By analyzing the size-resolved chemical composition of non-refractory fine particulate matter obtained at winter 2013-2014 and winter 2018-2019, this study investigated the potential effects of inorganics changes on aerosol water uptake and secondary organic aerosol formation. They found the increased water uptake at larger particles as the aerosol chemical profile shifted from a sulfate-rich to a nitrate-rich. Further, they reported that the enhanced aerosol water uptake at larger particle sizes resulted by the changed aerosol chemical profile, would facilitate the efficient aqueous-phase SOA formation, which provided an interesting perspective to evaluate the role of inorganics and organics in their multiphase processes. Yet, there are some issues that need to be addressed for further improving this work.

1. Line 24-25, it is better to give a specific value here for quantify the “higher relative-humidify”
2. Line 28-30, it is unclear what the implicit relation between SHAP value and the aerosol water.
3. Line 89-91, two kinds of instruments were used for the measurement of NR-PM₅ and its size-resolved chemical composition. Is there any difference for quantification of organic and inorganic components? A brief explanation should be provided here.
4. Line 97-99, the experimentally determined RIEs and standard RIEs were used for different components, why?
5. Line 174-175, in winter 2013-2014, chloride are obviously concentrated on the smaller particle size, while it maintained a relatively smaller and stable contribution in winter 2018-2019. Please explain it.
6. Line 215. As showed in Figure 2, the mass range of NR-PM2.5 in 2018-2019 are much smaller than that in 2013-2014, it is better to explore and compare the variations in the fractions of nitrate and sulfatein the same mass range of NR-PM2.5.  
7. Line 300-305, I agree thatthe fraction of organics in total NR-PM2.5 changed less in winter 2018-2019 compared to winter 2013-2014. However, the fraction of water-soluble organics, which contributed more to ALWC related to total organics, maybe varied a lot.
8. Line 319-325, a clear explanation for the intend implication of theSHAP value should provide first before the discussion on SOA formation.
9. Line 329-332, as the particle mass concentration had been largely reduced in 2018-2019, the gas-particle partitioning of water-soluble organic compounds would be also suppressed as the aerosol surface been decreased. What about the aerosol acidity with increased nitrate fraction? And what the role of aerosol acidity on the SOA formation through multiphase reactions?  

Comments on “Measurement Report: Size-resolved secondary organic aerosol formation modulated by aerosol water uptake in wintertime haze”

This study investigated the factors influencing SOA formation at different particle sizes. Data from two observations (2013–2014, 2018–2019) in Xi’an in winter revealed that the composition of inorganic aerosols changed significantly from sulfate-rich to nitrate-rich at different particle sizes. This transition resulted in changes in aerosol water uptake. Further analysis using random forest and the Shapley additive explanation algorithm (SHAP) elucidated the relative significance of aerosol water in SOA formation, particularly at larger particle sizes. This finding implies that the enhancement of aerosol water uptake at larger particle sizes and high RH may contribute to liquid-phase SOA formation. Before this article was published, there may have been some issues that needed to be corrected as follows:

General comments

• The detailed methodology of PMF analysis as a function of particle size is not clear. Is 3D-matrix or 2D matrix PMF used? How about the time series of the size-resolved PMF factor? Detailed information on how to obtain the best solution for the size-resolved PMF shall be shown.
• Are the fractions of SOA and POA in total OA based on sizes-resolved PMF consistent with the fractions obtained with MS data only? If not, what might cause the differences?
• Section 3.2 the organic aerosols account for a large fraction of total PM masses and the fraction increases with the particle sizes get higher. Thus, the aerosol water uptake contributed by organics cannot be ignored and shall be considered here. E.g., method in Guo et al. (2015).
• The methodology for SHAP analysis is not very clear as well. The validation of the output results from the machine learning model shall be displayed.
• The meaning of Figure 6 is not easy to follow for readers who are not farmilar with SHAP model, and more explanation is needed here. For the water in Figure 6b, it seems that the water contributes negative SOA formation when the water mass concentration is low and positive SOA formation when the water concentration is high. How to explain this? The importance of water to SOA formation seems to decrease from “225-317 nm” to “447-631 nm” and then increase when the particle size increases. What might cause this?
• How to explain the temperature effect on SOA formation in Fig. 6. More explanation is needed here since temperature is the most important factors which influence the SOA formation here.

Minor comments

• Line 121: How many factors are the signals with SNR with 0.2 <SNR<3 downweighed?
• Line 142: The formula is missing a sequence number.
• The color code of Fig. 2 is not very clear, especially for the small particle size (112-631nm), please modify it so that the readers can see it more clearly.
• Line 90, Is eptof data used here or the regular PToF data? Please clarify.

References:

Guo, H., Xu, L., Bougiatioti, A., Cerully, K.M., Capps, S.L., Hite, J.R., Carlton, A.G., Lee, S.H., Bergin, M.H., Ng, N.L., Nenes, A., Weber, R.J., 2015. Fine-particle water and pH in the southeastern United States. Atmospheric Chemistry and Physics 15, 5211-5228.