Unfortunately,  I cannot recommend it for publication in ACP in its present form. Feel free contacting me if you have any questions, or need more clarifications regarding the comments.

Best regards,

Alexei

The purpose of this study is to evaluate varying cloud properties during the evolution of mixed phase clouds (i.e., from the first appearance of ice within supercooled liquid clouds to their complete glaciation). It tests for this by determining the ratio of liquid, mixed, and ice phase samples within a continuous cloud sample. Some important findings are revealed by the analysis, such as the vertical air motion is much more variable at the first appearance of ice at temperatures less than -20°C compared with other parts of the evolution, suggesting dynamical factors may be a significant factor for ice initiation.

The research topic is very important, as there are still large uncertainties associated with the evolution of mixed phase properties; and the novel approach qualitatively determining the stage of macroscale mixed phase evolution could be valid for a statistical analysis as undertaken in the study (assuming that once ice occurs within a supercooled liquid cloud, it will tend towards complete glaciation). This stage classification is often combined with the spatial extent of mixed and ice phase samples (ice spatial ratio). Findings can vary from quite insightful and perhaps striking (e.g., similar rates of increase in ice water content within mixed phase samples and also ice phase samples with increasing ice spatial ratio), to rather speculative (e.g., a key step for the Wegener-Bergeron-Findeison process to occur is when pure ice segments are present).

Aside from the analyses performed, there are major concerns associated with their methodology listed below:

1)    Specific combinations of LCR, MCR, and ICR ratios may be rather ambiguous. For example, are we sure a phase 2 cloud segment with >97.5% MCR is still considered to be in the earlier stage of glaciation than a phase 3 cloud segment with <2.5% ICR and >97.5% LCR? There is the potential for extreme ambiguity in what part of the mixed phase evolution a cloud region is currently in based on the current framework.

2)    This study does not appear to account for whether the aircraft is sampling within the cloud or the precipitation underlying a given cloud. So a primarily liquid phase cloud (with few mixed phase samples) could be precipitating ice and if the aircraft samples the precipitation, it will be considered phase 3, although the cloud itself would be phase 2. Not to mention, the aircraft could have a majority of cloud samples/an entire length of continuous cloud sampling as precipitation. Including precipitation likely accounts for subsaturated conditions for both liquid and ice phase samples in Figures 10 and 11.

Minor concern:

1)    The lengths of the clouds used in the mixed phase evolution appear (note: the lengths are not clearly defined in the text) to not be set to a constant length, so some lengths could vary from two or three cloud samples (are single 1 Hz cloud samples without neighboring cloud samples removed?) to hundreds of samples. This means processes associated with different length scales will have different impacts on different cloud lengths and may not result in an apples-to-apples comparison. 

I think a good start for addressing these issues is to take a good look at the individual total cloud regions: the distribution of their lengths as well as their phase ratios and how they look for the individual “phases of mixed phase evolution.”

There are also other major concerns aside from the mixed phase evolution methodology:

1)    The authors use the UHSAS probe to discern aerosol concentrations within the clouds. However, it appears as though aerosol measurements are taken within the cloud samples. This is problematic as there is a high likelihood the probe is sampling the residuals of cloud particles, and I suspect the positive correlation of aerosols with diameters greater than 500 nm within greater ICR is the UHSAS sampling cloud particle residuals. In order to use the UHSAS in the cloud, it is vital to provide sensitivity tests to confirm no such biases are occurring for both liquid and ice particles.

2)    Phase frequencies are determined using a phase definition relating the phase fraction of liquid to the total number of particles using the UWLID product. However, it was stated that the 2DS has a size range of 40-5000um. Assuming the UWLID product provides phase information for this particle size range (does it? If so, the smaller particle sizes [<~200 um] should be associated with large uncertainties due to the loss of the 2DS imaging resolution), what about particles less than 40 um? It doesn’t seem like the CDP measurements are incorporated into the analysis, which could significantly impact number concentration ratios. 

There were quite a few grammatical errors as well, and multiple citations were in error. And although there are some interesting findings presented in this study, I unfortunately cannot recommend that this manuscript be accepted for publication. I recommend doing a thorough revision of the paper and resubmitting.