the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 26 Jun 2025
Secondary Ice Formation in Cumulus Congestus Clouds: Insights from Observations and Aerosol-Aware Large-Eddy Simulations
Abstract. Secondary ice production (SIP) was investigated in a cumulus congestus system observed during the Secondary Production of Ice in Cumulus Experiment (SPICULE) campaign. Large-eddy simulations were performed using UCLALES–SALSA, a model that explicitly resolves aerosol–hydrometeor interactions through a sectional representation of aerosols, cloud droplets, rain droplets, and ice crystals. Two scenarios were compared: one including only immersion freezing as an ice formation process, and another incorporating additional SIP mechanisms – namely droplet shattering, rime splintering, and ice–ice collisional breakup.
The SIP-inclusive simulation reproduced the evolution of the observed cloud microphysical structure in both warm and mixed-phase regions. Ice–ice collisional breakup generated substantially more secondary ice particles than droplet shattering; however, it was only initiated after droplet shattering provided a sufficient initial ice particle population to meet the SIP triggering conditions. Droplet shattering was triggered by the presence of large supercooled droplets, defined by an integral raindrop diameter exceeding 3.5 mm L-1 at temperatures below 265 K. Once formed, secondary ice particles enhanced riming and accretion, leading to auto-catalytic amplification of SIP through ice–ice breakup. This feedback rapidly depleted cloud liquid water within approximately 10 min.
Enhanced updrafts were identified in SIP-active regions, suggesting invigoration in the upper mixed-phase levels. SIP also intensified precipitation via the ice phase, resulting in a 26 % increase in domain-mean cumulative precipitation. The simulations reproduced key aspects of the observed ice multiplication, supporting the adequacy of the SIP representation in the model framework.
- Preprint
(12998 KB) - Metadata XML
-
Supplement
(9889 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-2730', Anonymous Referee #1, 24 Jul 2025
Overall comments:
Calderón et al. presents a detailed analysis of the process-level evolution of ice microphysics in a developing cumulus congestus cloud. They provide airborne in situ measurements from the SPICULE campaign and utilize the sophisticated UCLALES-SALSA model to simulate various secondary ice production (SIP) mechanisms, including rime-splintering, droplet shattering, and ice-ice collisional breakup. The authors find that the simulated cloud with the inclusion of SIP produced higher total water content and taller cloud top heights. They also find that there may be SIP-induced invigoration related to positive buoyancy from water phase changes.
The authors do an excellent job showing the evolution of the cumulus congestus cloud and the influence of SIP on the droplet size distributions and accumulated precipitation through the cloud’s lifecycle. This paper also provides new insights that expand upon recent modelling and observational studies highlighting the chain of events of ice microphysics prior to cloud glaciation.
This paper is suitable for publication after the following minor comments have been addressed.
Minor comments:
Line 28: “…in situ vertical profiling”, I think you mean sampling at different altitudes. Vertical profiling is not possible with aircraft in convective clouds.
Line 29-34: Patnaude et al. (2025) also provided evidence of fragmented droplets from a CPI. They also used in situ INP measurements to demonstrate the presence of SIP.
Patnaude, R. J., and Coauthors, 2025: On the Role of Airborne Ice Nucleating Particles in Primary and Secondary Ice Formation Processes in Convective Midlatitude Clouds. J. Atmos. Sci., 82, 869–892, https://doi.org/10.1175/JAS-D-24-0135.1.
Line 39: Korolev and Leisner 2020 is a review paper. It would be better to cite the specific laboratory studies you may be referring to.
Line 138 – 145: Does this scheme allow for heterogeneous freezing of cloud droplets only? Does it also include heterogeneous freezing of raindrops?
Line 152: “… including both modes (i.e. collision of drop with smaller crystal…”: If you are referring to the two modes of DS from Phillips et al. (2018) I would recommend citing this here since most studies do not consider multiple modes (i.e., Patnaude et al. 2025; Sullivan 2018, Grzegorczyk et al. 2025a). The more agreed upon representation and observed mechanism for DS is the fragmentation of the droplet during the primary freezing process (see Keinert et al. 2020), which is not mentioned here.
