Exploring PM2.5 and O3 disparities and synergies management through integrated natural and sociology-environmental drivers in the YRD

Identifying the main factors influencing PM_2.5 and O₃ concentrations is crucial for effective pollution control in urban areas. By combining optimal parameter-based geographical detector (OPGD) and multi-scale geographically weighted regression (MGWR) models, this study reveals the underlying mechanisms of PM_2.5 and O₃ spatial variation in the Yangtze River Delta (YRD), focusing on both natural and socioeconomic indicators. The OPGD model optimized the spatial scale and zoning of geographic data, enhancing the accuracy of identifying PM_2.5 and O₃ drivers compared to conventional methods. The results showed that the optimal spatial scale of PM_2.5 and O₃ concentrations in this study region was 9 km. Optimal discrete parameter combinations for most socioeconomic factors were quantile breaks with 9 intervals, while Natural Breaks or equal breaks were more suitable for natural factors. Both natural factors, such as precipitation, wind speed, dew point, temperature, solar radiation, and elevation, and anthropogenic factors, including land use types and vehicle numbers, were key drivers of variations in PM_2.5 and O₃ concentrations over the years. Combined natural and socioeconomic factors significantly enhanced the explanatory power of PM_2.5 and O₃ concentrations. The MGWR model’s fit for key factors was highest in spring, with adjusted R² values for PM_2.5 and O₃ both exceeding 0.8, indicating that the coordinated management of these pollutants should prioritize spring, particularly in areas with low wind speed, where wind interacted non-linearly with most of factors, strongly influencing PM_2.5 and O₃ variation. Even in summer, when O₃ and PM_2.5 concentrations differed significantly, elevation and land use types each explain over 40% of the variance. This suggests that optimizing land use structures in low-altitude urbanized areas and enhancing local dispersion conditions could improve air quality. However, during autumn and winter, no significant common factor was found to explain the variation in both PM_2.5 and O₃ concentrations. Vegetation-related factors, such as the Normalized Digital Vegetation Index (NDVI) and urban green coverage ratio, though weak individually, exhibited strong nonlinear interactions, highlighting their indirect role in pollutant dynamics, especially for O₃ in colder months and PM_2.5 during spring and summer. This study underscores the necessity for region-specific air pollution regulations to consider both natural and social factors across various time scales.