Understanding potential evapotranspiration (PET) is essential in planning economic uses of water resource and in assessing the potential ecological conditions for ecological restoration. In the semi-arid western Chinese Loess Plateau. This work tested three models commonly used to estimate monthly PET rate (i.e. Behnk-Maxey, Prestley-Taylor and Hargreaves). The major purposes are: (1) to determine the model used to calculate PET with small data requirements and high accuracy in the Zuli River Basin; and (2) to calculate the areal distribution of potential evapo-transpiration in the study area. From the approach, conclusion can be drawn that Hargreaves model may be the best way to estimate PET in the semi-arid area of the Loess Plateau with sparsity of weather stations. The general methods described here are expected to be applicable to the entire Loess Plateau.

Estimation of land surface evapotranspiration (ET) using remote sensing is increasingly becoming an effective approach to approximating the moisture and heat exchanges between soil and atmosphere systems. The ET in the western Chinese Loess Plateau is comprehensively estimated by inputting NOAA/AVHRR data-derived parameters into surface energy balance system (SEBS) model. The observed data in the western Chinese Loess Plateau are used to verify the estimates of input variables. ET estimation in this research shows that ET in the western Chinese Loess Plateau is primarily determined by vegetation coverage. Topography also plays an important role in controlling the estimated ET spatial distribution. This study also demonstrate the remote sensing-aided surface energy balance model can be used to estimate ET in semiarid areas.