the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Feb 2024
GREP reanalysis captures the evolution of the Arctic Marginal Ice Zone across timescales
Abstract. The recent development of data-assimilative reanalyses of the global ocean and sea ice enables a better understanding of the polar region dynamics and provides gridded descriptions of sea ice variables without temporal and spatial gaps. Here, we study the spatiotemporal variability of the Arctic sea ice area and thickness using the Global ocean Reanalysis Ensemble Product (GREP) produced and disseminated by the Copernicus Marine Service (CMS). GREP is compared and validated against the state-of-the-art regional reanalyses PIOMAS and TOPAZ, and observational datasets of sea ice concentration and thickness for the period 1993–2020. Our analysis presents pan-Arctic metrics but also emphasizes the different responses of ice classes, marginal ice zone (MIZ) and pack ice, to climate changes. This aspect is of primary importance since the MIZ has been widening and making up an increasing percentage of the summer sea ice as a consequence of the Arctic warming and sea ice extent retreat. Our results show that the GREP ensemble provides reliable estimates of present-day and recent past Arctic sea ice states and that the seasonal to interannual variability and linear trends in the MIZ area are properly reproduced, with ensemble spread often being as broad as the uncertainty of the observational dataset. The analysis is complemented by an assessment of the average MIZ latitude and its northward migration in recent years, a further indicator of the Arctic sea ice decline. There is substantial agreement between GREP and reference datasets in the summer. Overall, the GREP ensemble mean is an adequate tool for gaining an improved understanding of the Arctic sea ice, also in light of the expected warming and the Arctic transitions to ice-free summers.
-
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
(4838 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(4838 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-413', Anonymous Referee #1, 09 Apr 2024
The study by Cocetta et al. shows by comparison with regional reanalysis products (PIOMAS and TOPAZ) as well as some other satellite products (CS2/SMOS or SIC from OSI SAF and NSIDC) that the global Ocean Reanalysis Ensemble Product (GREP) is very well able to resolve regional and interannual variability and trends in ice cover and ice thickness. In particular, a comparison of the ice thicknesses of different reference products with GREP emphasizes the additional value of using ensemble means.
To my knowledge, the comparison of global ensemble means with regional reanalysis data is new and innovative. The good agreement is worth mentioning as reanalysis products are becoming increasingly important for a range of scientific application (e.g. training of models). The paper is well written, very clearly structured and has a number of interesting figures. This paper shows that existing sea ice reanalysis products can be trusted, at least with respect to the variables studied here.
In parts, the authors could try to better explain observed differences or refer to other studies if necessary. I find the paper very interesting and recommend it for publication if the following points are addressed:
I find the definition for MIZ difficult. As described by the authors, the MIZ is also characterised by its proximity to the open ocean. I am afraid that the threshold-based classification (15-80%), even if based on a 25 km product, is not sufficiently accurate and induces errors. I recommend that the authors either:
- A) to show that the definition used is sufficiently accurate and enables a MIZ classification even in years when the ice concentration in parts of the central Arctic falls below 80%.
- B) to expand/revise the definition of the MIZ and, for example, to take into account the distance to the ice edge
- C) To refrain from using the term MIZ and instead speak of "low ice concentration areas" or similar.
I also wonder whether the title of the manuscript should not be kept more general. The focus in the title on the MIZ is not really necessary, as a much broader comparison is made that goes well beyond the MIZ
More detailed comments:
Abstract is well-written and presents the main objectives and findings.
Line 4: GREP: The Climate Copernicus website usually refers to the Global ocean Reanalysis Ensemble Product. I would leave out "ocean".
Line 1: May be use “data-assimilating” instead
Line 8: Widening or expanding? Or both?
