Satellite Imagery reveals an accelerating rate of increase in land evapotranspiration

Abstract Quantifying Evapotranspiration (ET), a major component of the hydrologic cycle, is critical for monitoring agricultural water use at the field scale and enabling better water management. Despite the importance of a global field-scale ET product, there is limited information available on the inter- and intra-annual variability in recent years. Here, we generate a monthly 100-m resolution ET dataset from 1990 to 2021, using a validated single-source energy balance algorithm and more than four million thermal Landsat satellite imagery scenes. We find that global ET has significantly accelerated over the last two decades at an annual rate of 1.33 mm yr− 1 (0.2%), despite regional disparities. This rate has intensified to 0.47% and 1.97–2.15% in the most recent twelve and seven years of the study, respectively, primarily due to increases in summer ET over North America, African tropics, and Indochina. Our machine learning analysis indicates that air temperature, rainfall, and atmospheric moisture are the primary drivers of these ET trends. This publicly available ET datasets has the potential to advance our understanding of global and local water use dynamics and can inform water policy and management decisions in a warming climate.