Identifying and communicating the relationships between natural gradients, human activities, and aquatic habitat integrity is crucial to aquatic community conservation. The South Carolina Department of Natural Resources (SCDNR), in conjunction with Clemson University, has developed a novel web-based mapping application to support aquatic conservation activities. The application 1) communicates assessment findings to a broad audience, and 2) engages users in the process of modeling and forecasting stream conditions. Model prediction metrics were generated from over 200 measurements taken at approximately 400 stream locations during a statewide assessment of wadeable streams performed from 2006 to 2011. Assessment findings were related to spatial predictor data created under the National Fish Habitat Assessment. Predictions were based on classification and regression models generated using the Random Forest machine-learning technique. The mapping application provides users with a browser-based interface to modify predictors at catchment (local) and watershed (network) scales. A web service dynamically generates predictions based on these user inputs.
Results/Conclusions
Web mapping applications are commonly used as the basis for decision support tools. Traditionally, these applications rely on statically generated spatial layers. The dynamic execution of models broadens the utility of the application and opens the forecasting process to a non-technical audience. Providing an accessible means of forecasting the effects of management decisions encourages a watershed perspective towards aquatic conservation. The application is targeted to stakeholders at the policy making and conservation planning levels. The approach described was implemented for South Carolina but is applicable to stream assessment programs at the regional and national levels.