the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Jun 2025
Microphysical Parameter Choices Modulate Ice Content and Relative Humidity in the Outflow of a Warm Conveyor Belt
Abstract. Warm conveyor belts (WCBs) play a crucial role in Earth's climate by transporting water vapor and hydrometeors into the upper troposphere/lower stratosphere (UTLS), where they influence radiative forcing. However, a major source of uncertainty in numerical weather prediction (NWP) models and climate projections stems from the parameterization of microphysical processes and their impact on cloud radiative properties as well as the vertical re-distribution of water. In this study, we use Lagrangian data from a perturbed parameter ensemble (PPE) of a WCB case study to investigate how variations in microphysical parameterizations influence water transport into the UTLS and the outflow cirrus properties. We find that the thermodynamic conditions (pressure, temperature, specific humidity) at the end of the WCB ascent show little sensitivity to the explored parameter perturbations. In contrast, ice content and relative humidity exhibit substantial variability, primarily driven by the capacitance of ice (CAP) and the scaling of ice formation processes directly influenced by ice-nucleating particle (INP) concentrations. Different combinations of CAP and INP scaling yield vastly different ice and relative humidity distributions at the end of the ascent and in the subsequent hours. These differences are particularly pronounced in fast-ascending air parcels, where modifications to the saturation adjustment scheme (SAT) introduce small variations in pressure and temperature at the end of ascent. Our findings have potential implications for parameter choices in cloud models and considerations for geoengineering strategies. Future comparisons with high-quality observational data could help constrain the most realistic parameter choices, ultimately improving weather and climate forecasts.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Chemistry and Physics. The peer-review process was guided by an independent editor, and the authors also have no other competing interests to declare.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(4431 KB)
-
Supplement
(715 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(4431 KB) - Metadata XML
-
Supplement
(715 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-1816', Anonymous Referee #1, 18 Jun 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-RC1-supplement.pdfCitation: https://doi.org/
10.5194/egusphere-2025-1816-RC1 -
AC4: 'Reply on RC1', Cornelis Schwenk, 28 Jul 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-AC4-supplement.pdf
-
AC4: 'Reply on RC1', Cornelis Schwenk, 28 Jul 2025
-
RC2: 'Comment on egusphere-2025-1816', Anonymous Referee #2, 30 Jun 2025
Title: Microphysical Parameter Choices Modulate Ice Content and Relative Humidity in the Outflow of a Warm Conveyor Belt
Authors: Cornelis Schwenk, Annette Miltenberger, and Annika Oertel
Date: June 27, 2025
Recommendation: Minor RevisionsOverview:
This manuscript investigates how variations in cloud-microphysics parameters and SST influence the outflow properties of warm conveyor belts (WCBs) into the upper troposphere–lower stratosphere. 70 ensemble members are used to simulate a single WCB case with specific constant parameters of each ensemble member perturbed using Latin hypercube sampling. The authors show that the end-of-ascent temperature, pressure, and specific humidity are largely insensitive to these perturbations, whereas ice-phase processes—specifically ice-capacitance (CAP) and ice-nucleating-particle (INP) concentrations—strongly modulate ice content and relative humidity. CCN concentrations also have secondary impacts on outflow properties. Differences are greater in fast ascending parcels.
The literature review in the introduction is impressive in scope, and the results are compelling. There are no major issues with the manuscript but many minor ones that if addressed could improve it.
General Comments:
- The science questions are good, but radiative effects that were a big part of the motivation for studying WCB microphysical properties aren’t examined. Is there a reason for that? Connecting microphysical changes to radiative changes would be very insightful.
- A few questions on sensitivities to the methodology:
- There is a bit of justification provided for the range of some parameter values, but it isn’t clear how all of the CAP, INP, CCN, SAT, and SST ranges are chosen for the PPE. Presumably these are constrained by possible real world values dictated by previous studies?
- With no test dataset for the RF model, couldn’t that result in overfitting to this single case?
- WCBs need some definition so thresholds are to be expected, but are results sensitive to the 600-hPa ascent depth and ascent rate threshold used?
- For sensitivity to CCN, how much should we trust the snow and graupel changes in terms of applicability to the real world given arbitrary threshold conversions between these 2 categories as opposed to riming transitioning smoothly to produce a range of variably rimed precipitating ice?
- 6 also shows nonlinear relationships where CAP, INP, and CCN sensitivities are particularly low or high, so it should perhaps be noted that the distribution of values of these parameters in the real world is important for dictating whether overall sensitivities are large or small.
- With the greater sensitivities in fast ascending trajectories, it seems like model representation of convective processes could be important and there could be some model resolution dependence there. Should that be mentioned?
- The conclusions and discussion should include caveats. For example:
- This is a single case, and it isn’t clear how representative it is of WCB events in general.
- PPEs sample the uncertain multi-parameter phase space but still have the weakness of assuming constant parameter values for some parameters that are not real world physical constants. Thus, sensitivities can be overestimated relative to a potentially more realistic stochastic framework in which constant parameters may be varying (e.g., Stanford et al. 2019).
