The paper reports the new retrieval of temperature in the mesosphere using MIPAS latest realeased level 1b spectra. I am not a native English speaker, therefore I just made few language corrections.

It deserves to be published after the following comments have been addressed.

Line 29 : Add some reference to the ESA site or to the official document

Line 34: This part is officially known as Optimised Resolution, please add this name when you first mention the part of the measurements

Line 39: substitute ‘a few days in a row’ with ‘few (or write the number if it always the same) consecutive days’

Line 70: I assume each dataset refers to a single observation mode, I suggest making it clearer in the text here.

Line 114-117: please add some reference to documents or papers that describe this.

Line 124-125: there is a new release of HITRAN (2020) why did you not use it?

Line 137: But you do not retrieve continuum above 58 km, so why you find problems?

Line 144: substitute ‘km in just’ with ‘km just’

Line 154-155: The sentence should be changed in ‘In V5R_t_m21, the temperature ….etc.’

Line 192: Substitute ‘in each iteration’ with ‘at each iteration’

Line 193: Substitute ‘change slightly’ with ‘have been slightly changed’

Line 200: Substitute ‘is’ with ‘are’

Line 208-210: Hard to see what you say in the figure due to the adopted color scale. Also checking the AK it looks that very small information is contained in the measurements above 100 km, so this comment doesn’t hold very well

Lines 213-215: Again very hard to see in the figure

Lines 219-222: Could you please explain better this point, I could not follow it! What ‘elevated stratopause events’ are?

Line 224: ‘wrinkled’ is not very scientific, use oscillating or something more appropriate.

Line 230: what is the threshold you use of the AK peak to decide if there is information?

Lines 235-239: Could you explain why the vertical resolution oscillates so much?

Line 245: you say 34 scenarios: shouldn't them be 40? 5 latitude bands X 4 seasons X 2 (day/night)

Line 249: You say ‘in the following’, is it in the table or in the text?

Line 253: ‘propagation of measurement’ -> ‘propagation of the measurement’

Line 255: The noise values change very much from the region around 700 cm-1 to the region at 900 cm-1, I would rather use the range than the average value here

Lines 256-259: The calibration is performed as part of your analysis or as part of the level1 process?

Lines 260-261: do you analyse unapodised spectra? Otherwise the ILS shape should be dominated by the apodization used

Table 3: even if you explain it in the caption the word chief does not suggest what you have introduced in the column, could you use a different and more appropriate word?

Lines 275-278: HITRAN usually reports spectroscopic uncertainties, why haven’t you used them?

Lines 279-283: the whole paragraph is rather confusing, could you explain it better? An educated guess is not a quantity, could you explicit the guess you made?

Lines 296-297: Some of the random errors are systematic, but of a random nature….

Lines 305-306: Do you mean that you did not know the covariance matrix associated to those errors?

Lines 307-309: The whole sentence is not very clear, could you describe this better?

Line 313: what do you mean with ‘chiefly’?

Line 315 ‘for 34’ -> ‘for the 34’

Line 430: Which NLTE model does Saber use? If it is the same as in this paper it should be mentioned, as in this altitude range it is the main contributor to the retrieved temperatures.

Line 481 onward: I did understand that the comparison were done on the smoothed profiles only, so please clarify this better in the text!

In the references at  Line 699 ‘ Fera, S.D.’ -> ‘ Della Fera, S.’

Review of “Version 8 IMK/IAA MIPAS temperatures from 12-15 micron spectra: Middle and Upper Atmosphere modes” by Garcia-Comas et al.

General Comments:

This manuscript describes a new MIPAS temperature dataset from 18-115 km altitude, which the authors call version 8.03. In addition to improved calibrated spectra, the authors detail a variety of improvements to the retrieval and show several comparisons against previous versions of the data as well as comparisons against existing SABER temperature observations. The paper provides a comprehensive description of the uncertainties and overall, the figures are instructive and clear. Although the authors spend considerable effort comparing their results to SABER temperatures, comparisons with recently re-processed MIPAS temperatures using nitric oxide emission are quite limited and it is not really clear why. The manuscript is also lacking detail in some places, particularly in explaining the results in the upper altitude region of their dataset. The reviewer recommends publication provided that the authors address the concerns enumerated below.

