Grid-level source tracing and source variation of sand and dust storm affecting in Beijing

Abstract

Abstract In recent years, sand and dust storm (SDS) activity in Northern China has shown a notable increasing trend. Fine-scale source apportionment of SDS events affecting Beijing is crucial for understanding the mechanisms behind this increase, facilitating effective desertification control, and improving early warning systems. In this study, we developed a grid-level source apportionment model utilizing GPU-accelerated computing, distinguishing it from previous studies where source tracing was often restricted to broad geographical regions. Results show that Mongolia and Inner Mongolia contribute 47.3% and 39.7% of Beijing’s average dust concentration, respectively. During high-concentration events, Mongolia’s contribution surges to 71.9%, while the influence of western Inner Mongolia diminishes to 1.5%. These severe events are primarily driven by the specific corridor of the China–Mongolia border, specifically Dornogovi and Omnogovi in Mongolia, as well as the Sonid Left and Siziwang Banners in Inner Mongolia. Furthermore, three transport pathways were identified, with Central Inner Mongolia and Northeastern Hebei serving as critical transit corridors. The recent increase in SDS events affecting Beijing is directly attributed to rising emissions from these key regions, primarily driven by a higher frequency of gale days and exacerbated by fluctuations in vegetation coverage. These findings provide a solid scientific basis for targeted, cross-border desertification control to mitigate the impact of severe SDS on Beijing.