Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: July 9, 2004
Publication Date: November 1, 2004
Citation: Bosch, D.D., Sheridan, J.M., Batten, H.L., Arnold, J.G. 2004. Evaluation of the SWAT model on a coastal plain agricultural watershed. Transactions of the ASAE 47(5):1493-1506. Interpretive Summary: Under Congressional Act and court order, total maximum daily loads (TMDLs) are being developed for all impaired surface waters throughout the U.S. In most cases, the TMDL implementation plans are developed through a combination of observed data and computer model simulations. A commonly used simulation package, the Better Assessment Science Integrating point and Nonpoint Sources (BASINS) system and the Soil Water Assessment Tool (SWAT) watershed scale model, were evaluated for their utility as hydrologic predictors of the streamflow, groundwater contributions to streamflow, and evapotranspiration within watersheds in the Coastal Plain Region of the U.S. Findings indicate that the overall accuracy of the hydrologic component of the SWAT model is adequate (+/-20%) for developing average annual and monthly hydrologic budgets in the region, but may not yield the accuracy necessary for daily simulations necessary for developing TMDLs. Accurate land-use, soils, and topographic data are critical for producing accurate results. Low resolution readily available GIS coverages are unlikely to produce the accuracy necessary for TMDL development. Findings also demonstrate the necessity of testing model response using regionally-representative sites that have high resolution input data as well as long-term observations to compare with model predictions.
Technical Abstract: Under Congressional Act and court order, total maximum daily loads (TMDLs) are being developed for all impaired surface waters throughout the U.S. In most cases, TMDL implementation plans are developed through a combination of observations and model simulations. The Better Assessment Science Integrating point and Nonpoint Sources (BASINS) system was developed by the U.S. Environmental Protection Agency to facilitate developing TMDLs. The Soil Water Assessment Tool (SWAT) is one of the watershed scale simulation models packaged within BASINS. Because of the critical nature of the process, it is imperative that BASINS and SWAT be adequately validated for the regions they are being applied to. The BASINS system and the SWAT model were tested using hydrologic data collected on Watershed J of the Little River Watershed in South Georgia. A six-year simulation of streamflow was conducted. Comparisons were made between simulated water balance results obtained with default conditions set by the model and with modified initial conditions. Estimates were also made of the improvement that can be expected by increasing the resolution of the spatial input data. Simulated annual and monthly water balance data were within 20% of the observed values. Nash-Sutcliffe model efficiencies (E) for the monthly total water yields were calculated as 0.80 using the high resolution input data with the modified initial conditions while the efficiency was 0.64 using the low resolution input data with the default initial conditions. Comparisons with daily results indicated poor agreement with the observed daily data (E @ 0). The mean absolute average error between the observed and the simulated daily total flow was reduced by 11% by modifying the initial settings of three parameters describing the initial conditions of the shallow groundwater system. While the model simulated trends in the streamflow quite well, discrepancies were observed between observed and simulated hydrograph peaks, time to peak, and hydrograph durations. The results indicate the model should be a useful tool for this region. However, model modification and further calibration may be necessary to increase the accuracy of the daily flow estimates for TMDL development.