This paper analyzes data collected over a 3-y period by a ground-based elastic lidar located at Dome-C.  In particular, the authors derived lidar ratio (i.e., the ratio of the extinction to backscatter coefficients) from this dataset, processed the data to provide a weighted single value for each identified cloud, and then analyzed these cloud observations as a function of 4 different cloud types to characterize the median lidar ratio (LR) for each type. 

I found the paper well written, the methodology well explained, and the results intriguing. I have a few questions and points that I believe would strengthen the paper before it be accepted for publication.

• Ultimately, from the 144 measurement days that spanned the 3-y period, there were less than 400 individual clouds that were analyzed (Table 1). And line 219 states that “each detected cloud layer represents one independent realization for the lidar-ratio retrieval.”  However, there seems to be many, many more points that are displayed in the 2-d histograms in the figures.  I can only conclude they are showing individual profiles that were measured in each of these clouds in those histograms, which brings up questions about the temporal correlations within a cloud (i.e., these are not independent samples).  The authors need to be more clear about this sampling, and the impact of possible temporal correlation on their analyses.
• The authors reported (line 135) that there are ancillary radiosondes and model output that provided temperature and pressure, from which ancillary molecular backscattering coefficients could be computed. However, these temperature data were not used in any of their analyses, either to select the height of the tropopause (I do recognize they performed a sensitivity test here, but why not use the actual value?), to evaluate some of their hypotheses (e.g., line 340 when they suggest the difficulty of isolating pure ice layers from mixed-phase structures; knowing if the temperature is above/below -40 Celcius would be useful, or e.g., line 353 when they discuss small vs large NAT crystals), or analyzing the LR values (LR does not depend on height like Fig 4 or Fig 8, but it can be related to temperature which helps constrain moisture). 
• Line 466 suggests that analyzing the subset of Young-optimized retrievals results in only a modest shift in median LR. However, for LS_ice, the value went from 52 sr to 32 sr; this is not modest.  And given there are only 9 vs 2 clouds of ice PSC (Table 1), it seems these statistics are extremely uncertain (especially due to temporal autocorrelation; see first comment).  Please provide more discussion here.
• Line 479 leads to an intriguing thought: if the Young criteria can be used to identify cloud inhomogeneity, then it seems that it would be useful to indicate that fraction of each observed cloud type that could be considered homogeneous. This could be added to Table 1.  But more than just adding an additional column to that table, please see if there is supporting literature that supports the ‘estimated fraction that are homogeneous’ for each class.

General comments

This manuscript presents a methodology for estimating the lidar ratio (LR) of cirrus and polar stratospheric clouds (PSCs) over Antarctica, using observations from a backscatter lidar at Dome C. The study applies the Young inversion method (Young, 1995) to a strictly selected dataset and compares the results with those obtained from the full dataset.

Overall, the manuscript addresses an important topic for atmospheric sciences, and in particular for the lidar community. However, the presentation of the methodology lacks clarity in several parts, making it difficult to follow the processing steps and underlying assumptions. I believe the manuscript would benefit from a major revision before being considered for publication.

Please find below a list of major and minor comments.

Major comments

1.    The Young inversion method is not sufficiently explained in the methodology section. The current description is vague and would benefit from a dedicated paragraph clearly reminding of the principles of the method, its assumptions, and its implementation in this study. Please explain what it does, based on given inputs and outputs.

2.    The wording throughout the manuscript is sometimes overly complex or not aligned with standard terminology in the lidar community. The authors are encouraged to improve clarity by simplifying and using standardize the language. Examples include:

“polarization diversity Rayleigh lidars” → “elastic backscatter lidar with depolarization capability”

“extinction corrections” → “extinction retrievals’

3.    Key variables (e.g. Rfixed, Rcor) are not clearly defined when first introduced. A clear and consistent definition of all symbols and acronyms is necessary.

4.    The description of the methodology, particularly the iterative procedures involved in the LR retrieval, is difficult to follow. It appears that multiple iterative processes are used (e.g. cloud boundary detection and Young inversion), but these are not clearly distinguished. A schematic diagram or a step-by-step description of the full processing chain would greatly improve readability.

5.    The results are mainly compared to literature values. The manuscript would be significantly strengthened by including an independent evaluation, for example through comparison with satellite observations such as CALIOP (on CALIPSO mission, for the data until June 2023) or ATLID (on EarthCARE mission, for the data starting mid-2024), particularly for the year 2024 if data are available.

6.    Abstract: One important result of the comparison between Young-derived and full-dataset LR retrievals, is the cloud variability and mixing. This could be highlighted in the abstract.

