Confirmation of a Relative Risk Model Ecological Risk Assessment using Multivariate Statistics. J. F. Thomas, L. R. Bodensteiner, T. J. Hall, A. M. Obery, and W. G. Landis; Western Washington University, National Council for Air and Stream Improvement, and Oregon Department of Environmental Quality
Landscape-scale ecological risk assessment (EcoRA) is a rapidly developing resource management tool with a need for methodologies that result in analyses that are both testable and useable for decision-making. The purpose of our study was to test the risk hypotheses generated for the biological endpoints of a Relative Risk Model (RRM) EcoRA done on the Codorus Creek Watershed (CCW) in south central Pennsylvania. We were able to confirm the patterns of risk predicted for both tested endpoints by applying principal components analysis (PCA) to two data sets, a fish assemblage and a macroinvertebrate assemblage. We found that using a subset of the fish species allowed us to separate the sampling sites and examine patterns. Many of the fish that allowed this separation were also on the list of species of value generated by the stakeholders at the outset of the RRM EcoRA, which is an important point for resource managers. The first three components explained 73% of the variation and demonstrated similar patterns as the EcoRA. PCA also identified a pattern of reduced variation within and between the middle sites when compared to the outer sites, indicating these sites were more constrained by impacts. We were also able to use a subset of macroinvertebrate genera to separate the sampling sites. The first three components explained 58% of the variation but did not give any clear patterns, which also confirmed the RRM trend for this endpoint. All the patterns found using PCA were confirmed with hierarchical clustering and discriminant analysis.
Go to . . .
2001 SRA Annual Meeting Table of Contents
2001 SRA Annual Meeting Author Index
Main Abstracts Menu Page
RiskWorld Home Page