Dear Dr. Velders,
This is a very nice paper. You’ve done a good job of responding to the reviewers’ comments, and in particular the description of how the inventories are constructed is much improved. I think it’s nearly ready for publication. I have just a few minor points that I would like you to address, and then I will re-examine this myself (no further external review needed). Thank you.

Regards,
Drew Shindell

1) On page 3, lines 91-94 it states that previous studies included banks but ‘the focus in discussing climate change effects was only on emissions and radiative forcing ‘. This is a useful addition you made in response to reviewer 2, but I think could be even clearer, as your study also presents emissions and radiative forcing. You might add at the end of that line something like ‘only on emissions and radiative forcing and did not highlight (or point out) the lag between production and emissions’.

2) That same sentence states that previous studies such as Velders et al 2012 included banks but did not focus on them. However, on page 4, lines 102-104 it says: the benefits of earlier HFC phaseouts will be greater than previous estimates, where only emissions, concentrations, and radiative forcing were considered but not banks (UNEP, 2011a;Velders et al., 2012). This seems to contradict the prior statement about banks being included in the earlier work. The same sentence on p4 still seems not perfectly clear to me either. Did the previous estimates include emissions based on production with no banking? Wouldn’t that lead to about the same benefits in the long-term, and they would simply have been too early in the prior studies? The results of the current study seem clear and are well presented, but it seems very important to be as clear as possible about how they relate to the prior studies and that’s still not quite clear to me.

3) At the start of section 4 you state that you will use 100-yr GWP to create a CO2e scale. Here I think it is important to note that 100-yr GWP provides only a rough approximation to the climate impact at that point in time, and does not imply equivalent impact at either shorter or longer timescales. This is quite an important issue when you’re comparing gases with very different lifetimes, as you can get the same GWP100 from a short-lived HFC as from a long-lived gas like CO2, but the short-lived HFC will have all of its impact on RF (and hence much of its impact on climate) early on in the 100 yr integration period while the long-lived will be much more evenly distributed, and hence even the 100 yr temperature change is unlikely to be equal (let alone other times). I know this is not your primary point and I think it would be distracting to use multiple timescales or metrics, but it’s nonetheless important to point out that CO2e doesn’t really mean the emissions are equivalent to CO2 emissions or to each other for the various gases studied here (see, e.g., the discussion of CO2e in chapter 8 of IPCC AR5), especially for something like this paper that will hopefully be read by policy-makers.

Dear Dr. Velders,
Thank you for the last set of revisions. I am pleased to accept your paper.
Best,
Drew