the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
The impact of landfast sea ice buttressing on ice dynamic speedup in the Larsen-B Embayment, Antarctica
Anna E. Hogg
Stephen L. Cornford
Benjamin J. Wallis
Benjamin J. Davison
Heather L. Selley
Ross A. W. Slater
Elise K. Lie
Livia Jakob
Andrew L. Ridout
Noel Gourmelen
Bryony I. D. Freer
Sally F. Wilson
Andrew Shepherd
Abstract. We observe the evacuation of 11-year old land-fast sea ice in the Larsen-B Embayment on the East Antarctic Peninsula in January 2022, which was in part triggered by warm atmospheric conditions and strong offshore winds. This evacuation of sea ice was closely followed by major changes in the calving behaviour and dynamics of the ocean-terminating glaciers in the region. Following a decade of gradual slow-down, satellite measurements show that Hektoria, Green and Crane Glaciers have sped up by approximately 20–50 % since February 2022, each increasing in speed by more than 100 m a-1. Circumstantially, this is attributable to the loss of floating ice/mélange tongues and their transition into tidewater glaciers. However, a question remains as to whether the landfast sea ice itself could have acted to provide direct buttressing to the glaciers prior to its disintegration. We use diagnostic model simulations to estimate the buttressing effect of the land-fast sea ice in the Larsen-B Embayment and its impact on the speed of Hektoria, Green, Evans and Crane Glaciers. The results show that direct sea ice buttressing had a negligible impact on the dynamics of the grounded ice streams. Additionally, our results show that the loss of sea ice buttressing likely produced noticeable changes to the flow speeds of the rheologically weak ice tongues, which could have diminished their stability over time. However, as the accompanying changes in viscous stress were small compared to local spatial variation, this loss of buttressing is likely to have been a secondary process in the disintegration of the ice tongues compared to, for example, increased ocean melting or swell.
- Preprint
(5041 KB) - Metadata XML
-
Supplement
(3090 KB) - BibTeX
- EndNote
Trystan Surawy-Stepney et al.
Status: open (until 02 Nov 2023)
-
RC1: 'Comment on tc-2023-128', Anonymous Referee #1, 23 Sep 2023
reply
This paper presents satellite-derived glacier speeds from 2014 and landfast sea ice extent from 2002 till present. A numerical model is used to compute whether landfast sea ice (modeled as a thin meteoric ice shelf) could have provided sufficient buttressing to produce the observed glacier speedup. The answer to this question is no, following a standard modeling approach based on diagnostic runs quantifying instantaneous changes in glacier speed resulting from different boundary conditions. Additional hypothesis are then offered regarding potential causes of the glacier speedup, but these are no longer supported with numerical modeling.
Satellite estimates:There are a lot of statements throughout the paper about pre/post 2011. But something that hasn't been shown is what the glaciers and glacier fronts did in terms of speed between 2002 and 2011 - how quickly did the system stabilize past the fast speed up in 2002? Was there any change in glacier speed around 2011 when the landfast sea ice settled in? And did the ice tongues only form past 2011 as mentioned in the paper? I don't think any of that was shown here, but is important for understanding the role of sea ice in the glacier dynamics in this bay.
The paper talks in a lot of detail about trends from 2014 but not clear what that reflects. Is that still recovering from 2002 speedup? is it reacting to 2011 landfast sea ice presence?
Previous papers have already shown similar satellite-derived information as shown here (and the studies are cited in this manuscript), so I wasn't clear on what the novelty here was. Making that explicit could help.
Modeling:There is no model validation, although Larsen B offers a rare opportunity in glaciology to actually validate a model in some sense. If the consensus is, that the Larsen B ice shelf breakup caused acceleration of outlet glaciers (and some models have showed that using similar techniques as here), then I would expect to see that validation here. That is, to see the model in this study first show that the removal of the ice shelf from a model tuned to pre 2002 velocities reproduces observed post 2002 speedup. If it does so, then I think the conclusions about the relative insensitivity of the land ice to landfast sea ice buttressing will be more robust.
There is an inconsistency in the modeling. The inversion uses ice flow speeds from 2021 when landfast sea ice was present, but the effect of the landfast sea ice itself is not included in the model at this point. So the tuning to velocities is done with the wrong setup/geometry. Further, it is unclear why the authors do the opposite experiment to nature. They take an (inconsistently) tuned state and then look at the effect of addition of an ice shelf, rather than the effect of the removal of an ice shelf. At the very least it would be good to explain the reasoning for that choice and an argument for why this procedure is generally reversible (especially considering potentially long lasting transients).
The modeling emphasis is on the glaciers that showed large change in observations, and the statement is that the modeled response after sea ice removal is not enough. How about the response of glaciers that showed no change in observations? Is the response to the sea ice removal in model larger than the observed change? Maybe you show something for Evans in a figure but there is no mention of this at all in the text.
This is another opportunity to have a "control", or a validation of your model and setup.3.2.2 talks about sea ice buttressing effect near the termini, but nowhere it is shown whether the (observed) loss of the ice tongues produces the observed grounded ice acceleration. I think the paper could go a bit further with the modeling experiments and show whether the ice tongue loss does reproduce the observations of speedup.
In the end a hypothesis is suggested that basal melt rate change could have contributed to the glacier speed up - can you show with a model that the inferred change in melt rates over the duration of the change can indeed produce significant change in glacier speed over the observed time period?
Methods:
There is minimal methods section so it is sometimes quite hard to assess what the authors actually did.
For example, which points they chose to show time series, how representative those points are, what do the error bars (shaded ares) actually represent, etc.The sections 3.2.1 and 3.2.2 need to match corresponding sections/paragraphs in "Modeling setup" so that it is clear which modeling experiment will answer what question/hypothesis.
General:The study really tries to be concise but that is at the expense of clarity and detail. At many places justification of choices and any sort of reasoning is completely absent.
The paper tries to merge observations with modeling but doesn't succeed in joining the two together very well. Part of it is probably the presentation. It needs to be stated a bit more clearly what the role of the model is in this study and what exact question, provided by the observations, it aims to answer. And also how the model and observations complement each other. Part of this is probably the inconsistency between the time series from observations, but then diagnostic model, rather than time evolving one is used. I understand that this is what people have been doing now for a while and it is fairly standard, but at some point it would be good to start understanding transient responses.
A good example of the lack of joining model and observations together is that the modeling results are in no way compared visually to observations (e.g. in Fig 2, could you also add observed flow speeds before and after sea ice removal, in addition to the modeled ones?)
The language is a bit clumsy, some long, confusing (ambiguous) sentences are present. Often times unclear what authors mean, some repetition is probably necessary.
Terminology is also an issue, often times the authors use many different terms to refer to the same thing, introducing ambiguity. Also, when saying ice, in this particular study it is really important to specify every time whether you mean sea ice/meteoric ice.
Other comments (not necessarily minor) are in the attached pdf.
Trystan Surawy-Stepney et al.
Trystan Surawy-Stepney et al.
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
199 | 105 | 8 | 312 | 17 | 5 | 3 |
- HTML: 199
- PDF: 105
- XML: 8
- Total: 312
- Supplement: 17
- BibTeX: 5
- EndNote: 3
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1