Linking indices of biotic integrity to environmental and land use variables: multimetric clustering and predictive models

摘要

Advanced computerized methods and models of retrieving knowledge from large multiparameterndata bases were used to analyze data on fish and macroinvertebrate composition (metrics),nhabitat, land use and water quality. The research focused on the north central and northeasternnUnited States and involved thousands of sites monitored by the state agencies. The techniquesnand methodologies included supervised and unsupervised Artificial Neural Networks (ANN)nmodeling, Principal Component Analysis, Canonical Component Analysis (both linear andnnonlinear), Multiple Regression Analyses, and analyses of variance by ANOVA. The researchnresulted in defining a concept of clusters of sites based on their biotic (fish) communityncomposition, identified cluster dominating factors, and developed meaningful models fornpredicting fish composition based on environmental and in—stream habitat stresses.