Simulating hydrological and nonpoint source pollution processes in a karst watershed: A variable source area hydrology model evaluation

An ecohydrological watershed model can be used to develop an efficient watershed management plan for improving water quality. However, karst geology poses unique challenges in accurately simulating management impacts to both surface and groundwater hydrology. Two versions of the Soil and Water Assessment Tool (SWAT), Regular-SWAT and Topo-SWAT (which incorporates variable source area hydrology), were assessed for their robustness in simulating hydrology of the karstic Spring Creek watershed of Centre County, Pennsylvania, USA. Appropriate representations of surface water - groundwater interactions and of spring recharge - discharge areas were critical for simulating this karst watershed. Both Regular-SWAT and Topo-SWAT described the watershed discharge adequately with daily Nash-Sutcliffe efficiencies (NSE) ranging from 0.77 to 0.79 for calibration and 0.68-0.73 for validation, respectively. Because Topo-SWAT more accurately represented measured daily streamflow, with statistically significant improvement of NSE over Regular-SWAT during validation (p-value = 0.05) and, unlike Regular-SWAT, had the capability of spatially mapping recharge/infiltration and runoff generation areas within the watershed, Topo-SWAT was selected to predict nutrient and sediment loads. Total watershed load estimates (518 t nitrogen/year, 45 t phosphorus/year, and 13600 t sediment/year) were within 10% of observed values (−9.2% percent bias for nitrogen, 6.6% for phosphorous, and 5.4% for sediment). Nutrient distributions among transport pathways, such as leaching and overland flow, corresponded with observed values. This study demonstrates that Topo-SWAT can be a valuable tool in future studies of agricultural land management change in karst regions.