Review of

“Improved representation of volcanic sulfur dioxide depletion in Lagrangian transport simulations: a case study with MPTRAC v2.4

By Liu et al.

Review by Nina Kristiansen

General comments

The paper presents new developments to include new or improved SO2 removal processes in a Lagrangian transport model. Accurate representation of the emissions, transport and removal of atmospheric constituents is crucial for accurate modelling and forecasting, and therefore the paper is highly relevant for the field of atmospheric modelling. The paper shows a breakdown of the contribution of different removal processes for SO2 (as a % of total mass removed), similar to the Berglen 2004 paper, but with more details and focus on SO2. Such details have been lacking in the literature. The paper is very well written, figures are exceptional and extensive sensitivity tests are presented in a very clear and concise way. The paper is highly suited for publications with only a few minor comments to consider, as specified below:

• Main comment: The comparison between the modelled SO2 mass to the TROPOMI observations in Figure 3a shows a very good agreement. The comparison is done on the TROPOMI horizontal footprint and times. However, such a comparison also need to consider the Averaging Kernel (AK) which characterize the vertical sensitivity of the measurements, and which is required for comparison with other types of data. This AK should be applied to the model data before comparing to the observations (as in Kristiansen et al 2010 doi:10.1029/2009JD013286). As the SO2 emissions and plume height for the case considered (Ambae) is mostly >10 km, I do not think this will change the results significantly, however it should still be included. The AK is available as an output field of the TROPOMI retrieval.
• The comparisons to the SO2 removal processes in the FLEXPART and NAME models could be made clearer. In particular, the Introduction could include an explicit statement that the new developments in MPTRAC are compared to SO2 removal schemes in similar Lagrangian models with a short explanation of the removal schemes in other models. Also acknowledging that many other types of models (e.g., Eulerian CTMs) include quite complex chemistry schemes. Furthermore, the wet deposition scheme in FLEXPART is mentioned, but the SO2 chemistry in FLEXPART is not and should be highlighted (Eckhardt et al 2008  https://doi.org/10.5194/acp-8-3881-2008). Also, both FLEXPART and NAME use OH climatological background fields in their SO2 chemistry schemes and this should be mentioned as its highly relevant to your findings.
• It’s not entirely clear how the chemistry calculations are performed in MPTRAC. Are they done in Lagrangian or Eulerian (i.e., gridded) space? Please provide some more details.
• Please comment on whether the diurnal variability in OH is a significant factor. It gives a stepwise appearance on the plots, but the importance of this is not much discussed.
• The applicability/suitability of an atmospheric SO2 chemistry schemes to volcanic clouds could be discussed in more general. Often these schemes are developed for air quality purposes, which may involve different atmospheric species than those present in volcanic clouds, and also with a focus on the lower atmosphere. Volcanic eruptions that release significant amounts of water to the atmosphere (e.g., the HTHH 2021 eruption) could have a different chemistry in the plume, and the additional water/ice will likely not be included in driving meteorological fields and will need to be treated separately.

Specific comments

L51: Comparing the total SO2 emitted from the Ambae 2018 eruption to that released from the other eruptions mentioned (Raikoke, Sarychev, Nabro, Kasatochi) would give some perspective on the size of the SO2 emissions.

Technical comments

L365: The word ‘change’ should be ‘changed.

L375: Figure 12a should be Figure 12.

Review of manuscript “Improved representation of volcanic sulfur dioxide depletion in Lagrangian transport simulations: a case study with MPTRAC v2.4 ” by Liu et al.

General comments:

This paper introduces the implementation of several new chemical and physical SO₂ removal schemes into the Massive-Parallel Trajectory Calculations (MPTRAC) Lagrangian transport model. The paper demonstrates how different processes contribute to the depletion of volcanic SO₂ in the atmosphere. Using the 2018 Ambae eruption as case study, the authors compare the simulated SO₂ lifetime with the observed SO₂ lifetimes from the Atmospheric InfraRed Sounder (AIRS) and the TROPOspheric Monitoring Instrument (TROPOMI). By preforming multiple sensitivity tests on the new schemes through adjustment of the tuning parameters, it is found that the modelled SO₂ loss processes are highly sensitive to the release vertical profile and the presence of clouds.

Overall, the paper provides an useful evaluation of the new chemical and physical schemes in the MPTRAC model, and the topic of this paper will be of interest to the readers of this journal. Accurate representations of removal processes of atmospheric constituents are highly relevant for the skill of a transport model and its forecasting abilities. Therefore, this work is relevant for the wider atmospheric modelling community.

The paper is well structured, the quality of the figures is good, and the authors give a sufficient introduction to the new schemes and a clear and concise interpretation of the data and sensitivity studies.

I therefore recommend minor revisions to address the points outlined below before publication.

Main comments:

• When comparing the model results with the TROPOMI satellite retrievals in section 3.1, the model simulations should be adjusted using the Averaging Kernel (AK). As mentioned in the methodology section (L.244), the TROPOMI product assumes the SO₂ at a particular altitude in the atmosphere. This also results in a specific vertical sensitivity of the satellite SO₂ retrievals, where the satellite is in general more sensitive to higher altitudes. No such sensitivities are present in the model, and therefore one needs to apply the same vertical sensitivity profile from the satellite retrievals to the model data (i.e. the AK) to make the comparison correct. As most of the SO₂ is above 10km, the impact will most likely be small, but it should be included.
• In the introduction, I think the paper would benefit from a short discussion of the relevant SO₂ removal schemes present in other Lagrangian transport models (e.g. NAME and FLEXPART) to help to put your work in context with these existing schemes.

Specific comments:

L.69: How are the chemistry calculations done in the MPTRAC model? A short discussion of the implementation of the schemes into the model is missing here.

L.116: The implementation of the diurnal variability is mentioned here, but no discussion of the impact on the results in presented in the manuscript. In several figures (e.g. figures 3b, 5b and figure 8) the impact is visible, but it is not clear to me if this is actually having any significant impact on the overall depletion of SO₂. Is it essential to include the diurnal variation, or could this be ignored in future simulations?

Figure 2: Due to the large number of grey points, it is not possible to understand if the two regression lines are a good representation of the data. I think it would be better if the data could be shown as a density plot.

Figure 12: How is the tropopause height calculated?

 Technical corrections/suggestions:

L.14: physical of SO₂ -> physical SO₂

L.109: climatology(Inness et al., 2019)  ->  space is missing

L.365: change -> changed

L.375: Figure 12a -> Figure 12