Isostatic rebound Amundsen Bay: a negative feedback that acts on West Antarctic grounding line retreat

In our series ‘Understanding Sea Level Rise’ we’ve paid ample attention to positive melting feedbacks, mechanisms that accelerate ice melt and ice sheet dynamics as global temperatures keep rising. Now of course there are also negative feedbacks, like local relative sea level lowering around ice sheets (due to decreasing gravitational pull and isostatic rebound as the ice mass shrinks), a factor that can influence the position of the grounding line of West Antarctic glaciers, the line that separates where these glaciers move across bedrock and where they start floating, forming ice shelves. In general these grounding lines are retreating as a consequence of a warming water wedge and thinning of the ice shelves and the actual ice sheet, however the local relative sea level lowering as a result of this ice loss has the opposite effect, promoting an advance of these grounding lines – therefore acting as a stabilising factor for West Antarctic glaciers, like the Twaites Glacier and Pine Island Glacier, glaciers that are potentially important contributors to acceleration of 21st century sea level rise.

Now we presume you want to know which feedbacks are (likely to be) dominant in this area, the positive or the negative feedbacks. Well, apart from reading the title of this article you can also click it, to learn more…

Grounding lines under West Antarctic Ice Sheet
Map of West Antarctic ice sheet showing Holocene fluctuations of grounding lines (the point where ice shelves attach to the bedrock) illustrating its instability – at least over the Ross Sea and Weddell Sea, that host the two largest ice shelf complexes. To the left is Amundsen Sea.

Macroperspective: (on larger timescales) acceleration of ice melt makes sense

Before we get to the research [and a short interview with two Antarctic feedback researchers who are in an ideal position to answer our question – see below], let’s first try to use a macro common sense approach:

We know from paleoclimatology that Earth’s major ice sheets are very sensitive to changes in average global temperatures. The Eemian interglacial for instance (some 120,000 years ago) saw temperatures just a tad (about 0.8 degrees) above Holocene values, with sea levels 3 to 8 meters higher, resulting from decreased ice volumes on both Greenland and West Antarctica. The Pliocene, some 3 million years ago, is another powerful analogue, better resembing the future climate. Earth temperatures were about 2 or 3 degrees higher, and sea levels in the range of 12 to 32 meters higher – possibly close to +25m. That’s a planetary state where only on East Antarctica a substantial ice mass would remain, while Greenland and West Antarctica were essentially ice free.

Even within the Holocene the smaller West Antarctic ice sheet showed remarkable sensitivity to climatic fluctuations, as described in a study published June 13 in Nature – a publication that provides evidence of rapid retreat of the grounding lines of glaciers covering the Ross Sea and the Weddell Sea (sea image on top of this article) about ten thousand years ago. Now stay tuned for more about those grounding lines – what suffices for now is to say that West Antarctica, like Greenland, is covered by a relatively unstable ice sheet, with a potential for large-scale melting triggered by anthropogenic climate change – or rather our 20th and 21st century CO2 emissions, because that’s essentially what does the trick.

As shown in the graph below from another publication in Nature, dating from February 2016, this melting will take several millennia, and is still largely dependent on 21st century emission scenarios for the extent of global average sea level rise it will cause, with a possible outcome between 55 meters under business as usual emissions and about half that deluge (29 meters) under a scenario of ambitious global climate policy [other researchers are somewhat more optimistic about the potential of ambitious mitigation].

future sea level rise graph - Nature
On a macro time scale sea level rise will first accelerate, possibly quickly, after an initial temperature rise. Over centuries and millennia ahead negative melting feedbacks will slowly become dominant, most notably the decreasing availability of unstable (‘meltable’) ice. Yes, a no-brainer. Note that in a high 21st century emissions scenario essentially all the world’s mountain glaciers (+0.25m sea level rise) and most likely all of Greenland (+7m) and West Antarctica (East & West Antarctic contribution combined: 45 meters) will become ice-free. The speed of the associated sea level rise depends both on temperature (therefore also on future emissions) and on a myriad of complex feedback loops that are actively being researched by many climate scientists across the world – people like Valentina Barletta of the Space Institute at the Technical University of Denmark and her co-author Benjamin Smith of the Polar Science Center of the University of Washington, who recently published new insights about an important negative feedback (interview below).

Now if temperature is elevated due to an increased greenhouse gas concentration, which happens ‘relatively quickly’, then over the millennia to follow these major ice sheets will start melting until they reach a new equilibrium with the Earth temperature. The logical pattern here is first a phase of (rapidly) accelerating ice melt, followed by (multiple centuries of) ‘sort of steady’ ice melt, followed by deceleration towards the new equilibrium, as illustrated by the above graph.

On all accounts our current, 21st century, is one of accelerating sea level rise – kick-starting the process as various forms of atmosphere-ocean-ice thermal climate inertia are conquered.

That should essentially answer your question: at some point, centuries from now, negative melting feedbacks will become dominant. But not in our times. That does not mean negative feedbacks are not already active. (And fortunately so – our climate system would explode without them.)

