the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Modelling the evolution of Arctic multiyear sea ice over 2000–2018
Heather Christine Regan
Pierre Rampal
Einar Ólason
Guillaume Boutin
Anton Korosov
Abstract. Multiyear sea ice (MYI) cover in the Arctic has been monitored for decades using increasingly sophisticated remote sensing techniques, and these have documented a significant decline in MYI over time. However, such techniques are unable to differentiate between the processes affecting the evolution of the MYI. Further, estimating the thickness, and thus the volume of MYI remains challenging. In this study we employ a sea ice-ocean model to investigate the changes to MYI over the period 2000–2018. We exploit the Lagrangian framework of the sea ice model to introduce a new method of tracking MYI area and volume, which is based on identifying MYI during freeze onset each autumn. The model is found to successfully reproduce the spatial distribution and evolution of observed MYI extent. We discuss the balance of the processes (melt, ridging, export, and replenishment) linked to the general decline in MYI cover. The model suggests that rather than one process dominating the losses, there is an episodic imbalance between the different sources and sinks of MYI. We identify those key to the significant observed declines of 2007 and 2012; while melt and replenishment are important in 2012, sea ice dynamics play a significant role in 2007. Notably, the model suggests that convergence of the ice, through ridging, can result in large reductions of MYI area without a corresponding loss of MYI volume. This highlights the benefit of using models alongside satellite observations to aid interpretation of the observed MYI evolution in the Arctic.
Heather Christine Regan et al.
Status: final response (author comments only)
-
RC1: 'Comment on tc-2022-211', Anonymous Referee #1, 24 Nov 2022
The manuscript "Modeling the evolution of Arctic MYI cover over 2000 - 2018" by Heather Regan uses a sea ice-ocean model to examine how MYI area and volume change in the Arctic. The advantage of the presented approach is that, compared to other studies based on satellite data, budgeting of different sink and source terms such as ridging, melt and replenishment can be done and linked to the MYI retreat. The authors conclude that rather than one process dominating the observed losses. Furthermore, they take a closer look at the processes controlling MYI area and volume changes in anomalous years like in 2007 and 2012.
The study is definitely worth publishing. It is a great addition to existing (remote sensing) studies, excellently written and very well and clearly structured. The model validation and comparison with observations is extensive and convincing. The budgeting of the individual processes contributing to the decline of MYI in the Arctic is comprehensible. The focus on the two extreme years in 2007 and 2012 is also reasonable.
I believe that the paper will be widely cited and can make an important contribution to a better understanding of the changes in the last ice areas.
My comments are mainly "minor". I do, however, encourage a somewhat deeper investigation or at least better explanation of the post-2007 decline in ridging (see my comments below). If possible, please provide a Figure like Fig. 6 for the individual key regions (i.e. in the appendix).
Comments:
Line 24: Isn’t the statement that “MYI area anomalies being closely linked to anoamlies in Arctic ice volume” contradicting to what is said in the abstract? Line 11: “…can result in large reduction of MYI area without a corresponding loss of MYI volume”? Maybe you can emphasize this contradiction in the abstract since it points to the importance of this study.
Line 43: “ice type classification fails in summer”: I don't think this can be generalized, because of course different methods and sensors are used. Some of them are more robust in summer.
Line 46: A little bit outdated reference: There are more recent studies on reliability of motion products in summer available. Also the products get better and better. E.g. Hiroshi Sumata (Tromsoe) did quite some work on product intercomparison.
Line 50: “…challenging”. May be refer to “von Albedyll 2021, Linking sea ice deformation to cie thickness redistribution using HR satellite and airborne obs.
Line 52: I guess there are more recent publications assessing the accuracy of altimetry missions. E.g. the Nature Paper by Landy (2022). Avoid terms like “relative uncertain”
Paragraph 42 - 55: In this paragraph, the author focuses strongly on the drawbacks of the satellite-based methods. I think that this is not necessary, because the advantages of the models are obvious. Its just a comment, no need to change anything J
Line 89/70: MYI
Line 125: Just out of curiosity: How well do the results on temporal and spatial variability agree with satellite-based methods? Satellite-based methods probably rely on changes in surface properties, while temperature differences and heat fluxes come into play in models? However, in general, model and satellite data on freeze up should be comparable?
Line 143: “MYI is both thicker and stronger…”: This assumption probably holds one of the largest uncertainties: In particular, in the marignal ice zones and throughout the Transpolar Drift, FYI and SYI (i.e., MYI in this study) are likely to have similar thicknesses and are otherwise difficult to distinguish. When is ridging in the model considered to be completed in an ice age class? Or in other words: At what point has enough FYI been deformed for MYI to proceed?
Fig 2: All very interesting!
Line 191: May be you can get in touch with the producer and ask what the reason may be?
Line 201: “…conditions in certain years” and areas. See my comment to Line 143: I believe that in the marginal ice zones the assumption may not be correct
Line 209: CDR data.Using
Line 2014: “The model and satellite-based data?
