the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Coupled ice/ocean interactions during the future retreat of West Antarctic ice streams
David T. Bett
Alexander T. Bradley
C. Rosie Williams
Paul R. Holland
Robert J. Arthern
Daniel N. Goldberg
Abstract. The Amundsen Sea sector has some of the fastest-thinning ice shelves in Antarctica, caused by high, ocean-driven basal melt rates, which can lead to increased ice stream flow, causing increased sea level rise (SLR) contributions. In this study, we present the results of a new synchronously coupled ice-sheet/ocean model of the Amundsen Sea sector. We use the WAVI ice sheet model to solve for ice velocities and the MITgcm to solve for ice thickness and three-dimensional ocean properties, allowing for full mass conservation in the coupled ice/ocean system. The coupled model is initialised in the present day and run forward under idealised warm and cold ocean conditions. We find that Thwaites Glacier dominates the future SLR from the Amundsen Sea sector, with a SLR that is approximately quadratic in time. The future evolution of Thwaites Glacier depends on the life-span of small pinning points that form during the retreat. The rate of melting around these pinning points provides the link between future ocean conditions and the SLR from this sector and will be difficult to capture without a coupled ice/ocean model. Grounding-line retreat leads to a progressively larger Thwaites ice-shelf cavity, leading to a positive trend in total melting, resulting from the increased ice basal surface area. Despite these important sensitivities, Thwaites Glacier retreats even in a scenario with zero ocean-driven melting. This demonstrates that a tipping point may have been passed and some SLR from this sector is now committed.
- Preprint
(2501 KB) - Metadata XML
- BibTeX
- EndNote
David T. Bett et al.
Status: final response (author comments only)
-
RC1: 'Comment on tc-2023-77', Anonymous Referee #1, 13 Jul 2023
Bett et al. presented the ice evolution of the Thwaites, Pine Island, and Smith Glaciers in the West Antarctic in a century scale using a new synchronously coupled ice/ocean model. Three couped simulations were conducted with warm and cold forcings and another one with no sub-shelf melting. They found the Thwaites Glacier provides a much higher sea-level contributions with a sea-level rise increasing approximately quadratically with time. The ice mass loss from Thwaites is closely dominated by the formation and duration of isolated pinning points and ocean-driven melting is the key driver behind the loss of pinning points. Overall, the manuscript is generally well written. However, the model setup section and some of the descriptions need more improvements.
Here are some general comments:
There are some details missing in the model setup section, especially about how the coupled model handles the grounding line movement. It is also not clear how the model facilitates the information exchange between the WAVI ice sheet models and the MITgcm STREAMCE every coupling period. Grounding line movement is a very important process in the coupled ice ocean models. However, the authors did not explain much about how they handle the grounding line movement in the coupled setup. You mentioned that (P5, L150) ‘the WAVI drag field is passed to MITgcm to decide where melting can occur during each coupling period’ and ‘grounding-line retreat is accomplished naturally’. Do you mean the grounding line position is updated every ocean timestep and then passed to WAVI. Will the grounding line position be fixed in WAVI during each couple period? How do you decide the grounding-line should retreat in the ocean model?
The boundaries between Thwaites, PIG, and Smith defined in Figure 1a are not reasonable to me. The following analysis on SLR contributions, melting and grounding ice area loss are all based on these boundaries. I would suggest using the basin boundaries to divide these three glaciers.
The authors suggested that the melting rates around the pinning points will be difficult to be captured without a coupled ice/ocean model. I would suggest running a separate ice-only model with parameterised basal melting and compare the difference.
Specific Comments:
P3, L73: please provide the details (data name, version number, etc.) about the datasets (ice velocity and thickness change rates) for the inversion.
P3, l79: It’s not clear to me how you configure the englacial temperature in your ice model. I understand you cite Arthern et al., 2015 to cover the ice model setup but you should at least mention how the ice flow equation is represented in your ice model.
P4, L97: eastern à western?
P4, L105: When you say the initial melt rates, do you mean the melt rates you gor from the ocean only simulations at end of the 2 years spin-up?
P4, L111: About the ‘coupling shock’, do you mean the sudden changes in the water fluxes across the ice/ocean boundary? Could you explain it in the text when you first mention it?
P4, P114: what is the value derived from observation?
P5, L 130: why do you set the subglacial layer to be 4 m thick. Is this a empirical value from previous studies?
P5, L135: that in the 3D ocean grid that would never go afloat à that would never go afloat in the 3D ocean grid.
P5, L138: how do you decide the couple domain is ‘far enough’? Did you justify it based on a previous projection for this region?
P5, L156: This information about bathymetry and initial ice geometry should be firstly mentioned in Sect. 2.1 and 2.2.
P5, L160: what are ‘velocity grid points’?
P5, L161: when you talk about ‘no ice basal drag’, I understand you mean floating area. But in this study, you have two regions with basal drag: basal drag beneath the grounded ice and basal drag beneath the floating ice. Please clarify it across the text.
