Performance of real evapotranspiration products and water yield estimations in Uruguay

Abstract

Real evapotranspiration (ETR) is a key variable in socio-ecological systems since it is related tothe food supply, climate regulation, among others. Additionally, ETR plays a significant role indetermining water yield (WY) at the catchment level, which directly impacts water availabilityfor consumption and irrigation. Therefore, it is essential to quantify ETR and WY fluctuations inresponse to various human pressures to enable comprehensive water planning. In recent decades,remote sensing has become increasingly employed worldwide for hydrological monitoring andestimating ETR. In Uruguay, several approaches have been attempted to quantify ETR. However,there is still a lack of assessments concerning the performance of different products, particularlythose using remote sensing. The main objectives of this article were twofold: a) to evaluate theperformance of various spatial explicit approaches for estimating real ETR and b) to estimate andanalyse the variability in WY derived from the different ETR products for three climatically contrastingyears. To achieve these objectives, we utilized four remote sensing ETR products: thePenman–Monteith–Leuning model (PMLv2), the MODIS product, the Simplified Jackson Modelbased on Landsat images and INTA-SEPA model based on NOAA-AVHRR images. We also employedtwo water balance models at two scales: national (derived from the National Institute forAgricultural Research of Uruguay, INIA) and micro-watershed level. Our results indicate thatMODIS and PMLv2 remote sensing products exhibited better performances compared to the otherapproaches. These products provided the highest spatial (500 m) and temporal (8 days) resolution,effectively capturing seasonal differences between land-covers. Moreover, they showed positiveand strong correlations with annual precipitation and productivity. The discrepancies observedbetween products have direct implications on the estimation of WY, not only in terms ofquantity but also in terms of spatial patterns. Future studies should explore the application ofMODIS and PMLv2 ETR estimations for understanding hydrological and ecological processes,conducting climate change research, detecting and mitigating agricultural drought, and managingwater resources effectively.