Impact of Meteorology on PM 2.5 and O3 Pollution in Three Provincial Cities in Northeastern Provinces of China

In recent years, China has faced severe air pollution of fine particulate matter (with an aerodynamic diameter of smaller than 2.5 µm, PM2.5) and ozone (O3). It is importance to investigate the driving factors to air pollution. Numerous studies have emerged on the impact of meteorology on PM2.5 and O3, but most of them have concentrated on the highly developed and heavily populated areas of eastern China. There is insufficient research on PM2.5 and O3 pollution in the three northeastern provinces of China. To further understand the effect of meteorology on PM2.5 and O3 in the three northeastern of China, back-propagation neural network (BPNN) and random forest (RF) was used in this study. Firstly, meteorological factors (temperature, relative humidity, precipitation, wind speed, pressure and planetary boundary layer (PBL) height, cloud cover) were put into BPNN and RF models to investigate the influence of weather conditions on PM2.5 and O3 during 2015–2021. Results showed that RF performed better than BPNN. Among meteorological factors, precipitation and PBL significantly affect PM2.5; temperature and pressure (PRS) significantly affect O3. Then, when pollutants factors, such as PM2.5, O3, carbon monoxide (CO), sulfur dioxide (SO2), nitrogen dioxide (NO2) and formaldehyde (HCHO) were also inputted into BPNN and RF, it was illustrated that prediction accuracy of them had been improved. This study also revealed that the contribution of meteorology to PM2.5 and O3 were 16.73