Assessments of three evapotranspiration products over China using extended triple collocation and water balance methods

Abstract

Evapotranspiration (ET) plays an essential role in the water and energy balance between the surface and the atmosphere. Recent decades have witnessed the generation of various land evapotranspiration products. It is important to quantify their uncertainties, thus to identify which one exhibits the best performance. Therefore, to assess the performance of three popular ET products, namely the Global Land Evaporation Amsterdam Model (GLEAM), European Center for Medium-Range Weather Forecasts (ECMWF) Reanalysis (ERA-interim), and Moderate-Resolution Imaging Spectroradiometer (MODIS), across Chinese mainland on monthly scale, the root-mean-square error (RMSE) of each product was calculated at the basin scale using extended triple collocation (ETC) together with the water balance method. The results indicated that the RMSEs of GLEAM, ERA-interim and MODIS based on the ETC method (and water balance method) at the basin scale are 17.63 (12.04), 27.41 (22.90) and 40.55 (31.62) mm/month on average, while those at the grid scale are 8.47, 9.71 and 10.52 mm/month, respectively. This finding illustrates that GLEAM has the best performance (taking up ∼60 %) in China, while MODIS performs the worst. In addition, it indicates that the ETC method is a potential tool for evaluating different products and estimating their uncertainty without knowing the true value.