Review of “Impact of a strong volcanic eruption on the summer middle atmosphere in UA-ICON simulations” by Wallis et al.

The manuscript is focused on understanding the response of the summer mesosphere to a strong volcanic eruption. These effects are investigated using the UA-ICON model, with the volcanic effects included by simulating the influence of an injection of 20 Tg S into the stratosphere. The simulations indicate that a large response (~15-20 K) occurs in the summer mesosphere several months after the simulated eruption. Two sets of ensemble simulations with different gravity wave forcing are used to diagnose the mechanism by which the volcanic eruption influences the summer mesosphere, with a particular focus on inter- versus intra-hemisphere coupling. The manuscript provides insight into how the mesosphere responds to volcanic eruptions, and would be suitable for publication. However, I believe that there are a number of aspects that would first need to be addressed prior to publication. These are provided in the specific comments below.

Major Comments

1. The manuscript would benefit from additional description of how the volcanic eruption is simulated in the model. Although a description is provided in Section 2.2, the reviewer found it difficult to understand exactly how the effects of the volcanic eruption are included. My interpretation from the text is that this is done by specifying a modification of the aerosols in the model, which then influence the stratosphere heating. It is recommended that the authors revise the description of the simulation setup in order to make the description of how the volcanic eruption is included in the model clear to the reader. It would also be beneficial to explicitly state the timing of the simulated eruption, which can only be inferred from the text and figures currently. 

2. There are clear differences in the results for the two experiments with different gravity wave parameters. However, it is unclear how to interpret these results. My understanding is that the results in experiment 1 use the default gravity wave parameters, which were presumably tuned to obtain accurate model climatology, but that using modified gravity wave parameters provides responses to the volcanic eruption that are more consistent with expectations, especially in the response of the Northern Hemisphere polar vortex. The results would thus partly seem in conflict. That is, the tuned gravity wave parameters would give a better climatology but potentially worse volcanic eruption response, while the modified parameters give worse climatology but better response to the eruption. Is this correct? It is recommended that the authors include some additional discussion with regards to how to interpret the results with the two different specifications of the gravity wave parameters.

3. The UA-ICON model does not include interactive chemistry. This represents a possible limitation to the simulations. For example, the effects of the volcanic eruption on ozone are not included. This limitation is not discussed at all in terms of how to interpret the results. Additional discussion of the potential limitations of the study due to neglecting the chemical effects should be included.

4. The interpretation of the results in terms of the effects of a large volcanic eruption on the mesosphere are unclear. Should the effects in terms of the summer mesosphere cooling be considered only qualitatively? That is, the results of the study show the potential mechanisms that would lead to the summer mesosphere cooling, but the magnitude of the cooling is uncertain. 

Minor Comments:

1. Line 50: “below as as” should be “below as”

2. Line 66: “the dynamic core” should be “the dynamical core”

3. Line 86: The authors should clarify that the two reference experiments are also ensemble simulations.

4. Lines 111-112: The authors should consider moving this text to the beginning of Section 3 so that it is immediately clear to the reader why the results in Figure 3 are focused on November-February. 

Impact of a strong volcanic eruption on the summer middle atmosphere in UA-ICON simulations 

Sandra Wallis, Hauke Schmidt, and Christian von Savigny 

This paper focuses on the response of the mesosphere to heating in the lower tropical stratosphere that is caused by an injection of aerosols from a large volcanic eruption. To do this a high-top GCM is used where the volcanic forcing is prescribed. Analysis of the differences in ensemble means of runs between those with and without volcanic forcing indicates the mesosphere in the Southern Hemisphere summer changes. This is attributed to changes in momentum deposition from parameterized non-orographic gravity waves. The sensitivity to the response to changes in sub-scale orographic drag parameterization settings is tested. The paper is well written and the presentation of the results is, for the most part, clear. Its topic is suitable for ACP. However, the methodology needs to be better explained and I have some concerns about interpretation. Specifically, the paper could be improved in several areas by providing: 1) a better model description (esp. the handling of radiative transfer), 2) analysis of the mesopause temperature and location, rather than temperatures in the mesopause region, 3) analysis of the drivers of stratospheric wind changes (e.g., EP flux divergence vs thermal wind changes), 4) an explanation of the orographic gravity wave parameterization and motivation for changing it (with respect to inter vs into hemispheric coupling), and 5) an expanded description of the model volcanic forcing. Below are specific comments that I think if addressed will cover most of the issues listed above.

Specific comments:

The Abstract needs updating to reflect the results of the study in a quantitative sense. Almost all the text is devoted to motivation and saying what will be done ("This study will focus"), rather than what was done. There is no mention of the sensitivity of test with the orographic drag parameterization.

