the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 02 Sep 2024
Investigating ice formation pathways using a novel two-moment multi-class cloud microphysics scheme
Abstract. We developed a novel microphysics scheme to investigate the formation pathways of ice crystals in the atmosphere. The new two-moment scheme distinguishes between five ice classes (’ice modes’) each with their unique formation mechanism: homogeneous freezing of solution droplets, deposition nucleation, homogeneous freezing of cloud droplets and raindrops, immersion freezing and secondary ice from rime splintering. The ice modes interact with each other, e.g. in competition for growth by deposition of water vapor and aggregation, but also with the other cloud particle classes, i.e., cloud droplets, rain, snow, graupel, hail.
This scheme was employed to investigate the liquid origin vs in-situ formation in the fully glaciated parts of an idealised convective cloud. Liquid origin ice clouds stem from droplets that freeze close to water saturation. In-situ formed ice clouds form directly from the vapor phase below water saturation. The majority of the cloud ice in the deep convection cloud consisted of frozen droplets (liquid origin). This was caused by the high number concentration of cloud droplets available for freezing. In-situ formed ice was only relevant for the overshoot where ice from both formation pathways mixed.
The new scheme is also useful for investigation of the ice formation in the mixed-phase parts of the convective cloud. There is a vertical layering of ice modes in the cloud. The lower most layer consists of secondary ice from rime splintering and occurred near the updraft core at temperatures around the Hallet-Mossop zone. In an altitude between 6 and 9 km ice mostly stems from immersion freezing. We find a correlation between the abundance of ice from immersion freezing and snow. The majority of ice crystals above 9 km stems from homogeneously frozen cloud droplets since ice nucleating particles (INP) required for immersion freezing where quickly depleted.
-
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
(4643 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(4643 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-2157', Anonymous Referee #1, 08 Oct 2024
This manuscript provides a detailed overview of the development of a new bulk microphysics scheme that includes several classes of ice hydrometeor, as opposed to the typical two classes in older schemes. It also includes the evaluation of a test case in an idealized convective cloud that clearly shows how differences between ice microphysical assumptions impact key properties/processes within a cloud. The further development of ice microphysical schemes is important to better understand how the ice phase influences key cloud properties. However, that message does not come across clearly enough in this work, with the sparse amount of . Similarly, I recommend that the authors undergo more rounds of revision and proofreading before this manuscript is ready for publication. I have included specific comments/questions/suggestions below to aid in the revision.
Introduction
- Here is where I found that the discussion did not cover enough of the work to date to justify the need for a new multi-class scheme, with too little literature cited overall. A couple background topics to consider covering: What is it about two class schemes that make them unsuitable to simulate ice microphysics? What have others done with multi-class schemes and what have they found (consider reviewing the work on the P3 scheme by Morrison and Milbrandt, 2015 and others).
Model description
- Lines 142-149: liquid-origin and in situ were already defined in the introduction so I don’t think you need to redefine them here
- Line 166: why does the saturation adjustment mean that the direct numerical integration of the homogeneous freezing coefficient?
- Lines 243-245: the fact that the model does not include an explicit aerosol microphysics model for simulating CCN and INPs, this could be considered for future work
- Figure 2: please readjust panel a such that xtick labels are not overlapping
- Section 2.3, consider renaming as the contents cover depositional growth of ice particles so the title may confuse readers
- Section 2.2.4: consider stating that the differences between these heterogeneous nucleation parameterizations are tested below for an idealized convective case because of the implications on the ice phase in mixed-phase clouds (as stated in the text)
Idealised simulations
- The authors should make it clearer in the text that they tested the new multi-class ice scheme with the three heterogeneous nucleation options by HA, UL, and PH (with the additional PH3) as detailed in Section 2.
- Please make clearer that the REF scheme refers to the SB scheme with only two ice modes
- It is typical to explicitly state the model version, resolution, and domain (both space and time) up front in a description of experiments, or in the methods section before this. For example, there are several versions of the ICON model, so it should be made clear that the authors ran it in LES mode following Heinze et al. (2017). What other standard configurations are worth noting for your model version? This helps with reproducibility of your results.
- What is the Weisman-Klemp setup? Is it the prescribed temperature and relative humidity profile? Why this set up over others (if available)?
