Spatial heterogeneous associations and spillover effects of ozone pollution in China

Abstract

Surface ozone pollution has become a major environmental concern in China. Understanding its spatial associations at both intra-stratum and inter-strata levels is essential for effective pollution control. However, research on spatial stratified heterogeneity and spillover effects of ozone pollution remains limited. This study applies a hybrid model integrating random forest, SHapley Additive exPlanations, and geographically weighted regression to investigate the spatial heterogeneity and spillover effects of ozone pollution. Principal component analysis and geographically and temporally weighted regression were further used to assess local environmental drivers of ozone spillover. The results indicate a fluctuating upward trend in ozone concentrations with enhanced spatial clustering. Intra-stratum associations in South, Central, and East China and inter-strata associations in Northeast and North China exhibited strengthening trends. Ozone spillover levels were highest in North China, with a slight decline, followed by Northwest and Northeast China, where spillover decreased more markedly. In contrast, Southwest, East, South, and Central China—regions with previously low spillover levels—showed a fluctuating increase. Industrial structure was the primary driver of ozone spillover across all regions and years; regions historically dominated by the secondary sector experienced spillover reduction following structural adjustments. Technology expenditure mitigated ozone spillover in North China but intensified it in Southwest China. The effects of relative humidity, ventilation coefficient, and mean sea level pressure on ozone spillover emerged in Central and Eastern China.