Estimating wheat evapotranspiration through remote sensing utilizing GeeSEBAL and comparing with lysimetric data

Abstract

Abstract Accurate estimation of ET is vital for water resource management. In recent decades, researchers have focused on utilizing satellite imagery for this purpose. The use of RS data has enabled the development of new models that provide detailed spatial assessments. GeeSEBAL, an automated ET estimation tool, employs the SEBAL algorithm via GEE. The current version of GeeSEBAL utilizes Landsat images and ERA5 global reanalysis data to produce time series estimates. Landsat 8 images were processed into a 16-day time series spanning 2013–2022, specifically during the wheat growing season. To validate the GeeSEBAL model for 2013–2014, results were compared against lysimeter data. Subsequently, ET was calculated for the years 2015–2022. The evaluation of GeeSEBAL against lysimetric data, by metrics such as R 2, RMSE, MAE, NSE, and NRMSE, yielded values of 0.94, 0.98, 0.07, 0.86, and 0.62, respectively. Those findings underscore the importance of GeeSEBAL for estimating wheat ET in regions with limited data availability.

Keywords

• ET
• SEBAL algorithm
• GeeSEBAL
• GEE environment
• ERA5 weather data