Estimation of interactive controls and threshold levels of ozone under traffic-influenced near-road conditions

Ground-level ozone (O₃) pollution is an escalating concern for air-quality management in regions where rapid growth in traffic emissions combined with favorable meteorological conditions frequently lead to exceedances and increased health risks. In this study, we examined the drivers of O₃ variability using synchronized hourly measurements of traffic flow, meteorological variables, and gaseous air pollutants collected at a roadside site on National Highway-16 in Kharagpur, India. To resolve the nonlinear, threshold, and interaction controls on O₃ across contrasting seasons, we applied a Random Forest (RF) regression model coupled with SHapley Additive exPlanations (SHAP). We uncover a clear regime shift in O₃ drivers: O₃ variability in Summer is dominated by NO-driven titration and emission-chemistry coupling whereas O₃ in Autumn is primarily governed by temperature-dependent photochemistry. Distinct thresholds define rapid transitions in O₃ response, including NO (∼17-19 μg m^-3 in summer; ∼12-13 μg m^-3 in autumn), temperature (∼33°C vs. ∼26-27°C), and relative humidity (∼60%). Beyond these tipping points, O₃ sensitivity reverses or amplifies, highlighting strongly nonlinear system behavior. Interaction analysis reveals that O₃ formation is controlled by synergistic coupling between precursors and meteorology with NO-temperature and temperature-humidity interactions dominating in summer and consistent with a VOC-limited regime. In autumn, interaction strength weakens and shifts toward meteorological control. These results provide direct and observation-constrained evidence that O₃ formation pathways are seasonally reconfigured through nonlinear thresholds and interaction effects. This framework advances predictive understanding of traffic influenced O₃ variability and enables more effective, season-specific mitigation strategies under evolving emission and near-road conditions.