Plant hydraulics: A theory-rich but data-poor field

Abstract:

Recent advancements in our conceptual understanding of links between plant hydraulics and key ecophysiological processes are being rapidly incorporated into hydrologic and earth system models and being applied to understand patterns of plant growth and mortality across the landscape. However, the abundance of models (each with its own mathematical algorithms and assumptions) is not matched by a sufficient abundance of data necessary to benchmark and test these frameworks. Data on key plant hydraulic variables and traits, including water potential and hydraulic conductivity, are generally limited to discrete and often destructive observations made at individual sites, and occasionally shared to plant trait databases. Observations of soil water potential, which are critical for interpreting variation in plant water potential, are almost never measured in-situ; rather, this key control on a host of ecophysiological process is often inferred from so-called ‘pedo-transfer functions” which can disagree by orders of magnitude when compared to each other, injecting huge uncertainty in models of water and carbon fluxes and pools. In this presentation, we will highlight some opportunities for progress on predicting and mapping water potential gradients across the landscape that could be enabled by more systematic and continuous observations of key hydraulic variables, especially when combined with ecosystem-scale flux observations (i.e. eddy covariance) and remotely-sensed proxies for plant water content.