APPLICATION OF MULTIVARIATE TECHNIQUES TO END-POINT DETERMINATION, SELECTION AND EVALUATION IN ECOLOGICAL RISK ASSESSMENT

Ecological risk assessment has evolved so that the interaction among t he components is now an implicit assumption. Unlike single species-bas ed risk assessments, it is often crucial in environmental or ecologica l risk assessments to be able to describe a system with many interacti ng components. In addition, some quantifiable description of how diffe rent biological communities respond upon the addition of a toxicant or some other stressor is required to adequately describe risk at the ec osystem level. Three methods have been applied at this level: the mean strain measurement used by K. Kersting, the state-space analysis pion eered by A.R. Johnson, and the nonmetric clustering developed by G. Ma tthews for ecological data sets and for analysis of standardized aquat ic microcosm data. Each method has direct application to the descripti on of an affected ecosystem without reliance upon a single specific an d perhaps misleading endpoint. Each also can assign distance or probab ility measures in order to compare the control to treatment groups. No nmetric clustering (NMC) has the advantage of not attempting to combin e different types of scales or metrics during the multivariate analysi s and is robust against interference by random variables. Applications of these methodologies into an ecological risk assessment should have the benefit of combining large interactive data sets into distinct me asures to be used as a measure of risk and as a test of the prediction of risk. The primary impact of these methods may be in the selection and interpretation of assessment and measurement endpoints. Much recen t debate in toxicological studies has focused on appropriate measureme nt endpoints for tests. Nonmetric clustering and other multivariate te chniques should aid in the selection of these endpoints in ways meanin gful at the ecosystem level. We suggest that the search for assessment and measurement endpoints be left to the appropriate multivariate com putation algorithms in the case of multispecies situations. Applicatio n of these methods in the verification and validation process of risk assessment will serve to check the selection of endpoints during model ing exercises and to improve the presentation of assessment criteria.