Modeling potential evapotranspiration of two forested watersheds in the southern Appalachians
Global climate change has direct impacts on watershed hydrology through altering evapotranspiration (ET) processes at multiple scales. There are many methods to estimate forest ET with models, but the most practical and the most popular one is the potential ET (PET) based method. However, the choice of PET methods for AET estimation remains challenging. This study explored ways to identify appropriate PET models for two small forested watersheds, one dominated by conifer plantation and one dominated by native naturally regenerated deciduous hardwoods, by using long-term hydrometeorological data collected at the Coweeta Hydrologic Laboratory in the humid Appalachians in the southeastern U.S. Our specific objectives were to: (1) contrast three common PET models (FAO-56 grass reference ET, Hamon PET, and Priestley-Taylor PET) and compare these PET estimates with measured AET at monthly and annual temporal scales, and (2) derive correction factors for the FAO-56 grass reference ET and Hamon PET models at the monthly scale using the Priestley-Taylor equation as the standard method for estimating forest PET. We found that different PET models gave significantly different PET estimates. The Priestley-Taylor equation gave the most reasonable estimates of forest PET for both watersheds. We conclude that the uncorrected Hamon and FAO PET methods would cause large underestimates of forest PET. Annual PET rates of the conifer watershed were higher than those of the native deciduous watershed due to the lower albedo (thus higher net radiation) in the former compared to the latter. Monthly correction factors provided useful tools for forest PET estimation in those areas lacking climatic data (i.e., radiation, humidity, and wind speed).