Line 15: Transition
Line 16: Provide more recent reference
Line 20: Suggestion: “…are expected to continue unless anthropogenic greenhouse gas emissions are mitigated”
Line 25: This definition of MIZ seems somewhat outdated to me (see general comment above). In the summer months in particular, ice concentration values well below 80 % can occur temporarily throughout the Arctic. This is the result of intensified melting but sometimes also atmospheric effects (e.g. https://doi.org/10.1525/elementa.2023.00039 ). Is there a way to take this into account in the definition? And if not, what impact is this likely to have on the validation?
Line 30: “projected future expansion”…reference missing! Rolph primarily take a look at the satellite era. Won't the MIZ area automatically become smaller at some point the less ice remains in the Arctic in the summer months? For example, if ice is limited to the last ice areas in a few years' time, wouldn't the MIZ be much smaller than it is today?
Line 38: DA = Data Assimilation?
Line 31: Dissimilar: Consider using “different”
Comment: The research questions at the end are nicely shaped
Line 74: I was searching the CMS catalogue but could not find any product via the product reference.
Line 76: … but is there a reference that describes model setup and data assimilation method? Can you add it?
Tab. 1 and text: DA sea ice data: Please provide product reference and coverage/resolution for remote sensing observations. E.g. OSI-450a / CERSAT, etc…
Tab 1: The provided reference refers to a documentation of the model?
Chapter 4.1: I find the definition of the MIZ somewhat difficult. As the author points out in the intro, the MIZ is characterized by wave/tidal interaction and the influence of the open ocean. I believe that a simple threshold value (i.e. ice concentration of less than 80%) is not a sufficiently precise definition, but that other parameters should actually be included here. This could be, for example, the distance to the open ocean or from the ice edge. See my general comment above. I think a good summary of the MIZ definition problem is given in https://journals.ametsoc.org/view/journals/atot/34/7/jtech-d-16-0171.1.xml
General comment: If I have understood correctly, all the products used have the same resolution (1/4°, or 25 km). Consequently, no major differences in the product comparisons are to be expected that could be attributed to the resolution? Perhaps it would be good to emphasise this again.
Fig. 2b): It is somewhat difficult to see the differences. May be you show any anomalies here? Alternatively you can shorten the x-axis and make the y-axis larger.
Line 148: I could imagine that the reanalysis products together with the applied MIZ definition provide a better estimate of the MIZ area than the satellite data.
Fig 3: Very nice performance. Good and interesting to see
Line 173: Fig. – figure, Tab. – table
Line 189: I think family does not fit.
Line 201: interesting observation! Just out of curiosity: I wonder if the frequency or magnitude of such autumn peaks has increased over time?
Line 204: Again, I am not sure if satellite observations can provide a robust MIZ estimate with the applied definition.
Line 221: … add “in Fig. 4 c)
Line 222: I agree, although these are processes that affect the entire Arctic and not just the MIZ.
Line 225 and following: I believe that an interpretation of the trends is difficult due to the MIZ definition used.
Fig. 5: I may have missed it: but I would expect a smaller spread in January and March MIZ latitudes compared to July and September averages. Can you explain why this is? Is it because the area of the MIZ is smaller in March and limited to areas characterized by high variability (Fram Strait)?
Fig. 7: Is it possible to enlarge Y-axis?
Line 270: …it slowly narrows until 2011: Any idea why this is?
Line 275-285: Interesting scatter plot! However, I think it would be important to go beyond a description of the agreements and initiate a discussion about the larger spread in the summer months, for example. This could be done with reference to other studies that look at similar relationships (Fig. 3; lag correlation between sea ice volume and thickness: https://www.nature.com/articles/s41586-022-05058-5)
Fig. 9 and discussion: That's a very interesting comparison! Nice to see. Perhaps you can adjust the colour display in Fig. 9. Red/green/orange is difficult to distinguish.
Citation: https://doi.org/10.5194/egusphere-2024-413-RC1 -
AC2: 'Reply on RC1', Francesco Cocetta, 03 Jul 2024
Dear Reviewer,
We thank you for accurately reading and commenting on the manuscript and suggesting how to improve it. Detailed answers to each of your comments are provided in the supplement material. We hope you find them satisfactory. Reviewer comments are in black, followed by our response in blue, which includes changes and/or additions to the text.