- With only Hallett-Mossop rime splintering parameterized for secondary ice production, could mixed phase ice concentrations be biased low, potentially influencing the WCB outflow sensitivities? Recent studies by Alexei Korolev, Vaughan Phillips, and others have highlighted the potential importance of additional secondary ice mechanisms such as raindrop fragmentation upon freezing and ice collisional breakup.
- Could results be sensitive to the thresholds used to define the WCB (600-hPa depth) and their ascent rates?
- Sentences beginning with “This” could be made clearer by stating the object that it is referring to (for example, “this difference…” instead of just “this…”). These are several instances:
- Line 360: “This high correlation indicates that in both qv PPE members the specific humidity…”.
- Line 389: “Fewer supercooled liquid drops explain why…”
- Line 425: “This delay is also the reason…”
- Line 468: It is unclear what “This” refers to here. Could the authors perhaps rephrase the sentence to not use that word?
- Line 474: It is unclear what about the mean and medians of the two groups “This” is referring to. Is it the difference? Could the authors clarify?
- Line 477: “…because of this subsaturation.”
- Line 489: “Equation 2 shows that parameter…”
- Line 495: “The increase in RHi past tsat,ice=5*10^2 shows that…”
- Lines 516 to 517: “This bulge in turn…”
- Line 551: “This shift means…”
Additional sections that could be made clearer:
Lines 67-68: How far south are the authors referring to? Could the authors be more specific?
Lines 267-271: Are the authors saying that a high IBF score could be due to a parameter being highly correlated with another parameter instead of the parameter being “actually important”?
Here is a possible modification: “Instead, it indicates that the PPE parameter could contribute significantly to the RF model’s prediction due to a high correlation with another PPE parameter.
Lines 304-305: Since the authors are discussing temperatures in terms of Celsius, could the corresponding plots (e.g., Fig. 2) be modified to be in terms of Celsius? Using Celsius would make more intuitive sense in the framework of microphysics discussions including homogeneous freezing.
Line 325: For the first part of this sentence before the comma referring to all PPE members having a mean RHi > 100%, could the authors refer to Figure 2j?
Lines 325 to 327: What about the observation made is “particularly interesting”?
Lines 341 to 345: It seems like the argument here is that using the means for the RF model means that the spread of the distribution (5th to 95th percentiles) is not considered by the RF model. If so, the argument as stated appears a little convoluted and difficult to follow. Is what matters here the change in means between PPE members relative to the spread between the 5th to 95th percentiles (because all variables could be argued to have a large 5th to 95th percentile spread)?
Lines 347: It is unclear how Figure 2c shows a correlation between T95 and qv95. Perhaps the authors meant to refer to Fig. A1c?
Line 354: This statement (“The change is stronger for qv_95 than for T_95”) presumably refers to Figs. 2f and 2b. It is unclear how this change is computed and how the 2 different variable changes can be fairly compared against each other. Perhaps the max 95th value minus the min 95th value divided by the 5th to 95th percentile spread to compare changes relative to the range of variable values?
Line 359: It is clear visually that the highest and lowest qv value correlates strongly with the calculated saturation specific humidity. However, could the authors include a correlation coefficient to quantitatively support this claim?
Lines 357 to 361: Isn’t Fig. A1a or something similar to it plotting qv as a function of temperature a more straightforward argument than Fig. A1b (qv vs. qv_sat) that qv is strongly constrained by temperature? Qv correlates strongly with qv_sat, but they are not 1:1, and it isn’t clear from Fig. A1b alone how temperature vs. pressure modulate qv_sat to affect that relationship or how dynamics and microphysics affecting supersaturation.
Lines 376: “The spread… is unchanged…” The word unchanged seems a little too strong. Could the authors moderate it to “mostly unchanged”?
Lines 377 to 378: “… reduces the spread”. This reduction is not easily visible. Could the authors quantify this reduction?
Line 380: The second mode in the distribution is not a “peak” since it is not a local maximum. It would be more accurate to describe this as a “second mode.”
Line 390: “…many small cloud droplets reach the homogeneous freezing level.” How is “many” defined here? It is highly likely this is homogeneous freezing and glaciation temperatures in Fig. S11 provide some support, but could it be shown that this second mode is indeed due to homogeneous freezing, e.g., by examining drop concentrations at -38C or the change in ice concentration across that temperature level?
Figure S12 caption: “second peak” should be “second higher concentration mode.”
Lines 423 to 424: “This clear dependency… parcel evolution (Fig. 5e)”: In Figure 5e, the max_qc panel is dark red, and the max_qr panel is dark blue. Shouldn’t these large magnitudes mean that CCN strongly modifies the liquid content rather than not strongly affecting it? Also, is “in early stages of parcel evolution” inferred from the liquid mass mixing ratios being maximum values?
Line 424: “Therefore, cloud droplets are far smaller…” It is unclear how cloud droplets are proven to be smaller when qc strongly increases with the CCN scaling factor in Figure 5. Is there other evidence to support this assertion?
