Understanding how soil moisture affects evapotranspiration (ET) is essential for improving ET estimates. However, many global ET datasets overlook soil moisture constraints, causing large uncertainties. In this study, we developed an improved model that better captures the influence of soil moisture on vegetation and soil evaporation. Our model significantly improves ET estimation accuracy and provides a new long-term global ET dataset to support water cycle and climate research.

We have proposed a novel framework to generate the first publicly released 10-m global plastic-covered greenhouse (PCG) map for 2020, derived from Sentinel-2. The global PCG area is about 14 259.85 km² in 2020, concentrated between 30° N and 40° N. China has 8224.90 km², accounting for 57.67 % of global and 83.29 % of Asian PCGs. The Global-PCG-10 map shows producer's accuracy of 85.12 % ± 0.90 % and user's accuracy of 99.82 % ± 0.11 %.

This study uses an ensemble machine learning model to predict long-term, high-resolution cooking activity data, establishing China’s first county-level cooking emission inventory spanning 1990–2021. It covers key pollutants such as polycyclic aromatic hydrocarbons. It reveals emissions’ long-term spatiotemporal trends and driving factors, such as population migration and economic growth, offering efficient control strategies. This dataset is crucial for air pollution and health impact studies.