The paper uses (idealized) numerical simulations of trajectories during DCMEX to study effects of entrainment on ice multiplication activity in deep convective clouds (DCCs). The study investigates different secondary ice production processes: rime-splintering, spherical droplet shattering (mode 1), droplet shattering upon collision with larger ice particle (mode 2) and breakup upon ice-ice collision, using University of Manchester bin microphysics parcel model. The study claims the first systematic attempt to investigate the impact of homogeneous, and inhomogeneous entrainment on SIP. Multiple perturbation simulations, from 15 representative DCMEX cases, have been performed to analyze the impact of entrainment on cloud properties such as LWC, effective particle size, CDNC, and ice number concentrations. Comparison of the simulated microphysical properties with three observed days of DCMEX is presented. I found the topic timely and relevant, aligned with current community efforts to better represent and quantify the impact of SIP in atmospheric models. It also sufficiently discusses the recent developments and uncertainties in SIP research. However, I believe the manuscript could benefit from further clarification, and the major/minor comments below are intended to help the authors improve the overall clarity and organization.

Line 31: Recent experiments Seidel et al. should be Recent experiment by Seidel et al.

Lines 34-36: I found these lines quite misleading as if other SIP processes are “perhaps as uncertain”, what this study adds beyond considering other SIP processes over RS to explain ice enhancement in DCCs? These lines need to be re-written as it currently weakens the motivation of the study.

Line 34: change Ice-ice to ice-ice?

Line 35: parametrisations should be parameterisations

Line 42: What is HAIC/HIWC?

Line 43: As already defined above, ice-ice collisional breakup can be written as CB.

For the description of SIP processes in the introduction, the limitations reported in recent studies are described quite well. However, the physical mechanisms underlying these processes are not well described (also not in Sec. 3.3 Secondary ice parameterisations). What is crucial is to discuss how these operate at the microphysical levels (how fragment production in each SIP process depend on temperature, vertical velocity, particle size and mass, etc.). This is important to improve the scientific clarity of the manuscript.

Line 60: While the acronym DCMEX is introduced in the abstract, it is not defined anywhere in the introduction. Abbreviations should be defined at their first occurrence in the main text, as the abstract is considered standalone.

Lines 61-65: These lines can be best fitted in Sec. 2 DCMEX.

Lines 70-71: “While some …”. To my understanding, the activity of SIP processes varies significantly over the lifecycle of DCCs, can the author comment on the growth phase of DCCs they are comparing between these studies.

Lines 70-78: It is clear from the introduction that SIP processes are uncertain and that previous numerical studies have produced contrasting results regarding the dominant mechanisms. The study attempts to bring novelty by highlighting "entrainment impact on SIP" in the title, which is a valid and interesting point. However, this closing paragraph fails to deliver a convincing motivation for this specific novelty. The sentences do not connect logically, and the reader may be unable to follow the reasoning from the identified gaps in the literature to the scientific objectives of the present study. More critically, the study investigates three distinct entrainment regimes, adiabatic, homogeneous, and inhomogeneous, yet none of these are introduced or discussed anywhere in the introduction. I think the reader needs to understand what these entrainment types represent physically, how they differ and more importantly, why they are important for studies involving SIP.

Line 77: dependent should be depend?

Line 88: BAE should be BAe?

Line 89: I wonder if this “Kite-shaped” pattern was for a specific reason.

Line 96: What is  here?

Line 99: DMT?

Line 99: In CDP-2, 2 for CDP version?

Line 100: Space between (CDNC) and (Lance et al. 2010b).

Line 102: SPEC?

Lines 141-142: The authors should clarify what they mean by "zero of the Köhler equation"? Is this the critical supersaturation, the equilibrium radius at a given supersaturation, or something else?

It is also not clear how primary ice is activated in the current setup of BMM.

Line 139: The hygroscopicity parameter is kept constant across the bins. Can the authors justify this simplification and discuss how it may affect CCN activation and the resulting droplet size distribution? The mode 2 SIP process can be highly sensitive to the drop size thus any bias introduced in droplet size distribution could affect mode 2 activity.

Also, the use of SMPS is mentioned earlier in Sec. 2.1 (also in Fig. S1), I wonder if the modeled size distribution is constrained to the observed size distribution (e.g., with Fig. S1).

Sec. 3.1: In general, this section should aim to connect each technical choice back to its scientific implications for the results, rather than presenting the model formulation in isolation.

Lines 169, 173, 177-179: , μ d/dt are defined multiple times.

Line 188: 10,s should be 10 s?

