Topical Editor "decision stage review" re: reviewer comments on manuscript "The CAMS volcanic forecasting
system utilizing near-real time data assimilation of S5P/Tropomi SO2 retrievals", by Antje Inness et al.

Two reviewers submitted their comments on the manuscript in the discussion period, with also a 3rd review
from the Executive Editor, requesting the code be made available in a suitable open-access repository.

Both reviewers identify the manuscript of interest to the community, and within scope for GMD, as a model
evaluation paper, noting the particular focus on the challenges to assimilating vertically-integrated
information on the SO2 from space-borne sensors to forecast the transport and dispersion of a volcanic plume
from an explosive eruption such as 2019 Raikoke.

Both reviewers had quite a substantial number of minor corrections, which the authors have comprehensively
replied to, in many cases adding quite extensive new text to address the comments. One of the two reviewers
found the title to be "a bit misleading" re: the assimilation experiments being carried out in a different
system than the operational Near-Real-Time CAMS system ECMWF uses for volcanic forecasting. That reviewer
makes a specific suggestion for a change of title, which the authors have followed exactly in the revised
manuscript under consideration here.

In response to the executive editor's comment, the authors have added a DOI for the Tropomi layer height dataset
they applied in the analysis, providing also a link to the ESA S5P website where the Tropomi SO2 data (operational
S5P) can be downloaded. Information on the ECMWF's IFS model codebase and documentation, also with the specific
doi's for each of the 6 IFS experiments, within the "Code and Data Availablility" section.

I have read through the reviewer comments, and the revisions the authors have made, and whereas these certainly
address each of the points raised, and make appropriate edits to the text and Figures accordingly, in some
cases the text added is not specific enough, and requires some improvements.

Also, whilst the revised title the reviewer suggested is generally OK, it omits one of the main values of the
study, in documenting and evaluating the updated baseline capability for assimilating total-column-SO2, now
taking in observations from Tropomi. Although the operational volcanic SO2 plume forecasting system was described
quite comprehensively by Flemming and Inness (2013), the initial volcanic forecasting system there was 8 years
ago now, and assimilated OMI, GOME-2 and SCIAMCHY SO2. The updated system described here assimilates Tropomi
SO2, and the manuscript sets out some detailed specifics of the method for combining into the system.

For this reason, the first of my specific comments in the Decision-Stage Topical Editor review is to request
a minor revision to the title, adding "column SO2 and NRT" into the title, to highlight that both the column-SO2
and the layer-height product aspects are "new" since the original Flemming and Innes (2013) manuscript.

The only other of my major comments (see comment 4) is that section 3.1 of the manuscript "CAMS model" needs
to be extended, to communicate also the important specifics of the atmosphere model being used within the
particular version of the Integrated Forecasting System being used in this analysis. Reviewer 1 points out
that the experiments are carried out in a different system than the near real time (NRT) CAMS system used
operationally for volcanic forecasting.

I must admit that, from reading the manuscript, this is not apparent, and I was not aware of it when I carried
out my initial Topical Editor review. This is an example where the specifics of the broader modelling system
used in a study need to be set out clearly, and this being something that partly motivates why GMD required the
specific version numbers to be communicated at the outset, in the title of the manuscript/article. The 3rd & 4th
of my comments then requests an extra paragraph be added to section 3.1 to set out the specifics of the IFS model
used in the experiments -- and also to consider changing the "The CAMS model" in section 3 title and the
subsection 3.1 title to something more specific such as "The ECMWF Integrated Forecasting System" or similar.

The remaining revisions are very minor however, and I expect the authors will be able to make these changes
easily -- to then enable the manuscript to progress to publication in the GMD, within the joint ACP/GMD/AMT
special issue on the Raikoke eruption.

Note that line numbers below relate to the revised manuscript not the Author-Tracked-Changes manuscript.

Specific comments:
------------------

1) Title -- As I mentioned above, in the revised manuscript the authors have completely replaced the original
title of the article, following exactly the suggestion of Reviewer 1. The revised title mentions only the
assimilation of the SO2 layer height in the system, and although I agree that is the main focus of the article,
since this is (to my knowledge) the first time the ECMWF volcanic SO2 plume forecasting system has been described
with the assimilation of Tropomi SO2 data, it's important to mention this aspect of the paper also.

Suggest therefore to insert "column SO2 and NRT" after "S5P/Tropomi", also deleting "data" from the title.
Also suggest to insert "forecasting" between "global" and "system".

Suggest also a slight re-wording to improve the grammar and readibility -- change "Evaluation of" to
"Evaluating the", and "in the " to "into the", changing also "for the" to ", based on case study of the",
finally re-ordering "the Raikoke 2019 eruption" to "the 2019 Raikoke eruption" (putting the year before
the name of the volcano is the correct way to refer to a specific eruption).

I mean the title then to be "Evaluating the assimilation of S5P/Tropomi column SO2 and NRT layer height
into the CAMS global system, based on case study of the 2019 Raikoke eruption"

2) Title -- this is a follow-on comment, in relation to the GMD policy to request authors state explicitly
in the title the specific version number for modeling systems used in the analysis. With the revised
title, "the CAMS global system" needs to have a corresponding version number applied. As in my general
comments above, I think rather than the "CAMS global system", perhaps better to refer to the "Integrated
Forecasting System" -- and then I think providing the IFS cycle number would then satisfy the GMD policy
on version numbers.

3) Titles to Section 3 and section 3.1 -- related to comment 2, and explained in my general comments above,
please change "The CAMS system" in the title to section 3 to be more descriptive -- certainly the word "global"
is needed -- but as I suggested, maybe the "Integrated Forecasting System" is better in this case? And then
the title of sub-section 3.1 also to follow that revision of "CAMS system". In my initial Topical Editor review
I suggested the change to "volcanic forecasting system" -- and I note in Flemming and Inness (2013), that initial
system was referred to as "volcanic SO2 plume forecasting system" -- and that would be my preference, for the
title of this section 3 and subsection 3.1 at least. Whether the authors feel a change from "CAMS global system"
to "CAMS volcanic SO2 plume forecasting system" is also appropriate I leave to them to decide. Given the GMD
policy re: version numbers, perhaps better to have that as Integrated Forecasting System, but perhaps more suitable
in this section 3 to re-iterate the "volcanic SO2 plume forecasting" capability -- with then the specifics of the
IFS cycle used etc. then explained in the extra para added to section 3.1 (see comment 4 below).

4) Add extra paragraph to section 3.1 to briefly describe the specifics of the atmosphere model used.
As in my general comments, please add an extra paragraph here to communicate the specifics of the modelling system
used. This can be taken from a description of the specific releast cycle of the Integrated Forecasting System that
is used in these model experiemnts. I note that the 2013 Flemming and Innes JGR paper included (section 3.1) quite
a detailed description of the Integrated Forecasting System, and resolution of the model experiments etc.
Please provide similar (but perhaps briefer, to 1 paragraph) specifics of the IFS model and resolution uesd in these
experiments.

5) Section 2.1 (lines 132-133) -- The 3rd of reviewer 1's comments asks the authors whether there is a "super-obbing"
step when the total-column SO2 data is assimilated, and also whether the data are "thinned", furthermore about
specifics of any pre-processing. The authors replied pointing to the text they have already, but for Trompomi that
only says "The TROPOMI SO2 data are averaged to the model resolution (TL511, about 40km) before being used in the
CAMS system".

I am not at all clear what is meant by "super-obbing" but the reviewer's question suggests they're requesting more
information than simply that the data are averaged. If there is any further "thinning" of the data, or pre-processing
in this process, some specific mention should be made of this, or else to state "without any further pre-processing or
data thinning" if that is the case.

6) Section 2.1 (lines 135-136) -- this sentence was added in response to reviewer 1's 5th comment -- I just take issue
with the word "knowledge" here. It's true that the system needs to use a prior SO2 vertical profile -- but perhaps it
only needs to be approximate to broadly represent the particular class of eruption. The word "knowledge" suggests a more
precise profile for the eruption is needed, but that may well not be the case -- it might not be sensitive to the specifics
of the profile used -- it may adjust subsequently to be reasonably accurate even with only an approximate vertical profile.
Suggest to replace "knowledge of a prior SO2 profile" instead to "an assumption for the SO2 vertical profile".

7) Section 2.1 (lines 136-138) -- Starting a sentence "Because" is poor grammar. Replace with "Since".

8) Section 3.2.1 (lines 143-144) -- this sentence (beginning "However..") was also added re: reviewer 1's 5th comment.
The authors have written "these do not provide any real information". Please re-word this to be specific to what is meant.

9) Section 3.2.1 (lines 144-145) -- I am not familiar with the acronym "ATBD" -- please provide the full name here.
If the acronym is used later in the manuscript, also add "(ATBD)" -- but otherwise the acronym is not needed.

10) Section 2.1 (lines 145-146) --- this sentence was added in response to reviewer 1's 4th comment -- I think there needs
to be a few extra words added to be more specific than "observation errors as given by the data providers are used".
I mean to clarify exactly how they are used and where -- I guess you mean within the 4Dvar data assimilation, right?
If so please add a few words after "used" to clarify the specifics of where in the assimilation process those obs-errors
are used.

11) Section 4 (line 306) -- Change "All the satellite data available" to "All the TCSO2 satellite data available".

12) Section 4 (lines 371-373) -- The authors have added this sentence in response to the 2nd of reviewer 2's comments,
but the wording here needs to be improved. Please change "The numerical cost" to the "The CPU cost" or "The computational cost"
and ", with the largest increase coming from" changed to ", with ~50% of the increased cost coming from", then enabling also
to delete the text "which is about 50% numerically more expensive" from the end of the sentence. I'm assuming this change is
consistent with what the authors meant -- but please amend slightly if you meant something different. Main thing is that
"numerical cost" needs to be "computational cost" or "CPU cost". Also suggest to insert "stage of the" after "second" and
before "minimisation" -- if that's what was meant?

13) Abstract (lines 29-35) -- the grammar in this new text needs improving -- "which is better than what is obtained" is
colloquial English, and not appropriate for a manuscript article. Suggest to delete "what is" from that text, and also
add a comma after "the initial eruption". Similarly "forecast skill drops" is colloquial -- change "drops" to "reduces".

14) Introduction (lines 44-45) -- This sentence beginning "In the short term" needs revising -- the sentence is explaining
a potential health hazard of SO2 in the aircraft cabin. I think the term "short term" is referring to the type of exposure,
short-term exposure and long-term exposure, and suggest to replace "In the short term, SO2 in the aircraft cabin" instead
with "If sufficient SO2 is diffused into the aircraft cabin", and replace "is the biggest issue and can lead to" instead
to "this could potentially lead to".

15) Introduction (lines 59-60) -- This short sentence states that assimilating volcanic SO2 in NRT "is difficult".
I'm not sure it adds too much though, and needs either to be deleted or to make some specific about which aspect
is difficult. I'd recommend the former.

16) Section 3.2 (lines 233-236) -- This sentence was added in reply to one of Reviewer 2's comments.
Change "cost function that measures the differences" to "cost function for the total difference".
Also, suggest to delete the word "background" here -- you have "model's background fields" -- but it's
the actual model fields that are considered here -- I get that there are specific terms in the assimilation
referring to "background fields" etc. -- but when combined with "model" that is confusing to the more general
reader because they might be thinking about the non-volcanic component of the model SO2. Suggest simply to
delete the word "background" here and change "model's" to "model". See for example on line 294 (section 3.2.2)
the text states "the model's background SO2 concentrations" and there it means the non-volcanic part.

Also, you've written "by adjusting the initial conditions" -- but is that actually what's done in the 4DVAR
DA system -- isn't it more to identify the solution that gives that minimum in the cost function?
Suggest to delete "by adjusting the initial conditions".
Or if I am misunderstanding, please reply to explain this to me.

17) Section 3.2 (lines 294-295) -- you've added this sentence in response to another of Reviewer 2's comments.
But you've written "some vertical variation of the SO2 loading will be achieved if parts of the plume have
different altitudes" . Re-word this to more carefully explain what you're referring to here. I think you
mean that volcanic plumes in the troposphere tend to become become deformed into multiple layers at different
altitudes. Certainly the text "will be achieved" needs to be changed to something different.

Suggest maybe "We note however that some parts of the SO2 plume are likely to be deformed into multiple layers
at different altitudes..." or simlar text to explain better what you're communicating here.

References
----------

Flemming, J. and Innes, A. (2013) -- Volcanic sulfur dioxide plume forecasts based on UV satellite retrievals for the
2011 Grímsvotn and the 2010 Eyjafjallajokull eruption, J. Geophys. Res., vol. 118, 10,172–10,189.
https://doi.org/10.1002/jgrd.50753

Topical Editor "decision stage review" re: reviewer comments on manuscript "The CAMS volcanic forecasting
system utilizing near-real time data assimilation of S5P/Tropomi SO2 retrievals", by Antje Inness et al.

Two reviewers submitted their comments on the manuscript in the discussion period, with also a 3rd review
from the Executive Editor, requesting the code be made available in a suitable open-access repository.

Both reviewers identified the manuscript within scope for GMD (model evaluation paper), and requiring only
minor corrections, which the authors comprehensively replied to, in many cases adding quite extensive new text
to address the comments.

The title of the originally submitted manuscript referred to the model and data assimilation system as
"the CAMS system", but after my initial Topical Editor review, this was changed to "The CAMS volcanic
forecasting system", which was then the name of the modelling system in the manuscript seen by the
two reviewers. One of the two reviewers stated their opinion that the title was "a bit misleading",
in the sense that the assimilation experiments being carried out in a different system than the
operational Near-Real-Time CAMS system ECMWF uses for volcanic forecasting.

That reviewer made a specific suggestion for a change of title, which the authors initially followed exactly
in the 1st revised manuscript which I considered in my "Decision stage topical editor (TE) review".

In that initial Decision Stage TE review I noted that some parts of the new excerpts of text that were added
in response to the reviewer comments were (in several places) not specific enough, and required some improvement.

And the other aspect I queried was re: that change of title, and re: the related observation that section 3,
and specifically subsection 3.1, did not include any description of the atmosphere model used for the model
experiments to test improvements to the operational volcanic SO2 plume forecasting system.

The edits I suggested to the "added excerpts of text" have almost all been made as suggested, with an explanation
provided for where the authors have chosen to retain the existing text (e.g. re: the aspects of the 4DVAR system).
Since I am much less familiar with those aspects of the model, I am happy for the authors to proceed as they have
suggested, and then those new segments of text are now much improved, and suitable for publication.

Similarly, re: the title, the authors have added mention of the total column SO2 (in addition to the NRT plume
height aspects), both aspects being "new" since the Flemming and Inness (2013) description of the system.
With my request to provide specifics of the modelling system in the title, the authors have chosen to refer to the
modelling system as "the CAMS integrated forecasting system" (rather than "the CAMS volcanic SO2 plume forecasting
system" or "the CAMS volcanic forecasting system"). As I alluded to in my comments, having the manuscript title
refer to the integrated forecasting system then allows the specific version number (IFS cycle number) to be provided
explicitly there -- and for that reason I think that is fine for this article.

The alternative would be to refer to the system more specifically as a progression of the "volcanic SO2 plume
forecasting system", which was the original title of the modelling system it was given in Flemming and Inness (2013).

My opinion as Topical Editor is this is the authors choice, with either being OK, and perhaps that is the best decision
in this case, still referring to the "volcanic SO2 plume forecasting system" in subsection 3.1, but with the broader
host modelling system provided in the manuscript title, and as the overarching title of section 3.

However, whilst I am content with the new title and the new titles for section 3 and section 3.1, it's clear to me that,
although the edits the authors have made do now describe some aspects of the model well -- specifically the emissions,
(e.g. specifics of the anthropogenic and wildfire emissions now provided, with some other information in relation to
the emissions) -- still there are some aspects missing from that section 3.1, that mean the reader is not able to fully
understand some of the basic functionality in the model experiments being carried out.

The main aspects I think are missing, are regarding the SO2 tracer, and its oxidation, and how this aligns with the CB05
chemistry scheme described in the initial paragraph of section 3.1. With there being no information provided, the reader
would likely refer back to the Flemming and Inness (2013).

In that original description of this volcanic SO2 plume forecasting system, the text states "Volcanic SO2 is represented
as a transported tracer in the IFS, which is subject to wet deposition and chemical loss because of reaction with the
hydroxyl radical (OH) expressed as constant e-folding time. Dry deposition, which is an important loss process for SO2,
was not simulated because the SO2 plume was mostly located above the planetary boundary layer."

My understanding of the model experiments presented here, are that the oxidation of the SO2 is no longer based on this
fixed timescale approach, but rather is based on the SO2 tracer within the CB05 chemistry scheme described in the
1st paragraph of section 3.1.

And I wonder if that may be perhaps the main thing the reviewer querying the title was flagging up -- that there has been an
important progression in model capability from that initial Flemming and Inness (2013) description, namely that the volcanic
SO2 is now oxidised consistently with the tropospheric chemistry scheme that is now fully integrated into the IFS, within
this recent realisation of the IFS as the operational CAMS forecasting system.

Basically my main recommended revision here is to, assuming my understanding above is correct, add a sentence or two at
the end of that 1st paragraph of section 3.1, to explain this progression in capability re: the volcanic SO2.

I mean to add 2 sentences something similar to the following:

"In the original version of the volcanic SO2 plume forecasting system described by Flemming and Inness (2013), there
was a dedicated "volcanic SO2 tracer", with oxidation based on a simple fixed timescale approach. By contrast, in the
progression of the volcanic SO2 system described here, the volcanic SO2 emissions, and data assimilation of SO2, is
applied to the SO2 tracer within the CB05 chemistry scheme, with here oxidation to sulphate aerosol occurring, based
on the kinetics specified in the chemistry scheme (e.g. for SO2 reaction with the hydroxyl radical)."

Since this detail is still not provided in the article, I must admit that I am still not 100% sure that is the case,
but since the authors describe the CB05 chemistry scheme so prominently in section 3.1, I can only assume that is
the case.

With this being the primary functionality being evaluated in the paper, I must admit to feeling I should have pointed
this out in my initial TE review, but feel it is best in such initial reviews, to confine my TE comments to only
assess the scope of the paper, and any minor improvements that can be made to the Abstract and Introduction, the
peer review process to assess the main parts of the article.

However, providing this in this final stage of the review will still remedy this, to then properly describe the
functionality of the SO2 emission and oxidation within the system.

In addition to those 2 sentences above re: the chemistry, since the dispersion of the volcanic plume is also a
very important aspect of the system being evaluated, the authors also need to mention the advection scheme,
with also mention of whether a mass-fixer is used, the integration with turbulent diffusion, and the convection
scheme used. I'm referring to the Flemming and Inness (2013) article where it says:

"The IFS uses a semi-Lagrangian advection scheme, and a proportional global mass fixer was applied to ensure
exact mass conservation. The IFS model accounts for turbulent diffusion and convective transport. The former
is mainly confined to the boundary layer. Convective transport played a minor role because of low convective
activity in the study region"

If there have been no changes to the way that system operates (e.g. in relation to the integration of the
CB05 chemistry into the IFS) then it could be appropriate simply to summarise that content into 2 sentences
to then have a 4-sentence paragraph on these aspects. However, if the mass-fixer has any different functionality
with the SO2 tracer (than the separate volcanic SO2 tracer), then this should be mentioned. Perhaps the current
aspects are as described in the Huijnen et al. (2019) GMD article that discussed the chemistry in cycle 43R1.

Basically there needs to be a much better description of the sources, sinks and transport of the SO2 tracer in this new
realisation of the volcanic SO2 plume forecasting system. Perhaps this can just be to pick out the aspects
of the model from these more detailed papers, and refer to those for more information.

But please add this extra paragraph as the 2nd para of section 3.1 -- to then have an adequate description
of how the transport of the SO2 will be progressing in the model experiments, alongside also the DA increments
that adjust the SO2 to best match the satellite observations (where these are available and are flagged to be
volcanic in the satellite dataset).

Once this paragraph is added (between the current 1st and 2nd paras of section 3.1?), I will then be content to approve the manuscript proceed to publication.

References
----------

Flemming, J. and Inness, A. (2013) -- Volcanic sulfur dioxide plume forecasts based on UV satellite retrievals
for the 2011 Grímsvotn and the 2010 Eyjafjallajokull eruption, J. Geophys. Res., vol. 118, 10,172–10,189.
https://doi.org/10.1002/jgrd.50753

Huijnen, V, Puzzer, A, Artela, J., Brasseur, G. et al. (2019) -- Quantifying uncertainties due to chemistry
modelling – evaluation of tropospheric composition simulations in the CAMS model (cycle 43R1), Geosci. Mod. Dev.,
vol. 12, 1725–1752, https://doi.org/10.5194/gmd-12-1725-2019

Topical Editor "3rd decision stage review" re: reviewer comments on manuscript "The CAMS volcanic forecasting
system utilizing near-real time data assimilation of S5P/Tropomi SO2 retrievals", by Antje Inness et al.

Two reviewers submitted their comments on the manuscript in the discussion period, with also a 3rd review
from the Executive Editor, requesting the code be made available in a suitable open-access repository.

Both reviewers identified the manuscript within scope for GMD (model evaluation paper), and requiring only
minor corrections, which the authors comprehensively replied to, in many cases adding quite extensive new text
to address the comments.

The title of the originally submitted manuscript referred to the model and data assimilation system as
"the CAMS system", but after my initial Topical Editor review, this was changed to "The CAMS volcanic
forecasting system", which was then the name of the modelling system in the manuscript seen by the
two reviewers. One of the two reviewers stated their opinion that the title was "a bit misleading",
in the sense that the assimilation experiments being carried out in a different system than the
operational Near-Real-Time CAMS system ECMWF uses for volcanic forecasting.

That reviewer made a specific suggestion for a change of title, which the authors initially followed exactly
in the 1st revised manuscript which I considered in my "Decision stage topical editor (TE) review".

In that initial Decision Stage TE review I noted that some parts of the new excerpts of text that were added
in response to the reviewer comments were (in several places) not specific enough, and required some improvement,
including a further suggested change to the title, and to improve the titles/subtitles for section 3 and section 3.1.

The authors addressed each of the comments I made in that initial Decision Stage TE review, and the manuscript
was much improved at the 2nd initial Decision Stage, very close to being ready for publication.

However, there was still an important omission missing from that section 3.1 that meant that, even at that
2nd Decision Stage TE review, the reader was still not able to fully understand some of the basic functionality
in the model experiments being carried out.

Specifically the aspects still missing were regarding the SO2 tracer, and its oxidation, it not being explained
in section 3 how the SO2 tracer aligned with the CB05 chemistry scheme. With there being no information provided,
the reader would likely assume the same functionality existed as set out in Flemming and Inness (2013).

In that original description of this volcanic SO2 plume forecasting system, there was a dedicated "volcanic SO2 tracer"
its oxidation represented simply as constant e-folding time.

Although I was not certain, I explained my understanding that the model experiments presented no longer use a fixed
timescale oxidation approach for the volcanic SO2, rather the volcanic emissions link directly with the SO2 tracer
within the CB05 chemistry scheme, the chemical sinks then fully consistent with the online tropopsheric chemistry
in the IFS, which recently became fully integrated into the operational CAMS forecasting system.

