Evapotranspiration Methods Compared on a Sierra Nevada Forest Ecosystem

Evapotranspiration, as a major component in terrestrial water balance and net primary productivity models, is often difficult to measure with instruments and predict with models. This study compared five potential evapotranspiration models applied to a ponderosa pine forest ecosystem at an Ameriflux site in Northern California. We measured fluxes of water vapor and sensible heat by the eddy covariance method at the Blodgett Forest Research Station in the Sierra Nevada Mountains over two climatically different growing seasons (1997 was drier than the climatic mean, and 1998 was cooler and wetter than the climatic mean influenced by El Nino). We estimated actual evapotranspiration rates based on the latent heat flux measurements. We also logged environmental variables that are known to influence evapotranspiration rates, such as net radiation, wind speed, air temperature, and relative humidity. The evapotranspiration models ranged from simple temperature- and solar radiation-driven equations to physically-based combination approaches and included reference surface and surface cover-dependent algorithms. Vorosmarty et al. (1998) compared a similar set of models, but on a global scale. For each evapotranspiration model, results were compared against mean daily latent heat from half-hourly measurements recorded on a tower above the forest canopy at the Blodgett Forest Research Station in California. All models calculate potential evapotranspiration and thus overpredicted values from the summer seasons of 1997 and 1998 because soil moisture was not at field capacity. Development of a soil moisture function to connect potential with actual evapotranspiration resulted in significant improvement on three of the five models. In our study, Shuttleworth-Wallace, Penman-Monteith, and McNaughton-Black all yielded similar results. A modified Priestley-Taylor method performed well given its relative simplicity. As was found in Vorosmarty et al. (1998), the Penman method performed poorly.