Initial Submission:

Recommendation: Major Revisions.

Comments to Author(s):

Manuscript Number: egusphere-2023-2986

Manuscript Title: Microphysical processes involving the vapour phase dominate in simulated low-level Arctic clouds

Authors: Theresa Kiszler et al.

Overview and general recommendation:

This manuscript uses a new microphysics wrapper in the ICON-LEM model to extract microphysical tendencies related to low-level cloud phase in the Arctic. The authors wish to quantify the relative importance of different phase-changing processes. To simplify the analysis, all liquid and ice hydrometeors are binned together, respectively, so that only two categories are considered in the analysis. These tendencies are plotted as a function of temperature, vertical velocity, and ice saturation, and compared between polar day and polar night. The authors find that tendencies involving the vapor phase of water (e.g. evaporation, deposition, sublimation, and condensation) are especially important. The authors interpret the frequent occurrence of evaporation and deposition as evidence of the WBF process.

The approach used in this manuscript is novel and interesting. Instantaneous microphysical tendency fields are critical to mixed-phase cloud processes, but are rarely studied at this scale. The tool described and employed here gives granular insight into these processes in a valuable way. The writing is generally clear and I also find the manuscript to be well-organized. I find the results interesting, but their analysis and presentation leads to conclusions that feel vague and lack clear takeaways. Specifically, conclusions regarding the WBF process should explicitly quantify its importance as a function of other climate fields. I think that with some additional recommended analysis, the results will be more valuable to the broader community. As a result, I recommend major revisions. These concerns along with other more minor comments are included below.

Comments are formatted as:

Line number: “Text”

Specific Comment

1: “either”

Possible change: replace “either” with "both". "either" does not add to the information conveyed by the sentence since any bias must result from an over- or under-estimation of a model field.

6-8: “It was found that… vapour phase dominate.”

The first conclusion is vague, what processes specifically vary and how so?

The second conclusion seems more important to the manuscript. Should it be listed first and highlighted?

9-10: “Going a step further…occurrence.”

The meaning and importance of this sentence is unclear to me. Wouldn't we effect the WBF effect to occur in the mixed-phase regime and to be characterized by liquid evaporation and ice deposition? Is new understanding enabled by this perspective?

31-32: Citation of McGraw (2023) and Shaw (2022)

I would highlight that these studies adjust both the ice nucleation and WBF processes while investigating the cloud forcing, not just the WBF process.

McGraw shows that agreement with observations can be achieved with multiple nucleation schemes if WBF is adjusted, while Shaw shows that multiple configurations of the WBF process rate and nucleation rates can be used to achieve good cloud phase agreement with observations.

33: “new: Classical Nucleation Theory”

This citation format is not familiar to me. Please review.

53: “making the location…clouds.”

Given later comments about the importance of surface type, can you comment on representativeness of Svalbard to the Arctic as a whole?

53: “ICON-LEM”

Please introduce this acronym.

62 (and elsewhere): “approx.”

This does not need to be abbreviated.

64-65: “The forcing is…ICON-NWP runs.”

For those unfamiliar with this model: Within the model domain, what boundary conditions are applied vs. determined by the model?

e.g. Greenhouse gas concentrations, surface temperature/type/fluxes, radiation fields at model boundaries.

61-72 (entire paragraph)

Has this model/configuration been evaluated in its ability to capture MPCs and their radiative effects before? A simple, general overview of the model's performance would be useful here if available.

Convince the readers that this model is an appropriate tool for this study (e.g. fit for task). If the model has biases or limitations, how may they affect the conclusions of this study?

75: “10-8 kg/kg”

Later in the manuscript a threshold is described as 10^-18 kg/kg. Are these different thresholds or is it an error in the text?

89-94: Entire paragraph

How is the model vertical coordinate handled when identifying clouds and microphysical tendencies? Are outputs also produced on a vertical grid for cells where a cloud is present?

96-97: “This approach…single processes.”

Are the rates produced by this approach identical to those interactively seen by the model during the run?

98-101: “Another advantage…schemes have run.”

This is a very important note. So the saved output is not merely a subset for MPCs, it is taken during a different part of the model updating loop?

A flowchart figure showing this could be really helpful for a reader!

112-114: “In this study…compensate each other.”

Does this simplification create any potential shortcomings in the analysis? E.g. Phase partitioning is also dependent on sources and sinks of hydrometeors, so conversion of cloud ice to graupel/hail could quickly deplete ice and modify the phase partitioning.

128: “making the frozen mass increase due to riming smaller.”

meaning of "smaller" is unclear.

141-142: One can see…frozen hydrometeors.”

A clear example could make the meaning of this statement more clear.

148-149: During the polar day…less frequent.”

