Optimization of canopy conductance models from concurrent measurements of sap flow and stem water potential on Drooping Sheoak in South Australia

Canopy conductance (gc) is a critical component in hydrological modeling for transpiration estimate. It is often formulated as functions of environmental variables. These functions are climate and vegetation specific. Thus, it is important to determine the appropriate functions in gc models and corresponding parameter values for a specific environment. In this study, sap flow, stem water potential, and microclimatic variables were measured for three Drooping Sheoak (Allocasuarina verticillata) trees in year 2011, 2012, and 2014. Canopy conductance was calculated from the inversed Penman‐Monteith (PM) equation, which was then used to examine 36 gc models that comprise different response functions. Parameters were optimized using the DiffeRential Evolution Adaptive Metropolis (DREAM) model based on a training data set in 2012. Use of proper predawn stem water potential function, vapor pressure deficit function, and temperature function improves model performance significantly, while no pronounced difference is observed between models that differ in solar radiation functions. The best model gives a correlation coefficient of 0.97, and root‐mean‐square error of 0.0006 m/s in comparison to the PM‐calculated gc. The optimized temperature function shows different characteristics from its counterparts in other similar studies. This is likely due to strong interdependence between air temperature and vapor pressure deficit in the study area or Sheoak tree physiology. Supported by the measurements and optimization results, we suggest that the effects of air temperature and vapor pressure deficit on canopy conductance should be represented together. Key Points We selected the best canopy conductance model for a native species in Adelaide Selection of proper stress functions is important for conductance modeling Temperature effect is different from other studies and nonignorable