Understanding the Control of Hydraulic Traits in Tropical Forests Using a Hydrodynamic Model Within a Demographic Vegetation Model

Vegetation plays a key role in global carbon cycles and thus is an important component within Earth System Models (ESMs) to project future climates. A recent trend for ESMs is to incorporate vegetation demography, making it feasible for a more realistic representation of key size- and succession-structured processes that mediate plant responses to climate, and thus, ecosystem processes. In this study, we report on a new hydrodynamics (HYDRO) model within the DOE-sponsored dynamic vegetation model, the Functionally Assembled Terrestrial Simulator (FATES-HYDRO). The HYDRO model is built on the size and canopy structure representation within FATES and is expected to better capture the control of hydraulic traits on both vegetation dynamics and carbon/water fluxes. As a first step, we conducted a parameter sensitivity analysis using the distribution of biologically-interpretable and measurable plant hydraulic traits. We focused on tropical forests, where co-existing species have been observed to possess large variability in hydraulic traits. We first assembled 10 distinct datasets of plant hydraulic traits of stomata, leaves, stems, and roots, determined the best-fit theoretical distribution for each trait, and linked these based on taxonomically-standardized species names to generate a rank correlation matrix, which quantified the degree of interspecific (between-species) trait-trait coordination. Our analysis showed that hydraulic traits that determine the soil-root connection and the stomata control are more important for transpiration and GPP during dry periods, while hydraulic traits that determine the whole tree conductance are more importance for wet periods. Our analysis suggests that hydraulic traits could play an important role in carbon and water fluxes and vegetation dynamics in tropical forests and further measurements to capture the hydraulic control on stomata, root-soil interface and whole tree resistance could improve our prediction of future tropical forests within ESMs.