A global terrestrial evapotranspiration product based on three-temperature model from 2001 to 2020

Accurate estimation of global terrestrial evapotranspiration (ET) is essential for better understanding the earth’s energy and water cycles. Although there are many existing global ET products, recent studies show that there is still great uncertainty in ET estimates. In addition, with the increasing trend of extreme climate hazards (e.g., drought and heatwave), accurate estimation of ET under such extreme conditions remains challenge. To overcome these challenges, we use 3-hourly and 0.25° GLDAS datasets (i.e., net radiation, land surface temperature, and air temperature) and the three-temperature model (3T), excluding resistance and parameter-calibration, to develop a global terrestrial ET product. Results show that the ET product retrieved from the 3T model agreed well with both observations from global EC towers (RMSE = 24.9 mm month-1) and basin scale water balance (RMSE = 116.0 mm yr-1). Moreover, the 3T-based global terrestrial ET product was comparable to other commonly used ET products, i.e., MOD16, P-LSH, PML, GLEAM, GLDAS, and Fluxcom, retrieved from varied models, but the 3T model performed better in cropland under extreme weather conditions than GLDAS did, decreasing RMSE by 9.0-20%. The proposed ET product can provide periodic and large-scale information to support water cycles related studies. The proposed ET product, covering 2001-2020, will be provided at daily and monthly scales with .nc4 data format. The filename for daily data is “3T_ET_YYYY_MM” whereas it is “3T_ET_YYYY” for monthly. “YYYY” refers to a given year and “MM” refers to a given month. For ocean areas and no-data pixels, the digital numbers (DN) are filled with -9999. To obtain the actual ET values, the DNs should be multiplied by a scale factor of 0.01.