Please revised your manuscript according to the thorough recommendations by the reviewer given below. A final review before publishing will be done by me as editor.

Review comments:

I find the revised version more digestible but I'm still unclear on a few aspects. Please see below:

I don't think the words 'corrected' and 'corrections' are the best choices for describing gas transfer suppression and associated adjustments. maybe 'adjusted' or 'unsuppressed' instead.

Please be precise in the language. For example, captions for Figs 2 and 3 should read "adjustment to the k v U relationship", rather than "correction of fluxes"

Eq 1 and 2 still from the perspective of air concentrations. i.e. k shown = Ka, not Kw
should be flux = k(Cw - Ca*H) when k = Kw

Eq 3. earlier k is used, and now ktotal is used. please be consistent. I suggest Kw or Kt for Eq. 1-3

top of p. 3. the dual tracer (3He/SF6) method likely doesn't capture the full extent of bubble-mediated gas transfer. see: Asher and Wanninkhof 1998, vol103, 10555-10560, JGR. accounting for bubbles will likely increase the N00 curve

p. 4, line 12. 17% reduction due to inclusion of DMS air concentration seems large. 100 ppt of DMS in air converts to an equilibrium concentration 0.06 nM at 20 deg C, while a typical seawater DMS concentration is about 1.5-2 nM

p. 5, line 29. so this assumes that N00, W14, and TAK09 aren't suppressed already

p. 6, line 24. the two approaches are equivalent in the case that they result in the same k vs. u slope. but of course they will result in two different mean K values, for example

p. 7, line 26. "we interpolated the wind and wave data to 1 h time steps..." so the authors used the 2014 wave information, and not modelled wave data from the actual year of the dual tracer measurements, correct? are the modelled wave data taken from the same locations as the actual dual tracer measurements? are the "1 h time steps" around the actual times of the dual tracer measurements? or they're 1 h time steps of the entire year?

p. 10. line 18. however a gas transfer 'suppression' was observed for the Southern Ocean GasEx cruise

p. 10, line 21. Most critically, this is where I'm lost. the authors already show that the original N00 is heavily suppressed (by "22%"). So why is the subtraction of deltaK in Eq. 14 necessary in the global estimates?

If anything, it seems that if the global average k suppression is 10%, and N00 is suppressed by 22%, one should get a HIGHER, not LOWER global flux when using the adjusted N00 relationship (compared to using the original N00 relationship)

The Tak09 relationship, similar to W14, is based on multiple years of observations and presumably includes the effect of any gas transfer suppression already.
The authors state the following and I think the same applies to TAK09.

W14 parameterization is just calculated for comparative and example reasons to show the effect on this parameterization type. The W14 parameterization already includes an average estimate for gas transfer suppression and no suppression needs to be added on top.
I don't think it makes sense to apply the gas transfer suppression model to W14 and TAK09 and get revised global CO2 fluxes. I think those values should be removed from Table 3.

The authors have satisfactorily addressed the questions raised by the reviewer in the final round; thus this manuscript can now be published. While the reviewers are now satisfied, I think the paper would profit very much if it were read and language corrected by somebody with a scientific background and proficient in the English language. This would be prior to the copy-editing in the Copernicus office. Sentences like "For CO2 we estimated the magnitude of suppression for the DMS the magnitude and the global reduction of DMS efflux." (in the Abstract) make no sense. Also, the sentence in the abstract "An unsuppressed Nightingale 2000 parameterization leads to an average increase of 22 % of its predicted gas transfer velocity. The increase for Wanninkhof 2014 is 9.85 %.", while technically correct, is awkward: Can this not be turned around: e.g. "our proposed correction based on the suppression effect reduces the predicted gas transfer velocity by 22% if using the Nightingale 2000 parametrisation, and 9.83% if using the Wanninkhof 2014 parametrisation." or similar. It is also unfortunate, that the proposed mechanism behind the suppression effect is nowhere described in the text, but one has to go to the original Zavarsky et al., 2018 paper to understand it. Adding perhaps a small paragraph in the introduction where the basics of the flow separation in the wave field are explained would greatly help readers who are not from the field. Without going to the Zavarsky et al., 2018 paper it is incomprehensible why the suppression effect occurs at low values of the transformed Reynolds number, e.g. at low significant wave height. This is counterintuitive since the effect is happening at high wind speeds, which imply also high waves... Explaining this in clear language would greatly enhance the impact of this quite important paper.