P7, L179: delete the first ‘(e)’ here.
P8, L182: The colorbar for Fig3c melt rate is not good enough to show the increased basal melt near the new grounding line. Please adjust it to the visible range. Please also add the grounding lines in the caption.
P9, L194: The Thwaites Glacier also shows a sign of deceleration by comparing the snapshots of 100 years and 125 years in Fig. 3b, which corresponds to the drop down in Fig 4b and 4c after year 100. I realise you mentioned this on P12L266-269, but I think you should at least point it out here and leave the explanation later.
P9, L200: There are lots of noise on Figs. 4f and 4j but I did not see this noise in Fig. 4b. You ran the model for the whole domain and extract the rate of change rate of SL rise for each of the glaciers, right? Then why?
P10, L220: For Smith, it is closer to 0.15 mm/yr rather than 0.1.
P10, L227: It’s ‘almost’ 600 rather than ‘over’ 600. Actually, it is around 500 at end of 125 years.
P10, L236-237: You mentioned that one of the limitations in this study was the idealised constant ocean forcings applied to the boundaries. Not just the decadal variability but also the climatology related changes under different emission scenarios in the future. How could you make the statement that ‘the future SLR from this region is only weakly influenced by variations within the plausible range of ocean conditions’?
P17, L361: ‘with only more modified CDW’? I think you mean ‘only limited CDW’.
P19, L389: the SLR rate did not continue over the 125-year time period based on Fig. 4b if you are talking about the red line.
P19, L407: 2 km mesh near the grounding line is not seen as a very high-resolution model. A coarser mesh near the grounding line may have underestimated the mass loss in the marine ice sheet systems.
L20, P427: how about the subglacial freshwater discharge?
L20, P438: The structure of discussion could be better organised. You’re talking about another limitation here. It looks messy in the structure of the discussion. Why don’t you talk about all these limitations together rather than separated by some other points like the paragraph above?
P21, L478: it would be interesting to see the differences by conducting the ice-only experiments with parameterised ocean-driven melting.
Citation: https://doi.org/10.5194/tc-2023-77-RC1 -
RC2: 'Comment on tc-2023-77', Anonymous Referee #2, 18 Jul 2023
The manuscript “Coupled ice/ocean interactions during the future retreat of West Antarctic ice streams” by D. T. Bett and colleagues simulates the evolution of glaciers in the Amundsen Sea sector over a 125 year period using a synchronously coupled ice-ocean model. They find limited grounding line retreat and mass loss over Pine Island and Dotson-Crosson ice shelves but very large retreat on Thwaites glacier under conditions similar to warm ocean conditions observed over the past decade. The retreat rate varies spatially and with forcing conditions, and many pining points form during this retreat, suggesting a strong control of these pinning points and the importance of knowing the bathymetry precisely to accurately simulate the retreat of this glacier.
Overall, the manuscript is interesting and presents a detailed study of this region, however several aspects need to be improved and are sometimes misleading. In particular, the description of the models and processes used needs more clarification, as well as the limitations of this set-up. One particular aspect, since the pinning points play such an important role, is how the model handles parameters close to the grounding line: what is done if there are partially grounded cells, what is the friction and melt in this case, and what is the impact of the resolution in these areas on the retreat rate. These questions should be better investigated and discussed. One other potential problem is that there is no calving or ice front retreat and a thin layer of ice is “artificially” maintained; so melt continued under these thin parts, while there should not exist anymore. How does this impact the overall melt simulated and the simulations in general? Another missing part is the impact of the long spin-up of the ice model: how different is the configuration of the glaciers compared to observations after the 4000 year spin-up and how does this impact the possible retreat? Also, the forcing is highly idealized, which is clearly explained in the manuscript, but misleading in the title and abstract. There are a few places were previous studies are misrepresented. Also, it remains unclear what the coupled model brings to this study and what was learned that could not have been done with a standalone model. Finally, some figures need to be improved. Specific points are listed below.
l.1: the title is misleading: given that the scenarios are highly idealized, it seems inappropriate to talk about “future retreat”. Similarly, the work is done for the Amundsen Sea sector, which is much narrower than the West Antarctic ice streams. The title should be rephrased to better capture the study done.
l.19: I would have liked to see a sentence about calving or the ice front retreat in the abstract.
l.28: should be: “with Thwaites Ice Shelf ...”
l.32-34: Thwaites ice shelf is rather unconfined, it might be worth mentioning it here.
l.48: I am surprised to see here “simple ocean melt parameterisations” being mentioned: studies such as Reese et al., 2020 use parameterizations that are as complex as can be with today’s knowledge, so I am not sure what the authors suggest could be more complex than that.
l.51: “De Rydt and Gudmundsson, 2016” (and same in the rest of the manuscript)
l.55: “e.g., Goldberg …”
l.69-80: What about the rheology of the ice?
l.75: What is the impact of the long spin-up? How similar/different are the geometry, velocity, etc. compared to the initial configuration before the spin-up? How does it impact of the simulations?
