Comparison of evapotranspiration estimation methods based on remote sensing technique and SWAT model

Evapotranspiration is one of the key components in calculating water balance and determining water consumption in agriculture for planning and management of water resources. Various methods have been proposed for estimating evapotranspiration so far, which depends on different algorithms and techniques governing them, their calculation process and structure are different. The purpose of this study is to compare the performance of the hydrological-vegetation SWAT model and global products of WaPOR and MOD16 based on remote sensing technique for estimating potential and actual evapotranspiration. Hence, Aji Chay watershed, including one of the important sub-basins of Lake Urmia, is selected as case study to investigate the purpose of this study. The Google Earth Engine platform is used to estimate evapotranspiration through global products, which have been developed based on different algorithms. Moreover, the SWAT model according to the streamflows data was set up, calibrated and validated for the watershed and the values of potential and actual evapotranspiration were simulated by the model for the period 2009-2019 with R2 = 0.95 and RMSE = 20 mm. The results of the SWAT model showed that the average simulated actual evapotranspiration has the highest agreement with the corresponding values of the main agricultural products presented in the national water document for the main plains of the watershed. Comparison between the models showed that the estimated values of potential and actual evapotranspiration both on an annual and monthly scales for the WaPOR have the highest agreement and consistency with the SWAT model compared to the MOD16.