the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Updated merged SAGE - CCI- OMPS+ dataset for evaluation of ozone trends in the stratosphere
Viktoria F. Sofieva
Monika Szelag
Johanna Tamminen
Carlo Arosio
Alexei Rozanov
Mark Weber
Doug Degenstein
Adam Bourassa
Daniel Zawada
Michael Kiefer
Alexandra Laeng
Kaley A. Walker
Patrick Sheese
Daan Hubert
Michel Roozendael
Christian Retscher
Robert Damadeo
Jerry D. Lumpe
Abstract. In this paper, we present the updated SAGE-CCI-OMPS+ climate data record of monthly zonal mean ozone profiles. This dataset covers the stratosphere and combines measurements by 9 limb and occultation satellite instruments – SAGE II, OSIRIS, MIPAS, SCIAMACHY, GOMOS, ACE-FTS, OMPS-LP, POAM III and SAGE III/ISS. Compared to the original version of the SAGE-CCI-OMPS dataset (Sofieva et al., 2017), the update includes new versions of MIPAS, ACE-FTS, and OSIRIS datasets, and introduces data from additional sensors (POAM III, SAGE III/ISS) and retrieval processors (OMPS-LP).
In the paper, we show detailed intercomparisons of ozone profiles from different instruments and data versions, with a focus on the detection of possible drifts in the datasets. The SAGE-CCI-OMPS+ dataset has a better coverage of polar regions and of the upper troposphere and the lower stratosphere (UTLS) than the SAGE-CCI-OMPS dataset.
We also studied the influence of including new datasets on ozone trends, which are estimated using multiple linear regression. The changes in the merged dataset do not change the overall morphology of post-1997 ozone trends: statistically significant trends are observed in the upper stratosphere. The largest changes in ozone trends are observed in polar regions, especially in the Southern Hemisphere.
The updated SAGE-CCI-OMPS+ dataset contains profiles of deseasonalized anomalies and ozone concentrations from 1984 to 2021, in 10° latitude bins from 90° S to 90° N, and in the altitude range from 10 km to 50 km. The dataset is in open access at https://climate.esa.int/en/projects/ozone/data/ and at ftp://cci_web@ftp-ae.oma.be/esacci (ESA Climate Office, last access: 10 August 2022).
- Preprint
(3635 KB) -
Supplement
(2166 KB) - BibTeX
- EndNote
Viktoria F. Sofieva et al.
Status: closed
-
RC1: 'Good description of updated ozone profile dataset', Anonymous Referee #1, 06 Dec 2022
This is a good and comprehensive description of an updated ozone profile data set, generated by merging all major relevant single satellite instrument data sets. The paper shows the small changes arising from using slightly different combinations and/or versions. The provided results are very plausible, and the changes in derived ozone trends are minor and consistent with previous knowledge.
The presentation is clear, but the English could be improved / copy-edited.
There is a minor glitch in the title of the bottom panel of Fig. 10. Should be 0.5*(OMPS UBr + OMPS USask).
I think this is a good paper for AMT and can be published as is.
Citation: https://doi.org/10.5194/amt-2022-313-RC1 -
AC1: 'Reply on RC1', Viktoria Sofieva, 03 Feb 2023
Dear reviewer,
Thank you very much for your very positive comments on our paper.
In the revised version, we corrected the title of the bottom panel of Figure 10.
Citation: https://doi.org/10.5194/amt-2022-313-AC1
-
AC1: 'Reply on RC1', Viktoria Sofieva, 03 Feb 2023
-
RC2: 'Comment on amt-2022-313', Anonymous Referee #2, 02 Jan 2023
Review of manuscript amt-2022-313, “Updated merged SAGE - CCI- OMPS+ dataset for evaluation of ozone” by V. Sofieva et al.
This manuscript describes an updated version of the merged profile ozone data set SAGE-CCI-OMPS, which is one of several merged ozone records used widely within the ozone science and trends community. The updated merged record includes several new data versions of data used in the current record as well as new data sources. The authors provide a thorough summary of the version changes for the relevant instruments, and an analysis of the impact of each change on the final trend estimates. The manuscript is well written (some copy-editing suggestions included below), logically organized and easy to follow. I recommend publication after some very minor comments are addressed.
Minor Comments:
When comparing the versions of the data for MIPAS and especially for ACE-FTS, I suggest cutting the figures off at 55km, or at least no higher than 60km. The ACE-FTS plots go up to 90km, which is far above the range of the merged record and thus not relevant to these results. Expanding the vertical scale will allow the authors to better highlight the regions that matter most to the merged product. In particular in ACE-FTS, there appears to be a seasonal component in the differences, with the mid-stratospheric increase from the old to new version over a broader vertical range in summer compared to winter.
There also seems to be a bit of a trend relative to MLS in the new ACE-FTS version compared to MLS (Figure 4 bottom panel, also Fig. S1 but not as clear) starting in mid-2019 that might be worth noting as this would also contribute to a more positive trend.
Page 18, Lines 15-20: For the details of reconstructing the merged record in absolute units (adding a seasonal cycle back in) and the error analysis, the readers are referred to Section 3.1 and the description of the original merged SAGE-CCI-OMPS. There is no discussion that I see concerning the error estimates, do the authors mean to refer to the 2017 paper? Did the inclusion of the new data lead to any notable changes in estimated uncertainties in the merged record? While the trend analysis including each instrument is very informative, a comment about any changes (or no changes) in uncertainty would be useful.
As for the seasonal cycle, section 3.1 says “For the SAGE-CCI-OMPS, the amplitude of the seasonal cycle is estimated using MIPAS measurements because they provide all season pole-to-pole measurements with dense sampling. The absolute values of the seasonal cycle are estimated from SAGE II and OSIRIS in the overlapping period (which are very close to each other and to GOMOS measurements), thus preserving the consistency in the dataset through the whole observation period.” I’m not sure I follow the use of SAGE II and OSIRIS here, are SAGE II and OSIRIS seasonal cycles checked against the MIPAS seasonal cycle in their respective overlap periods with MIPAS, thus verifying using the MIPAS seasonal cycle over the full record is valid, or is the seasonal cycle from SAGE II and OSIRIS used directly? In any case, it appears there may be seasonal changes in some of the new data versions, an update to the seasonal cycles in Figure 4 of the 2017 paper would be useful as a supplemental figure that can be referred to in the text to support the representativeness of MIPAS to establish the seasonal cycle in the new version.
When discussing trend results I assume all results are for the second portion of the piecewise linear fit (since 1997) as opposed to a linear fit over the full ‘hockey-stick’ proxy, but this should be specified in the text.
Typos/Editorial Suggestions
Page 2
L2 The importance of monitoring stratospheric ozone and its vertical structure is well recognized …
L8 The main advantages of satellite …
L10 … instruments is limited, data from several instruments …
Page 3
L4 … ozone profiles are retrieved on a geometric altitude grid …
L5 presented on an altitude grid from 10 to 50 km.
L6 (upper troposphere and lower stratosphere)
L11 … we used ozone profile datasets …
L14 … we used altitude gridded datasets (HARMOZ_ALT), available …
L18 Add space before “Below”
Page 4
L9 described in Boone et al. (2020),
L14: “Sheese et al. (2022) showed that v4.1 ozone data bias with respect to data sets” do the authors mean with respect to independent data sets?
L23-24: This leads to less instrument drift in the retrieved ozone values.
Page 5
L6: OSIRIS measurements are used to produce three long term data records: vertically …
L7: … upper troposphere; recently these processing chains …
L21: authors of Brion et al. (1993), Daumont et al. (1992) and Malicet et al. (1995),
L23-24: described by Rieger et al. (2019).
L25: https://arg.usask.ca/docs/osiris_v7/index.html (last access: 09 October 2022).
Page 6
L24: the “AO3” ozone product used here is derived from measurements
Page 7
L6-7: within 5% in the stratosphere, increasing …
L9: “which results in random errors of more than about 10%” do the authors mean ‘less than’ 10% here, or is ‘more than’ correct? It reads as though the less than 10% of ozone profiles suffer from sunspot-related artifacts should lead to lower random errors, if the more than 10% is correct, this should be re-worded.
Page 8
L30: “retrieval grid width” consider adding vertical for clarity… retrieval vertical grid width
Page 9
L2: “(~10 %)” suggest changing to (~ +/- 10%) to clarify positive and negative range
L6: ACE-FTC -> ACE-FTS
L15: ozone profiles cover a larger altitude range
Page 14
Fig. 7 Caption: typo - Panel B is OMPS UBr.
Page 16
L5: within the ~60°S–60°N latitude range or within~60°S–60°N latitude (remove range)
L6: POAM III and SAGE III/ISS
Page 18
L3: How is good agreement defined? Consider re-wording as We use only the data that do not exhibit significant offset or drift with respect to …
L6: consider (illustrations and discussion of these data exclusions are presented in Sofieva et al. (2017)).
Page 21
L5 and L8: Figures S7-S12 should be Figures S8-S13
L10: (consistent with … )
Page 22
Fig. 15 caption: at the 95% confidence level
Page 23
L16: do not change the overall morphology of trends in ozone profiles: statistically significant trends …
L19: The profiles of ozone concentrations and deseasonalized anomalies are presented on an altitude grid …
Page 24
L4: framework of the ESA ALGOM project
Figure S3 caption: NLS -> MLS
Figure S5 and S6: specify this is OMPS UBr in the figure or caption
Citation: https://doi.org/10.5194/amt-2022-313-RC2 - AC2: 'Reply on RC2', Viktoria Sofieva, 04 Feb 2023
-
RC3: 'Comment on amt-2022-313', Anonymous Referee #3, 17 Jan 2023
The SAGE – CCI – OMPS data set is one of several important merged satellite data sets that have contributed to our understanding of ozone variability and trends in recent decades. The current paper describes improvements to the dataset made to create the new SAGE – CCI – OMPS+ data set, including new versions of existing data products as well as new instruments contributing to the data set. This straightforward paper is well-written, and AMT readership will appreciate the clear documentation of changes made since the previous version of the data. I have no major comments, but have a list of suggested changes and grammar revisions to improve the clarity of the manuscript.
Page 3, line 11 – “ozone profiles datasets” -> “ozone profile datasets”
Table 1: I suggest including a column with the reference to the appropriate instrument paper for the version of the data you are using for each instrument. Also, can you please check the vertical resolution numbers here? Saying these instruments have ~1 km vertical resolution seems awfully optimistic. Just because the vertical retrieval grid is at 1 km doesn’t mean the vertical resolution is, as I’m sure the authors are aware. For example, I believe the OMPS vertical resolution should be more like 2-3 km, not ~1 km.
Page 3, line 16: I was originally confused by why you were mentioning both the UBr and USask retrievals. The reasoning for this is made clear later in the paper (i.e., that they have slightly different coverage that you are trying to exploit). In this section, it might be nice to have a sentence listing the new data sets you are using, and mentioning why you are adding a second OMPS data set.
Page 3, Table 1: Put a comma after Ubr v 3.3
Page 4, line 9 – Fix parentheses on Boone reference
Page 6, lines 14-16: You state the systematic uncertainty below 20 km and above 50 km here – what about 20 – 50 km?
Page 8, eq 1: Is there a motivation for expressing anomalies as a fraction rather than absolute values (i.e., why divide by rho_m)? Can you briefly comment/justify this choice?
Page 8, lines 20-24: These two sentences are confusing. Is the second sentence referring to the new SAGE – CCI – OMPS+ data set, and saying you are doing something different than the origina SAGE-CCI-OMPS data set (as described in the first sentence)? If so, you could start the sentence with something like “In SAGE-CCI-OMPS+, …”
Page 9, line 6: “ACE-FTC” -> “ACE-FTS”
Page 9, line 6 (and elsewhere): Why do you refer to version 4.1/4.2 of the ACE-FTS data? Shouldn’t it be one or the other?
Page 12, Figure 5: Panel a has a different x-axis, please make x-axes the same in this figure.
Page 13, line 23: Drifts and steps are very different to me. Can you point more specifically to where possible steps occur.
Fig. S3 caption: “NLS” -> “MLS”
Figs S5-S6: I think this is UBr OMPS. Please specify in captions.
Page 17, line 6: A colon is more appropriate here than a comma.
