Diagnosing LE fluxes, models, sensors and other culprits

Despite its omnipresence in atmospheric models, the Penman-Monteith (PM) equation often fails to represent the latent heat (LE) flux accurately. Deviations of several tens of % between modelled and observed LE flux are not an exception. The original PM equation assumed a constant stomatal resistance in time, but most current atmospheric models implement a varying resistance that depends on atmospheric conditions such as radiation, temperature and vapor pressure, while more recent models account for plant physiological stomata control to some degree.

In this study, we present a diagnosis of LE fluxes modelled based on the PM equation combined with a fixed, an environmentally driven and a plant physiology driven stomatal conductance model versus observed LE fluxes by Eddy-Covariance. The analysis covers a decade of observations for a grass and three years for a forest site in the Netherlands. We identify atmospheric conditions where the model and observations match and most strongly disagree and evaluate the contribution of stomatal resistance models in reproducing flux observations. In this study, we demonstrate that implementing models that account for varying stomatal conductance in response to atmospheric and soil conditions does not help to improve LE model estimates for these two datasets. We investigate whether eliminating some of the assumptions underlying the PM equation improves flux estimates and zoom in on the role of aerodynamic versus stomatal conductance in controlling LE flux. The aim is to provide suggestions for conceptual improvements that can help resolve some of the shortcomings in the PM-based LE flux estimation .