This is a very interesting and useful study on the impacts of wind-driven changes to Ku and Ka radar returns from ground-based observations during MOSAiC. The study is quite thorough and rigorous, I just had a few minor comments and suggestions for the text as noted below. I would suggest publication subject to some minor revisions.

Overall, I find the results to be quite interesting to ponder as they show a very detailed look at wind and roughness induced changes in Ku and Ka radar returns. The authors make clear this is a step towards interpreting what factors influence the often-times complex radar returns found in airborne and satellite radar altimeter data, there is not necessarily definitive conclusions to be determined for going from these results to altimeter data but the results are certainly intriguing and worthwhile to publish. I do wonder what this might mean for next steps in terms of future experiments with radar systems on field sites such as this, perhaps this could be added to the end discussion to further highlight what the significance of the data and results may be.

Minor comments

L38: “snow redistribution events increased the dominance of the air/snow interface at nadir as the dominant radar scattering surface” Is the use of the term “dominant” here redundant, or purposeful?

L73-74: I’m not sure the term “originate” is applicable here, perhaps stating they are assumed to be the dominant scattering surface is more appropriate.

Figure 1(a) and (b): What is “foot” in the figures? I think the caption may be describing this, but it would be good if consistent terminology is used.

Also L130-131: Is the pedestal here just the platform the radar sits on? Not the phase center of the radar or some determined zero range point?

L128: Isn’t the center frequency a bit off from the CryoSat-2 frequency? Or do you mean the frequency ranges overlap?

L137-138: Given the small range of incidence angles of radar altimeters like CryoSat-2 and AltiKa (mentioned previously in the text), I’m curious what motivated the reasoning for the 5 degree incidence angle steps? Do you not expect there to be much variation at smaller incidence angles or was this an instrument limitation?

L189-191: What approximate time interval did this integration over the amplitude thresholds typically occur? I think including that could give further evidence to support the fact that you expect the returns to capture the entire snow interval.

L193: A similar statement about the time interval could apply here too. So long as there weren’t sea ice ridges or very high surface features in the footprint I think a smaller interval as stated here is appropriate.

Figure 8: Are these plots from data averaged between -5 to 45 degrees? It seems so from the caption. Is there much difference in plots showing only data from the nadir direction? I think to some degree this is shown in Figure 9, but I do like the waveform plots in the bottom panel of Figure 8 as a way to show the waveform structure in similar manner to altimetry data.

Review, "Wind Transport of Snow Impacts Ka- and Ku-band Radar Signatures on Arctic Sea Ice" by V. Nandan et al. 

The manuscript presents waveform and backscatter time series aquired over snow covered sea ice with a ground-based radar altimeter and scatterometer. The instrument has an exceptional bandwidth of 6 GHz at Ku-band (13.6 GHz) and 10 GHz at Ka-band (35 GHz) resulting in centimeter slant-range resolution. The instrument can provide detailed and relevant information to interpret data from existing and future radar altimeter missions like CryoSat-2, CRISTAL, SARAL, etc. Publishing such data and time series is a valuable contribution suitable for the journal The Cryosphere (TC). The manuscript, however, requires thoroughly revision (major) to meet scientific publishing standards and to focus more concisely on the most relevant aspects of the valuable work.

The most relevant points of manuscript are: 

- Very few ground-based radar data especially over sea ice exists. The manuscript contains detailed plots of such data which provides valuable observations.

 - The time series plots of very high (centimeter) resolution Nadir backscatter returns reveal the dynamics of the relative scattering contribution of the main interfaces air/snow and snow/ice. The data show that air/snow interface is not necessarily the strongest scattering contribution , even at Ka-band, which is valuable for future radar altimeter missions. 

 - during pre-wind conditions the dominant radar return (at 13 and 35 GHz) at nadir alternates between the air/snow interface and the snow/ice interface.

 - The creation of hard Wind slabs causes an increase in backscatter at Nadir and a decrease in backscatter for non-nadir angles (observed at 15...50°).

 - The comparison of Nadir radar returns with off-nadir (theta_inc > 15°) is a valuable information regarding penetration depth of radar sounders/altimeters (less penetration) vs. SAR systems (larger penetration due to slant-looking gemometry).

General comments: 

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

The manuscript needs considerable improvements in readability and focus. 

 - Try to focus/condense on most relevant aspects (see above), 

 - consider to shorten where possible, e.g. the polar plots seem not to reveal too much information;

 - results/discussion/conclusion: check for redundant, irrelevant information. 

