In recent years, the PM2.5 pollution in China has become a top environmental and health concern, involving the characterization of healthy effects over a broad spatial area with uneven geographical distribution. This research aims to explore the influential factors for the PM2.5 distribution from a socio-economic perspective, based on the observations from China's 1497 monitoring sites in 2015. First, the Moran's I index and the local indicators of spatial association (LISA) were computed to outline the distribution of PM2.5 on a national scale using provincial-level divisions. Second, the correlation between the spatial distribution of PM2.5 and socio-economic factors were analyzed by ordinary least squares (OLS) and geo-weighted regression (GWR) models. The results indicated that the GWR model explained the causal relationships better. Generally, Beijing, Tianjin, and Hebei had peak levels of PM2.5, while Guangxi, Sichuan, and several other southern provinces had the lowest levels. Particularly, forest coverage rate and electricity consumption per capita were negatively correlated with the concentration of PM2.5. In this study, the vehicle ownership per capita proved to be the most significant factor that positively contributed to the concentration.
Keywords: PM2.5; socio-economy; spatial autocorrelation; spatial regression; spatial statistics.