Please find attachment.

This is the review of the manuscript entitled “Surface-based observations of cold-air outbreak clouds during the COMBLE field campaign” written by Mages et al.

Generally, I found the study quite interesting and scientifically worth to publish. However, in quite many passages the authors stay vague with their statements. They leave open questions, which need to be resolved before one can actually trust the argumentation line of the authors. Respective passages in the text hint to deficiencies in the applied techniques, which either need to be addressed, discussed or negated. That’s the reason why I had to provide a rather long list (of partly short) major concerns. I need to see detailed answers in order to be convinced that the presented results and conclusions are indeed reasonable and defendable. I thus strongly recommend a second review round.

Major concerns:

L137: Please prove the statement “These updraft occurrences CLEARLY CORRELATE WELL with periods when the MWR detects the presence of columns with liquid water exceeding 0.25 kg m-2.” using statistical methods. Or do you mean that the visual inspection of the time series suggests a correlation?

L142-144: Wouldn't also the Doppler lidar data (doi: 10.5439/1178583 ) help to validate the retrieval? One could use the Doppler-lidar-derived vertical velocity at cloud base to obtain the motion of the high number of small cloud droplets, which is most-likely the air motion. These values could then be correlated to the Dopper velocity and reflectivity observations of KAZR.

L156: Why do different Z-V_D relationships apply to each day? What are the potential meteorological drivers?

L157-158: In lines 137-140 you motivate why you need to remove values exceeding 0.25 kg/m². Here, you did the test also for higher LWP values and it also works? How and Why?

L159-160: Shouldn't also the observed particle phase state play an important role in the selection of valid datapoints? ice/snow falls very different and produces very different reflectivities, compared to liquid water droplets.

L177: Can you provide a literature basis for the retrieval? It is similar to what was done by Li et al. in the group of D. Moissev?Li, H., Möhler, O., Petäjä, T., and Moisseev, D.: Two-year statistics of columnar-ice production in stratiform clouds over Hyytiälä, Finland: environmental conditions and the relevance to secondary ice production, Atmos. Chem. Phys., 21, 14671–14686, https://doi.org/10.5194/acp-21-14671-2021, 2021. 

L178-179: Where is the spectral energy density defined? You might just add this variable name to Line 100, where Ka-SACR is introduced and where you only write about 'Doppler spectra'.

L179-180: How comes this value of 0.28 m/s? Is turbulence always the same so that this correction is always the same? I would expect that different corrections are required, depending on the convective situation. How would a variable ‘correction value’ affect the overall retrieval?

L191-192: When clouds only extend over 0.5-3 km (this was mentioned earlier), how many 20-minutes samples could be derived given these short spatial scales?

L212-213: Which ‘typical’ studies used the ‘typical’ value of 2m/s to discriminate stratiform from convective situations?

L220-222: What makes the authors assume that the updraft regions travel with the horizontal wind? Can they rule out the presence of standing/rolling waves? The frequent appearance of cloud straits in cold-air outbreaks make we wonder whether the clouds actually travel parallel to the wind vector or if they are trapped in a roll-over vortex. Satellite images might help to demonstrate the validity of the assumption of the authors.

L232: How was cloud top derived? Are there uncertainties related to the derived cloud height values?

L233: To which height region does the statement about wind shear apply? Or was it the same in the whole troposphere (or at least at all height levels)? This is rather unlikely in convective situations.

L237: How can a cloud with top temperatures of below -40°C show no ice formation? Could it be that the cloud was snow-dominated, but melting was not detected due to the cold cloud base and low (below 0°C) dew point? In general, dew point needs to exceed 0°C in order to trigger melting of the particles. This is, e.g. the general assumption in the ACTRIS Cloudnet algorithms.

L244-245: Can Mie-scattering conditions be excluded? Or could it be that large cloud droplets or rain produced artificially high Brightness temperatures and consequently overestimated values of LWP?

L268-273: Really nice approach! Just wanted to point this out.

299-302: Why does v_air drop for high LWP values? Mie effects? Can Mie effects be negated?

Fig. 1: What is the saturation region of KAZR? i.e., what is the upper limit of detectable reflectivity? How would any saturation effects of the cloud radar affect the presented retrieval?

Minor comments:

L34-36: I suggest to provide some quantitative information. I.e, how large were the eddy dissipation rates? How was the evidence of secondary ice formation derived? How intense was it during COMBLE?

L97-102: KAZRs provide usually General Mode and Sensitive Mode. Which one was used? Or, even better - link to the utilized datasets in the ARM database. Same holds for the other instruments.

L104-106: How often were sondes launched?

L109: Further questions arise about the ceilometer dataset: (1) which of the three cloud bases was selected? (2) The doi links to many different CEIL datasets. It is possible to only link to the COMBLE dataset? (3) which type of ceilometer was operated?