Proposed hydrodynamic model increases the ability of land-surface models to capture hysteretic nature of transpiration

Hydraulic limitations are known to control transpiration in forest ecosystems when the soil is drying or when the vapor pressure deficit between the air and stomata is very large, but they can also impact stomatal apertures under conditions of adequate soil moisture and lower evaporative demand. We use the NACP dataset of latent heat flux (LE) measurements and model observations for multiple site/model intercomparisons to evaluate the degree to which currently un-resolved high-frequency (sub-daily) hydrodynamic stresses affect the error in model prediction of latent heat flux. Particularly, we see that models have difficulty resolving the dynamics of intra-daily hysteresis. We hypothesize that this is a result of un-resolved afternoon stomata closure due to hydrodynamic stresses. We find that although no model or stomata parameterization was consistently best or worst in terms of ability to predict LE, errors in model-simulated LE were consistently largest and most variable when soil moisture and VPD were moderate to limiting. This suggests that models have trouble simulating the dynamics that cause stomata to close due to high VPD and moderate to low soil-water availability. The majority of models tend to underestimate LE in the pre-noon hours and overestimate in the late evening. These diurnal error patterns are consistent with models' diminished ability to accurately simulate the natural hysteresis of transpiration. Nearly all models demonstrate a marked tendency to underestimate the degree of maximum hysteresis which, across all sites studied, is most pronounced during moisture limited conditions. The current formulations for stomatal conductance and the assumed empirical or semi-empirical coupling between stomatal conductance and soil moisture used by these models does not resolve the hydrodynamic process of water movement from the soil to the leaves. This approach does not take advantage of advances in our understanding of water flow and storage in the trees, or of tree and canopy structure. A more thorough representation of the tree-hydrodynamic processes could potentially remedy this significant source of model error. In a forest plot at the University of Michigan Biological Station, we use measurements of sap flux and leaf water potential to demonstrate that trees of similar type – mid-late successional deciduous trees – have very different hydrodynamic strategies that lead to differences in their temporal patterns of stomatal conductance and thus hysteretic cycles of transpiration. We found that red oak trees continue transpiring despite a large stem-water deficit while red maple trees regulate stomata to maintain a high water potential. Red oaks, which are ring porous, are also able to access more soil water, assumingly from deeper ground layers and have higher conductivity, compared with the maples, which are diffuse porous. These differences will lead to large differences in stomata conductance and water use based on the species composition of the forest. We also demonstrate that the size and shape of the tree stem-branch system may lead to differences in the extent of hydrodynamic stress, which may change the forest respiration patterns as the forest grows and ages. We propose a framework to resolve tree hydrodynamics in global and regional models. It is based on the Finite-Elements Tree-Crown Hydrodynamics model (FETCH) – a tree hydrodynamic model that can resolve the fast dynamics of stomatal conductance. FETCH simulates water flow through the tree as a simplified system of porous media conduits. It explicitly resolves spatiotemporal hydraulic stresses throughout the tree's hydraulic system that cannot be easily represented using other stomatal-conductance models. It uses a physical representation of water flow in a 3-D tree-stem-branch system assuming the xylem is a porous media. Empirical equations relate water potential at the branch-tips to stomata conductance at leaves connected to these branches. FETCH calculates the hydrodynamic stress related closure of stomata, provided the atmospheric and biological variables from the global model, and could replace the current empirical formulation for stomatal adjustment based on soil moisture. We propose that coupling FETCH to other land-surface models would reduce intra-daily errors in LE and improve atmospheric and hydrologic simulations.