Evaluation of SWAT for Estimating ET in Irrigated and Dryland Cropping Systems in the Texas High Plains

Abstract. Hydrologic models such as SWAT are used extensively for predicting water availability and water quality responses to alternative management practices. Modeling results have been used by regulatory agencies for developing remedial measures for impaired water bodies and for water planning purposes. However, comprehensive calibration and testing of these models for predicting daily ET are noticeably absent from most modeling efforts. This is largely due to the limited number of quality, long-term ET and related management datasets. Consequently, more readily available and easily measured components of the water balance, including runoff and stream flow, are often used as calibration targets. ET is often simply ignored or adjusted by manipulating sensitive ET model parameters in order to match other simulated water balance components to measured values. This approach can provide a false sense of confidence that the model is performing properly although significant errors in modeled ET may exist. In essence, the model may provide the right answer for the wrong reasons. This condition may go unnoticed until the calibrated model is used to simulate other components of the water balance such as ET and compared with measured data. Additionally, model errors in ET simulation may vary considerably under irrigated and dryland management practices. The use of default crop database parameters in SWAT‘s embedded production functions appears to be inadequate for accurately simulating ET under dryland conditions, resulting in inaccurate estimations of crop biomass, yield, and irrigation demand. Improper or incomplete calibration and testing of water balance components and the use of uncalibrated crop database parameters can be problematic and lead to improper evaluation of management strategies. In this presentation, we discuss calibration and testing efforts of SWAT simulated ET under both irrigated and dryland cropping systems of the semi-arid Texas Panhandle using lysimetric data from the USDA-ARS Conservation and Production Research Laboratory in Bushland, Texas.