Simple models for stomatal conductance derived from a process model: cross-validation against sap flux data

Abstract

Representation of stomatal physiology in models of plant-atmosphere gas exchange is minimal, and direct application of process-based models is limited by difficulty of parameter estimation. We derived simple models of stomatal conductance from a recent process-based model, and cross-validated them against measurements of sap flux (176-365 d in length) in 36 individual trees of two age classes for two Eucalyptus species across seven sites in the mountains of southeastern Australia. The derived models - which are driven by irradiance and evaporative demand and have two to four parameters that represent sums and products of biophysical parameters in the process model - reproduced a median 83-89% of observed variance in half-hourly and diurnally averaged sap flux, and performed similarly whether fitted using a random sample of all data or using 1 month of data from spring or autumn. Our simple models are an advance in predicting plant water use because their parameters are transparently related to reduced processes and properties, enabling easy accommodation of improved knowledge about how those parameters respond to environmental change and differ among species.