The process rates seem to depend heavily on the mean state of the cloud phase. Does this present a chicken-egg problem? i.e. Is the cloud phase the result of the process rates or are the process rates the result of the cloud phase? This may be more of a question for the discussion, but I think it is important to address.

Also, do the rates generally balance each other? If sublimation decreases during the day does a vapor to condensate or ice transition also decrease to maintain balance?

If equilibrium is not achieved, is this the result of ignoring advection?

I guess what I am asking here is what does the budget for vapor, liquid, and ice look like using this analysis? I think that this would be a helpful way to visualize the balance of processes at play and also to understand processes that may be missing from this analysis (i.e. a not-closed budget implies important roles from advection, etc).

158: This is likely…is used here.”

See question about the representativeness of results above. How well can the results be generalized for the entire Arctic if most is ocean? Could you repeat the analysis for an ocean gridcell to see if the conclusions change? Would you expect the conclusions to change with a different surface type?

162-163: “while processes…not as relevant.”

I would consider clarifying where the WBF process falls here. I assume that it is not considered a liquid to frozen process because in the model there is an intermediate vapor phase step. But in some models the WBF process directly shifts liquid water to ice, right?

165-166: To get an…the hydrometeors.”

Would you consider also looking at the process rates as a function of the vertical coordinate? E.g. height above cloud base or optical depth?

What about the vertical coordinate? Given the strong vertical phase partitioning of mixed-phase clouds, I might expect a very different balance of process rates between liquid cloudtops and the icier cloud interiors.

177-180: “Of the total…31% contained ice.”

This is really interesting! Could you include histograms of liquid water content and ice water content for polar day and night as well?

201-202: “The homogeneous…further included.”

Are the statistics bad? I would be interested in seeing these values plotted if they are interpretable. As you previously noted, the mass tendencies from these processes are small, but the number tendencies are very important.

Figure 3:

Plots in the left column would be easier to read with more standard temperature increments (e.g. 5C) and ticks.

220-221: “Reasons for…temperature range.”

I think these are good hypotheses, especially given previous arguments that PD/PN differences result in part from differences in temperature and cloud phase. To test this hypothesis, you could plot 2-d histograms of deposition frequency and tendency as a function of both vertical velocity and temperature.

223: “Ny-Å”

Abbreviation is not needed.

225-226: “For the liquid mass…upward motion.”

Could you include these described results as supplementary figures?

240: “The introduction mentions that”

Remove this text.

241-242: “As shown in…underestimated.”

This evaluation of ICON-LEM should be included in the introduction so readers have an understanding of the performance of the model and if it is fit for task.

Underestimated by how much? Can the model be trusted for this study? See fit-for-task comment in the methods.

244-246: “The hypothesis…phase cloud.”

The wording here is unclear to me. The WBF process is active when the air is saturated with respect to ice but not water, so condensation does not occur. Evaluating the role of the WBF process requires knowledge of the temperature, right?

Figure 4:

Recommendation for readability: Updates x-ticks in subplot (a) to be centered around zero. Matching panel b would be easiest for readers.

247-249: “was produced…deposition occurs.”

What is the relative frequency of each set? i.e. How often is this simple definition of the WBF process happening when both cloud phases are present?

261: “For both…statistically significant.”

Can you state what method of significance testing was used?

263: “one-order-of-magnitude”

This is a fourfold increase, not quite an order of magnitude.

263-264: “increase…same order of magnitude”

If the absolute increase in evaporation is not matched by the increase in deposition, are there other important terms in the moisture budget?

240-267: Section on the WBF process.

I think that there is an excellent opportunity here to explicitly quantify the importance of the WBF process and evaluate it as a function of the variables used previously (vertical velocity, temperature, etc). Similar to the methods used to separate the different cases, you could calculate a simple estimate of liquid-to-ice flux due to the WBF process by taking the minimum of the evaporation and deposition tendencies when temperature is between 0 and -38C (or using another approach if one is more valid). This new WBF tendency could be compared with the others (as in figure 1) and also plotted as a function of vertical wind speed, temperature, and saturation. Additionally, you could compare this WBF tendency to the sum of all liquid-to-ice tendencies to see what fraction of liquid-to-ice transitions can be attributed to the WBF process and how that fraction changes as a function of vertical wind speed, temperature, and saturation. I think that this would allow you to quantify your results in a clear way for the audience.

In summary, I recommend:

• Calculating a new WBF tendency as described above.
• Including that tendency in Figure 1.
• Plotting both the WBF tendency and the fraction of liquid-to-ice flux it accounts for as a function of temperature, vertical wind speed, and saturation.

Figure 6:

Suggestions:

- Use consistent labelling of the sets in the figure legends.

