the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 03 Dec 2024
Quantifying the decay rate of volcanic sulfur dioxide in the stratosphere
Abstract. The injection of sulfur dioxide (SO2) into the stratosphere and its subsequent oxidation to form sulfate aerosols after large volcanic eruptions can have profound effects on Earth’s climate. The lifetime of volcanic SO2 in the stratosphere is thought to be determined by its gas-phase oxidation by the hydroxyl radical (OH); once oxidized, it goes on to form sulfate aerosols. However, it has also been suggested that heterogeneous oxidation on ash could also be important or even dominant, which would imply faster formation of aerosols at least in ash-rich plumes. Additionally, recent work uses an assumed exponential fit to determine the total SO2 mass loading following large eruptions; the quality of this fit translates directly to the accuracy of the mass loading estimate. It is therefore of interest to examine how accurately the SO2 lifetime can be determined from observations, and compare observations to models. Here we evaluate the SO2 lifetime and its uncertainties following several significant eruptions using three different sets of satellite observations and compare these to the CESM-WACCM6 model. We show that defining an accurate baseline against which a volcanic injection can be quantified limits accuracy in the estimated lifetime for some satellite data sets. We find that uncertainties in lifetimes across different altitudes and eruptions make it difficult to attribute variations in lifetime to specific SO2 removal processes for the events examined.
-
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
(2526 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2526 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
EC1: 'Preliminary Editor Comment on egusphere-2024-3525', Matthew Toohey, 13 Dec 2024
-
AC1: 'Reply on EC1', Paul Nicknish, 29 Dec 2024
The authors thank Dr. Toohey for his insightful comments on the paper. He brings up an excellent point regarding the wording on the removal of volcanic SO2, and the authors will adjust the language as suggested in the manuscript in future drafts. The point of the second comment is also well taken, and the authors are planning to incorporate a deeper analysis and discussion of mean-lifetime vs decay rate in future versions of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2024-3525-AC1 - AC5: 'Reply on AC1', Paul Nicknish, 09 Apr 2025
- AC6: 'Reply on AC1', Paul Nicknish, 09 Apr 2025
-
AC1: 'Reply on EC1', Paul Nicknish, 29 Dec 2024
-
RC1: 'Comment on egusphere-2024-3525', Anonymous Referee #1, 03 Jan 2025
In this manuscript, Nicknish et al. describe their investigations on stratospheric decay timescales of sulfur dioxide (SO₂) following explosive volcanic eruptions, focusing on three specific events: Kasatochi in August 2008, Sarychev in June 2009, and Nabro in June 2011. The authors employ datasets from two space-borne limb-sounders (MLS and MIPAS), one nadir-sounder (OMI), as well as model simulations using WACCM. To my knowledge, this study represents the first comprehensive analysis integrating and comparing these datasets to assess the decay of SO₂ in the stratosphere. Given potential scenarios of climate engineering involving the deliberate injection of SO₂ into the stratosphere, understanding its conversion into sulfate aerosols and improving the fidelity of related models is of critical importance. The manuscript provides a clear and transparent description of the methodologies and procedures applied to analyze the datasets. The sensitivity analyses conducted to quantify uncertainties in the decay timescale calculations are robust and methodologically sound. The results are discussed in a manner that elucidates the strengths and limitations of each measurement technique, offering valuable insights to the reader. However, some conclusions would benefit from clearer articulation and some of a more nuanced presentation. These issues are detailed in the specific comments provided below. While the study does not yield definitive recommendations for specific model improvements, the work enhances our understanding of the specific characteristics of different stratospheric SO₂ measurements and provides a valuable framework for future research. I support the publication of this manuscript following revisions to address the comments and suggestions outlined below.