Page 18, line 3: “…only the data, which…” -? “… only data that …”
Page 18, line 7: “As mentioned above…” – I don’t see where it was mentioned that you exclude UBr data 14 – 16 km and 20S-20N. But maybe I’m missing something. Although you do identify UBr as problematic in this region, I think this is the first explicit mention that you are not using data in that region. Also, when I came to this sentence I began to question both the choice of masking region as well as the entire motivation for using UBr at all. It seems like the main motivation for using UBr is to get the additional data in the tropics that USask doesn’t have, but then you say you are not actually using it here. Also, is the exact choice of masking really sufficient? It would be pretty easy to answer this by showing a latitude-altitude plot of the trend in the difference between UBr and USask. This would much more clearly help justify (or not) the choice of what levels and altitudes to remove from the analysis.
Page 23, line 17: What about the negative trends in the LS. It looks like some of those are significant, and that there are maybe a few more grid boxes that are significant in the new data set in comparison to the previous data set?
Citation: https://doi.org/10.5194/amt-2022-313-RC3 - AC3: 'Reply on RC3', Viktoria Sofieva, 04 Feb 2023
Status: closed
-
RC1: 'Good description of updated ozone profile dataset', Anonymous Referee #1, 06 Dec 2022
This is a good and comprehensive description of an updated ozone profile data set, generated by merging all major relevant single satellite instrument data sets. The paper shows the small changes arising from using slightly different combinations and/or versions. The provided results are very plausible, and the changes in derived ozone trends are minor and consistent with previous knowledge.
The presentation is clear, but the English could be improved / copy-edited.
There is a minor glitch in the title of the bottom panel of Fig. 10. Should be 0.5*(OMPS UBr + OMPS USask).
I think this is a good paper for AMT and can be published as is.
Citation: https://doi.org/10.5194/amt-2022-313-RC1 -
AC1: 'Reply on RC1', Viktoria Sofieva, 03 Feb 2023
Dear reviewer,
Thank you very much for your very positive comments on our paper.
In the revised version, we corrected the title of the bottom panel of Figure 10.
Citation: https://doi.org/10.5194/amt-2022-313-AC1
-
AC1: 'Reply on RC1', Viktoria Sofieva, 03 Feb 2023
-
RC2: 'Comment on amt-2022-313', Anonymous Referee #2, 02 Jan 2023
Review of manuscript amt-2022-313, “Updated merged SAGE - CCI- OMPS+ dataset for evaluation of ozone” by V. Sofieva et al.
This manuscript describes an updated version of the merged profile ozone data set SAGE-CCI-OMPS, which is one of several merged ozone records used widely within the ozone science and trends community. The updated merged record includes several new data versions of data used in the current record as well as new data sources. The authors provide a thorough summary of the version changes for the relevant instruments, and an analysis of the impact of each change on the final trend estimates. The manuscript is well written (some copy-editing suggestions included below), logically organized and easy to follow. I recommend publication after some very minor comments are addressed.
Minor Comments:
When comparing the versions of the data for MIPAS and especially for ACE-FTS, I suggest cutting the figures off at 55km, or at least no higher than 60km. The ACE-FTS plots go up to 90km, which is far above the range of the merged record and thus not relevant to these results. Expanding the vertical scale will allow the authors to better highlight the regions that matter most to the merged product. In particular in ACE-FTS, there appears to be a seasonal component in the differences, with the mid-stratospheric increase from the old to new version over a broader vertical range in summer compared to winter.
There also seems to be a bit of a trend relative to MLS in the new ACE-FTS version compared to MLS (Figure 4 bottom panel, also Fig. S1 but not as clear) starting in mid-2019 that might be worth noting as this would also contribute to a more positive trend.
Page 18, Lines 15-20: For the details of reconstructing the merged record in absolute units (adding a seasonal cycle back in) and the error analysis, the readers are referred to Section 3.1 and the description of the original merged SAGE-CCI-OMPS. There is no discussion that I see concerning the error estimates, do the authors mean to refer to the 2017 paper? Did the inclusion of the new data lead to any notable changes in estimated uncertainties in the merged record? While the trend analysis including each instrument is very informative, a comment about any changes (or no changes) in uncertainty would be useful.
