Estimation of regional PM2.5 concentration in China based on fine-mode aerosol optical thickness (AODf) and study of influencing factors

Abstract

In recent years, rapid industrialization and urbanization in China have resulted in severe air pollution, with fine particulate matter (PM_2.5) being a major issue. PM_2.5 estimation typically relies on aerosol optical depth (AOD) data, while PM_2.5 is primarily composed of fine-mode aerosols, better represented by fine-mode aerosol optical depth (AODf). This study constructed PM_2.5 estimation models using both AODf and AOD data to obtain long-term PM_2.5 concentration datasets for China. SHAP and biased dependence algorithms were applied to analyze influencing factors and interactions, along with regional differences in PM_2.5 estimation based on multimodal AOD. The results indicate that AODf-based PM_2.5 estimation slightly improves accuracy compared to AOD. PM_2.5 concentrations showed an increasing trend from 2001 to 2013, peaking during this period, followed by a decline after 2013. Seasonally, the highest concentration was observed in winter (64.49 ± 19.8 μg/m³), followed by spring and autumn, with the lowest in summer (33.07 ± 8.8 μg/m³). The main influencing factors include AODf (26.97%), relative humidity (14.33%), 2m temperature (10.75%), and total evaporation (9.93%). Regional differences are evident: in the west, coarse-mode aerosols dominate, limiting the accuracy of AODf-based estimation, while in the east, fine-mode aerosols play a larger role. Furthermore, the continued decline in PM_2.5 is attributed to the decreasing proportion of fine-mode aerosols. This study is of great significance for a comprehensive understanding of the changing pattern of PM_2.5 and the formulation of air pollution control policies according to local conditions.