For the authors,
Francesco Cocetta
-
AC2: 'Reply on RC1', Francesco Cocetta, 03 Jul 2024
-
RC2: 'Comment on egusphere-2024-413', Anonymous Referee #2, 11 May 2024
The authors extend the evaluation of sea ice in the CMEMS GREP Ensemble Reanalysis Product from the Antarctic (Iovino et al., 2022) to the Arctic in this manuscript, focusing on panArctic scale performances and the Marginal Ice Zone (MIZ) properties. As MIZ is an increasing proportion of the Arctic sea ice regime under climate change, accurately representing their spatiotemporal variabilities is becoming a key benchmark for sea ice modeling skills.
Overall, this manuscript is well-structured and well-written. More importantly, the proposed scientific questions are sound and adequately discussed. I like the authors' indication of ocean-sea ice reanalysis application scenarios, especially for the hottest machine learning techniques at the moment, where reanalyzed data is not only complementary to observed data, but also may be a more recommended dataset for training models. This is another valuable guideline beyond the data quality assessment for reanalysis users.
I recommend the publication of this excellent manuscript after addressing the following detailed issues, which are rather minor:
- Introduction: The authors have thoroughly reviewed the Arctic sea ice changes under climate change and ORAs' role in this subject. However, I am unfortunate to find Chevallier et al. (2017) and Uotila et al. (2019), two very comprehensive assessments of sea ice in global/regional reanalyses, surprisingly uncited. I believe the Introduction will be completed by connecting your work with theirs.
- Line 129: "(SMOS)" firstly appears in line 122.
- Line 135: "NSIDC" has been defined in line 110.
- Line 144: "wide" should be "wider".
- Line 147: "(f) to (e)" should be "(f) to (h)".
- Line 158: "Figure 3(a)" should be "Figure 2(a)"?
- Line 160: The seasonal cycles in Fig. 2(a) contain information on both climatology and interannual variability, with the former dominating the curve. This makes it difficult to signal interannual variability directly from the figure. I wonder if drawing anomalies would be more intuitive.
- Figure 3: Although you have mentioned the meaning of the individual line types in the text, I suggest that it would be clearer to the reader if you also specify it in the figure legend.
- Line 163-164: It seems to me that September has the smallest spread, within the range of two observations, while March has the largest.
- Line 200: Does "1 106 km2" mean "1x106 km2"?
- Line 198: "Fig. 2(b)" should be "Fig. 4(b)".
- Line 204-205: I find it difficult to understand this sentence, please rephrase it.
- Figure 4: It might be better to rearrange the subfigures in Fig. 4 in the order in which they were written in the main text.
- Table 3: Why not directly list these four quantities in Fig. 5?
- Line 225: I do not get what "MIZ SIA" means.
- Line 255: Should the section number read "4.3"?
- Table 4: Also, I would recommend directly listing these three quantities in Fig. 7.
- Additionally, I recommend labeling the SICs throughout the text in %, including the colorbars in figures.
References
Chevallier, M., Smith, G. C., Dupont, F., Lemieux, J.-F., Forget, G., Fujii, Y., et al. (2017). Intercomparison of the Arctic sea ice cover in global ocean–sea ice reanalyses from the ORA-IP project. Climate Dynamics, 49(3), 1107–1136. https://doi.org/10.1007/s00382-016-2985-y
Uotila, P., Goosse, H., Haines, K., Chevallier, M., Barthélemy, A., Bricaud, C., et al. (2019). An assessment of ten ocean reanalyses in the polar regions. Climate Dynamics, 52(3–4), 1613–1650. https://doi.org/10.1007/s00382-018-4242-z
Citation: https://doi.org/10.5194/egusphere-2024-413-RC2 -
AC1: 'Reply on RC2', Francesco Cocetta, 03 Jul 2024
Dear Reviewer,
We thank you for accurately reading and commenting on the manuscript and suggesting how to improve it. Detailed answers to each of your comments are provided in the supplement material. We hope you find them satisfactory. Reviewer comments are in black, followed by our response in blue, which includes changes and/or additions to the text.
