Integrating leaf functional traits improves modelled estimates of carbon and water fluxes at a subtropical evergreen conifer forest

Simulations of gross primary productivity (GPP) and evapotranspiration (ET) by terrestrial biosphere models (TBMs) are subject to significant uncertainty, in part due to the spatiotemporal variability in leaf photosynthetic capacity, which is not well represented in models. Recent studies have shown the potential for using leaf chlorophyll content (Chlleaf) to constrain GPP and ET modeling in deciduous vegetation with a strong seasonal phenology. However, little is known about how integrating physiological trait information affects modelled GPP and ET in evergreen plants. In this study, we investigated the feasibility of incorporating Chlleaf and leaf age into a TBM, as a proxy for leaf maximum carboxylation rate at 25 °C (Vcmax25) for improving GPP and ET simulations. Measurements of Chlleaf and Vcmax25 from different leaf age classes (current-year and 1-year-old) for Masson pine and Slash pine species, and leaf area index (LAI) were made in a subtropical Evergreen Needleleaf Forest (ENF) eddy covariance flux tower site. The parameterization of Vcmax25 using combined information on Chlleaf and leaf age considerably reduced the biases in simulated GPP and ET, relative to the cases of i) constant Vcmax25 and ii) Chlleaf based Vcmax25. The largest improvements in GPP and ET simulations were found in growing season (May to August), when monthly absolute errors (AEs) of modeled GPP were ∼40 % reduced, from 120.5 to 71.2 g C m−2 mon−1, with a 25 % decrease of monthly AEs of modeled ET from 52.3 to 39.1 mm mon−1. Chlleaf plays a different role in modelled photosynthesis and transpiration between sunlit and shaded leaves. The modeled water use efficiency (WUE) and light use efficiency (LUE) of the shaded leaves were both higher than those of sunlit leaves. This study presents the newly use of Chlleaf and leaf age as a proxy for improving Vcmax25 modeling at an ENF stand, which highlights the importance of using plant physiological traits and leaf age for improving ecosystem carbon-water coupling simulations.