This paper details work on implementing a HOM chemistry scheme into a large-scale model and comparing the results with an updated inorganic-only nucleation scheme. The inclusion of organic nucleation and growth at small particle sizes due to HOM formation leads to better agreement overall with measurements of high altitude CCN and frequency of NPF events globally. This work provides an interesting look into incorporating HOM chemistry into global models. The results seem reasonable and fit within the scope of ACP. My main concern stems from a lack sensitivity studies surrounding many of the uncertainties in the mechanism as well as a lack of discussion of the limitations of the mechanism. Given HOM chemistry is an active area of research, it makes these results incredibly significant to the community, but also means communicating clearly the limitations of the work given the present understanding of HOM chemistry is all the more important. Overall, I think this paper presents an important contribution to the field and I would support publication if the following comments are addressed.

Major comments:

1. The sensitivity studies done in this work seem informative, but more should be done on the other unknown parameters related to HOM formation such as the autoxidation rate coefficient, the temperature-dependence of this rate coefficient, and the dimerization rate coefficients.
2. Additionally, Liu et al (2024) identifies the branching in the NO termination pathway as being highly uncertain, but here the sensitivity to the rate of NO reaction is investigated. Why was the sensitivity to the rate rather than the branching ratio studied?
3. In Table 1, it is shown that all C20 and C15 compounds are represented by one species each with one volatility each. As seen in previous work (Stolzenburg 2018, Ye et al 2018, Schervish and Donahue 2020, etc) not all accretion products lead to ULVOCs or even ELVOCs. The assumption that they do seems like it would dramatically overestimate the role of organic nucleation and growth.
4. While the low branching ratio to accretion products may somewhat account for the concern brought up in point 3, experimentally many C20s end up in the E/LVOC range, allowing them to contribute to small particle growth, and the mechanism seems to indicate products from C10+C10 accretion reactions can only be C20 ULVOCs or non-HOM species.
5. Why are only 2 steps of autoxidation simulated? Laboratory evidence (Heinritzi et al 2020, Simon et al 2020, etc) shows products with very high oxygen content that likely underwent more than 2 steps of autoxidation. Would allowing more autoxidation lead to a higher organic contribution to nucleation, or perhaps this model step up could suggest how many steps are likely to occur in the actual atmosphere prior to termination or condensation.
6. Are any particle-phase processing of HOMs considered such as particle phase oligomerization or decomposition of accretion products and organic hydroperoxides?
7. In Liu et al (2024), it seems that the HOM chemistry mechanism overpredicts surface HOM mass concentrations. While, I agree with the argument in that work that this could be due to limited detection of HOMs in the observations, more recent lab work (including Stolzenburg et al 2018  referenced earlier) has focused on closing that gap and could be simulated as well. It seems difficult to justify moving forward in the current work with this mechanism without more thorough validation.
8. Overall, I think much more discussion needs to be provided into what the uncertainties in this model are and how they limit the abilities and applications of this mechanism in the context of this paper. For example, Liu et al (2024) finds that temperature is a significant parameter affecting HOM formation, but this work does not discuss at all the temperature-dependence of autoxidation and how this uncertain parameter can affect the results in this work.

Minor comments

1. The title describes the chemical mechanism as “explicit.” This to me indicates that you are representing the full chemistry with specific chemical species. However radical species are lumped leading to a reduced, perhaps, semi-explicit mechanism. I would consider editing the title to make that clearer.
2. Line 30-31: What does this sentence mean? I think this is referencing the finding that turning off organic initial growth leads to a greater decrease in the aerosol number, but the wording is very confusing.
3. A more detailed description of the mechanism should be provided with important assumptions made clear as a reference to another paper does not provide enough context.
4. Table 1: Are these C* (300 K) values?
5. Eqn 7-8: Are all accretion products considered as part of [ACC]? Heinritzi et al (2020) showed that the formation of C15s from isoprene and monoterpene cross-reactions led to lower nucleation rates suggesting these do not participate in pure organic nucleation, (at least under (organic) supersaturation conditions in that work).
6. Paragraph at line 260: Can this analysis be made clearer? Why does an overestimation of H2SO4 lead to and underestimation in growth rate overall, but then also explains overestimated growth rates in specific cities?
7. Over- or underestimations in condensation sink are occasionally used to justify a corresponding under- or overestimation in NPF. This makes sense, however, can CS be prescribed in order to validate this?
8. Figure 4c,d can a label and the units be places under the x-axis?
9. Line 340: Is the “organic nucleation rate” mentioned here just the sum of the neutral (J_Org,n) and ion-induced pure organic nucleation (J_Org,i) or does it also include inorganic-organic nucleation (J_SA-Org)?
10. Figure 9: There are 2 panels labeled g.
11. Figure 9: There is only 1 unit given in the caption, but the left and right plots appear to have different units.
12. Figure 10: There is reference in the caption to up to panel h, but the figure only contains up to panel d.
13. If Liu et al (2024) is not published, please provide the full mechanism in this paper.

References

Liu, Y. et al., A Modeling Study of Global Distribution and Formation Pathways of Highly Oxygenated Organic Molecules Derived Secondary Organic Aerosols (HOMs-SOA) from Monoterpenes, J. Geophys. Res.-Atmos., (under review), 2024

Stolzenburg, D. et al, Rapid growth of organic aerosol nanoparticles over a wide tropospheric temperature range PNAS, 2018, 115, 9122–9127, https://doi.org/10.1073/pnas.180760411  

Ye, Qing et al Environmental Science & Technology 2019 53 (21), 12357-12365 DOI: 10.1021/acs.est.9b03265

Schervish, M. and Donahue, N. M.: Peroxy radical chemistry and the volatility basis set, Atmos. Chem. Phys., 20, 1183–1199, https://doi.org/10.5194/acp-20-1183-2020, 2020.