Sec. 3.3: The current description of SIP parameterizations in the BMM is insufficient. The manuscript repeatedly refers the reader to previous studies for the mathematical representation of the SIP processes considered, but this is not acceptable as a substitute for a self-contained description in the present paper. Although the literature describing derivations of these SIP processes cited here, the authors should, at minimum, present the key equations governing each SIP mechanism, along with a clear explanation of how each process is implemented in the BMM. This includes their bin representation, the favorable conditions, the fragmentation rates, the temperature dependencies, rime-fraction (for CB from Phillips et al. 2017a), and any threshold parameters used. Without this, it is impossible for the reader to critically evaluate the model results or assess whether the implementation is physically reasonable. A paper whose central scientific contribution is the comparison of SIP parameterizations under different entrainment conditions must provide a transparent and complete description of those parameterizations within the manuscript itself. The same holds true for heterogeneous ice nucleation parameterizations.

Line 203-208: The authors treat mode 2 as a fourth independent SIP mechanism, but in Phillips et al. (2018), mode 1 and mode 2 are both represent a single process called fragmentation of freezing drops (or drop shattering, DS). The authors should justify why mode 2 is classified as a separate mechanism rather than as a part of DS, or restructure their classification accordingly to remain consistent with the original Phillips et al. (2018) framework.

Lines 113-114:  is defined multiple times earlier in the manuscript.

From Table 1, There is an inconsistency between Section 2.2, where only three flight cases (C300, C303, C309) are stated to be considered in this study, and the table presented here, whose caption suggests that all 15 listed flights are considered. The authors should clarify which flight cases are actually used in the simulations and ensure consistency between the text, table, and caption throughout the manuscript.

Line 234: D_eff  is not defined earlier?

Lines 235-236: Just a thought. Ice enhancement can also be defined as N_ICE/N_INP which I believe can be a more standard way of defining it (See also Hobbs et al. 1980, Waman et al. 2022, Han et al. 2024, etc).

Lines 243, 245: contributions of individual SIP mechanisms to what?

Line 246: I guess LWC, CDNC are defined earlier?

Lines 249-250: This sentence seems incomplete as it is not clear which simulation refer to these values of LWC?

Lines 251-252: Interesting, can this (300% more LWC) be also highlighted in the abstract?

Line 261: ADIA simulations instead of adiabatic (ADIA) simulations?

Lines 261-285: The description in these paragraphs is difficult to follow without clear references to the relevant figures. The authors should ensure that Fig. 1 and any other supporting figures are explicitly cited at appropriate points throughout these paragraphs to guide the reader. This applies to almost all the paragraphs in the result section.

Lines 262-264: The authors refer to a decrease in CDNC but it is not clear what this decrease is relative to.

Lines 274-275: Terms likes RA and EA are repeatedly defined.

Lines 285: At this point, the reader may lose track of the flow of information as multiple cases are being discussed simultaneously. Can the authors explicitly attribute the peak D_eff to each of the three cases.

Line 287: representative cases to representative DCMEX cases?

Lines 291-292: For dispersion, proper figure should be cited here.

A suggestion: Can the authors describe briefly the term “dispersion” in the methodology?

Lines 292-293: The term “instrumental effects” introduced suddenly without prior discussion and dropped without any follow-up. Also, instrumental effects “by” Lance et al. (2010a).

The phrase "three cases" is used ambiguously throughout the manuscript. At times it refers to the three selected DCMEX flight cases (C300, C303, C309), and at other times it appears to refer to the three entrainment scenarios used in the ice-phase simulations (ADIA, HOM+EA, and INHOM+EA+RA). This inconsistency makes it difficult to follow the text without repeatedly cross-referencing earlier sections. Can the authors clarify this in the text?

Lines 302-304: The claim that HOM+EA cannot explain the observed dispersion "during the early stages of cloud development" is unclear (if it is in the context of Figure 3, which shows vertical profiles against temperature rather than temporal evolution). Can the authors clarify what they mean by early stages, and if necessary, support this claim with a time series or clearer reference to the relevant figure panels?

Lines 65, 98, 306 (Wu et al. 2025): This article appears to be listed in the reference list as "in preparation" and does not appear to be available as a preprint. Citing unpublished and unavailable manuscripts makes it impossible for the reader to verify the claims supported by this reference.

Line 319-321: Again, no clear mathematical formulation is given in the manuscript to describe heterogeneous freezing by Daily et al. (2025) and DeMott et al. (2010).

Line 320: ICNC is used without defining it earlier (the same at other places).

Lines 319-322: I do not see which figure is actually referred here (Fig. 8?).

I noticed a recurring structural issue throughout the results section that paragraphs tend to open with a main finding or conclusion without first pointing to the figure that supports it (e.g., lines 323-330). This makes it harder to follow the argument, as readers encounter the conclusion before understanding what evidence it is based on. I would encourage the authors to pay closer attention to this throughout the manuscript, ideally introducing the relevant figure and briefly describing its contents before stating the result drawn from it.

