Technical Report

TR 051 – Environmental Hazard Assessment of Substances

TR 051 : Environmental Hazard Assessment of Substances | January 1993

The main objective of the work described in this report was the development of a process for the Environmental Hazard Assessment of Substances, including principles and practical approaches. The concept is applicable to all substances, whether "new" or "existing" and to all environmental compartments.

The assessment of whether a substance presents a hazard to organisms in the environment is based on a comparison of the predicted environmental concentration (PEC) with the predicted no effect concentration (PNEC) to organisms in ecosystems.

The extent and nature of the information available to make such a comparison varies for different substances and environments. A practical hazard assessment scheme must allow for this. The process described in this report comprises a universal approach in which the estimation of PEC values is based on exposure model calculations and PNEC values on toxicological data. It should be recognized that where more precise information is available, for example direct observations linking measured concentrations in the environmental compartments of concern with effects, or absence of effects, on organisms in the corresponding ecosystem, this should take precedence.

The proposed scheme involves a stepwise of sequential process in which increasingly refined estimates of PEC and PNEC can be introduced if necessary and as data become available. Less refined estimates represent more conservative values. Comparison between PEC and PNEC can be made at different stages in the sequence. This makes it possible to reach a conclusion and terminate the assessment at the earliest possible point in the sequence, thus optimising effort and resources. Where PEC exceeds PNEC at an early point in the sequence, it is necessary to proceed further until a later comparison using more refined information proves satisfactory or risk management must be considered. It is not necessary that PEC and PNEC values are at the same level of refinement for PEC:PNEC comparisons.

In the sequential process, PEC values are estimated using predictive mathematical models for two geographical scales: regional and local. For diffuse sources, regional models based on representative environments estimate average environmental concentrations and provide information for the particular substance on environmental compartments of concern (air, water, sediment, soil or biota). Local models are used to estimate concentrations near the point of entry into the environment. The sequence comprises three levels of refinement: a preliminary or "Screening" phase, a "Confirmatory" phase and a final or "Investigative" phase. Different release patterns are considered (dispersive, diffuse and localised) and the losses of the substance to the environment during the full life cycle (manufacture, formulation, use and disposal) are addressed.

PNEC values can be estimated from toxicological data for the substance concerned. Such data may comprise acute or chronic toxicity measurements for various species in laboratory tests or (less frequently) toxicity measurements made in the field. PNECs may be derived from these data by extrapolation using appropriate "application factors" to allow for differences between the conditions under which the measurements were made and those applying in practice. In the present scheme, these application factors were determined empirically by comparing reliable, validated acute, chronic and field toxicity data for sub-sets from a database relating to 360 industrial and agricultural chemicals. This comparison indicated that for the majority of the substances listed, to estimate a PNEC from acute toxicity studies (which should involve at least three species from at least two taxa), the lowest of the median effect concentrations (EC50s) should be divided by an application factor of 200. Where chronic toxicity data are available, the PNEC may be derived by dividing the lowest no-observed-effect concentration (NOEC) by 5, whereas the NOEC from a multi-species ecosystem study can be taken directly as the PNEC. Certain categories of substances, for example metals and metal-containing compounds do not appear to fit the general relationships and have been excluded. Other product groups such as petroleum products may also require further consideration.

The Environmental Hazard Assessment follows a pragmatic approach in evaluating PEC/PNEC ratios. Where PNEC exceeds PEC in the evaluation for the regional case, the next step in the assessment is to determine a PEC for the local situation. If PNEC exceeds PEC at the local level, because of the conservative approach it can be concluded that the substance does not constitute an environmental hazard and hence the assessment can be finalised.

Environmental Hazard Assessment of substances is a demanding task. The central challenge is to develop practicable and valid means of reflecting behaviour in highly complex real environments on the basis of inevitably limited test data and models. It is believed that the approach proposed here represents the current state of the art (particularly with respect to application factors and regional models). However, although the basic concepts should be universally applicable, it is not yet possible to work out detailed applications of the scheme for all situations. Environmental Hazard Assessment is evolving rapidly and significant further advances in models or in availability of ecotoxicological data can be expected.

Thus, to date most Environmental Hazard Assessment schemes have concentrated on freshwater systems and much less attention has been paid to sediments, soil or the marine environment. As a reflection of this, for this report it proved possible only to undertake the ecotoxicological evaluation and derivation of PNECs for aquatic organisms.

Furthermore, the concepts presented in the report are at present only worked out for steady-state environmental concentrations resulting from continuous emissions and should not be used for non steady-state situations resulting from intermittent releases such as batch processes.

The proposed approach was evaluated for several representative chemicals with differing properties and for which sufficient data were available. PECs and PNECs were derived from all available data so that for evaluation purposes PEC/PNEC comparisons could be made at as many stages in the sequential process as possible, including greater levels of refinement than needed to indicate absence of hazard according to the step-wise scheme.

This evaluation demonstrated that, for the limited number of substances tested, the proposed approach represents a workable system for the aquatic environment. Based on PEC:PNEC comparisons, further refinement did not alter the final judgement once the criteria for absence of hazard had been satisfied.

For both the exposure estimation and the assessment of effects, further work is required to demonstrate and verify the applicability of the scheme and to develop it further. This would include evaluating the scheme with additional chemicals, extending to other environmental compartments and taking into account non steady-state conditions. The present document should therefore be seen as an initial report.