Line 250: Just a general comment: I find the evaluation and comparison with the CDR and NSIDC data highly interesting and well done. Thanks, I appreciate reading
Fig. 6e) Hard to read. Can you make it bigger or refer to Fig. 8
Line 280: I get what you saying and I find it interesting. May be you can clarify statement a bit better
Fig 8: Can you add notations of Fig. 6e? (A,B,C,etc)
Line 291: The CE is the largest contributor because this is where most of the FYI turns into SYI?
Line 294/295: I am sceptical about the conclusion. May be move it to the discussion.
Line 296: I guess you refer to Fig. 8 in this statement, but are you sure there is no trend if looking at the Arctic wide sources and sink terms? To me it seems that ridging is reduced, although it has this staircase appearance.
Comment to export: How well does the modelled (Fram Strait) export compare with exports from others? Ricker, Smedsrud, etc? Does it capture the seasonal and interannual variability correctly?
Chapter 4.2.1: Great chapter. However, it took me a while to understand it all. Maybe you can refer to Figures more often in one or the other place? Same for 4.2.2
Line 332: I guess the replenishment rate is directly related to the FYI area available at the end of the summer. May be just state that more FYI was melted then usual, such that replenishment rate was reduced? I hope I did not get this wrong though. Note that according to this study (https://doi.org/10.1038/s41598-019-41456-y) there is a generally reduced survival rate of FYI, and hence a generally reduced replenishment rate (at least in the TP Drift)? Can you confirm this?
Line 354-356: I guess this sentence is not needed since it is well described in 4.2.1?
Line 378: I think the Smedsrud study is not solely based on observations, but observations were used to establish a relationship between pressure gradients across Fram Strait and export?
Chapter 5.2: This is all very interesting! However, I have a few questions related to the chapter
Line 401: Not sure if I got this correct: You mean, as the ice cover shrinks, less FYI survive the summer and hence, there is less replenishment taking place in the Central West region (and others)? Again, this would support the Krumpen story of a reduced survival rate and the timing of the drop-down in ridging and FYI survival rate (Krumpen) is about the same.
And Line 400: “mostly unaffected” I would expect that ridging in the Laptev See and other Russian shelf seas went to almost zero, since MYI production zones are shifted elsewhere (north)? May be it would be a good idea to provide a Fig like Fig 6 for each section (A,B,C,D… ) in the appendix. This would also be a valuable information for other studies
Line 401: You mean: As the MYI cover shrinks….
Line 402 – 405: Still this does not explain the stepwise decrease in ridging after 2008 (or may be I just did not get it). If this would be solely related to a shrinking MYI cover and shifting replenishment zones, it would be a gradual change, right?
Conclusion: Line 447: “no one process stands out”… I think that the stepwise reduction of ridging somewhat stands out…
Line 457: “This change in behaviour related to the general reduction…”. I don't think this fact has actually been explored deeply enough in the discussion to make that statement.
Citation: https://doi.org/10.5194/tc-2022-211-RC1 - AC1: 'Reply on RC1', Heather Regan, 09 Feb 2023
-
RC2: 'Comment on tc-2022-211', Anonymous Referee #2, 13 Dec 2022
The study uses the NeXtSIM sea ice model to trace MYI area and volume and compare to satellite-based observations. MYI budgets for different regions of the Arctic are constructed and the relative contribution of source and sink terms is examined. While they find good agreement in MYI area of the model and observations w.r.t. to magnitude and trends from 2000 to 2018, the interannual variability is less well explained. They analyze 2007 and 2012 to contrast the different mechanisms.
I find this paper to be put together very well. The design of the study is and the key findings are presented clearly. I commend the authors for making the limitations of their results clear rather than sweeping them under the rug. I think the study makes a nice contribution towards characterizing the fate of MYI. I would have liked to see a bit more investigation on why MYI area anomalies between model and observations don’t seem to match up particularly well, but maybe that’s for another paper.
I have some comments and suggestions for their consideration, but I think the study is a rare case of “publishable as is”.
Well done!
Detailed Comments:
Line 59: “more comparable”
comparable to what?
Line 103 “do online”
Does this mean “in real time”?
Line 109: “average ice growth”
Average for a specific location or over entire domain?
Line 134 “likely an upper bound”
Why is this an upper bound? If MYI is in fact still thicker than FYI its melt rate should be higher (Bitz and Roe, 2004)? If I'm mistaken, please clarify. Maybe a brief discussion of feedbacks between thickness, ridging and melt-rates might be useful?
Line 165: grids cells with an uncertainty of < 0.02
I assume that’s a probability? Previously probabilities were given in % (ok a bit nitpicky)Line 214. “the model struggles to capture the variability of the data”.
I am glad you are pointing this out and provide the numbers! You might want to add that the two observational data sets struggle similarly to replicate each other’s anomalies so that there is also some “observational” uncertainty. It seems though you have settled on OSI-SAF to be the data set of "reference" with better accuracy.
Line 230. “Insufficient replenishment”
The ERA-5 tends to be too warm in the winter, is that a possibility?