Specific comments:
l.22: ‘…is produced naturally in seawater’:
This is only one of the several sources of OCS. It is even not clear if the major source is direct emission from sea-water, or if OCS is mainly produced by oxidation of CS2 and DMS (which originate in sea-water) (e.g. Kremser et al., 2016). Please be either more specific (or more general) here.
l.81: ‘…they potentially provide greater sensitivity to volcanic SO2’; l.86: ‘the greater sensitivity of limb sounders may be advantageous’; l.332: ‘This is likely to be a bias in the OMI data, perhaps due to the limited sensitivity of nadir instruments as the plume disperses…‘; l.399: ‘This may be a bias…’
I acknowledge the scientific caution when attributing biases in stratospheric SO₂ decay timescales to nadir-viewing instruments. However, the evidence presented in this study, as well as in previous works, strongly supports the existence of such a bias. This bias arises from the limited sensitivity of nadir-viewing instruments to diluted stratospheric SO₂ amounts compared to limb-viewing instruments, which benefit from significantly longer optical path lengths (several hundred times greater) through the dispersed plume. Given the robustness of this evidence, I recommend adopting clearer language to describe this phenomenon. Such clarity is important to avoid potential misinterpretation of nadir-viewing data, particularly in the future, when a substantial volume of nadir-derived data remains available while limb-measurements might no longer be conducted.
l.98: ‘high spectral resolution’
Should this not read ‘high spatial resolution’?
l.106: ‘are reported on pressure levels with an approximate spacing of…’
The reported retrieval-grid of remote sensing data is generally not equal to its spatial resolution. Therefore, I would suggest to add the information on resolution from here: https://mls.jpl.nasa.gov/data/v5-0_data_quality_document.pdf, p. 157
l.146: ‘to the pressure coordinate of the MIPAS data’
Shouldn’t this read ‘to the altitude coordinate…’?
l.236: ‘The reason for this remains unclear, and no explanation or documentation for this difference was found in the literature.’
You might try to contact the MLS-team for a possible explanation(?)
Chapter 3.2: ‘Background seasonal cycle in the MLS data’
Is there any possibility to infer the disturbing seasonal cycle from comparing different years with less volcanic influence?
l.290, Table 1:
I strongly recommend consolidating all the information from Table 1 and Table B1 into a single comprehensive table. Additionally, I suggest including the uncertainties reported in Höpfner et al. (2015) and the results from the WACCM simulations, including those for the 18–22 km layer if available. This enhancement would greatly improve the readability of the discussion in Chapter 3.3, allowing readers to follow the analysis more easily without needing to consult multiple tables.
l.290: ‘Their values show a clear increase of e-folding time with height, which is not as apparent in our results.’
However, the decay timescales provided here also show increasing values (with one exception of MLS in case of Kasatochi). Further, I suggest also to try to provide best-estimates from your analysis of the decay-timescales at 18-22 km for MIPAS in case of Kasatochi and Sarychev – when looking at Fig. 2 and Fig. A1, there seems to be a clear signal.
l.295-314: comparison to WACCM
The discussion may give the impression that the uncertainties in the decay timescales derived from the measurements are too large to allow for meaningful comparisons with the model. However, upon examining Figures 5, C1, and C2, it appears that for the Sarychev and Nabro eruptions, the model significantly underestimates the decay timescales compared to the limb-sounding datasets, whereas this discrepancy is less pronounced for the Kasatochi eruption. Could you provide possible explanations for this observed difference?
l.355: ‘our main focus here is on the decay times of the stratospheric inputs of the indicated eruptions and not the total stratospheric mass.’
The entire chapter 5 of the manuscript (as well as section 2.5 ‘Calculation of total stratospheric SO2 burden’) is dedicated to the ‘Estimating the stratospheric SO2 burden’. Therefore, I don’t understand this statement. I would suggest to delete this sentence and extend chapter 5 a bit by extending Table 2 to include the estimations of stratospheric SO2 mass by the extrapolation methods used by Pumphrey et al. (2015) and Höpfner et al. (2015). I would also suggest to add the results (M(t0)) from the fits performed in the present work. It should be made clear that the method described here in section 2.5 is not adequate to calculate the total stratospheric SO2 burden in case of MIPAS and, to a less extend, also for MLS.
l.400: ‘…and should be considered when analyzing volcanic SO2 with OMI’
I would suggest to add here: ‘and other nadir-sounding instruments’.
l.413-417: eruptions with ash
The eruption of Puyehue in June 2011 was also rich in ash. Have you tried to inspect that one for any effects on SO2 lifetime? (e.g. Griessbach et al., 2016, doi:10.5194/amt-9-4399-2016)
l.424: ‘Our work suggests that the current SO2 data reported by available observational products are subject to significant uncertainty when examining the stratospheric lifetime of volcanic SO2 and suggests that more precise data is needed if chemical mechanisms and SO2 mass loading following an eruption are to be elucidated using observed decay times.’