- Why is the comparison made only between REF and HA at the start of Section 3.1?
- In section 3.1 it might be worth noting that you discuss the other panels in figure 3 later on
- Line 511, sentence starting with “However, it is heavier…”. Does the “it” refer to ice, as in total ice mass? Maybe consider making this clearer.
- Figures 5, 6, and 7 are generally great. Where did the 0.1 mg/m3 threshold come from?
- Lines 543-544 and Figure 7: “There is significantly more snow present in the ice mode simulation than in the reference simulation.” That does not seem to be the case looking at figure 7. Can the authors explain what they mean more clearly?
- Line 547, is this sentence referring to the HA scheme?
- Line 588: please state which figure again, and I think the authors might mean panels c and g instead?
Discussion
- Are there any limitations with this new scheme that weren’t addressed by this study that could be considered in future work?
General
- Several instances of using “an” where there should have been and “a” and vice versa
- Several acronyms weren’t defined before their use or the longname was used after the acronym was defined, check for consistency
- Are there any observations of precipitation rates that the authors could compare to their new model setup as a test of their new scheme?
Citation: https://doi.org/10.5194/egusphere-2024-2157-RC1 - AC1: 'Reply on RC1', Tim Lüttmer, 28 Nov 2024
-
RC2: 'Comment on egusphere-2024-2157', Anonymous Referee #2, 18 Oct 2024
This work introduces new ice classes into an existing cloud microphysics scheme to better elucidate the role of different ice formation mechanisms on the cloud evolution. Since the role of the ice phase on the cloud evolution is still fraught with uncertainty this work and its conclusions are of significance to the scientific community. It is also well written, however some issues must be addressed before publication, as described below.
General comments.
The introduction is almost completely devoid of any citations, despite being full of overarching generalizations. It makes me wonder whether the authors conducted a proper literature review before writing the paper. Please add citations and put the work within the context of the current literature.
It is important to recognize that this is a very idealized model, that although illustrative it might have limited application in the modeling of real situations. The size distribution of the ice crystals is a single function and can’t be separated by origin as ice grows. It would be impossible to measure or even estimate the parameters of the distribution for the different ice classes. It is also not recommended to excessively add tracers to be advected by the host model since it may lead to numerical diffusion issues. For host models with lower resolution it would be difficult to relate the different classes to macro-scale variables like cloud fraction and total condensate. Finally, the INP classes used in the work are somehow artificially separated by design while in real cloud they tend to be active at the same time. These considerations must be made clear in the discussion section of the work.
Specific comments.
Line 19. Clouds are uncertain in models not in the Earth, please rephrase.
Lines 24, 29, 44, 45, 54, 55 and many other places. These statements are not obvious and need references backing them up.
Line 52. Cloud types.
Line 59. Remove “will”.
Line 70. Please rewrite this sentence in clearer terms.
Line 74. What is f? Should it be normalized to the total mass instead?
Line 76. Define “the particle mass distribution”.
Line 104. Remove “Actually”.
Line 125. This needs a reference as well.
Line 135. Remove “huge”.
Line 140. Please clarify what this means.
Line 154. Should contact ice nucleation be included?
Line 200. Since competition between homogeneous and heterogeneous nucleation is a significant feature of cloud ice formation, it is not clear how this approach would represent real clouds.
Line 210. This expression needs to be weighted by the size distribution of the aqueous solution droplets. In fact, is nhom limited at all by the available droplets?
Line 244. This is a strong assumption that makes this a very idealized model. Please clarify how it might affect the results.
Line 269. This only works for INP immersed within cloud droplets. It is not clear whether it is applied that way.
Line 283. There is an issue here since “na” for deposition and immersion is different (i.e., whether they are inside/outside of cloud droplets).
Line 344. Why is this not dependent on the mass of other cloud species?
Liner 440. How many new tracers are added to the dynamics scheme? What is the domain of the simulation?
Figure 3: Are these lines averaged over the whole domain?
Line 477, 497, 498 and other places. Using words like “weak”, “stronger”, “catches up” is ambiguous. Be precise on the description.
Line 477. Not sure what the mass density means here as it has not been used to this point.