And my remaining required revision in the 2nd Decision Stage TE review was to, assuming my understanding above
was correct, add a sentence or two at the end of that 1st paragraph of section 3.1, to explain this progression
in capability re: the volcanic SO2. Specifically, I suggested to add 2 sentences there similar to the following:

"In the original version of the volcanic SO2 plume forecasting system described by Flemming and Inness (2013), there
was a dedicated "volcanic SO2 tracer", with oxidation based on a simple fixed timescale approach. By contrast, in the
progression of the volcanic SO2 system described here, the volcanic SO2 emissions, and data assimilation of SO2, is
applied to the SO2 tracer within the CB05 chemistry scheme, with here oxidation to sulphate aerosol occurring, based
on the kinetics specified in the chemistry scheme (e.g. for SO2 reaction with the hydroxyl radical)."

In addition to the requested sentence(s) re: the chemistry, also the other element of the modelling system I identified
was still not sufficiently explained (yet critical to the dispersion of the volcanic plume being evaluated/validated)
was re: the advection scheme, and to mention the mass-fixer approach used, retaining similar breadth (but potentially
less depth) than the original Flemming and Inness (2013) article.

The authors have replied to that "2nd Decision Stage TE review" confirming my understanding was correct re: the online
SO2 oxidation as simulated within the online tropospheric chemistry, and have added the paragraph of text below accordingly,
to now explain both of these key parts of the volcanic SO2 forecasting system's functionality.

"In the original version of the volcanic SO2 plume forecasting system described by Flemming and Inness (2013), there was
a dedicated "volcanic SO2 tracer", with oxidation based on a simple fixed timescale approach. By contrast, in the
progression of the volcanic SO2 system described here, the volcanic SO2 emissions, and data assimilation of SO2,
is applied to the SO2 tracer within the CB05 chemistry scheme (Flemming et al., 2015), with oxidation to sulphate aerosol
occurring, based on the kinetics specified in the chemistry scheme. The main SO2 loss in the coupled chemistry-aerosol
system is the conversion to sulfuric acid. There are two pathways for this (i) in the gas phase via the hydroxyl radical (OH)
and (ii) in clouds (aqueous phase). Pathway (i) via OH is the main pathway in the stratosphere. Heterogenous conversion
on ash particles, is not directly modelled in the IFS. Other loss processes are wet deposition and surface dry deposition.
As described in Flemming et al. (2015) the IFS uses a semi-Lagrangian advection scheme. Since the semi-Lagrangian advection
does not formally conserve mass, a global mass fixer is applied to the chemical tracers, including to the SO2 tracer,
and a proportional mass fixer as described in Diamantakis and Flemming (2014) was used for the runs presented in this
paper. More details about the CB05 chemistry scheme can be found in Flemming et al. (2015, 2017), Remy et al. (2018)
and Huijnen et al. (2019)."

With that info now included, I can confirm I am now content then to approve the manuscript proceed to publication,
except that I noticed there are four typo-level corrections in that added paragraph of text (see below), with these then
just needing to be remedied. Although there are 4 of these, and I've included some detail to explain why these typo-corrections
need to be made, these definitely still fall within the scope of the "technical corrections" classification for approving
to proceed to publication in GMD.

Four remaining technical corrections
-----------------------------------

1) The added text is quite lengthy, and (aside from point 2) is good to include.
But this added text should be added as a new paragraph, otherwise the paragraph becomes too long.
Please start a new paragraph at the point the new text is added.

2) One of the added sentences is not necessary and should be deleted. Specifically I mean the sentence:
"The main SO2 loss in the coupled chemistry-aerosol system is the conversion to sulfuric acid.
This does not really add anything, serving only as a potential source of confusion, and deleting the
sentence does not affect what is already communicated within the paragraph.

3) It is definitely beneficial that the authors have provided some additional specifics about the two pathways for
SO2 oxidation (in-air oxidation via OH, and aqueous phase oxidation within cloud droplets), and furthermore that
they also identify a distinction between volcanic SO2 plumes explosive enough to reach the stratosphere, and
those that remain within the troposphere. I think I understand the authors meaning when they state that, for
stratospheric-injecting eruptions, the oxidation proceeds mainly by the 1st of the 2 oxidation pathways, some
readers might (not unresonably) infer that the authors mean there is still a small proportion of the SO2 oxidation
that occurs via the aqueous phase pathway. Although within the volcanic plume there may indeed be some aqueous
processing, and the oxidation on ash particles is rightly identified as a potentially important pathway for
conversion to sulphate aerosol, the text in this section 3 should be primarily explaining the oxidation occurring
within the modelling system, with then 100% of the SO2 oxidation occurring through the "in-air" pathway.

