Long-term monitoring of evapotranspiration using the SEBAL algorithm and Google Earth Engine cloud computing

The geeSEBAL application estimates and displays evapotranspiration maps times series based on Landsat images global meteorological data from ERA5 Land reanalysis. Codes applications are available at https://github.com/et-brasil/geesebal https://etbrasil.org/geesebal , respectively. Accurate estimation of ( ET ) is essential for several in water resources management. models using remote sensing have flourished recent years allowing spatial temporal assessments unprecedented resolutions. This study presents geeSEBAL, a new tool automated the Surface Energy Balance Algorithm (SEBAL) simplified version CIMEC (Calibration Inverse Modeling Extreme Conditions) process endmembers selection, developed within Google Earth Engine (GEE) environment. framework introduced, case studies across multiple biomes Brazil presented by comparing daily with eddy covariance (EC) 10 flux towers. Based 224 as inputs, yielded an average root mean squared difference (RMSD) 0.67 mm day −1 when compared to EC corrected energy balance closure. Additional analyses indicate low sensitivity yielding RMSD 0.71 driven situ measurements. On other hand, we found higher algorithm selection internal calibration. For instance, adjusting percentiles tropical error that was 36% lower standard percentiles. Finally, assessed long-term effects (1984–2020) land cover changes surface fluxes use agriculture key areas Brazil, deforested Amazon irrigated crops Pampas Cerrado biomes. A comparison temperature-based (SSEBop) vegetation-based (MOD16) model also performed assess relative advantages disadvantages. analysis showed has significant potential assessment data-scarce areas, due its inputs. codes written Python JavaScript freely GitHub ). includes graphical user interface ), important advances management regional scales.