An improved method for estimating global evapotranspiration based on satellite determination of surface net radiation, vegetation index, temperature, and soil moisture

We proposed a method in an earlier study to estimate latent heat of evapotranspiration (ET). However, the influence of soil moisture (SM) on ET was not well considered and is addressed in this paper by incorporating the Diurnal land surface temperature (T_s) Range (DTsR). ET, measured at twelve sites in the U. S. during 2001-2006, is used to validate the improved method. Site land cover varies from grassland, native prairie, cropland, deciduous forest, to evergreen forest. The correlation coefficient between the measured and predicted 16-day daytime-average ET is about 0.92 for all the sites, the bias is -1.9 W m^-2 and the Root Mean Square Error (RMSE) is 28.6 W m^-2. We calculated global monthly ET from 1986 to 1995 at a spatial resolution of 1degtimes1deg from the International Satellite Land Surface Climatology Project (ISLSCP) Initiative II global interdisciplinary monthly dataset and compared it with the fifteen land surface model simulations of the Global Soil Wetness Project-2. The results of the comparison of 118 months global daily ET show that the bias is 4.5 W m^-2, the RMSE is 19.8 W m^-2 and the correlation coefficient is 0.82.