Analysis of Variations in PM2.5 Concentration and Meteorological Factors in Harbin

The combustion of fossil fuels has led to increasingly prominent air pollution problems. A major air pollutant, PM2.5, has been gradually tackled around the world. This study investigated the relationship between PM2.5, a number of influencing factors, and their temporal changes using a machine-learning method in Harbin from 2015 to 2019. It can be seen from the analysis that the random forest model has good performance in predicting PM2.5 concentration. In this model, the mean Relative humidity and aerosol optical depth have a high impact on PM2.5 concentration, but there was negligent correlation with PM2.5. The results indicated that the level of PM2.5 pollution continuously decreased from 2015 to 2019, and there were significant seasonal differences in PM2.5 concentration and its variations. There were significant seasonal differences in the PM2.5 concentration over the past five years. The PM2.5 concentration was highest in winter but declined in a wavelike pattern, slightly lower in autumn, declining in a cliff-like manner, low but stable in spring, and lowest in summer, decreasing gradually year by year. In 2019, due to the impact of heating and adverse meteorological conditions, PM2.5 pollution during the heating period increased significantly. This study provides theoretical and data support for the analysis of PM2.5 pollution in Harbin and formulation of air pollution control policies. Due to the different accuracy of the different data, this article can only be used for calculation, which may cause deviation of the analysis results. In future work, we will continue to improve this issue.