 - Give the main points more weight by improving the graphics showing these results.

 - I really don't understand what is the exact extend and overlap of the observed area with Ku- and Ka band. Different ranges of azimuth angles are given in the text and figures; please clarify this using an observation-setup-sketch showing all relevant geometric parameters.

 - Try to provide more convincing conclusions where possibly and try to reduce speculative interpretations throughout the manuscript.

 - There are several interpretations/speculations related to surface roughness: consider removing them where not really necessary.

Specific comments: 

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

35/36: It would be good to mention in the abstract what kind of radar is used and for which observation which mode was used: scatterometer or altimeter or both modes of the instrument? 

37: "waveforms and backscatter": I guess, they were measured in altimeter mode? Please mention.

39/40: "...altimeter. The relative ... decreases, ... with increasing incidence angle." For the described observations, it seems like the scatterometer mode was used here? The sentence before ends with altimeter; this sentence mentions "increasing incidence angle" which does not match with altimeter. Please clarify; maybe start with "With increasing incidence angle of the scatterometer, the relative scattering contribution ...  ..." if that agrees with what you mean.

124 and Fig. 1b: It would be good to 

 - 1) refer/use Fig. 1b to better explain/clarify how the "azimuth range" is scanned, at discrete incidence angles. I guess, Fig 1b shows the azimuth-elevation scan pattern of the instrument; 

 - 2) this could be mentioned in the caption and as a axis-label on Fig. 1b (at least azimuth).

 

131 and Caption of Figure 1a: What is the difference between radial distance and range and why is that important? Please mention. I would expect parameters like height of the antennas, slant-range.

132: How is the beam-width defined? full-width-at-half-maximum (FWHM) equivalent to the "3dB-beam width"? Strove 2020 writes "Antenna 6 dB two-way beamwidth".

135: "The overlapping footprint is between -5 and +45° for Ku-band, and -45 to +5° for Ka-band": I don't understand this sentence. To me, this seems to contradict the sentence before and the geometric radar setup as shown by Strove 2020. The overlap in % (as given by Strove) of course depends on incidence angle, but why is the azimuth angle mentioned here? Why is the "overlapping footprint" for Ku-band at positive angles and at negative az-angles for Ka-band?

Related to that, I don't understand the yellow-orange-red color in Figure 6; I don't understand why Figure 7 shows az-angles of -65..+25° for Ku-band and -25° to +65 for Ka band; I don't understand why Figure 10-12 show azimuth angles for -45 to +45 for both Ku and Ka band. Consider drawing a measurement-setup figure indicating all relevant geometric parameters like footprints, scan areas, etc. for both, Ku and Ka band.

137: what are the scan-increments (in degree) in the azimuth direction? For incidence angle it seems to be 5°.

Section 2 or 2.2: I miss here an overview map figure (e.g. the blue sub-figure in Fig. 4 could be used) showing the location of instruments. Such a figure is very helpful for the reader to imagine where which instrument is installed and where which sampling/measurement was done.

196-202: I don't understand the azmimuth sectoring method: 

- why are negative and positive theta_az sectors mentioned separately? Is there any special relation between negative and positive angles?

- How is the number of independent samples calculated? I would expect something like azimuth-angle-width / antenna-beamwidth * analyzed-range / range_resolution. What is "theta_az width"? Why do you devide the antenna beam-width? Why half of it? What are the range-gates? It this the range-sample spacing or the range-resolution? (note: increasing the sampling of a band-limited signal does not increas the number of *independent* samples).

Table 1: Why does the number of independent samples change with $\theta_az$ at Nadir? Is is $theta_az$ a range of azimuth angles or does $\theta_az$ describe a specific angle? Or are the numbers in the table given for a 5° theta_az bin? 

207-216: Definition of the CPD: Note, that in Leinss 2016, the CPD is derrived from S_VV * S_HH* while here it is derrived from S_HH * S_VV*, resulting in a sign change. The definition in Leinss 2016 is motivated by the desire for a positive change for increasing snow, while your definition makes more sense in terms of radar terminology. Make sure, you describe the increase/decrease of the CPD in agreement with the used definition (HH * VV* vs. VV * HH*).

220: At which time did WE1 start? 

Figure 2: I understand that the daily wind-rose plots might be important but I miss temporally-resolved plots indicating hourly or sub-hourly resolved wind speeds. This could be plotted together with either temperature or air-pressure.

Figure 2: Could you indicate in the plots, similar to Fig. 3, when the wind-events occured? This would simplify the understanding. To save space you could also plot temperature as a continuous time series (one plot, like Fig. 3a and b) with the series of daily wind-roses above or below.