On one hand, I support this statement, particularly considering the imminent loss of limb-sounding capabilities for stratospheric SO₂ observations, which will create a significant gap in our ability to monitor the stratosphere. On the other hand, I find the statement somewhat overly general. As noted in the manuscript, each observational technique has specific advantages and limitations in quantifying stratospheric SO₂. Therefore, to effectively evaluate and refine models, it may be more appropriate to tailor comparisons to align with the strengths of each dataset. For example, model results could be compared directly with nadir and MLS data closer to the eruption time, while comparisons with IR limb-sounding datasets might be more suitable for periods several weeks after the eruption.
Citation: https://doi.org/10.5194/egusphere-2024-3525-RC1 -
AC2: 'Reply on RC1', Paul Nicknish, 07 Feb 2025
Thank you for your time and your detailed comments on the paper! Please see our responses attached in the pdf.
- AC4: 'Reply on AC2', Paul Nicknish, 09 Apr 2025
- AC7: 'Reply on AC2', Paul Nicknish, 09 Apr 2025
-
AC2: 'Reply on RC1', Paul Nicknish, 07 Feb 2025
-
RC2: 'Comment on egusphere-2024-3525', Anonymous Referee #2, 05 Jan 2025
Review of “Quantifying the decay rate of volcanic sulfur dioxide in the
Stratosphere” by Nicknish et al.
This is an interesting paper, but needs some additional explanations, justification of the approach and some data analysis.
The basic idea is to quantify the decay rate for SO2 following volcanic eruptions. The authors looked at MLS and MIPAS SO2 retrievals and did some additional comparisons with OMI SO2 observation. SO2 converts to H2SO4 which forms aerosols. Stratospheric aerosols have been linked to surface cooling so understanding the conversion process is important.
Major issues:
- I was hoping to see a ‘back of the envelope’ check against the total aerosol eruption mass burden – the end product of SO2 oxidation. This would be a useful addition. (See Schulte et al., 2023,https://doi.org/10.5194/AMT-680 16-3531-2023 on computing the total mass). There are a number of stratospheric aerosol sources you can use – but probably GLOSSAC is the best. This sort of ‘stupidity check’ would confirm that the SO2 estimates agree with aerosol production - which is why we care about this.
- The MLS algorithm, as I understand it, generates negative mixing ratios for data on the edge of observability. The correct way to deal with these are to average the data over larger regions including both positive and negative mixing ratios. I was looking for a discussion of this and mention of MLS validation also found in Livesey et al. (2022) (found at https://mls.jpl.nasa.gov/eos-aura-mls/documentation.php). Discussion of how to use the data including quality flag screening that is appropriate for SO2 is also found there and should be mentioned in the data description (lines 65-75). An equivalent MIPAS discussion is needed.
- The division of the SO2 into three separate regions (10-14), (14-18), (18-22) made me uncomfortable. At high latitudes in winter these regions are all in the stratosphere – in the summer the 10-14km may include the troposphere. In the tropics (as with Nabro and many other eruptions), only 18-22km is entirely in the stratosphere. This distinction can play an important part since the water vapor content and OH concentration (eq. R1)of these layers can be quite different – upper troposphere vs lower stratosphere – and thus will affect the decay rate. Since MLS and MIPAS also make water vapor measurements, the water vapor content can be added to the analysis. It seems to me that the the authors should have used two layers - below the tropopause and above the tropopause - rather than what was done here. It is easy to get tropopause height information from reanalysis data sets (GFS, MERRA2, ERA5).
- The authors neglect the transport between the layers. Exchange between layers needs to be discussed as possibly influencing the decay rate.
- The authors are using a zonal mean SO2 on a 10° latitude grid (ln 153). It seems like they could also construct a tighter latitude grid (say 5°) and a longitude grid and select high SO2 regions which might reduce the uncertainty (e.g. Fig. 2). I would like to see how this affects their decay rate and agreement between the two satellite instruments.
- I don’t think adding OMI SO2 helps the paper at all. In fact, it just adds noise, not signal. This is because the OMI total column includes massive amounts of tropospheric SO2 (for most eruptions) which – as the authors note – probably explains the significant differences in total SO2 mass and decay rate. If you add OMI you might as well take a look at SO2 measurements from AIRS (mentioned line 71) as well for completeness. Anyway, I suggest you just drop this section – it really adds nothing.
- The WACCM simulation is interesting but barely discussed (Fig. 5). Take a look at the water vapor in WACCM. Was it the same as MLS observations? This might explain the accelerated decrease.
Minor comments:
Add layer labels to Fig. 2
Line 287 ‘less uncertain’ - how about ‘better’
Fig 4. Why not connect the dots vertically. The figure – as is, is a little hard to read.
Line 418 “Honga-Tonga” - the APARC group recommends using ‘Honga’ not HTHH or HT or other acronyms. The Honga eruption is a good example where hydrolysis probably played a critical role in accelerating the decay of SO2 and conversion to aerosols as noted. This is why I recommend the authors also take a look at H2O in other regions.
Citation: https://doi.org/10.5194/egusphere-2024-3525-RC2 -
AC3: 'Reply on RC2', Paul Nicknish, 07 Feb 2025
Thank you for your time and the detailed comments on the paper. Please see our responses in the attached pdf.
- AC8: 'Reply on AC3', Paul Nicknish, 09 Apr 2025
Interactive discussion
Status: closed
-
EC1: 'Preliminary Editor Comment on egusphere-2024-3525', Matthew Toohey, 13 Dec 2024
As part of the open discussion, I am posting a brief editor comment which I hope will be useful to the authors, as well as for the referees to consider as they review the manuscript.
The main topic of interest of the study is the persistence of volcanic sulfur dioxide in the stratosphere. The title of the manuscript refers to a “decay rate”, however this term is used sparingly in the rest of the manuscript, instead we see many uses of the terms “lifetime” and “e-folding time”. This seems to imply that these three terms are synonymous. Indeed, in much of the relevant literature, one finds these terms (as well as related terms like “residence time”) used interchangeably. They should not be. These terms refer to measures of persistence that are related, but distinct from each other.If a conservative tracer is injected into a reservoir at time t=0 and we measure the total amount of tracer in the reservoir as a function of time, our measurement time series might be called a “washout function”. The “e-folding time” is simply the time elapsed from t=0 when the amount of tracer reaches 1/e=0.368 of the injected amount. The “mean lifetime” is the average time that one molecule of tracer spends in the reservoir, which can be computed from the washout function. In both cases, it does not matter what functional form the washout function takes—it could be an exponential decay, a linear decay, or any other form, the e-folding time and mean lifetime can be computed from the washout function. Finally, the “decay timescale” describes the rate of exponential decay of the washout function. If the decay is perfectly exponential in nature, this fitting procedure will produce a single decay timescale value no matter the period over which the fit is performed. If the decay is not purely exponential, the fitting procedure will produce different decay timescale values depending on the period over which the fit is performed. Importantly, the e-folding time and mean lifetime are always single-valued quantities while the decay timescale could be a single value if applied to the full washout function, or could itself a function of time if the fit is performed over a moving window.
Some of these ideas are elaborated on in a manuscript currently in review of which I am an author (Toohey et al, 2024) in the context of analysis of volcanic stratospheric aerosol. The definitions of these quantities are certainly can be found in many sources, but I point out this particular manuscript in hopes that it might be useful to the authors.
To improve the analysis of SO2, I suggest:
1. What is calculated through the analysis presented in the present manuscript (e.g., as shown in Fig. 1) is I think best called a decay timescale, not an e-folding time, and to avoid confusion the authors should in potential future versions of the manuscript adjust all descriptions of their results as a computation of a decay timescale rather than other terms.
2. A main conclusion of the study (translated in terminology) is that since the decay timescale shows a range of values as the fitting window is varied for any individual eruption, the mean lifetime of SO2 in the stratosphere (for each eruption) is uncertain. This conclusion erroneously equates decay timescale with mean lifetime, which would only be true for a perfectly well-mixed reservoir, which is clearly not the case for the stratosphere. I imagine that there is considerable uncertainty in the mean SO2 lifetime, but that its quantification and our understanding of its physical underpinnings could be better described with some adjustments to the analysis, and a more carefully constructed framework in which to interpret variations in decay timescale and its implications.
ReferencesToohey, M., Jia, Y., Khanal, S., and Tegtmeier, S.: Stratospheric residence time and the lifetime of volcanic stratospheric aerosols, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2024-2400, 2024.
Citation: https://doi.org/10.5194/egusphere-2024-3525-EC1 -
AC1: 'Reply on EC1', Paul Nicknish, 29 Dec 2024
The authors thank Dr. Toohey for his insightful comments on the paper. He brings up an excellent point regarding the wording on the removal of volcanic SO2, and the authors will adjust the language as suggested in the manuscript in future drafts. The point of the second comment is also well taken, and the authors are planning to incorporate a deeper analysis and discussion of mean-lifetime vs decay rate in future versions of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2024-3525-AC1 - AC5: 'Reply on AC1', Paul Nicknish, 09 Apr 2025
- AC6: 'Reply on AC1', Paul Nicknish, 09 Apr 2025
-
AC1: 'Reply on EC1', Paul Nicknish, 29 Dec 2024
-
RC1: 'Comment on egusphere-2024-3525', Anonymous Referee #1, 03 Jan 2025
In this manuscript, Nicknish et al. describe their investigations on stratospheric decay timescales of sulfur dioxide (SO₂) following explosive volcanic eruptions, focusing on three specific events: Kasatochi in August 2008, Sarychev in June 2009, and Nabro in June 2011. The authors employ datasets from two space-borne limb-sounders (MLS and MIPAS), one nadir-sounder (OMI), as well as model simulations using WACCM. To my knowledge, this study represents the first comprehensive analysis integrating and comparing these datasets to assess the decay of SO₂ in the stratosphere. Given potential scenarios of climate engineering involving the deliberate injection of SO₂ into the stratosphere, understanding its conversion into sulfate aerosols and improving the fidelity of related models is of critical importance. The manuscript provides a clear and transparent description of the methodologies and procedures applied to analyze the datasets. The sensitivity analyses conducted to quantify uncertainties in the decay timescale calculations are robust and methodologically sound. The results are discussed in a manner that elucidates the strengths and limitations of each measurement technique, offering valuable insights to the reader. However, some conclusions would benefit from clearer articulation and some of a more nuanced presentation. These issues are detailed in the specific comments provided below. While the study does not yield definitive recommendations for specific model improvements, the work enhances our understanding of the specific characteristics of different stratospheric SO₂ measurements and provides a valuable framework for future research. I support the publication of this manuscript following revisions to address the comments and suggestions outlined below.
Specific comments:
l.22: ‘…is produced naturally in seawater’:
This is only one of the several sources of OCS. It is even not clear if the major source is direct emission from sea-water, or if OCS is mainly produced by oxidation of CS2 and DMS (which originate in sea-water) (e.g. Kremser et al., 2016). Please be either more specific (or more general) here.
l.81: ‘…they potentially provide greater sensitivity to volcanic SO2’; l.86: ‘the greater sensitivity of limb sounders may be advantageous’; l.332: ‘This is likely to be a bias in the OMI data, perhaps due to the limited sensitivity of nadir instruments as the plume disperses…‘; l.399: ‘This may be a bias…’
I acknowledge the scientific caution when attributing biases in stratospheric SO₂ decay timescales to nadir-viewing instruments. However, the evidence presented in this study, as well as in previous works, strongly supports the existence of such a bias. This bias arises from the limited sensitivity of nadir-viewing instruments to diluted stratospheric SO₂ amounts compared to limb-viewing instruments, which benefit from significantly longer optical path lengths (several hundred times greater) through the dispersed plume. Given the robustness of this evidence, I recommend adopting clearer language to describe this phenomenon. Such clarity is important to avoid potential misinterpretation of nadir-viewing data, particularly in the future, when a substantial volume of nadir-derived data remains available while limb-measurements might no longer be conducted.
