The manuscript proposes a methodology for correcting SF6-based age-of-air (AOA) to remove biases associated with chemical sinks.  I commend the authors for tackling this difficult problem, which I feel they address rigorously within the limitations of their approach.  However, I have several concerns about certain fundamental assumptions that are made and these concerns subsequently lead me to encourage publication only after major revisions have been performed. In particular, I am concerned about the approximation of the AOA spectrum as a delta function (major comment below), which is clearly violated throughout the entire stratosphere.  I ask that the authors provide more solid physical evidence for the assumptions underlying their methodological approach.

Major Comment:

I have major concerns regarding the assumption introduced in developing equation (2) that the AOA spectrum (i.e., G) be approximated as a delta-function.  Physically, this assumption only approximates the true physical medium when the flow is purely advective (i.e., mixing = 0).  This is simply *not* the case in the stratosphere and, indeed, is the whole reason why the age spectrum was developed thoroughly in Hall and Plumb (1994).  While the authors seem to appreciate that this is a “crude assumption” I find it too inconsistent with the physical nature of stratospheric transport to be used.  I ask that the authors more rigorously justify their assumption and/or modify their derivation to account for the effects of mixing, as this assumption renders the whole methodology seriously limiting in its applicability.

Hall, Timothy M., and R. Alan Plumb. "Age as a diagnostic of stratospheric transport." Journal of Geophysical Research: Atmospheres 99, no. D1 (1994): 1059-1070.

Minor Comment 1:

How does the “path-averaged lifetime” introduced in equation (2) relate to the “path-dependent lifetime” introduced in Holzer and Waugh (2015)?  It appears to me that the authors may be acting unconstructively (though, perhaps, unwittingly) here by adding unnecessarily new terminology to the field. Please provide a rigorous definition of this lifetime and relate it to that from this 2015 study. Note that the path-dependent lifetime presents the advantage that it does not assume zero mixing.

Holzer, Mark, and Darryn W. Waugh. "Interhemispheric transit time distributions and path‐dependent lifetimes constrained by measurements of SF6, CFCs, and CFC replacements." Geophysical Research Letters 42, no. 11 (2015): 4581-4589.

Minor Comment 2:

Equation (1): If you are defining t’ as transit time (somewhat unconventional, as transit time is often denoted using the notation tau=t-t’, see Holzer and Hall (2000) for more), then your use of G(t’) suggests that the transport operator is stationary, i.e., not dependent on either the initial pulse time or receptor time.  This assumption needs to be stated clearly as several studies have shown that this assumption is actually *not* valid, even in the stratosphere. In particular, see the study by Li et al. (2012), which directly quantifies the large non-stationarity of G (and associated mean age):

Holzer, Mark, and Timothy M. Hall. "Transit-time and tracer-age distributions in geophysical flows." Journal of the atmospheric sciences 57, no. 21 (2000): 3539-3558.

Li, Feng, Darryn W. Waugh, Anne R. Douglass, Paul A. Newman, Steven Pawson, Richard S. Stolarski, Susan E. Strahan, and J. Eric Nielsen. "Seasonal variations of stratospheric age spectra in the Goddard Earth Observing System Chemistry Climate Model (GEOSCCM)." Journal of Geophysical Research: Atmospheres 117, no. D5 (2012).

Minor Comment 3:

Is it really prudent to use the balloon data from 2000 (grey stars on Figure 2)?  These large values seem unrealistically high.

Typos:

Line 36: remove “and” after “surface”

Line 47: typo “particualr”

Line 56: Do you mean “ideal age”, not simply “ideal”? If so, please diligently acknowledge the following references:

Thiele, G., and J. L. Sarmiento. "Tracer dating and ocean ventilation." Journal of Geophysical Research: Oceans 95, no. C6 (1990): 9377-9391.

England, Matthew H. "The age of water and ventilation timescales in a global ocean model." Journal of Physical Oceanography 25, no. 11 (1995): 2756-2777.

Line 61: Again, please clarify the use of “ideal” here.

This is a well-written and thorough paper that is well suited for publication in ACP. The derivation of AoA from observed SF6 is an important topic in stratospheric research and this paper is a novel very useful contribution to the challenge of how to best quantify AoA from a tracer with chemical sinks. Overall I think that this paper represents a significant amount of intellectual and modelling work and I only have a few minor comments before I think it is publishable (see below).

There are likely further details and approximations which could be explored by different model setups but I think the current paper is largely sufficient to present the proposed methodology

Specific comments.

The model uses a timeslice setup which will give a steady circulation (with variability). I understand the reasons for that. However, it would be good to mention the possible implications (if any) for the work if a transient run had been used to simulate up to the present day, assuming the model would capture any realistic trend in AoA.

The work is based on the chemical scheme of EMAC. Other models will likely give different SF6 loss rates and the range of lifetimes is given in the final section. What would be the implications is the shortest estimates of the SF6 lifetime were correct?

Line 46. ‘tracer’ (singular)

Line 47. ‘particular’

Line 48. ‘often-measured’, ‘has become increasingly clear’.

Line 70. ‘next section. Here it would be better to give the explicit section number.

Line 147. ‘constrained’.

Line 173. The caption of Figure 1 could do with some references for the data and model values.

Line 181. 10^6.

Line 231. MLS N2O is known to have a drift, especially in the older versions. Please clarify what version is use here (in GOZCARDS), whether it is subject to a drift and if so what is the impact on the AoA derived using it.

Line 258. Figure 3 left panel. You might as well have a legend which labels the 5 lines explicitly.

Line 285. Space before ‘+’.

Line 286. I think the partial lifetimes of the EMAC SF6 tracer should be stated more clearly and comprehensively (in Section 3.1). Here we are told 1900 years for stratosphere and mesospheric loss. (Assuming no loss in the troposphere that would be the total lifetime?). The conclusions (line 533) say 2100 for just the stratosphere (why just the stratosphere for this part?). These values may be consistent but it would be better to summarise up front what partial and overall lifetimes are produced by this version of EMAC.

Line 294 (and many other places) ‘3^rd-order’.

Line 300. ‘evaluated as well’.

Line 303. You are mixing US and UK spellings. (Also for color / colour). Also ‘non-physical’.

Line 323. Need to rephrase, e.g. ‘this spin up effect …. is not only an internal property of the model..’?

Line 340. ‘lifetime’

Line 401. ‘sections’.

Line 416. Degrees symbols in latitude range.

Line 437 (and line 464). Better to say ‘previous section’ or give the number.

Line 457. ’21-year’.

Line 543. ‘Plumb’ is misplaced.