235, Caption: 

 - Surface plots -> I would say this is a 2D color plot, it does not show any 2D surface in 3D space (what I would expect for a "surface plot" in visualization-terminology. I would also avoid the word surface here, as the plot shows a time-series not a snow(or other physical)-surface.). 

 - "Yellow pixels represent snow volume": I know what you mean, but this statement does not make sense. Better indicate by a box or line where the snow volume is (see comment below). Yellow represents strong temperature gradients (within the snow pack).

 - with the top 20 cm representing the distance between the first sensor located above... and at...: I don't understand this sentence. How many sensors are above the air/snow interface? Do the top-20cm represent the height above the air/snow interface, i.e. everything above 20cm is air? How did the snow height change within the shown 8 days? Could you draw the (possibly estimated) air/snow interface into the plot? 

235, Figure: Could you indicate in the plot what you consider as near-surface, snow/snow volume, sea-ice, and ocean? Could you also indicate that Fig. 3d is (I think it is) a zoom/subsection of Fig- 3c? You could indicate this by drawing the outline of the zoom into Figure 3c.

252 - 253: From figure 3d there seem to be temperature gradients up to 7 or 8 K/cm during WE1, in the text I read 3 °C/cm for WE1. For WE2 fig. 3d indicates gradients of around 2-3 K/cm. Please check consistency.

254: 0.25 °C/m: do you mean 2.5 °C / cm? 

257: uppermost: could you provide a number like e.g.: uppermost ... cm? 

274: I doubt that breakup of snow particles decreases the SSA. No matter how SSA is defined, as surface area per volume or kg, breaking up crystals would increase the SSA because the grain size get's smaller by the breaking events. King 2020 describes the SSA decrease in wind slaps rather as a "product of mechanical wind rounding and subsequent sintering".

305: "superimposed on the TLS data in yellow ... and orange where the two overlap": I understand the yellow-to-blue colors in Figure 6 so that black indicates no TLS data (count=0). Why are there then a yellow or orange colored radar footprints or TLS data that are located on black pixels? Or does yellow and red indicate two different radar observations? Please clarify in the caption what is yellow, orange and red. As mentioned earlier, I did not understand the difference in radar observations in the positive and negative azimuth angles theta_az. 

Figure 6 and Figure 7: The TLS data and the radar-Nadir data indicate a possibly considerable slope within the observation area. Looking at both figures, I can estimates slopes of 2-5°. Would it be possible to make any statement about changes of the local slope in the observation area? As observed later, in Figure 9, the incidence angle has a very significant effect on the radar waveform, hence the local incidence angle must also have an significant effect. I can't tell of 2-5° are already significant, but I belive they are.

Figure 7: I find it quite confusing that Figure 7 shows TLS profiles from different dates in every figure. Why not showing the TLS profile from 08 Nov (and possibly 01 Nov) for the radar data from Nov 09 , 11, and possibly 11 and the TLS data from Nov 15 ontop of the radar data from Nov 15? If you perfer to keep all three TLS profiles, then please mention why different TLS profiles are shown ontop of the radar data for each date.

378, caption 8: 

 - "red, yellow, black": I do not see a yellow line in the figure. Consider refering only to the dashed red and black lines. 

 - But keep the sketched yellow arrows in the figure (very interesting!) and the sentence refering to them.

 - Mention the meaning of the vertical gray lines in the caption, I guess they constrain WE1. You could also add a label "WE1" to the figure.

 - "Time series of the interfaces NRCS values are snown below the echograms" -> "NRCS time series of the two interfaces (red, black) are snown below the echograms for HH and HV (dashed/dotted)". Also add a legend to the timeseries indicating the colors and line-styles.

 - Figure 8 has many sub-panels. Consider labeling them with (a, b, c...). Is figure 8, bottom right mentioned in the text? If not, it seems not to be too important and could be removed.

 - Figure 8 (and other figures): Consider labeling the time-axis according to the shown tick labels, e.g. "Date (YYYY-MM-DD)" or "Date / time (MM-DD HH:mm)"

356: "It is interesting ...can still be seen ... 10 November, .."  add a reference to "yellow arrows, Figure 8".

359-360: "During WE2, ... the air/snow interface moved upwards... (bottom right of Figure ... and 8)": I do not see any effect of WE2 in Figure 8. The x-axis of Fig. 8 ends likely at end of the day 2019-11-15. Could you indicate, similar to WE1 by gray vertical bars, where WE2 is located in Figure 8? 

