Estimating Evapotranspiration Using Coupled Remote Sensing and Three SEB Models in an Arid Region

Abstract

Τhree surface energy balance (SEB) models were used to estimate ET based on remote sensing (RS) techniques: surface energy balance algorithm for land (SEBAL), mapping evapotranspiration at a high-resolution with internalized calibration (METRIC™), and simplified surface energy balance index (S-SEBI). The study was conducted on the Todhia Arable Farm (TAF), located in the Riyadh province of the Kingdom of Saudi Arabia. Both RS data and meteorological data were used. RS data were obtained by processing 20 Landsat 8 satellite images for 2014 focused on alfalfa cultivation. The estimated ET was validated by comparing it with the measured ET using an eddy covariance (EC) device. The ET values estimated by the SEBAL, METRIC, and S-SEBI models were 1.7–18.9 mm·d−1 with a mean of 8.3 mm·d−1, 2.8–22.4 mm·d−1 with a mean of 11.0 mm·d−1, and 6.5–21.6 mm·d−1 with a mean of 15.7 mm·d−1, respectively. The highest ET values were observed in cultivated areas, particularly during the periods of increased irrigation. The comparison between the estimated and measured ET indicated that the METRIC model exhibited the best performance, with a root-mean-square deviation (RMSD) of 1.7 mm·d−1. The S-SEBI model was less accurate, followed by the SEBAL model, with an RMSD of 2.9 and 3.1 mm·d−1, respectively. The METRIC model exhibited moderate ET results as well and is considered to be the most suitable ET SEB model for arid regions.