Contribution of expanded marine sulfur chemistry to the seasonal variability of DMS oxidation products and size-resolved sulfate aerosol

Tashmim, Linia; Porter, William C.; Chen, Qianjie; Alexander, Becky; Fite, Charles H.; Holmes, Christopher D.; Pierce, Jeffrey R.; Croft, Betty; Ishino, Sakiko

doi:https://doi.org/10.5194/egusphere-2023-1056

Preprints

https://doi.org/10.5194/egusphere-2023-1056

Preprints

12 Jun 2023

| 12 Jun 2023

Contribution of expanded marine sulfur chemistry to the seasonal variability of DMS oxidation products and size-resolved sulfate aerosol

Linia Tashmim, William C. Porter, Qianjie Chen, Becky Alexander, Charles H. Fite, Christopher D. Holmes, Jeffrey R. Pierce, Betty Croft, and Sakiko Ishino

Abstract. Marine emissions of dimethyl sulfide (DMS) and the subsequent formation of its oxidation products methane sulfonic acid (MSA) and sulfuric acid (H₂SO₄) are well-known natural precursors of atmospheric aerosols, contributing to particle mass and cloud formation over ocean and coastal regions. Despite a long-recognized and well-studied role in the marine troposphere, DMS oxidation chemistry remains a work in progress within many current air quality and climate models, with recent advances exploring heterogeneous chemistry and uncovering previously unknown intermediate species. With the identification of additional DMS oxidation pathways and intermediate species influencing its eventual fate, it is important to understand the impact of these pathways on the overall sulfate aerosol budget and aerosol size distribution. In this work, we update and evaluate the DMS oxidation mechanism of the chemical transport model GEOS-Chem by implementing expanded DMS oxidation pathways into the model. These updates include gas- and aqueous-phase reactions, the formation of the intermediates dimethyl sulfoxide (DMSO) and methane sulphinic acid (MSIA), as well as cloud loss and aerosol uptake of the recently quantified intermediate hydroperoxymethyl thioformate (HPMTF). We find that this updated mechanism collectively decreases the global mean surface-layer gas-phase sulfur dioxide (SO₂) mixing ratio by 38 % and enhances sulfate aerosol (SO₄^2-) mixing ratio by 16 %. We further perform sensitivity analyses exploring the contribution of cloud loss and aerosol uptake of HPMTF to the overall sulfur budget. Comparing modeled concentrations to available observations we find improved biases relative to previous studies. To quantify impacts of these chemistry updates on global particle size distributions and mass concentration we use the TOMAS aerosol microphysics module, finding changes in particle formation and growth affect the size distribution of aerosol. With this new DMS-oxidation scheme the global annual mean surface layer number concentration of particles with diameters smaller than 80 nm decreases by 12 %, with cloud loss processes related to HPMTF mostly responsible for this reduction. However, global annual mean number of particles larger than 80 nm increases by 4.5 %, suggesting that the new scheme promotes seasonal particle growth to these sizes capable of acting as cloud condensation nuclei (CCN).

Received: 19 May 2023 – Discussion started: 12 Jun 2023

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 1869 KB)

Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
Preprint (1869 KB)

Download & links

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Journal article(s) based on this preprint

19 Mar 2024

Contribution of expanded marine sulfur chemistry to the seasonal variability of dimethyl sulfide oxidation products and size-resolved sulfate aerosol

Linia Tashmim, William C. Porter, Qianjie Chen, Becky Alexander, Charles H. Fite, Christopher D. Holmes, Jeffrey R. Pierce, Betty Croft, and Sakiko Ishino

Atmos. Chem. Phys., 24, 3379–3403, https://doi.org/10.5194/acp-24-3379-2024,https://doi.org/10.5194/acp-24-3379-2024, 2024

Short summary

Linia Tashmim, William C. Porter, Qianjie Chen, Becky Alexander, Charles H. Fite, Christopher D. Holmes, Jeffrey R. Pierce, Betty Croft, and Sakiko Ishino

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-1056', Anonymous Referee #3, 30 Jun 2023