393 and 394: Both sentences: Add reference to the corresponding panels in Figure 8.

415-416: "... at all theta_inc. VV and HH backscatter primarily originates as surface scattering at the air/snow interface" I think this might be an inadmissible generalization. The fact, that Nadir observations indicate the strongest return at the air/snow interface cannot be generalized to all incidence angles. On the contrary, the backscatter at the air/snow interface might even be reduced for non-zero incidence angles due to specular reflection away from the radar.

420-429: This seems more an interpretation of the Nadir observations in Figure 7 and 8 rather and seems less relevant or even misleading for the interpretation of off-Nadir observations. Consider moving to the interpretation of figure 7 and 8 and refer only briefly to this interpretation when discussing the non-nadir angles in Figure 9.

436-439: I think observation of different layers at non-zero incidence angles is difficult due to the slant imaging geometry. In the slant geometry each slant-range bin samples the backscatter from all targets located at the same range (bin). This includes contributions from the air/snow surface, the volume and the snow/ice interface. As for each slant-range distance a slant cross section through the scattering volume is measured, the measured profile is rather a representation of the beam-pattern weighted by the incidence-angle dependent backscatter intensity from the different surfaces illuminated by the beam. 

Similar to Figure 9 and 10 in [Leinss et al. EUSAR 2014] I interpret the strong near-range return in Figure 9c (at theta_inc=15°), possibly also 9d (30 deg) as a Nadir-return from the air/snow interface (consider the beam-width of 12-17°). However, the increasing slant-range distance to the strongest return is a good indicator that the snow above the ice delays the radar signal.

439/440: "The waveform analysis shows": Please be more specific. Similar to the observations with the SnowScat instrument, where dry arctic snow up to 1m depth appeared almost transparent at incidence angles between 30 and 60° and X to Ku-band frequencies [https://doi.org/10.1109/JSTARS.2015.2432031], I interpret the increasing range to the strongest return caused by 1) increased delay to to increased accumulation and 2) caused by the snow/ice interface. See also Figure 9 and 10 in [Leinss et al. EUSAR 2014] where the snow/soil interface is detected. This interface appears further away from the sensor (due to increasing delay) with increasing snow depth. 

Reference: Leinss, Silvan and Lemmetyinen, Juha and Wiesmann, Andreas and Hajnsek, I.. (2014). Snow Structure Evolution measured by Ground Based Polarimetric Phase Differences. in Proceedings of European Conference on Synthetic Aperture Radar (EUSAR)] https://www.researchgate.net/publication/312171448_Snow_Structure_Evolution_measured_by_Ground_Based_Polarimetric_Phase_Differences

Figure 10, 11, 12: These figures take up a lot of space but seem to reveal little information. Consider removing them (and the associated discussion) if you agree that they are not crucial for the main points of the manuscript.

451: "Next, we show..." Sentence should be used as introductory sentence for section 3.3.2 and not at the end of section 3.3.1.

482: "stable snow metamorphism": metamorphism is a dynamic process. Do you mean "stable snow conditions with little metamorphism" or "continuous metamorphism"?. Variations of the CPD indicate metamorphism, however a single observation of the CPD does not provide any information about the dynamics of the snow pack. Nevertheless, it can give an idea about the history of methamorphism of the snow pack.

499: What is "phase reversal"? Do you mean "phase wrapping"? 

480-502: I have some doubts on the results and interpretation of the CPD.

- First of all, the backscatter results in the previous sections are most convincing when averaged over azimuth (but not over incidence angle!) and plotted with temporal resolution (like Figure 8 and 9). Why not showing such plots for the CPD first (at different incidence angles, because the CPD is strongly incidence angle dependent)? If temperature-gradient-metamorphism induces variations of the structural anisotropy then these variations should be well visible during the extreme temperature gradients up to 800 K/m shown in Figure 3.

- Could the authors ensure that they use the same definition of the CPD as in Leinss 2016? See comment above: Currently, there might be a sign error. The sign of the CPD might also dependent on the data processing and chosen side-bands in the electronics of the instrument. 

- Did the authors ensure that the CPD of the instrument was well calibrated? Snow-free data and open water should have zero CPD.

- An analysis of CPD time series, averaged over the whole area might also indicate calibration issues.

- The instrument seems to allow processing of the data at user-defined bandwidth and central frequencies. Phase wraps can be easily detected by using two slightly different frequencies. See approach in [https://doi.org/10.1109/JSTARS.2015.2432031].