‘Observed rapid bedrock uplift in Amundsen Sea Embayment promotes ice-sheet stability’

After that somewhat extensive disclaimer we think it’s time to discuss new research into a very interesting negative melting feedback – and an actual much-welcomed piece of good news, that was published in Science on June 22 by an international team of 17 specialised climate researchers, including a number of geodesists, because possible land movement as a response to climate change is what these people were after.

The mechanism is isostatic rebound, the uplift of land surface (or rather, bedrock) as it partly recovers from the immense weight of an ice sheet pressing the rock formations of the Earth crust down into Earth’s mantle.

Together with decreasing gravitational pull of a shrinking ice mass (that leads to a state of elevated sea levels around ice sheets) this rebound leads to local sea level decline – a paradoxical phenomenon that occurs around all ice sheets with significant ice loss (therefore sea level rise is higher the further you are away from a melting ice sheet – Antarctic melting will be most felt on the northern hemisphere, Greenland melting on the southern).

The mechanism is well documented in science, as similar isostatic movements can also still be observed around areas of former Pleistocene glaciations (where land surface is still rising today, rebounding from a long-gone ice sheet, for instance Scandinavia) and their collapsing ‘glacial forebulges’ (that were actually elevated during the glaciation, and where compensating subsidence is still taking place today, for instance the adjacent North Sea basin).

The news therefore is not that this uplift will also take place under the decreasing West Antarctic Ice Sheet, but that it can already be monitored – and that it is happening much faster than was anticipated, at least in the area of Amundsen Bay (where Twaites and Pine Island Glacier mouth) – a piece of West Antarctic coastline that is usually situated ‘left-below’ on most maps of Antarctica.

The researchers presume the uplift is going faster than anticipated due to local geological circumstances under Amundsen Bay (relatively low viscosity of the mantle, that presses the Earth crust back as the weight of the ice decreases) – and also because probably the speed of ice melt taking place has been underestimated (as of course satellite measurements of decreasing ice height – in itself btw a positive melting feedback that receives little attention – needs to be calibrated for correct isostacy values.

Isostatic rebound West Antarctica - a negative melting feedback
Nice illustration (from supplementary materials of Science publication) of how relative sea level decline (caused by isostatic rebound and decreasing gravitational pull) can help stabilise glaciers, by grounding buttressing ice shelves. For sake of illustrating this negative melting feedback the image is a bit overpronounced – in reality it is not expected grounding lines (on average) will advance, but rather keep retreating, as melting continues and ice shelves shrink. Some model studies also project under continued melting the floating ice shelves will (locally) disappear completely, leaving a situation with increasingly unstable ice cliffs. Note that despite the isostatic uplift the height of the ice sheet actually keeps decreasing, making it more susceptible for (summer) surface melt. All in all West-Antarctic ice melting is accelerating on a decades-timescale (threefold increase over last 10 years, yet another recent Nature publication states) – although multi-annual fluctuation is still large and GRACE data show a more or less linear decline of ice mass for West Antarctica over the last 10 years (see graph below), where in recent years for instance in the Amundsen area the grounding line retreat of the Pine Island Glacier has stopped.

Now why this is good news, is because a stronger rebound leads to stronger relative local sea level decline – and that can possibly stabilise glaciers, by putting a break on the retreat of their grounding lines, the bedrock anchor points where the West Antarctic glaciers start floating.

Let’s try to get these insights in context – a short Q & A with two of the authors

So how should we interpret these results? No one better equipped than the lead author of the actual publication in Science, Valentina Barletta of the National Space Institute at the Technical University of Denmark – who together with co-author Benjamin Smith (‘the glacier guy’ on the paper) of the Polar Science Center of the University of Washington – was kind enough to guide us through by answering a few impossible questions:

Q – Rolf: “Oftentimes individual ice sheet studies focus on single mechanisms influencing ice sheet dynamics, and therefore lead to headlines like ‘it is going to be worse’ – or ‘it may be all right’. In recent years most of the news from Antarctica seemed to lean towards the side of ‘worse’. Now of course as a science journalist, I hope to make sense of all of the research – and therefore I wonder, what do you expect as an outcome if you were to try to combine possible positive melting feedback mechanisms (like albedo, hydrofracturing, cascading ice cliffs) and negative feedbacks, like the isostatic rebound you have studied: do you think positive feedbacks will dominate, or negative feedbacks?”

A – Valentina: “What you are asking is actually the edge of the research. We have a pretty good overall picture but we have lots of uncertainty (we don’t understand the details). Therefore more research has to be done to understand which feedback is more effective in ice mass loss.

However the climate forcing (which is not a feedback) is the most important factor in the ice mass loss: if the global warming suddenly increase (more than expected) then I’m sure the ice mass loss will rapidly increase too.”

Q: “Related: do you think overall West Antarctic Ice Sheet grounding lines are likely to retreat towards ice sheet (as documented in other research) or advance towards the Southern Ocean?”

A: “Most likely the grounding lines in general will not advance. They can at best stay stable where they are or retreat less rapidly. All this depends very much on the global warming effect (warming ocean, warming atmosphere, increasing precipitations and so on).

