Perennial coolwater streams (maximum daily mean water temperature 20.7–24.6°C; 90th percentile annual exceedence flow≥0.0283m3/s) are common in the Laurentian Great Lakes region of North America, including the state of Wisconsin, but they lack specific and effective bioassessment tools. I used landscape-scale ecological modeling of geographically referenced data on stream location, stream and valley slope, watershed surficial geology, air temperatures, riparian and watershed land cover, watershed road and population densities, and watershed nutrient and wastewater inputs to classify Wisconsin stream sites based on water temperature, flow, and degree of environmental degradation. Fish data from sites classified as perennial coolwater were then analyzed with objective standardized procedures to develop coolwater indices of biotic integrity (IBI). Because of substantial differences in their fish assemblages, separate IBI's were formulated for cool-cold transition (20.7–22.5°C) and cool–warm transition (22.6–24.6°C) streams. The cool–cold transition IBI had five metrics: numbers of darter, madtom, and sculpin species, of coolwater species, and of intolerant species, and the percentages of individuals as tolerant species and as generalist feeders. Scoring criteria differed between northern and southern Wisconsin (boundary at 44.6°N latitude) for the coolwater species metric. The cool–warm transition IBI also had five metrics: numbers of native minnow, of intolerant, and of benthic invertivore species, and percentages of individuals as tolerant species and as omnivores. Scoring criteria differed between small and large streams (boundary at 200km2 watershed area) for the intolerant species, benthic invertivore species, and percentage tolerants metrics. In validation tests, scores for both IBI's were significantly related to independent measures of human disturbance, indicating that the IBI's would be useful tools for assessing the environmental quality of coolwater streams. However, for any given level of human disturbance, scores for both IBI's varied substantially, suggesting that multiple IBI samples from multiple sites would be needed to provide a reliable evaluation of a particular stream segment.
Read full abstract