The manuscript by Kolonjari et al. is a thorough elaborated validation paper for the

recent retrieval (version 5.2) for the HCFC-22 observations in the UTLS by the ACE-FTS.

I recommend the paper for publication, however, some clarifications are needed.

General issues

• Mixing ratio and concentration are different quantities. All of the
  
    paper seems to focus on HCFC-22 mixing ratios.  Please omit
  
    "concentrations" in the title as the validation is on mixing ratios.
  
    And use mixing ratio instead of concentration throughout the paper.
•  If one has 2 quantities A and B, the "percent difference relative to A"
  
    would be (B - A) / A x 100% in my understanding.  This seems not,
  
    what is plotted in most figures.  The sign of the difference is
  
    extremely important for the understanding the validation paper.
  
    There should not be possible mis-understandings. Please correct
  
    this (either in the plots or in the description).

Specific issues

page 2, line 5: mention the year of the Montreal Protocol, 1988.

p. 6, l14.: from the JPL-2006 to JPL-2011 there is a change in

recommendation for the HCFC-22 photolysis. They recommend using the

extrapolation of absorption cross sections up to 220 nm. This has a

significant impact on the calculated photolysis rate. Not so much on

the HCFC-22 depletion as photolysis is not the dominant loss

process. However, as the model also shows a high bias at 25 km, which

may point to a missing sink, please confirm that you really used this

recommendation.

p. 8, sampling. few negative observation values are omitted from the

comparison as they could cause strange numbers in the relative

difference. What about the near-zero positive values?

p. 13, l12 "It is unlikely that a singular event occurred in 2012"

I think what is seen here is the reduction of emissions forced by the

Montreal Protocol and its amendments. This is also visible in

fig. 1.3. of the 2022 WMO ozone assessment.

p. 38 fig 12 caption: To what does "NOAA (1996" with missing ")" refer to?

Kolonjari et al. report assessment and validation of ACE-FTS v5.2 HCFC-22 limb observations in the UTLS by using a model simulation with specified dynamics (CMAM39) and correlative satellite limb observations (MIPAS), in situ and limb observations from balloons (BONBON and MkIV) and airborne in situ observations (CARIBIC). Biases between the datasets in different altitude regimes are quantified and discussed. Tropospheric trends are analysed for the time frames 2004-2012 and 2012-2018 by using ACE-FTS data, CMAM39 data, and surface data from NOAA, and are found to be consistent. Furthermore, consistent interhemispheric differences are derived from these datasets and their decadal trends are analysed.

The study is interesting, well-organized, and the data set is very valuable, since ACE-FTS is currently the only instrument providing global vertical profiles of HCFC-22 from space. The manuscript has a clear structure and the tables and figures are clear and appropriate. I have only a few remarks regarding the comparison with the model data and correlative observations and some suggestions for improvement of details.

Specific comments

P7L11 Here or elsewhere, the vertical resolution of the MIPAS data product should be provided.

P7L26 Also here it would be interesting to provide the vertical resolution. Furthermore, it would be helpful to provide information on the uncertainties of the MkIV data.

P8L4 Also here it would be helpful to provide information of the uncertainties of the CARIBIC data.

P5L25 It would be helpful if the authors could provide some more information on the comparison between ACE-FTS and the model instead of just referring to the Kolonjari et al. (2018). The simulation uses a T47 resolution, re-gridded to 3,75°x3,75°, with 71 levels, while the ACE-FTS data have a vertical resolution of 2-6 km and a certain resolution along the viewing direction. Furthermore, the satellite passes a certain horizontal range while recording a set of observations resulting a vertical profile. In how far need these different characteristics of the datasets need to be taken into account in the comparisons? I guess they have only a small effect, since the vertical gradient of HCFC-22 is weak and effects in the horizontal domain probably largely cancel out due to the large amount of coincidences. However, e.g. at the edge of the polar vortex or in the vicinity of monsoon systems, this might be relevant. I would appreciate if the authors could comment on this.

P8L22 The MIPAS, Bonbon, MkIV and CARIBIC data have different vertical and horizontal resolutions, and also the temporal and geographical match plays a role. I guess that in the case of the MIPAS data and possibly also for CARIBIC, these effects largely cancel out due to the large number of coincidences. However, for Bonbon and MkIV, only a few coincidences are available, which could be affected by these factors if vertical structures in the HCFC-22 distribution would be present in the region around the tropopause (e.g. tropopause variations, tropopause folds, intrusions). Should this be reminded when interpreting the biases between ACE-FTS and these observations?

P9L19 What do you mean by “the presence of polar vortexes”. Are you referring to strong gradients in trace gas distributions at the vortex edge due to downwelling within the vortex?

P10L23 See above, it would be interesting if you could provide additional information on the errors that are taken into account in the MkIV error bars. Furthermore, it is discussed that the one sigma “error” bars of the ACE-FTS data show the zonal standard deviation. Therefore, I am wondering whether the word “error” is suitable in this context.

P11/L6 To my understanding, only CMAM39 and MkIV data are shown in Fig. 7. In the text, ACE-FTS zonal mean profiles are mentioned. Please clarify.

Technical points

P3L12 specttra

P30F4 Does the “-5.2°” etc. information in the top of the panels refer to coordinates of BONBON or ACE data? See also P31F5.

P36F10: Is the thermally-defined tropopause derived from the model data or the observations?

P38F12: Part of the last sentence seems to be missing: (1996-2021…?