IMPROVED AVGWLF MODEL FOR SEDIMENT/PESTICIDE PREDICTIONFROM AGRICULTURAL WATERSHEDS

摘要

AVGWLF is a watershed-based, hydrological, pollutant load model linking ArcView GIS with the GWLFmodel. The GWLF model simulates runoff, sediment, and nutrient loads from non-spatial source areas of rural watersheds.However, the model does not predict pesticide losses and has been shown to have large errors in sediment prediction. InPennsylvania initial TMDL assessments have used the AVGWLF model and there is interest in enhancing the model byadding an improved erosion/sediment transport capability and a pesticide sub-model. This paper briefly describes theseefforts and testing of the sediment and pesticide algorithms.To predict sediment loads the GWLF model uses a modified USLE combined with an area-based sediment delivery ratioand a transport capacity factor. Efforts to improve sediment prediction of the AVGWLF model include modifying sedimentdelivery ratio functions and incorporating RUSLE- and MUSLE-based technology. Changes in magnitude and timing ofsediment predictions caused by model revisions are evaluated and results are compared against observed sediment load datafor East Mahantango Watershed in Pennsylvania. The added pesticide sub-model has simple surface and subsurfacecomponents to predict pesticide loads to surface waters. Percentages of different crops on a watershed-level are generatedfrom county-level Census of Agriculture data. Pesticide use data are inferred from county-based pesticide data.Comparisons of predicted and observed atrazine and metolachor loads in streamflow for three watersheds in Pennsylvaniawere evaluated. Results indicate that the pesticide sub-model and the modified sediment yield algorithms have potential toimprove AVGWLF as a tool for TMDL assessments.