Use of the Soil and Water Assessment Tool (SWAT) for simulating hydrology and water quality in the Cedar River Basin, Iowa, 2000--10

The U.S. Geological Survey, in cooperation with the Iowa Department of Natural Resources, used the Soil and Water Assessment Tool to simulate streamflow and nitrate loads within the Cedar River Basin, Iowa. The goal was to assess the ability of the Soil and Water Assessment Tool to estimate streamflow and nitrate loads in gaged and ungaged basins in Iowa. The Cedar River Basin model uses measured streamflow data from 12 U.S. Geological Survey streamflow-gaging stations for hydrology calibration. The U.S. Geological Survey software program, Load Estimator, was used to estimate annual and monthly nitrate loads based on measured nitrate concentrations and streamflow data from three Iowa Department of Natural Resources Storage and Retrieval/Water Quality Exchange stations, located throughout the basin, for nitrate load calibration. The hydrology of the model was calibrated for the period of January 1, 2000, to December 31, 2004, and validated for the period of January 1, 2005, to December 31, 2010. Simulated daily, monthly, and annual streamflow resulted in Nash-Sutcliffe coefficient of model efficiency (E_NS) values ranging from 0.44 to 0.83, 0.72 to 0.93, and 0.56 to 0.97, respectively, and coefficient of determination (R²) values ranging from 0.55 to 0.87, 0.74 to 0.94, and 0.65 to 0.99, respectively, for the calibration period. The percent bias ranged from -19 to 10, -16 to 10, and -19 to 10 for daily, monthly, and annual simulation, respectively. The validation period resulted in daily, monthly, and annual E_NS values ranging from 0.49 to 0.77, 0.69 to 0.91, and -0.22 to 0.95, respectively; R² values ranging from 0.59 to 0.84, 0.74 to 0.92, and 0.36 to 0.92, respectively; and percent bias ranging from -16 for all time steps to percent bias of 14, 15, and 15, respectively.

The nitrate calibration was based on a small subset of the locations used in the hydrology calibration with limited measured data. Model performance ranges from unsatisfactory to very good for the calibration period (January 1, 2000, to December 31, 2004). Results for the validation period (January 1, 2005, to December 31, 2010) indicate a need for an increase of measured data as well as more refined documented management practices at a higher resolution. Simulated nitrate loads resulted in monthly and annual E_NS values ranging from 0.28 to 0.82 and 0.61 to 0.86, respectively, and monthly and annual R² values ranging from 0.65 to 0.81 and 0.65 to 0.88, respectively, for the calibration period. The monthly and annual calibration percent bias ranged from 4 to 7 and 5 to 7, respectively. The validation period resulted in all but two E_NS values less than zero. Monthly and annual validation R² values ranged from 0.5 to 0.67 and 0.25 to 0.48, respectively. Monthly and annual validation percent bias ranged from 46 to 68 for both time steps. A daily calibration and validation for nitrate loads was not performed because of the poor monthly and annual results; measured daily nitrate data are available for intervals of time in 2009 and 2010 during which a successful monthly and annual calibration could not be achieved.

The Cedar River Basin is densely gaged relative to other basins in Iowa; therefore, an alternative hydrology scenario was created to assess the predictive capabilities of the Soil and Water Assessment Tool using fewer locations of measured data for model hydrology calibration. Although the ability of the model to reproduce measured values improves with the number of calibration locations, results indicate that the Soil and Water Assessment Tool can be used to adequately estimate streamflow in less densely gaged basins throughout the State, especially at the monthly time step. However, results also indicate that caution should be used when calibrating a subbasin that consists of physically distinct regions based on only one streamflow-gaging station.