Use of SWAT to compute groundwater table depth and streamflow in the Muscatatuck River watershed

Groundwater table depth oscillation over time is an important issue for planning drainage systems in rural watersheds. Its proximity to the soil surface impacts soil properties, crop development, and agricultural chemical transport. Even though hydrologic models are valuable tools for simulating and predicting changes in water dynamics, groundwater table depth is usually difficult to predict. The Soil Water Assessment Tool (SWAT) is a continuous and distributed hydrologic model created to simulate the effect of land management practices on water chemicals, and sediment movement for large watersheds. However groundwater table depth is not computed by the model. A procedure to compute perched groundwater table depth using SWAT outputs is proposed, based on the theory used by DRAINMOD, in order to expand SWAT's capabilities. SWAT was calibrated and validated for streamflow for three watersheds, and for groundwater table depth for three soils, at sites located within the Muscatatuck River basin in southeast Indiana. The Nash-Sutcliffe model efficiency (R-N(2)) for monthly streamflow was 0.49, 0.61, and 0.81 for the three watersheds for the validation period (1995-2002). SWAT-predicted groundwater table depths provided R-N(2) values of 0.61, 0.36, and 0.40 for the three soils in the calibration period (1992-1994) and 0.10, -0.51, and 0.38 for the validation period (1995-1996). Even though the model performance for predicting groundwater table depth was not as good as for streamflow, SWAT predictions resembled the seasonal variation of the groundwater table with correlation coefficients (r) of 0.68, 0.67, and 0.45 for the three wells during the validation period.