Climate science

How Antarctic ice retreats


New records of iceberg-rafted debris from the Scotia Sea reveal episodic retreat of the Antarctic Ice Sheet since the peak of the last glacial period, in step with changes in climate and global sea level. See Letter p.134

About 19,000 years ago, the ice sheets that covered large areas of North America and Eurasia began to melt. The ice on Antarctica also melted, but to a lesser extent — most of it still exists today. Understanding past ice-sheet instability and melting is important for predicting future ice behaviour in a warming world. On page 134 of this issue, Weber et al.1 present new well-dated records of iceberg-rafted debris from two marine sediment cores from the Scotia Sea that reveal at least eight episodes of ice loss from Antarctica between 20,000 and 9,000 years ago. These records allow examination of interactions between temperature, ice melt and the water masses of the Southern Ocean, which are central to the carbon cycle and to climate change between glacial and interglacial periods.

Since the Last Glacial Maximum (LGM) — the peak of the most recent glacial period — which occurred between about 26,000 and 19,000 years ago2, melting ice sheets have raised global sea level by about 130 metres (ref. 3). But there are significant uncertainties in the timing and amount of ice lost from Antarctica. For example, estimates of sea-level rise resulting from Antarctic ice melt have ranged from about 8 m to 30 m, with the most recent estimates around the lower end of this range4.

Currently, Antarctica loses ice by two main processes: melting of the underside of floating ice shelves and calving of icebergs. The icebergs melt slowly as they are carried westwards along the coast of Antarctica, and icebergs that reach the Weddell Sea are blocked by the Antarctic Peninsula and turn northwards, where they melt more rapidly in the warmer waters of the Scotia Sea. Icebergs themselves are ephemeral, but they carry mineral grains and rock fragments scoured from Antarctic bedrock. As icebergs melt, this iceberg-rafted debris (IBRD; Fig. 1) falls to the seabed and is steadily buried in marine sediments to form a record of iceberg activity.

Figure 1: Iceberg-rafted debris from marine sediment offshore of Porpoise Bay, East Antarctica.

Elizabeth Pierce

Weber et al.1 describe similar iceberg-rafted debris from two marine sediment cores from the southern Scotia Sea that document several episodes of ice loss from Antarctica between 20,000 and 9,000 years ago. Scale bar, 2 mm.

Some icebergs are more debris-rich than others: large tabular icebergs calved from floating ice shelves have already lost much of their debris-rich bases to melting, whereas icebergs that calve close to the line between grounded and floating ice tend to retain their debris, and are probably more common during ice-sheet retreat than during times of ice-sheet stability. Nevertheless, the main interpretive link is sound — more IBRD is a sign of more icebergs and greater ice loss from the Antarctic Ice Sheet. The two IBRD records reported by Weber and colleagues are similar, even though the sediment-core sites are separated by 2° of latitude and are subject to different local oceanographic conditions. This similarity increases confidence that the IBRD records represent an iceberg signal.

Until now, the timing of ice retreat has been constrained by radiocarbon dating of marine sediments and by dating of land surfaces that were uncovered as the ice sheets thinned5,6. Here, the eight episodes of iceberg discharge were dated by matching the record of wind-blown dust in the same sediment cores to wind-blown dust in an already-dated Antarctic ice core. This approach results in a continuously dated record, which provides a significant advance in knowledge of when the Antarctic Ice Sheet retreated over the time since the LGM. Weber et al. find that the first of the Antarctic ice discharges took place 20,000–19,000 years ago and was followed by a series of larger episodes between 17,000 to 9,000 years ago.