The climate can be seen as an external variable acting on the ice, and I’m not an expert on that. The stabilizing feedbacks are triggered when those external forcing are already playing their role. The stabilizing feedback can only do so much, if the external forcing is too strong, the stabilizing feedback will not help.

To better explain how this feedback works let me make an analogy with the ‘Chinese Finger Trap’: the more you pull, the more your fingers are trapped. In this analogy the feedback is the cylinder for the fingers. The climate forcing (that produces ice melting) is the fingers pulling apart. The more the ice melts, the stronger the feedback becomes, which in turns tends to counteract the ice melting.

To break free from the finger trap you can always pull applying an extreme strength so much to rip the trap apart. Having an extreme global warming is like ripping the finger trap apart.”

[Interesting one. And it may be especially strong in a context of other studies that project possible complete (local) disappearance of floating ice shelves – like the cascading ice cliffs study of DeConto & Pollard, see image below:
Ice cliffs feedback DeConto Pollard
These cliffs would form after the relatively fragile ice shelves have disintegrated and melted away (for live coverage: Wilkins, Larsen) – and then add a completely new feedback on progressive grounding line/ice margin retreat.]

Q: “Do you expect isostatic rebound to be strong for all major West Antarctic glacier mouths, or do you expect this is confined by geological differences?”

A: “No. For now only Amundsen sector is experiencing such fast GIA [glacial isostatic adjustment], because the Earth structure is soft there but also because the ice melting is exceptional there.”

Q: “Could local relatively strong uplift lead to subsidence elsewhere, possibly under adjacent parts of the ice sheet?”

A: “Yes, but tens of times smaller in amplitude.”

Q: “Are there somehow ramifications for East Antarctica?”

A: “East Antarctica is an independent ice sheet on top of a very different Earth structure. So in the short term, there are no implications for East Antarctica. However there are profound implications in the study of the last ice age and the associated GIA and that has implication on the understanding of the ice history of the whole Antarctica.”

With some additional questions we also spoke to co-author and Antarctic glacier expert Benjamin Smith, who agrees on the likely outcome of overall forces on West Antarctic grounding lines:

A – Benjamin: “I think the West Antarctic Ice Sheet grounding lines are likely to retreat, episodically, toward the ice sheet. We will likely see temporary readvances, but overall they’re headed inland.”

About placing the new findings in context of previous Antarctic glacier research:

“The June 22 paper infers something new about the structure of the lithosphere and the viscosity of the mantle in the Amundsen region, and compares a projection of that response into the future with the rates of change that have been shown to stabilize some ice-sheet models. Previous papers have shown that effects like these are important over millennia, but the mantle model suggested by our paper responds so quickly that it looks like the effects could be important sooner, maybe over a few hundred years.

The caveat is that not every ice-sheet model captures the same processes. The three papers cited in the June 22 paper that showed the strong sensitivity of mass loss to bedrock uplift all use models that are designed to model the response of the whole ice sheet to long-term climate change, and don’t necessarily resolve the fine-scale details of the ice sheet and how it melts at the bed, near the grounding line.

When we’ve run small-scale models that just capture the behavior of Pine Island and Thwaites over a few hundred years, the details of melt right at the grounding line seem to be really important, because the ice speeds up a lot when it loses contact with small hills on the bed. This leads to quick episodic retreats, which add up to a net mass loss that’s more important than long-term gradual retreat. I suspect that episodic retreat, which involves local draw-down rates of a few meters per year, would easily outpace the centimeters per year of bedrock uplift predicted by the earth model.

The other thing that a fine-scale model can do, which a full-ice-sheet model usually can’t, is focus intense melt rates directly at the grounding line. This also tends to produce local changes that are going to strongly overpower the effects of viscoelastic rebound.

I find the viscoelastic feedback somewhat comforting for the very long-term future of the ice sheet, but not comforting at all for what’s going to happen during my (theoretical) grandchildren’s lifetime. For strong first-order effects, like the way ice melts when it’s bathed in warm salty water, second-order feedbacks, like the bedrock’s response to unloading, shouldn’t be overstated. I think that first-order effects are bad enough that for me, they tend to overshadow questions about feedbacks.”

Your chat in the bar about climate science

Wow. Well, that should do for now, because we’ve run out of letters. Main thing we apparently have to underline again is that ice sheets are very complex and that properly modelling their dynamics in response to climate change remains a big challenge, as any individual feedback works in conjunction with so many others. Fortunately there are so many hard-working and dedicated scientists progressing on this important topic!

We also learn that despite these complexities it’s probably even more important to repeat that at their core there is still a very straight-forward correlation between temperature and ice melt – so, dear colleague science communicators, let’s not overemphasize feedbacks when perhaps the bigger yet simpler story is the initial climate forcing – and associated ‘first order effects’ (warm water, warm air).

Thank you very much Valentina Barletta and Benjamin Smith for sharing your thoughts and insights – and thank you for reading, and helping to keep the public climate debate science-based, by having a chat with friends about grounding lines and very complicated climate feedbacks that can be completely torn apart by excess heat. So let’s cut emissions on all accounts. Cheers.

© Rolf Schuttenhelm |

Comments are closed.