Figure 5. Are these zonal means? On what domain?
Line 517. Remove “in an altitude”.
Line 531. This separation is by design and may not occur in real clouds.
Line 537. Correct “cloude”.
Line 539. Repeated “are”.
Line 544. Correct the units.
Line 551. Please elaborate on this, not sure why there would be an increased number of collisions.
Line 557. What bias?
Line 581. Use “concentration”
Line 589. Correct “then”
Section 4: please add more discussion on the highly idealized nature of the setup and its implications (see general comment).
Line 648-650. What is the ground truth here?
Citation: https://doi.org/10.5194/egusphere-2024-2157-RC2 - AC2: 'Reply on RC2', Tim Lüttmer, 28 Nov 2024
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-2157', Anonymous Referee #1, 08 Oct 2024
This manuscript provides a detailed overview of the development of a new bulk microphysics scheme that includes several classes of ice hydrometeor, as opposed to the typical two classes in older schemes. It also includes the evaluation of a test case in an idealized convective cloud that clearly shows how differences between ice microphysical assumptions impact key properties/processes within a cloud. The further development of ice microphysical schemes is important to better understand how the ice phase influences key cloud properties. However, that message does not come across clearly enough in this work, with the sparse amount of . Similarly, I recommend that the authors undergo more rounds of revision and proofreading before this manuscript is ready for publication. I have included specific comments/questions/suggestions below to aid in the revision.
Introduction
- Here is where I found that the discussion did not cover enough of the work to date to justify the need for a new multi-class scheme, with too little literature cited overall. A couple background topics to consider covering: What is it about two class schemes that make them unsuitable to simulate ice microphysics? What have others done with multi-class schemes and what have they found (consider reviewing the work on the P3 scheme by Morrison and Milbrandt, 2015 and others).
Model description
- Lines 142-149: liquid-origin and in situ were already defined in the introduction so I don’t think you need to redefine them here
- Line 166: why does the saturation adjustment mean that the direct numerical integration of the homogeneous freezing coefficient?
- Lines 243-245: the fact that the model does not include an explicit aerosol microphysics model for simulating CCN and INPs, this could be considered for future work
- Figure 2: please readjust panel a such that xtick labels are not overlapping
- Section 2.3, consider renaming as the contents cover depositional growth of ice particles so the title may confuse readers
- Section 2.2.4: consider stating that the differences between these heterogeneous nucleation parameterizations are tested below for an idealized convective case because of the implications on the ice phase in mixed-phase clouds (as stated in the text)
Idealised simulations
- The authors should make it clearer in the text that they tested the new multi-class ice scheme with the three heterogeneous nucleation options by HA, UL, and PH (with the additional PH3) as detailed in Section 2.
- Please make clearer that the REF scheme refers to the SB scheme with only two ice modes
- It is typical to explicitly state the model version, resolution, and domain (both space and time) up front in a description of experiments, or in the methods section before this. For example, there are several versions of the ICON model, so it should be made clear that the authors ran it in LES mode following Heinze et al. (2017). What other standard configurations are worth noting for your model version? This helps with reproducibility of your results.
- What is the Weisman-Klemp setup? Is it the prescribed temperature and relative humidity profile? Why this set up over others (if available)?
- Why is the comparison made only between REF and HA at the start of Section 3.1?
- In section 3.1 it might be worth noting that you discuss the other panels in figure 3 later on
- Line 511, sentence starting with “However, it is heavier…”. Does the “it” refer to ice, as in total ice mass? Maybe consider making this clearer.
- Figures 5, 6, and 7 are generally great. Where did the 0.1 mg/m3 threshold come from?
- Lines 543-544 and Figure 7: “There is significantly more snow present in the ice mode simulation than in the reference simulation.” That does not seem to be the case looking at figure 7. Can the authors explain what they mean more clearly?
- Line 547, is this sentence referring to the HA scheme?
- Line 588: please state which figure again, and I think the authors might mean panels c and g instead?
Discussion
- Are there any limitations with this new scheme that weren’t addressed by this study that could be considered in future work?
General
- Several instances of using “an” where there should have been and “a” and vice versa
- Several acronyms weren’t defined before their use or the longname was used after the acronym was defined, check for consistency
- Are there any observations of precipitation rates that the authors could compare to their new model setup as a test of their new scheme?
