Persistent Reduction in Snow Mass Over China: Evidence From 40 Years of Satellite Observations

Snow water equivalent (SWE) serves as a key parameter for assessing snow hydrology, climate regulation, and water resource management. Despite its critical role, high spatiotemporal resolution SWE datasets remain limited, particularly for vast and topographically diverse regions such as China. Here, we present an innovative daily, 500-meter resolution SWE dataset spanning 1980–2020 across China, developed through multi-source remote sensing data fusion and machine learning-based downscaling. This dataset overcomes the spatiotemporal constraints inherent in conventional passive microwave SWE retrievals. Validation with ground-based measurements yields an RMSE of 10.85 mm, demonstrating its enhanced accuracy relative to existing products. Analysis of the dataset reveals a clear seasonal snow mass cycle in China, with a four-decade daily mean of 2.05 × 10¹⁰ m³ and peak values reaching 6.90 × 10¹⁰ m³ in January or February. Longitudinal trends indicate a significant and persistent decline in snow mass accumulation from 1980 to 2020. Furthermore, spatial variability analysis highlights marked regional differences in SWE trends during the snow season. This study advances the understanding of snow cover dynamics over China, offering a robust, high-resolution SWE dataset with applications in snow hydrology, climate change studies, and water resource planning.