Quantitative lines of evidence for screening-level diagnostic assessment of regional fish community impacts: a...

Determination of local ecological impacts and stressor identification in aquatic ecosystems is increasingly needed and required for designing effective watershed management plans in various jurisdictions. Spatial database evaluation methods applied to available environmental data resources can provide screening-level, geographically based stressor identification hypotheses. Cross-comparison of the output of independent methods is a critical step to address common concerns with the interpretation of output, identify strengths and weaknesses, and reduce uncertainty. Two current approaches were compared in this study: (1) the "Effect and Probable Cause" (EPC) method and (2) a GIS "Weights-of-Evidence/Weighted Logistic Regression" (WOE/WLR) method. The methods were applied to the same Ohio (U.S.) spatial data resources to link impacts on local fish assemblages with various natural and anthropogenic stressors. The methods generally yielded significantly similar results in the identification of stressors and their relative influence. However, key differences were also observed between the methods which reflected the distinctive objectives and sensitivities of each. The findings show that scientific interpretation of analysis output requires an understanding of method characteristics, and suggests the potential value of utilizing multiple methods as quantitative lines of evidence in screening-level regional diagnostic assessment.