Line 331: (As said earlier) kindly refer to a figure supporting this statement.

Figs. 5-7: The ice number concentrations derived from the CPI are consistently between 1–100 L⁻¹. The FAAM BAe-146 is also equipped with other probes such as the CIP and 2D-S that made observations of ice particles. By plotting data from all three probes, can the authors confirm whether these probes reported similar ice number concentrations, or whether there are discrepancies between the instruments? If differences in the measurement, which probe is reliable and why? Also, what conditions are imposed on the simulated and observed ice number concentration shown in these plots. In the current plots, how the CPI data is filtered, for example, to avoid possible bias from artificial shattering of ice particle upon impact?

Figs. 5-7: In these figures, particularly Figs 6-7, the general claim is that including SIP in the simulations explains the observed ice number concentrations. However, I do not see this true for most of the vertical levels, particularly levels warmer than -20°C (for C303 flight) and -10°C (for C309 flight). Although RS is included, which is active at much warmer subzero levels, for what reasons the observed ice concentrations is consistently higher than simulated? Also, the current x-axes range in all these plots somewhat hinders the information of ice numbers from simulations for lower values.

Lines 345-346: No figure is cited to support this ice enhancement from RS (the same is true for other occurrences). Also, as ice enhancement is used to define SIP activity (also lines 353-354), which is the main focus of this study, I would suggest to bring these supplementary figures into the main text and discuss them thoroughly.

Fig. 5 (right panel) is not discussed in the main text?

Line 355: Please mention temperature -22°C in the main text. Additionally, the motivation for presenting the CPI imagery alongside the ice concentration profiles is not clearly stated, are the two panels intended to be interpreted together, or is the CPI imagery provided purely for illustrative purposes? If purely for illustrative purpose, I would recommend to show these imageries in the case description section for all flights (right panels of Figs. 5, 6, 7) and refer them in the main text whenever needed. Also, the authors claimed that the hexagonal shape visible in CPI imagery in Fig. 5 (at -22°C) is likely from ice-ice collision.

Lines 356-357: Conceptual clarification: the phrase “fragmentation between supercooled droplets and more massive ice particles” should be revised to “fragmentation during the freezing of supercooled droplets upon collision with a more massive ice particle,” which more accurately reflects the physical mechanism involved. Additionally, as said before, the particle images contain substantial microphysical information and therefore should be described in detail within the case description section.

Line 358: Again, a statement is made without referring to proper figure. As well, kindly consider my previous comment on ice enhancement figures (moving them to and citing them properly in the main text).

Lines 364: Following this, the wording at line 9 in the current version of the abstract can be misleading as it can simply mean three entrainment conditions. But the main text considers two entrainment representations. Please correct the text for better clarity and consistency. I am unsure whether the authors also consider adiabatic case as a type of entrainment?

Lines 368-369: What are the shallower cases here?

Lines 369-370: I wonder whether the CPI imagery alone is sufficient to support this conclusion. While the images demonstrate the presence and morphology of hydrometeors, it is not clear how they can be used to infer that collisions between large supercooled droplets and small ice crystals were rare. This statement, in its current form, appears to go beyond what can be directly inferred from the particle images. Clarification on how this conclusion was derived would be helpful.

Line 391: LES?

Line 395: secondary ice production can be written as SIP?

Line 396: C306 and C312 are suddenly introduced here without supporting figure/discussion in the main text/supplement.

Line 427: What is ICE-T? Similarly, HAIC/HIWC at other places in the manuscript?

References

Hobbs, P.V., Politovich, M.K. and Radke, L.F., 1980. The structures of summer convective clouds in eastern Montana. I: Natural clouds. Journal of Applied Meteorology and Climatology, 19(6), pp.645-663.

Waman, D., Patade, S., Jadav, A., Deshmukh, A., Gupta, A.K., Phillips, V.T., Bansemer, A. and DeMott, P.J., 2022. Dependencies of four mechanisms of secondary ice production on cloud-top temperature in a continental convective storm. Journal of the Atmospheric Sciences, 79(12), pp.3375-3404.

Han, C., Hoose, C. and Dürlich, V., 2024. Secondary ice production in simulated deep convective clouds: A sensitivity study. Journal of the Atmospheric Sciences, 81(5), pp.903-921.

Review of the manuscript titled “Impacts of entrainment on secondary ice production in deep convective clouds” by Bowen et al.