- At Ka-band, there seem to be large phase differences at Nadir. It could be speculated that this could be caused by an wind-induced anisotropy in the x-y-plane (horizontal) rather than processes that act in the vertical direction (settling, metamorphism). However, a non-zero CPD at Nadir might also indicate a not well calibrated instrument. Please check.

504: "The dominant radar scattering surface": add: "at nadir and for both Ku- and Ka-band"

513-514: "provide contextual information for reliable interpretation": this is a very vague statement. Try to draw more specific conclusions. One point I see is: The fact that the dominant radar return for incidence angles > 15 degree, possibly even smaller, is not the air/snow interface anymore has an important impact for Altimeter-observations on snow-covered slopes. For such slopes, the surface height is likely to be underestimated. If you agree, please mention it in the discussion.

517: "which indicates that snow density and surface roughness contrasts (Figure 4) existing prior to ..." -> "which indicates that snow layers existing prior to ...": I don't understand why surface roughness is mentioned here. The manuscript shows no data on surface roughness; Especially not in Figure 4. What does the word contrasts refer to? I don't understand. Consider revising this sentence.

523: "The relatively small backscatter indicates ..." I think, the relatively small backscatter is a consequence from the specular reflections away from the sensor at non-nadir angles. It's rather the increasing delay (slant-range) to the observed echo that indicates that most scattering is associated with the snow/ice interface. See comment above (436-439).

524: what are "shallow theta_inc"? Incidence angles close to zero or close to 90°? 

524/525: What does "change" mean here? What changes under which conditions? 

535: "a two-scale function of the microscale surface roughness" what do you mean? What is a two-scale function? 

569/570: I doubt that "strong contributions from snow grain volume scattering at C-band" exists.

584: why would the anisotropy induce "scale-dependent" properties? I think scale-dependent should be removed here.

585: Why would the anisotropy alter surface and interface roughness? I would rather agree that snow metamorphism can alter interface roughness.

586-596: Revise according to the outcomes resulting from adressing comment 480-502.

599: "the first-ever recording of" a matter of taste if you need that or not. As publications should naturally contain new and original data all main results should be first-ever anyway. I would remove this statement.

603-604: Revise sentence to make it accurate. 

 - As observed in Figure 9, the air/snow interface can be hardly detectable with non-zero incidence angles. I even presume that the air/snow signal at 15° is a Nadir-return of the 12-17° beam-width (6 dB two-way beamwidth according to Strove 2020). As the beam-width does not have sharp edges it is certain that a Nadir-return can be observed.

 - "buried air/snow interface remains detectable" well, yes, this is shown in the nadir-looking data in Figure 8. However, this figure also shows that the intensity of the buried air/snow interface seems to be 10-20 dB below the intensity of the air/snow interface.

613: "strong spatial variability in backscatter": The figures 10 and 11 indicate temporal changes of less than 3 dB for WE1, and around 5 dB for WE2 with an decrease in Ka and an increase in Ku band. However as change and no absolute backscatter is shown one cannot speak about a strong spatial variability, rather than about a temporal variability of 3-5 dB at least for the second wind event.

614-617: Revise according to the outcomes resulting from adressing comment 480-502.

Technical comment: 

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

36/37: "changes in ...backscatter coincident with ... are observed": The sentence is correct, however, I had to read it a few times to make sure that really "backscatter coincident with..." is meant and not "backscatter coefficient". You might want to move the verb before coincident: "changes in ...backscatter (coefficients) are observed, coincident with ... "

39: detect -> detected

47: "Our results reveal the imprtance of wind, through its geophysical impact on Ka- and Ku-band (...) and has implications..." I'd suggest: "Our results reveal the impact of wind on Ka- and Ku-band (...) which has implications..."

68: I'd suggest: -> will result in the formation of heterogenities on different scales, from cm-scale ripple marks to snow bedforms ... 

131: from "nadir to theta_inc = 50°" maybe, from "theta_inc = 0 - 50°"

142: "denote d" -> "denoted \textit{d}" (italic d)

175 "across the theta_az range" -> "across theta_az" (here and other places: Even though linguistically correct, I would avoid using the word "range" together with "azimuth" to describe a span/interval/sector to avoid confusion with slant-range of the radar)

234: "are spaced every 2 cm" ->  "are placed every 2 cm" or more accurate: "are spaced by 2 cm".

256: increase and decrease in density and SSA -> increase in density and decrease in SSA

525: "dominant changing scattering mechanism" add: "at nadir".

527: "changed by more " -> "increased by more"