l.98: ‘high spectral resolution’
Should this not read ‘high spatial resolution’?
l.106: ‘are reported on pressure levels with an approximate spacing of…’
The reported retrieval-grid of remote sensing data is generally not equal to its spatial resolution. Therefore, I would suggest to add the information on resolution from here: https://mls.jpl.nasa.gov/data/v5-0_data_quality_document.pdf, p. 157
l.146: ‘to the pressure coordinate of the MIPAS data’
Shouldn’t this read ‘to the altitude coordinate…’?
l.236: ‘The reason for this remains unclear, and no explanation or documentation for this difference was found in the literature.’
You might try to contact the MLS-team for a possible explanation(?)
Chapter 3.2: ‘Background seasonal cycle in the MLS data’
Is there any possibility to infer the disturbing seasonal cycle from comparing different years with less volcanic influence?
l.290, Table 1:
I strongly recommend consolidating all the information from Table 1 and Table B1 into a single comprehensive table. Additionally, I suggest including the uncertainties reported in Höpfner et al. (2015) and the results from the WACCM simulations, including those for the 18–22 km layer if available. This enhancement would greatly improve the readability of the discussion in Chapter 3.3, allowing readers to follow the analysis more easily without needing to consult multiple tables.
l.290: ‘Their values show a clear increase of e-folding time with height, which is not as apparent in our results.’
However, the decay timescales provided here also show increasing values (with one exception of MLS in case of Kasatochi). Further, I suggest also to try to provide best-estimates from your analysis of the decay-timescales at 18-22 km for MIPAS in case of Kasatochi and Sarychev – when looking at Fig. 2 and Fig. A1, there seems to be a clear signal.
l.295-314: comparison to WACCM
The discussion may give the impression that the uncertainties in the decay timescales derived from the measurements are too large to allow for meaningful comparisons with the model. However, upon examining Figures 5, C1, and C2, it appears that for the Sarychev and Nabro eruptions, the model significantly underestimates the decay timescales compared to the limb-sounding datasets, whereas this discrepancy is less pronounced for the Kasatochi eruption. Could you provide possible explanations for this observed difference?
l.355: ‘our main focus here is on the decay times of the stratospheric inputs of the indicated eruptions and not the total stratospheric mass.’
The entire chapter 5 of the manuscript (as well as section 2.5 ‘Calculation of total stratospheric SO2 burden’) is dedicated to the ‘Estimating the stratospheric SO2 burden’. Therefore, I don’t understand this statement. I would suggest to delete this sentence and extend chapter 5 a bit by extending Table 2 to include the estimations of stratospheric SO2 mass by the extrapolation methods used by Pumphrey et al. (2015) and Höpfner et al. (2015). I would also suggest to add the results (M(t0)) from the fits performed in the present work. It should be made clear that the method described here in section 2.5 is not adequate to calculate the total stratospheric SO2 burden in case of MIPAS and, to a less extend, also for MLS.
l.400: ‘…and should be considered when analyzing volcanic SO2 with OMI’
I would suggest to add here: ‘and other nadir-sounding instruments’.
l.413-417: eruptions with ash
The eruption of Puyehue in June 2011 was also rich in ash. Have you tried to inspect that one for any effects on SO2 lifetime? (e.g. Griessbach et al., 2016, doi:10.5194/amt-9-4399-2016)
l.424: ‘Our work suggests that the current SO2 data reported by available observational products are subject to significant uncertainty when examining the stratospheric lifetime of volcanic SO2 and suggests that more precise data is needed if chemical mechanisms and SO2 mass loading following an eruption are to be elucidated using observed decay times.’
On one hand, I support this statement, particularly considering the imminent loss of limb-sounding capabilities for stratospheric SO₂ observations, which will create a significant gap in our ability to monitor the stratosphere. On the other hand, I find the statement somewhat overly general. As noted in the manuscript, each observational technique has specific advantages and limitations in quantifying stratospheric SO₂. Therefore, to effectively evaluate and refine models, it may be more appropriate to tailor comparisons to align with the strengths of each dataset. For example, model results could be compared directly with nadir and MLS data closer to the eruption time, while comparisons with IR limb-sounding datasets might be more suitable for periods several weeks after the eruption.