For the authors,
Francesco Cocetta
-
AC1: 'Reply on RC2', Francesco Cocetta, 03 Jul 2024
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-413', Anonymous Referee #1, 09 Apr 2024
The study by Cocetta et al. shows by comparison with regional reanalysis products (PIOMAS and TOPAZ) as well as some other satellite products (CS2/SMOS or SIC from OSI SAF and NSIDC) that the global Ocean Reanalysis Ensemble Product (GREP) is very well able to resolve regional and interannual variability and trends in ice cover and ice thickness. In particular, a comparison of the ice thicknesses of different reference products with GREP emphasizes the additional value of using ensemble means.
To my knowledge, the comparison of global ensemble means with regional reanalysis data is new and innovative. The good agreement is worth mentioning as reanalysis products are becoming increasingly important for a range of scientific application (e.g. training of models). The paper is well written, very clearly structured and has a number of interesting figures. This paper shows that existing sea ice reanalysis products can be trusted, at least with respect to the variables studied here.
In parts, the authors could try to better explain observed differences or refer to other studies if necessary. I find the paper very interesting and recommend it for publication if the following points are addressed:
I find the definition for MIZ difficult. As described by the authors, the MIZ is also characterised by its proximity to the open ocean. I am afraid that the threshold-based classification (15-80%), even if based on a 25 km product, is not sufficiently accurate and induces errors. I recommend that the authors either:
- A) to show that the definition used is sufficiently accurate and enables a MIZ classification even in years when the ice concentration in parts of the central Arctic falls below 80%.
- B) to expand/revise the definition of the MIZ and, for example, to take into account the distance to the ice edge
- C) To refrain from using the term MIZ and instead speak of "low ice concentration areas" or similar.
I also wonder whether the title of the manuscript should not be kept more general. The focus in the title on the MIZ is not really necessary, as a much broader comparison is made that goes well beyond the MIZ
More detailed comments:
Abstract is well-written and presents the main objectives and findings.
Line 4: GREP: The Climate Copernicus website usually refers to the Global ocean Reanalysis Ensemble Product. I would leave out "ocean".
Line 1: May be use “data-assimilating” instead
Line 8: Widening or expanding? Or both?
Line 15: Transition
Line 16: Provide more recent reference
Line 20: Suggestion: “…are expected to continue unless anthropogenic greenhouse gas emissions are mitigated”
Line 25: This definition of MIZ seems somewhat outdated to me (see general comment above). In the summer months in particular, ice concentration values well below 80 % can occur temporarily throughout the Arctic. This is the result of intensified melting but sometimes also atmospheric effects (e.g. https://doi.org/10.1525/elementa.2023.00039 ). Is there a way to take this into account in the definition? And if not, what impact is this likely to have on the validation?
Line 30: “projected future expansion”…reference missing! Rolph primarily take a look at the satellite era. Won't the MIZ area automatically become smaller at some point the less ice remains in the Arctic in the summer months? For example, if ice is limited to the last ice areas in a few years' time, wouldn't the MIZ be much smaller than it is today?
Line 38: DA = Data Assimilation?
Line 31: Dissimilar: Consider using “different”
Comment: The research questions at the end are nicely shaped
Line 74: I was searching the CMS catalogue but could not find any product via the product reference.
Line 76: … but is there a reference that describes model setup and data assimilation method? Can you add it?
Tab. 1 and text: DA sea ice data: Please provide product reference and coverage/resolution for remote sensing observations. E.g. OSI-450a / CERSAT, etc…
Tab 1: The provided reference refers to a documentation of the model?