Tashmim et al report global model simulations of DMS oxidation. The model includes an advanced DMS oxidation scheme that accounts for recent insights into DMS oxidation chemistry. This work builds on the work of Novak et al., where the gas and multiphase chemistry of HPMTF was explored initially in GEOS-Chem. This work significantly advances beyond the study of Novak et al. to investigate the role of DMS+O₃ multiphase chemistry and the impact of the new DMS oxidation mechanisms on particle number and size distributions. The manuscript is well written and should be published following the authors attention to the following points:
General comments:
The modified version of the model has HPMTF and DMS cloud chemistry. It was not clear to me how DMSO and SO₂ cloud chemistry was treated in the model. This seems to be an important component of sulfur cycling in the MBL that could be addressed here. Was this included, but not discussed or was this chemistry not included in the revised mechanism.
The percentage of DMS lost to each reaction pathway (e.g., OH, BrO, O₃(aq), NO₃) is cited in the conclusions and features in figure. It is not abundantly clear how these percentages were calculated. Are these the fraction of DMS emitted that is lost to each of these reaction pathways? Or is this the average of the fractional losses (e.g., f(DMS_OH)/total loss) averaged spatially over the entire map? I think it should be (and probably is) the former, but it would be helpful to have confirmation.
Specific Comments
Line 48: Cl and BrO should be in parentheses rather than brackets.
Line 81: Are you referring to a multiphase DMS+OH addition pathway or multiphase DMSO chemistry, or a DMS+O3 pathway. In either case, it would be helpful to be specific.
Line 87: I appreciate that MSP is used in the literature, but I don’t know why. What is that an acronym for? I would suggest MTMP.
Line 90: I don’t think all of these references are for the last statement (30-50% of DMS ends up as HPMTF). Perhaps distribute the references through the sentence so they refer to the correct statements?
Figure 1: In my version there are no green boxes as referenced in the figure caption (they are orange). Are these the only species and reactions used? More specifically, is DMSO chemistry included? It is discussed in the text surrounding Figure 1, but not highlighted in the figure caption. I appreciate that this may complicate the figure (and I am not suggesting it needs to be added), but if DMSO features in the model, it would be good to state it in the figure caption.
Line 125: The numbers cited here are from the global model simulation across all cloud fields, not just for the cloudy case. Perhaps this was the intent of the sentence, but maybe breaking this into two sentences would help get this point across that the 24% reduction in MBL SO2 is a global, annual average not from the case study.
Table 4: What is MSP + MO2?
Table 5 caption: It would be helpful to fully explain what HPMTF =SO42- means. I think you mean there is a 100% S-yield of SO42-. Also, is gamma here really the activity coefficient? I think you mean uptake coefficient.
Line 224: OH+HPMTF was measured in Jernigan et al. it would be best to cite that.
Line 236: I don’t think it maters at all (since loss is diffusion limited in the cloud) but the HPMTF uptake coefficient to dilute cloud droplets should not be faster than that to the aerosol. I would use the experimentally determined value from Jernigan for both. Again, I don’t think it matters for the simulation.
Line 280: How are these fractions of DMS loss calculated? Is this taking the map (in Figure 3) and calculating and average % or is this weighted by the amount of DMS that is lost. Given the strong spatial gradients in DMS I think this makes a difference.
Line 286: What are the “two possible pathways” Shouldn’t DMS+NO3 make MTMP with 100% yield? I am really surprised that DMS+NO3 accounts for 15% of the total DMS loss? That seems big to me as I’d expect [NO3] to be almost zero at the surface over the ocean. Perhaps some more discussion on this point is needed.
Figure 5, Line 352: These DMS measurements look very, very low. I think it is appropriate to question whether they are correct. Also, what measurements are used to create Figure 5?
Figure 6: Without constraining the DMS flux, I don’t think it is possible to attribute the improvement in model-measurement of [DMS] to inclusion of DMS+BrO. It is very likely that the DMS emissions are driving this.

Citation: https://doi.org/10.5194/egusphere-2023-1056-RC1
- AC2: 'Reply on RC1', Linia Tashmim, 04 Oct 2023
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1056/egusphere-2023-1056-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-1056-AC2
RC2:
'Comment on egusphere-2023-1056', Anonymous Referee #2, 02 Jul 2023
This paper explores, within the GEOS-Chem CTM, the impact of a more complex description of the oxidation of DMS on the concentration of sulfur compounds and size resolved aerosol. This is an important area of research with the oxidation of DMS providing a significant, natural background source of sulfur in both the present and past atmospheres. Having a robust understanding of this chemistry is thus vitally important for us to understand both the present day atmospheres and any changes from the preindustrial to the present day. The current representation of the chemistry scheme in this model (the three reactions given in Table 1) is outdated and it is good to see that some development work is taking place.
I however I have two significant concerns about this paper and then a number of smaller ones (described below). Until these major concerns are addressed I don’t think the paper is suitable for publication.
Major issues:
The new chemistry scheme. The mechanism used for the model is a merging of a number of different mechanisms available in the literature. However, I have some concerns about how this has been done.

The DMS + NO3 reaction appears to be in twice. It is in the OH addition pathway section and in the H-abstraction pathway section. The rate constant is the same for both pathways but is given different references. I think this essentially means that this reaction is double counted and the DMS+NO3 channel is twice as fast as it should be. Both the latest IUPAC and NASA data evaluation has a single NO3 + CH3SCH3 → CH3SCH2 + HNO3 reaction for this. Thus any subsequent chemistry needs to come from the further oxidation of CH3SCH2. Thus I think that there has been double counting by having this reaction in twice.
Similarily, I am confused by the DMS+Cl reaction. It has two channels, an abstraction channel (DMS+ClCH3SCH2 + HCl) and an addition channel (DMS+ClDMS-Cl). The IUPAC recommendation gives the recommendation of 3.6e-10 for both reactions with a 50:50 ratio between the two. This paper seems to follow this recommendation with a reaction of DMS+Cl0.5SO2+0.5DMSO+0.5HCl+0.5ClO. However, an additional reaction DMS+Cl0.45MSP+0.55C2H6SCl+0.45HCl is also included in the scheme. This is again is a split between the addition and abstraction reactions (0.55:0.45). But it appears that the overall DMS+Cl reaction is in the mechanism twice. I’m also then a bit confused by the C2H6SCl chemistry. I think the only thing that can happen to this in the mechanism is that it falls apart back to DMS+Cl. Thus the addition channel in this part of the chemistry is effectively a null cycle for DMS oxidation whereas for the other DMS+Cl reaction there is an assumption that it leads to the continued oxidation of the DMS.
The OH-addition reaction between OH and DMS gives SO2, MSA and CH3O2 as the products. Quoteing Pham and Spracklen. Looking at Spracklen they have that channel for the DMS oxidation giving 0.6SO2 and 0.4DMSO. The DMSO can then react with OH to give MSA. The mechanism inclulded in the model seems to have lumped this together to avoid having to have DMSO as a tracer. However there is DMSO as a tracer in place for the oxidation of BrO.
The basis for some of these rates is some rather old complications of recommended rates (Saunders et al., 2003, Burkholder et al., 2015). There are more upto date recommendations in the the literature by both IUPAC and JPL. It would be very useful to update the mechanism to these recommendations rather than relying on some rather elderly rate constants.
Overall, I feel that the new chemistry scheme has rather crudely merged previously developed chemistry schemes without much thought to the underlying assumptions in these scheme. These previous schemes have made various approximations, but the new mechanism doesn’t seem to have understood these approximations and developed a scheme which is capable of either removing these approximations or by dealing with them appropriately. It has just patched things on top of each other. It would be advantageous to read the primary literature, the IUPAC and NASA recommendations for rate constants and use these as the basis of creating a consistent mechanism which uses the latest current thinking for this oxidation Unless the mechanism can be better updated and it then better explained I don’t think the basis of this work is built on weak foundations.
DMS and HPMTF Concentrations.