Grzegorczyk, P., Wobrock, W., Canzi, A., Niquet, L., Tridon, F., and Planche, C.: Investigating secondary ice production in a deep convective cloud with a 3D bin microphysics model: Part I - Sensitivity study of microphysical processes representations, Atmospheric Research, 313, 107774, https://doi.org/10.1016/j.atmosres.2024.107774, 2025.
Keinert, A., Spannagel, D., Leisner, T., and Kiselev, A.: Secondary Ice Production upon Freezing of Freely Falling Drizzle Droplets, Journal of the Atmospheric Sciences, 77, 2959–2967, https://doi.org/10.1175/jas-d-20-0081.1, 2020.
Sullivan, S. C., Barthlott, C., Crosier, J., Zhukov, I., Nenes, A., and Hoose, C.: The effect of secondary ice production parameterization on the simulation of a cold frontal rainband, Atmos. Chem. Phys., 18, 16461–16480, https://doi.org/10.5194/acp-18-16461-2018, 2018.
Line 174-179: “It is particularly useful…” Okay I see the fragmentation of droplets is mentioned here. I think that this should be mentioned earlier and perhaps in the introduction. It is also not clear if this mechanism of DS was used in the simulations. Later you mention that in the SIP-ON simulations, you include mode 1 and 2 of Phillips (2018) but in table S2, there are other parameterizations listed (Lawson 2015, Sullivan 2018) that would account for this representation of DS. Please clarify.
Line 192 – 200: A kappa value of 0.54 is quite low. Do you have a sense of how much this changes the wet diameter? Did you consider using the clear-sky CDP for measurement of supermicron aerosols?
Line 238 – 247: Patnaude et al. (2025) also showed evidence of DS and CPI images of fragmented frozen droplets in fresh updrafts from RF06 of the SPICULE campaign.
Line 263: “.. too low to reproduce observed ice microphysics despite…” Could you be clear on what you mean here? Too few ice crystals? The INP measurements from RF04 were anomalously high compared to the rest of the campaign (see Patnaude et al. 2025 Figure 4). It may be at there were cases that the INP and ice crystal number concentrations were much lower, and SIP was still occurring as was shown in Patnaude et al. (2025).
Line 264 – 265: “observed ice number concentrations.” Is this in reference to the one cloud penetration from Lawson (2023) with > 2000 L ice?
Line 295 – 305: Similar to my comment above, it would be helpful to provide more context to the GV and Learjet observations. I assume it is the measurements from Lawson 2023 Table 3 and 4?
Line 301: “… they increased at higher altitudes…” What is meant by “they”? LWC or the differences?
Line 303 – 305: It is not clear to me which simulation is being referred to for “lower LWCs in warm cloud…” Please be more specific here.
Line 324 – 326: “we considered that modeled ice…” In the SIP-OFF simulation, does the model not allow for any interactions between liquid and ice hydrometeors? Meaning there would be no freezing of rain droplets via a collision with an ice particle? That would be another source of ice particles besides heterogeneous nucleation of INPs.
Line 361 – 362: Also agrees with Patnaude et al. (2025) who showed fragmented droplets at -17C during RF06 of SPICULE.
Line 374 – 375: This sentence is a bit confusing as it reads like Figure 7g-I is showing concentrations of large droplets, when I think the point is that the other SIP mechanisms are constrained to smaller areas.
Line 375 – 377: Same as previous comment. I think you are making the point that Figure 8c-d is showing the limited existence of liquid water, but it reads like you are stating that those figures are showing strong updrafts carrying liquid water upwards.
Line 380: Below freezing level sounds like you are saying colder, when I think you are stating the opposite. I would suggest saying “higher” or “warmer”, or “lower altitudes”.
Line 499 – 502: It is difficult to discern in Figure 12 where 40 and 50 minutes are occurring. It may be helpful to highlight those lines on the figure so the reader can more easily observe the DSD.
