Appendix H
Workshop Summary
The workshop lasted only 3 days, and it was impossible to achieve consensus on every issue. There was general agreement on the need for ecological risk assessment to be broadly defined. As noted in the plenary presentations by Drs. Yosie, Lovejoy, and North, the policy needs that must be served are broad. Despite the diversity of environmental problems and the complexity of the science needed to address them, decision-makers need common frameworks for comparison and common procedures to ensure credibility.
Retrospective studies, such as those of TBT and the spotted owl controversy, which involve identification and resolution of existing problems, and predictive studies, such as those of agricultural chemical regulation and biological control, which are aimed at preventing new problems, involve different scientific approaches and rest on different information bases. The technical issues discussed at the workshop include the following:
-
Selecting among numerous possible end points at different levels of biological organization;
-
Extrapolating effects from one species or level of organization to others;
-
Discontinuities and nonlinear responses;
-
Spatiotemporal scaling;
-
Accounting for background variability;
-
Evaluating both quantitative and qualitative uncertainties.
Despite the complexities, there was a clear consensus that it is feasible to talk about assessing ecological risks in a manner analogous to human health risk assessment and that most, and perhaps all, types of ecological risk assessment can be accommodated within a single conceptual framework.
There was also a consensus that the health risk assessment framework presented in the NRC's 1983 report, although useful as a point of departure, is too narrowly defined for ecological risk assessment. Its most obvious weakness is its orientation toward toxic chemicals. Ecological risk assessment must be applicable to a much broader array of stresses. The discussion groups on exposure assessment and dose-response assessment agreed, however, that the concepts of exposure and dose-response could be generalized in a straightforward way to accommodate nonchemical stresses.
A clear theme running through nearly all the case studies and discussion groups was that the links between management and risk assessment are much stronger and more pervasive in ecological risk assessment than is indicated in the 1983 report. Subjects of particular importance include the role of policy, in the form of legal mandates and regulatory procedures, in defining an ecological hazard, the kinds of information to be used to assess risks, and the complexity of risk characterization in ecological risk assessment. Ecological risk assessment must include evaluations of kinds of uncertainty usually absent from health risk assessment, expression of risks in terms useful for decision-making (including economic valuations), and communication between risk assessors and risk managers, many of whom are not trained as ecologists. Several groups discussed possible modifications of the Red Book paradigm, but workshop participants as a whole were divided over whether to modify the paradigm or to develop a new one that is explicitly ecological.
A number of research needs themes surfaced at the workshop. These are too numerous to list here and are noted in the summaries of individual discussion groups. However, we note several common themes, some of which were discussed in more than one group:
-
Extrapolation across scales. Effects of interest to risk managers usually involve changes in populations or ecosystems. However, many stressors (including toxic chemicals and exploitation) act through direct
-
effects on individual organisms. Alternatively, risk managers might be interested in effects of large-scale regional change over long periods (e.g., logging of old-growth forest), but individual studies are restricted to relatively short periods and small areas. Some form of modeling appears generally necessary to make these extrapolations, but few models have been used.
-
Quantitative and qualitative analysis of uncertainty. It was amply noted that uncertainty in ecological risk assessment extends far beyond uncertainty in individual parameter values. Many of the uncertainties are related to extrapolations across scales and other kinds of qualitative gaps in knowledge. Those knowledge-based uncertainties result in many assessments being based principally on professional judgment, rather than on quantitative analysis. Evaluating the uncertainty inherent in professional judgments is as important as quantifying the uncertainty in model-based assessments.
-
Validation of predictive tools: Needs for validation were mentioned specifically in the risk characterization, uncertainty, and modeling groups and by plenary session speakers (Yosie and North). Validation could include both designed experiments and retrospective monitoring of the outcome of risk management decisions.
-
Expression of risks in policy-relevant terms. This topic was debated at length in discussions of risk characterization, the regulatory process, and valuation and was mentioned in plenary session presentations (by North, Yosie, and Slimak). Many difficulties were noted. Terms used by ecologists (such as ecosystem, stability, and resilience) are unfamiliar to decision-makers and the public, and their value is not immediately obvious. Expression in economic terms is attractive and is favored by decision-makers. However, the valuation discussion made it clear that many aspects of valuing ecological resources—especially nonuse values, such as biodiversity—involve economic theories and measurement methods that themselves are highly uncertain.
Some of the above issues might never be fully resolved. However, workshop participants familiar with health risk assessment often noted that the same or similar difficulties also affect health risk assessment, and the existence of difficulties has not precluded health risk assessments. In his closing statement, Dr. Barnthouse noted that the terms of discussion about ecological risk assessment have changed. In past years, the discussion was about whether the concept of risk and the methods of