Chapter 4.1: I find the definition of the MIZ somewhat difficult. As the author points out in the intro, the MIZ is characterized by wave/tidal interaction and the influence of the open ocean. I believe that a simple threshold value (i.e. ice concentration of less than 80%) is not a sufficiently precise definition, but that other parameters should actually be included here. This could be, for example, the distance to the open ocean or from the ice edge. See my general comment above. I think a good summary of the MIZ definition problem is given in https://journals.ametsoc.org/view/journals/atot/34/7/jtech-d-16-0171.1.xml
General comment: If I have understood correctly, all the products used have the same resolution (1/4°, or 25 km). Consequently, no major differences in the product comparisons are to be expected that could be attributed to the resolution? Perhaps it would be good to emphasise this again.
Fig. 2b): It is somewhat difficult to see the differences. May be you show any anomalies here? Alternatively you can shorten the x-axis and make the y-axis larger.
Line 148: I could imagine that the reanalysis products together with the applied MIZ definition provide a better estimate of the MIZ area than the satellite data.
Fig 3: Very nice performance. Good and interesting to see
Line 173: Fig. – figure, Tab. – table
Line 189: I think family does not fit.
Line 201: interesting observation! Just out of curiosity: I wonder if the frequency or magnitude of such autumn peaks has increased over time?
Line 204: Again, I am not sure if satellite observations can provide a robust MIZ estimate with the applied definition.
Line 221: … add “in Fig. 4 c)
Line 222: I agree, although these are processes that affect the entire Arctic and not just the MIZ.
Line 225 and following: I believe that an interpretation of the trends is difficult due to the MIZ definition used.
Fig. 5: I may have missed it: but I would expect a smaller spread in January and March MIZ latitudes compared to July and September averages. Can you explain why this is? Is it because the area of the MIZ is smaller in March and limited to areas characterized by high variability (Fram Strait)?
Fig. 7: Is it possible to enlarge Y-axis?
Line 270: …it slowly narrows until 2011: Any idea why this is?
Line 275-285: Interesting scatter plot! However, I think it would be important to go beyond a description of the agreements and initiate a discussion about the larger spread in the summer months, for example. This could be done with reference to other studies that look at similar relationships (Fig. 3; lag correlation between sea ice volume and thickness: https://www.nature.com/articles/s41586-022-05058-5)
Fig. 9 and discussion: That's a very interesting comparison! Nice to see. Perhaps you can adjust the colour display in Fig. 9. Red/green/orange is difficult to distinguish.
Citation: https://doi.org/10.5194/egusphere-2024-413-RC1 -
AC2: 'Reply on RC1', Francesco Cocetta, 03 Jul 2024
Dear Reviewer,
We thank you for accurately reading and commenting on the manuscript and suggesting how to improve it. Detailed answers to each of your comments are provided in the supplement material. We hope you find them satisfactory. Reviewer comments are in black, followed by our response in blue, which includes changes and/or additions to the text.
For the authors,
Francesco Cocetta
-
AC2: 'Reply on RC1', Francesco Cocetta, 03 Jul 2024
-
RC2: 'Comment on egusphere-2024-413', Anonymous Referee #2, 11 May 2024
The authors extend the evaluation of sea ice in the CMEMS GREP Ensemble Reanalysis Product from the Antarctic (Iovino et al., 2022) to the Arctic in this manuscript, focusing on panArctic scale performances and the Marginal Ice Zone (MIZ) properties. As MIZ is an increasing proportion of the Arctic sea ice regime under climate change, accurately representing their spatiotemporal variabilities is becoming a key benchmark for sea ice modeling skills.
Overall, this manuscript is well-structured and well-written. More importantly, the proposed scientific questions are sound and adequately discussed. I like the authors' indication of ocean-sea ice reanalysis application scenarios, especially for the hottest machine learning techniques at the moment, where reanalyzed data is not only complementary to observed data, but also may be a more recommended dataset for training models. This is another valuable guideline beyond the data quality assessment for reanalysis users.
