Summary:

This paper focuses on an inter-comparison between polarized lidar data in New Zealand, Chile and Germany. The different locations of these lidar systems allows the authors to examine the impact of different aerosol environs on heterogeneous ice formation. The relationship between aerosol and ice formation is a complex one and an extremely high priority for improving model simulations. The data underlying this paper provides a unique opportunity to separate out the influences of different aerosol loads and allows the authors to extend existing research on heterogeneous ice formation. I think this paper is a valuable contribution to the scientific literature and hope that future work will be done to extend this line of research.

Minor comments:

Some of the paragraphs are excessively long, particularly in the introduction and conclusions, which serves to obscure some of the points of the paper. Splitting some of these paragraphs would lead to an improved experience for future readers.

The sentence starting on line 28 (“Even though.....”) is confusing to read. I would appreciate it if this sentence could be rewritten for clarity.

The two sentences starting on line 54 (“Too few.... . Via too strong....) are separated in a way which muddles the meaning. Once again, I think this should be rewritten for clarity.

Line 96 change “are” to “have been”

The sentence starting on line 118 (“By considering.....”) is confusing to read due to the complex clausal structure. This should be rewritten for simplicity.

For completeness, a reference should be included after the statement that dew point spreads of 2k are needed for cloud formation (line 217)

This discussion of figure 10 is a little underdeveloped (lines 298 - 301). I’m unsure if an unfamiliar reader would be able to interpret the results of figure 10 based on the existing discussion. I also think additional discussion here is warranted, given the results of the figure.

General comments

In this paper, the efficiency of ice formation in clouds over New Zealand has been investigated, focusing on the effect of the different aerosol loads and their origin above the area of study. The study is focused in clouds for which the top temperature falls within the heterogeneous freezing range (-40^o to 0^o C) making use of a combined dataset comprising of ground-based lidar observations for detection of liquid and ice-containing clouds, radiosondes and Global Data Assimilation System (GDAS) for profiles of atmospheric parameters, CAMS-MACC model runs for aerosol re-analyses data, HYSPLIT model runs for air-masses backward trajectories, and TRACE profiles for air-mass source attribution analysis. A case study is selected to demonstrate how the different type datasets are/can be used to study the relationship between the mixed-phase cloud formation and the aerosol load and type, before presenting the overall results and statistics from the analysis on the clouds and the air-masses using the available long-term datasets. Overall, the manuscript is well-structured, well-written, and its scientific significance makes it suitable for publication, after some minor revisions to be kindly considered from the authors.

Specific comments

• Section 2.1: I think a brief description of the lidar specifications (e.g. laser energy and repetition rate, telescope diameter, FOV, spatial resolution, measured optical products) could be added here.
• Lines 103-104: I kindly suggest to name also the lidar products that are used in the study (backscatter coef., depolarization ratio) instead of the lidar signals.
• Lines 110-111: The authors state that the volume depolarization ratio is used along with the backscatter coef. to identify the cloud base and top, but in line 160 and Fig. 1 the particle depolarization ratio is used instead in the visual inspection for clouds in the scene. Which product is really used? If both, then please clarify.
• Lines 113-114: Do the authors use a threshold in the volume depolarization ratio values to discriminate the clouds (liquid or ice-containing) from other depolarizing particles (e.g. dust)? Do the authors account for the contribution of molecular depolarization in the volume depolarization ratio? Maybe a more detailed description is needed here.
• Line 165 “heights up to above 14 km”: It is not clear if these elevated layers were observed during the case study. If yes, what is the height range that the elevated smoke layers are detected, up to 14 km or above 14 km? And, what is the base of the elevated smoke layer?
• Section 3.3: It is not so clear that only the WD (well-defined) clouds are used in the cloud statistics and also in the following section (sec. 3.4 and Fig. 9 and 10) for the separation of clouds based on airmass statistics. Please clarify which sample of clouds (total or WD) the authors use
• Figure 9b: Kindly consider to use a different line color for clusters 3&4, since lightblue might be challenging for some readers to discriminate it from the cluster 2 blue line.

Technical corrections

• Line 307 “the Seifert et al. (2010, 2015), methods to asses …”: is the coma after the parenthesis necessary?
• Figure 8: duplicated for in “Fraction of ice-containing clouds as function of cloud-top temperature in intervals of 5 K for for Leipzig”
• Line 295: typo in “…ice formation efficiency, albeit within statistical uncertainty.”?
• Section 4: Since the full “New Zealand” is used throughout the manuscript I would suggest to use it also here and avoid the abbreviation NZ (e.g. lines 304, 306, 310)