Using Hierarchical Models to Evaluate Factors Affecting Occupancy, Abundance, and Detection Probabilities in Southern Coastal Plain Streams

The southeastern United States contains one of the most diverse freshwater fish assemblages in North America.  Species-habitat interactions are still poorly understood in southern coastal plain streams at the local scale, and this information is critical for the management of coastal plain assemblages.  To fill in some of these data gaps, we investigated the relative importance of local scale habitat covariates on fish occupancy within 264 South Carolina coastal plain stream reaches.  We evaluated species-habitat relationships by developing hierarchical, multi-species models for 50 species using 3-pass depletion data collected by the South Carolina Department of Natural Resources.  We also investigated covariates for detection and abundance, as well as interactions between sites and species.  Modeling results suggest species-specific occupancy, abundance, and detection probabilities differ by local scale habitat covariates, and detection probabilities vary considerably across species and sites.  This information provides much needed context on how assemblages are organized at the local scale. Our future work will expand this analysis to more species (especially those that are rare), and include larger spatial scales through landscape level covariates to better characterize these stream reaches.