The paper demonstrates an expansion of the radar-lidar synergy retrieval product varpy, the algorithm underlying the successful DARDAR-CLOUD and related CloudSat-CALIPSO retrieval products, to include supercooled liquid and mixed-phase clouds. The principle is that in clouds identified as mixed-phase by the DARDAR-MASK synergistic radar-lidar target classification produce, the radar observations are used to constrain the retrieval of ice, and the lidar measurements are used to constrain the retrieval of liquid cloud. The retrieval is demonstrated using a single 10 minute case study of high-latitude mixed-phase boundary layer clouds observed by a coordinated aircraft underflight of the A-Train during the ASTER campaign. 

This is an important issue in spaceborne detection and retrieval of clouds, especially relating to cloud-radiation interactions, and an expansion of the family of DARDAR products is welcome, as is any chance to evaluate spaceborne retrievals with field campaigns.

Overall the paper is well-written, the algorithm is thoroughly described, and the plots are produced to a good standard and are easy to understand. Some of the limitations of this study are discussed by the authors; however, there could be a clearer distinction between the limitations of the retrieval and those of the evaluation using the field campaign data, and both could do with expanded discussion to consider their implications, such as for application to a DARDAR product. 

I recommend this paper for major revisions, subject to addressing the following comments.

Major comments:

• What are the macrophysical implications of the in-situ profile demonstrating that neither the radar nor the lidar are sampling the entire cloud? Is there an indication from the in-situ data whether the cloud is mixed-phase through its entire depth, or does it transition to ice cloud? This is obviously a wide-ranging issue relating to all spaceborne radar-lidar observations, but it seems there’s a bit of information here from the in-situ measurements that’s worth commenting on.
• Further to this, what are the implications for DARDAR-CLOUD’s existing ice retrievals in mixed-phase cloud (Table 5)? I’m not asking for a change of scope of the present paper; however, it would be interesting to compare varpy-mix and varpy-ice in the present case study to explore the effect of interpreting lidar extinction in these two very different ways.
• In the purple part of the case study, where varpy-mix ice extinction is much higher than that measured in-situ but the IWC is about right, does this imply a significant bias in the retrieved ice effective radius? 
• The evaluation of the retrieval is summarized using the mean absolute and mean percentage errors, indicating impressive performance in most cases; however, more information is available here that could be expanded upon, especially relating to the bias. In addition to the table, is it possible to use a plot that shows either the PDF of the retrieval compared to the in-situ data, or the distribution of errors, to retain more of this information. If the liquid water content retrieval is biased low, as appears to be the case here, is this a systematic bias or could we expect a different result in other case studies? Please comment on these uncertainties.
• There is some ambiguity about the mass-size relation that is used. It is clear from the method section and conclusion that both the Heymsfield and Brown & Francis relations are implemented as LUTs in VarPy, and the coefficients for both are given. In the algorithm description we are told “For both VarPy-ice and -mix, both LUT are used to retrieve the ice properties” (L156-7) but in the result section the results are shown for the Heymsfield mass-size relation only (L360-2). Does the selection of the Heymsfield relation indicate that these results are better than those using Brown & Francis, and can the authors (or the reader) conclude anything from this? Is there any value in showing both results? Are there any cases, e.g. in the purple section where a large error in the retrieved effective radius is implied, where the other mass-size relation may improve the performance? It is a strength of the VarPy algorithm to be able to answer these questions, so please expand on this or resolve the ambiguity.

Minor comments:

• Fig 2 (a) there is some contouring around noisy features in the lidar backscatter that make this panel difficult to interpret
• There are minor differences between the figures that surprise and confuse when comparing features between plots: 
• In Fig. 6 the colorbar labels face the other way to those on all other figures.
• In Fig. 6 the format of the time and latitude labels, and their positions, are different from Fig. 5. This makes it difficult to line up the features visually.

