This paper evaluates how the newly updated CRACMMv1.0 chemical mechanism does at predicting northeastern US O3 concentrations when compared to older CMAQ mechanisms and surface observations of O3 across the northeast. They find the new mechianism predicts lower concentrations of O3 than the RACM2_ae6 mechanism did, which was closer to surface network observations. Evaluating this mechanism in a box model, they explain these improvements arise (1) largely because of rate constant updates more in line with the state of the science and (2) better representations of aeromatic and monoterpene chemistry. The analysis shown was convicing, well explained, and insightful. Although it is primarily a mechanism evaluation and comparison, the paper also has important policy implications- that controlling aeromatics may be a particularly useful way to reduce urban O3 pollution without rural disadvantages.  

Overall, I found the paper to be exceptionally well written and of high scientific quality and reccommend accepting it with a minor addition to the code and data availability section.

In this section, the authors highlight that the mechanism is available in CMAQ and on github and have pointed users to the F0AM and AMET github sites where users can access the surface O3 data and the box model used in their analysis and state that specific model inputs and code will be archived on data.gov following its final publication. The only minor addition I'm requesting is to make sure the authors intend to include a functioning F0AM mechanism file with the CRACMMv1.0 mechanism they developed in this final data so other F0AM users can utilize this (now field constrained), mechanism to expand the scientific utility of this work. Furthermore, I would encourage the authors to include the SMILES strings for all compounds (and how those relate to their "short names" in the mechanism) as part of this archived data, in order to allow the broader community to use computational tools to compare such a large mechanism to other mechanisms that might lump organic species differently. SMILES codes are provided for the compounds in mechanisms like the MCM and Bates et al., 2022 to enable this exact kind of comparison between different chemical mechanisms. For an example of what I'd like to see to continue this trend of being able to easily compare new chemical mechanisms with many different organic species and lumping schemes, I'd point the authors to the supplement of Bates et al., 2022:  https://acp.copernicus.org/articles/22/1467/2022/acp-22-1467-2022-supplement.pdf . For lumped compounds, a few examples of SMARTS strings that would match a lumped compound or a SMARTS string that would return compounds matching that lumping scheme would be sufficient. If these details are added the final data archive, the paper would enable more researchers to easily build off this work and use this new mechanism for other scientific insights. 

This paper presents a new, expanded chemical mechanism for use in the EPA regional model CMAQ. The mechanism is evaluated in CMAQ simulations of the northeast U.S. for summer 2018 through comparison to surface ozone observations and 2 other established chemical mechanisms.  The mechanisms are also compared using the F0AM box model to quantify the importance of several groups of VOCs at high and low NOx.

The paper is very clearly written and the figures and tables clearly present the results and support the conclusions that the CRACMM mechanism shows improvements in regions of importance for exposure. 

I recommend publication after consideration of a few minor points.

Section 3.2 states some statistics for differeces between urban/suburban and rural sites, but these are not easily distinguished in maps or given in a table.  It would be helpful to show separate maps of just the urban and rural sites, or distinguish urban vs rural in the current maps with different symbols (circle and star, for example).  

Section 4.1 describes sensitivity simulations where the emissions from different categories of VOCs are set to zero.  These cases provide some interesting results and give an indication of the importance of the different sources to ozone production. However, the section should start with a discussion of the challenges of this sort of sensitivity simulation due to the non-linearity of ozone production.  It might have been more realistic to perform sensitivity tests with a partial reduction (20%) of emissions.

'zero out' -> 'zero-out' or maybe some other wording in some cases.  Such as l.353, which could be written '... a series of zeroed emissions cases.' 

l.480: CRAMM -> CRACMM

I am sure the Editors will point out that they do not accept Github for the archiving of code and require a copy to be put on Zenodo.