Specific Comments:

1. Line 97 and elsewhere. The authors spend considerable effort comparing to existing SABER temperature observations. Could the authors please comment on and show how their results compare to MIPAS thermospheric temperatures derived from nitric oxide spectra at 5.3 microns (e.g. Funke et al., Atmos. Meas. Tech., 16, 2167-2196, 2023)? There is mention of day-night differences on lines 384-387 and latitude variations on lines 394-396. However, a direct comparison between the two MIPAS temperature datasets would be particularly useful for the reader. If there is a reason for this limited comparison or if this has been done in other publications then please say so and give a reference. Thank you.

1. Lines 165-179. Could the authors show how the model results used for atomic oxygen compare to the recent MSIS 2.0 for atomic oxygen (Emmert et al., Earth and Space Science, 7, e2020EA001321. https://doi.org/10.1029/2020EA001321, 2020)? More importantly, how do any differences in MSIS 2.0 atomic oxygen affect the temperature retrieval?

1. Line 230 and Figure 2. Are these kernels available to the reader? They would be useful for comparing with models or other datasets, but it is not clear from the data availability statement at the end of the paper that they are available. Note also that the reviewer could not connect using the link to the supplement at the end of the paper.

Technical Corrections:

1. There are many acronyms throughout this abstract that are not spelled out. Please correct these when first mentioned. Thank you.

1. Lines 25-26. Please give a number here (+/- X K) and indicate the altitudes at which this applies. SABER faces similar non-LTE challenges in that temperature retrieval. From Figures 12-15 the agreement gets worse and the combined systematic uncertainties get much larger above 90 km so the authors need to be more explicit about this in the abstract.

1. Line 30. Please give a reference for these v5 spectra.

1. Line 80. A brief paragraph describing the Envisat mission, as well as the orbital inclination and the equator crossing local times would be very helpful to the uninitiated reader here in order to give valuable context to the dataset.

1. Line 149. What do the authors mean by “bias-corrected”? Is this a bias in the model and if so, why is this done? Given the importance of atomic oxygen in the temperature retrievals, one or two sentences explaining this would be helpful.

1. Figure 2 caption. To what does 30 degrees refer?

1. Line 270. The reviewer does not understand how CO2 uncertainties can be calculated from the WACCM results. Why can this uncertainty not be estimated by ACE and/or SABER data alone? Please explain in the text.

1. Lines 289-295. Again, the reviewer does not understand how an atomic oxygen uncertainty can be calculated from the WACCM results. Please explain in the text. Furthermore, it is well known that the NRLMSIS00 atomic oxygen values above 90 km are inaccurate (e.g. Sheese et al., J. Geophys. Res., 116, D01303, doi:10.1029/2010JD014640). The reviewer requests that the authors either crop their Fig. 4 at 97 km or use the newer more accurate values from MSIS 2.0 (Emmert et al., 2020).

1. Line 355-356. How is does the quenching rate used compare to what is used by SABER? Please state this quenching rate here and state both if different.

1. Figure 9 caption. If the “M61” title is V8 and “M21” is V5 then please say so in the caption.

1. Lines 383-387. Given the importance of tidal variability, it would be important here to (re)state the equatorial crossing local times of the Envisat orbit.

1. Lines 431-432. Do the authors mean +/- 1000 km and +/- 2 hours? Please be explicit. Also, if “2 hours” means 2 hours in Universal Time rather than Local Time, the authors should say that as well. Given the importance of tides in the MLT, if “2 hours” means +/-2 hours UT then a statement in the text or the captions about the local times used for the MIPAS data and those used for the SABER data would also be useful.

1. Line 450. Here it would also be helpful to the reader to also state how CO2 densities and O densities, as well as their diurnal variations, are specified in the MIPAS and SABER datasets. A few sentences with references would be very useful in diagnosing the comparisons between the two instruments. This is particularly important for altitudes above 100 km and also particularly important if the values used are different.

1. Figures 12-15. In all of these figures, the reader cannot see the extent of the combined systematic errors at the top. Please either crop the figures at the top or expand the x-axes so the reader can see how big these are.

1. Figures 12-15. To what does the “Nc” refer? If these are the number of coincident profiles, then please say so in the caption. Also, does that mean that the coincidences are pairs such that there are exactly Nc SABER profiles and Nc MIPAS profiles? If there are different numbers of profiles that fall within the coincidence criteria then please state that explicitly. If the numbers of profiles are different for each instrument then please give both.