Abstract
Increasing drought frequency and intensity present significant challenges for the accurate simulation of terrestrial gross primary productivity (GPP) and evapotranspiration (ET) in terrestrial biosphere models (TBMs). Traditional soil water stress schemes often rely on empirical scalars or threshold-based parameterizations that fail to represent plant physiological acclimation during drought. In this study, we integrated the unified theory of plant photosynthesis and hydraulics (PH) into the Biosphere-atmosphere Exchange Process Simulator (BEPS), named the BEPS-PH model. This framework mechanistically couples plant hydraulics, stomatal regulation, and leaf photosynthetic capacity using trait-based optimality principles. We evaluated the performance of BEPS-PH against two widely used empirical approaches across four eddy covariance sites representing contrasting climates and vegetation types: tropical savanna, deciduous broadleaf forest, oak savanna, and grassland. Our results demonstrate that BEPS-PH consistently enhances the simulation of seasonal and interannual GPP and ET variations, achieving significantly lower root mean square errors and reduced biases than empirical models, particularly under moisture-limited conditions. Notably, the BEPS-PH accurately reproduces the observed declines in GPP and ET with decreasing soil water potential during dry-down periods, reflecting realistic hydraulic constraints on carbon–water coupling. These findings indicate that incorporating hydraulically explicit, trait-based parameterizations into TBMs provides a more robust and physically based framework for predicting ecosystem carbon and water fluxes under soil water stress.

This publication has 44 references indexed in Scilit: