Estimating evapotranspiration fraction by modeling two-dimensional space of NDVI/albedo and day–night land surface temperature difference: A comparative study

Three simple models estimating the evapotranspiration fraction (EF) were tested, validated and compared with available data from NDVI/albedo and day–night land surface temperature difference (DT). They were analyzed using ground based measurements collected by the energy balance Bowen ratio system at the 11 enhanced facilities located at the Southern Great Planes of the USA between April 2001 and May 2005. One model was superior in predicting the EF. It used an extension of the Priestly–Taylor equation and a relation between the normalized difference vegetation index (NDVI) and DT.

Research highlights
► Three simple models including PT–ΔTs model, PT–NDVI–ΔTs model, and S-SEBI–ΔTs model for estimating the evaporative fraction (EF) at a regional scale are tested, validated and compared with available data over Southern Great Plains, USA.
► The EF computed with the PT–NDVI–ΔTs model is the best.
► We also analyzed the reasons for the differences among three models and believe that these results will lead to better understanding of EF retrieval from remote sensing and its capability.