This study focused on the relationship between large-scale circulation patterns and daily PM2.5 concentrations, and their sensitivity to emissions in YRD over 1999–2019 based on UKESM1 model. The authors constructed an index and projected future changes in PM2.5 under SSP3-7.0 scenario and found the effect of climate change would partly offset that of emissions. The manuscript was well-written and properly organized, which was within the scope of ACP. However, there are also some major issues that have to be addressed before it can be considered for publication. Please see my detailed comments below.

General comments:

1.    The results of this study are primarily based on UKESM1 model simulations. Although the authors compared the modeled PM2.5 concentrations with those from CAQRA reanalysis data, they have not verified whether the model could well reproduce the circulation conditions during the polluted days. The modeled PM2.5 only covers 20 years and the nationwide PM2.5 observations have already had 10-year data. Comparing the model simulation and observations in term of large-scale circulation and the index defined can give a more robust conclusion.

2.    The future predictions of climate change impact of PM2.5 is investigated in this study based on changes in meteorological fields under SSP3-7.0 pathway. First of all, SSP3-7.0 is not a representative scenario for future air quality or climate change, at least in China. The scenario assumes the anthropogenic emissions of air pollutants continue to increase for a long time after 2015, but the emissions in China have significantly reduced since 2010s, which largely affect regional climate and cause the inaccurate of regional climate under SSP3-7.0. Also, China has committed to achieve carbon neutrality in 2060 and the results under the low forcing scenarios should be considered or discussed.

3.    In addition, for the future predictions, how is the UKESM1 performed compared to other climate models in CMIP6. Different models tend to predict different regional circulation response. Does the conclusion that “a weaker pressure gradient between the Siberian High and the Maritime Continent Low” also exist in other CMIP6 models?

Specific comments:

1.    Many studies have examined the circulation pattern and regional transport of air pollution over eastern China from the past to the future and they have similar or different conclusions. For example, Ren et al. (2021) quantified the sources of PM2.5 in many subregions of China and they found that PM2.5 pollution in eastern China is dominated by local emissions using an aerosol source tagging technique in an aerosol-climate model. Yang et al. (2021) examined the atmospheric circulation patterns conducive to severe haze in eastern China based on observations, modeling results and CMIP6 future predictions. They found that during the extreme pollution month the PM2.5 was mainly from aerosol transport from the North China Plain, although they also reported a future increase in the atmospheric circulation pattern conducive to the pollution under high forcing scenarios. Li et al. (2022) also highlighted the importance of climate change in regulating future air quality. They found that climate-driven aerosol changes are comparable to those contributed by changes in emissions over many regions of the world in high forcing scenarios. The authors are suggested to compare their results with previous studies.

2.    Why only SSP3-7.0 scenario is selected? 

3.    Will the lack of SOA affect the conclusion, since that the circulation pattern is accompanied by temperature/relative humidity changes, affecting the formation of aerosols?

4.    I noticed the authors have published a very similar paper in ACP using observations (Jia et al., 2022). They should clarify the new scientific findings in this study rather than the data used (model results and observations). The index defined in this study is not the same as Jia et al. (2022). Does that mean the model and observations will draw different results?

5.    There are many uncertainties in this studies that should be discussed. I strongly recommend the authors to add a discussion section. For example, the model has biases in reproducing mean aerosol concentrations and the correlation coefficients reported in this study are not high enough. These may influence the results. The

References:

 

Ren, L., Yang, Y., Wang, H., Wang, P., Chen, L., Zhu, J., and Liao, H.: Aerosol transport pathways and source attribution in China during the COVID-19 outbreak, Atmos. Chem. Phys., 21, 15431–15445, https://doi.org/10.5194/acp-21-15431-2021, 2021.

 

Yang, Y., Zhou, Y., Li, K., Wang, H., Ren, L., Zeng, L., et al. (2021). Atmospheric circulation patterns conducive to severe haze in eastern China have shifted under climate change. Geophysical Research Letters, 48, e2021GL095011. https://doi. org/10.1029/2021GL095011

 

Li, H., Yang, Y., Wang, H., Wang, P., Yue, X., and Liao, H.: Projected Aerosol Changes Driven by Emissions and Climate Change Using a Machine Learning Method, Environ. Sci. Technol., 56, 7, 3884–3893, https://doi.org/10.1021/acs.est.1c04380, 2022.

Jia, Z., Doherty, R. M., Ordóñez, C., Li, C., Wild, O., Jain, S., and Tang, X.: The impact of large-scale circulation on daily fine particulate matter (PM2.5) over major populated regions of China in winter, Atmos. Chem. Phys., 22, 6471–6487, https://doi.org/10.5194/acp-22-6471-2022, 2022.

This study investigated the influence of large-scale circulation on daily PM2.5 concentrations in China during winter using simulations with an Earth system model. Indexes based on circulation patterns corresponding to regional PM2.5 levels were proposed by the authors to demonstrate the circulation impacts on PM2.5 for both history and the future, especially for YRD. Sensitivities of regional pollution to emission reduction were also quantified. In general, the results reported in this study are interesting and have reasonable implications for effective PM2.5 projections. However, some clarifications and major revisions are suggested before it can be considered for publication.

General comments:

1. The authors published a highly relevant paper on ACP earlier this year, using the same model and a very similar analysis approach. However, the circulation-based indices defined in the two studies differ from one another, both in variable choices and key region detections. The authors claimed that the differences might be attributed to the different time spans. Does it mean that the index definition is sensitive to years and models? What do the uncertainties originate from, model instabilities or interannual-to-decadal variability of the relationship between PM2.5 and circulations? What are the dominant variables (e.g. SLP, V850, Z500, etc.) in characterizing the day-to-day variability of PM2.5 over different regions? Without a consistent definition of indices, the implication would be quite limited.
2. The projections of the circulation-based indices in the future should be ensembled from more CMIP models to test the robustness of the trend. Besides, more SSP-RCP scenarios should be considered when evaluating the possible response of PM2.5 both to circulations and emissions, since the authors consider exploring the daily pollution responses to climate change in the title.

Specific comments:

1. Why is the correlation map shown in Figure 2 different from Figure 1 in Jia et al., 2022, especially for YRD? The PM2.5 levels in YRD were closely connected with maritime airmasses in Jia et al., 2022, but correlated better with a broader region of inland in 1999-2018.
2. I notice that the model failed to capture some heavily polluted episodes for the three regions, even for the year 2014. Since the indices and composite analysis are all based on the model simulations, would it significantly impact the results?
3. In Line 20-23, Page 8, why are heavily polluted days impacted more strongly by emission reductions? In addition to the absolute levels change (from 3.1ug/m3 to 8.5 ug/m3 from local contribution for example), the relative values with respect to daily mean also increase (from 6.7% to 15.8% in this case). How to explain it?
4. In section 4 the authors claimed that the index over YRD could “distinguish effectively between different levels of air pollution”, especially for heavily polluted and clean conditions. While in section 5, the PM2.5 response to emission reductions in I<-1 days (Line 28, Page 8) shows similar values to the situation when considering winter daily mean PM2.5 (Line 18, Page 8), but largely below the actual polluted conditions (Line 23, Page 8). How to address the contradiction?
5. In Figure 9(b) and 9(f), why do reductions in emission over YRD lead to a PM2.5 increase over coastal regions? The emission and compositions in PM2.5 changes are suggested to be investigated since it shows large nonlinearity.
6. Would it be possible to include the PM2.5 projection in future scenarios in Figure 10 or an additional figure?