the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Toward a marginal Arctic sea ice cover: changes to freezing, melting and dynamics
Rebecca Caitlin Frew
Daniel Feltham
David Schroeder
Adam William Bateson
Abstract. As summer Arctic sea ice extent has retreated, the marginal ice zone (MIZ) has been widening and making up an increasing percentage of the summer sea ice. The MIZ is defined as the region of the ice cover that is influenced by waves, and for convenience here is defined as the region of the ice cover between ice concentrations (area fractions) A of 15 to 80 %. The MIZ is projected to become a larger percentage of the summer ice cover, as the Arctic transitions to ice free summers. We compare individual processes of ice volume gain and loss in the ice pack (A>80 %) to those in the MIZ to establish and contrast their relative importance and examine how these processes change as the summer MIZ fraction and amplitude of the seasonal sea ice growth/melt cycle increases over decadal timescales. We use an atmosphere-forced, physics-rich sea ice-mixed layer model that includes a prognostic floe size distribution (FSD) model including brittle fracture and form drag. The model has been compared to FSD observations, satellite observation of sea ice extent and PIOMAS.
The MIZ fraction of the July sea ice cover, when the MIZ is at its maximum extent, increases by a factor of 2 to 3, from 14 % (20 %) in the 1980s to 46 % (50 %) in the 2010s in NCEP (HadGEM2-ES) atmosphere-forced simulations. In a HadGEM2-ES forced projection the July sea ice cover is almost entirely MIZ (93 %) in the 2040s. Basal melting accounted for the largest proportion of melt in regions of pack ice and MIZ for all time periods. During the historical period, top melt was the next largest melt term in pack ice, but in the MIZ top melt and lateral melt were comparable. This is due to a relative increase of lateral melting and a relative reduction of top melting by a factor of 2 in the MIZ compared to the pack ice. The volume fluxes due to dynamic processes decreases due to the reduction in ice volume in both the MIZ and pack ice. As the ice cover becomes marginal (MIZ), it melts earlier: in the region that was pack ice in the 1980s and became marginal in the 2010s, peak melting starts 20/12 days earlier (NCEP/HadGEM2-ES). This continues in the projection where melting in the region that becomes MIZ in the 2040s shifts 14 days earlier.
- Preprint
(1958 KB) - Metadata XML
- BibTeX
- EndNote
Rebecca Caitlin Frew et al.
Status: final response (author comments only)
-
RC1: 'Comment on tc-2023-91', Anonymous Referee #1, 21 Jul 2023
General comments –
The paper provides an assessment of the mass budgets in the MIZ/non-MIZ regions in Arctic sea ice simulations. It uses a comprehensive sea ice model which has been well utilized. It acknowledges that the lack of active atmosphere and dynamic ocean components affects feedbacks in the system, although some deeper discussion of how this might affect the results could be helpful. Overall, I believe that the experiment design is reasonable and can provide some valuable insights on changing Arctic sea ice mass budgets and their projection into the future. However, as noted below, I believe that there is a need for a somewhat deeper analysis and more interpretation of the results.
For example, the study explores differences between simulations with different atmospheric forcing. However, it provides little information on why the NCEP forced vs HadGEM forced runs simulate different behavior. Additionally, the comparison of MIZ versus non-MIZ mass budgets is given but again there is limited information on why the mass budgets differ across these different regions. I do believe that there can be value in delimiting the analysis into MIZ/non-MIZ/changing to MIZ regions. However, I felt that there was a missed opportunity within the manuscript to better articulate why this approach was useful and what new insight it provided relative to previous studies on sea ice mass budgets. Additionally, it would have been useful to discuss the implications from these new insights on broader questions such as the future evolution of the sea ice and/or discrepancies across models, for example. Overall, I believe that some deeper analysis (including possible new analysis) and deeper discussion of the study implications are needed. With changes in these areas, I believe that the originality and significance of this work would be clearer and this would result in a more impactful study.
Specific comments.
- Please provide more information throughout the manuscript on the value of separating analysis into MIZ/non-MIZ regions. How do you expect processes to differ in these regions? Why do things differ across these regions both in their mean state and in their response over time? What value is added by looking at the system from this perspective?
- Please provide more information on what new information/insights were learned from this study relative to previous work.
- Please provide more information on the implications of these new findings for bigger questions on the future evolution of the ice cover and/or discrepancies across models (for example).
- Line 25, “if” should be “is”
- Lines 52-53: I appreciate that you acknowledge the limitations of the forced model framework. I think that it would be useful to return to this in the results or conclusions section to discuss how these limitations may specifically affect the results from this study.
