The work by Garcia et al. focus on the impact of the instrumenta line shape of the high resolution FTIR spectrometer on the retrieval of vertical  ozone profiles. The authors evaluate ver systematically more than 8 retrieval steregies to use y estimate the instrumental lineshape and evaluate these strategies using the long term measurements of FTIR spectra together with measurements from Brewer and  the frequent ozone  soundings. These collocated measurements at the Izaña observatorio are an unique posibility to evaluate the different retrieval strategies with respect to  the instrumenta line shape and  to study the sensitivity of the ozone profile retrieval on the estimation of the ILS.

The optimization of  retrieval strategies for vertical gas profiles and especially of ozone by a very focused study  is going to improve this remote sensing method towards a more exact measuremet method which is an important step for measurements in  the NDACC network. The work fits perfectly in the scope of AMT, the manuscript is well written and clear presented.

In my opinion the manuscript is ready and paper should be published as is.

Just some few specific comments:

I would recomend to include a headline like "relative difference" in table 1,3,4 andit would be helpfull to see in addition to the infromation with respect to the  brewer column measurements, a relative diference between the retrieval sterategies, maybe with respect to the favourite retrieval sterategies of the authors. This would ilustrate clear how sensible the  ozone profile depend on the retrieval sterategy with respect to the ILS.

I would recomend to include a little more basic information  about the modulation eficiency and the phase error in the begining of the section 3.1, so that the reader get the important information which  and how the ILS depends on both parameters without consulting the cited papers.

It is very interessting that the authors recomend to fit the pase error at ZPD, but use the modulation efficiemncy from the cell measurements. It would be very nice if the authors might try to give a possible  physical  explanation, why a fit of the "pase error" at ZPD from an individual spectrum imporves the retrieval? Does it depend on alignment the temperature of the beamsplitter the phasecorrection duringt the calculation of the spectrum?

Comment on amt-2022-44 “Impact of instrumental line shape characterisation on ozone monitoring by FTIR spectrometry” by Omaira E. García et al.

General comments

The IRWG of NDACC is an international network bringing together more than 20 observational FTIR sites over the world since the 1990s. In comparison to other observational methods used for the investigation of gaseous composition of the atmosphere, FTIR spectrometry provides a unique advantage of simultaneous measurements of the total columns (or even profiles) of a number of climatically active gases. The significant efforts are being made by the IRWG community to develop the unified retrieval strategies for deriving total columns/profiles of atmospheric species including O3. The main target of these efforts is verifying and harmonizing the results obtained by different groups operating FTIR sites. To get reliable information on O3 trends in the stratosphere which are currently nearly zero, it is necessary to provide the FTIR products of high accuracy and precision. Achieving this goal requires knowledge of the parameters characterizing the alignment of FTIR spectrometer (instrumental line shape function, ILS) and correct accounting of ILS in the retrieval procedure. Paper by García et al. is devoted to the detailed study of the influence of several ILS approaches (used in the procedure of inverse problem solving) on the O3 retrieval results (focusing the stratosphere). FTIR instruments having different alignment status are considered.

The manuscript corresponds to the AMT main subject areas and can be recommended for publication (minor revision is required) after a few points are addressed (please, see specific comments section).

Specific comments

Results presented in Appendix B deserve to be moved to the main text of the paper as a separate section. But the final decision is up to the authors.

Authors tested several approaches of ILS for the getting best retrieval results on O3. Is there a “universal” recipe for the processing FTIR observations (for example, archive spectra) in the absence of information on instrument alignment (ILS function)? Is it correct that in the case of the ideal ILS function should be used for overall spectra processing?  Is it possible to create a homogeneous O3 row by stitching separate O3 time series obtained as a result of processing FTIR spectra using different ILS approaches? Analysing such a complex time series can be an additional challenge to reveal long-term O3 trends close to zero.

It is not quite clear whether the AVKs (averaging kernels) were taken into account when comparing the O3 results obtained by the FTIR and Brewer techniques?