After developing this new chemistry oxidation scheme the modelled DMS and HPMTF concentrations are compared to those from the ATOM-4 mission. The model does pretty poorly for DMS and surprisingly well for HPMTF. This leaves the authors in a difficult position. The DMS emissions could be wrong, but that would imply that the HPMTF is then right for the wrong reason. The DMS emissions could be right but the DMS lifetime was too long but that would imply an error in the chemistry mechanism. Or the DMS observations could be incorrect. They show the seasonal cycle from one surface site which looks pretty good as additional justification, but this doesn’t seem sufficient.
It would be useful to discuss the DMS emissions in the model more. What is the emission in the model? How does this compare to previous studies? Are the model emissions higher / lower than other studies etc? How much wiggleroom is there here for improving the model performance?
If think that more analysis is needed to show that the DMS concentration calculated by the model are 'reasonable' and that the ATOM DMS observations can be reconciled with the model. I would suggest that more comparisons with surface sites would provide the increased confidence here. It seems difficult to go onto the next stage of the analysis (the impact on aerosols), without having confidence in the ability of the model to get the DMS concentrations right. At the moment there is some doubt.
Minor issues:
Table 1. Can the rate constants be put into this table?
Figure 1. Where do the numbers come from for this table. It would be useful to point towards the simulation that is being used?
Page 6. It might be benefitial to start with a description of the gas phase aspects of the model before moving to the aerosol scheme?
Page 8. The literature contains other DMS oxidants (IO, Br, etc) Why were these not included in the scheme? They may be considered small but it would be good to explain that.
Page 10. A table of DMS emissions and global (hemispheric sinks) would be useful here. Lifetimes to different oxidatants would also provide some useful way of comparing the different oxidation routes in the BASE and the MOD simulations. It would also be useful to provide information on the global (hemispheric) mean concentration of important oxidatants (OH, Cl, NO3, O3, BrO etc).
Line 280. Does the MOD simulation have wet and dry deposition of DMS? Could more information be provided about that?
Line 323. I would put the model / measurement comparisons section before the budget details. I would start with an analysis of the model’s ability to simulate DMS (both from aircraft and from the group) and then move onto HPMTF.
Line 388. Is there a 37% reduction in the global SO2 burden with the change of chemistry? Or is that a spatially averaged fractional change?
Citation: https://doi.org/10.5194/egusphere-2023-1056-RC2
- AC3: 'Reply on RC2', Linia Tashmim, 04 Oct 2023
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1056/egusphere-2023-1056-AC3-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-1056-AC3
RC3:
'Comment on egusphere-2023-1056', Anonymous Referee #1, 05 Jul 2023