- Line 90-92: “The ocean temperature and salinity below the mixed layer …” Does this lead to a weaker sea ice response, especially near the ice edge? How do the results from this study compare to the sea ice simulated in the HadGEM2-ES runs?
- Line 99-100: “and a prognostic floe size distribution model” What is the wave forcing used to drive this model? Are there any feedbacks from the model onto the wave forcing? I assume not since this coupling is not mentioned. The discussion of the limitations of this, particularly for studying things in an MIZ/non-MIZ perspective should be discussed.
- Line 102-103: “The HadGEM2-ES product …” Could you say a bit more about this product? Is it from a single ensemble member? How does this forcing (for example, Arctic mean surface air temperature, precipitation, etc.) compare to observations?
- Line 116-118: “The maximum sea ice extent …”: The winter ice edge (and maximum ice extent) are particularly sensitive to ocean conditions. How does the use of fixed T/S below the mixed layer influence ice loss to 2040? How do the results in this forced model framework differ from the HadGEM2-ES simulations and what does that tell you about the role of coupling? Similar questions are also relevant for the summer ice loss in the HadGEM2-ES forced runs and I’d suggest that you compare those also to the coupled HadGEM2-ES simulation that was used to obtain the atmospheric forcing data.
- Line 118-119. “The NCEP forced simulation shows a stronger declining summer sea ice extent trend …”. Why is this the case? Is it consistent with internal variability? Is it due to biases in HadGEM2-ES forcing?
- Table 1. For the 1980s and 2010s, it would be useful to also include observations on this table. Adding the range of values for the individual years in the 5-year periods that are analyzed would also be helpful for putting the differences between runs into context. It would also be helpful to show this information visually with a bar chart or something similar.
- Figure 1. The quality of this figure should be improved. It is hard to distinguish the MIZ lines on the plots (particularly for Aug and Sept). What do the vertical bars on the observed extent signify?
- Line 121. “NASA Team and NASA Bootstrap” – please provide a reference for these datasets.
- Lines 133-134. “By the 2010s the MIZ becomes the dominant part of …” Please clarify for what months this is true.
- Line 134-135. “the summer sea ice cover is almost entirely MIZ …” Please be more specific on what months this is for.
- Line 138. “PIOMAS” – please provide a reference and brief information on the PIOMAS product. In particular, it should be noted that PIOMAS is a model product and not direct observations.
- Line 165. “By the 2010s the MIZ” – please provide a comparison to the observations here.
- Line 175. “basal growth makes up the largest proportion of melting” - Do you mean “basal melt” here?
- Line 177: “significantly more” – what is the significance level used to determine this?
- Line 186. “This means that it is likely that MIZ closer to the pack ice has a different balance of processes to the outer MIZ that has lower ice concentrations.” Given this, what is the value in separating things into a MIZ/non-MIZ framework? Is it useful? What are the limitations? Would it be beneficial to look at the mass budgets as a function of ice concentration instead?
- Line 192-194. “A possible explanation for the increase in top melting in the future MIZ…” Is this inconsistent with the results from Keen et al. that did include active ocean models? Are there other factors that could explain increased top melt? Do factors like the change in seasonality, regionality play a role? Could you provide more analysis (for example of heat budgets) to gain a better understanding of the factors at play?
- Line 194-197. “This means that the results here could be seen as a lower estimate on the ice loss…” I would encourage you to also assess the HadGEM2-ES runs to determine if those exhibit different ice loss characteristics. I appreciate that the ice models are different in these runs which will affect the results but it would nonetheless provide a useful point of comparison on the sea ice response with coupled ocean feedbacks.
- Line 198. “convection and ridging”. Isn’t it just advection of the ice that can cause a mass flux? Won’t ridging affect the distribution of ice but conserve ice volume/mass?
- Line 229-230. “Peak total melting rates …” Why are there these differences between the NCEP and HadGEM2-ES runs?
- Line 234-236. “In the NCEP simulation there is a 17% increase, whilst in the HadGEM2-ES simulation there is a 6% decrease, likely due to the larger melt rates in the NCEP simulation.” Why “due to the larger melt rates”? How do the differences in melt rates drive the different ice growth response?
- Line 268-270. “Top melt was twice as important …” Please provide information on why these differences are present.
- Conclusions section. Please provide information on what new insights were gained in this study relative to previous work and the broader implications of this study.
