Exploring the spatial heterogeneity and temporal homogeneity of ambient PM10 in nine core cities of China

Abstract

We focus on the causes of fluctuations in wintertime PM₁₀ in nine regional core cities of China using two machine learning models, Random Forest (RF) and Recurrent Neural Network (RNN). RF and RNN both show high performance in predicting hourly PM₁₀ using only gaseous air pollutants (SO₂, NO₂ and CO) as inputs, showing the predominance of the secondary inorganic aerosol and implying the existence of thermodynamic equilibrium between gaseous air pollutants and PM₁₀. Also, we find the following results. The correlation of gaseous air pollutants and PM₁₀ were more relevant than that of meteorological conditions and PM₁₀. CO was the predominant factor for PM₁₀ in the Beijing-Tianjin-Hebei Plain and the Yangtze River Delta while SO₂ and NO₂ were also important features for PM₁₀ in the Pearl River Delta and Sichuan Basin. The spatial heterogeneity and temporal homogeneity of PM₁₀ in China are revealed. The long-range transported PM₁₀ was substantiated to be insignificant, except in the sandstorms. The severity of PM₁₀ was attributable to the lopsided shift of thermodynamic equilibrium and the phenology of indigenous flora.