4.2.3.3 Marine biodegradation simulation tests

4.2.3.4 Use of biodegradation screening test data

Freshwater 1) Estuaries 4) Other marine

Notes to Table 24:

4.2.4 Local Assessment

4.2.4.2 Calculation of PEClocal for the aquatic compartment

The concentration at the regional scale (PECregionalseawater) is used as background concentration

PEClocal seawater = Clocal seawater + PECregional seawater (85)

PEClocal seawater,ann = Clocal seawater,ann + PECregional seawater (86)

Clocalseawater local concentration in seawater during episode [mg.l-1] eq. (83)

4.2.4.3 Calculation of PEClocal for the sediment compartment.

PEClocalseawater concentration in seawater during emission episode [mg.l-1]

4.2.5 Regional assessment

4.3 MARINE EFFECTS ASSESSMENT

4.3.1 Effects Assessment for the aquatic compartment

4.3.1.2 Evaluation of data

4.3.1.3 Derivation of PNEC

Data set Assessment factor

Notes to Table 25:

4.3.2 Effects assessment for the sediment compartment

4.3.2.2 Strategy for effects assessment for sediment organisms

PNECsaltwater Predicted No Effect Concentration in saltwater [mg.l-1]

4.3.2.4 Calculation of PNEC for marine sediment using assessment factors

Available test results Assessment factor PNECmarine sediment

Available test results Assessment factor a)

One long-term freshwater sediment test 1000

Table 28 Acute and chronic whole sediment toxicity tests

Test organism Acute or Duration Endpoints Reference

Nereis/Neanthes sp subacute/ 12 d - 28 d survival - ASTM (1994) Distributed widely throughout the

Table 28 continued overleaf

Table 28 continued Acute and chronic whole sediment toxicity tests

Test organism Acute or Duration Endpoints Reference

Arenicola marina chronic 28 d Survival ASTM (1994) Degrader, wide tolerance of

1) Standard operating procedure

4.3.3 Assessment of secondary poisoning

4.3.3.2 Assessment of bioaccumulation and secondary poisoning

The assessment scheme

of different types of aquatic organisms). Therefore, both a bioconcentration factor (BCF)

Marine water Fish Predator

Figure 16: Secondary poisoning food chain

Calculation of PEC in food of predators

PEC oral , predator = PEC seawater BCF fish BMF1

PECoralpredator concentration in the food of the predator [mg.kg-1]

log Kow BCF (fish) BMF1 BMF2

Application of different spatial scales

PEC seawater = 0.5 (PEClocal seawater ,ann + PECregional seawater )

When PECseawater is substituted in equation 89 this results in the following equation:

PECoralpredator concentration in the food of the predator [mg.kg-1]

PEC water = 0.1 PEClocal seawater ,ann + 0.9 PECregional seawater

When PECwater is substituted in equation 90 this results in the following equation: