Borthakur et al. present a study where they use a 1-D chemistry model to study the atmospheric impact of solar proton events. Particularly, they consider the October 2003 event and extreme event case, with special focus on the impact from chlorine ion chemistry. The model results are compared to observations of the MIPAS satellite instrument.

The Authors are neglecting quite many relevant recent publications on ion chemistry and its atmospheric importance, some regarding also chlorine species. In my detailed remarks I am pointing out a few, which I hope will help the Authors.

Considering previous publications, and the fact that MIPAS data and the 2003 event have already been extensively studied before, the novelty of the presented work remains somewhat unclear and should be carefully and clearly stated in the abstract and summary when the text is revised. Because the model has been used for similar studies before, it would be useful to describe its development since previous studies and how that contributes to the novelty of this work. Wider implications or recommendations based on the results are not currently discussed. 

The figures and the data shown should be carefully re-evaluated with the aim of providing more focus and clarity on the main results. Unnecessary data/figures should be identified and removed from the manuscript. The quality of some figures could be improved. Some of the results clearly require more discussion and explanation.

The Authors should address these issues before the manuscript can be published. 

Detailed remarks:

L 91-92: "The chlorine negative ion is the most important ion of the lower D region during day and night time." I don't think this is true. In the winter polar region, large amounts of NOx and lower level of dissociating solar radiation contribute to creation of NO3- and its clusters, for example. For daytime and at lower latitudes, chlorine ions are abundant in the mesosphere but not dominant, I would say. The Authors could actually check this easily from their model results.

L 170: "The ion-chemistry is based on the UBIC (University of Bremen ion-chemistry) model developed by Winkler et al. (2009) for the terrestrial middle atmosphere." The Authors could briefly describe the main development and changes between UBIC and Exotic, if any.

Section 3.1. There are indeed methods for estimation of the polar vortex edge vortex from the gradient of daily CO mixing ratio (e.g. Harvey et al. 2018, https://doi.org/10.1029/2018JD028815). Also, Funke (2005) and the gradient method is mentioned by the Authors, so why use a fixed CO mixing ratio? Also, CO is good for the mesosphere but the Authors also show a lot of results from the stratosphere where the potential vorticity would be a better vortex measure, I assume. Please comment on this.

Figure 5 (and similar figures of MIPAS comparisons). For clarity, I would suggest to remove the model data without-averaging-kernels-applied and harmonize the color scales. Firstly, the x axis is almost unreadable. Secondly, if averaging kernel impact needs to be shown then one good example should do it. Thirdly, changing the color scales makes the plot confusing (at least to me). Finally, it's not necessary to plot the same MIPAS data four times. Maybe it would make sense to plot the difference between the model runs and MIPAS?

Figure 5 (and similar figures). It seems to me that plotting the results from the "full ion chemistry" simulation makes no sense because that model setup cannot handle O(1D) properly. The Authors basically admit this and it's something that was already noted by Winkler (2009, and the follow-up correction paper). Therefore, I would suggest to remove all data from the "full ion chemistry" simulation and clarify in the model description why "O(1D) in photo-chemical equilibrium" is necessary and sound approach. 

Figure 5. Why is there a strong peak in HOCL increase at about 35 km in the daytime simulation panels?

Figure 6. The data above 50 km could be excluded from the figures because nothing is going on.

Figure 8. Is the modeled ozone data sampled at MIPAS measurement local times? There could be variability with local time in the mesosphere.

Figure 9. I don't understand why the parameterised NOx production makes more NOy than the ion chemistry. The parameterization uses a fixed value of 1.25 NOx molecules / ion pair while it was shown by e.g. Nieder et al. (J. Geophys. Res. Space Physics, 119, 2137-2148, doi:10.1002/2013JA019044, 2014) that this number is an underestimation in the upper mesosphere and lower thermosphere. Also, Andersson et al. (J. Geophys. Res. Atmos., 121, 10328–10341, https://doi.org/10.1002/2015JD024173, 2016) and Kalakoski et al. (Atmos. Chem. Phys., 20, 8923-8938, https://doi.org/10.5194/acp-20-8923-2020, 2020) have shown that in the mesosphere the ion chemistry produces more NOx than the fixed parameterization during SPEs. Please discuss and clarify this issue.

Figure 10. The ionic HOx production depends on the level of H2O at least in the upper mesosphere, but the parameterised HOx production is calculated with a fixed H2O profile (Andersson et al., J. Geophys. Res. Atmos., 121, 10328–10341, https://doi.org/10.1002/2015JD024173, 2016). Could H2O explain some of the differences between the parameterized-HOx and the other runs?

Figure 10. The Authors discuss HOx recovery but I think it's worth to note that the SPE ionization stays at an elevated level for the duration of the simulation, i.e. also after the peak on day 302, as shown in Figure 1. Therefore, the recovery is only partial as EPP-HOx production continues albeit with a lower rate.

L 458: "due to the same" => due to the chlorine ion chemistry

Figure 11. Also Andersson et al. (J. Geophys. Res. Atmos., 121, 10328–10341, https://doi.org/10.1002/2015JD024173, 2016) have previously reported a HCl decrease from ion chemistry between 40 and 50 km in during an SPE. However, the Authors show here that a quite similar decrease is seen below 50 km even without ion chemistry.  Also, the decrease at 40-50 km seems to be much larger than what was reported by Andersson et al. Perhaps the Authors can briefly discuss this matter.

Figures 12,13,14. It seems to me that the extended ClO decrease below 50 km seen in the EXT run is balanced by increase and buildup of ClONO2, not HOCl. Could it be that NOx plays a more important role at these altitudes than HOx?

Figure 17. "NO is quite long lived down at 40 km altitude and below so the ozone loss due to the NO catalytic cycle seems to be persistent." Based on the figure, it looks more like "at 60 km and below".

Figure 17. It could be useful the extend the altitude range upwards to fully see the ozone depletion region.

Line 503. "ozone depletion which stays on and produces a diurnal cycle as seen from the Figure 17." The diurnal variability of ozone depletion has previously been discussed by, e.g., Verronen et al. (J. Geophys. Res., 110, A09S32, https://doi.org/10.1029/2004JA010932, 2005).

Line 504. "continuing ozone loss in the middle and upper stratosphere, after the event stops, is found to be still 80-100% for the extreme scenario". Based on the figure, I would say it's 60-80%. I think a brief discussion on this would be needed because ozone changes quite a lot  compared the the impact calculated, e.g., for the Carrington event. See, e.g., Calisto et al. (Environ. Res. Lett. 8 (2013) https://doi.org/10.1088/1748-9326/8/4/045010), Rodger et al. (J. Geophys. Res., 113, D23302, https://doi.org/10.1029/2008JD010702, 2008).

Figure 18. I don't quite see any use for this figure. There is no comparison to observations, so what advantage is gained from averaging the model data for daytime and nighttime? I would suggest to remove Figure 18. 

Figure 19. "A very small impact of the chlorine ions around 10-20 % is observed on the event day." I would not call a 10-20% impact small. Also there should be an explanation why the EXT run shows no impact from chlorine ion chemistry at 70-75 km while the other run does. I suspect it would better to use the same control run as the reference for all runs when calculating ozone depletion percentages, and then compare those numbers. Otherwise some of the differences shown could be simply from using different references.

Figure 19. Kalakoski et al. (Atmos. Chem. Phys., 20, 8923-8938, https://doi.org/10.5194/acp-20-8923-2020, 2020) have reported stronger ozone depletion around 60 km when D-region ion chemistry is included. The results presented here seem to agree qualitatively, particularly in the EXT case. Perhaps a brief comment on this can be added.

The authors have studied the impact of inclusion of Chlorine ion chemistry in a 1D model, particularly in the case of a solar proton event. They contrast he Halloween SPEs to MIPAS observations, as well as contrasting the Halloween SPEs to an event from 775AD. Overall this is an interesting study that should be published, but some clarifications are necessary, particularly to the figures included in the manuscript. 

Major comments:

Figures 5-8: The panels are far too small, the font is too small, the changing colour scale in each column makes comparisons extremely difficult, and the contour intervals are impossible to read in the figures (please add interval to caption). When using set contour intervals, I recommend also using set intervals in the colourbar. Please don’t change the colour scale between the vertical panels.

Figures 9-15: These figures are easier to read, but why are the full ion chemistry results not shown? Again there contour intervals are not clear here. 

Figure 17 and 19: Why is an interpolated colour scale used here? However, the fontsize in the axis used in these figures is the best of all the many contour panel figures and I encourage you to to implement this in the above mentioned figures.

All figures from Figure 4 onwards: Please consider the use of “_” in the figure titles, this does not help the reader.

1D ion-chemistry modelling of the Halloween event has been done in the past, but these studies (e.g. Verronen, et al (2005), Diurnal variation of ozone depletion during the October – November 2003 solar proton events, J. Geophys. Res., 110, A09S32, doi:10.1029/2004JA010932) were not discussed. The previous studies focused on here were those involving 3D models, with limited or no ion chemistry. Some discussion on ion-chemistry studies would be relevant to add.

Overall references should be added to where previous work is discussed. For example where in the text you discuss “Recent studies….” you need to actually include the references. One such example is the sentence on lines 39-40.

Minor comments/typos:

Pay attention to use of “Figure” and “figure” throughout the text.

There are several mentions of “significant” as a measure in the text. What do you mean by significant, did you use some measure for this? For example in section 3.2 you talk about “significant low values”, you probably mean clearly low values etc.

“Earth” (the planet) always starts with a capital letter. 

67.5°N. There should not be a space after the number and before the degree sign, or between the degree sign and N/S.

Check where you need to use comma before the word “which”.

Abstract: Last sentence: This currently reads as there was only 10/20% ozone loss and it was all due to chlorine chemistry. I know you mean that you specifically found that the inclusion of chlorine ion chemistry ADDED 10/20% to the ozone loss, so this sentence ghouls be revised.

Line 33: “..and HEAVIER nuclei”

Introduction: SPEs are currently defined to mean Solar Particle Event and well as Solar Proton Event. In your case you I think you only mean the latter.

Introduction: Can you comment on observed amounts of chlorine ions in the mesosphere? For example, how do we know that Cl^- and Cl^-(H2O) are the most abundant?

Line 106: Shanklin (1985), remove the full stop after Shanklin and before bracket.

Lines 125-126: This sentence needs to be rearranged to that the three statements are in correct order: circumpolar cyclone, formed due to solar insolation, dominates circulation (rather than circulation, use dynamics).

Line 185: Space before the reference.

Section 2.2.2 and conclusions: So why is the rate of O(¹D) too large and how would you assure this was not an issue if the ion chemistry was applied to a climate model? That is, can you offer a working solution from the results of this paper which you could state in the conclusions?

Averaging kernels: The text uses both A and AK, figures use AK. Please change these so that the use is consistent.

Section 3:

Using the averaging kernels: Do you mean that the model is first sampled only at MIPAS altitude grid locations, then averaging kernels are applied. This just need a clarification in the text where currently it’s a bit mixed up.

Same section, you need a reference here for the MIPAS tracer aspects.

Page 13: 

numbered point 1. What do you mean by this?

numbered point 2. degrees North.

numbered point 3. What do you mean by the model temperature profile being fixed? Fixed the whole time?

Section 3.1

Lines 294-295: This needs rearranging or clarifying.

Later you talk about latitude bands relating to vortex edge and deep in the vortex, but you never actually define these regions. Please add the relevant information and how these were determined here.

Section 3.2

Line 310: Initially, MIPAS looks to be much higher. However, this is a good example where the current figure is difficult to read with the changing colour scale and multiple panels. After careful examination the statement in the text is probably correct, but you need to clarify the figures.

Line 315-316: I don’t see this removal of a peak in the figure.

Section 3.3. I do not understand where the 57N-77N = “vortex edge” and 70N-90N = “deep in the vortex” comes from. This is something you need to clarify in the earlier section according to my previous comment. The overlapping latitudes are particularly strange here.

Lines 387-388: Where does this arise from? Adding a reference could be sufficient.

Lines 395-396: You need to include references here.

Lines: 401-402: Does this information come from examination of the reaction rates, or from a reference (not included)?

Lines 410-412: Why do you expect the AISstorm ionisation rates being difference to real ones for the event? Particularly the precipitating proton fluxes are were well known for the Halloween event.

Section 4:

What latitude is modelled here?

Why are both NO_y and HO_x defined different to previous sections of the paper for this comparison? Why not use the same definition (the initial ones) throughout? Seems like the macron was added to NO_y to distinguish it from the previous definition, but the same was not done for HO_x?

Lines 445-446: The results for the full ion-chemistry model results are not shown so we can’t see where this comparison result comes from. 

Lines 463-464: This is repetition of previous sentences, remove.

Figure 14: Looking at this figure, it would seem that the chlorine ion chemistry is not important for ClONO₂? Perhaps you can comment in the text?

Figure 15-16: It’s impossible to see any detail in these figures. I think you should be able to easily clarify these b making lines thicker and adjusting the y-scale. The labelling font is far too small.

Lines 488-489: Not clear. Perhaps: “Because there are not enough chlorine atoms during nighttime…”?  

Section 4.3: The discussion of ozone loss in the mesosphere and stratosphere is a bit mixed. Perhaps you can clearly distinguish between the HO_x and NO_x driven losses to clarify this section. 

In this section it would also be useful to contrast the Halloween ozone loss to previously published model results (e.g. Verronen et al. I mentioned earlier) as well as observations since this is a very well studied event.

Lines: 516-517:  Why is this expected?

End of Conclusions: Based on your work, can you discuss lessons learned that should be considered for inclusions of D-region ion chemistry in EMAC.