Line 503-504: “The larger the supercooled droplet…” This is true but I think this statement ignores the fact that the II-BR SIP mechanism had already completely taken over as the dominant SIP mechanism by 48 minutes. II-BR mechanism is likely consuming many of the large droplets not DS. Also, in laboratory studies of droplet fragmentation, Keinert et al. (2020) found that at most ~45 % of the large droplets underwent any kind of breakup, likely not enough to significantly deplete the raindrops.
Technical comments:
Line 212: Add a space after the degrees symbol and before “to”
Line 261: AOP should be OAP.
Line 262: “…SIP mechanism of droplet shattering and other adding the mechanism of rime splintering.” Please revise.
Figure 3. It is quite difficult to see the differences in the model percentile lines. I would recommend using more different line styles or widths for each to make it more obvious.
Figure 5. It appears that this figure is missing a legend.
Figure 6. I do not see a line for 258.15K.
Line 434: Do you mean Figure 10 e-f?
Line 448: I do not think this reference is needed here.
Line 514: Change droplet size distribution to DSD
Line 541 – 543: missing a second parentheses.
Citation: https://doi.org/10.5194/egusphere-2025-2730-RC1 -
RC2: 'Comment on egusphere-2025-2730', Anonymous Referee #2, 22 Aug 2025
Reviewer comments on the manuscript egusphere-2025-2730 by Calderon et al.
The manuscript is undoubtedly a work of very high degree of complexity, so it was a challenge for me (not a modeler) to follow the discussion of various processes and feedback mechanisms associated with secondary ice formation. Nevertheless, I feel that this is exactly the type of modelling study needed to bridge the gaps between the patchy snapshots obtained via airborne in-situ observations, lab studies conducted under poorly-defined conditions, and simplified cloud modelling where parameterizations of SIP rates are based on incomplete lab data and wild extrapolations. I therefore recommend publication of the manuscript after several minor issues have been addressed.
1. I have a general comment on the choice of parameterization of h(T) for DS mechanism: since (Leisner 2014) there was a number of experimental studies of DS mechanism (Keinert et al., 2020; Kleinheins et al., 2021; Lauber et al., 2018) that, even if not covering the whole parameter space, have specifically addressed more realistic freezing conditions. These studies have demonstrated, that in comparison to parameterizations based on (Lawson et al., 2015) and later used by (Sullivan et al., 2018), the number of secondary ice particles produced by a single freezing droplet should be enhanced for free fall conditions (Keinert et al., 2020) and could be even higher by factor 4 to 7 if considering the pressure release events as indicators for secondary ice production (Kleinheins et al., 2021). So maybe the parameterization used here is not the greatest choice.2. Lines 163-166: “In this study, we keep the conventional assumption that the rime splintering mechanism generates single-size ice crystals in the shape of hexagonal columns with a density of 917 gm−3 and a maximum length of 10 μm … (Buhl et al., 2019)).
(Bühl et al., 2019) does not discuss the size and shape of ice splinters produced in rime-splintering mechanism. I am also not aware of any other study citing “conventional assumption” that such crystals are generated in an RS SIP. In fact, from the mechanistic point of view this is not very plausible assumption, since RS is supposed to produce secondary ice upon riming of ice by droplets that are themselves in the size range between 10 µm and 30 µm. This would require splintering of the whole frozen droplets or quite large fragments of the frozen droplets upon collision with ice. Neither that nor freezing of a supercooled droplets upon glancing contact with the ice, another mechanism that would produce secondary ice in this size range, has not been observed in the recent experimental study of (Seidel et al., 2024).3. Line 164: The density of “917 gm^-3” doesn’t make any sense. It should be kg x m^-3 or 0.917 g x cm^-3.
4. Lines 165-166: “Instead, we assume that the mechanisms of droplet shattering and ice–ice collisional breakup generate ice crystals that can range in size from 2 μm to nearly as large as the fragmenting hydrometeor (i.e. supercooled droplet or ice crystal).”
Why “instead”? This sentence relates to a different mechanism, doesn’t it?5. Lines 167-169: “We distribute the total number of secondary ice particles NSIP (T,Dl,Dm) produced by a collision between two hydrometeors of size Dl and Dm between size bins smaller than the fragmenting hydrometeor in such a way that each bin gains the same amount of mass, similar to Lawson et al. (2015).”