Tashmim et al. report results from the GEOS-Chem global chemical transport model incorporating many recent findings on reactive intermediates in dimethyl sulfide (DMS) oxidation chemistry and quantify impacts on terminal products and aerosol particle size and abundance. This type of integrated analysis is necessary for evaluating the combined impact of the numerous recent revisions to our understanding of DMS chemistry and this work therefore has a high potential value. However, this is dependent on a thoughtful synthesis of reaction mechanisms from various sources which I believe needs some further work in this manuscript. In particular, I have concerns about how details of the reaction of DMS with NO3 and Cl were implemented (see major comments below). Otherwise I find the work to generally be of a high quality and the results and discussions are well supported. If these apparent issue with the reaction mechanism are resolved along with the other comments below, then I believe this manuscript will likely be suitable for publication in ACP.
Major Comments:
1.) My primary concern with the manuscript is what appears to me to be a mistake in the reaction mechanism resulting in the DMS + NO3 and DMS + Cl reactions being included twice, which impacts all of the results presented in this analysis. In table 2 the following reaction is listed:
DMS + NO3 → SO2 + HNO3 + CH3O2 + CH2O
rate: 1.90e-13*exp(530/T) reference: (Burkholder et al., 2015)
And in table 4 the following reaction is listed:
DMS + NO3 → MSP + HNO3
rate: 1.9e-13*exp(520/T) references: (Novak et al., 2021; Wollesen de Jonge et al., 2021)
These are not two distinct chemical reactions. Both reactions are an H-abstraction from DMS by NO3 with the same rate constant. The only difference is in the assigned products where the reaction in table 2 makes the simplifying assumption that SO2 is formed at unit yield, while table 4 instead goes through the reactive intermediate species MSP. In reality the reaction in table 2 also proceeds through MSP, this was likely just neglected in the referenced compilation of Burkholder et al., 2015 because the significance of the MSP intermediate for HPMTF chemistry was not know at the time of that data evaluation. Only the Reaction in table 4 should be included in the model. By including both you are double counting this reaction pathway and incorrectly increasing the modelled significance of NO3 chemistry.
Similarly, for DMS + Cl the following reaction is given in Table 2:
DMS + Cl → 0.5SO2 + 0.5DMSO + 0.5HCl + 0.5ClO
rate: 3.40e-10 reference: (Barnes et al., 2006; Burkholder et al., 2015)
While in Table 4 the following reaction is listed:
DMS + Cl → 0.45MSP + 0.55C2H6SCl + 0.45HCl
Rate: 3.40e-10 reference: (Fung et al., 2022)
Again, these are fundamentally the same reaction resulting in this chemistry being double counted. The only difference is in the simplifying assumptions made about product yields.
2.) The results of Jernigan et al. (2022) show that HPMTF is the primary precursor to OCS formation from DMS oxidation with HPMTF + OH -> 0.13 OCS + 0.87 SO2. At a minimum, this should be considered as it will reduce the overall SO2 production from DMS oxidation which will impact the results presented in this manuscript. The overall yield of OCS is also therefore highly dependent on HPMTF multiphase loss processes. With minimal additional analysis, this work could also provide a valuable update on to the GEOS-Chem modeling results from Jernigan et al. (2022). I do not feel strongly that extended analysis of OCS production should be included, but do feel that some comment on the impacts on SO2 production are necessary beyond what is included at lines 91-95.
3.) SO2 mixing ratios were measured during the ATom-4 campaign at suitable precision to be informative in background marine air masses (https://daac.ornl.gov/ATOM/guides/ATom_SO2_LIF_Instrument_Data.html). A comparison of measured and modelled SO2 could be a very useful addition.

Other Comments:
What SO₂ heterogenous chemistry is included in this work?
The sensitivity runs with and without sea-salt aerosol debromination are appreciated given remaining uncertainties in BrO measurements and model implementations. Is it correct that the revised debromination mechanism of Wang et al. 2021 was not used here? If so what is the motivation for this? This comment is based on the references included in the methods section in lines 171-172.
Can you show a figure of the global distribution of BrO in the MOD and MOD without sea salt debromination model cases? Otherwise it is difficult for the reader to make absolute comparisons for either model case to measurements of BrO.
Table 3 and lines 224 - 230: Jernigan et al. (2022) provides an experimental value for k(HPMTF + OH) of 1.4E-11 cm3 molec^-1 s^-1 which is a useful validation of the assumed value of 1.1E-11 cm3 molec^-1 s^-1 used here and in Vermeuel et al. (2020) and Novak et al. (2021). This should be referenced.
Line 388 and onward: You should make clear what the altitude range is for the quoted reductions and enhancements are in the simulation. Presumably these are for some near surface range and not total column?
Figure 9. It appears that much of the particle number increase is for Dp > 200 nm. What is the size range where CCN abundance is most sensitive to particle growth? Some additional context for the reader may be useful in connecting changes in particle size bins to potential changes in CCN abundance.

Citation: https://doi.org/10.5194/egusphere-2023-1056-RC3
- AC1: 'Reply on RC3', Linia Tashmim, 04 Oct 2023
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1056/egusphere-2023-1056-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-1056-AC1

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-1056', Anonymous Referee #3, 30 Jun 2023

