UC Berkeley

Evapotranspiration, a major component in terrestrial water balance and net primary productivity models, is difficult to measure and predict. This study compared five models of potential evapotranspiration (PET) applied to a ponderosa pine forest ecosystem at. an AmeriFlux site in Northern California. The AmeriFlux sites are research forests across the United States, Canada, Brazil, and Costa Rica with instruments on towers that measure carbon, water, and energy fluxes into and out of the ecosystems. 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. For each evapotranspiration model, results were compared against mean daily latent heat from half-hourly measurements recorded on a tower above the forest canopy. All models calculate potential evapotranspiration (assuming well-watered soils at field capacity), rather than actual evapotranspiration (based on soil moisture limitations), and thus overpredicted values from the dry summer seasons of 1997 and 1998. A soil moisture function was integrated to estimate actual evapotranspiration, resulting in improved accuracy in model simulations. A modified Priestley-Taylor model performed well given its relative simplicity. (c) 2004 Elsevier Ltd. All rights reserved.