GIS-BASED HYDROLOGIC MODELING IN THE SANDUSKY WATERSHED USING SWAT

Typically, simulation output generated by computer simulation models is calibrated and validated at the drainage outlet of a watershed. Shortcomings of such an approach include the validation at only one geographic location without explicitly accounting for variable hydrologic patterns within the watershed. Our objective was to conduct a spatially distributed calibration and validation of water flow using the Soil and Water Assessment Tool (SWAT). We simulated surface, groundwater, and total flow in the Sandusky watershed in Ohio, located in the Lake Erie watershed and the Great Lakes basin. The calibration and validation of simulated water flow was conducted concurrently at four subwatersheds and the drainage outlet of the Sandusky watershed. We used measured streamflow data at five U.S. Geological Survey (USGS) gauge stations from water years 1998 and 1999 for calibration and from water years 2000 and 2001 for validation. The surface water simulations at all monitoring stations were better than the groundwater simulations. The validation of total water flow showed a range in mean error of 0.03 to 4.00 m3 s-1, a root mean square error of 0.06 to 2.56 m3 s-1, correlation coefficients of 0.70 to 0.90, and Nash-Sutcliffe coefficients of 0.40 to 0.73. In conclusion, spatially distributed calibration and validation accounted for hydrologic patterns in the subwatersheds nested within the relatively large Sandusky watershed. Overall, simulations of water flow in the Sandusky watershed and subwatersheds were satisfactory except for winter rainfall-runoff events. This study showed the importance of spatially distributed calibration and validation.