Line 428 to 431: the word “presumably” is used twice in this sentence. Suggest replacing one of them with a synonym for improved readability.
Lines 458 to 459: How robust are these red and cyan best fit lines? Could the authors include correlation values?
Line 462: “… the spread of RHi values reduces strongly…” Is there a way to quantify this spread and its reduction?
Line 473: “The mean and median for the first group…” Could the authors clarify what “first group” and “second group” are referring to?
Line 501: Referencing specific visual cues would greatly help the readability of the sentence: “as indicated by the red line, CAP reduces…”
Line 509: “… below this threshold for the second group…” Could the authors clarify which line in the figure they are referring to?
Line 510: “…higher abundance of trajectories above the threshold also explains the long tail”. Could the authors clarify that they are referring to the lines in Fig. 8a?
Line 553: The authors mention fast vs all trajectories, but the figures the authors are referring to within this sentence only contain fast trajectories. Could the authors reference back to the appropriate figure for comparison between the two trajectory groups? For example: “…for fast ascending trajectories than for all trajectories (compare xxx in Figures X and X)”.
Lines 557 to 560: “This is because the peak t for fast trajectories in Fig. 12b is below 5*10^2…”
Line 586: “The large Ni implies smaller Tsat…” Recommend referencing equation 2 here. Also, since decreasing ri has the opposite impact on Tsat, the authors should mention why the effect of Ni on Tsat is greater than ri.
Line 605: Can the authors describe which behavior they are referring to?
Line 606: Could Fig. 4e be referenced right after “bimodal distribution” for improved clarity?
Lines 610 to 611: “Alternatively, some small ice crystals also persist when CAP is high” How does Fig. 15b alone support this statement?
Line 643: “… a small shift in the lower Ni, which…”. It appears that there is a word missing after “Ni”.
Line 697: Did the authors mean “Ni 5 hours after ascent”?
Throughout: When using terms like mean maximum variable, it would help to clarify what the mean and maximum correspond to, so it is clear how the variable is being computed.
Typos:Line 378: “…which is a result of a large difference…”
Line 552: did the authors mean “Fig. 11c and f”?
Lines 634: “…a peak at about 106 %” Fig 8a makes the value seem more like 101%.
Line 650: “Oertel et al. (2025) found altered…”
Line 663: “…larger fraction of the ascent spent in mixed-phase conditions, where water vapor-condensate…” (for consistency with the spelling of vapor elsewhere in the manuscript)
Line 693: “…transport should keep CAP and the INP…”
Figures:
(Optional) The authors should consider re-spelling out the acronyms at key locations (e.g., figures, conclusion) to aid readers.
Figure 2: Make sure that to clearly state that largest and smallest “means” refer to the distributions of the variable for ensemble members with the smallest and largest mean value of that variable. The caption is a bit confusing as currently worded.
Figure 3: Could the authors mention in the caption that these distributions correspond to those at the “end of ascent”?
Figure 5e: (optional) Could the correlations be shown in a table format to add additional information such as the correlation coefficient value? Currently, max_qc and max_qs are very similar in shading.
References:
Stanford, M. W., Morrison, H., Varble, A., Berner, J., Wu, W., McFarquhar, G., & Milbrandt, J. (2019). Sensitivity of simulated deep convection to a stochastic ice microphysics framework. Journal of Advances in Modeling Earth Systems, 11, 3362–3389. https://doi.org/10.1029/2019MS001730
Citation: https://doi.org/10.5194/egusphere-2025-1816-RC2 -
AC3: 'Reply on RC2', Cornelis Schwenk, 28 Jul 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-AC3-supplement.pdf
-
RC3: 'Comment on egusphere-2025-1816', Anonymous Referee #3, 03 Jul 2025
GENERAL
This is an interesting paper examining water vapor transport to high levels and the sensitivity to specific parameters in the microphysics scheme of the ICON model. The authors make a good case about the scientific importance of this type of study in terms of global climate modeling. The paper is very well written, nicely presented, the figures are excellent, and the methodology is scientifically sound. That said, I must declare that I am not really qualified to comment on much related to the PPE method itself. Overall I think this is a great paper and probably nearly ready for publication; my one main concern is the generality of the conclusions in terms of the specific microphysics scheme used (see below). I think the authors should comment on this in a revised manuscript, taking into account the comments below.
SPECIFIC COMMENTS
The “ice” and “snow” categories are very important in this study. First, these should be clearly defined in the paper; these are categories of small and large unrimed ice crystals in bulk and bin microphysics (mp) schemes. In traditional bulk schemes such as ICON’s SB two-moment scheme, these ice-phase categories are predefined and with prescribed physical properties (e.g. capacitance, mass-fall speed parameters) and have the necessary but purely artificial process of conversion between categories. In nature, there is no such thing as “conversion from ice to snow”. In traditional category-based scheme, the way this process is “parameterized” ultimately impacts the relative distribution of ice and snow, subsequently impacting the mp growth rates and distribution of hydrometeor mass.
