Decoding the spatial pattern of PM2.5 pollution from the perspective of socioeconomic factors and regional connectivity

Research on the spatial pattern of PM_2.5 pollution has achieved progress recently, but there are still shortcomings in the effects of macro socioeconomic factors and regional connectivity of PM_2.5 emissions. To address these issues, our study followed an analytical framework integrating empirical orthogonal function, Morlet wavelet analysis, time series decomposition, back propagation neural network, geographical detector and social network analysis. This framework was applied to a dataset embodying PM_2.5, meteorology and socioeconomics over seven years (2015-2021) across 279 cities of mainland China. The results indicated that the dominant spatial mode could well coincide with both pollutant emission field and atmospheric transportation background field in China. The emission-driven PM_2.5 concentrations were mainly impacted by socioeconomic factors of per capita natural gas supply, industrial added value and highway freight traffic volume. The urban PM_2.5 emissions in the Beijing-Tianjin-Hebei and its surrounding areas, Yangtze River Delta and Pearl River Delta all exhibited complex and multi-threaded spatial connectivity. Our study demonstrated the efficacy of this framework in PM_2.5 pollution research across large geographical regions, highlighting the importance of regional joint prevention and control of air pollution under routine regulation, not merely during pollution episodes.