Midwestern U.S. regional bioassessment: Using multiple data sources and a standardized modeling approach

We conducted a regional bioassessment for rivers across a three state region, (IL, MI, and WI), using various invertebrate metrics and a standardized modeling approach. An advantage of this modeling approach is its ability to identify reference conditions for a continuum of stream types and in severely altered landscapes. Integrating data from different state agencies and research projects was challenging. Consequently we stratified modeling by state, data source, and sampling approach. Our models performed well, generally accounting for between 25-70% of the variability in invertebrate metrics. For each site and metric we calculated a score comparing expected versus existing condition and assigned a status class (above expectation to very poor). The standardized metrics were more sensitive to landscape stressors (urban, agriculture, and flow alteration) and accounted for variation due to non-stressors (drainage area, surficial geology, temperature, and flow regime). Within a state, we also averaged single-metric based scores resulting in a composite multimetric score for each site. We combined the scores from the three states into a regional GIS model to examine patterns associated with land development and other landscape scale variables. Urban land was strongly associated with degraded stream condition while the impact of agriculture varied by state.