Evaluation of three evaporation estimation methods in a Canadian prairie landscape

Three approaches to estimating actual evaporation (evaporation from water, soil, and transpiration from plants) are evaluated against eddy covariance observations taken during the summer period of 2006 over an upland mixed-grass site in the St Denis National Wildlife Area, central Saskatchewan. The Penman-Monteith (P-M) combination approach explicitly takes into consideration the influence of surface resistance and available energy in order to calculate evaporation from non-saturated surfaces. The Granger and Gray (G-D) expression is an extension of the Penman equation to the case of non-saturated surfaces using a complimentary approach that considers the relative evaporation G, or the ratio of actual to potential evaporation as an inverse function of the relative drying power of the air, D. D is a function of the humidity deficit and available energy. The Dalton-type bulk transfer (BT) approach typically applied in land surface schemes considers turbulent transfer along the humidity gradient between the surface and atmosphere as diagnosed from the land surface temperature. In this case, surface temperature was observed radiometrically rather than modelled. The models were evaluated for several temporal scales from 15 min to seasonal, and compared with measured evaporation data obtained by an eddy covariance system. Results suggest that all three approaches have 'reasonable' applicability for estimating evaporation at point-scales for periods longer than daily, but none of the methods provide consistently reliable daily or sub-daily estimates of evaporation. Copyright (c) 2008 John Wiley & Sons, Ltd.