Proposed Hydrodynamic Model Improves Resolution of Species-Specific Responses to Drought and Disturbance

Abstract:

Plant functional types in land surface models (LSMs) are broadly defined, and often represent species with different physiologies within the same category. For example, trees of opposing hydraulic strategies and traits are commonly grouped together, as is the case of red oak and red maple. As a result, LSMs generate typical patterns of errors in predictions of transpiration and production. We studied sap flux, stem water storage, stomatal conductance, photosynthesis, rooting depth, and bole growth of these species at disturbed and undisturbed field sites in Michigan. Species-specific differences significantly impact temporal patterns of stomatal conductance and overall transpiration responses to both drought and disturbance. During drought, maples relied heavily on stem-stored water, while oaks did not. After disturbance, oaks increased stomatal conductance while maple conductance declined. Isotopic analysis of xylem water revealed that oak roots can access a deep groundwater source, which maple roots cannot. This deep rooting strategy permits transpiration and growth to continue in oaks during periods of water limitation, even when maples cease transpiration. Using 16 years of bole growth data, we show that maple growth is strongly correlated with mean annual precipitation, yet oak growth is not. We propose a framework to incorporate these species-specific differences into LSMs using the Finite-Element Tree-Crown Hydrodynamics model version 2 (FETCH2) that resolves the fast dynamics and diurnal hysteresis of stomatal conductance at the tree level. FETCH2 uses atmospheric and biological forcings from the LSM, simulates water movement through trees as flow through a system of porous media conduits, and calculates realistic hydraulic restrictions to stomatal conductance. This model replaces the current, non-physical link which empirically connects soil moisture to stomatal conductance in LSMs. FETCH2 resolved transpiration is then easily scaled to the plot level using remote sensing data. By incorporating species-specific constraints on water flux into predictions of transpiration, growth, and mortality, we can improve simulations of the surface energy budget and global carbon and water balances.