The manuscript computes radiative effects of ice-nucleating particles (INPs) including crystalline ammonium sulfate based on Beer et al. (2022). Although I do not quite understand why these radiative effects were not already reported in Beer et al. (2022), I do think that this is an interesting subject worthwhile another publication in ACP. The authors find that ammonium sulfate INPs contribute nearly as much to the total radiative effect as mineral dust and soot aerosol combined. Similar to Righi et al. (2021), they find that prescribing a fairly large globally constant vertical velocity can result in a larger effect. Potential reasons for the larger difference compared to Righi et al. (2021) should, however, be discussed. Regarding the INP seeding section, I am not sure about the added value compared to previous studies. I suggest that in order to address this point, the authors include a sensitivity study in which they consider ammonium sulfate, glassy organics, soot, and dust as background INPs to Sect. 3.5, and focus their discussion on comparing the results from this sensitivity study to the results of a run in which they include only dust and soot in the background. I also think that Sect. 3.3 requires further analysis as suggested below. Sect. 3.4 does not present results and should be merged with other sections.

Major comments:

1. Line 9: Please try to better motivate the assumption underlying the sensitivity experiments that yield the -55 mW m⁻² radiative effect and to explain the physical reasoning behind this assumption in the main text, perhaps around Lines 115ff. 

2. Line 16: "one order of magnitude": Please discuss this in the light of Righi et al. (2021) in the results or the conclusions section (e.g. around Line 436). A quick and superficial look only at the abstract of Righi et al. (2021) suggests that the difference in the present study may be much larger here. Righi et al. (2021) mention a factor two.

3. Line 40: Please specify what you mean by "very large". In line 218, you state that it is on the order of -100 mW m⁻². Providing a typical range (possibly mentioning the magnitude of the outliers) may be better. Regarding the magnitude of the anthropogenic forcing due to INP, I agree with your assessment in Line 262, which is repeated in Lines 421f. Although you are not comparing the same quantities in Line 40 and 262, I think there may be an inconsistency hidden in the qualitative assessments in Lines 40 and Line 262. Please consider rephrasing your statement in Line 40 and/or explain relative to what effect an effect on the order of -100 mW m⁻² is very large. Please also check that your recommendation in Line 17 is consistent with your overall assessment. I think that repeating the discussion in Line 262 in lines 420f was a good idea. I think that this comparison is important for understanding the results of this study in the context of the existing literature.

4. In Beer et al. (2020), the authors evaluated simulated dust number concentrations for the upper troposphere at different resolutions with observations. They found very large differences in upper tropospheric dust concentrations depending on resolution (possibly partially due to different tunings), with very large overestimates of dust number concentration relative to observations in some low resolution setups. Such large differences could be expected to influence the sensitivity to anthropogenic INPs found in this study. At the moment, the authors point to Beer et al. (2020) for a model evaluation. However, as far as I can see, Beer et al. (2020) did not evaluate the setup used in the present manuscript. Because of the strong resolution dependence, an evaluation for the resolution (and the tunings) used in this study would, however, be important. The authors should either point the readers to a published evaluation or else provide an evaluation of upper tropospheric dust number concentrations. The authors should also discuss uncertainties related to dust number concentration and/or point the reader to a published discussion. I think that this becomes especially important when looking at radiative effects of anthropogenic INPs.

5. More generally, the authors should briefly explain why they chose such a low horizontal resolution (perhaps because it allowed them to perform many sensitivity runs) and also discuss how this choice could affect their results.

6. Line 330: Nudging may also suppress responses to changes in INP concentration. Please discuss. 

7. Line 333: I would argue that 40% larger does not automatically imply "comparable". Please discuss. If at all feasible, I would strongly recommend to run several additional free running simulations with small perturbations in order to check whether some of the difference could be due to internal variability. Something like slightly changing the stratospheric diffusion (e.g. enstdif=0.99, enstdif=1.01, etc. in ECHAM) during the first month of spin-up or so would be enough. This will give you a range of values for the free running runs. If the value for the nudged run is somewhere near the range from the free-running ensemble (but not necessarily near the average), this would then be sufficient to imply "comparable".