Citation: https://doi.org/10.5194/egusphere-2024-3525-RC1 -
AC2: 'Reply on RC1', Paul Nicknish, 07 Feb 2025
Thank you for your time and your detailed comments on the paper! Please see our responses attached in the pdf.
- AC4: 'Reply on AC2', Paul Nicknish, 09 Apr 2025
- AC7: 'Reply on AC2', Paul Nicknish, 09 Apr 2025
-
AC2: 'Reply on RC1', Paul Nicknish, 07 Feb 2025
-
RC2: 'Comment on egusphere-2024-3525', Anonymous Referee #2, 05 Jan 2025
Review of “Quantifying the decay rate of volcanic sulfur dioxide in the
Stratosphere” by Nicknish et al.
This is an interesting paper, but needs some additional explanations, justification of the approach and some data analysis.
The basic idea is to quantify the decay rate for SO2 following volcanic eruptions. The authors looked at MLS and MIPAS SO2 retrievals and did some additional comparisons with OMI SO2 observation. SO2 converts to H2SO4 which forms aerosols. Stratospheric aerosols have been linked to surface cooling so understanding the conversion process is important.
Major issues:
- I was hoping to see a ‘back of the envelope’ check against the total aerosol eruption mass burden – the end product of SO2 oxidation. This would be a useful addition. (See Schulte et al., 2023,https://doi.org/10.5194/AMT-680 16-3531-2023 on computing the total mass). There are a number of stratospheric aerosol sources you can use – but probably GLOSSAC is the best. This sort of ‘stupidity check’ would confirm that the SO2 estimates agree with aerosol production - which is why we care about this.
- The MLS algorithm, as I understand it, generates negative mixing ratios for data on the edge of observability. The correct way to deal with these are to average the data over larger regions including both positive and negative mixing ratios. I was looking for a discussion of this and mention of MLS validation also found in Livesey et al. (2022) (found at https://mls.jpl.nasa.gov/eos-aura-mls/documentation.php). Discussion of how to use the data including quality flag screening that is appropriate for SO2 is also found there and should be mentioned in the data description (lines 65-75). An equivalent MIPAS discussion is needed.
- The division of the SO2 into three separate regions (10-14), (14-18), (18-22) made me uncomfortable. At high latitudes in winter these regions are all in the stratosphere – in the summer the 10-14km may include the troposphere. In the tropics (as with Nabro and many other eruptions), only 18-22km is entirely in the stratosphere. This distinction can play an important part since the water vapor content and OH concentration (eq. R1)of these layers can be quite different – upper troposphere vs lower stratosphere – and thus will affect the decay rate. Since MLS and MIPAS also make water vapor measurements, the water vapor content can be added to the analysis. It seems to me that the the authors should have used two layers - below the tropopause and above the tropopause - rather than what was done here. It is easy to get tropopause height information from reanalysis data sets (GFS, MERRA2, ERA5).
- The authors neglect the transport between the layers. Exchange between layers needs to be discussed as possibly influencing the decay rate.
- The authors are using a zonal mean SO2 on a 10° latitude grid (ln 153). It seems like they could also construct a tighter latitude grid (say 5°) and a longitude grid and select high SO2 regions which might reduce the uncertainty (e.g. Fig. 2). I would like to see how this affects their decay rate and agreement between the two satellite instruments.
- I don’t think adding OMI SO2 helps the paper at all. In fact, it just adds noise, not signal. This is because the OMI total column includes massive amounts of tropospheric SO2 (for most eruptions) which – as the authors note – probably explains the significant differences in total SO2 mass and decay rate. If you add OMI you might as well take a look at SO2 measurements from AIRS (mentioned line 71) as well for completeness. Anyway, I suggest you just drop this section – it really adds nothing.
- The WACCM simulation is interesting but barely discussed (Fig. 5). Take a look at the water vapor in WACCM. Was it the same as MLS observations? This might explain the accelerated decrease.