l.79: What values are used for the accumulation? What is the impact on the simulation and especially the mass gain/loss since this is a first order control on sea level contribution.
l.69-80: This description is missing a description of what happens in the reason close to the grounding line. How is the grounding line included? How are partially grounded cells treated if there are some? How is the melt and the friction close to the grounding line? This information seems key given the role of the pinning points and should be better described.
l.89: What is the impact of having no sea ice? How does it impact the ocean and in particular the stratification?
l.95: What period is used for these observations?
l.109: What is the form of the drag coefficient? Is it velocity dependent?
l.119 and l.135 seem to be contradictory
l.127: Does it also need to exist anywhere there is ice since this is how the ice thickness is computed?
l.139: How do you ensure a smooth transition in the ice thickness and that it does not diverge between the two models over time?
l.143: What is the impact of using shorter or longer time steps for the coupling?
l.153: It would be better to put this information (and detail it a lot more) in the description of the ice model.
l.160: What does it mean “on velocity grid points”? Where are they?
l.162: What happens when the grounding line retreat and new floating cells are formed?
Fig.1: the green color for Smith is hard to see. Also it would be best to use different colors for the top left and the top right since these things are not related. It would be best to use a loop to describe the coupling in b).
Fig.2: it would be good to also put the mean melt rate on the figure. It is confusing to have positive numbers for the melt rate but negative numbers for the total melt. I am surprised but the shape and extent of Thwaites ice shelf, where does it come from? In the caption: should be “the ocean domain for …” and “melt ate under PIG”
Fig.3: Unit for the second column is “m/yr” which is very surprising. The third column has columns very hard to see (everything is blue), maybe change the scale or use a log scale to make it easier to see. How does the front evolve? Are there regions where the ice becomes really thin and therefore have no ice? How does that impact the total melt?
l.191: How much retreat?
Fig.4: There is a large difference between the no melt case and the melt cases for Thwaites glacier compared to the other glaciers, this should be better discussed in the text (maybe around l.240). Caption: “(blue line), and warm (red line)”. Maybe “Cumulative grounded area”
l.225-230: Some discussion about the role of the fixed ice front and the thin shelves that therefore keep melting would be important. I would also comment more on the difference between the melt and no melt scenarios for the different ice shelves (e.g., PIG very large while Thwaites is limited).
Fig.5: What is the bathymetry impacted by the scenario (panel d: bathymetry under warm retreat). Caption: “presence of isolated pinning points”
l.275: Maybe add a sentence about the difference between flat and elevated regions, or between regions sloping inland vs downstream.
l.279: “shallow” -> “shallower”
Fig.6: It’s not very easy to see the differences in panels e and f since everything is blue. Caption: “The area shown in Figures b-f is shown with a black box in Figure 3.”
l.311: “pinning points – the time …”
l.313: How much lower?
l.314: Why does it increase the length? If everything retreats faster, it is not clear why it would cause this kind of changes.
Fig.7: Add missing titles on panels e, f, k, l, q, and r. The dark blue for panels n and o does not seem needed. It’s a bit confusing to have the simulation years and the actual years. Maybe it would be better to start the caption with something like: “Evolution of conditions as grounding line retreats over pinning points.” Before going into the details.
l.328: “show the ice geometry …”
l.340: “The resulting loss …”
Fig.8 caption: should be left and right instead of to/bottom. Add figures are every 25 years in the caption.
l.370: slope does not really seem to become shallower on Figure 8.
l.373-376: add numbers to this description
l.404: How does that differ from the effects of pinning points in the standalone model?
l.406: What is needed to correctly model these small features?
l.412 (“very weak”) and l.415 (“important buttressing”) seem contradictory.
l.434: “spatial variability”: How does it compare with observations and parameterizations?
l.439: How about ocean conditions?
l.446-447: rephrase
The discussion is missing about what really is different with the coupled model and with a synchronous coupling.
Also, it could be good to compare this study to the results of Urruty et al. (2022) or Reese et al. (2022) as they seem quite different and suggest relatively stable grounding lines around Antarctica.
l.449: the forcing is really idealized so it is a bit pushing to call these projections, maybe simply evolution.
l.461: “For Pine Island …”
l.461: provide numbers
l.472: “temporal variability”: it seems relatively constant on Fig.9c. Is it about the total or the area averaged?
l.486: What is the form of the equation with drag coefficient? What is the unit?
Fig.A1: It is confusing to see the difference compared to b) and then the difference compared to observations.
l.505: How do you deal with the very thin water column thickness that is created when the grounding line retreat? Why is this not a problem when it seemed a large problem in the initialization.
l.527: “strong correlation” but the slopes and relationships are different for the different oceanic conditions, so what does this suggest?
Citation: https://doi.org/10.5194/tc-2023-77-RC2
David T. Bett et al.
David T. Bett et al.
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
504 | 178 | 14 | 696 | 8 | 9 |
- HTML: 504
- PDF: 178
- XML: 14
- Total: 696
- BibTeX: 8
- EndNote: 9
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1