| Abstract: |
Modeling Nutrient Loadings from Ungauged Watersheds in Chesapeake Bay�s Eastern Shore
Watershed models are extremely valuable tools for exploring human interactions with the landscape and predicting effects of different management options. Models often need to be calibrated and validated for improved performance, which require some sort of observed data available from the watershed. When there is insufficient or no data in a study watershed, one can build and calibrate the model in a nearby watershed having observed data, and then transfer the calibrated model parameters to the study watershed. In this study we used the SWAT model to predict flow and nutrient loadings (N, and P) from several small watersheds of the Chesapeake Bay coastal community of Queenstown, MD, where the Town Planning Commission is in the early phases of developing a comprehensive land use plan. Since neither flow nor nutrients data was available, SWAT model was calibrated with hydrologic and water quality data collected from three drainage basins on the Chesapeake Bay�s Eastern Shore (SERC 304, SERC 310 and USGS Greensboro). The calibrated SWAT model had a relatively high Nash-Sutcliffe Efficiency (ENS) and R2 (over 0.67 and 0.7 respectively) for all three watersheds in flow prediction. The Model�s performance for N was acceptable with ENS and R2 values around 0.6 and 0.4 respectively. The model did not predict P well with lower R2 and ENS values compared to flow and N. The calibrated SWAT model was independently tested on three additional watersheds close to the study area (SERC 305, SERC 306 and USGS Ruthsburg). The performance statistics R2 and ENS for Flow, Nitrogen and Phosphorus were quite similar to the results from calibration, verifying the reliability of the SWAT model. The calibrated parameters were used to build up the Queenstown watershed model.
|
