A Web-Based Geographic Information System For Visualizing and Forecasting The Biological Condition Of Wadeable Streams In South Carolina

Stream assessment programs can inform management decisions and guide conservation priorities by revealing relationships between biological condition, natural gradients, and human activities. Effectively communicating the nature of these dynamic, nonlinear relationships to non-technical audiences poses a challenge to natural resource agencies. Traditional reporting methods rely on tabular results or static spatial visualizations that require statistical expertise to interpret.  To reach a broader audience, a web-based geographic information system is introduced that allows users to engage in the process of modeling and forecasting stream conditions. Users are presented with an interface that allows the adjustment of spatial indicators reflecting human activities at the catchment and watershed scale. These modifications are used to forecast metrics describing biological condition. The underlying models were generated using data from a statewide assessment of wadeable streams performed by the South Carolina Department of Natural Resources. Assessment findings were related to spatial indicator data using Random Forest machine-learning methods. A web service executes a prediction and generates a spatial representation of the results. The target audience is primarily non-technical policy makers and conservation planners. The approach described focuses on South Carolina, but has applicability to stream assessment programs at the regional and national level.