Citation: https://doi.org/10.5194/egusphere-2024-2157-RC1 - AC1: 'Reply on RC1', Tim Lüttmer, 28 Nov 2024
-
RC2: 'Comment on egusphere-2024-2157', Anonymous Referee #2, 18 Oct 2024
This work introduces new ice classes into an existing cloud microphysics scheme to better elucidate the role of different ice formation mechanisms on the cloud evolution. Since the role of the ice phase on the cloud evolution is still fraught with uncertainty this work and its conclusions are of significance to the scientific community. It is also well written, however some issues must be addressed before publication, as described below.
General comments.
The introduction is almost completely devoid of any citations, despite being full of overarching generalizations. It makes me wonder whether the authors conducted a proper literature review before writing the paper. Please add citations and put the work within the context of the current literature.
It is important to recognize that this is a very idealized model, that although illustrative it might have limited application in the modeling of real situations. The size distribution of the ice crystals is a single function and can’t be separated by origin as ice grows. It would be impossible to measure or even estimate the parameters of the distribution for the different ice classes. It is also not recommended to excessively add tracers to be advected by the host model since it may lead to numerical diffusion issues. For host models with lower resolution it would be difficult to relate the different classes to macro-scale variables like cloud fraction and total condensate. Finally, the INP classes used in the work are somehow artificially separated by design while in real cloud they tend to be active at the same time. These considerations must be made clear in the discussion section of the work.
Specific comments.
Line 19. Clouds are uncertain in models not in the Earth, please rephrase.
Lines 24, 29, 44, 45, 54, 55 and many other places. These statements are not obvious and need references backing them up.
Line 52. Cloud types.
Line 59. Remove “will”.
Line 70. Please rewrite this sentence in clearer terms.
Line 74. What is f? Should it be normalized to the total mass instead?
Line 76. Define “the particle mass distribution”.
Line 104. Remove “Actually”.
Line 125. This needs a reference as well.
Line 135. Remove “huge”.
Line 140. Please clarify what this means.
Line 154. Should contact ice nucleation be included?
Line 200. Since competition between homogeneous and heterogeneous nucleation is a significant feature of cloud ice formation, it is not clear how this approach would represent real clouds.
Line 210. This expression needs to be weighted by the size distribution of the aqueous solution droplets. In fact, is nhom limited at all by the available droplets?
Line 244. This is a strong assumption that makes this a very idealized model. Please clarify how it might affect the results.
Line 269. This only works for INP immersed within cloud droplets. It is not clear whether it is applied that way.
Line 283. There is an issue here since “na” for deposition and immersion is different (i.e., whether they are inside/outside of cloud droplets).
Line 344. Why is this not dependent on the mass of other cloud species?
Liner 440. How many new tracers are added to the dynamics scheme? What is the domain of the simulation?
Figure 3: Are these lines averaged over the whole domain?
Line 477, 497, 498 and other places. Using words like “weak”, “stronger”, “catches up” is ambiguous. Be precise on the description.
Line 477. Not sure what the mass density means here as it has not been used to this point.
Figure 5. Are these zonal means? On what domain?
Line 517. Remove “in an altitude”.
Line 531. This separation is by design and may not occur in real clouds.
Line 537. Correct “cloude”.
Line 539. Repeated “are”.
Line 544. Correct the units.
Line 551. Please elaborate on this, not sure why there would be an increased number of collisions.
Line 557. What bias?
Line 581. Use “concentration”
Line 589. Correct “then”
Section 4: please add more discussion on the highly idealized nature of the setup and its implications (see general comment).
Line 648-650. What is the ground truth here?
Citation: https://doi.org/10.5194/egusphere-2024-2157-RC2 - AC2: 'Reply on RC2', Tim Lüttmer, 28 Nov 2024
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 356 | 146 | 142 | 644 | 19 | 18 |
- HTML: 356
- PDF: 146
- XML: 142
- Total: 644
- BibTeX: 19
- EndNote: 18
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Tim Lüttmer
Peter Spichtinger
Axel Seifert
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(4643 KB) - Metadata XML