As for the seasonal cycle, section 3.1 says “For the SAGE-CCI-OMPS, the amplitude of the seasonal cycle is estimated using MIPAS measurements because they provide all season pole-to-pole measurements with dense sampling. The absolute values of the seasonal cycle are estimated from SAGE II and OSIRIS in the overlapping period (which are very close to each other and to GOMOS measurements), thus preserving the consistency in the dataset through the whole observation period.” I’m not sure I follow the use of SAGE II and OSIRIS here, are SAGE II and OSIRIS seasonal cycles checked against the MIPAS seasonal cycle in their respective overlap periods with MIPAS, thus verifying using the MIPAS seasonal cycle over the full record is valid, or is the seasonal cycle from SAGE II and OSIRIS used directly? In any case, it appears there may be seasonal changes in some of the new data versions, an update to the seasonal cycles in Figure 4 of the 2017 paper would be useful as a supplemental figure that can be referred to in the text to support the representativeness of MIPAS to establish the seasonal cycle in the new version.
When discussing trend results I assume all results are for the second portion of the piecewise linear fit (since 1997) as opposed to a linear fit over the full ‘hockey-stick’ proxy, but this should be specified in the text.
Typos/Editorial Suggestions
Page 2
L2 The importance of monitoring stratospheric ozone and its vertical structure is well recognized …
L8 The main advantages of satellite …
L10 … instruments is limited, data from several instruments …
Page 3
L4 … ozone profiles are retrieved on a geometric altitude grid …
L5 presented on an altitude grid from 10 to 50 km.
L6 (upper troposphere and lower stratosphere)
L11 … we used ozone profile datasets …
L14 … we used altitude gridded datasets (HARMOZ_ALT), available …
L18 Add space before “Below”
Page 4
L9 described in Boone et al. (2020),
L14: “Sheese et al. (2022) showed that v4.1 ozone data bias with respect to data sets” do the authors mean with respect to independent data sets?
L23-24: This leads to less instrument drift in the retrieved ozone values.
Page 5
L6: OSIRIS measurements are used to produce three long term data records: vertically …
L7: … upper troposphere; recently these processing chains …
L21: authors of Brion et al. (1993), Daumont et al. (1992) and Malicet et al. (1995),
L23-24: described by Rieger et al. (2019).
L25: https://arg.usask.ca/docs/osiris_v7/index.html (last access: 09 October 2022).
Page 6
L24: the “AO3” ozone product used here is derived from measurements
Page 7
L6-7: within 5% in the stratosphere, increasing …
L9: “which results in random errors of more than about 10%” do the authors mean ‘less than’ 10% here, or is ‘more than’ correct? It reads as though the less than 10% of ozone profiles suffer from sunspot-related artifacts should lead to lower random errors, if the more than 10% is correct, this should be re-worded.
Page 8
L30: “retrieval grid width” consider adding vertical for clarity… retrieval vertical grid width
Page 9
L2: “(~10 %)” suggest changing to (~ +/- 10%) to clarify positive and negative range
L6: ACE-FTC -> ACE-FTS
L15: ozone profiles cover a larger altitude range
Page 14
Fig. 7 Caption: typo - Panel B is OMPS UBr.
Page 16
L5: within the ~60°S–60°N latitude range or within~60°S–60°N latitude (remove range)
L6: POAM III and SAGE III/ISS
Page 18
L3: How is good agreement defined? Consider re-wording as We use only the data that do not exhibit significant offset or drift with respect to …
L6: consider (illustrations and discussion of these data exclusions are presented in Sofieva et al. (2017)).
Page 21
L5 and L8: Figures S7-S12 should be Figures S8-S13
L10: (consistent with … )
Page 22
Fig. 15 caption: at the 95% confidence level
Page 23
L16: do not change the overall morphology of trends in ozone profiles: statistically significant trends …
L19: The profiles of ozone concentrations and deseasonalized anomalies are presented on an altitude grid …
Page 24
L4: framework of the ESA ALGOM project
Figure S3 caption: NLS -> MLS
Figure S5 and S6: specify this is OMPS UBr in the figure or caption
Citation: https://doi.org/10.5194/amt-2022-313-RC2 - AC2: 'Reply on RC2', Viktoria Sofieva, 04 Feb 2023
-
RC3: 'Comment on amt-2022-313', Anonymous Referee #3, 17 Jan 2023
The SAGE – CCI – OMPS data set is one of several important merged satellite data sets that have contributed to our understanding of ozone variability and trends in recent decades. The current paper describes improvements to the dataset made to create the new SAGE – CCI – OMPS+ data set, including new versions of existing data products as well as new instruments contributing to the data set. This straightforward paper is well-written, and AMT readership will appreciate the clear documentation of changes made since the previous version of the data. I have no major comments, but have a list of suggested changes and grammar revisions to improve the clarity of the manuscript.
