Ice-shelf stability questioned

Surface lakes and streams are forming on Antarctica's ice shelves, making them susceptible to instability and possible collapse. But rivers could mitigate this effect by efficiently exporting meltwater to the ocean. See Letters p.344 & p.349

The hundreds of floating ice shelves surrounding Antarctica have a crucial buttressing effect on inland glaciers1 — if an ice shelf breaks up, the glaciers feeding it will flow more rapidly to the sea2, contributing to global sea-level rise3. Ice shelves are rapidly melting not only from beneath, owing to contact with warming ocean water4, but also from above, because of increasing air temperatures5. Until now, meltwater lakes and streams on Antarctic ice shelves were considered a rarity. But on page 349, Kingslake et al.6 provide evidence that widespread lake and stream networks have been present on top of many Antarctic ice shelves for at least 70 years. Because lakes are thought to be hazardous to ice-shelf stability, these observations suggest that the risk of ice-shelf collapse will be amplified if the extent and intensity of surface melting increases. However, on page 344, Bell et al.7 suggest that the increased risk will not be so great if large river networks export a sizeable fraction of the meltwater to the ocean, as observed on Antarctica's Nansen Ice Shelf.

Surface lakes result from the ponding of meltwater in topographic depressions, and can be dangerous to ice shelves because they act as loads that can flex and weaken the ice, causing it to crack8. If a lake suddenly drains through a crevasse to the ocean below, the load deficit from the ice shelf's surface can induce more crevasses, potentially triggering a chain reaction of further lake-drainage events9. This process might have been responsible for the large-scale break-up of Antarctica's Larsen B Ice Shelf in 2002, when more than 2,000 lakes drained in just a few days9. An increase in the coverage of ice-shelf lakes also enhances surface melting, because water absorbs more of the Sun's radiation than the surrounding, more reflective ice or snow.

Kingslake et al. use historical satellite imagery and aerial photography to show that about 700 large-scale lake and stream systems have persisted for decades on Antarctica's ice shelves (and on some of the glaciers that feed them), often transporting water by as much as 120 kilometres. For example, the authors find that meltwater lakes and streams have existed on the Roi Baudouin Ice Shelf since 1947. A handful of previous studies5,10 have documented surface lakes and streams on individual ice shelves over a span of a few years. But the authors' work is the first to extensively map meltwater features and drainage systems on all of Antarctica's ice shelves, over multiple decades.

Streams and rivers, like lakes, act as surface loads on ice shelves, but they also play a crucial part in the movement and distribution of meltwater. As discussed by Kingslake and colleagues, a stream can form when a lake overflows, allowing meltwater to be transported to lower elevations, and perhaps into another lake. Alternatively, Bell et al. show that a large river (or river network) can enable a substantial proportion of an ice shelf's total meltwater volume to be exported to the ocean — often by a large waterfall at the ice edge, as observed on the Nansen Ice Shelf by the authors. This process of meltwater export therefore mitigates the risk that meltwater-induced ponding will lead to ice-shelf break-up.

Surface meltwater on the Nansen Ice Shelf was first detected in 1909 by Ernest Shackleton and his Nimrod team, who repeatedly had to cross, and navigate around, lakes and streams on their way to the magnetic South Pole11. Now, Bell and colleagues use satellite imagery to show that six of the eight summer melt seasons between 2006 and 2015 were warm enough to allow the formation of a large-scale river network and ice-edge waterfall, facilitating surface-meltwater export to the ocean. Once the waterfall formed, it persisted for 5–25 days, and for longest when air temperatures were highest.

Because of rising air temperatures, melt rates on almost all Antarctic ice shelves are expected to increase two- to threefold, on average, by 2050 (ref. 12). However, the extent to which meltwater production will enhance ice-shelf instability is debatable. It was previously suggested9 that the development of hundreds of surface lakes could provide the tipping point, which, once reached, would trigger the break-up of an ice shelf. If this is true, many ice shelves could be exposed to an ever-increasing risk of break-up, given their already extensive lake coverage. However, such a threshold might not be reached if surface water can instead be efficiently exported from the ice shelf to the ocean through large river networks.

