Suitability of SWAT for the Conservation Effects Assessment Project: Comparison on USDA Agricultural Research Service Watersheds

Recent interest in tracking environmental benefits of conservation practices on agricultural watersheds throughout the United States has led to the development of the U.S. Department of Agriculture’s (USDA) Conservation Effects Assessment Project (CEAP). The purpose of CEAP is to assess environmental benefits derived from implementing various USDA conservation programs for cultivated, range, and irrigated lands. Watershed scale, hydrologic simulation models such as the Soil and Water Assessment Tool (SWAT) will be used to relate principal source areas of contaminants to transport paths and processes under a range in climatic, soils, topographic, and land use conditions on agricultural watersheds. To better understand SWAT’s strengths and weaknesses in simulating streamflow for anticipated applications related to CEAP, we conducted a study to evaluate the model’s performance under a range of climatic, topographic, soils, and land use conditions. Hydrologic responses were simulated on five USDA Agricultural Research Service watersheds that included Mahantango Creek Experimental Watershed in Pennsylvania and Reynolds Creek Experimental Watershed in Idaho in the northern part of the United States, and Little River Experimental Watershed in Georgia, Little Washita River Experimental Watershed in Oklahoma, and Walnut Gulch Experimental Watershed in Arizona in the south. Model simulations were performed on a total of 30 calibration and validation data sets that were obtained from a long record of multigauge climatic and streamflow data on each of the watersheds. A newly developed autocalibration tool for the SWAT model was employed to calibrate eleven parameters that govern surface and subsurface response for the three southern watersheds, and an additional five parameters that govern the accumulation of snow and snowmelt runoff processes for the two northern watersheds. Based on a comparison of measured versus simulated average annual streamflow, SWAT exhibits an element of robustness in estimating hydrologic responses across a range in topographic, soils, and land use conditions. Differences in model performance, however, are noticeable on a climatic basis in that SWAT will generally perform better on watersheds in more humid climates than in desert or semidesert climates. The model may therefore be better suited for CEAP investigations in wetter regions of the eastern part of the United States that are predominantly cultivated than the dryer regions of the West that are more characteristically rangeland.