Deep learning-derived anthropogenic and meteorological drivers of surface ozone change in China

Urban air pollution continues to pose a significant health threat, despite regulations to control emissions. Here we present a comparative analysis of the anthropogenic and meteorological drivers of surface ozone (O3) change in China by integrating deep learning (DL) and chemical transport model (CTM) methods. The DL method suggests volatile organic compound (VOC)-limited regimes in urban areas over northern inland China in contrast to strong nitrogen oxides (NOx)-limited regimes in GEOS-Chem simulations. Sensitivity analysis indicates that the inconsistent O3 responses are partially caused by the inaccurate representation of O3 precursor concentrations at the locations of urban air quality stations in the simulations. The DL method exhibits possible weakened anthropogenic contributions to surface O3 rise in the North China Plain, for example, 1.53 and 0.54 ppb/y in 2015-2019 and 2019-2021, respectively. Similarly, GEOS-Chem simulations suggest an accelerated decrease in surface O3 concentrations driven by the decline in nitrogen dioxide (NO2) concentrations. Furthermore, both DL and GEOS-Chem models suggest the reverse of meteorological contributions to the observed O3 change in the North China Plain in 2019-2021, which is mainly resulted from the reversed changes in meteorological variables in surface air temperature and relative humidity. This work highlights the importance of DL as a supplement to CTM-based analysis. The derived O3 drivers are helpful for making effective regulatory policies to control O3 pollution in China.