Scenario-dependent O3 driver identification and formation responses to VOCs/NOx emission reduction pathways from integrated and sectoral sources in the Beijing-Tianjin-Hebei region, China.

Persistent surface ozone (O3) in China's Beijing-Tianjin-Hebei region, driven by complex meteorology-precursor interactions, has become a critical air quality crisis. A longstanding challenge in O3 pollution mitigation is identifying drivers of O3 formation and its response to precursor emissions. Using explainable ensemble machine learning, scenario-based modeling and empirical kinetic modeling approach, we quantified O3 formation drivers and evaluated scenario-dependent responses to integrated/sectoral precursor emission reductions across multiple dynamic scenarios. We demonstrate that temperature, solar radiation, and nitrogen dioxide dominate O3 formation across pollution levels. Meteorological conditions for O3 dispersion varied significantly between 2013-2023, with 2021/2023 representing the most favorable/unfavorable years, highlighting strong interannual variability. Spatiotemporal O3-precursor sensitivity exhibited volatile organic compound (VOC)-limited dominance, modulated by dynamic meteorology and pollution levels. While Only-VOCs reductions significantly decreased O3, they have inevitable limitations. VOCs/nitrogen oxides (NOx) = 3:1 pathways enhanced O3 mitigation in Beijing and Shijiazhuang but remained suboptimally efficient. Notably, deepening NOx reductions promoted a shift toward NOx-limited regimes, further improving the efficiency of O3 decrease. Extreme meteorology and heavy pollution increased O3 reduction potential compared to baseline scenarios, with industrial VOCs control proving most effective for sectoral interventions. These findings inform scenario-specific precursor emission policies to address severe O3 pollution.