The presented work addresses a significant gap in current climate and weather forecasting models by proposing a novel method for retrieving microphysical properties of mixed-phase clouds. Recognizing the limited representation of these clouds due to the complexity of underlying processes, the study leverages airborne or satellite radar and lidar measurements within an extended variational method. The key innovation lies in the simultaneous retrieval of ice and supercooled water properties, achieved by exploiting the distinct sensitivities of lidar and radar signals. The method is rigorously developed, considering the sensitivity of these instruments and relying on a robust a priori framework. Validation using DARDAR-MASK products and airborne in situ measurements demonstrates promising results, showcasing the method's potential to enhance our understanding of mixed-phase cloud processes. The work is well-structured, providing clear insights into the motivation, methodology, and validation approach, and it contributes significantly to the broader goal of improving climate and weather prediction models. Overall, the research exhibits a commendable quality, combining scientific rigor with practical applicability.

The abstract effectively introduces the paper's focus on addressing the challenges in representing mixed-phase clouds in climate and weather forecasting models. The proposed method for retrieving microphysical properties of mixed-phase clouds is clearly outlined. The use of airborne or satellite radar and lidar measurements, along with a variational method, enhances the understanding of cloud processes. The abstract sets a comprehensive context for the reader, emphasizing the importance of accurate representation of mixed-phase clouds in models. Section 2 is well-written and provides a comprehensive explanation of the methodology. In section 3 the introduction of the ASTAR campaign and the specific context of the study is clear and sets the stage effectively. The integration of in situ measurements and remote sensing data, especially the use of the Polar Nephelometer as a reference, enhances the robustness of the comparison. Section 4 appropriately acknowledges the limitations and challenges of the VarPy-mix methodology, such as missing data in the lower part of the cloud, spatial and temporal shifts, and biases in the retrieval process. This transparency enhances the scientific rigor of the study.

Specific Comments:

The abstract is well-structured, progressing logically from the motivation to the proposed method and concluding with the validation approach. The use of technical terms is appropriate, but consider providing brief explanations for readers less familiar with terms like "attenuated lidar backscatter β" to enhance accessibility.

Section 2: In some sentences, the detailed technical content and the complex structure might pose a challenge for readers unfamiliar with the specific terminology or mathematical formulations. Consider breaking down a few of the more complex sentences for better readability without sacrificing technical accuracy.

The scientific rigor of the proposed method is established in the abstract by clearly outlining the underlying assumptions and the methodology. It effectively communicates the dual sensitivity of lidar and radar in retrieving supercooled water and ice crystals, respectively.

Towards the end of Section 2, you might consider adding a brief summary or concluding paragraph that synthesizes the key aspects of the methodology. The connection between observations (Y) and the state vector (X) is effectively established.

The validation approach using DARDAR-MASK products and airborne in situ measurements is a strong point, showcasing the method's applicability and performance. Mentioning the mean percent error provides a quantitative measure of the retrieval performance, enhancing the credibility of the results.

The adaptation of the cloud phase classification for mixed-phase retrieval is well-motivated and aligns with the goal of correctly identifying each phase of the cloud.

In general, section 3 contains a significant amount of technical detail, which is appropriate for a scientific paper, but ensure that the language remains accessible to the target audience.

Section 4 effectively summarizes the VarPy-mix methodology, discusses its strengths and limitations, and lays the groundwork for further discussion and conclusions. I would suggest to provide more context on the RALI-THINICE campaign and the HALO-(AC)3 campaign, as these are mentioned but not explained in detail.

Lines 23-25: To gain a clearer understanding, could you provide more specific information on the largest differences observed in both time and space between remote sensing and in situ measurements? Is there a quantitative measure or range that characterizes these differences or any strategies employed to address or mitigate these differences in the analysis?"

Lines 72-73: I would recommend to introduce the VarPy-mix method earlier in the paper, possibly in the abstract or introduction, to provide readers with a clear understanding of the methodology from the outset. This can enhance the overall coherence and comprehension of the research, ensuring that the significance of VarPy-mix is highlighted prominently.

Technical Corrections:

Consistency in Terminology:

Ensure consistency in the use of terms such as "ice crystals" and "ice particles" for clarity.

Abbreviations:

Consider introducing abbreviations like IWC, LWC, and Ntot upon their first use in the abstract.

Line 78: “Additionally, this flexible algorithm can be apply on several radar-lidar platforms,”

The review is provided in the attached document.