Spatiotemporal variation and driving effects of evapotranspiration in China during 2001–2022

The Chinese mainland. Evapotranspiration (ET) plays a crucial role in global water and energy exchange. Diverse remote sensing-based ET products provide a valuable data foundation for research at regional to global scales. However, the accuracy of these products varies significantly across different regions, particularly in China—a country characterized by complex and diverse landscapes. This variability considerably limits our ability to accurately understand ET dynamics and their driving mechanisms. Therefore, this study integrates three mainstream remote sensing ET products, combines monthly data from 20 flux observation stations in China, conducts collaborative validation, and uses Random Forest and SHapley Additive Explanations to identify common driving patterns, providing a more comprehensive assessment of the nonlinear effects of climate and vegetation on ET changes, addressing gaps in current research. This study identifies Global Land Evaporation Amsterdam Model as the preferred ET product for China. From 2001–2022, ET significantly increased nationwide, whereas future trends are expected to follow a pattern dominated by anti-persistent increases, with persistent increases as a supplement. ET is likely to continue rising in ecological restoration zones. Nonlinear analysis revealed that precipitation is the primary driver of ET in China, with leaf area index as the secondary overall factor and temperature as the key meteorological variable, ranking third in total influence. • China's ET rose over 20 years, with sharp gains in restoration areas. • Future ET in China mainly shows anti-persistent increase, partly persistent change. • PRE dominates ET variation; LAI ranks second in overall influence.