With this in mind, how general are your conclusions with respect to the mp scheme used? Would you expect the same results/conclusions using a different category-based mp scheme (with different ice-to-snow conversion) or with a property-based scheme (like P3) which uses generic ice-phase categories with no artificial conversion? I think it would be useful to add some discussion (e.g. in the Conclusion section) on this topic.MINOR POINTS
- The description of saturation adjustment in ICON (line 216) seems not quite right. Presumably supersaturation with respect to ice can remain. This seems to be better explained on line 479, but perhaps it should be clarified earlier.
- Line 315: If the units for ice number are in # kg^-1, this is “number mixing ratio”, not number concentration. Concentration units in # m^-3 (as correctly indicated in eqn. (2)).
- 5: In the first color bar, the two dark red colors for “max_qc” and “max_qs” are very hard to distinguish.
- The paragraph starting on line 671 seems more like it should belong in the Introduction section.
Citation: https://doi.org/10.5194/egusphere-2025-1816-RC3 -
AC1: 'Reply on RC3', Cornelis Schwenk, 28 Jul 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-AC1-supplement.pdf
-
RC4: 'Comment on egusphere-2025-1816', Anonymous Referee #4, 03 Jul 2025
General Comments:
Current models have uncertainty in vertical redistributing of water and in cloud radiative properties. To address this, the authors study the effect of microphysical parameterizations on thermodynamic conditions, ice content, and relative humidity at the end of the ascent of warm conveyor belts in a model. This will allow a future study to improve weather and climate forecasts by constraining microphysical parameterizations to match high-quality observational data.
The authors do an outstanding job at contributing to scientific progress. The material they present is original and meaningful and supported by an extensive body of literature. The methods are valid, the results and conclusions are clear, and English is used appropriately. I stronly recommend publication of this paper.
Minor revisions are necessary prior to publication, however. The most important revisions to make are in the results section where more literature should be referenced and cited and where the font sizes on some figures should be tweaked. Additionally, I saw in the introduction that the authors discussed how previous studies addressing UTLS moisture transport focused mainly on the tropics, with more studies needed for the extratropics. I would like to see the authors cite more studies that looked at UTLS moisture transport in the extratropics and explain how their work expands upon and differs from each prior study.
Specific comments:Lines 42-48 ("While most of the aforementioned...contributor to extratropical UTLS moisture."): I would like to see more literature on UTLS moisture transport in the extratropics cited, with explanations of how the authors' study relates to this literature. Examples of articles the authors may want to cite include Weigel et al. 2016 ("UTLS water vapour from SCIAMACHY limb measurements V3.01 (2002-2012)"), Heller et al. 2017 ("Mountain waves modulate the water vapor distribution in the UTLS"), and Sun et al. 2017 ("Characteristics of water vapor in the UTLS over the Tibetan Plateau based on AURA/MLS observations").
Table 1: It would help to show in the table whether perturbations for each parameter between the min and max are spaced linearly (e.g., -2, -1, 0, 1, 2 for SST), logarithmically (e.g., 0.01, 0.02, 0.04, 0.08, ..., 5, 10, 20 for INP), or otherwise, as well as the number of distinct values tested for each parameter.
Lines 339-345 ("We interpret this inability...with parameter perturbations."): The authors should try to compare their results to related findings in at least one study besides Oertel et al. (2025) and see whether any explanation for their results can be found in past literature, and if so, whether the explanation given in past literature is plausible or whether a new explanation is needed. If the authors' results contradict past literature, or if no past literature exists for comparision, that too would be important to mention.
Lines 384-393 ("We interpret these findings...ice mass mixing ratio qi."): See comment for lines 339-345.
Lines 406-417 ("We interpret...(heavier ice particles fall out more quickly).): See comment for lines 339-345.
Lines 422-432 ("This clear dependancy...dominate the ice-phase cloud microphysics."): The authors should check to see whether any related findings exist in any past literature. If so, the authors should cite the literature and relate their analysis to the analysis done in the literature. If not, the authors should clarify that their findings and analysis are original.
Lines 460-462 ("The same is true the other way around...(red dots in Fig. 7 b)."): Please see my comment about figures 7b and 7e. I am not convinced that any linear fit between INP and RH (tao_sat) is statistically significant. Once a relationship between INP and RH (tao_sat) is decided upon, the statement in lines 460-462 should be updated.
Figures 7b and 7e: I am not convinced that the linear fits plotted are worthy of publication. As the authors mention, the INP scaling axis is logarithmic, yet the relative humidity and timescale for supersaturation over ice are fit to linear functions of INP scaling. This creates a situation where only a small fraction of the data has statistical significance in estimating the slope of the fits. Further fuelling my skepticism, the random variability in the median relative humidity and timescale appears to be at least as significant as the total variance captured by the linear fits. If the authors choose to fit relative humidity and timescale for supersaturation over ice to INP scaling, I urge them to consider non-linear fits and test whether any best fit, particularly for any non-linear relationships supported by previous literature, has a significant Spearman correlation coefficient.
