Consideration of Drainage Ditches and Sediment Rating Curve for SWAT Model Performance.

Water quality models usually require a considerable amount of data for proper configuration, but in many cases this requires additional procedural steps prior to model applications.  We examined two different scenarios of such input issues in a small watershed using the Soil and Water Assessment Tool (SWAT).  The German Branch (GB; ~50 km2) is a sub-watershed of the Choptank, a USDA’s Conservation Effects Assessment Project (CEAP) benchmark watershed located on Maryland's Eastern Shore.  To provide the most accurate model estimates of watershed response, consideration must be given to the unique features of the watershed as well as the quality of available input data. One set-up scenario addressed the issue of accounting for copious drainage ditches within the study area by comparing stream flow estimates derived using a conventional Digital Elevation Model (DEM) versus a DEM, hand-edited to include drainage ditches. The second set-up scenario examined the issue of estimating measured sediment loads for time-periods lacking data using a sediment rating curve. Preliminary results showed a slight improvement in modeled flow estimation (a 4% increase in NSE value) during the calibration period and a significant decrease in modeled flow estimation (15% in NSE value) during the validation period when using the DEM edited to include the extensive ditch network of the GB.  Although the method used to derive sediment loading data was quick and rudimentary, it provided a good calibration of the sediment component of SWAT in the GB considering the limited amount of measured data available. This method, however, is not expected to always give good results especially with more drastic changes in land use and climate in the region.