Citation: https://doi.org/10.5194/tc-2023-91-RC1 -
RC2: 'Comment on tc-2023-91', Anonymous Referee #2, 27 Jul 2023
The paper makes use of a sophisticated sea-ice model (CICE-Icepack) to investigate the processes affecting ice mass gain/loss in the marginal ice zone (MIZ) compared to the pack ice. Two simulations with different atmospheric forcing are performed: one using the NCEP reanalysis, available from 1979 to 2020 and the other using the HadGEM2-ES model, available from 1980 to 2050. Such a framework allows for validations of the model with observations in the 1980s, where the ice cover is dominated by the presence of thick multiyear ice (low-MIZ) and in the 2010s when the ice cover is much thinner and younger (high-MIZ). Then, the model is used to assess the processes affecting the ice mass balance in the 2040s, where the sea-ice cover is projected to be seasonally ice-free (all-MIZ). While the research topic is important, the methods are sound and the results are interesting, the text and the structure have significant flaws. Therefore, I suggest that major revisions should be conducted by the author and co-authors to improve the flow and the clarity of the text.
General Comments
1- There are several structural issues within the text that require attention. First, the introduction is mostly composed of method statements with some discussion statements. It should focus more on presenting the existing literature, which would help emphasize the paper's significance and delineating its original contributions. Second, the method section appears to duplicate information already mentioned in the introduction. A data section, presenting the two observational datasets (i.e., Nasa Bootstrap and the Nasa Team) is required as they are an essential part of the results. Third, more effort is needed to explain the results clearly. Specifically, the figures are almost never referenced when commenting about a result which makes it hard to follow. An excess number of percentages are used to compare different decades, atmospheric forcings, and observation datasets, making it arduous to follow the narrative. Finally, many clarifications should be moved to the figure captions to enhance clarity. Last, the conclusion also appears disorganized primarily due to an excessive emphasis on stating percentage changes, which detracts from highlighting the main study results.
2- Is there any wave forcing in the model? If so, a detailed description of wave forcing used in the simulations should be mentioned in the method section as waves plays a critical role in the processes affecting the FSTD - via wave-induced fracture leading to enhanced lateral melt and from the wave-dependent new-ice formation proposed by Roach et al. 2019. In fact, the FSTD model developed by Roach et al. 2018, 2019 should not be used in the absence of a wave field as mentioned in the model documentation. If there is no wave forcing, the authors should seriously reconsider the analysis or/and justify how the absence of a wave field affects their results.
Specific comments
1) Lines 2-3, 4-5 and 7-8: Theses sentences are repetitive : " [...] the MIZ has been widening and making up an increasing percentage of the summer sea-ice", " The MIZ is projected to become a larger percentage of the summer ice cover, as the Arctic transitions to ice free summers" and " [...] as the summer MIZ fraction and amplitude of the seasonal sea ice growth/melt cycle increases over decadal timescales ". This is not appropriate in an abstract.
2) Line 4: Use SIC instead of A for sea-ice concentration. If you are to define an acronym, use it consistently through the manuscript.
3) Line 9: The name of the model used (CICE) should be mentioned in the abstract.
4) Line 12: The notation that you start using to compare NCEP and HadGEM2-ES atmosphere-forced simulation is NCEP (HadGEM2-ES) (e.g., "from 14 % (20 %)"). Please use this notation consistently through the abstract.
5) Line 18: Why redefine MIZ?
6) Line 18: Saying that "the ice cover becomes more marginal" is incorrect. Use thinner, younger, mobile or "MIZ-like".
7) Line 22: Remove "traditionally". This is its definition. You could cite an older paper.
8) Line 24: Cite from the oldest to the newest. This comment is valid for the whole manuscript.
9) Line 26: What kind of fragmentation are you referring to here? Wave-induced? "[...] higher concentration of smaller floe" I would use "number" or "fraction", to avoid confusion with sea-ice concentration.
10) Lines 27-29: You say: "As the sea ice cover shrinks [...] the MIZ contracts" followed by "As summer Arctic Sea ice has retreated [...] the MIZ fraction [...] has been increasing". These two sentences appear to say the exact opposite. This needs to be clarified.
11) Lines 32-44: This is the closest from an introduction that we get. You need to expand this section. Explain clearly what has been done in previous studies. What method did they use? What were their results? Then finish the introduction by clearly stating which gap in the literature you are trying to fill.
12) Lines 47-53: This is all method section, there is no need to describe in depth all the components of the model here.
13) Lines 53-60: What is this? Why use the future tense? Is this a general statement or are you discussing the results of your model? If so, this goes into the result/discussion section.
14) Lines 61-74: Again, this is part of the method.
15) Line 85: This is a joint floe size and thickness distribution model (FSTD). Not an FSD. This is valid for the whole manuscript.
16) Line 85-90: This should go to the introduction. We should have more background information about the previous study that used this model, so we have a better understanding of why this model is adequate at this point.
17) Line 91: "MYO-WP4-PUM-GLOBALREANALYSIS-PHYS-001-004" What is that name? There must be a simpler, more understandable way to say that?
18) Line 94-100: This is a repetition of the previous paragraph in many ways. Is it all necessary? You could just mention the most important components for the study, and perhaps put a namelist file in supplementary material to avoid making such a long enumeration of all the model components that breaks the flow of the paper.
19) Lines 101-105: This needs to be more detailed as it is the heart of your method. Why use those two specific atmospheric forcing? What question are you specifically trying to answer with that?
20) Lines 106-109: These sentences are redundant, especially if you are ending the introduction with an enumeration of content of the paper.
21) Line 112: "[...] initialised with a 6-year spin-up". This is part of the method. Also is 6 years of spin-up enough? Please add a citation that supports this.
22) Line 112-113: The content of the caption should not be repeated in the text. This comment applies to all figures.
23) Lines 114-116: Here you are talking about the method used to analyze the data (i.e., using the last 5 years of each decade). I would move that to the method.
24) Table 1: I do not really like the use of a table here. I would prefer a plot of the average monthly SIE and MIZ extent, with the 5 years minimum and maximum as a shading. If you keep the table, add a top row with only the title 1980s, 2010s and 2040s.
25) Line 121: What are those data (NASA Team, NASA bootstrap) ? They are key for the validation of your model. They need a separate Data section. Same for PIOMAS.
26) Line 121-135: Refer to the figure that you are talking about more often. Not only once, at the beginning of the paragraph, but each time you are pointing at something new in the figure. This will make it easier to follow. This is valid for the whole result section.
27) Lines 117-118: The narrative of your result is extremely hard to follow. This is because you are meticulously comparing all the values between 2 different simulations with observation using different atmospheric forcing (NCEP and HADGEM2-ES), in 3 different periods (the 1980s, 2010s, and 2040s) and in 3 different regions ("always MIZ", "become MIZ" and "always pack ice"). For a general reader, it is extremely confusing to read. I would try to avoid mentioning the exact values when you are making comparisons. For example, in this case instead of saying "The maximum sea ice extent [...] shows a modest decline from 1.22e7 km2 in the 1980s to 1.15e7 km2" in the 2040s" I would just say something like "The difference between the maximum sea-ice extent in the 2040s compared to the 1980s is modest in the HadGEM simulation (Table 1)." Then the reader that is really interested in the exact values can investigate the table or the figure. This is a general comment that should be applied to all the result sections. The author should unearth the “story” in their simulations and present that instead.
28) Line 137: Why only North of 66.5 N? There are large MIZ regions south of that (e.g., Labrador Coast, Bering Sea, etc.)
29) Figure 1: Put units in millions of km2 instead of having a 1e13 factor. This is valid for all plots. Why are there vertical bars in the NASA team and Bootstrap SIE data? Reduce the number of x ticks. Increase the size of x and y labels for all figures.
30) Figure 2: add a) and b) for each panel and refer to each of the specific panels in the text. This is noisy, plot a running average?
31) Table 2: Again, I do not really like the use of a table here. I would prefer an histogram. This would facilitate visual comparison for the reader.
32) Figure 4: I would add larger title on top of each column (e.g., on the top of the first column put "NCEP" and the left of the first row put "Always pack". No need to keep a long subtitle for each panel. Remove the y ticks label in each of the subplot and keep only the one in the first column. Merge the third column with the second column (i.e., add 2040s with comparison with both the 1980s and 2010s). These comments are valid for figures 5 and 6.
33) Lines 251-281: Same comments as in the results there is no need to say all the percentages. This will help to highlight what are the main results of this study. We need to clearly see how they are original compared to what was previously done. What could be done in future work?
As I mentioned earlier, I believe that this is truly interesting and useful work. However, it is not yet a finished product that is ready for submission to a scientific journal. Here, the specific comments will help to improve the manuscript, but are certainly not the only changes that are required for the future submission. I believe that more efforts, especially from the co-authors, are required for the organization of the material and for a clearer writing of the results.
Citation: https://doi.org/10.5194/tc-2023-91-RC2
Rebecca Caitlin Frew et al.
Rebecca Caitlin Frew et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 238 | 74 | 15 | 327 | 7 | 8 |
- HTML: 238
- PDF: 74
- XML: 15
- Total: 327
- BibTeX: 7
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1