Improvement of evapotranspiration simulation in a physically based ecohydrological model for the groundwater–soil–plant–atmosphere continuum

Abstract

Accurate quantification of terrestrial evapotranspiration (ET) is essential to understanding the interaction between land and atmosphere, as well as the feedback response of vegetation dynamics. In our previous work, a physically based ecohydrological model called the simple terrestrial hydrosphere (SiTH) model was developed to estimate ET and the other ET-related variables based on the groundwater–soil–plant–atmosphere continuum (GSPAC). However, the SiTH model (SiTHv1) still has some deficiencies in the model structure and parameters, which can result in potential uncertainty in the estimation of terrestrial ET. In this study, we aimed to address these limitations by developing a new version of the SiTH model (SiTHv2). The main modifications of the SiTHv2 model include: (1) the vegetation moisture constraint module is updated with vegetation optical depth observations; (2) the critical model parameters associated with root distribution are constrained using flux observations; (3) the soil module is extended to a three-layer module with 5 m of total depth; (4) an irrigation input water strategy is applied in the cropland areas; and (5) the latest ERA5-Land reanalysis data with a finer spatial resolution are used as the meteorological forcing data. The estimated ET of the SiTHv2 model was validated/compared at multiple scales (i.e., site/plot, basin, and global) with flux data, basin water balance data, and other mainstream global ET products, respectively. The results demonstrate that the SiTHv2 model performs better than the SiTHv1 model, with an improvement in the overall model root-mean-square error of 0.66 mm day−1 (plot scale) and 98.58 mm year−1 (basin scale), representing 27% and 22% improvements over the SiTHv1 model in the same circumstances, respectively. In addition, the performance of the SiTHv2 model ranks well when compared to the existing terrestrial ET models and products. The improvements to the SiTH model should allow improved estimation of terrestrial ET and provide support to potential studies in water transfer within the GSPAC.