Dear Bill and co-authors,

First of all, I would like to thank the two anonymous reviewers and other commenters for their constructive feedback on the AerChemMIP manuscript. I also appreciate that you've taken their comments into account and provided a revised manuscript.

On the basis of the revised manuscript and on your responses to the reviewer comments, I have two remaining concerns:

1. A number of comments related to one of the stated aims of AerChemMIP, namely, how do uncertainties in historical NTCF emissions affect radiative forcing estimates. In your response, you clearly state that the proposed simulations rely on central estimates for emissions and that exploring the full spectrum of uncertainties would be unrealistic due to uncertainties being region-, species-, and sector-specific. You state "For that purpose, we will make use of perturbation (pre-industrial to present-day) simulations. This is likely to provide an upper bound on the impact of emission uncertainties". However, no PI-PD perturbation experiments other than those based on the central emissions estimates are being proposed here. And although I accept that you cannot increase the computational burden of AerChemMIP further, I would question whether the proposed simulations alone can really address the stated aim. Can you be more explicit about what you mean by "we will make use of perturbation (pre-industrial to present-day) simulations"? Who are "we" in this context? And what perturbation simulations are you referring to?

2. The second concern I have is in relation to the AerChemMIP diagnostic request. The excel spreadsheets have been updated and have clearly improved since the manuscript was first submitted. Thank you. However, in your response to Martine Michou's comments, you indicate that "We will keep the excel tables updated on the wiki page and they shall be consistent with the data request on November 1st". Despite the targetted date, this suggests to me that the AerChemMIP data request is still subject to change but with modelling groups due to start CMIP6 simulations imminently, it's important to finalise the data request. The second thing is that at the Met Office Hadley Centre, we ourselves have recently found inconsistencies between the AerChemMIP data request (dated 26 Oct 2016) and one of the recent iterations of the CMIP6 data request (v42: dated 15-11-2016). Therefore, I would ask that you upload the excel spreadsheets as Supplementary Material to accompany the AerChemMIP paper as a means of finalising the AerChemMIP data request which would then serve as the definitive request while CMIP6/CEDA irons out any remaining discrepancies.

I also have a number of very minor comments:

1. Page 2, line 10: Change CO2 to carbon dioxide (CO2)
2. Page 2, line 26: Give full names for IPCC and AR5
3. Page 2, line 31: Give full names for NOx, CO and VOC
4. Page 2, line 34: Add in "(CFCs)" and "(HCFCs)" here so that you can refer to CFCs/HCFCs elsewhere (e.g. in Table 1)
5. Page 3, line 8: Give full name for ACCMIP
6. Page 4, line 7: Can you expand further on the differences between ACCMIP and AR5 models?
7. Page 5, line 34: Add closing bracket
8. Page 9, line 22: Replace "ODS halocarbons" with "ozone-depleting halocarbons"
9. Page 9, line 27: Replace "describe" with "described"
10. Page 10, line 15: Put "histSST" in italics and make sure that all experiment names are italicized throughout the manuscript (e.g. Section 3.2, Section 3.2.1), including Table captions (e.g Table 1) for consistency.
11. Page 10, lines 19/20: Please correct "These experiments cover on the time frame from 2015 to 2055"
12. Page 10, line 32: Give full name for AIM
13. Page 11, line 18: Please give full name for TOA here or else in previous section, where referring to taking the net difference in radiative fluxes (e.g. Opening paragraph in Section 3)
14. Page 12, line 2: Change "present" to "present-day"
15. Page 13, line 4: Remove CMIP5 from list, given that you're discussing model versions relative to those used in CMIP
16. Page 13, lines 17/18: Can you add a brief explanation for aerfixed, aermonthly-ed, aerdaily etc..?
17. Page 13, line 25: Give full name for COSP - same is true for CALIPSO and CALIOP
18. Page 14, line 19: Replace "semi-direct" with "semi-direct effect"
19. Page 15, line 3: Can you add explanations for variables rsutca, rsutcsca, rlutca, and rlutcsca?
20. Page 15, line 23: Change "The stratosphere has been identified as potentially important contributor to model-differences" to "The stratosphere has been identified as a potentially important contributor to model differences" i.e. add "a" and remove hyphen between model and differences
21. Page 15, line 29: Correct "nitrous oxides"
22. Page 16, line 33: Given full name for PCMDI
23. Page 42: In response to the comment that the more complex configuration is always encouraged for AerChemMIP, although you respond that you have added the minimum model configuration in each table, I would suggest strengthening the point that the most complex configuration be used in the relevant Table captions.
24. Page 44: No diagnostics are being requested for anthropogenic emissions as these are assumed to be CMIP6. However, is this a valid assumption for VOCs, where modelling groups have to make decisions about how to map CMIP6 emissions onto those VOCs represented in their
chemical mechanisms?
25. Page 45: Suggestion of having a composition DECK - Whilst I agree that composition lacks equivalents of the DECK experiments, I do feel that this suggestion is outside the scope of the AerChemMIP paper. There is a need for community input on this matter and I support
your proposal that Aerocom and CCMI are the projects by which this initiative should be taken forward.

Once these comments are addressed, I will recommend that your manuscript be accepted for publication in GMD.

Many thanks,
Fiona O'Connor

Dear Bill and co-authors,

Many thanks for addressing the minor comments that I raised following your response to the reviews submitted during the open discussion of the AerChemMIP manuscript. I am pleased to inform you that the paper has now been accepted for publication in Geoscientific Model Development.

Regards,
Fiona O'Connor