L8: "The simulation" does not reflect that ensembles were run with different gravity wave parameterization settings.

L63-75: The description of the UA-ICON is insufficient to determine the suitability of the model for the current study. Many details are missing. For example, what was changed in the dynamical core to go "from shallow to deep atmosphere dynamics" (L67). For what date was IFS analysis used to initialize the model and how was the model initialized all the way up to 150 km? How are the prescribed constituents used (L71)? How is radiative transfer handled, are non-LTE effects and chemical heating accounted for? What is the source of the solar spectral irradiance and what wavelength region does it cover? If this is the first time this model has been presented, some validation should be shown (e.g., comparison to URAP or another satellite climatology).

L74: described -> parameterized or implemented?

L76: A more detailed description of the 'volcanic forcing' is needed. It was not clear if the aerosols distribution is prescribed or prognostic. Does it react at all to the change in tropical upwelling rates shown Figure 3? How does the vertical distribution evolve with time? More importantly, is it realistic? Unless I missed it, the date of the eruption is not described, nor how long the model ran before the eruption. 

L87-91: I found the sensitivity experiment (settings in the Lott parameterization) poorly motivated? Why is it necessary to explore the impact of "relatively different representations of the polar vortex". The reference to Figures 4 and 8 do not actually show the state of the polar vortex, so it is difficult to judge what or why this matters. Saying the G and Cd changes also is not helpful to a reader who is familiar with the parameterization. Please describe what these parameters do and describe how the model state has changed rather than asking the reader to try to determine the differences between two figures. For example, I would guess that Cd is a (surface?) drag coefficient? Is it reasonable to change it by a factor of 2? How long was the model allowed to adjust to the new settings from its IFS initialization?

L93: It would be good to provide a reference for formulation of the TEM calculation used (e.g., Gerber and Manzini [2016] or Andrews, Holton and Leovy, etc.)?

L99: It would be good to state the differences from the Ref1 case as well as the absolute temperature. It's not clear if this is simply an upward shift in the location of the mesopause (as indicated by dipole patter) or a cooling of the mesopause.

L103: Perhaps showing the w* anomaly would be more convincing, rather than streamfunction. Note, also, that Figure 3 has no units.

L107: Be sure to be consistent in the font used for v*. Also, if it is the residual velocity, I think it should have an overbar.

L119: Is the change in the zonal wind in the stratosphere only from the change in temperature gradient? I would think any change in the stratosphere would impact the wave driving. The EP flux divergence in Figure 7b seems to be non-negative but the scale does not really reveal the magnitude of the forcing. 

L128: Since you are talking about non-zero phase speeds, it would be good to be explicit that you are referring to the non-orographic gravity wave forcing. Regarding Figure 5 and the 'zonal wind tendency', again, it's not clear if this is only from the non-orographic parameterized gravity waves or all GWs.

L140: Picking a fixed altitude in Figure 6 does not show the mesopause change, which is the cold point. What should be shown is the mesopause temperatures for each case and their differences. It would also be useful to show the height variation.

L153: I think it is generally acknowledged that wave-mean flow interactions drive the stratospheric circulation and gravity waves are responsible for the driving of the deep branch of the B-D circulation and the cold summer mesopause. It is therefore not surprising that the EP flux divergence is relatively small at the mesopause. What I think it missing here is a quantification of the impact of the resolved wave driving on stratospheric zonal winds that then impacts the wave filtering of the parameterized gravity wave forcing. It seems too simplistic to simply say the heating in the tropics causes a thermal wind response at mid-latitudes that impacts gravity wave filtering.

L156: Can you clarify why "the simulated weakening response of the polar vortex in the winter stratosphere seems unrealistic"?

L156: "the gravity wave parameterization was therefore changed" is ambiguous - please be explicit that you are changing the orographic GWs and why.

L160: This is confusing - the volcanic forcing causes the anomaly. Did the mesopause really warm 21K? In other words if the mesopause was 140K it only gets down to 161K?

L164: It is not clear if this difference of differences is coming from the difference in the Refs (very likely) or that the *response* was modified by the choice of the orographic gravity wave forcing. Which is it?

L212: Do you really mean blocking here? Or filtering? Blocking typically refers to a surface pressure condition.

L266: The explanation of how the Ref2 and Vol2 experiments can isolate the interhemispheric coupling needs to be expanded - can you describe how this works? Preferably much earlier in the paper. Apologies if I missed the argument. As far as I can see this just changes the strength of the zonal wind in the Southern Hemisphere.