Page 3, line 11 – “ozone profiles datasets” -> “ozone profile datasets”
Table 1: I suggest including a column with the reference to the appropriate instrument paper for the version of the data you are using for each instrument. Also, can you please check the vertical resolution numbers here? Saying these instruments have ~1 km vertical resolution seems awfully optimistic. Just because the vertical retrieval grid is at 1 km doesn’t mean the vertical resolution is, as I’m sure the authors are aware. For example, I believe the OMPS vertical resolution should be more like 2-3 km, not ~1 km.
Page 3, line 16: I was originally confused by why you were mentioning both the UBr and USask retrievals. The reasoning for this is made clear later in the paper (i.e., that they have slightly different coverage that you are trying to exploit). In this section, it might be nice to have a sentence listing the new data sets you are using, and mentioning why you are adding a second OMPS data set.
Page 3, Table 1: Put a comma after Ubr v 3.3
Page 4, line 9 – Fix parentheses on Boone reference
Page 6, lines 14-16: You state the systematic uncertainty below 20 km and above 50 km here – what about 20 – 50 km?
Page 8, eq 1: Is there a motivation for expressing anomalies as a fraction rather than absolute values (i.e., why divide by rho_m)? Can you briefly comment/justify this choice?
Page 8, lines 20-24: These two sentences are confusing. Is the second sentence referring to the new SAGE – CCI – OMPS+ data set, and saying you are doing something different than the origina SAGE-CCI-OMPS data set (as described in the first sentence)? If so, you could start the sentence with something like “In SAGE-CCI-OMPS+, …”
Page 9, line 6: “ACE-FTC” -> “ACE-FTS”
Page 9, line 6 (and elsewhere): Why do you refer to version 4.1/4.2 of the ACE-FTS data? Shouldn’t it be one or the other?
Page 12, Figure 5: Panel a has a different x-axis, please make x-axes the same in this figure.
Page 13, line 23: Drifts and steps are very different to me. Can you point more specifically to where possible steps occur.
Fig. S3 caption: “NLS” -> “MLS”
Figs S5-S6: I think this is UBr OMPS. Please specify in captions.
Page 17, line 6: A colon is more appropriate here than a comma.
Page 18, line 3: “…only the data, which…” -? “… only data that …”
Page 18, line 7: “As mentioned above…” – I don’t see where it was mentioned that you exclude UBr data 14 – 16 km and 20S-20N. But maybe I’m missing something. Although you do identify UBr as problematic in this region, I think this is the first explicit mention that you are not using data in that region. Also, when I came to this sentence I began to question both the choice of masking region as well as the entire motivation for using UBr at all. It seems like the main motivation for using UBr is to get the additional data in the tropics that USask doesn’t have, but then you say you are not actually using it here. Also, is the exact choice of masking really sufficient? It would be pretty easy to answer this by showing a latitude-altitude plot of the trend in the difference between UBr and USask. This would much more clearly help justify (or not) the choice of what levels and altitudes to remove from the analysis.
Page 23, line 17: What about the negative trends in the LS. It looks like some of those are significant, and that there are maybe a few more grid boxes that are significant in the new data set in comparison to the previous data set?
Citation: https://doi.org/10.5194/amt-2022-313-RC3 - AC3: 'Reply on RC3', Viktoria Sofieva, 04 Feb 2023
Viktoria F. Sofieva et al.
Viktoria F. Sofieva et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 340 | 83 | 16 | 439 | 37 | 8 | 6 |
- HTML: 340
- PDF: 83
- XML: 16
- Total: 439
- Supplement: 37
- BibTeX: 8
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1