This manuscript describes an approach for triple frequency radar retrievals of precipitating ice microphysical parameters. It is overall an interesting study worth of publishing after the authors address comments below.

Comments.

1. Equation (2): m-D relations. Usually, size of particles is defined in terms their major dimension. Since you define it here differently, it would be useful if you briefly discuss what typical differences can be expected between your size definition and that which uses the major dimension. You are suggesting on line 68 that your relation agrees with the one from Leinonen and Szyrmer (2015) but these authors (unlike you) use maximum size for D, so discussion about differences in size definition would be helpful. Also, it appears that your relation m=0.015D^2.05 [SI units] provides particle mass values, which are quite a bit different than many existing relations for aggregates (see for example, Mitchell JAS 1996, p.1716, relations from Heymsfield et al. JAS 2010, p.3303 and many others). Again, Fig. 1 shows mass dependence on Dmax not the size you use according the statement on line 61. Can you address these issues?
2. Do you account for particle orientations and shapes? Observations show that DWR depends quite strongly on particle shape and orientation (e.g., Matrosov et al. JAMC 2019 p. 2005). For vertical beam measurements, more spherical particles produce larger DWR than less spherical particles.
3. I think the NRC aircraft can also provide measurements with side view radar beams. Did you compare side and vertical measurements?
4. From what I know, the NRC aircraft microphysical suite has two Nevzorov probes (at least it was the case with the flights I know about). The IWC estimates form the two probes can differ. Did the flights, which you analyzed, have measurement from two Nevzorov probes? If yes, what were the differences?
5. Did you try to calculate X, Ka, and W reflectivities using your scattering data base and compare them to the radar observed values?
6. Please clarify in more details how the “truth” in Fig. 4c was obtained.
7. Line 229: what are the uncertainties of estimating Doppler velocity from a moving aircraft?

Editorial comments:

1. Line 10 and elsewhere: provide units for Dm and IWC to better understand RMSE values here and statistical metrics results given in terms of logarithmical values.
2. Equation (1): provide integration limits.
3. Line 270: If only one frequency (W) has non-Rayleigh scattering it is already dual-frequency not triple-frequency approach.

This manuscript describes microphysical retrievals using multi-frequency radar. Retrievals of ice water content, mean mass diameter, and degree of riming are demonstrated. The manuscript shows the difficulty in retrieving the degree of riming, particularly from a single-wavelength. The authors conclude by demonstrating the retrievals from single, multi-frequency and multi-frequency with Doppler velocity in a research flight. The single frequency has little information content in the degree of riming or in volumes with large drops. The multi-frequency retrievals improve upon the single frequency, but still struggle with degree of riming. By adding a Doppler velocity to multi-frequency, the retrievals are in better agreement with the measured quantities.

Overall this manuscript is straightforward and shows interesting results for a difficult problem. I have a few main comments as well as some smaller things I noticed.

• It seems one of the main take-aways the authors would like the reader to have is the improvement to the microphysical retrievals, particularly the degree of riming, when a Doppler velocity is included. However, it was unclear to me how this is used in the retrieval and seems to come out of nowhere when Fig. 4 is introduced. Is it from nadir or zenith (or some interpolated mean like reflectivity)? Does the Doppler velocity include air motions? How is it included in the actual retrieval?
• Figure 2 is important but I had a hard time interpreting it. What is the difference between the top row and the bottom row? Please clarify.
• Figure 3: I’m confused about how the uncertainties are presented in this figure. If I am interpreting this correctly the retrieval of degree of riming has very little uncertainty middle of the flight leg (~1950 – 2110 UTC) (i.e. no black bars), and similarly the measured IWC in panel b (very small green bars)?

Minor comments:

Ln 98: Remove or expand ‘(REF)’.

Ln 145: Which WC is this – IWC or LWC?

Section 3: It is odd to have “3.0.1” etc. for the sections—remove the .0. and make 3.1, etc.

Figure 3: I’m curious why the IWC are presented in kg/m3? In the text this is stated in g/m3 (such as the uncertainty in ln 210). Similar comment for all plots of IWC.

Figure 4: panel b is misaligned compared to the other 2 panels.

Ln 241: Please check the sentence beginning “For negligible DWRs, multi-frequency information is reduced so the difference between the algorithms.” This is not a complete sentence.

Ln 255: “estimates is available” should be “estimates are available”.

This manuscript describes a snow microphysical property retrieval algorithm that employs multi-frequency radar simulations.    A novel aspect of this study is that no a priori particle size distribution (PSD) parametrization is used as part of the retrieval scheme.   Instead, direct airborne PSD measurements, combined with state-of-the-art ice scattering models, are used in forward three-frequency radar simulations to retrieve microphysical properties that are then compared to independent and concurrent airborne microphysical observations.   Two key findings that are not entirely unexpected, but still extremely valuable as quantifiable evidence for the community's benefit, are that multi-frequency radar and Doppler velocity measurements are key observables needed to produce optimal snow microphysical property retrievals. 

Despite being limited to one case study, this study is an extremely useful addition to the literature as a proof-of-concept study that will provide useful guidance on future sensor development to ultimately produce more accurate snow property retrievals.   The snowfall remote sensing community will benefit from lessons learned in this study.   I find the manuscript written in a succinct and easily understandable fashion, yet provides sufficient analytical heft that conveys valuable results.  I encourage its eventual publication after the following minor comments are considered by the authors.

1. Line 48:   Should a different dielectric factor of liquid water be applied to the W-band radar reflectivity forward model simulations?   This is a very basic methodological question, but causes much consternation among researchers applying or modeling radar simulations.  The fact that the authors state that the 0.93 value is appropriate for “standard temperatures and frequencies below the Ka-band” might cause some confusion as to why this value is not altered for W-band simulations.

2. Line 52: Minimizing W-band attenuation complications is another novel aspect of this study.   The authors rightly highlight that W-band attenuation must be considered at longer distances from the radar under many circumstances, but the fact that these simulations are created using microphysical observations allows the authors to simplify the proof-of-concept message in the study.  

3. Figure 1 caption: I suggest adding the explicit year of the Morrison and Grabowski reference to the caption for completeness.

4. Lines 56-59 elicit a general methodological question: over what time span are the binned PSD observations aggregated?   I cannot offer a strong opinion of the optimal time sampling needed to produce robust binned PSDs, but it would be good to advertise this value to the community.   I presume PSD variability over short time scales is deemed somewhat muted for this stratiform event, but I would still appreciate the authors advertising the time scale used for PSD measurements that are utilised in the forward radar reflectivity model.

5. Line 60: Is the snowflake size automatically measured by analysis software?   If so, do appropriate references exist for this procedure?

6. Line 84: Minor typographical error.  Change to “density of ice particles”

7. Line 98: It appears as if a reference is missing (REF).

8. Lines 98-110: This seems like a reasonable and creative method to deal with the complexity of possible snowflake morphology and atmospheric conditions. 

9. Figure 2 caption: I suggest explicitly writing that the top row are expected values and the bottom row standard deviations of the quantities.

10. Figure 3: I suggest enlarging the font contained in various figure legends.   The values are very difficult to read.

11. Figure 3: Panels d, e, and f show reflectivity observations for each radar frequency.  Might it be better to show DWR values instead since DWR is explicitly shown in Fig. 2?  Or somehow creatively combine DWR with the single frequency values shown?   This is not a mandatory suggestion by any means, but I am left wondering if showing DWR observations might also be beneficial to better connect with meaningful information contained in the observations.