I recommend the publication of this excellent manuscript after addressing the following detailed issues, which are rather minor:
- Introduction: The authors have thoroughly reviewed the Arctic sea ice changes under climate change and ORAs' role in this subject. However, I am unfortunate to find Chevallier et al. (2017) and Uotila et al. (2019), two very comprehensive assessments of sea ice in global/regional reanalyses, surprisingly uncited. I believe the Introduction will be completed by connecting your work with theirs.
- Line 129: "(SMOS)" firstly appears in line 122.
- Line 135: "NSIDC" has been defined in line 110.
- Line 144: "wide" should be "wider".
- Line 147: "(f) to (e)" should be "(f) to (h)".
- Line 158: "Figure 3(a)" should be "Figure 2(a)"?
- Line 160: The seasonal cycles in Fig. 2(a) contain information on both climatology and interannual variability, with the former dominating the curve. This makes it difficult to signal interannual variability directly from the figure. I wonder if drawing anomalies would be more intuitive.
- Figure 3: Although you have mentioned the meaning of the individual line types in the text, I suggest that it would be clearer to the reader if you also specify it in the figure legend.
- Line 163-164: It seems to me that September has the smallest spread, within the range of two observations, while March has the largest.
- Line 200: Does "1 106 km2" mean "1x106 km2"?
- Line 198: "Fig. 2(b)" should be "Fig. 4(b)".
- Line 204-205: I find it difficult to understand this sentence, please rephrase it.
- Figure 4: It might be better to rearrange the subfigures in Fig. 4 in the order in which they were written in the main text.
- Table 3: Why not directly list these four quantities in Fig. 5?
- Line 225: I do not get what "MIZ SIA" means.
- Line 255: Should the section number read "4.3"?
- Table 4: Also, I would recommend directly listing these three quantities in Fig. 7.
- Additionally, I recommend labeling the SICs throughout the text in %, including the colorbars in figures.
References
Chevallier, M., Smith, G. C., Dupont, F., Lemieux, J.-F., Forget, G., Fujii, Y., et al. (2017). Intercomparison of the Arctic sea ice cover in global ocean–sea ice reanalyses from the ORA-IP project. Climate Dynamics, 49(3), 1107–1136. https://doi.org/10.1007/s00382-016-2985-y
Uotila, P., Goosse, H., Haines, K., Chevallier, M., Barthélemy, A., Bricaud, C., et al. (2019). An assessment of ten ocean reanalyses in the polar regions. Climate Dynamics, 52(3–4), 1613–1650. https://doi.org/10.1007/s00382-018-4242-z
Citation: https://doi.org/10.5194/egusphere-2024-413-RC2 -
AC1: 'Reply on RC2', Francesco Cocetta, 03 Jul 2024
Dear Reviewer,
We thank you for accurately reading and commenting on the manuscript and suggesting how to improve it. Detailed answers to each of your comments are provided in the supplement material. We hope you find them satisfactory. Reviewer comments are in black, followed by our response in blue, which includes changes and/or additions to the text.
For the authors,
Francesco Cocetta
-
AC1: 'Reply on RC2', Francesco Cocetta, 03 Jul 2024
Peer review completion
Journal article(s) based on this preprint
Model code and software
Scripts for "GREP reanalysis captures the evolution of the Arctic Marginal Ice Zone across timescales" F. Cocetta https://doi.org/10.5281/zenodo.10651276
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 381 | 129 | 33 | 543 | 24 | 21 |
- HTML: 381
- PDF: 129
- XML: 33
- Total: 543
- BibTeX: 24
- EndNote: 21
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Francesco Cocetta
Lorenzo Zampieri
Julia Selivanova
Doroteaciro Iovino
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(4838 KB) - Metadata XML