Tashmim et al report global model simulations of DMS oxidation. The model includes an advanced DMS oxidation scheme that accounts for recent insights into DMS oxidation chemistry. This work builds on the work of Novak et al., where the gas and multiphase chemistry of HPMTF was explored initially in GEOS-Chem. This work significantly advances beyond the study of Novak et al. to investigate the role of DMS+O₃ multiphase chemistry and the impact of the new DMS oxidation mechanisms on particle number and size distributions. The manuscript is well written and should be published following the authors attention to the following points:
General comments:
The modified version of the model has HPMTF and DMS cloud chemistry. It was not clear to me how DMSO and SO₂ cloud chemistry was treated in the model. This seems to be an important component of sulfur cycling in the MBL that could be addressed here. Was this included, but not discussed or was this chemistry not included in the revised mechanism.
The percentage of DMS lost to each reaction pathway (e.g., OH, BrO, O₃(aq), NO₃) is cited in the conclusions and features in figure. It is not abundantly clear how these percentages were calculated. Are these the fraction of DMS emitted that is lost to each of these reaction pathways? Or is this the average of the fractional losses (e.g., f(DMS_OH)/total loss) averaged spatially over the entire map? I think it should be (and probably is) the former, but it would be helpful to have confirmation.
Specific Comments
Line 48: Cl and BrO should be in parentheses rather than brackets.
Line 81: Are you referring to a multiphase DMS+OH addition pathway or multiphase DMSO chemistry, or a DMS+O3 pathway. In either case, it would be helpful to be specific.
Line 87: I appreciate that MSP is used in the literature, but I don’t know why. What is that an acronym for? I would suggest MTMP.
Line 90: I don’t think all of these references are for the last statement (30-50% of DMS ends up as HPMTF). Perhaps distribute the references through the sentence so they refer to the correct statements?
Figure 1: In my version there are no green boxes as referenced in the figure caption (they are orange). Are these the only species and reactions used? More specifically, is DMSO chemistry included? It is discussed in the text surrounding Figure 1, but not highlighted in the figure caption. I appreciate that this may complicate the figure (and I am not suggesting it needs to be added), but if DMSO features in the model, it would be good to state it in the figure caption.
Line 125: The numbers cited here are from the global model simulation across all cloud fields, not just for the cloudy case. Perhaps this was the intent of the sentence, but maybe breaking this into two sentences would help get this point across that the 24% reduction in MBL SO2 is a global, annual average not from the case study.
Table 4: What is MSP + MO2?
Table 5 caption: It would be helpful to fully explain what HPMTF =SO42- means. I think you mean there is a 100% S-yield of SO42-. Also, is gamma here really the activity coefficient? I think you mean uptake coefficient.
Line 224: OH+HPMTF was measured in Jernigan et al. it would be best to cite that.
Line 236: I don’t think it maters at all (since loss is diffusion limited in the cloud) but the HPMTF uptake coefficient to dilute cloud droplets should not be faster than that to the aerosol. I would use the experimentally determined value from Jernigan for both. Again, I don’t think it matters for the simulation.
Line 280: How are these fractions of DMS loss calculated? Is this taking the map (in Figure 3) and calculating and average % or is this weighted by the amount of DMS that is lost. Given the strong spatial gradients in DMS I think this makes a difference.
Line 286: What are the “two possible pathways” Shouldn’t DMS+NO3 make MTMP with 100% yield? I am really surprised that DMS+NO3 accounts for 15% of the total DMS loss? That seems big to me as I’d expect [NO3] to be almost zero at the surface over the ocean. Perhaps some more discussion on this point is needed.
Figure 5, Line 352: These DMS measurements look very, very low. I think it is appropriate to question whether they are correct. Also, what measurements are used to create Figure 5?
Figure 6: Without constraining the DMS flux, I don’t think it is possible to attribute the improvement in model-measurement of [DMS] to inclusion of DMS+BrO. It is very likely that the DMS emissions are driving this.

Citation: https://doi.org/10.5194/egusphere-2023-1056-RC1
- AC2: 'Reply on RC1', Linia Tashmim, 04 Oct 2023
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1056/egusphere-2023-1056-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-1056-AC2
RC2:
'Comment on egusphere-2023-1056', Anonymous Referee #2, 02 Jul 2023
This paper explores, within the GEOS-Chem CTM, the impact of a more complex description of the oxidation of DMS on the concentration of sulfur compounds and size resolved aerosol. This is an important area of research with the oxidation of DMS providing a significant, natural background source of sulfur in both the present and past atmospheres. Having a robust understanding of this chemistry is thus vitally important for us to understand both the present day atmospheres and any changes from the preindustrial to the present day. The current representation of the chemistry scheme in this model (the three reactions given in Table 1) is outdated and it is good to see that some development work is taking place.
I however I have two significant concerns about this paper and then a number of smaller ones (described below). Until these major concerns are addressed I don’t think the paper is suitable for publication.
Major issues:
The new chemistry scheme. The mechanism used for the model is a merging of a number of different mechanisms available in the literature. However, I have some concerns about how this has been done.

The DMS + NO3 reaction appears to be in twice. It is in the OH addition pathway section and in the H-abstraction pathway section. The rate constant is the same for both pathways but is given different references. I think this essentially means that this reaction is double counted and the DMS+NO3 channel is twice as fast as it should be. Both the latest IUPAC and NASA data evaluation has a single NO3 + CH3SCH3 → CH3SCH2 + HNO3 reaction for this. Thus any subsequent chemistry needs to come from the further oxidation of CH3SCH2. Thus I think that there has been double counting by having this reaction in twice.
Similarily, I am confused by the DMS+Cl reaction. It has two channels, an abstraction channel (DMS+ClCH3SCH2 + HCl) and an addition channel (DMS+ClDMS-Cl). The IUPAC recommendation gives the recommendation of 3.6e-10 for both reactions with a 50:50 ratio between the two. This paper seems to follow this recommendation with a reaction of DMS+Cl0.5SO2+0.5DMSO+0.5HCl+0.5ClO. However, an additional reaction DMS+Cl0.45MSP+0.55C2H6SCl+0.45HCl is also included in the scheme. This is again is a split between the addition and abstraction reactions (0.55:0.45). But it appears that the overall DMS+Cl reaction is in the mechanism twice. I’m also then a bit confused by the C2H6SCl chemistry. I think the only thing that can happen to this in the mechanism is that it falls apart back to DMS+Cl. Thus the addition channel in this part of the chemistry is effectively a null cycle for DMS oxidation whereas for the other DMS+Cl reaction there is an assumption that it leads to the continued oxidation of the DMS.
The OH-addition reaction between OH and DMS gives SO2, MSA and CH3O2 as the products. Quoteing Pham and Spracklen. Looking at Spracklen they have that channel for the DMS oxidation giving 0.6SO2 and 0.4DMSO. The DMSO can then react with OH to give MSA. The mechanism inclulded in the model seems to have lumped this together to avoid having to have DMSO as a tracer. However there is DMSO as a tracer in place for the oxidation of BrO.
The basis for some of these rates is some rather old complications of recommended rates (Saunders et al., 2003, Burkholder et al., 2015). There are more upto date recommendations in the the literature by both IUPAC and JPL. It would be very useful to update the mechanism to these recommendations rather than relying on some rather elderly rate constants.
Overall, I feel that the new chemistry scheme has rather crudely merged previously developed chemistry schemes without much thought to the underlying assumptions in these scheme. These previous schemes have made various approximations, but the new mechanism doesn’t seem to have understood these approximations and developed a scheme which is capable of either removing these approximations or by dealing with them appropriately. It has just patched things on top of each other. It would be advantageous to read the primary literature, the IUPAC and NASA recommendations for rate constants and use these as the basis of creating a consistent mechanism which uses the latest current thinking for this oxidation Unless the mechanism can be better updated and it then better explained I don’t think the basis of this work is built on weak foundations.
DMS and HPMTF Concentrations.

