Under the backdrop of climate change, various countries/regions across East Asia have witnessed severe ground-level ozone (O₃) pollution, which poses potential health risks to the public. The complex relationships between O₃ and its drivers, including the precursors and meteorological variables, are not yet fully understood. Revealing the formation regime of O₃ is crucial for providing evidence-based information for pollution control. In the present study, we evaluated the influence of key chemical  (e.g., volatile organic compounds, PM_2.5, NO_x) and meteorological drivers (e.g., air temperature, relative humidity) on ground-level O₃ pollution at Tucheng site in New Taipei, Northern Taiwan, using fine-resolution atmospheric composition measurements and machine learning. The developed random forest machine learning models performed well, with 10-fold cross-validation R² values above 0.867. The results reveal seasonal disparities on the formation regimes of ground-level O₃ between winter and summer. Chemical drivers contributed 82.4% and 62.1%, respectively, to O₃ concentrations in winter and summer. Based on the random forest models, temperature, 1,2,3-Trimethylbenzene, NO_x, t-2-Butene, and relative humidity were identified as the dominant drivers of O₃ formation. The machine learning-based modelling framework developed in this study can be easily adapted to new sampling sites with minor modifications if necessary.

O3; VOCs; Meteorology; Machine learning; Taiwan