Improved simulation of gross primary production and evapotranspiration in a drought-prone temperate deciduous forest with the BEPS-EcoHydro

Climate extremes, particularly drought, severely affect ecosystem functions. Most terrestrial biosphere models use empirical soil moisture stress factors to represent the impacts of drought on stomatal conductance and photosynthesis, which lack a mechanistic representation of water flow in the soil-plant-atmosphere continuum (SPAC) and result in uncertainties in simulated carbon and water fluxes. In this study, a plant hydraulics module was integrated into the process-based Biosphere-atmosphere Exchange Process Simulator, i.e., the BEPS-EcoHydro, and comprehensively evaluated in a drought-prone temperate deciduous forest in the central USA. BEPS-EcoHydro considers SPAC water flow driven by the soil-leaf water potential gradient, potential transpiration, and plant water storage. Building on these hydraulic processes, the effect of water stress on photosynthesis in BEPS-EcoHydro was quantified via a linkage to leaf water potential. The results showed that BEPS-EcoHydro effectively captured variations in predawn leaf water potential at the ecosystem scale with a coefficient of determination (R2) of 0.54 (p < 0.01), and outperformed the original BEPS in simulating soil moisture with an improvement of R2 by 34%. Additionally, evapotranspiration (ET) and gross primary production (GPP) simulation performance has been improved with BEPS-EcoHydro, especially at the hourly scale. Importantly, BEPS-EcoHydro captured drought impact better than the original BEPS and detected the hysteretic responses of GPP and ET to leaf water potential during drought intensification and recovery periods. These results suggest that consideration of plant hydraulics in process-based ecosystem models is necessary to better understand mechanisms in vegetation responses to climate extremes.