The partitioning of canopy transpiration (Ec) and soil evaporation (Es) is often important to estimate evapotranspiration for hydrologic, ecologic, forest, and agricultural applications. Although Shuttleworth-Wallace (SW) model can estimate the partitioning of Ec and Es, some parameters due to the complicated model structure are not easily obtained. Therefore, this has limited the further development of SW model. In this study, an improved dual-source model (SSW) was proposed for ET estimation. This simpler model structure with less parameters than SW model can estimate Ec and Es, separately, and only requires conventional meteorological data. The estimations of Ec and Es for the observation period were consistent with the data measured from sap flow and micro-lysimeter, respectively, in a commercial cherry orchard in Beijing, China. Also daily ET estimated from the SSW, SW, and Penman-Monteith (PM) model were compared and analyzed. The results showed there was a small difference between SSW and SW model, but their accuracy were both higher than PM model's, and the root mean square error (RMSE) was 0.641, 0.613 and 0.955 mm/day, respectively. Sensitivity analysis for SSW model showed that the variation of ET was less than 2% when parameters varied in the range of ±20%. So, the SSW model is not only simple but also high accurate and stable, and it appears suitable to estimate cherry orchard ET.
