Review of Wugofski and Kumjian, AMTD 2025 (amt-2025-672)

 General comments to the manuscript https://doi.org/10.5194/egusphere-2025-672

 In the study titled “Detection of Multi-Model Doppler Spectra. Part 2: Evaluation of the Detection Algorithm and Exploring Characteristics of Multi-Modal Spectra Using a Long-term Dataset” by S. Wugofski and M. Kumjian, the multi-modal radar Doppler spectrum detection algorithm developed in Part 1 was applied to a 3-yr-dataset of Ka-band vertically-pointing Doppler cloud radar observations at the North Slope of Alaska. The authors set the found multi-modal layers into context with height, depth, and temperature of these layers and used ERA-5 temperature data to assess which microphysical process could be responsible for the multi-modal events. Statistical results related to multi-modal layer detection as well as three detailed case studies are presented. The study closes with a short section on conclusions. 

Recommendation:

I would suggest the manuscript to be published after major revisions considering the remarks below. The authors should address the following points:

General/Major comments:

While there is an advantage of only relying on data from one instrument, KAZRs are usually not operated on a stand-alone basis but deployed with other instrumentation and the manuscript would greatly benefit from using these to substantiate the conclusions. – Currently, the potential microphysical processes leading to the observed multi-modal radar Doppler spectra like primary ice production, growth of column or plates, and Hallett-Mossop rime splintering are here only assessed based on ERA-5 temperature data instead of making use of the extensive ARM instrumentation suite (MWR, depolarization lidar, etc.). Adding observational evidence from other instrumentation would be very beneficial in interpretation of the temperature-only based hypothesis of microphysical processes. For example, to substantiate the hypothesis of potential occurrence of Hallett-Mossop rime-splintering, variables like liquid water path as derived from microwave radiometer observations help to see if liquid water was available for riming as prerequisite for rime-splintering. The points made in the manuscript would be much stronger if additional information from other instruments are added.  

Line 289 - 290: In line 289 it is correctly stated that at T between -3 and -8°C both, primary ice nucleation of columnar ice crystals or Hallett-Mossop-rime splintering can occur. Please include the possibility of primary ice nucleation in this T-range in the abstract and in line 299 + line 314 etc. as well instead of only restricting to Hallett-Mossop-rime splintering. Furthermore, Section 5.1 should also be labeled accordingly.

Line 299, 444 and elsewhere: Dendritic growth temperature zone employed here seems very narrow, often -20 to -10 °C are used instead -12 to -8°C, see e.g. https://acp.copernicus.org/articles/22/11795/2022/ and references therein.

In line 233 dendritic growth zone temperature regime is reported as -18 to -12°C – use consistent T-ranges throughout the manuscript.

 Section 3.1 (Case Verification) I think the study would benefit from adding two additional parameters “distance of multi-modal layer top from cloud top” (to see where the multimodalities occur with respect to cloud top) as well as cloud depth.

Line 190-194: Please explain why you choose to set monthly flag count into context with ERA-5 monthly thermodynamic, kinematic, and microphysical variables – Fig.4 shows that high flag counts are mostly related to single continuous events. Why not analyze the thermodynamic, kinematic, and microphysical variables for those events instead? Or instead contrast ERA-5 variables for flagged events vs. non-flagged cloudy periods? I struggle seeing the benefit of the correlation of the monthly multi-modal flag count with ERA-5 variables as it is presented unless it is set into context with existing literature, e.g. on seasonal mixed-phase cloud occurrence at the NSA site etc.

Minor comments:

Throughout the text, it would be helpful to always state which “flag” (Multi-modal flag/LDR flag) is meant to avoid confusion (e.g. Line 444f “the algorithm detected flags that aligned with the multi-modal layers”.)

Line 10: please check the sentence, what does the word “verifying” refer to?

Line 16 and elsewhere: Clearly state where you refer to spectral or integrated LDR

Line 55: How does the manual verification work?

Line 76: replace “being a secondary mode” with “containing a secondary mode”

Line 80: Motivate why you choose two hours as minimum threshold for case identification? – Depending on cloud type and cloud lifetime, microphysical processes with pronounced multi-modal Doppler spectra occur on much shorter time scales.

Line 88: Explain how this consolidation of cases is done. Manually?

Line 230-231: Please give references for the statement that “primary mode” refers to the faster falling one and “secondary mode” to the slower falling one. In KAZR terms, I think primary mode is the one with the highest reflectivity independent of fall velocity.

Line 234: Rephrase: “The distinction between liquid droplets and columnar ice/needles” …

Line 282: add “columnar” before ice crystals

Line 352: replace “profile” with “atmosphere”

Line 431: Explain why would you limit the algorithm to vertically-pointing Ka-band radar observations and don’t propose to also use it for other radar bands (X-/W-/G-band)?

Line 434: add “columnar” before ice

Line 448: As stated above, I strongly suggest including additional remote-sensing observations existing at ARM KAZR sites

Comments on Figures:

Fig. 1: Please add time-height panels of KAZR reflectivity, mean Doppler velocity, and spectrum width to contextualize the two case studies for which the flag-criteria depicted are matched. In this way the readers can also see if vertical spectrograms are applicable or if Doppler spectrum evolution seems rather plausible along slanted fall streaks.

Fig 1: Is Oct 31 2023 (title) or 2024 (caption) shown?

Section 5: Add time-height plots of KAZR reflectivity, MDV, spectrum width and LDR for all three presented case studies to contextualize the presented height spectrograms.

Fig 10: What are the horizonal lines in panel d, e, f? (also in Fig. 12)

Fig 10: extend your x-axis to lower velocities to capture the entire observed Doppler spectra

This study applies a recently-developed algorithm from Part I of this series to a multi-year record of radar observations for identifying multimodal Doppler spectra of precipitation echoes from a vertically-pointing radar to evaluate its performance. Case studies help to interpret the meaning of bulk results and are an important component of the study. Ultimately, I find the study to be generally well-reasoned and constructed. It contains a worthwhile blend of analyses and would benefit from a handful of revisions which are suggested below. Most of my comments are minor in nature, but at least one may require a decent amount of revision to the analysis so I've flagged it as major. 

General Comments

The standard abbreviation of the 5th generation ECMWF reanalysis is "ERA5". Thus, all instances of "ERA-5" throughout the manuscript should be revised. 

Relatedly, the only major concern I have about the construction of the effort relates to the reliance on monthly mean ERA5 fields for the broad correlation analysis used. The precipitation events comprising the multimodal cases occur sporadically throughout the year and are seldom evenly distributed throughout the month. Reducing the information content of the reanalysis to a monthly mean would seem to be counterproductive to identifying prevailing relationships with environmental characteristics and the cases evaluated. Namely, I would imagine that in some months a multimodal case could occur within an environmental extreme compared to the monthly mean and thus the correlations assessed would be mostly meaningless. I recommend utilizing the higher-rate (i.e., hourly) and higher-resolution ERA5 output sampled representatively from each event as the basis for this work.

Though it could be argued to be beyond the scope of the study, an example with colocated aircraft data would be a tremendously meaningful addition to the manuscript, primarily as a more robust validation exercise for the algorithm. I recognize that such data may not exist for the site used, but it would be helpful for that to be acknowledged as justification for not executing such work, should that be the case.  

Specific Comments

Line 13: "access" should be "assess"

Lines 36-39: this is a complicated and confusing sentence. Please clarify and split into 2 sentences.

Line 150: simplify "the vast majority of" to "most"

Lines 233-234: this was already said. Add citations to previous mention or delete. 

Figure 13: this image is very blurry. Please improve resolution. 

Line 439: Unnecessary comma after "cases"