The manuscript by Yang et al. analyzes the meteorological characteristics and processes of severe ozone pollution events in four major cities in China using the GEOS-Chem model, and examines the historical and future changes of the ozone-favorable meteorological characteristics from ERA5 re-analyses and CMIP6 simulation results. The topic is within the scope of ACP. The methods are valid and the findings are valuable for understanding the historical and future change of ozone pollution in China. In general, it is a concise and well-written manuscript. I recommend publication after some clarification.

1.    Introduction: I recommend further introducing what is missing in the above studies and what this study adds on then.

2.    Table 1: It is quite interesting to see that ozone chemical net production is negative in Chengdu and Guangzhou. This requires some further discussions. Showing some detailed patterns of these diagnostics terms might be helpful. Is it possibly due to the relatively coarse resolution of the simulation (2x2.5)? It might be helpful to discuss the possible influence of model resolution on model diagnostics.

3.    Figure 8 (and future projection) uses the spatial correlation of SLP and GPH for each year with that in the targeted ozone pollution month to examine the similarity of weather patterns. While this is mostly sound and efficient, I wonder whether the magnitude of SLP and GPH should be considered. In addition, it should be clear whether the correlation is calculated for each month in 1980-2020 to compare with the targeted ozone pollution month.

4.    Line 161：misspelling of “simulation” 

The authors analyzed meteorological anomalies in June 2018, July 2017, September 2019 and July 2015 corresponding to severe O3 pollution events in the NCP, YRD, PRD and SCB, respectively. The four meteorological variables (T2M, RH, SLP and GPH) which are associated with severe O3 pollution events in these four months are then analyzed to investigate their changes in 1980-2019 and 2021-2100, and their changes are predicted to lead to more severe O₃ pollution in the future. The topic is interesting, and the conclusion is important for better control of O3 pollution. However, there are still issues that need to be addressed before the paper can be considered for publication.

Comments:

• Lines 28-31: The readers may assume the conclusion is based on the most severe O3 pollution months over these regions, but they are June 2018, July 2017, September 2019 and July 2015 for the NCP, YRD, PRD and SCB, respectively, that is, one month per region. It is unclear whether one month can represent the meteorological conditions of severe O3 pollution events sufficiently.
• Furthermore, the severe O3 pollution events in the SCB are interpreted as regional transport from Northern China (as shown in Figure 5b). Consequently, meteorological variables SLP and GPH were selected to describe severe O3 pollution in the SCB. However, it is hard to believe regional transport from Northern China could be the dominant factor for severe O3 pollution events in the SCB which has strong local anthropogenic emissions and is surrounded by mountains. In lines 160-162, the authors indicated “In the top three O3 polluted months in Chengdu, only in July 2015 the higher concentrations than the long-term averages can be captured by the simulations with fixed emissions”, and it is the reason why July 2015 was selected to represent the SCB in the analysis. However, does this imply inaccuracy in the model simulations to interpret severe O3 pollution events in the SCB? Are SLP and GPH suitable meteorological variables to describe severe O3 pollution events in the SCB in the past and the future?
• Lines 152-155: Why is Chengdu not mentioned in this sentence?
• Figure 2: How are these anomalies calculated?
• Figure 3a: Is it 850 hPa (figure title) or 1000 hPa (caption) wind anomalies? The same question for Figures 4-6.
• Section 3.2: The discussion in this section is simple. It could be better to provide a quantitative comparison of the impacts of each meteorological variable on the high O₃
• Figure 8c-d: These two panels are associated with the conclusion “In SCB and PRD, the occurrence of atmospheric circulation patterns similar to those during the polluted months increased, together with the more frequent hot and dry conditions, contributing to the increases in severe O3 pollution in SCB and PRD during 1980–2019”. First, it is suggested to add uncertainty estimates particularly for July 2015 (SCB) to determine whether there are significant trends in the meteorological anomalies. Moreover, is it possible to have quantitative estimations for the possible contributions from meteorological changes to O3 changes in 1980-2019 to confirm the importance of meteorological changes?
• Section 7: It could be helpful to provide a comparison of the impacts of meteorological changes in 1980-2019 and 2021-2100 to extend the discussion in this section.