The largest iceberg release lasted from 14,800 to 14,400 years ago and overlapped, within dating uncertainty, with a period of sea-level rise known as meltwater pulse 1A (MWP-1A), which occurred 14,650 to 14,310 years ago7. During this period, sea levels rose by about 14–18 m at the astonishing rate of 4 m or more per century. Weber and colleagues' iceberg-discharge data clearly show a contribution to MWP-1A from Antarctica, but how much meltwater does this represent? Recent work4 puts the total budget for sea-level rise from Antarctic ice melt since the LGM at about 9 m, and this melt budget has to be shared among the eight iceberg-discharge events, including MWP-1A. By this reckoning, the Antarctic contribution to MWP-1A is relatively minor compared with the contribution from Northern Hemisphere ice, which must provide the balance of the roughly 14–18-m rise in sea level. This result contrasts with modelling of differences in amplitude of the MWP-1A sea-level rise at different locations (caused by the gravitational and rotational effects of removing the ice mass), which estimates that half or more of MWP-1A comes from Antarctica7. This mismatch has yet to be fully resolved.

Questions remain about which areas of the Antarctic Ice Sheet became unstable and produced these iceberg-discharge events. Did major ice-drainage sectors retreat simultaneously, perhaps in response to external triggering such as an initial sea-level rise from the north or southward migration of relatively warm Circumpolar Deep Water, as modelled by the authors? Or did different sectors retreat independently, as individual thresholds for instability were crossed in each sector? Principal sources of icebergs were probably the nearby Antarctic Peninsula and Weddell Sea embayment, where ice streams drain about a quarter of Antarctic ice by area. Icebergs are also likely to have travelled from other ice outlets around East Antarctica. The provenance of the IBRD, and the icebergs that carried it, can be found by matching the geochemical fingerprint (such as characteristic argon-isotope ages) of individual mineral grains in the IBRD to the corresponding geochemical fingerprint of the different source areas8 — a topic for future study.

The episodic iceberg discharges described by Weber and colleagues shed light on the question of how the Antarctic ice sheets melt. Will there be similar iceberg releases in the future? Ice streams in the Amundsen Sea sector of the West Antarctic Ice Sheet are already in the early stages of retreat9,10. During the last interglacial period, about 125,000 years ago, sea levels reached 6–9 m higher than today11,12, much of this attributable to an Antarctic meltwater source, at global temperatures only 1–2 °C warmer than those of today. The planet is on course for a temperature rise exceeding this value, so we can expect similar ice-sheet instability and retreat to that described by Weber et al. in the future.


  1. 1

    Weber, M. E. et al. Nature 510, 134–138 (2014).

    CAS  ADS  Article  Google Scholar 

  2. 2

    Clark, P. U. et al. Science 325, 710–714 (2009).

    CAS  ADS  Article  Google Scholar 

  3. 3

    Austermann, J., Mitrovica, J. X., Latychev, K. & Milne, G. A. Nature Geosci. 6, 553–557 (2013).

    CAS  ADS  Article  Google Scholar 

  4. 4

    Whitehouse, P. L., Bentley, M. J. & Le Brocq, A. M. Quat. Sci. Rev. 32, 1–24 (2012).

    ADS  Article  Google Scholar 

  5. 5

    Heroy, D. C. & Anderson, J. B. Quat. Sci. Rev. 26, 3286–3297 (2007).

    ADS  Article  Google Scholar 

  6. 6

    Hillenbrand, C.-D. et al. Quat. Sci. Rev. (2013).

  7. 7

    Deschamps, P. et al. Nature 483, 559–564 (2012).

    CAS  ADS  Article  Google Scholar 

  8. 8

    Pierce, E. L. et al. Paleoceanography 26, PA4217 (2011).

    ADS  Article  Google Scholar 

  9. 9

    Joughin, I., Smith, B. E. & Medley, B. Science (2014).

  10. 10

    Rignot, E., Mouginot, J., Morlighem, M., Seroussi, H. & Scheuchl, B. Geophys. Res. Lett. (2014).

  11. 11

    Kopp, R. E., Simons, F. J., Mitrovica, J. X., Maloof, A. C. & Oppenheimer, M. Nature 462, 863–867 (2009).

    CAS  ADS  Article  Google Scholar 

  12. 12

    Dutton, A. & Lambeck, K. Science 337, 216–219 (2012).

    CAS  ADS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Trevor Williams.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Williams, T. How Antarctic ice retreats. Nature 510, 39–40 (2014).

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.


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing