This paper seems quite thorough and generally well written, except for a few patches where the English usage could be improved (discussed below).  Detailed comments are provided below, none of them particularly concerning.

> Line 60: “…we have maintained the same retrieval setup as in the O3 NOM dataset A comparison…”

                Missing period after ‘dataset’

> Line 106: “…along the LOS.”

                The acronym LOS is not defined.

> Line 150: “and the apodization of calculated spectra used a wider frequency range.”

                What apodization is being used?

> Line 201: Daytime

                No numbering on the subsection.  Same for line 213 (Nighttime)

> Line 255: “The updated figures for version V8 are shown in Sec. D.”

                Appendix D

> Line 268: “data with smaller values are considered non-trustful”

                …non-trustworthy

> Line 309: missing paragraph indent

> Line 369: “…information on the confidence limits of the error margins…”

                I cannot decipher what this means.  Are you saying some uncertainties (e.g., for intensities) are unavailable, or you don’t know if the errors provided are 1-sigma or 2-sigma, or something else?

> Line 333: “The uncertainties of the spectrally interfering molecules with ozone which are not jointly fitted (e.g. as in the case of water vapour), as well as their vertical covariances, are estimated from the error covariance matrices of previous MIPAS data version V5.”

                I do not see anywhere an indication of which interferers (other than H2O and a passing mention of CO2 laser lines) are being included in the analysis.

> Line 382: “…requires in the photochemical model described above…”

                required

> Line 382: “…was estimated in 7%.”

                in -> to be

> Line 384: “…ingoing uncertainties affecting the retrieval were reported…”

                Perhaps “…various contributions to retrieval uncertainties were reported …”?  Starting in the last few paragraphs, the level of English seems worse than the text up to that point.  The issue continues for this section (4.2).  I will provide no further specific comments on English usage, but I would suggest this section be rewritten to improve grammar and phrasing.  The level of English improves again following this section, although Section 6 could also use a mild rewrite for English usage.

> Line 395: “In the figures…”

                In Fig. 5

> Line 396: “T-LOS”

                Terminology is not used consistently.  Sometimes T-LOS, sometimes T+LOS (in the figures), sometimes TLOS.  It is not made clear that LOS refers to the line of sight and relates to the tangent altitude determination.

> Line 410: “(so-called “headache errors” (see von Clarmann et al., 2022).”

                unmatched parentheses

> Page 21, caption to Figure 7: Titles on the leftmost panels explicitly assume m = 5, MA measurements (i.e., they indicate 561-522), and the caption assumes m=5 for V8 (V8 561) but leaves m as a variable elsewhere.  This creates some confusion.  Are you comparing V8 MA results only to MA results from V5 or to results from all other V5 measurement modes?  The same question applies to Figure 8.

> Figure 7: The color scale for the rightmost panels extends from -10 to +10 %, but there are contours in the plots labeled as 20 and -20.  Is the expectation that the reader will mentally extrapolate the color scale (e.g., brighter red means even more positive)?

> Line 566: “…the differences are caused by inconsistencies between the spectroscopic data in those spectral regions.”

                It has been shown that O3 intensities in HITRAN 2016 (the basis of ACE v4 retrievals) are too weak by about 3% (doi:10.1016/J.JQSRT.2019.01.004), consistent with this assumption.

> Line 620: “…Microwindows and spectroscopic data were also update in order to be consitent…

                update -> updated, consitent -> consistent

> Line 666: data availability should probably include all the data plotted in the paper: MLS, ACE-FTS, SMILES, etc.  Some links are provided in the text but are more appropriately listed here.  ACE-FTS has no data availability information provided.  The link provided for MLS is broken.

Referee Comments, 7 August 2023

MIPAS ozone retrieval version 8: middle atmosphere measurements

By Lopez-Puertas et al.

Summary

The paper presents an update to the high-altitude ozone measurements retrieved

using the MIPAS/Envisat instrument between 2002-2012. The update makes use of

recalibrated L1B spectra as well as several modifications to the non-LTE 

retrieval scheme. Apart from describing the changes to the processing, the paper

presents a revised, TUNER-compliant, error budget. 

The paper also updates a useful intercomparison exercise involving five other

instruments and their latest data versions, which could have been a

separate paper in itself.

The authors appear to have been careful and thorough, and the paper is

essentially publishable as it stands.

I have just one significant concern which is the ad-hoc adjustment of

the k3 chemical quenching reaction rate (L187-189) which seems to have

been largely on the basis of bringing the retrieved O3 more into line

with other instruments.  While this is not unreasonable, I would like

this highlighted in the abstract so that future researchers, perhaps

also with similarly high O3 measurements, will not automatically assume that

they must be incorrect since it disagrees with all previous results.

