This manuscript presents a novel data set of stratospheric δ13CO2 dating back to 1985 and examines the processes controlling its distribution and its potential use as an age tracer. I think this is a very useful dataset. The results presented will be of interest to ACP readers, and I think they will be acceptable for publication after relatively minor changes. 

My main concern with the manuscript is the discussion of the δ13CP age results. I think the authors are overstating the agreement with the CO2 age estimate and the potential use of the δ13CP age estimate. Specifically, in the abstract it is stated that "the mean age derived from δ13CP was consistent with that derived from the CO2 mole fraction, suggesting its usefulness for further investigation of stratospheric transport processes.", while in the conclusions it is started that "it [δ13CP] should prove a useful tool for investigating stratospheric transport processes" and "d13CP age is slightly larger than CO2 age ... by about 1.1 years on average".

The stated consistency occurs only because there is a very large (30%) uncertainty in δ13CP, and consistency is in terms of the mean over all data. Figure 10 shows that there is no statistical agreement for many individual measurements.  Also, I don't think an average bias of 20% (which is what 1.1 yrs is) can be considered small.   Given this large uncertainty and bias, I am doubtful that δ13CP age is a useful estimate/tool, as stated in the abstract and conclusions. It has the potential to do this only if the uncertainty can be greatly reduced.  

I think the text needs to be modified to better indicate that there is large uncertainty and bias, and this needs to be reduced if this is to be a useful age tracer.  

Following on from this, I think there could be more discussion of potential errors. On line 549 it is stated that "it is likely that d13CP age is overestimated because the effect of methane oxidation was underestimated in the calculation of d13CT." Why was methane oxidation underestimated, and can an estimate of the impact of this be made?

Finally, the age calculation from δ13CP is buried at the end of Appendix C. Rather than having this appendix focused on CO2 and then meaning the new aspect (δ13CP) at the end, both tracers should be discussed at the start. Highlighting when differences in approach, or steps were could be large uncertainty. 

This paper uses flask sample measurements from balloon flights in the stratosphere over Japan to investigate the stable carbon isotopic ratio of CO2 and other molecules involved in CO2 photochemistry in the stratosphere.  The measurements and analytical methods are described in some detail and as far as I could follow it the techniques seemed reasonable, but I am not an expert in this topic as the authors clearly are.  The discussions of the mechanisms responsible for the delta13CO2 and CO2 stratospheric profiles are well done and the schematic shown in Figure 6 provides a nice summary of the competing processes.  The 2D modeling is interesting in general and provides some support for the mechanisms described here as the primary drivers of the observed distributions.  The model transport is inaccurate, as the authors attest, but that is a common limitation of most stratospheric models.

The use of delta13Cp as an age of air tracer is somewhat dubious.  The large uncertainties on the ages with this quantity make it essentially unusable on its own.  The authors do suggest that it could be used in combination with other trace gas measurements for multi-component age estimates but it seems unlikely that this new quantity will help constrain any of the current age of air trace gas estimates.  My main suggestion would be to considerably shorten Section 3.5 by removing much of the detail of the age of air calculation with delta13Cp.

Overall, this paper presents novel measurements and analysis suitable for publication in ACP.  I recommend publication with consideration of the main comment above and the specific comments below.

Specific comments:

Line 18:  add ‘the’ after ‘investigate’

Line 45:  For those of us not as familiar with the details of isotopic studies, perhaps a brief explanation of MIE would be helpful here.

Figure 2c:  Appear to be missing the dotted line in this figure that would show the model results without the tropospheric trend, GS or airborne sources.

Lines 380-5:  The GS correction of CO2 is interesting, although not as significant as for delta13CO2 as is mentioned.  The example of 22 August 2010 is said to have a GS correction of 0.4 ppm for CO2 at the 34 km altitude level but it doesn’t look that large in Fig. 5a.  Certainly, all of the points below the 34 km level have very small GS corrections for CO2.  But 0.4-0.6 ppm would have an effect on the age of air calculation with CO2.  Would you recommend that all age of air calculations with CO2 use a GS correction?  If so, would an average profile of deltaG as in Fig. A1 be appropriate to use in Eqn. 8?  The implication here is that age of air from CO2 without taking into account GS, which is essentially all age of air calculations done thus far, has an old age bias that increases with height.  If this bias is quite small, say less than a month, then this is not significant.  But it appears to be larger than that at high altitudes.  A brief statement here about age of air implications would be useful.

Figure 10:  I would suggest removing the trend line for the delta13Cp age since the uncertainty on the values are large and the trend is not significant.  The insignificance of the trend is mentioned in the text but the figure implies there is a discrepancy in the age trends rather than that they are in agreement within uncertainties.

Line 598:  add ‘is’ before ‘important’