Evaluation of three evapotranspiration models in terms of their applicability for an arid region

Abstract

Evapotranspiration has a major controlling influence on all hydrological processes. Unfortunately, reliable estimates are extremely difficult to obtain, especially in arid regions. In this research, three evapotranspiration models are evaluated, namely, the Advection-Aridity, the Complementary Relationship Areal Evapotranspiration (CRAE), and the Penman-Monteith models. By applying these models to an arid river basin in Xinjiang, China, the consistency and reliability of the models are examined. It is found that the Penman-Monteith model is particularly sensitive to the canopy resistance r c, but is not sensitive to the surface albedo. In arid regions canopy resistance may range from 400 to 1000 sm −1 for crops and grassland and a 10% error in r c results in a 10% error in the estimated evapotranspiration. The use of the surface roughness length and surface albedo have significant effects on the estimates of evapotranspiration in the Advection-Aridity model. To estimate regional evapotranspiration, a weighted average method is proposed. By using this method, regional evapotranspiration is obtained from point estimates (Advection-Aridity and Penman-Monteith models) which are weighted according to their relative surface areas. The actual evapotranspiration for the river basin obtained by this method is very close to water balance estimates. From this study, it can be concluded that the Advection-Aridity, the CRAE, and the Penman-Monteith models are applicable to arid regions, and that the Penman-Monteith model yields the best results.