Comment on acp-2021-787

L11 and throughout: "amount fraction" is not a term I'm familiar with, I suggest "abundances" for referring to both O2/N2 and CO2, and "mole fraction" for referring to CO2. L11: "Observations were corrected for significant..." L18 and throughout: usually "northern hemisphere" and "southern hemisphere" are capitalized, but I would refer to the specific style guidelines of the journal. L22: "indicated a clear evidence of influence" -suggest change to "indicated a clear influence" L24: What is a "C equivalent"? Do you mean simply petagrams of carbon? If so, Pg C a-1 is a widely used unit.


INTRODUCTION
L27: "ratio" should be "ratios" L27: "marine biospheric activities" is a little unclear to me. I might suggest simply "...early 1990s, for the primary application of constraining the marine and terrestrial exchange of CO2." L31: suggest "terrestrial biosphere exchanges" L36: suggest "carbon dissociation effect" read something like "the carbonate buffer system" L39: I think there is a missing sentence here explaining that airborne observations are useful because they quickly map a large spatial area. Suggest cutting "from this point of view" and moving the sentence to the first sentence of the paragraph beginning with "Aircraft observations" L40-45: Steinbach (2010) might be worth citing here, since it pre-dates the Ishidoya and van der Laan references. Bent (2014) might be worth citing here as well, since he also reported Ar/N2.

METHOD
This section could use subsections for easier reading.
L76: Suggest changing "Method" to "Methods" in the section heading.
L78: "The cruising altitude is about 6km" L79: "titanium" (not capitalized) L101: Could the authors specify the scales these species are measured on, at least for O2/N2 and CO2? L102-103: The authors cite Ishidoya et al 2014, which cites Niwa et al 2014, which cites Tsuboi et al 2013. I will admit I only scanned the papers but it seems Tsuboi is the only refernce that describes the intake and flask sampling apparatus. So I would point directly to this paper on L81 to save the reader time. This paper does not seem to have a diagram of the flask sampler, so unless I missed it in one of these papers I think it would be nice to include either in the text or in a supplement. I think this is important because the fractionation of the samples is quite considerable. I am still not clear on what kind of inlet is actually outside the airfract. From Stephens et al 2021 it was clear that the design and orientation of the inlet is critical for avoiding fractionation. Where is the air conditioning inlet located on the plane, and what does it look like? The thermal fractionation the authors identify is so massive it is probably obscuring other sources of artifacts, like at the inlet or somewhere in the air conditioning system. Since sample air passes through this, could the authors include it in the plumbing diagram? I am also surprised the authors don't have serious problems with surface effects, given that the teflon tubing is used, flasks are only partially dried, they are pressurized fairly high to 0.4 MPa, and then analyzed (I think) without a push gas. I am sure the authors have worked all this out, and probably have already published details on it, but without details here or specific citations it's hard to understand the sampling conditions. L113: Perhaps cite here that the ratio of the scaling factors 4.57/16.2 is close to the Keeling et al 2004 diffusion factor for (Ar/N2)/(O2/N2) and results in the same tracer d(O2/N2)* L115: Since all of the samples are from the same region, why not use the monthly mean at Tsukuba? I expect there will be some lag to consider, but otherwise it seems like this is introducing an unnecessary approximation.
L116: How was the uncertainty evaluated, and what terms are contributing to the total uncertainty? Are you accounting for natural variations in Ar/N2? Is this what is meant in L117? How much does the annual mean of Ar/N2 vary? Have the authors considered forcing to a constant value of Ar/N2? This might be preferable since the paper deals with interannual trends.
L119: This seems like the beginning of the "Results" section to me, since data is presented. L156: I don't understand this sentence: "driven by an annual mean air-sea O2 and N2 fluxes...that was considered by Tohjima...was ignored". Also, in the Tohjima et al 2012 reference there is an unnecessary hyphen in "annual". In the Tohjima paper this seems refer to the Gruber et al ocean inversion O2 fluxes? Was there a separate run of the Gruber fluxes? Or just two products: 1) Garcia and Keeling + Takahashi + CDIAC and then 2) simulated -observed APO?

RESULTS
I suggest to cut some of the L161-177 text, the decrease in O2/N2, the rise in CO2, and their seasonal cycles are well known. I would start the paragraph at "The average rates of change..." with figure citation. L171/Fig4: there are multiple fit lines to the data but this is not explained in the caption. Maybe a legend? I think it L186/ Fig 6: I think it would be better to plot each latitude bin as a separate plot with observations and model together, it is a little difficult to compare as is. It would also be nice to see the seasonal cycle in the detrended observations along with the fits. L194: Suggest changing "anti-phase nature in the seasonal APO cycles" to the "opposing phase of the seasonal APO cycles" L197: I don't understand "superimposing the anti-phase seasonal cycles through the interhemispheric mixing of air". From this I would think you are running only the SH flux, but the "w/o SH flux run" would imply it was northern hemisphere fluxes only. Earlier in L192 it says "northern hemisphere flux only". L198/199: should read "the seasonal cycle in CO2 mole fraction" or just "seasonal cycle in CO2". L205: From figure 6 it looks to me that most of the seasonal cycle is due to NH fluxes, as one would expect...it does not look that much smaller to me. Can you give the amplitudes of the two runs? L213-215: I don't fully follow this--it's the gradients in the fluxes, with contributions from atmospheric transport, that causes a gradient in the amplitude in the atmosphere. I think it would be clearer to say simply that the SH makes a significant contribution to the amplitude and phase of the lower latitude observations. I would also caution against over interpreting the model results--the transport model could be over or underestimating the interhemispheric transport.
L224: Suggest changing "highly" to "more" Fig8/9: Typically one plots altitude on the y axis, but I leave it up to the authors.
Fig10: This looks like a fit to the data, can you include the points as well? The data look a little odd, I would expect observations of the APO growth rate to look noisier.
L265: This is a rough approximation of the thermal component of APO, which is a combination of air-sea fluxes of O2, N2, and CO2 caused by solubility changes. The "netbio" APO will also have a contribution from fossil fuel burning/CO2. L273-287: I am not fully convinced that this exercise accomplishes more than very roughly constraining the global APO and O2 flux. The number of simplifying assumptions is extensive, and using aircraft observations from a comparatively small region, sparse data coverage, and enormous sampling artifacts is not an ideal approach. To me what this shows is that the corrected data have a coherent signal the authors can explain, and helps to prove that the data are of good quality. But I would caution against over interpreting and overselling the data. L301: I thought 1.35 was used? L312: Where does the 4-5 year period come from? Just visual inspection? Missing here is a citation for the Nevison et al 2008 study, which (also) showed that the errors from assuming a constant zeff over short time scales (e.g. 5 years) are quite large. CONCLUSIONS I would change this section heading to "Summary" L327: Better something like "Regular air samples were taken on cargo aircraft flights from..." L341 and throughout: "Superposition" is a slightly strange choice of words, I might suggest something simple like "combination".