A scalable spatial decomposition of air pollution data into intercity- and neighborhood-scale components: Application to PM2.5 in South Korea

Despite ongoing efforts to mitigate air pollution, the effectiveness of policies often varies across regions due to the differing spatial scales of air pollution variability, which arise from the characteristics of pollution sources as well as geographical and meteorological factors. Understanding air pollution by isolating its components at different spatial scales is crucial for designing effective mitigation strategies. In this study, we propose a simple and intuitive method for the scalable spatial decomposition of spatiotemporal air pollution data into intercity-scale (tens of kilometers) and neighborhood-scale (several kilometers) components. To separate the intercity-scale from the neighborhood-scale component, we introduce a spatially varying ‘effective range’ for intercity-scale variability, based on the distance-decaying spatial autocorrelation of background-removed components. This effective range is influenced by emissions and geographical features. We applied this method to hourly PM_2.5 data from 535 air quality monitoring stations (AQMSs) across South Korea for 2021–2022. Our findings reveal that the intercity-scale component contributes most significantly to PM_2.5 concentrations in urbanized and industrial regions, such as the Seoul metropolitan area. In contrast, the neighborhood-scale component is more prominent near emission hotspots, such as industrial complexes. These results suggest that in regions where intercity-scale contributions are significant, effective air pollution mitigation strategies should prioritize intercity-scale regulations, which should be managed by the central government or through inter-local agreements, rather than focusing solely on local hotspots. This study provides a robust approach for quantifying both intercity-scale and neighborhood-scale air pollution contributions using ground-based AQMS data, facilitating the development of multi-spatial-scale strategies for air quality management.