Abstract: To overcome the shortage of actual evapotranspiration simulation studies in areas with data deficiency, we constructed a coupled model (BC₂₀₂₁) based on the Budyko theory and complementary model to simulate actual evapotranspiration of the Hunhe River basin from 1982 to 2020. The Hunhe River basin is a typical vegetation restoration area in the Mongolian section of the Yellow River basin. We analyzed the spatiotemporal variations and influencing factors of evapotranspiration. The results showed that: (1) The established BC₂₀₂₁ model could well simulate the actual evapotranspiration at different temporal scales in the Hunhe River basin. The parameters of the model were positively correlated with precipitation and negatively correlated with sunshine duration. (2) The average actual evapotranspiration was 413.24 mm in recent 40 years, and slowly decreased with a rate of 0.44 mm·a^-1 from 1982 to 1999. However, it began to significantly increase with a rate of 2.53 mm·a^-1 since 1999, due to large-scale afforestation. The actual evapotranspiration changed most in spring at the intra-annual scale, while the interannual variation showed a pattern of first decrease and then increase. Spatially, actual evapotranspiration in the Hunhe River basin gradually increased from the northwest to the southeast under the combined influence of terrain, climate, and vegetation coverage. (3) Precipitation was the main factor affecting annual actual evapotranspiration variation. The driving factors for the variations of actual evapotranspiration differed among different seasons. Precipitation, NDVI, and sunshine duration were the main factors affecting the actual evapotranspiration in summer, autumn, and winter, respectively. Our results provide reliable basis for long-term simulation of actual evapotranspiration in areas with deficient data, and have important implications for rational utilization and management of water resources in arid and semi-arid areas with successful vegetation restoration.

Key words: evapotranspiration, Budyko theory, complementary model, spatiotemporal variation.