Spatiotemporal Evolution Patterns of Global Actual Evapotranspiration and Its Influencing Factors

Abstract

Understanding the spatiotemporal dynamics of actual evapotranspiration (AET) and its drivers is critical for addressing climate change and ensuring ecosystem sustainability. Here, we analyzed global AET trends from 2001 to 2019 and assessed the relative contributions of six key influencing factors. Our findings reveal that AET exhibits a significant positive trend across 31.6% of the global land surface, predominantly in the Amazon Plain and the Loess Plateau of China, whereas a significant negative trend is observed over 5.2% of the land area, concentrated in eastern Brazil and southern Africa. The normalized difference vegetation index (NDVI) showed the strongest partial correlation with AET, influencing 26.6% of the global land area. Multiple linear regression (MLR) analysis indicates that precipitation exerts the greatest influence on AET in 39.9% of the world, followed by wind speed (WS) at 37.9%, while soil moisture (SM) is the dominant factor in only 0.2% of the global land area. Notably, WS drives 23.5% of the observed AET trends, whereas precipitation contributes most to trends in just 8.6% of the land area. Among the factors evaluated, NDVI emerges as the primary driver of AET changes, followed by precipitation, while surface net solar radiation (SNSR) has the weakest influence. These insights advance the understanding of global AET's spatiotemporal evolution and its driving mechanisms, offering a foundation for devising adaptive strategies to mitigate climate change impacts and enhance ecosystem resilience.