8. Sect. 3.4 discusses uncertainties but does not present results. I think this discussion should be merged with the introduction and the conclusion section. Especially the uncertainties related to ammonium sulfate should be discussed in the in a prominent place so that this discussion becomes difficult to overlook. The information in Line 342 regarding cloud cover parametrization may also fit into the methods section. I suggest to present this particular piece of information prior to the results. 

 

9. Line 369: Why is the purely homogeneous case used as a reference here? 

10. Sect. 3.5. What is the added value compared to previous studies? Please explain.

11. Lines 397-406: Instead of providing a belated introduction, the conclusion section should highlight novel findings. Read together, the introduction and the conclusion section should clearly explain how these new results enhance our knowledge from previous studies. The next paragraph (Lines 407 to 423) nicely summarizes and explains the results in the context of previous studies.

12. Line 491: Does this mean your next manuscript will be based on changing the assumption of a purely homogeneous case in Line 369? I think that adding this analysis to the present manuscript could potentially provide the added value compared to previous studies that I currently have trouble finding in Sect. 3.5.

  

Other comments:

1. Line 12: in Line 9 you report two values for assuming a smaller and a larger ice nucleating potential. In line 12 you report only one value. I think you should clarify in line 12 whether the value in line 12 was derived assuming a larger or a smaller ice nucleating potential.

2. Line 124: "... due to the uncertain freezing properties of aviation BC" -> I don't understand the logic behind this argument. Please explain.

3. Line 146: Does "possible" imply realistic? Please explain or rephrase.

4. Line 199: "[P]ossibly due to increased cloud lifetime effects in the presence of INPs": I think you explain what you mean in Lines 236f and also in Lines 405f. Please explain already here. If you are aware of existing references for this, please mention, perhaps already in the introduction. Unless you are very sure about your suggestion and can explain it better, you could perhaps consider rephrasing this and simply write more frequent (or increased) cloud formation due to INPs instead of increased cloud lifetime effects, or else write more frequent (or increased) cloud formation due to INPs and possibly increased cloud lifetime. The best may be to cite literature. I think you may find something for both statements, although I am not sure, and different studies may yield qualitatively different results.

5. Line 254: Please re-iterate that the concentrations were unchanged from near present-day (2014) levels and perhaps also state this even more clearly already in Sect 2. Around line 254, you could briefly mention masking effects by anthropogenic greenhouse gases. I think that if you used 1850 instead of near present-day greenhouse gas concentrations, the INP effect may increase slightly because of anthropogenic greenhouse gases masking the cloud changes. I also suggest to change "present-day and pre-industrial conditions" to "present-day and pre-industrial aerosol conditions" in line 254. 

6. Line 286: How do these assumptions compare with the default simulation? I suggest to once more point the reader to Righi et al. (2021) and to also add a plot, perhaps of the 95th percentile of total vertical velocity (resolved plus parameterized) to the Supplement. 

7. Line 436: Could you please discuss potential reasons for the disagreement with the statement in the abstract of Righi et al. (2021)? In the abstract Righi et al. mention a factor two. 

8. Lines 451f "recent advancements ... (Kärcher, 2022)": please be more specific. I suggest to briefly explain what the improvement that you are referring to consists of.

9. Table 2: Should there be a base run that does not include mineral dust and soot? Or did I misinterpret line 226 and Figure 3?

 

10. Figures 1, 3, etc. show cloud occurrence frequency instead of cloud cover. Can this be justified by the modifications to the cloud cover parametrization mentioned in line 341? Or would it make sense to show cloud cover? Please explain or modify. Is cloud occurrence frequency defined using a threshold value for cloud cover? 

  

Technical:

Table 1: Stating that all runs were run for 10 years might be enough. I am sure you need an extra column in the table for this. 

Line 4: Does "(aviation) soot" mean "soot including soot from aviation"?

L. 30: ...influence these climatic impacts significantly... -> please rephrase

Lines 238, 238: increased cloud frequency -> more frequent cloud formation?

