Technical Report

TR 082 – Risk Assessment in Marine Environments

TR 082 : Risk Assessment in Marine Environments | December 2001

Within Europe, risk assessments for new and existing substances for the freshwater aquatic environment are based on the standardised procedures embodied in the Technical Guidance Document (TGD) and the EUSES model. These enable PECs to be determined for generic environments covering a range of spatial scales, when suitable measured data or site-specific models are not available or not applicable. Methods to address risk assessment procedures for marine environments are currently being developed and it is desirable to define marine assessment schemes which parallel and complement this freshwater approach.

The extension of the well-established principles of the risk assessment framework elaborated in the TGD (i.e. exposure assessment, hazard identification, risk characterisation, risk characterisation revision/reiteration, risk reduction, risk acceptance) should provide a consistent approach which may be readily implemented and accepted and which can be used to address issues surrounding the impact of anthropogenic emissions to the marine environment. Additional concerns associated with the marine environment, i.e. secondary poisoning of marine predators, need to be accommodated within this framework. To avoid confusion, in all cases a PEC/PNEC ratio of greater than or equal to 1 should be adopted as an indicator of unacceptable risk. Any precautionary allowances for uncertainty should be applied to either the PEC or PNEC, as appropriate.

In certain instances, some or all of a marine risk assessment will be site specific (relating to a known geographical area) and in such cases monitoring data representative of that area, or site specific models, may be available that can be used to define the distribution of a chemical resulting from a single or from multiple emission sources. When relevant monitoring data are not available or when a site specific model is not appropriate, the assessment scheme is based on the generic model used in the TGD to represent the freshwater and terrestrial environment of Europe, but identifying when and how this model can be modified to address the marine environment. Although a number of different mathematical models exist to predict the behaviour and fate of chemicals in the marine environment, for both specific and generic scenarios, these cannot be adopted or refined until the underlying assumptions have been examined more thoroughly.

As in the TGD, the spatial scales proposed are (marine) local, regional and continental, but these are conceptually nested within a further "global" scale, representing the world's oceans, although the proposed risk assessment techniques are restricted to the three smaller scales. Generic local scenarios, comparable with those described in the existing TGD, are defined for the marine environment. A conceptual marine region is proposed as the estuarine and coastal sea area that receives the river water flow and atmospheric load emanating from the terrestrial/freshwater "default" region defined in the TGD, although it may be beneficial to consider two or more different regional scenarios. The dimensions of the continental scale will depend on the size selected for the regions within it; an example is given which is similar in size and characteristics to the North Sea. The report demonstrates how the marine regional and continental scales can be modelled by adjusting the parameters of the current EUSES model.

The modelling approaches used in risk assessment rely on the accuracy of key physico-chemical data. To assess the fate and effects of organic solutes in marine waters, it is important to consider the apparent increase in hydrophobicity of the substance, or its tendency to partition out of the aqueous phase, beyond that which applies in fresh water. However, the data suggest that, for nonpolar organic substances, the magnitude of difference in solubility and partition coefficients between sea water and fresh water is relatively small and will generally be of little significance for elucidating environmental risk.

Persistence needs to be addressed when estimating the presence of a chemical in the marine environment. In principle, half-life criteria should be used to determine the rate of degradation in the environment and both biotic and abiotic degradation pathways should be taken into account. It is difficult to assess degradation in the marine environment on the basis of existing laboratory tests, however, it is reasonable to conclude that biodegradation rates in marine environments, especially in estuarine and coastal regions with higher nutrient levels and sediment disturbance, can be as great as in fresh water. The same translation of rate constants from standard test data used for freshwater assessments should be employed for the marine regional model.

In the absence of half-life values for biodegradation, positive ready biodegradation test results should be used with care to provide rate constants (from the TGD) for use in the marine assessment. When data from specific inherent tests are available, a less conservative approach should be considered, and the data should be taken as evidence of marine biodegradation. A strategy for marine biodegradation testing is proposed.

Although further research is required in a number of areas, a framework is proposed, for marine risk assessment, using existing tools and knowledge, that allows exposure (PECs) to be defined at the local, regional and continental scale.

Bioconcentration from water into marine organisms is only likely to differ significantly from that in fresh water when the form or speciation of the substance is directly affected by the ionic composition of sea water. For neutral, hydrophobic organics, bioconcentration measurements or predictions based on freshwater data will provide reasonable estimates of marine bioconcentration potential.

As for bioconcentration, the biomagnification potential in marine food chains may differ significantly from that in fresh water when the chemical speciation of the substance is directly affected by the chemistry of sea water. It is also possible that the greater complexity of marine food chains, and the more variable lipid content of marine organisms, requires special consideration in risk assessment, compared with the existing EU process for fresh water. To allow for such uncertainty, modifications are proposed to the current method for determining the exposure (PEC) for birds and mammals feeding on fish at the local-regional level and, at the regional-continental level, for larger mammals preying on these fish-eating birds and mammals. Whenever possible the estimated PEC values in marine organisms and their predators should be supported by existing measurements in biota.

The ecotoxicity database relating to the effects of chemical substances on marine and estuarine organisms is comparatively limited, especially for organic compounds. Although it is desirable that the ecotoxicity database for such species is extended, 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 classical aquatic taxa (i.e. fish, crustacea, algae). There does not appear to be any marked difference in sensitivity between freshwater and saltwater biota that systematically applies across all three trophic levels considered. Where evaluated, differences between trophic levels within each medium were generally as significant or even more marked. Such variation is implicitly assumed in the use of assessment factors in current risk assessment practice. Overall, the use of freshwater acute effects data in lieu of, or in addition to, saltwater effects data for risk assessment purposes is not undermined by the empirical data reviewed in this report. The use of pooled freshwater and seawater data is therefore recommended. Under such circumstances, PNEC values should be derived from the most sensitive result regardless of the medium.

Within the marine environment (although not necessarily exclusively) there is a need to consider the impact of the unknown long-term effects of bioaccumulation and biomagnification in complex food chains which include larger marine species of mammal or birds (i.e. high trophic level species). In order to characterise the risk to biota, it is necessary to obtain an estimate of the concentration of the chemical to which the organisms are exposed, considering all potential routes of exposure. The exposure concentration may be measured by analytical monitoring or predicted by modelling, or a combination of both. The ways in which monitoring data and different modelling techniques can be used to describe the distribution of a chemical depend on the spatial scale considered. For example, preliminary release from a point source necessitates a local assessment in the immediate proximity of the point of discharge, whilst further distribution and fate is dealt with at the regional scale. A similar differentiation may also be applied to the marine environment, and a multi-stage/scale approach is proposed for the risk assessment of chemicals in marine ecosystems:

  • at the local scale, concentrations of contaminants from point sources will generally be greatest in close proximity to the discharge and the assessment of risk within the water and sediment compartments will therefore be necessary;
  • secondary poisoning effects are considered based on the local-regional mean exposure level, for birds and mammals feeding on aquatic biota;
  • dependent upon the properties of the chemical substance under assessment, it may also be necessary to determine risk at a regional scale in particular when a local risk has been established or when residual concentrations from multiple emissions may combine. It is then necessary to establish the extent of potential impact. Under such circumstances the risk will be assessed in water and sediment compartments;
  • the secondary poisoning effects on higher tier mammalian predators (e.g. bears eating seals) are considered based on the regional-continental mean exposure level.

At continental scale, risk can be characterised as for the regional scale, as appropriate for substances which are more persistent and capable of being bioaccumulated.