Influencing factors analysis and prediction of near-surface ozone in Henan Province from 2015 to 2022

Abstract. This study analyzed factors influencing near-surface ozone (O₃) in Henan Province from 2015 to 2022 using real-time pollutant data from the China National Environmental Monitoring Centre and daily meteorological data from the Henan Provincial Ecological Environment Monitoring and Safety Center. Regression and machine learning models (including multiple linear regression (MLR), support vector machine (SVM), random forest (RF), ridge regression (RR), BP neural network, and extreme gradient boosting (XGBoost)) were used to predict O₃ concentrations. The results showed that among the major pollutants (CO, NO₂, SO₂, PM_2.5, and PM₁₀), there was a consistent negative correlation with O₃. Notably, NO₂ had the strongest negative correlation (r = -0.825), while PM₁₀ showed the weakest (r = -0.687). From the perspective of meteorological factors, temperature showed a strong positive correlation with O₃, while wind speed, relative humidity, and precipitation showed weak negative correlations, influencing regional variations in O₃ concentrations. Among the six prediction models constructed to predict O₃ concentrations, the most accurate model for predicting concentrations for the next day was the extreme gradient boosting (XGBoost) model (R² = 0.883). For the next 3 days, the random forest (RF) model demonstrated the highest accuracy (R² = 0.704). Similarly, the random forest model (RF) also exhibited the highest accuracy for predicting the next 7 days (R² = 0.651). In summary, over the past 7 years, there has been a strong correlation observed between O₃ concentration and other major pollutants, as well as meteorological factors in Henan Province. Therefore, it is essential to implement targeted measures for O₃ pollution prevention and control based on specific weather conditions.