Technical comments:Figure 1: The text in the legend above the subplots should be made larger.
Figure 1c: The text on each axis and on the colorbar should be made larger.
Figure 8: The text in the legend above the subplots and the numbers on each subplot's axes should be made larger.
Figures 9-10: The numbers and text in colorbars and the numbers along axes should be made larger.
Citation: https://doi.org/10.5194/egusphere-2025-1816-RC4 -
AC2: 'Reply on RC4', Cornelis Schwenk, 28 Jul 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-AC2-supplement.pdf
-
AC2: 'Reply on RC4', Cornelis Schwenk, 28 Jul 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-1816', Anonymous Referee #1, 18 Jun 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-RC1-supplement.pdf
-
AC4: 'Reply on RC1', Cornelis Schwenk, 28 Jul 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-AC4-supplement.pdf
-
AC4: 'Reply on RC1', Cornelis Schwenk, 28 Jul 2025
-
RC2: 'Comment on egusphere-2025-1816', Anonymous Referee #2, 30 Jun 2025
Title: Microphysical Parameter Choices Modulate Ice Content and Relative Humidity in the Outflow of a Warm Conveyor Belt
Authors: Cornelis Schwenk, Annette Miltenberger, and Annika Oertel
Date: June 27, 2025
Recommendation: Minor RevisionsOverview:
This manuscript investigates how variations in cloud-microphysics parameters and SST influence the outflow properties of warm conveyor belts (WCBs) into the upper troposphere–lower stratosphere. 70 ensemble members are used to simulate a single WCB case with specific constant parameters of each ensemble member perturbed using Latin hypercube sampling. The authors show that the end-of-ascent temperature, pressure, and specific humidity are largely insensitive to these perturbations, whereas ice-phase processes—specifically ice-capacitance (CAP) and ice-nucleating-particle (INP) concentrations—strongly modulate ice content and relative humidity. CCN concentrations also have secondary impacts on outflow properties. Differences are greater in fast ascending parcels.
The literature review in the introduction is impressive in scope, and the results are compelling. There are no major issues with the manuscript but many minor ones that if addressed could improve it.
General Comments:
- The science questions are good, but radiative effects that were a big part of the motivation for studying WCB microphysical properties aren’t examined. Is there a reason for that? Connecting microphysical changes to radiative changes would be very insightful.
- A few questions on sensitivities to the methodology:
- There is a bit of justification provided for the range of some parameter values, but it isn’t clear how all of the CAP, INP, CCN, SAT, and SST ranges are chosen for the PPE. Presumably these are constrained by possible real world values dictated by previous studies?
- With no test dataset for the RF model, couldn’t that result in overfitting to this single case?
- WCBs need some definition so thresholds are to be expected, but are results sensitive to the 600-hPa ascent depth and ascent rate threshold used?
- For sensitivity to CCN, how much should we trust the snow and graupel changes in terms of applicability to the real world given arbitrary threshold conversions between these 2 categories as opposed to riming transitioning smoothly to produce a range of variably rimed precipitating ice?
- 6 also shows nonlinear relationships where CAP, INP, and CCN sensitivities are particularly low or high, so it should perhaps be noted that the distribution of values of these parameters in the real world is important for dictating whether overall sensitivities are large or small.
- With the greater sensitivities in fast ascending trajectories, it seems like model representation of convective processes could be important and there could be some model resolution dependence there. Should that be mentioned?
- The conclusions and discussion should include caveats. For example:
- This is a single case, and it isn’t clear how representative it is of WCB events in general.
- PPEs sample the uncertain multi-parameter phase space but still have the weakness of assuming constant parameter values for some parameters that are not real world physical constants. Thus, sensitivities can be overestimated relative to a potentially more realistic stochastic framework in which constant parameters may be varying (e.g., Stanford et al. 2019).
- With only Hallett-Mossop rime splintering parameterized for secondary ice production, could mixed phase ice concentrations be biased low, potentially influencing the WCB outflow sensitivities? Recent studies by Alexei Korolev, Vaughan Phillips, and others have highlighted the potential importance of additional secondary ice mechanisms such as raindrop fragmentation upon freezing and ice collisional breakup.
- Could results be sensitive to the thresholds used to define the WCB (600-hPa depth) and their ascent rates?
- Sentences beginning with “This” could be made clearer by stating the object that it is referring to (for example, “this difference…” instead of just “this…”). These are several instances:
- Line 360: “This high correlation indicates that in both qv PPE members the specific humidity…”.
- Line 389: “Fewer supercooled liquid drops explain why…”
- Line 425: “This delay is also the reason…”
- Line 468: It is unclear what “This” refers to here. Could the authors perhaps rephrase the sentence to not use that word?
- Line 474: It is unclear what about the mean and medians of the two groups “This” is referring to. Is it the difference? Could the authors clarify?
- Line 477: “…because of this subsaturation.”