Other than that my comments are minor and typographical, as follows.

Minor comments

1) L32: Rather than 'ample', it would be more informative to give the actual

   spectral range. 

2) L51-52: It would be helfpul to have a table giving the details of

   the scan patterns used for these three modes (which could, for

   example, incorporate additional information such as total number of

   days of each and horizontal spacing). Also instead of L81-83.

3) I found the nomenclature for the different versions of L1b and the

   ozone product inconsistent and confusing. Eg  L74 implies

   that 'V8' is being used to refer to L1B v8.03 and in the next line

   it is also referred to as 'V8R'. Similarly does V5 refer to L1B or O3?.

   L79 then refers to O3 'version 7' but it is unclear which L1B data is used.

   It is not helped by varying references to 'L1b', 'level 1b' and 'level-1b'

   and 'level 1' data. L92-93 - again unclear whether V8R, V5R refer to 

   L1b or ozone products. Some figure captions (7,8) also revert to long 

   notation.

4) Is a '2-points horizontal temperature gradient' the same as a 'linear 

   temperature gradient' ? 

5) L111-112: this is confusing. Do you mean that the offset is assumed to

   be same for all microwindows in each band, with just altitude variation?

   This seems a strange assumption given that the offset is presumably of

   unknown origin and therefore probably has some spectral dependence. 

   Anyway, please rephrase more clearly. Also, there should be some 

   definition within this paper (eg Table A1) of the spectral ranges of the A 

   and AB bands.

6) The spectral structure that the forward model is required to resolve is

   the Doppler broadened O3 line. It would be useful to have this mentioned

   at this point, along with a value for this.

   I didn't understand what you meant by the grid for absorption cross-sections

   - I assumed this would be the forward model grid, but if not that suggests

   some further interpolation is required (and the figure 0.00097656... seems

   oddly specific - is there a reason for this number?)

   By my calculation the 0.001cm-1 grid resolution just matches

   the O3 Doppler half-width, so I'm slightly surprised this is adequate unless

   your grid is adapted to the line-centres rather than a fixed grid.

7) I may be wrong, but if your Tikhonov regularisation is applied to the shape

   of the VMR profile then it seems it would change the total ozone amount

   retrieved due to the non-linearity between VMR and partial column amount,

   even if the regularisation maintains the same average local VMR value.

8) Table 1: it would be clearer to put the actual rate constants in this table, 

   along with assumed uncertainties used in the error budget. 

9) Eq 1 shows the k2 reaction being neglected but the text seems to refer to 

   this reaction simply as 'chemical production', which seems an obscure way

   of referring to it. 

10) L406. Does this mean by two 'identical' instruments? which includes 

   calibration errors, forward model assumptions etc. In that case errors 

   due to interfering species would also be the same, so not contribute to 

   the random error. 

Typography/Grammar

General comment: frequent use of hyphen (-) rather than en-dash (-- in LaTeX)

to indicate a range of numbers. 

L2: Unecessary repetition of 'altitudes' in same sentence.

L3: 'data' used as both singular and plural in same sentence.

L2: 'data' omitted after 'MIPAS'.

L27: I think you can, eg, 'provide a review' but not 'report a review'

L28: Superfluous comma after 'measurements'.

L40: Might be less confusing just to say 'latter' rather than 'latter three'.

L46: 'middle/atmosphere' - presumably 'middle/upper atmosphere'.

L55: I think 'respects' sounds more conventional than 'aspects'

L61: Missing '.' after dataset.

L143: 'is the concentrations' doesn't sound right. Perhaps 'is the handling

     of the concentrations' ...

L178: Did you really mean 'e.g.' here? Or should it be 'i.e.'? In other words,

     are there species other than O and H?

L180: Here you give O3(v1,v2,v3) but elsewhere you only refer to O3(v1,v3).

     Is that intentional?

Fig 1: 'solstice' would better describe NDJ rather than DJF, similarly 

     'equinox'.

 

L221 & 253: 'night-time' (no hyphen when used elsewhere)

L279/280:  'in Clarmann2022)' - missing 'von ... et al (' 

L310: I'm not sure 'notoriously' is the right word to use here. Maybe just

     'notably' or 'significantly'.

L311: 'so is' should go before 'MIPAS sensitivity'.

L322: '0.00029' - oddly specific, particularly for an error estimate. Perhaps 

      better just rounded to 0.0003?

L368: 'as it was required' ? perhaps 'as would be required' was meant. 

L410: Extra ')' required.

L485: 'not in vain' - not sure what this means here, perhaps just remove it

     and the bracketed phrase still makes sense).

L620: 'consistent' (spelling).

-------------------