Monthly 10-m evapotranspiration rates retrieved by SEBALI with Sentinel-2 and MODIS LST data

Abstract

Actual and potential evapotranspiration (ET) are important variables for regional and global environmental modelling. These elements provide better understanding and management of the hydrological cycle, particularly in relation to several environmental stresses affecting ecosystems. With the advancement of airborne techniques, remote sensing approaches have enabled an accurate estimation of crop actual ET with limited field visits. In this study, an enhanced version (i.e. m-SEBALI) of the improved surface energy balance model for land, surnamed SEBALI, will be presented. Main improvements concern the retrieval of 10-m ET values using Sentinel-2 images and MODIS Terra LST 1-km datasets. The other enhancement considers the usage of monthly-only climatic datasets, instead of the required hourly, daily and monthly data in SEBALI. Calibrations and validations were made in China, Belgium, Germany, Lebanon and Italy between 2013 and 2017, yielding adequate RSME (i.e. 20.06 mm/month) and AME (i.e. 15.69 mm/month) values. These countries represent four different climatic regions (i.e. continental semi-arid, Monsoon-influenced warm-summer humid continental, Mediterranean hot summer and Oceanic climates). Thus, Landast-8 thermal bands will be unnecessary to implement SEBALI. Also, the conversion between hourly and daily climatic data and monthly climatic datasets in SEBALI showed a Pearson value of 93.3 %. Thus, seven of the required inputs in SEBALI were eliminated in m-SEBALI, saving on time and resources. Furthermore, the wheat seasonal trend is produced in the Bekaa valley, Lebanon, between November 2018 and July 2019, showing an average ET value of 620 mm in a region with 510 mm of precipitations during the same period. Two peaks were visible in January and May signaling the different wheat phenological stages. The importance of m-SEBALI lies in providing an automated retrieval of 10-m ET, biomass production and water productivity (WP), among other variables, in diverse climatic regions. Such improvements shall improve the assessment of ET and WP towards better management of water resources, particularly in regions lacking some required inputs.