Shifting the Plant Functional Type Paradigm to Reflect Hydraulic Properties may Improve Model Simulations of Drought and Disturbance

Ecosystem models of plant responses to climate classify trees into functional types by biome, phenology, and leaf traits, while excluding their hydraulic properties. Frequently, trees are grouped within the same plant functional type despite having opposing hydraulic strategies. Errors in the prediction of transpiration and carbon uptake by land surface models have been linked to the coarse resolution of plant functional types. We pair a field and a modeling study comparing two species typically classified together within the mid-successional temperate broadleaf deciduous plant functional type in order to highlight the divergent responses to drought and disturbance which result from their contrasting hydraulic strategies.

We studied sap flux, stem water storage, stomatal conductance, photosynthesis, rooting depth, and bole growth of red oak and red maple at disturbed and undisturbed field sites in Michigan. Within our research site, these species represent opposing ends of the hydraulic safety-efficiency spectrum. Red oak employs an efficient but high-risk hydraulic strategy - pairing anisohydric stomatal regulation with ring porous, embolism vulnerable xylem. Conversely, red maple relies on an ‘ultra' safe strategy of isohydric stomatal regulation and diffuse porous xylem. Species-specific differences significantly influenced temporal patterns of stomatal conductance and overall transpiration responses to both drought and disturbance. During non-limiting soil moisture conditions, transpiration from red maple typically exceeded that of red oak. However, during a 20% soil dry down, transpiration from red maple decreased by more than 80% while transpiration from red oak fell by only 31%. Stem water storage in red maple also declined sharply, while storage in red oak was only nominally changed. After disturbance, red oak increased stomatal conductance while maple conductance declined. Isotopic analysis of xylem water revealed that the deeper rooting strategy of oak allows transpiration and growth to continue during periods of water limitation, even when maple ceases transpiration. Bole growth data demonstrates that while red maple growth correlates strongly with mean annual soil moisture, growth of red oak does not.

Based on the cavitation risk-prone xylem architecture and leaf hydraulic strategy employed by red oak, the species would be expected to exhibit large water stress responses during soil drought. However, red oak did not experience the anticipated water stress. In fact, the risk-adverse red maple proved more sensitive to soil water content than oak. This result coupled with those from the xylem isotope analysis indicates that, during conditions typical to a hydrologically regular year, the rooting strategy of red oak offsets and may overcome the risks associated with its leaf and xylem hydraulic traits. These findings highlight the importance of a synergistic, whole-tree approach to the study and representation of tree hydrodynamics and the safety-efficiency trade off.

Use of these emergent tree-level hydraulic traits has the potential to improve model predictions of ecosystem-level transpiration and growth, particularly during periods of drought and disturbance. The implementation of such functional properties could be accomplished through either the recasting the plant functional type classification system to include whole-plant hydraulic traits, or explicitly representing plant hydrodynamics within land-surface models. Databases of species-specific hydraulic traits, such as the TRY Global Plant Trait Database, provide biologically relevant constraints for the governing hydraulic properties and will facilitate the implementation of both methods. Here, we propose a framework to incorporate whole-tree hydraulic strategies into land-surface models through the Finite-Element Tree-Crown Hydrodynamics model version 2 (FETCH2). FETCH2 resolves the fast dynamics of stomatal conductance at the tree level through a multi-layer canopy. FETCH2 uses atmospheric forcing from the land-surface model, simulates water flow through trees as a system of porous media conduits, and calculates realistic hydraulic restrictions to stomatal aperture on the basis of xylem water potentials. This model acts as a replacement for the current empirical link connecting soil moisture directly to stomata. FETCH2 resolved transpiration can be readily statistically scaled from the tree- to the plot-level using remote sensing data.

We use hydraulic traits from TRY and our field measurements to constrain FETCH2 parameters defining stomatal response to branch/stem water potential, xylem conduit properties, and rooting depth. Through these properties, we are able to capture the effects of iso- and anisohydric stomatal regulation, xylem conductance, and stem water storage on simulated sap flux and transpiration. Incorporation of these functional traits into FETCH2 allows us to replicate the disparate patterns of water acquisition and use of species with contrasting hydraulic strategies. Holistic plant hydraulic representation in models, through informed plant functional groupings or the incorporation of plant hydrodynamic models like FETCH2, will not only improve predictions of transpiration, growth, and mortality, but also simulations of the surface energy budget and the global carbon and water balances.