Minor comments:
Add layer labels to Fig. 2
Line 287 ‘less uncertain’ - how about ‘better’
Fig 4. Why not connect the dots vertically. The figure – as is, is a little hard to read.
Line 418 “Honga-Tonga” - the APARC group recommends using ‘Honga’ not HTHH or HT or other acronyms. The Honga eruption is a good example where hydrolysis probably played a critical role in accelerating the decay of SO2 and conversion to aerosols as noted. This is why I recommend the authors also take a look at H2O in other regions.
Citation: https://doi.org/10.5194/egusphere-2024-3525-RC2 -
AC3: 'Reply on RC2', Paul Nicknish, 07 Feb 2025
Thank you for your time and the detailed comments on the paper. Please see our responses in the attached pdf.
- AC8: 'Reply on AC3', Paul Nicknish, 09 Apr 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 874 | 141 | 41 | 1,056 | 25 | 38 |
- HTML: 874
- PDF: 141
- XML: 41
- Total: 1,056
- BibTeX: 25
- EndNote: 38
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Kane Stone
Susan Solomon
Simon A. Carn
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2526 KB) - Metadata XML
As part of the open discussion, I am posting a brief editor comment which I hope will be useful to the authors, as well as for the referees to consider as they review the manuscript.
The main topic of interest of the study is the persistence of volcanic sulfur dioxide in the stratosphere. The title of the manuscript refers to a “decay rate”, however this term is used sparingly in the rest of the manuscript, instead we see many uses of the terms “lifetime” and “e-folding time”. This seems to imply that these three terms are synonymous. Indeed, in much of the relevant literature, one finds these terms (as well as related terms like “residence time”) used interchangeably. They should not be. These terms refer to measures of persistence that are related, but distinct from each other.
If a conservative tracer is injected into a reservoir at time t=0 and we measure the total amount of tracer in the reservoir as a function of time, our measurement time series might be called a “washout function”. The “e-folding time” is simply the time elapsed from t=0 when the amount of tracer reaches 1/e=0.368 of the injected amount. The “mean lifetime” is the average time that one molecule of tracer spends in the reservoir, which can be computed from the washout function. In both cases, it does not matter what functional form the washout function takes—it could be an exponential decay, a linear decay, or any other form, the e-folding time and mean lifetime can be computed from the washout function. Finally, the “decay timescale” describes the rate of exponential decay of the washout function. If the decay is perfectly exponential in nature, this fitting procedure will produce a single decay timescale value no matter the period over which the fit is performed. If the decay is not purely exponential, the fitting procedure will produce different decay timescale values depending on the period over which the fit is performed. Importantly, the e-folding time and mean lifetime are always single-valued quantities while the decay timescale could be a single value if applied to the full washout function, or could itself a function of time if the fit is performed over a moving window.
Some of these ideas are elaborated on in a manuscript currently in review of which I am an author (Toohey et al, 2024) in the context of analysis of volcanic stratospheric aerosol. The definitions of these quantities are certainly can be found in many sources, but I point out this particular manuscript in hopes that it might be useful to the authors.
To improve the analysis of SO2, I suggest:
1. What is calculated through the analysis presented in the present manuscript (e.g., as shown in Fig. 1) is I think best called a decay timescale, not an e-folding time, and to avoid confusion the authors should in potential future versions of the manuscript adjust all descriptions of their results as a computation of a decay timescale rather than other terms.
2. A main conclusion of the study (translated in terminology) is that since the decay timescale shows a range of values as the fitting window is varied for any individual eruption, the mean lifetime of SO2 in the stratosphere (for each eruption) is uncertain. This conclusion erroneously equates decay timescale with mean lifetime, which would only be true for a perfectly well-mixed reservoir, which is clearly not the case for the stratosphere. I imagine that there is considerable uncertainty in the mean SO2 lifetime, but that its quantification and our understanding of its physical underpinnings could be better described with some adjustments to the analysis, and a more carefully constructed framework in which to interpret variations in decay timescale and its implications.
References
Toohey, M., Jia, Y., Khanal, S., and Tegtmeier, S.: Stratospheric residence time and the lifetime of volcanic stratospheric aerosols, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2024-2400, 2024.