Assessment of Emissions and Meteorological Contributions to PM_2.5 and O₃ Concentration Changes in a Typical Industrial City Based on Random Forest and SHAP Analysis

The reduction of pollutant concentrations is influenced not only by emission control measures but also by meteorological changes. It is therefore necessary to decouple meteorological factors to separately evaluate the impacts of emissions and meteorology on pollutants. Taking Jincheng, a typical industrial city, as an example, this study employs a machine learning method (Random Forest, RF) combined with SHAP analysis to quantitatively assess the contributions of emissions and meteorology to PM<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</inf> and O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf> concentrations during 2017 and 2020–2023, exploring the driving mechanisms behind pollutant concentration changes. The results show that both PM<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</inf> and O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf> concentrations were primarily driven by emissions, with 67%-70% of PM<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</inf> changes and 51%-53% of O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf> changes attributed to emissions of precursors. In 2021, emissions significantly contributed to PM<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</inf> and O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf>, with SO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>, NO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>, and CO accounting for 5.5%, 16.7%, and 33.2% of PM<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</inf> changes, respectively, and NO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> contributing 27.7% to O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf>. In 2022, PM<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</inf> changes were jointly driven by emissions and meteorological factors, while the influence of meteorology on O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf> intensified, with temperature (31.1%) and relative humidity (8.2%) making notable contributions. By 2023, the impact of meteorological factors further increased, with relative humidity contributing 15.0% to PM<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</inf> and net solar radiation peaking at 3.9% for O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf>. Additionally, the contributions of emission sources on polluted days were significantly higher than on clean days, as stronger net solar radiation enhanced secondary O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf> formation, while lower boundary layer heights and wind speeds exacerbated ground-level pollutant accumulation. This study provides scientific support for precise pollution control in industrial cities and regional air quality improvement. It also highlights the need for further optimization of models to investigate the secondary formation and nonlinear changes of particulate matter.