An Enhanced MOD16 Evapotranspiration Model for the Tibetan Plateau During the Unfrozen Season

Abstract Evapotranspiration (ET) is composed of soil evaporation (ET s ), canopy transpiration (ET c ), and canopy intercepted water evaporation (ET w ), and it plays an indispensable role in the water and energy cycles of land surface processes. More accurate estimations of variations in ET are essential for natural hazard monitoring and water resource management. An improved algorithm for ET s with soil moisture and texture (SMT) and an optimized ET c based on the MOD16 model (MOD16‐STM) are proposed for the cold, arid, and semiarid regions of the Tibetan Plateau (TP). The nonlinear relationships between the soil surface resistance ( r s s ) and soil surface hydration state in different soil textures were redefined by using five flux tower measurements over the TP. The value of the mean potential stomatal conductance per unit leaf area ( C L ) used for ET c calculation in grasslands was optimized at 0.0038 (m s −1 ). The MOD16‐STM was compared with the MOD16 and a soil water index‐based Priestley‐Taylor algorithm (SWI‐PT) at five independent sites. The results showed that the half‐hourly, daily, and monthly estimations from the MOD16‐STM were closer to observations than those from the other two models. The results indicated that the MOD16‐STM increased R 2 by 0.2/0.05 and the index of agreement by 0.37/0.23 and deceased RMSE by 43.89/42.77 W m −2 and the absolute mean bias (|MB|) by 46.23/45.89 W m −2 , when compared to the results of the MOD16 at daily/monthly scales. The results suggested that the MOD16‐STM will produce more accurate ET estimates over the flux sites and has great potential for ET estimation and land surface model improvement for the TP region.