- Check if colors in panel (a) are colorblind friendly.

- Report the fractional occurrence of each set in the figure label.

276-278: “The results suggest…active enough.”

This conclusion feels pretty vague, could you be more specific? Do the conclusions of this study point to one approach versus another?

278-279: “Further…large dataset.”

What does this say about the importance of the mean atmospheric state? How does ICON-LEM's representation of the mean state influence the conclusions of this work?

280-282: “It is worth…process rates.”

So mass tendencies from nucleation processes are small but the number tendencies are important? Could this help guide model development and tuning? For example, could nucleation only influence ice and liquid number tendencies and not mass tendencies without a large effect? (just for low-level mixed-phase clouds?)

305-306: “and the results suggest…frozen phase.”

This language is too vague. How much of the mass tendency occurs via the WBF process?

A simple estimate of the WBF process could be calculated as minimum(liq-to-vap,vap-to-ice) for individual timesteps.

318-319: “In this study…completely correct.”

I think this should be included in the introduction/model description. Are the aerosols prognostic, do they evolve in time?

Review of “Microphysical processes involving the vapour phase dominate in simulated low-level Arctic clouds” by T. Kiszler et al.

This study presents an analysis of phase-change process rates in Arctic clouds above Ny-Alesund motivated by a need to better understand phase-partitioning in mixed-phase clouds.  I think that analysis of process rates can be a valuable way to understand clouds and that is something that should be done more often. That said, I have some concerns about the applicability of their results to mixed-phase clouds generally. Even aside from these concerns, I’m not sure what the author’s main conclusions are. The only specific conclusion in the abstract is that “the importance of a process varies for the polar night and polar day … phase changes that involve the vapour phase dominate.” I’d argue that neither of these are particularly novel results. That said, I think that with some additional analysis, the paper could provide more insight than it does in its current form.

Major Comments:

1. Figure 1 shows that the liquid water budget is clearly not closed. The evaporation and condensation conditionally-averaged rates are essentially equal (my understanding is that averages are taken over all points that meet the minimum rate threshold), but evaporation is 4 – 8 times more frequent. That the budget is not closed is because the authors analyze rates through only a single grid column over Ny-Alesund. The results suggest then that clouds are advected over Ny-Alesund and not generally forming over Ny-Alesund. In short, Ny-Alesund does not capture the full life-cycle of clouds. This is a major limitation of the study, and a limitation that is not discussed by the authors. I find it very difficult to interpret the results without any sense for the underlying distribution of lifecycle stages of the clouds (and I realize that even quantifying the lifecycle is non-trivial). I think that this concern could be partly mitigated by focusing less on the tendencies in the dataset as a whole and more on subsets of data that include only clouds that meet certain criteria. For example, subset by clouds that are growing and clouds that are dissipating, or by clouds that are becoming more glaciated and clouds that are becoming less glaciated.

1. The authors are motivated by the need to understand phase partitioning, but aside from the WBF analysis toward the end of the manuscript, there is no explicit analysis of phase partitioning. Why not restrict analysis throughout the entire manuscript to mixed-phase clouds? And/or examine, say, tendencies of ice/liquid water fraction and identify the processes that are most important for changing this fraction? These processes may or may not be the same as the processes that are most important for the total liquid change and total ice change, if for example, two processes are well correlated and offsetting one another, or if for example, a process such as condensation is important only when a cloud is predominantly liquid. If the authors were to go down that route, I think it would be useful to additionally include sedimentation fluxes in and out of grid boxes since that is also a process that could change the local partitioning. It could also be interesting to examine the phase partitioning processes as a function of height in the cloud – perhaps phase-partitioning processes important near cloud top are not important in the precipitating regions of the cloud and vice versa.

In short, I think that the analysis could have been more creative to mitigate weaknesses in the data sampling and to provide more insight into phase partitioning. I don’t think the authors need to take all of my suggestions, but I think that major revisions could address these weaknesses and produce a study that more directly addresses the gap in knowledge that they identify and that will be ultimately more impactful for the community.

Minor Comments:

Line 45: I was uncertain whether “deposition rate” here referred to the deposition of snow to the ground surface, or the deposition of water vapour to the crystal surface. I think now that it is the former.

Line 98-101: I don’t understand this advantage. Are you writing the thermodynamic variables as they exist immediately before the call to microphysics? If so, I don’t believe that this was stated explicitly.

Line 128: confusing wording, please rephrase

Line 195: Is “increases” supposed to be “decreases”?

Lines 200-204: Most of these results are not shown, is that correct? If so, please say so explicitly.

Lines 234-238: Are any of these results shown?

Line 249: Presumably “no deposition occurs” is the same thing as “sublimation occurs” since we are only examining mixed-phase clouds?