Heinritzi, M., et al Molecular understanding of the suppression of new-particle formation by isoprene, Atmos. Chem. Phys., 20, 11809–11821, https://doi.org/10.5194/acp-20-11809-2020, 2020.

Simon, M., et al Molecular understanding of new-particle formation from a-pinene between −50 and +25 °C, Atmos. Chem. Phys., 20, 9183–9207, https://doi.org/10.5194/acp-20-9183-2020, 2020.

A version of the CESM climate model with state-of-the-art new particle formation (NPF) mechanisms is presented, with a focus on the production of Highly Oxygenated organic Molecules (HOMs). The model demonstrates improved agreement with observations. The authors find that organic molecules play a more important role in global NPF than previous studies suggested. Table 5 suggests 83.44% of nucleation proceeds via the mixed H2SO4-organic pathway below 5.8km, a result that, if nothing else, highlights the importance of further studying this possible NPF pathway.

While not emphasized in the paper, the authors also include an upgraded inorganic NPF mechanism, a potentially very useful innovation.

The article documents a significant effort and it is novel for this level of complexity in new particle formation to be included in a global climate model. The analysis and model evaluation is of high quality with some useful innovations such as the NPF event threshold.

I recommend the paper for publication, subject to responses to the comments below. I also appreciate that, while I do suggest some more sensitivity studies, it is surely not within the scope of the paper to explore all possible uncertainties, as long as the limited nature of the sensitivity studies is properly discussed.

Major comments

1. What are the uncertainties associated with using the Kerminen and Kulmala (2002) approximation for aerosol growth rates up to 20nm in diameter? Many models, including some GCMs that participated in CMIP6, resolve aerosols prognostically and advect them at much smaller sizes. It seems that using the Kerminen and Kulmala (2002) formula in this way for precise studies of NPF might incur quite large biases. Growth rates are not constant with size (Stolzenburg et al, ACP 2020, e.g. Eq. 16), and the more accurate version of the equation that takes better account of the size dependence of the loss rates is given by Lehtinen et al (J. Aerosol Science 2007). I suspect reasonable bounds on these potentially large uncertainties could be determined with some sensitivity studies, without changing the structure of the model. Also, (L173) should the 0.66 factor be modified between default and updated NPF mechanisms to account for the difference between the 1.7nm starting size for the updated scheme and the 1nm starting size for the default scheme? I believe the parameterization depends on (1/1.7 – 1/20) or (1/1 – 1/20) so this likely makes quite a big difference.

         2.   The model evaluation has some missing details. How were number concentrations of 10nm particles calculated in the model for comparison with measurements? At what temporal frequency were variables written out of the model for calculating nucleation rates and aerosol number concentrations? What kind of interpolation was done to match model with observation stations and aircraft measurements?

        3 . At line 330, I think the results suggest most ACC does not undergo autoxidation even though it has 8 or 9 oxygen atoms (Table 1), because ACC are unaffected by the NO sensitivity study. That may be possible, though it seems rather surprising. Also I would imagine ACC which are autoxidized will usually have more oxygen han those which don’t, so will be more likely to be ULVOC and nucleate. Either way it doesn’t make sense to me that uncertainties in organic chemistry (ACC as well as HOM) do not affect CCN number in Amazonia where pure biogenic NPF dominates – it seems more likely that the authors were unable to sample the uncertainty space sufficiently, and this could be discussed more.

Indeed, I would speculate despite the authors’  findings, CCN formation in general could still be very sensitive to either the chemical formation of organic molecules (HOMs and ACC) or to the choices made by the modelers on which species participate in NPF.  In particular, it is speculative that all HOMs (as defined by the authors) participate in NPF with H2SO4 at the rate specified in the Riccobono et al (2014) CLOUD study, as in Eq 6 of this paper. The species defined by Riccobono et al was called ‘BioOxOrg’, not HOM. The large uncertainties associated with these parameterizations should be acknowledged in this paper, and ideally quantified with sensitivity studies.

I would also speculate that the results are likely sensitive to uncertainties in concentrations of H2SO4 and NH3. The sources of these uncertainties are not really discussed in the main text yet (despite documentation that a bias in H2SO4 exists and a helpful table of budgets in the supplement).

In the light of these remarks, while I do agree that the uncertainty due to the omission of anthropogenic organics is potentially important, it is far from the only important uncertainty, and the last paragraph of the paper could be more balanced.

Minor comments

The citation for CMIP6 emissions is a bit vague, or insufficient to determine which version of the emissions were used. Feng et al, GMD 2020 is relevant. There are also significant differences between the most recent (post-CMIP6) releases of the “CMIP6 emissions” and the versions actually used. See https://github.com/JGCRI/CEDS/. If the most recent releases were not used, some comments may be needed about the deficiencies of the CMIP6 emissions, for example SO2 emissions were overestimated in the western USA.

It would be interesting to know how many chemical species are advected and how many are otherwise represented.

For the ion concentrations, could radon be important? Would ion-induced nucleation be important close to the land surface if ion production rates were a factor 5 or so higher there?

L167 I believe the correct citation for this organic parameterization is Kirkby et al, Nature 2016, not Dunne et al, 2016.

L174 What is the reduced condensation sink?

L320 misspelling of species