Specifically, the authors state "Pathway (i) via OH is the main pathway in the stratosphere."

It's a technical point, but the wording of that para there does not adequately distinguish between the pathways
for SO2 oxidation actually occuring within volcanic plumes (i.e. including those that aren't included in model),
and the specific oxidation pathways included in the model simulations the manuscript is evaluating/validating.

Taken on its own, that short sentence above is technically correct (for example re: heterogeneous SO2 oxidation
occurring also on ash particles). But as a follow-on from the previous 2 sentences explaining the model capability,
this would likely be confusing to many readers, a common inference likely that there must also be other oxidation
pathways occurring in the simulations.

Clearly there are no clouds in the stratosphere (aside from the initial volcanic plume, and polar stratospheric
clouds), but then since the model does not resolve SO2 oxidation in volcanic plumes, or (as is stated later in
the para) not on ash particles, readers will be puzzled as to what this could mean, with PSCs obviously not a
pathway for in-cloud SO2 oxidation.

To remedy this just required to incorporate that short sentence as a continuation of the preceding sentence,
and I'm suggesting specifically to rename:

"in clouds (aqueous phase). Pathway (i) via OH is the main pathway in the stratosphere.

to instead say:

"within cloud droplets (aqueous phase), with only pathway (i) occurring in the stratosphere (in the model)."

or similar wording.

4) The other remaining very-minor/technical correction is re: that subsequent wording on the SO2 oxidation on
ash particles -- just need to re-word to reduce that slightly, and to cite a paper that has shown these
effects to be important (e.g. the recent paper by Zhu et al., 2020). My specific suggested edit is also
advising the authors mention the other ash effect recently identified to be important -- namely the
heating effect from the ash in lofting the plume to higher altitude (see Muser et al., 2020).

I leave it to the authors whether to mention that 2nd effect from the ash, but in my opinion it would help
to mention this another effect, which I would argue is of equal importance, and shown in that paper to likely
also have been an influential process in determining how the Raikoke plume dispersed.

Since I also noticed that the short sentence mentioning dry and wet deposition is currently out-of-sequence
(that referring to the SO2 tracer), the specific suggested edit I'm advising to make here includes also
bringing that sentence forward, the re-worded sentence on the ash effect(s) then following ofter that.

The current ash text is actually split into 2 sentences, first stating "Heterogenous conversion on ash
particles, is not directly modelled in the IFS." and that dry/wet deposition being within the two,
both of which makes the text difficult to read.

These can be remedied with the simple following technical correction -- to change the existing text:

"Heterogenous conversion
on ash particles, is not directly modelled in the IFS. Other loss processes are wet deposition and surface dry deposition."

to instead say:

"In the troposphere, the model includes also the SO2 loss processes of wet deposition and surface dry deposition.
Although heterogenous SO2 oxidation on ash particles, and the self-lofting effect from the ash heating effect, have
both been shown to be important for the SO2 dispersion from Raikoke (Muser et al., 2020) and also from the
2015 Kelud eruption (Zhu et al., 2020), ash particles are not included in these IFS simulations."


References
----------

Muser, L. O., Hoshyaripour, G. A., Bruckert, J. et al. (2020)
Particle aging and aerosol–radiation interaction affect volcanic plume dispersion: evidence from the Raikoke 2019 eruption
Atmos. Chem. Phys., 20, 15015–15036, https://doi.org/10.5194/acp-20-15015-2020

Zhu, Y, Toon, O.B., Jensen, E. J. et al. (2020)
Persisting volcanic ash particles impact stratospheric SO2 lifetime and aerosol optical properties,
Nature Communications, vol. 11, 4526, https://doi.org/10.1038/s41467-020-18352-5