ABSTRACT
We propose a formal causal evaluation inference process that uses a local-scale in situ approach and provides statistically defensible correlative conclusions for the biological impairment in 303(d)-listed water bodies. Fish assemblage, water chemistry, and aquatic habitat data from 66 locations in the Limberlost watershed, Indiana, were collected in August 2003 to evaluate the condition of the watershed. Anthropogenic stressors were identified from a numerical classification analysis of fish assemblage data using a biological integrity response gradient. Physical and chemical stressors were evaluated relative to the biological integrity response gradient using the Kruskal Wallis analysis of variance by ranks test. Three water chemistry variables (sodium, chloride, and barium) were associated with biological integrity. The percent run habitat, channel morphology score, and Qualitative Habitat Evaluation Index score varied with the degree of dredging. We observed a negative response between the number of species and the biological integrity gradient and a positive response with the percent tolerant individuals. This local-scale in situ approach to physical and chemical data analysis, combined with concurrent biological data collection, provided correlative relations to the impaired condition with localized environmental stressors.
