Evaluation of the storm event model DWSM on a medium-sized watershed in central New York, USA

DWSM is a dynamic watershed simulation model that predicts distributed hydrograph and associated sediment discharge graph (sedigraph) of a watershed for a given storm event. Its performance, however, is not extensively tested in medium and large watersheds. Here, we applied DWSM to Upper Oneida Creek watershed located in central New York, USA with an area of 311 km² by dividing it into topographically connected 42 overland elements and 21 channel sections. Field-measured water discharge and sediment concentration data during two storm events, one on 9/30/2010 and the other on 6/28/2010, were used to test the performance of DWSM. Model simulation was performed by calibrating the key adjustable parameters in the input file till the best outcomes were achieved. The final results showed that during calibration for the 9/30/2010 event, DWSM successfully predicted the peak water discharge and its arriving time with the errors of -3.3% and 0%, respectively, and peak sediment discharge and its arriving time with the errors of -0.6% and -0.03%, respectively. For the whole event, DWSM under-predicted total water volume and event sediment load by 10.7% and 22.3%, respectively. Sensitivity analysis indicated that DWSM is most sensitive to the curve number adjustment factor, as well as factors representing flow resistance and flow detachment ability. During validation using the 6/28/2010 event, DWSM showed even better performance in predicting not only the peak values, but also event total values. These results showed that DWSM has the potential of successfully predicting event hydrology and sediment transport in the study watershed.