Efficient Diagnosis of Spatiotemporal Evolution and Driving Factors of Surface Ozone Pollution Episodes: An Application in Jiangsu Province, China

Severe ozone (O₃) pollution is a major challenge for air quality improvement in China, primarily due to its spatiotemporal heterogeneity and nonlinear formation regimes. Here, we developed a novel framework to illustrate the spatiotemporal evolution of pollution episodes and to identify their recurrent driver modes and formation sensitivities in Jiangsu Province during the warm season from 2020 to 2023. The framework integrated joint analysis of the synoptic weather type (WT) and the regional spatiotemporal pattern (SP) of O₃ episodes, explainable machine learning, and geostationary satellite observation of O₃ precursors. We showed that O₃ pollution episodes occurred preferentially with a limited number of WT × SP combinations, particularly those involving subtropical high-pressure systems coincident with pollutant accumulation patterns in southern Jiangsu. Three key driver modes (thermal-driven, radiation-moisture synergy, and transport-accumulation) were identified, with significant associations with specific WT × SP combinations. The transport-accumulation mode commonly corresponded to the VOC-limited regime, while thermal-driven and radiation-moisture synergy modes shifted O₃ formation toward the NO_x-limited regime more frequently. This work demonstrated an effective, efficient, and adaptable tool of detecting the source and evolution of multiple O₃ episodes and could support flexible and timely design of the pollution control strategy, with an improved understanding of heterogeneous and changing causes of O₃ episodes.