After developing this new chemistry oxidation scheme the modelled DMS and HPMTF concentrations are compared to those from the ATOM-4 mission. The model does pretty poorly for DMS and surprisingly well for HPMTF. This leaves the authors in a difficult position. The DMS emissions could be wrong, but that would imply that the HPMTF is then right for the wrong reason. The DMS emissions could be right but the DMS lifetime was too long but that would imply an error in the chemistry mechanism. Or the DMS observations could be incorrect. They show the seasonal cycle from one surface site which looks pretty good as additional justification, but this doesn’t seem sufficient.
It would be useful to discuss the DMS emissions in the model more. What is the emission in the model? How does this compare to previous studies? Are the model emissions higher / lower than other studies etc? How much wiggleroom is there here for improving the model performance?
If think that more analysis is needed to show that the DMS concentration calculated by the model are 'reasonable' and that the ATOM DMS observations can be reconciled with the model. I would suggest that more comparisons with surface sites would provide the increased confidence here. It seems difficult to go onto the next stage of the analysis (the impact on aerosols), without having confidence in the ability of the model to get the DMS concentrations right. At the moment there is some doubt.
Minor issues:
Table 1. Can the rate constants be put into this table?
Figure 1. Where do the numbers come from for this table. It would be useful to point towards the simulation that is being used?
Page 6. It might be benefitial to start with a description of the gas phase aspects of the model before moving to the aerosol scheme?
Page 8. The literature contains other DMS oxidants (IO, Br, etc) Why were these not included in the scheme? They may be considered small but it would be good to explain that.
Page 10. A table of DMS emissions and global (hemispheric sinks) would be useful here. Lifetimes to different oxidatants would also provide some useful way of comparing the different oxidation routes in the BASE and the MOD simulations. It would also be useful to provide information on the global (hemispheric) mean concentration of important oxidatants (OH, Cl, NO3, O3, BrO etc).
Line 280. Does the MOD simulation have wet and dry deposition of DMS? Could more information be provided about that?
Line 323. I would put the model / measurement comparisons section before the budget details. I would start with an analysis of the model’s ability to simulate DMS (both from aircraft and from the group) and then move onto HPMTF.
Line 388. Is there a 37% reduction in the global SO2 burden with the change of chemistry? Or is that a spatially averaged fractional change?
Citation: https://doi.org/10.5194/egusphere-2023-1056-RC2
- AC3: 'Reply on RC2', Linia Tashmim, 04 Oct 2023
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1056/egusphere-2023-1056-AC3-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-1056-AC3
RC3:
'Comment on egusphere-2023-1056', Anonymous Referee #1, 05 Jul 2023

