Comparison of Potential Evapotranspiration Methods Based on Results from FLUXNET Sites in the United States

Evapotranspiration is the largest single component of the terrestrial hydrological cycle and its accurate estimation is important in water resource management and understanding the links between land use change and climate (Wallace 1995). Evapotranspiration regulates the links between stomatal conductance, carbon exchange, and water use efficiency (Vorosmarty et al 1998). However, estimation of evapotranspiration is difficult, and a process-based understanding of evapotranspiration is needed to quantify likely changes in evapotranspiration due to climate and land surface change (Choudhury and DiGirolamo 1998; Hutjes et al 1998). One current approach for estimating evapotranspiration is to calculate potential evapotranspiration (PET) using methods driven by meteorological data and/or vegetation characteristics, and to scale this estimate down to actual evapotranspiration (AET) based on limitations in available water (i.e., soil moisture) (Stannard 1993, Federer et al. 1996, Vorosmarty et al. 1998). The primary objective of this study is to compare a suite of PET methods at the fetch scale in a variety of forest ecosystems. PET methods under investigation include Turc (1961), Hamon (1963), Jensen and Haise (1963), Priestley and Taylor (1972), McNaughton and Black (1973), and Shuttleworth and Wallace (1985). In order to parameterize and compare these PET methods, we use data from a variety of FLUXNET sites within the United States. This study allows comparison of PET methods at the fetch scale in a variety of ecosystem types. We critique the methods based on parameterization issues and a direct comparison of modeled versus measured AET estimates. The findings may facilitate the use of PET methods in larger scale modeling efforts.