- Line 489: “Equation 2 shows that parameter…”
- Line 495: “The increase in RHi past tsat,ice=5*10^2 shows that…”
- Lines 516 to 517: “This bulge in turn…”
- Line 551: “This shift means…”
Additional sections that could be made clearer:
Lines 67-68: How far south are the authors referring to? Could the authors be more specific?
Lines 267-271: Are the authors saying that a high IBF score could be due to a parameter being highly correlated with another parameter instead of the parameter being “actually important”?
Here is a possible modification: “Instead, it indicates that the PPE parameter could contribute significantly to the RF model’s prediction due to a high correlation with another PPE parameter.
Lines 304-305: Since the authors are discussing temperatures in terms of Celsius, could the corresponding plots (e.g., Fig. 2) be modified to be in terms of Celsius? Using Celsius would make more intuitive sense in the framework of microphysics discussions including homogeneous freezing.
Line 325: For the first part of this sentence before the comma referring to all PPE members having a mean RHi > 100%, could the authors refer to Figure 2j?
Lines 325 to 327: What about the observation made is “particularly interesting”?
Lines 341 to 345: It seems like the argument here is that using the means for the RF model means that the spread of the distribution (5th to 95th percentiles) is not considered by the RF model. If so, the argument as stated appears a little convoluted and difficult to follow. Is what matters here the change in means between PPE members relative to the spread between the 5th to 95th percentiles (because all variables could be argued to have a large 5th to 95th percentile spread)?
Lines 347: It is unclear how Figure 2c shows a correlation between T95 and qv95. Perhaps the authors meant to refer to Fig. A1c?
Line 354: This statement (“The change is stronger for qv_95 than for T_95”) presumably refers to Figs. 2f and 2b. It is unclear how this change is computed and how the 2 different variable changes can be fairly compared against each other. Perhaps the max 95th value minus the min 95th value divided by the 5th to 95th percentile spread to compare changes relative to the range of variable values?
Line 359: It is clear visually that the highest and lowest qv value correlates strongly with the calculated saturation specific humidity. However, could the authors include a correlation coefficient to quantitatively support this claim?
Lines 357 to 361: Isn’t Fig. A1a or something similar to it plotting qv as a function of temperature a more straightforward argument than Fig. A1b (qv vs. qv_sat) that qv is strongly constrained by temperature? Qv correlates strongly with qv_sat, but they are not 1:1, and it isn’t clear from Fig. A1b alone how temperature vs. pressure modulate qv_sat to affect that relationship or how dynamics and microphysics affecting supersaturation.
Lines 376: “The spread… is unchanged…” The word unchanged seems a little too strong. Could the authors moderate it to “mostly unchanged”?
Lines 377 to 378: “… reduces the spread”. This reduction is not easily visible. Could the authors quantify this reduction?
Line 380: The second mode in the distribution is not a “peak” since it is not a local maximum. It would be more accurate to describe this as a “second mode.”
Line 390: “…many small cloud droplets reach the homogeneous freezing level.” How is “many” defined here? It is highly likely this is homogeneous freezing and glaciation temperatures in Fig. S11 provide some support, but could it be shown that this second mode is indeed due to homogeneous freezing, e.g., by examining drop concentrations at -38C or the change in ice concentration across that temperature level?
Figure S12 caption: “second peak” should be “second higher concentration mode.”
Lines 423 to 424: “This clear dependency… parcel evolution (Fig. 5e)”: In Figure 5e, the max_qc panel is dark red, and the max_qr panel is dark blue. Shouldn’t these large magnitudes mean that CCN strongly modifies the liquid content rather than not strongly affecting it? Also, is “in early stages of parcel evolution” inferred from the liquid mass mixing ratios being maximum values?
Line 424: “Therefore, cloud droplets are far smaller…” It is unclear how cloud droplets are proven to be smaller when qc strongly increases with the CCN scaling factor in Figure 5. Is there other evidence to support this assertion?
Line 428 to 431: the word “presumably” is used twice in this sentence. Suggest replacing one of them with a synonym for improved readability.
Lines 458 to 459: How robust are these red and cyan best fit lines? Could the authors include correlation values?
Line 462: “… the spread of RHi values reduces strongly…” Is there a way to quantify this spread and its reduction?
Line 473: “The mean and median for the first group…” Could the authors clarify what “first group” and “second group” are referring to?
Line 501: Referencing specific visual cues would greatly help the readability of the sentence: “as indicated by the red line, CAP reduces…”
Line 509: “… below this threshold for the second group…” Could the authors clarify which line in the figure they are referring to?
Line 510: “…higher abundance of trajectories above the threshold also explains the long tail”. Could the authors clarify that they are referring to the lines in Fig. 8a?
Line 553: The authors mention fast vs all trajectories, but the figures the authors are referring to within this sentence only contain fast trajectories. Could the authors reference back to the appropriate figure for comparison between the two trajectory groups? For example: “…for fast ascending trajectories than for all trajectories (compare xxx in Figures X and X)”.