Tashmim et al. report results from the GEOS-Chem global chemical transport model incorporating many recent findings on reactive intermediates in dimethyl sulfide (DMS) oxidation chemistry and quantify impacts on terminal products and aerosol particle size and abundance. This type of integrated analysis is necessary for evaluating the combined impact of the numerous recent revisions to our understanding of DMS chemistry and this work therefore has a high potential value. However, this is dependent on a thoughtful synthesis of reaction mechanisms from various sources which I believe needs some further work in this manuscript. In particular, I have concerns about how details of the reaction of DMS with NO3 and Cl were implemented (see major comments below). Otherwise I find the work to generally be of a high quality and the results and discussions are well supported. If these apparent issue with the reaction mechanism are resolved along with the other comments below, then I believe this manuscript will likely be suitable for publication in ACP.
Major Comments:
1.) My primary concern with the manuscript is what appears to me to be a mistake in the reaction mechanism resulting in the DMS + NO3 and DMS + Cl reactions being included twice, which impacts all of the results presented in this analysis. In table 2 the following reaction is listed:
DMS + NO3 → SO2 + HNO3 + CH3O2 + CH2O
rate: 1.90e-13*exp(530/T) reference: (Burkholder et al., 2015)
And in table 4 the following reaction is listed:
DMS + NO3 → MSP + HNO3
rate: 1.9e-13*exp(520/T) references: (Novak et al., 2021; Wollesen de Jonge et al., 2021)
These are not two distinct chemical reactions. Both reactions are an H-abstraction from DMS by NO3 with the same rate constant. The only difference is in the assigned products where the reaction in table 2 makes the simplifying assumption that SO2 is formed at unit yield, while table 4 instead goes through the reactive intermediate species MSP. In reality the reaction in table 2 also proceeds through MSP, this was likely just neglected in the referenced compilation of Burkholder et al., 2015 because the significance of the MSP intermediate for HPMTF chemistry was not know at the time of that data evaluation. Only the Reaction in table 4 should be included in the model. By including both you are double counting this reaction pathway and incorrectly increasing the modelled significance of NO3 chemistry.
Similarly, for DMS + Cl the following reaction is given in Table 2:
DMS + Cl → 0.5SO2 + 0.5DMSO + 0.5HCl + 0.5ClO
rate: 3.40e-10 reference: (Barnes et al., 2006; Burkholder et al., 2015)
While in Table 4 the following reaction is listed:
DMS + Cl → 0.45MSP + 0.55C2H6SCl + 0.45HCl
Rate: 3.40e-10 reference: (Fung et al., 2022)
Again, these are fundamentally the same reaction resulting in this chemistry being double counted. The only difference is in the simplifying assumptions made about product yields.
2.) The results of Jernigan et al. (2022) show that HPMTF is the primary precursor to OCS formation from DMS oxidation with HPMTF + OH -> 0.13 OCS + 0.87 SO2. At a minimum, this should be considered as it will reduce the overall SO2 production from DMS oxidation which will impact the results presented in this manuscript. The overall yield of OCS is also therefore highly dependent on HPMTF multiphase loss processes. With minimal additional analysis, this work could also provide a valuable update on to the GEOS-Chem modeling results from Jernigan et al. (2022). I do not feel strongly that extended analysis of OCS production should be included, but do feel that some comment on the impacts on SO2 production are necessary beyond what is included at lines 91-95.
3.) SO2 mixing ratios were measured during the ATom-4 campaign at suitable precision to be informative in background marine air masses (https://daac.ornl.gov/ATOM/guides/ATom_SO2_LIF_Instrument_Data.html). A comparison of measured and modelled SO2 could be a very useful addition.

Other Comments:
What SO₂ heterogenous chemistry is included in this work?
The sensitivity runs with and without sea-salt aerosol debromination are appreciated given remaining uncertainties in BrO measurements and model implementations. Is it correct that the revised debromination mechanism of Wang et al. 2021 was not used here? If so what is the motivation for this? This comment is based on the references included in the methods section in lines 171-172.
Can you show a figure of the global distribution of BrO in the MOD and MOD without sea salt debromination model cases? Otherwise it is difficult for the reader to make absolute comparisons for either model case to measurements of BrO.
Table 3 and lines 224 - 230: Jernigan et al. (2022) provides an experimental value for k(HPMTF + OH) of 1.4E-11 cm3 molec^-1 s^-1 which is a useful validation of the assumed value of 1.1E-11 cm3 molec^-1 s^-1 used here and in Vermeuel et al. (2020) and Novak et al. (2021). This should be referenced.
Line 388 and onward: You should make clear what the altitude range is for the quoted reductions and enhancements are in the simulation. Presumably these are for some near surface range and not total column?
Figure 9. It appears that much of the particle number increase is for Dp > 200 nm. What is the size range where CCN abundance is most sensitive to particle growth? Some additional context for the reader may be useful in connecting changes in particle size bins to potential changes in CCN abundance.

Citation: https://doi.org/10.5194/egusphere-2023-1056-RC3
- AC1: 'Reply on RC3', Linia Tashmim, 04 Oct 2023
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1056/egusphere-2023-1056-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-1056-AC1

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

AR by Linia Tashmim on behalf of the Authors (04 Oct 2023) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (08 Oct 2023) by Ashu Dastoor

RR by Anonymous Referee #2 (23 Oct 2023)

Suggestions for revision or reasons for rejection

Mechanism
I still have some concerns about the mechanism used. As I said in my previous review it has been pulled together from different mechanisms making different assumptions. It is mainly a puling together of mechanisms in pervious models with the inherent assumptions and implications that they have made.

It would be useful for the authors to reference their reactions either by the original lab study, from the IUPAC / JPL compilation or when not available indicate that the rate constant has been estimated. For example Novak and Wollesen de Jonge use the DMS+NO3 rate constant but the "original reference" for this comes from the IUPAC recommendation based on rate constant measured in the 1980s. The Saunders et al., 2003 reference doesn't discuss DMS.

The H abstraction path chemistry I'm a bit confused about. We get to CH3S and then this can primary react with O2 to form CH3S(OO) with a rate constant of 1.2e-16*exp(1580). However, the back reaction for this CH3S(OO)-->CH3S is missing from the reaction scheme but is in the MCM and is substantially faster. This wouldn't matter so much if the only fate of CH3S was reaction with O2 but it can also react with O3. Why is this reaction missing?

