Assessing the efficacy of the SWAT auto-irrigation function to simulate Irrigation, evapotranspiration and crop response to irrigation management strategies of the Texas High Plains

Authors: CHEN, YONG - Texas A&M University; Marek, Gary; MAREK, THOMAS - Agrilife Research; Brauer, David; SRINIVASAN, R - Texas A&M University

Submitted to: Water
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/7/2017
Publication Date: 7/12/2017
Citation: Chen, Y., Marek, G.W., Marek, T.H., Brauer, D.K., Srinivasan, R. 2017. Assessing the efficacy of the SWAT auto-irrigation function to simulate Irrigation, evapotranspiration and crop response to irrigation management strategies of the Texas High Plains. Water. 9(7):509. doi:10.3390/w9070509.

Interpretive Summary: Water scarcity due to drought and groundwater depletion has led to an increased emphasis on irrigation strategies for extending limited water resources. Models are commonly used to assess the impacts of such strategies. The Soil and Water Assessment Tool (SWAT), a widely used hydrologic model, is increasingly being used to evaluate the impacts of irrigation strategies at both field and watershed scales. However, concerns about the ability of the auto-irrigate function in SWAT to simulate actual irrigation practices have tempered results. Scientists from ARS and Texas A&M AgriLife compared simulated irrigation, crop water use (ET), plant growth, and yield to measured values for crops grown in the Texas High Plains. Results showed that the auto-irrigate function was unable to represent irrigation practices of the region, prompting the need for revision of the auto-irrigation algorithm in SWAT.

Technical Abstract: The Soil and Water Assessment Tool (SWAT) model is widely used for simulation of hydrologic processes at various temporal and spatial scales. Less common are long-term simulation analyses of water balance components including agricultural management practices such as irrigation management. In the semi-arid Texas High Plains, the underlying Ogallala Aquifer is experiencing continuing decline due to long-term pumping for irrigation with limited recharge. Appropriate irrigation management is crucial for extending remaining groundwater resources. Accurate simulation of irrigation water use and other associated water balance components are critical for meaningful evaluation of the effects of irrigation management strategies. However, the lack of long-term observations of irrigation magnitude and frequency is a fundamental barrier to accurate irrigation simulation. In the absence of such data, modelers often employ the auto-irrigation functions within the SWAT model. However, some studies have raised concerns as to whether the function is able to adequately simulate irrigation representative of actual irrigation practices. In this study, long-term observations (2000-2010) of climate, irrigation management, daily evapotranspiration (ET), seasonal leaf area index (LAI), and annual crop yield derived from an irrigated lysimeter field at the USDA-ARS Conservation and Production Research Laboratory (CPRL) at Bushland, Texas were used to assess the efficacy of the SWAT auto-irrigation functions. The SWAT model was first calibrated for daily ET, seasonal LAI, and annual crop yield using actual, full irrigation management inputs designed to minimize plant water stress. The calibrated model was then used to assess the auto-irrigation function to simulate actual irrigation using multiple soil water deficit thresholds under the soil water content stress method. Results indicated an excellent match between simulated and observed daily ET during both the model calibration (2001-2005) and validation (2006-2010) periods for the manually input irrigation management with a Nash-Sutcliffe efficiency (NSE) greater than 0.75 and percent bias (PBIAS) within plus or minus 10%. Analysis of the auto-irrigation function resulted in reasonable ET simulations under all SWD thresholds as indicated by NSE values greater than 0.5. However, this indicated that ET simulation was not sensitive to changes in SWD thresholds as the auto-irrigation function tended to over apply irrigation under several SWD thresholds as compared to actual irrigations. Use of average values of planting date, crop database parameters, and harvest date for each crop resulted in only a slight decrease in NSE as compared to the baseline scenario. This study highlights the need for the improvement of the SWAT auto-irrigation functions to better represent simulations of water balance and crop growth under irrigated conditions. We propose that a more sensitive and intuitive auto-irrigation algorithm representative of actual irrigation management would greatly improve simulation of irrigated acreages.