Main comments on WCS-2021-53

The paper is mainly focused on the development of a mesocyclone detection algorithm based primarily on radar data (plus COSMOS model data). The method is then applied to a 5-year data set for Switzerland, climatological aspects are investigated, and influencing variables such as large-scale flow, diurnal variation, and orography are considered. In addition, some relationship is made to one past hail study.

Basically I enjoyed the paper, I think it is informative and well done, and it should be made available to the community. Especially in the first part (development of the algorithm/climatological consideration), I had primarily minor comments (but in sum quite many). However, in the second part I have some deeper comments, so the paper needs to be reviewed again (Major revision).

My major points are:

• Minor major point, L61-66:
  
  The purpose and objectives of the paper should be better formulated in the introduction. The authors mainly describe the methodological approach. It would be more interesting to formulate the question about the physics (e.g. question about influence orography, synoptic...). What scientific questions should be addressed by the obtained data set?
• Minor major point, L75:
  
  The description of the COSMO data is inadequate. Which time period is used, which variable exactly? Which area? Resolution? Data generated by yourself, or from others?
• In some places it needs to be described more clearly how exactly the sample set is composed (see comments in the text; e.g. one track vs. all detection time steps).
• Section 4.1 to 4.3:
  
  Please put the results more in context with other work in Switzerland and similar studies in Europe; compare it (not only with Nisi et a., 2018; see PDF for more details / suggestion).
• Figure 5:
  • Please rethink the colorbar; NW, N, NE, SE is hard to see and distinguish (also think of potential Red-green color blindness).
  • Suggestion: As a supplement you could make very well a wind rose in dependence on the frequency. Maybe you can split the analyses over the 7 months to clarify if there are time dependencies?
• Section 4.2 and Figure 6b:
  
  I would be more cautious with your interpretation. Overall, Figure 6b shows a large variability. Your interpretation of 4 UTC sample should also take into account that there are very few data (8 or 9 cases) and you should be careful not to interpret too much into your data.
  
  Furthermore, which time regarding the track is the basis of your evaluation? Start time? The time in the middle of the track? All this can have an influence. When interpreting the figure, it is important to consider how the duration of the individual tracks is.

• Figure 7:
  
  Maybe it is interesting to split Figure 7 regionally? 1x north and 1x south of the alpine ridge?
  
  Perhaps your statement "the majority of rotating storms move uphill" is purely coincidental and has no real systematic background, as your statement can be attributed only to the region and associated synoptic flow direction.
• The authors should better explain how the thresholds in Table A1 were chosen.
  
  Are there any comparable values in the literature? Have sensitivity studies been performed with varying thresholds? How sensitive are the results due to the choice of thresholds?

Minor comments:

• In the attached annotated PDF-file, you will find several further minor comments, suggestions and reference suggestions.

Review

Manuscript Number: wcd-2021-53

Title: A Characterisation of Alpine Mesoscale Mesocyclone Occurrence

Recommendation: Major revisions required.

General comments:

Overall, the manuscript is of interest to the science community and within the scope of Weather and Climate Dynamics. The scientific approach is mainly good. However, the results are not discussed thoroughly and more relevant references are needed. The presentation is mainly good. The figure captions are often to short and not precise;  important information that is needed to understand and interpret the analysis is “hidden” in the rather long appendix.

Specific comments:

• The discussion of the influence of the radar quality index on the derived spatial distribution of (mesocyclonic) storms is not very clear. E.g. What does a quality index of 0.5 mean? Should the results in this area be considered correct? Or underestimated?

• Meteorological expectations, derived distributions, limits of the radar network as well as assumed correlations between mesocyclones and hail are often mixed or used to support findings or assumptions in a rather ambiguous way.

• In the discussion it says “The accompanying relative quality index map helps interpret, where the mesocyclone detection is impeded by the physical limitations of the radar network.” However, this information is not thoroughly used in the presentation and discussion of the results.

• The discussion of the “synoptic weather situation” is rather short. This could be elaborated. Also, it should be clearly stated which sentences are hypotheses and which are based on analyses (references).

Technical corrections:

L. 62: Including these place names in a map would help those readers not familiar with the study domain.

L. 88: 1 Apr to 31 Oct

L. 123: distance from the radar.

Fig. 1 Caption: Is this the quality index only for mesocyclone detection or for all radar derived products?

L. 151: Typically, negative velocities indicate a component of the wind towards the radar and positive values indicate a component away from the radar. Please clarify.

Fig. 3 Caption: Individual mesocyclone detections (there could be detections in successive radar scans – thus multiple detections per storm) or mesocyclone storm tracks per km²? Wording should be clear and consistent throughout the manuscript.

Fig. 5 Caption: Spatial distribution of detected mesocyclones classified by the synoptic flow. The colours denote the prevailing wind at …. level. The percentages given in the box ….

All maps: Could you use latitude / longitude instead of “swiss x-/y-coordinates”?

L. 224-225: Please explain.

L. 239: How are storm area and track length defined?

L. 245: The influence of the low number of cases on the results should be discussed.

L. 255: What is the influence of the radar data quality on the mesocyclone detections at high altitudes and its intensity estimation?

Fig. 7: Use a boxplot instead. Also provide the number of cases per selected altitude bin.