L. 50: extent of the INP population -> please rephrase

L. 148: provieded -> provided

L. 267: This suggests -> Again, this suggests (because of Line 238)

L. 315: Please add "as explained above" (because of Line 283) or omit.

 

L. 322: Please check that this has indeed been "described above".

L. 325: create much less internal noise -> suppress differences between simulations due to internal variability 

L. 495: Please correct "ThePhysical" 

L. 506: I suggest to cite the final published ACP paper instead of the preprint.

General comment

In this manuscript, Beer et al, investigate the radiative forcing of ice nucleating particles on cirrus clouds through the EMAC global model, including sensitivity tests on some crucial parameters (e.g. the vertical velocity), with a particular emphasis on the potential influence of crystalline ammonium sulfate, which shows a radiative forcing (RF) magnitude comparable to that of other key INP species such as dust (DU) and black carbon (BC).

However, a comprehensive and detailed overview of the microphysics (e.g. INP particle number concentration) and the RF calculations, which are key points for the paper, should be emphasized in the manuscript. A complete and detailed description of the number, mass concentration and global distribution of INPs over the study period (2001-2010) would have been helpful to understand the different contributors to the RF, which I suggest to include in the manuscript. Furthermore, a section in the methods describing the specific reference cases used for RF calculations (i.e. homogenous freezing, pre-industrial…) is beneficial for a better interpretation of the RF results.

The inclusion of sulfate in the RF calculation represents the added value of this paper. However, these findings could have been suitably incorporated into (Beer et al, 2022), where a simulation with a higher spatial resolution has been performed, demonstrating a better representation of the aerosol concentrations, especially for mineral dust. Please find below my comments on your paper.

Major comments

• The simulations were conducted with a lower (higher) spatial (vertical) resolution than those in the study by Beer et al (2022), which presents a contradiction. Is there any reason to decrease the resolution to 2.8°x2.8°, while a better resolution of 1.9°x1.9° has been addressed and proven to provide better simulated aerosol concentrations in (Beer et al, 2022)? A better representation of the aerosol concentration (BC, dust, and sulfate) is crucial to rely on the estimate of the RF. A sensitivity of your results to spatial resolution have to be addressed.
• The spatial resolution can impact also the cloud formation and coverage as well as the representation of vertical updraft, please comment.
• what kind of aerosol feedbacks are activated in the simulation (direct, semi-direct or indirect)? Are the emissions (e.g. biogenic) in your model adjusting to temperatures and wind changes due to feedbacks on meteorology? Pleas comment.
• You mentioned the mixing state, did you investigate the impact of the different mixing states on the radiative forcing? How you treat the mixing state of BC, mineral dust, organics and sulfate? Addressing the impact of the mixing state on the radiative calculations is a major concern.
• How do you calculate the optical properties? Could you provide a description of how you perform the calculation?
• How you define the control cases to compare with your RF results (e.g. Fig. 1 where you compare only with homogeneous freezing)? Adding a section on all the control cases chosen would rather simplify the RF interpretation.
• The paragraph 3.2 shows the sensitivity to the updraft velocity, compared to (Righi et al, 2021), where a similar model configuration has been chosen (e.g. spatial resolution). I would have expected a more detailed comparison between the two papers, given that the model configuration is similar, covering partially the same study period, with the exception of the inclusion of ammonium sulfate.
• A section commenting on the source and variability of the number and mass concentration, as well as size distribution of potential INPs (e.g. DU, BC, organics, sulfate) is missing and it would be appropriate to add it in the manuscript in order to compare the magnitude to the radiative forcing. Which is the major contributor between DU, BC, organic and sulfate to the RF? Given the key role of radiative calculations in this manuscript, I believe it would be more suitable to include a thorough discussion of these aspects within the paper.
• In Sec. 3.1.1 you compare the impact of sulfate to dust and BC. Did you compare the impact of sulfate compared to aviation soot only?
• Did you validate your simulated concentrations with observations? In the conclusions you highlight the importance of the usage of observations as a constraint to the radiative calculations.

L.1-4 Please specify the study period.

L.15 Please quantify.

L.17 Please detail.

L.35 The definition of “radiative forcing” is missing. I would rather add here a description of what is the definition of radiative forcing. At some point in the manuscript you compare your result with the effective radiative forcing of aerosol-cloud interaction which I assume is a different defined quantity compared to your RF.