Why isn't the reaction between CH3SH2OOH and OH included as it is the the MCM? Is this not competative against the photolysis? What is J(41)

The mechanism then produces CH3S(OO) which in this mechanism has 2 fates. Decomposition to give CH3(O2) or decomposition to give CH3SO2. So I think CH3(O2) is meant to be CH3O2 as used previously here? The CH3S(OO) in this mechanism can also decompose to give CH3SO2. I don't see this reaction in the MCM (https://mcm.york.ac.uk/MCM/species/CH3SOO). There is a CH3SO2 in the MCM but its formed from CH3SOO2 rather than from CH3SOO. Where does the rate constant of 1s-1 for the rearrangement of CH3S(OO) Into CH3SO2 come from? Can the authors clarify what is going on here?

What happens to the CH3SCH2O which is formed from the CH3SCH2OOH photolysis?

Where the authors have ignored reactions in the MCM chemistry they should indicate why and give new references. It would be useful if the text of the paper described the choices use but the authors in constructing the scheme. What is the basis of the reactions?

Also can there consistent representation of the numbers in the tables. Sometimes the multiplication is represented by a * sometimes x sometime exponentials are represented as exp sometimes as e. There is a missing subscript on the O2 on the addition pathway. Why is there units on this and not the other reactions? The table labels these as rates but they are rate constants. Is there a need to give 5 significant figures on the HPMTF rate constants and 3 on the others?

Perhaps the authors could use the model given by the reaction table in https://acp.copernicus.org/preprints/acp-2023-42/acp-2023-42.pdf to give the appropriate sources for the reactions? It would also probably help if they considered this mechanism and thought about whether theirs is consistent?

A diagram of the reaction mechanism would help to clarify the mechanism.

Other aspects.
There is quite a large change in DMS burden (38%) from the inclusion of the new mechanism. The BrO only constitutes ~20% of the total loss now. It would be useful to assess how the OH, NO3, O3, BrO concentrations have changed between the model simulations. How much of the change in the burden is the additional routes and how much is a change in the oxidant concentrations? This isn't clear to me and without some sense of how the oxidants have changed it is hard to tell.

Hide

RR by Anonymous Referee #1 (23 Oct 2023)

ED: Reconsider after major revisions (05 Nov 2023) by Ashu Dastoor

AR by Linia Tashmim on behalf of the Authors (16 Dec 2023) Author's response Author's tracked changes Manuscript

ED: Publish as is (05 Jan 2024) by Ashu Dastoor

AR by Linia Tashmim on behalf of the Authors (12 Jan 2024)

Journal article(s) based on this preprint

19 Mar 2024

Contribution of expanded marine sulfur chemistry to the seasonal variability of dimethyl sulfide oxidation products and size-resolved sulfate aerosol

Linia Tashmim, William C. Porter, Qianjie Chen, Becky Alexander, Charles H. Fite, Christopher D. Holmes, Jeffrey R. Pierce, Betty Croft, and Sakiko Ishino

Atmos. Chem. Phys., 24, 3379–3403, https://doi.org/10.5194/acp-24-3379-2024,https://doi.org/10.5194/acp-24-3379-2024, 2024

Short summary

Linia Tashmim, William C. Porter, Qianjie Chen, Becky Alexander, Charles H. Fite, Christopher D. Holmes, Jeffrey R. Pierce, Betty Croft, and Sakiko Ishino

Viewed

Total article views: 857 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
557	262	38	857	18	21

HTML: 557
PDF: 262
XML: 38
Total: 857
BibTeX: 18
EndNote: 21

Views and downloads (calculated since 12 Jun 2023)

Month	HTML	PDF	XML	Total
Jun 2023	186	92	10	288
Jul 2023	88	41	8	137
Aug 2023	51	21	1	73
Sep 2023	34	20	2	56
Oct 2023	65	34	12	111
Nov 2023	18	12	0	30
Dec 2023	22	11	3	36
Jan 2024	38	6	1	45
Feb 2024	30	17	1	48
Mar 2024	25	8	0	33
Apr 2024	0
May 2024	0
Jun 2024	0
Jul 2024	0
Aug 2024	0
Sep 2024	0

Cumulative views and downloads (calculated since 12 Jun 2023)

Month	HTML	PDF	XML	Total
Jun 2023	186	92	10	288
Jul 2023	88	41	8	137
Aug 2023	51	21	1	73
Sep 2023	34	20	2	56
Oct 2023	65	34	12	111
Nov 2023	18	12	0	30
Dec 2023	22	11	3	36
Jan 2024	38	6	1	45
Feb 2024	30	17	1	48
Mar 2024	25	8	0	33
Apr 2024	0
May 2024	0
Jun 2024	0
Jul 2024	0
Aug 2024	0
Sep 2024	0

Viewed (geographical distribution)

Total article views: 847 (including HTML, PDF, and XML) Thereof 847 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 18 Sep 2024

Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (1869 KB)
Metadata XML

Short summary

Dimethyl sulfide (DMS) is mostly emitted from ocean surfaces and represents the largest natural source of sulfur to the atmosphere. Once in the atmosphere DMS forms stable oxidation products such as SO₂ and H₂SO₄ which can subsequently contribute to airborne particle formation and growth. In this study we update the DMS oxidation mechanism in the chemical transport model GEOS-Chem and describe resulting changes in particle growth as well as the overall global sulfur budget.


Total:	0
HTML:	0
PDF:	0
XML:	0