(Lawson et al., 2015) distributes the number of crystals evenly across all the bins, not the mass: The secondary ice produced is distributed evenly over a range of ice size bins that are smaller than the diameter of the original frozen drop (their page 2442). Also, (Lawson et al., 2015) considers exclusively droplet shattering (DS) as the sole SIP mechanism in their study, the RS and the IIBR are not discussed.6. Section 2.2, Table 1 and discussion thereof: It is difficult to follow the comparison between various descriptions of size and number distribution of secondary ice particles. Why don’t you give the functional form of these distributions for all three SIP mechanisms as functions of hydrometeors’ size, ambient parameters, and factors affecting collision events?
7. Figure 5 seems to have no legend, which makes it impossible to follow the discussion.
8. Figure 6: I don’t see three “…Continuous black lines indicate altitudes at which temperatures are 273.15 K, 265.15 K and 258.15 K”. There are only two black lines per panel, I assume the coldest one is missing?
9. Figure 7: why are the SIP-RS rates in the temperature range below -10°C (above the black line corresponding to 263.15 K in the panes d) and g) of the figure 7) non-zero? Should it not be strictly allocated to the temperature range between -3°C and -8°C (as indicated by parameterization in the table 1)? Same questions regarding Figure 11, panel a).
10. The difference in color in the color-coded lines in the Figure 12 panels a) and b) is hardly visible. Maybe use a different color palette?
Keinert, A., Spannagel, D., Leisner, T., and Kiselev, A.: Secondary Ice Production upon Freezing of Freely Falling Drizzle Droplets, Journal of the Atmospheric Sciences, 77, 2959-2967, 10.1175/jas-d-20-0081.1, 2020.
Kleinheins, J., Kiselev, A., Keinert, A., Kind, M., and Leisner, T.: Thermal Imaging of Freezing Drizzle Droplets: Pressure Release Events as a Source of Secondary Ice Particles, Journal of the Atmospheric Sciences, 78, 1703-1713, 10.1175/JAS-D-20-0323.1, 2021.
Lauber, A., Kiselev, A., Pander, T., Handmann, P. V. K., and Leisner, T.: Secondary Ice Formation during Freezing of Levitated Droplets, Journal of the Atmospheric Sciences, 75, 2815-2826, 10.1175/jas-d-18-0052.1Lawson, R. P., Woods, S., and Morrison, H.: The Microphysics of Ice and Precipitation Development in Tropical Cumulus Clouds, Journal of the Atmospheric Sciences, 72, 2429-2445, 10.1175/jas-d-14-0274.1, 2015.
Sullivan, S. C., Hoose, C., Kiselev, A., Leisner, T., and Nenes, A.: Initiation of secondary ice production in clouds, Atmos. Chem. Phys., 18, 1593-1610, 10.5194/acp-18-1593-2018, 2018.
Seidel, J. S., Kiselev, A. A., Keinert, A., Stratmann, F., Leisner, T., and Hartmann, S.: Secondary ice production – no evidence of efficient rime-splintering mechanism, Atmos. Chem. Phys., 24, 5247-5263, 10.5194/acp-24-5247-2024, 2024.Citation: https://doi.org/10.5194/egusphere-2025-2730-RC2
Data sets
UCLALES-SALSA Simulation Data for the SPICULE-RF04b Cloud Case from "Secondary Ice Formation in Cumulus Congestus Clouds: Insights from Observations and Aerosol-Aware Large-Eddy Simulations" Silvia Margarita Calderón et al. https://fmi.b2share.csc.fi/records/50ab4fa1e7b94909b29e31c9ef0e8618
Model code and software
UCLALES-SALSA: large-eddy-simulations with aerosol-cloud-ice-precipitation interactions Silvia Margarita Calderón et al. https://zenodo.org/records/15179737
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 422 | 49 | 15 | 486 | 28 | 13 | 25 |
- HTML: 422
- PDF: 49
- XML: 15
- Total: 486
- Supplement: 28
- BibTeX: 13
- EndNote: 25
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1