Decoding the urban-suburban ozone disparities in Shandong, China: A machine learning-driven attribution analysis (2019-2023).

Urban-suburban disparities in ozone (O3) pollution have received growing attention due to their implications for environmental equity and public health. However, the underlying drivers of these disparities remain poorly understood, especially the contrasting mechanisms between coastal and inland cities. Based on observations at 267 air quality monitoring sites in Shandong during 2019-2023, spatiotemporal variability of O3 and the drivers of the urban-suburban disparities were explored through the de-weathered method, ensemble machine learning models, and SHAP analysis. Results show that both the maximum daily 8-hour average (MDA8) and mean O3 concentrations were consistently higher in urban areas than in suburban areas, with the differences particularly pronounced during summer and nighttime. Between 2019 and 2023, urban-suburban disparities in MDA8 (Δ[MDA8]) and mean O3 (Δ[O3]) increased by 0.4 and 1.0 μg/m3, respectively. Variations in emission sources were identified as the primary drivers of the increase in Δ[O3] across 16 cities, accounting for 84 % of the observed changes, with nitrogen dioxide (NO2) being the most influential contributor. Furthermore, disparities in PM2.5 further amplified Δ[O3] in coastal regions. Superimposed on anthropogenic emissions, meteorological factors that intensified urban-suburban differences in surface temperature and wind speed were associated with increased Δ[O3] in coastal cities, whereas variations in mean sea level pressure were strongly linked to decreased Δ[O3] in inland cities. Therefore, strengthening NOx emission controls in inland areas and enhancing PM2.5 mitigation in coastal cities are expected to reduce O3 disparities across and within cities, thereby improving overall air quality in Shandong Province.