Line 240: “strengthens our confidence that the model has a good ability”
So you think the error sources are therefore mostly thermodynamic? Quick scan of Boutin 2022 paper shows some numbers, is this sufficient to eliminate drift, dynamics?
Line 249… The model captures about 80% of the observations…
This is a useful analysis and the 80% number is good to know. However, a more stringent evaluation of the model skill to correctly label ice type might be relative to climatology as it isn’t all that hard to correctly predict ice type for some regions (what's the accuracy of labeled anomalies in MYI area?) It is perhaps also noteworthy that the error rate doesn’t seem to change as the MYI fraction decreases, which is encouraging.
lIne 276: remains large
See also Moore et al. 2022
Line 285 “ relative MYI area loss”
Maybe better to say “increased contribution to the total MYI loss”. I was scratching my head a bit what "relateive MYI loss meant". The sentence could use some rephrasing.
Line 265… which represents over one third…
Those are good numbers for perspective. Are those derived from the budget “nets” (sum of nets) or a trend like in figure 3b?
Line 287 … “there is no trend in any of the source and sink terms”
As well as their sum (‘net’)? I’m also not quite understanding the argument about the “episodic imbalance”. If the net is negative, even without a temporal trend, that will yield declining MYI area and volume. Is the idea that the “mean net negative” over the period is the result of a few years? It might be useful to try to quantify this X % of “imbalance” over the 18 year period arises from Y years? Or with without years X,Y,Z, the net would be balanced?
Line 307 “ two anomalous years 2007 and 2012”
I don’t really see this in Figure 6a? Maybe 6b/d but also doesn’t exactly jump out ( I suspect within 1 sigma?). 2007 and 2012 are of course notable minima in September ice cover, maybe that’s the better motivation. I do like the “case study” analysis for those two years and haven’t seen the 2007 case nicely documented as here. I also wonder a bit how the analysis would look like if you defined years not from January 1 but as “ice years” from October 1 through September 30th . I wonder if that would not show the years a bit more “anomalous” as processes prior and subsequent to annual minima might be better separated (haven’t thought this through, just wondering)
Line 363 so the 212 August Storm
I don’ t think the storm was mentioned before? It would probably be helpful to add the information that this storm occurred and provide a reference (I see Parkison and Comiso 2013 already cited could cited again here, also, Simmonds and Rudeva, 2012, Zhang et al. 2013, Stern et al. 2020)
Line 367 well with observations, then there is
I think “then” is a typo and can be deleted
Line 409: surface signature in MYI
Does this refer to the microwave signature? Sticklers might argue that this isn’t strictly related to surface alone. Maybe “remote sensing” signature?
Line 434 the modeled ice drift is very good
This isn’t really shown in the paper, is it? There is a reference to Boutin 2022 which I only scanned and which has some numbers. It should be made clearer what is found in the present paper that may draw on conclusions from other papers. I guess the message is that ice-drift was found to be accurate by Boutin et al 2022 and isn’t the source of the model error. The finding in the present paper is that the model MYI extent mismatch in the fall is the primary source of error? As I said above, that could probably be worked out more clearly and the specific open questions regarding the source of the error could be more definitive (thermodynamic physics, forcings?). From my quick scan of Boutin et al. 2022, it appears that while drift stats look good, I wonder how well this translates to eliminating dynamics as a significant error source? The NeXtSIM model seems to be particularly well suited for the assimilation of sea ice motion, isn’t it? Using an equivalent analysis of MYI budgets with assimilation could isolate the effect of inadequately simulated sea ice motion (winds/drag) on the results (future work!)
Suggested References:
Moore, G. W. K., Steele, M., Schweiger, A. J., Zhang, J., & Laidre, K. L. (2022). Thick and old sea ice in the Beaufort Sea during summer 2020/21 was associated with enhanced transport. Communications Earth & Environment, 3(1), 198. doi:10.1038/s43247-022-00530-6
Simmonds, I., & Rudeva, I. (2014). A comparison of tracking methods for extreme cyclones in the Arctic basin. Tellus Series a-Dynamic Meteorology and Oceanography, 66. doi:10.3402/tellusa.v66.25252
Stern, D. P., Doyle, J. D., Barton, N. P., Finocchio, P. M., Komaromi, W. A., & Metzger, E. J. (2020). The Impact of an Intense Cyclone on Short-Term Sea Ice Loss in a Fully Coupled Atmosphere-Ocean-Ice Model. Geophysical Research Letters, 47(4), e2019GL085580. doi:https://doi.org/10.1029/2019GL085580
Zhang, J., Lindsay, R., Schweiger, A., & Steele, M. (2013). The impact of an intense summer cyclone on 2012 Arctic sea ice retreat. Geophysical Research Letters. doi:10.1002/grl.50190
Citation: https://doi.org/10.5194/tc-2022-211-RC2 - AC2: 'Reply on RC2', Heather Regan, 09 Feb 2023
Heather Christine Regan et al.
Heather Christine Regan et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 422 | 139 | 14 | 575 | 4 | 4 |
- HTML: 422
- PDF: 139
- XML: 14
- Total: 575
- BibTeX: 4
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1