Ecological risk assessment: from book-keeping to chemical stress ecology

Ecotoxicology emerged in the 1970s as the environmental branch of the field of toxicology. As a consequence, its major focus was on investigating the impacts of chemicals on individuals, rather than populations, communities, or ecosystems. Typical ecotoxicity experiments involve testing the effects of a chemical under standard laboratory conditions on individuals of a standard test species. This has yielded a massive amount of historical data, compiled in databases (e.g., www.epa.gov/ecotox) that are used in the ecological risk assessment (ERA) to estimate the environmental consequences of anthropogenic use of chemicals. Because of its focus on standardization, single-species tests, prescribed risk assessment methodologies, and flowcharts, ERA often seems more based on book-keeping than on science. Ecotoxicology, as the science underpinning ERA, should permit itself to grow out of its ¿single-species¿ shell because it should focus on the protection of populations and communities in the field (1, 2). The discrepancy between the question posed in ERA and the answer provided by single-species tests is concealed by the use of assessment factors (1).This historical background explains why only a very limited amount of ecological theory has become integrated into the field of ecotoxicology and ERA. As a result, science-based, ecosystem-level risk assessment methodologies have hardly been developed. To counteract this ecological deficiency in ERA, frameworks have been proposed to integrate some level of ecology into decision making (e.g., 3, 4). During the past decade, great progress toward this integration has been made on the experimental side (e.g., 5, 6) and some also on the modeling side (e.g., 7, 8). The understanding of how populations, communities, and ecosystems are affected by chemicals could be increased by integrating the fields of toxicology, chemistry, ecology, and bioinformatics at different levels of biological organization. The development of methods to extrapolate this improved understanding to untested situations would then greatly improve the ERA of chemicals (9); Figure 1)