Two Alternatives to the Two-Source Energy Balance Evapotranspiration Model

Environmental models are sensitive to inaccuracies in their approximation algorithms, which can bias model simulations and even lead to incorrect concepts. We present two alternatives to the two-source evapotranspiration (ET) model, which utilize a more accurate approximation of the Clausius-Clapeyron relation. Model performance was evaluated through a comparison with observed half-hourly eddy covariance fluxes. Modeled representations of sensible heat dynamics and the limiting behaviors were also evaluated to identify the causes of model inaccuracies. Our analysis shows that the new exponential approximation used here significantly reduces errors that stem from solutions to the energy balance equations. The proposed parallel and series models are both more accurate than the Shuttleworth-Wallace (SW) model under the conditions evaluated here, but we cannot conclude that the new models are consistently more accurate over a broader range of conditions. The new models can correctly reproduce several theoretical limiting behaviors, whereas the SW model generates conceptually incorrect results for several important limiting cases in which some key conductances approach both zero and infinity. Importantly, these models seriously underestimate the observed latent heat fluxes at night; the major causes of the nighttime model inadequacies are thought to be uncertainties of the estimated conductances and forcing data errors. The central concepts and processes related to new models are described via plots of temperature versus specific humidity to explore the theory behind an ET model. Through these efforts, we anticipate making more reliable ET predictions that will lead to a complete understanding of vegetation-atmosphere interactions.

Automated summary

This article presents two alternatives to the two-source evapotranspiration model that utilize a more accurate approximation of the Clausius-Clapeyron relation. Model performance was evaluated through a comparison with observed fluxes. The new exponential approximation significantly reduces errors in the energy balance equations. The proposed models outperform the Shuttleworth-Wallace model under the conditions evaluated, but their accuracy under broader conditions is still uncertain. Importantly, these models underestimate nighttime latent heat fluxes, likely due to uncertainties in estimated conductances and forcing data errors.