Various physical factors will determine which of the above processes dominates for individual ice shelves. For example, relatively flat topography and extensive snow coverage will encourage surface-water ponding, which is likely to increase instability (Fig. 1a). Conversely, steeper slopes and bare ice surfaces will encourage water flow and stream development, potentially offsetting some of this increased instability (Fig. 1b). The latter scenario is not currently accounted for by ice-sheet models, which assume that all meltwater is stored on top of ice shelves, making them increasingly unstable. For example, the results of these models suggest that some of Antarctica's major ice shelves — such as the Amery, Filchner–Ronne, Larsen C and Ross — will disintegrate after melt rates exceed 1.5 metres per year, in the next century3. However, Bell and colleagues' analysis of the surface topography of these four ice shelves puts this prediction into question.

Figure 1: Ice-shelf dynamics.

Kingslake et al.6 show that widespread lake and stream networks have existed on many Antarctic ice shelves for decades. These drainage networks are usually thought to be hazardous to ice-shelf stability, but Bell et al.7 suggest that this instability can be mitigated if river networks export a large fraction of surface meltwater to the ocean. a, An ice shelf that has a relatively flat topography is likely to have numerous lakes and small streams. Some of these streams will transport water from the glacier, across the 'grounding line', and onto the ice shelf — but few will export water to the ocean. Such an ice shelf could be unstable because of the weight of water on its surface, leading to the formation of crevasses, particularly near the ice front, and enhanced iceberg calving. b, Conversely, an ice shelf that has a steeper slope might have fewer lakes and a large river network, offsetting some of the instability caused by the presence of meltwater by exporting surface water to the ocean through an ice-edge waterfall.

These two studies suggest that the surface hydrology on Antarctica's ice shelves will play a crucial part in deciding their individual fates, and those of the outlet glaciers that feed them. However, the authors do not explicitly address the likely additional role of increased melting on the undersides of ice shelves caused by ocean warming4. Given that the Antarctic Ice Sheet contains enough ice to raise global sea levels by 60 m (ref. 13), identifying and quantifying the role of all surface and subsurface processes on the potential stability of ice shelves is becoming increasingly important. Footnote 1


  1. 1.

    See all news & views


  1. 1

    Fürst, J. J. et al. Nature Clim. Change 6, 479–482 (2016).

    ADS  Article  Google Scholar 

  2. 2

    De Rydt, J., Gudmundsson, G. H., Rott, H. & Bamber, J. L. Geophys. Res. Lett. 42, 5355–5363 (2015).

    ADS  Article  Google Scholar 

  3. 3

    DeConto, R. M. & Pollard, D. Nature 531, 591–597 (2016).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Pritchard, H. D. et al. Nature 484, 502–505 (2012).

    ADS  CAS  Article  Google Scholar 

  5. 5

    Lenaerts, J. T. M. et al. Nature Clim. Change 7, 58–62 (2017).

    ADS  Article  Google Scholar 

  6. 6

    Kingslake, J., Ely, J. C., Das, I. & Bell, R. E. Nature 544, 349–352 (2017).

    ADS  CAS  Article  Google Scholar 

  7. 7

    Bell, R. E. et al. Nature 544, 344–348 (2017).

    ADS  CAS  Article  Google Scholar 

  8. 8

    Scambos, T. et al. Earth Planet. Sci. Lett. 280, 51–60 (2009).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Banwell, A. F., MacAyeal, D. R. & Sergienko, O. V. Geophys. Res. Lett. 40, 5872–5876 (2013).

    ADS  Article  Google Scholar 

  10. 10

    Langley, E. S., Leeson, A. A., Stokes, C. R. & Jamieson, S. S. R. Geophys. Res. Lett. 43, 8563–8571 (2016).

    ADS  Article  Google Scholar 

  11. 11

    David, T. W. E. & Priestley, R. E. Geological Observations in Antarctica by the British Antarctic Expedition 1907–1909 (Lippincott, 1909).

    Google Scholar 

  12. 12

    Trusel, L. D. et al. Nature Geosci. 8, 927–932 (2015).

    ADS  CAS  Article  Google Scholar 

  13. 13

    Vaughan, D. G. et al. in Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds Stocker, T. F. et al.) 317–382 (Cambridge Univ. Press, 2013).

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Alison Banwell.

Related links

Related links

Related links in Nature Research

Palaeoclimate science: Pulsating ice sheet

Climate science: Cooling in the Antarctic

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Banwell, A. Ice-shelf stability questioned. Nature 544, 306–307 (2017).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Sign up for the Nature Briefing newsletter for a daily update on COVID-19 science.
Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing