A comparison of remote sensing evapotranspiration products over agriculture crops

Evapotranspiration is an essential element in the water cycle, being estimated through various methods such as the Penman-Monteith equation, the Priestley-Taylor equation, the FAO-56 Penman-Monteith equation, the Surface Energy Balance Algorithm etc. With the advent of satellite imagery, constant efforts have been made to improve the spatial resolution of ET estimates through the integration of high-resolution remote sensing data. Evapotranspiration estimated based on MODIS data utilizes the Penman-Monteith equation, resulting in a product at a resolution of 500 meters over a period of over 20 years (2001-2023). Recently, new estimation tools for ET based on Sentinel 2 and Sentinel 3 data have been developed, such as Sen-ET, which uses the Two-Source Energy Balance (TSEB-PT) model and a Data Mining model to bring the product to a resolution of 20 meters. Furthermore, Python libraries developed for the Google Earth Engine platform provide the capability to estimate ET at 30 meters based on Landsat data using the Surface Energy Balance Algorithm. In this study, a comparison of ET will be performed, based on products derived from the integration of satellite imagery obtained from Sentinel 2 and Sentinel 3, Sentinel 2 and Landsat 8, Landsat 8 alone, as well as the ET product from MODIS. The analysis extends over a period of 8 years between 2016 and 2023, which is the common timeframe for all datasets. The study area is located in southeastern Romania, where extensive agricultural fields frequently suffer from drought. The results obtained from satellite imagery will be compared with those obtained from meteorological data, as well as with other satellite products capable of highlighting moisture deficits or drought.Acknowledgments: This study has received funding from the European Union Agency for the Space Programme under the European Unions Horizon Europe research and innovation programme under grant agreement No. 101082189 (MAGDA project).