Optimization And Validation Of Two-source Energy Balance Model Coupling With Soil Moisture

Surface evapotranspiration is an important process of the terrestrial water cycle,which relates to the terrestrial surface ecosystem and hydrological processes.The two-source energy balance model(TSEB)and its revised model named Dual Temperature Difference(DTD)have been widely used to estimate land surface evapotranspiration under various landcover types and environmental conditions.However,due to the differences in the theoretical mechanism of model constructions,the models’performance will vary along with the variation of underlying land surface type and environmental condition.Meanwhile,land surface temperature may not provide accurate boundary conditions to constrain soil evaporation and plant transpiration under water stress conditions.In this study,the two models of TSEB and DTD were compared using the long time ground observation data collected from three different landcover types and environmental conditions,including alpine grassland,semi-arid irrigated farmland and arid riparian forest over Heihe River Basin in the Northwest China.The results showed that TSEB model was more sensitive to the information of land surface temperature.Given the processes of soil evaporation and plant transpiration both could be constrained by the soil water supply while the land surface was under soil water stress,a refined TSEB model was proposed by coupling surface soil moisture information(TSEB-SM).TSEB-SM model could simulate the surface fluxes accurately under various hydrothermal conditions,and separate surface evapotranspiration and its component fluxes reasonably.The main research contents and results were as follows:(1)DTD model and TSEB model showed different performances under different hydrothermal conditions.By comparing the performances of DTD model and TSEB model under various landcover types and environmental conditions,the results showed that the accuracy of instantaneous surface fluxes estimated by the TSEB model were variable,the model performed worst at the arid riparian forest site with sparse vegetation coverage.The latent heat flux estimated by the DTD model had a better agreement with ground measurements at the alpine grassland site,with the RMSE value of 62.00 W/m~2,while TSEB model showed a higher RMSE value of 75.49 W/m~2.While,at the semi-arid cropland site,the performances of the two models were more consistent where they produced closer RMSE values.But at the arid riparian forest site,both of the models significantly underestimated the latent heat flux.The DTD model showed a worse agreement with the ground measurements in both sensible and latent heat fluxes compared to the performance at the grassland and cropland sites.The DTD modeled latent heat flux had a higher RMSE value of 136.74 W/m~2where the TSEB model had a RMSE value of 86.40 W/m~2.The better agreement of the TSEB model latent heat fluxes may associate with the greater underestimation of modeled sensible heat flux which could partly compensate for underestimation of net radiation.However,at the daily scale,the performances of the DTD model and TSEB model were more similar.Additionally,the ratios of plant transpiration to evapotranspiration partitioned by the two models were in good agreement with results simulated from the water use efficiency(u WUE)model using ground measurements,but the DTD model performed slightly better than the TSEB model.Finally,the TSEB model was more sensitive to the model input of land surface temperature.(2)TSEB-SM model had better performance in simulating surface fluxes and separating surface evapotranspiration and its component fluxes under different hydrothermal conditions.By comparing the model outputs of the TSEB and TSEB-SM models with ground measurements,the results showed that the TSEB-SM model could better estimate the surface fluxes at the instantaneous scale over various landcover types and environmental conditions especially at the arid riparian forest site with sparse vegetation coverage.The latent heat flux estimated by the TSEB-SM model had a better agreement with ground measurements at the alpine grassland site,it produced an RMSE value of 58.44 W/m~2,while the MAPE value of TSEB-SM model was significantly decreased with a value about 10%when compared with TSEB model.While,in the semi-arid cropland site,the performances of the two models were more consistent.However,at the arid riparian forest,TSEB-SM model significantly improved the accuracy in estimating latent heat flux,it produced an RMSE value of 54.54 W/m~2,and the MAPE value decreased by 18%when compared with TSEB model.At the daily scale,the performance of TSEB-SM model was significantly better than that of TSEB model,especially at the arid riparian forest site with sparse vegetation coverage,where the MAPE of daily evapotranspiration was reduced from40%to 20%.The TSEB-SM model also produced more reliable values of the ratio of plant transpiration to evapotranspiration and had better agreement with the values partitioned by the u WUE model.At the alpine grassland and irrigated cropland sites,and the MAPE values produced by TSEB-SM model decreased by 16%and 23%,respectively when compared with the results produced by TSEB model.