L.35 References to “several global modelling studies” are missing. Please provide references.

L.39-42 I would rather reformulate evidencing the range of number reported by the references you cited. “very large negative forcing”, how much?

L.54 you provide the term “dynamic forcing”. Please provide a definition.

L.85 how much the resolution is impacting your result? 2.8°x2.8° is very low and may increase the uncertainties on cloud formation, the particle size distribution and updraft velocity. Please comment.

L.86 Is there a particular reason the choose the 2001-2010 study period? It seems that you use the anthropogenic inventory of 2014 to represent the “present day”. Please discuss.   

L.88 Please specify the resolution of the meteorological data.

L.94 “and mixing state”, which are the specific mixing states you are talking about? External? Internal? Can you provide more information about how you dealing with particle mixing state? This is a crucial part for the radiative calculations.  

L.96 Please provide a resume of the setup used in this study.

L.97 Please specify the anthropogenic and biomass burning inventory resolution. How you regrid the emissions to your low spatial resolution? Please comment.

L.104 Please detail “all other natural emissions”

L.162-166 are not results. I suggest to include it in the radiative calculation and relative uncertainties section.  

L.122 Please details how you are discerning between aviation soot and others sources. Which are the “other sources” you are talking about?

L.126 “Natural secondary”, biogenic SOA are not only formed by terpene emissions. Can you please detail on which are the species you are taking into account for the SOA formation?

L.144 what do you mean by “dynamic influence”?

L.170-171 Please add a reference.

L.173 -28 mWm^-2 is the result of a non-significant and a significant RF. Is this result reliable? Please comment.

L.177 the shortwave positive is not significant…

L.187-189 Please add references.

L.198 “regional reduction in cloud occurrence” not true over the Equator, please comment.

L.214 “global cooling” how much? Please quantify.  

L.217 “-100mW/m2” to what RF is referred?

L.219 Please quantify.

L.235 “large number concentration of ammonium sulfate INPs in that region”, how the spatial resolution impacts the magnitude of the RF of this result? You are referring to a temporal and spatially different simulation. And what about the non-significant RF at the tropics? Please comment.  

L.236-237 “Notably…soot”, please add a reference

L.260-262 you compare your results with the results from the IPCC. Are they comparable? RF and ERF are different in the definition.

L.269 please provide a reference.

L.320-335 is this paragraph necessary? we choose the nudging to improve the simulation.

L.360 “highly efficient INP”, could you please detail what you mean with “highly efficient”?

L.397-406 This paragraph does not contain conclusions…

L.413 horizontal resolution may impact this “low glassy organic INP”. Please discuss.

L.437-448 Here you highlight the importance of integrating measurements as a constraint to radiative forcing results. It would have been very interesting to integrate measurements in the current paper. In the manuscript you give an overview of your simulated results, but no validation/constraint with observations is used to validate the order of magnitude of your result. Please discuss.

Minor Comments

L.5 “Several sensitivity experiments”

L.11 Please specify the area of interest for the -29 mWm^-2

L.90 I would rather remove “climatological”, 10 years are not such a long period to be considered climatology.

L.93 “nine log-normal modes”, please specify the range of diameters.

L.38 “sudies” please correct.

L.45 “these processes” to “sub-grid processes”

L.111 “whole freezing spectrum”, please specify what “whole” means.

L.173 the area for the RF of -28 mWm^-2 is missing, you mean global?

L.204 “In order to explore”                                                   

Technical

• “As in Fig.” is recurrent in the manuscript. I would rather suggest to explicitly write a complete description of the figure in the caption.
• Could you please provide the boundaries of your averaged areas? SH-Ext, Tropics, NH-Ext?
• 1 “Confidence levels …respectively”. This part should go on methods where you detail how you perform the radiative calculations/comparisons
• In Tab. 2 the activated fractions are reported as (onset, central). I would rather suggest to quantify the values putting the associated label in parenthesis or vice versa.
• 3 and Fig. 4 could be merged into one figure.
• The magnitude of the clear-sky RF is not readable in the plots.