Evaluation and comparison of OpenET models for estimating soil water depletion of irrigated alfalfa in Arizona

Authors: ATTALAH, SAID - University Of Arizona; ELSADEK, ELSAYED AHMED - University Of Arizona; WALLER, PETER - University Of Arizona; HUNSAKER, DOUGLAS - Former ARS Employee; Thorp, Kelly; Bautista, Eduardo; Williams, Clinton; Wall, Gerard; ORR, ETHAN - University Of Arizona; ELSHIKHA, DIAA ELDIN - University Of Arizona

Submitted to: Agricultural Water Management
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/23/2025
Publication Date: 10/3/2025
Citation: Attalah, S., Elsadek, E., Waller, P., Hunsaker, D.J., Thorp, K.R., Bautista, E., Williams, C.F., Wall, G.W., Orr, E., Elshikha, D.M. 2025. Evaluation and comparison of OpenET models for estimating soil water depletion of irrigated alfalfa in Arizona. Agricultural Water Management. 320:109850. https://doi.org/10.1016/j.agwat.2025.109850.
DOI: https://doi.org/10.1016/j.agwat.2025.109850

Interpretive Summary: Water supply limitations in the US Southwest challenge crop growers to improve irrigation water use efficiency. In these areas, alfalfa is a major agronomical crop that needs adequate irrigation to satisfy crop evapotranspiration (ET) demand and attain marketable yields. Scheduling alfalfa irrigations based on actual ET demand could help growers achieve yield and efficient irrigation use. In this study, ARS researchers in Maricopa, AZ, and Temple, TX, evaluated a free web tool, called OpenET, that provides estimates of the actual ET in fields as calculated by six different ET models that incorporate remote sensing data from satellites. The OpenET data for models were compared to the measured ET determined in an alfalfa field in AZ. While all models were correlated to measured ET, two of the six stood out as being more accurate. However, the average ET data of the six models provided the best agreement to measured ET. Use of the average OpenET data, called the ensemble in the web tool, is recommended for guidance in alfalfa irrigation scheduling in this region. This information is of interest to producers, irrigation water managers, and irrigation professionals.

Technical Abstract: Crop growers in the desert USA Southwest and other arid regions need tools that provide actual crop evapotranspiration (ET) estimates to apply efficient irrigation. Such tools include the web-based, OpenET platform that provides accessible ET data in real-time based on six, satellite-based models. This study evaluated simulated ET of alfalfa (Medicago sativa L.) for the suite of OpenET models under a center pivot irrigation system in an arid climate (Buckeye, Arizona, USA). The models evaluated were ALEXI/DisALEXI, eeMETRIC, geeSEBAL, PT-JPL, SIMS, and SSEBop, as well as the Ensemble, which is an amalgamation of all models. Field data taken within the center pivot field were used to calculate “observed” alfalfa ET (ETobs) by soil water balance methods for 20, 7-21-day periods between 05/23/2023 and 12/1/2023. OpenET models suitable for alfalfa ET estimation for this environment were assessed by five statistical metrics, index of agreement (Dindex), Nash-Sutcliffe efficiency coefficient (NSE), coefficient of determination (r2), mean bias simulation error (MBE), and prediction error (PE), as applied against the ETobs derived at four, equal-area, quarter-circle plots under the pivot, as well as for plot-averaged data. The Dindex, NSE, and r2 indices were found acceptable for all models and Ensemble, except PT-JPL, which had negative NSE. Based on MBE and PE criteria for model acceptability, i.e., ±5.0 mm/period and ±15%, respectively, overestimation of ETobs by the eeMETRIC and SSEBop models and underestimation by the PT-JPL model were excessive and those models were deemed unreliable as singular models for alfalfa. The SIMS model had low MBE and PE. However, the SIMS model was deemed inadequate because its ET estimates over the 20 periods typically swung between excessive over and underestimation. The ALEXI/DisALEXI and geeSEBAL models were both considered acceptable based on the statistical metrics. However, the Ensemble was best overall having MBE and PE within ±2.7 mm/period and ±6.9%, respectively, for individual plots and plot-averaged values of -0.9 mm/period and -2.4%, respectively. Use of OpenET Ensemble ET should provide farmers and decision-makers with guidance for efficient irrigation management of alfalfa in arid climates.