Lines 557 to 560: “This is because the peak t for fast trajectories in Fig. 12b is below 5*10^2…”
Line 586: “The large Ni implies smaller Tsat…” Recommend referencing equation 2 here. Also, since decreasing ri has the opposite impact on Tsat, the authors should mention why the effect of Ni on Tsat is greater than ri.
Line 605: Can the authors describe which behavior they are referring to?
Line 606: Could Fig. 4e be referenced right after “bimodal distribution” for improved clarity?
Lines 610 to 611: “Alternatively, some small ice crystals also persist when CAP is high” How does Fig. 15b alone support this statement?
Line 643: “… a small shift in the lower Ni, which…”. It appears that there is a word missing after “Ni”.
Line 697: Did the authors mean “Ni 5 hours after ascent”?
Throughout: When using terms like mean maximum variable, it would help to clarify what the mean and maximum correspond to, so it is clear how the variable is being computed.
Typos:Line 378: “…which is a result of a large difference…”
Line 552: did the authors mean “Fig. 11c and f”?
Lines 634: “…a peak at about 106 %” Fig 8a makes the value seem more like 101%.
Line 650: “Oertel et al. (2025) found altered…”
Line 663: “…larger fraction of the ascent spent in mixed-phase conditions, where water vapor-condensate…” (for consistency with the spelling of vapor elsewhere in the manuscript)
Line 693: “…transport should keep CAP and the INP…”
Figures:
(Optional) The authors should consider re-spelling out the acronyms at key locations (e.g., figures, conclusion) to aid readers.
Figure 2: Make sure that to clearly state that largest and smallest “means” refer to the distributions of the variable for ensemble members with the smallest and largest mean value of that variable. The caption is a bit confusing as currently worded.
Figure 3: Could the authors mention in the caption that these distributions correspond to those at the “end of ascent”?
Figure 5e: (optional) Could the correlations be shown in a table format to add additional information such as the correlation coefficient value? Currently, max_qc and max_qs are very similar in shading.
References:
Stanford, M. W., Morrison, H., Varble, A., Berner, J., Wu, W., McFarquhar, G., & Milbrandt, J. (2019). Sensitivity of simulated deep convection to a stochastic ice microphysics framework. Journal of Advances in Modeling Earth Systems, 11, 3362–3389. https://doi.org/10.1029/2019MS001730
Citation: https://doi.org/10.5194/egusphere-2025-1816-RC2 -
AC3: 'Reply on RC2', Cornelis Schwenk, 28 Jul 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-AC3-supplement.pdf
-
RC3: 'Comment on egusphere-2025-1816', Anonymous Referee #3, 03 Jul 2025
GENERAL
This is an interesting paper examining water vapor transport to high levels and the sensitivity to specific parameters in the microphysics scheme of the ICON model. The authors make a good case about the scientific importance of this type of study in terms of global climate modeling. The paper is very well written, nicely presented, the figures are excellent, and the methodology is scientifically sound. That said, I must declare that I am not really qualified to comment on much related to the PPE method itself. Overall I think this is a great paper and probably nearly ready for publication; my one main concern is the generality of the conclusions in terms of the specific microphysics scheme used (see below). I think the authors should comment on this in a revised manuscript, taking into account the comments below.
SPECIFIC COMMENTS
The “ice” and “snow” categories are very important in this study. First, these should be clearly defined in the paper; these are categories of small and large unrimed ice crystals in bulk and bin microphysics (mp) schemes. In traditional bulk schemes such as ICON’s SB two-moment scheme, these ice-phase categories are predefined and with prescribed physical properties (e.g. capacitance, mass-fall speed parameters) and have the necessary but purely artificial process of conversion between categories. In nature, there is no such thing as “conversion from ice to snow”. In traditional category-based scheme, the way this process is “parameterized” ultimately impacts the relative distribution of ice and snow, subsequently impacting the mp growth rates and distribution of hydrometeor mass.
With this in mind, how general are your conclusions with respect to the mp scheme used? Would you expect the same results/conclusions using a different category-based mp scheme (with different ice-to-snow conversion) or with a property-based scheme (like P3) which uses generic ice-phase categories with no artificial conversion? I think it would be useful to add some discussion (e.g. in the Conclusion section) on this topic.MINOR POINTS
- The description of saturation adjustment in ICON (line 216) seems not quite right. Presumably supersaturation with respect to ice can remain. This seems to be better explained on line 479, but perhaps it should be clarified earlier.
- Line 315: If the units for ice number are in # kg^-1, this is “number mixing ratio”, not number concentration. Concentration units in # m^-3 (as correctly indicated in eqn. (2)).
- 5: In the first color bar, the two dark red colors for “max_qc” and “max_qs” are very hard to distinguish.
- The paragraph starting on line 671 seems more like it should belong in the Introduction section.
Citation: https://doi.org/10.5194/egusphere-2025-1816-RC3 -
AC1: 'Reply on RC3', Cornelis Schwenk, 28 Jul 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-AC1-supplement.pdf
-
RC4: 'Comment on egusphere-2025-1816', Anonymous Referee #4, 03 Jul 2025
General Comments:
Current models have uncertainty in vertical redistributing of water and in cloud radiative properties. To address this, the authors study the effect of microphysical parameterizations on thermodynamic conditions, ice content, and relative humidity at the end of the ascent of warm conveyor belts in a model. This will allow a future study to improve weather and climate forecasts by constraining microphysical parameterizations to match high-quality observational data.