This study investigates the importance of different secondary ice production (SIP) processes, including rime splintering (RS), ice–ice collisions (CB), fragmentation during spherical freezing of drops (M1), and fragmentation resulting from collisions between supercooled droplets and larger ice particles (M2), using both model simulations and aircraft campaign observations. The University of Manchester bin-microphysics parcel model (BMM) was employed to conduct the simulations, together with in situ cloud microphysics and aerosol measurements collected during the Deep Convective Microphysics EXperiment (DCMEX) field campaign. One focus of the study is the impact of different entrainment-mixing scenarios on cloud droplet formation and the subsequent influence on SIP processes. The results show that droplet size distributions are better simulated when entrainment is included. In most cases, homogeneous mixing combined with aerosol entrainment provides the best agreement with observed ice-cloud microphysical properties, and no evidence was found that inhomogeneous mixing dominates in the simulated cases. Regarding the relative importance of the different SIP mechanisms, M2 appears to be the dominant process. However, uncertainties remain in the simulation of the other SIP processes due to the limitations of the idealized parcel-model framework. Overall, this study is timely and provides a valuable contribution to the cloud microphysics community, and I have the following suggestions:

General:

In the results shown in Figures 5–7, the simulated ice properties are comparable to the observations, but only for the upper portion of the clouds. The authors suggest that the absence of simulated ice in the lower part of the cloud is due to the lack of vertical transport processes in the parcel model. However, could other processes, such as lateral mixing or turbulent mixing, also contribute to this discrepancy? More importantly, it would be helpful to clarify how the absence of sedimentation and vertical transport in the parcel model may affect the simulated microphysical properties in the upper portion of the cloud.

The authors discuss the uncertainties associated with the parameterizations of the different SIP processes, as well as the limitations arising from the use of a parcel model in representing these processes. The results suggest that DS M2 is the most active SIP mechanism for the DCMEX cases. However, given the large uncertainties associated with all investigated SIP processes, the conclusion may need to be stated more cautiously.

The description of the SIP processes in Section 3.3 could be expanded, particularly for CB and DS. In addition, M1 and M2 are treated here as two separate mechanisms, whereas they are initially introduced as two modes of the DS mechanism. This distinction should be clarified for consistency.

Specific:

L20-21: Please also include this paper as one of the references: Ladino, L. A., Korolev, A., Heckman, I., Wolde, M., Fridlind, A. M., and Ackerman, A. S.: On the role of ice-nucleating aerosol in the formation of ice particles in tropical mesoscale convective systems, Geophysical research letters, 44, 1574–1582, https://doi.org/10.1002/2016GL072455, 2017.

L42: HAIC/HIWC is not explained.

L71: Qu et al., 2022 instead of 2020.

L96: κ is not explained here which appears later in L139.

L99: "DMT" is not explained.

L188: "10,s" should be 10 s.

Table 1: Please clarify what is meant by ‘neighbouring case’. Does this refer to cases with similar atmospheric conditions that are close in time or geographical location?

Figure S5: The color coding appears inconsistent with that used in Fig. S2. In Fig. S2, the ‘blue’ used for the simulations appears closer to purple, whereas in Fig. S5 both solid and dashed blue lines are shown, and the color is similar to that used for the observational dots in Fig. S2. Please ensure that a consistent color scheme is used across all figures.

Figure 3: The observational data shown as dots overlap substantially, making it difficult to discern the distribution from these plots. It may be more effective to present the results using boxplots, both for all cases and for the subset corresponding to the convective core. In the caption of Fig. 3, the order of CDNC and LWC appears to be reversed. In addition, the caption states that ‘Blue dots indicate observation,’ but the observational points appear in multiple colors. Please adjust this. Finally, many observational data points have values of zero. Could these be screened out if they correspond to clear-sky conditions?

L271-272: “average observed CDNC”, again, difficult to tell based on the overlapped dots.

Figure 3 & 4: The font is a little bit too small. Please increase it.

L320: “ICNC” is not defined.

L325: Do you mean Figure 8 here instead of Figure 9?

Figure 5: The orange solid line is difficult to distinguish from the red line, as the colors appear too similar.

L350: In “INHOM and INHOM+EA”, Should INHOM+EA be INHOM+RA?

L352: In “INHOM+EA+RA and INHOM+RA”, should INHOM+RA be INHOM+EA?

L354-357: Could the authors provide additional cases to support this claim? The current example shows only one frozen drop with an irregular shape. More evidence would be needed to substantiate the conclusion.

L369-370: “the collision of large supercooled drops and small ice particles are rare” Could the author provide more evidence to support this? The observation of small ice particles is largely uncertain. Are there any fractured frozen drops in the CPI images? Could the large irregular ice particle in Fig. 5 the remnant of a fractured large frozen drop? 

L414: It seems HM and BR were not explained at this point.

Figure S11-S13: Could the authors explain why a cut-off minimum droplet diameter appears in the ADIA case?