Development of a predictive stream assessment model using the tolerances of Maryland freshwater fishes to physical habitat, chemistry and land use

Scott Stranko, Martin Hurd, Anthony Prochaska, Paul Kazyak
Maryland Department of Natural Resources, Monitoring and Non-Tidal Assessment Division, 580 Taylor Avenue, C-2, Annapolis, MD 21401; 410/260-8603; FAX 410/260-8620

A model was developed that predicts the fish species expected to occur in a stream if there were no impacts to the stream. The model uses tolerance limits of fishes to nine parameters related to naturally occurring stream conditions, including width, depth, gradient, discharge, current velocity, watershed area, Dissolved Organic Carbon, drainage basin, and physiography, to make predictions. The model was applied to 905 stream sites throughout Maryland. Results indicate that most streams have many fewer fish species than they could have if it were not for human impacts. Only 118 of the 905 sites (13%) had all or most of the fish species that we would expect to find based on naturally occurring stream conditions. A low percent of expected species (<25%) was observed at 149 of 905 sites (16%). A low percent of expected species present was used to indicate that a stream was impacted. Once a stream was identified as impacted, species tolerance ranges to 26 physical habitat, chemical, and land-use parameters were used to diagnose potential sources of impact. If a value for a parameter was not in the tolerance range for a species and the species was absent, but expected to occur, then the parameter was identified as a potential source of impact. The most prevalent sources of impacts responsible for the absence of fish species in Maryland streams where they were expected to occur based on this model were related to urbanization and large amounts of impervious land cover.