Search

Evapotranspiration (ET) plays an indispensable role in water-carbon-energy exchange between the land surface and atmosphere. Biophysical constraints like the canopy conductance lie the key to accurate prediction of plant transpiration (T) and soil evaporation (E) in ET models. Given the lack of comprehensive evaluation for resistances network in conductance-based two source energy balance model (TSEB), three scenarios (Gc-TSEBO, Gc-TSEBOSM, and Gc-TSEBSWC) were assessed on the frame of Gc-TSEB: one relying on Jarvis canopy conductance, other on optimal stomatal behavior, and other on soil moisture content, with forcing data from MODIS datasets and in-situ FLUXNET sites over six biome types. The ET simulated by the model combined with the optimal stomatal behavior and the complementary concept had higher values of R2 and Taylor skill scores, lower values of RMSE and BIAS, than the other two scenarios, particularly over cropland sites and arid/semi-arid sites where the other two failed. The sensitivity analysis for Gc-TSEB resistances elucidated that the effect of canopy conductance is more noticeable on ET simulation than the other resistances. When effective biophysical constraints on the canopy are lacking, a large deviation exists in ET partitioning mainly due to the systematic error of soil evaporation through the partitioning algorithm for soil and plant fluxes. We assessed the performance of calibration from MODIS LST instead of flux data. The LST-calibrated model performed better in croplands and grasslands compared to forest sites, which opens a good perspective to advance the model for operational applications in agriculture and herbaceous vegetation.