The authors do an outstanding job at contributing to scientific progress. The material they present is original and meaningful and supported by an extensive body of literature. The methods are valid, the results and conclusions are clear, and English is used appropriately. I stronly recommend publication of this paper.
Minor revisions are necessary prior to publication, however. The most important revisions to make are in the results section where more literature should be referenced and cited and where the font sizes on some figures should be tweaked. Additionally, I saw in the introduction that the authors discussed how previous studies addressing UTLS moisture transport focused mainly on the tropics, with more studies needed for the extratropics. I would like to see the authors cite more studies that looked at UTLS moisture transport in the extratropics and explain how their work expands upon and differs from each prior study.
Specific comments:Lines 42-48 ("While most of the aforementioned...contributor to extratropical UTLS moisture."): I would like to see more literature on UTLS moisture transport in the extratropics cited, with explanations of how the authors' study relates to this literature. Examples of articles the authors may want to cite include Weigel et al. 2016 ("UTLS water vapour from SCIAMACHY limb measurements V3.01 (2002-2012)"), Heller et al. 2017 ("Mountain waves modulate the water vapor distribution in the UTLS"), and Sun et al. 2017 ("Characteristics of water vapor in the UTLS over the Tibetan Plateau based on AURA/MLS observations").
Table 1: It would help to show in the table whether perturbations for each parameter between the min and max are spaced linearly (e.g., -2, -1, 0, 1, 2 for SST), logarithmically (e.g., 0.01, 0.02, 0.04, 0.08, ..., 5, 10, 20 for INP), or otherwise, as well as the number of distinct values tested for each parameter.
Lines 339-345 ("We interpret this inability...with parameter perturbations."): The authors should try to compare their results to related findings in at least one study besides Oertel et al. (2025) and see whether any explanation for their results can be found in past literature, and if so, whether the explanation given in past literature is plausible or whether a new explanation is needed. If the authors' results contradict past literature, or if no past literature exists for comparision, that too would be important to mention.
Lines 384-393 ("We interpret these findings...ice mass mixing ratio qi."): See comment for lines 339-345.
Lines 406-417 ("We interpret...(heavier ice particles fall out more quickly).): See comment for lines 339-345.
Lines 422-432 ("This clear dependancy...dominate the ice-phase cloud microphysics."): The authors should check to see whether any related findings exist in any past literature. If so, the authors should cite the literature and relate their analysis to the analysis done in the literature. If not, the authors should clarify that their findings and analysis are original.
Lines 460-462 ("The same is true the other way around...(red dots in Fig. 7 b)."): Please see my comment about figures 7b and 7e. I am not convinced that any linear fit between INP and RH (tao_sat) is statistically significant. Once a relationship between INP and RH (tao_sat) is decided upon, the statement in lines 460-462 should be updated.
Figures 7b and 7e: I am not convinced that the linear fits plotted are worthy of publication. As the authors mention, the INP scaling axis is logarithmic, yet the relative humidity and timescale for supersaturation over ice are fit to linear functions of INP scaling. This creates a situation where only a small fraction of the data has statistical significance in estimating the slope of the fits. Further fuelling my skepticism, the random variability in the median relative humidity and timescale appears to be at least as significant as the total variance captured by the linear fits. If the authors choose to fit relative humidity and timescale for supersaturation over ice to INP scaling, I urge them to consider non-linear fits and test whether any best fit, particularly for any non-linear relationships supported by previous literature, has a significant Spearman correlation coefficient.
Technical comments:Figure 1: The text in the legend above the subplots should be made larger.
Figure 1c: The text on each axis and on the colorbar should be made larger.
Figure 8: The text in the legend above the subplots and the numbers on each subplot's axes should be made larger.
Figures 9-10: The numbers and text in colorbars and the numbers along axes should be made larger.
Citation: https://doi.org/10.5194/egusphere-2025-1816-RC4 -
AC2: 'Reply on RC4', Cornelis Schwenk, 28 Jul 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1816/egusphere-2025-1816-AC2-supplement.pdf
-
AC2: 'Reply on RC4', Cornelis Schwenk, 28 Jul 2025
Peer review completion
Journal article(s) based on this preprint
Data sets
Trajectory data Annika Oertel https://www.doi.org/10.35097/ecgs4f56mp3ymjmt
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 789 | 87 | 27 | 903 | 36 | 11 | 31 |
- HTML: 789
- PDF: 87
- XML: 27
- Total: 903
- Supplement: 36
- BibTeX: 11
- EndNote: 31
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
Cornelis Schwenk
Annette Miltenberger
Annika Oertel
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(4431 KB) - Metadata XML
-
Supplement
(715 KB) - BibTeX
- EndNote
- Final revised paper