Summary:

This manuscript is a followup to an earlier paper which described the theoretical basis of a combined radar+microwave/sub-mm radiometer retrieval of ice cloud profiles. The earlier paper was focused on the method and evaluation with simulated data, whereas the current paper is focused on in-situ validation with actual measurements. Cases with airborne and a combination of airborne and satellite remote sensing measurements are considered, with the former presenting more consistent results due to better co-location of measurements. 

General Comments:

This is a nice follow-up to the earlier paper on the theoretical method. The issues with getting precise co-location (particular with space and airborne obs) are under-appreciated and it’s good that this paper draws attention to them. The results are noteworthy and my takeaway is that single-frequency radar and radiometer combination (even with wideband channels)  is still under-constrained to retrieve two PSD parameters and determine an optimal particle shape. I have a few minor specific comments but I have found nothing problematic with the general formulation of the investigation or interpretation of results.

Specific Comments:

Line 87: Although it probably doesn’t matter much to the passive frequencies being simulated, it would make more sense to me to extend the retrieved hydrometeor content at the sixth bin above the surface downward rather than the reflectivity, since the reflectivity is likely not constant (due to attenuation).

Table 2: For the Dm corresponding to IWC, what is the meaning of the “A priori mean” value of “IWC = 10^-6”? Shouldn’t the a priori mean be in units of length (as with RWC)?

Typographical errors:

Line 59: should be “profiles of ice hydrometeor…”

Line 142: should be “water vapor is calculated”

Line 254: should be “shapes for which the best agreement”

Line 298: should be “Secondly, a clear backscattering signal…”

This study develops a new retrieval algorithm combining radar with passive millimeter and sub-millimeter observations for the estimates of ice hydrometeor distributions. The retrievals are validated using the observations from aircraft measurements, which provides confidence on the method. This work is very important for the satellite community and for the preparation of ICI. The paper is well written, and I enjoy reading it. I only have several minor questions/comments.

Lines 69-71. It may be helpful to add the time periods of these campaigns in the text or in the table.

Line 100. It seems that you do not have the same measurements to perform the retrievals for different flights. How is the missing information handled in the retrieval method?

Figure 10. The discrepancies are quite large for 243 GHz channel for flight C161 compared to the other two flights. Could you comment on that?

Lines 220-222. As you also mentioned, the largest uncertainties correspond to the higher-level clouds where we have smaller ice particles. Are these uncertainties also related to the lack of representativeness of the particle shape/type used in the models?

Summary                                          

This study presents the results of a synergic retrieval (radar + mw/sub-mm radiometer) of ice cloud hydrometeor vertically-resolver distributions. A previously published paper presented the technical details of the retrieval and tested it with simulation-based results. The present paper validates this retrieval using real co-located radar and microwave radiometer observations. This paper is very well written and organized. Results are divided mainly in two: an analysis of the accuracy of the retrieval by comparing the retrieved results with in-situ measurements of bulk IWC and PSDs, and an analysis of the validity/consistency of the forward model active and passive simulations (i.e., the RT framework). 

General comments

The publication is fit for publication as the results are novel and the limitations and complexities involved are analyzed with care and detail. The main results I take is that particle habits are still un-constrained even with sub-mm frequencies and this is an important result on the edge of the ICI era. I recommend publication with only a few minor details/comments to be considered. 

Minor comments and Typos

Line 69: Please include the time periods of these campaigns for completeness.

Line 107 and 108: Title 2.2 missing capital letter (In situ measurements). Also missing the capital letter in the first sentence of the paragraph below.

Line 109: For completeness I would recommend a better description of what it is meant by “high-level” and the “lower parts” of in situ-sampling. Perhaps it makes more sense to introduce in-situ measurements before the analysis of the co-location of flight tracks. 

Line 130: Background properties of the atmosphere and the surface […]

Line 132-133: Although readers are referred to Pfreundschuh et al. (2020) for a detailed description of the retrieval, perhaps a better synergy between the text and table 2 would add to the description here regarding the parameters of the PSD for the different species. 

Line 136: [...] Atmospheric Radiative Transfer Simulator (ARTS, Buehler et al., 2018) is used [...]

Line 145: “In the first one, the bulk properties”. Use lower case in “in the first one [...]” 

Line 151: After the sentence that starts with “The updated values of […]”, for completeness perhaps the cloud ice PSD equation could be included in parentheses with the Dm, N0* and the shape parameter. 

Line 152: single particle optical properties. 

Line 154: “Since this is difficult”. Please expand on this for completeness. 

Line 175: retrieved hydrometeor size distributions. 

Line 224: for which the best agreement. 

Line 286: typo in “For, flight B984 [...]”. Move the comma please. 

Line 298: typo in “Secondly, the a clear backscattering”

Line 349: Just comment here. I wonder if for a real scenario, with the complexities of the vertical differences in particle orientation, habit phase, etc, this would add to 20% in the resultant observations at nadir. 

Figure 10. The discussion mentions the large differentes for 243 GHz, however flight B984 shows smaller residuals at this channel than the other two channels. Is there anything to add to the discussion regarding this? Also, there are large differentes for 448 +- 7.2 GHz, specially for flight C159. Could you comment on that too?       

Figure A1: Have you looked at a similar figure for the other channels?            

Figure 14: How do these results translate to IWP? Perhaps a general summary of such a Figure could add to the discussion of Figure 14.                

Figure 15: I have a tough time with the gray of in-situ (sample) 

Figure 14 and Figure 15. Results are presented for flight B984, which uses the HAMP MIRA 35 GHz cloud radar. Could you comment on possible differences with CloudSat CPR were to be used.