Assessing the Impact of Climate and Land-Use Changes on Midwest River Systems

The goal of this Midwestern regional study is to assess vulnerability of individual river reaches to potential impacts of projected climate and land-use changes using fish assemblages as indicators in the states of Michigan, Minnesota, and Wisconsin.  Because fish assemblages are strongly influenced by river temperature and flow regimes which are in turn affected by climate and land-use conditions, we have attempted to model the response of fish to climate and land-use changes through changes in river temperature and flow regimes.  Stream temperature, flow, and fish data were linked to individual confluence-to-confluence stream reaches along with land use data, landscape data (local and network buffers and catchments), and downscaled climate data.  We used an artificial neural network to model daily mean stream temperature for all reaches in the three-state region.  A modified version of the AFINCH (USGS) model was used to predict mean monthly stream flows.  Temperature and flow metrics calculated from the predictive models were then linked with fish assemblage measures.  Indicator analysis was used to discover useful temperature and flow metrics as well as fish metrics “sensitive” to climate and land-use changes.  We used boosted regression trees to determine current fish distribution and abundance, and then projected fish distribution and abundance under projected climate and land-use conditions for all river reaches in the study region.  The magnitude of changes between the current and projected future river conditions represents changes in species abundance, number of species, and/or fisheries assemblage type and hence the vulnerability of fish in a given stream reach.  Spatially-explicit determinations of vulnerabilities to climate and land-use changes provide managers with a tool to prioritize management scenarios at the scale of individual river reach, river type, state, or the entire three-state region.