Minor comments

1.    Line 32 (and throughout): “LR in the range 40–70 sr”: please specify the wavelength.

2.    Line 35: “S = 40 sr”: ensure consistency in notation for the lidar ratio throughout the manuscript. In most of the paper lidar ratio appears to be denoted as LR.

3.    Line 52: “extinction correction”: consider using “extinction retrieval,” which is more standard terminology.

4.    Line 56: “physically based alternative”: please clarify what is meant by “physically based.”

5.    Line 60: “not yet fully characterized”: if previous studies exist, they should be cited to properly position this work.

6.    Line 60: Dome C is mentioned for the first time. Consider adding a brief description (e.g. altitude, Antarctic Plateau context), so that the reader can better follow the subsequent paragraph.

7.    Lines 62–63 vs 87–88: Some repetition is present

8.    Section 2.1: The section “Concordia Station” includes extensive discussion of lidar characteristics. Consider splitting into subsections (e.g. ‘…Dome C, located at ~3km a.s.l on an ice dome on the East Antarctic Plateu’). Then you can continue with the next paragraph which provides more information on the conditions.

9.    Line 93: What is the maximum range of the lidar?

10.    Line 94: “which is the normal procedure”: clarify if this refers to 32-minute averaging or to some other aspect.

11.    Section 3: Consider splitting into subsections such as “Lidar processing” and “Young inversion.”

12.    Line 128: “uncertainty profiles are smoothed”: specify the smoothing scale and method (e.g. 300 m moving window?).

13.    Line 132: “weighting function produces a smooth transition”: specify the altitude range over which gluing is applied.

14.    Lines 139–143: Given known calibration issues (e.g. Di Liberto et al 2024), it would be helpful explain why the depolarization calibration is important, before describing the procedure.

15.    Line 144: Symbol usage may be misleading compared to standard conventions (e.g. Freudenthaler et al., 2016 that uses G for cross-talk). Please ensure consistency.

16.    Line 167: “attenuation-induced curvature of Ratt”: this wording is unclear and should be clarified.

17.    Line 178: Is the choice of 240 m window size and 180 m offset based on some criteria? Please answer the question in the paper.

18.    Lines 164–181: The processing steps are difficult to follow. A diagram or structured description is recommended.

19.    Line 199: “180 ms” → likely “180 m”.

20.    Lines 213–215: The meaning is unclear; please improve text for clarity.

21.    Lines 232–243: Have the authors considered including a cloud thickness criterion to improve classification? Please answer the question in the paper.

22.    Line 272: What is the minimum number of points required per bin? Please answer the question in the paper.

23.    Line 297: “emanating” → consider “extending.”

24.    Lines 319–321: Have the authors tested relaxing the selection criteria to increase the number of Young-compatible layers? Please answer the question in the paper.

25.    Line 331: “LR increased with both backscatter and depolarization”: this statement appears too strong given the observed variability. Looking at the plot I must say that the linear relationship described here is not always the case.

26.    Line 347: ‘…composed of H2SO4/…’, please add citations here.

27.    Line 398: The described oblique lines are not clearly visible.

28.    Line 399: “increase” → “increases”.

29.    Lines 396-402: The particle backscatter (beta_aer) as well as particle depolarization (delta_a) are discussed in this paragraph but not shown, which makes the description of the plot difficult to follow. Please consider rewriting this paragraph in a way that is easier for the reader to follow what is seen in the figure.

30.    Line 419: “R ~ 50 sr” → should refer to LR.

31.    Lines 438-440: Could the authors say a bit more about this plot? (e.g. what was expected from the literature). Please answer the question in the paper. Also the profile seemed to be relatively constant especially when reading the description of Fig.4, where STS is described constant with height.

32.    Line 440: Typo at the end of line.

33.    Line 449: ‘An increase in the lidar ratio generally reflects a shift toward larger or more aspherical particles’, the authors should clarify whether they  are discussing only about cloud particles as certain particle types do not necessarily follow this relation. (e.g. biomass burning).

34.    Lines 450–451: “complex particles”: please clarify (e.g. shape, composition).

35.    Line 466: In addition to LR shifts, the presence of multiple modes in Young retrievals appears important and could be discussed. Also should discuss a bit more in the text whether this modes are realistic.

36.    Line 521: Typo (“structures .”).

37.    Line 530: “real cloud retrievals” → consider “standard retrievals.”

38.    Line 540: The EarthCARE mission is already in orbit (not upcoming).

39.    Figures 1, 2, 5, 6: Consider consistently showing classification regions across figures.

40.    Figures 3, 7: Clarify dashed lines in captions (e.g. meaning of distribution ).

41.    Table 1: Clarify the definition of “cloud layers,” as the numbers do not appear consistent with the number of points shown in the figures.