News & Views

Climate science: The long future of Antarctic melting

Simulations show that melting of the Antarctic ice sheet in response to climate change could raise the global sea level by up to 3 metres by the year 2300 and continue for thousands of years thereafter. See Letter p.421

Most projections of Antarctic ice melting in response to climate change extend a maximum of a few centuries into the future, a timescale that has clear relevance to immediate human affairs. But to capture the total Antarctic contribution to sea-level rise caused by climate change, it is necessary to consider the possibility that ice-sheet mass loss will continue for thousands of years. On page 421 of this issue, Golledge and colleagues1 present multi-millennial ice-sheet simulations in which Antarctica continues to contribute significantly to global mean sea-level rise for more than 1,000 years, long after ocean and air temperatures have stopped increasing. The authors' simulations also show that ice-shelf melting driven by ocean warming over the next 100–300 years is a critical factor in determining the total future rise in global sea level.

The spectre of the imminent and rapid loss of ice from the West Antarctic Ice Sheet has been raised by dramatic events such as the collapse of the Larsen B ice shelf in 2002 (Fig. 1). Nevertheless, the contribution of the Antarctic ice sheet to global sea-level rise is currently small in comparison with that of other sources, although it is increasing at an accelerating rate2.

Figure 1: Ice-shelf collapse.
Figure 1

These satellite images show the rapid break-up of 3,250 square kilometres of the Larsen B ice shelf in Antarctica in 2002. Golledge et al.1 have simulated the retreat of the Antarctic ice sheet in response to climate change over the next few millennia. Image: Ted Scambos, National Snow and Ice Data Centre

The glaciologist John Mercer was the first to suggest that past recessions of the West Antarctic Ice Sheet were driven by fluctuations in climate that increased air temperatures over ice shelves to above freezing point3. He proposed that ice-shelf melting causes an increase in the flow of ice from land that occurs through other processes. In many coastal regions of Antarctica, ice shelves are laterally constrained by bays. Confirming Mercer's suspicions, numerical models4,5 have since shown that contact between the ice shelf and bay sidewalls exerts a restraining force on the flow of ice from land, leading to slower flow and reduced ice calving into the ocean. When the climate warms, ice shelves may melt, reducing the restraining contact with bay sidewalls, accelerating ice flow from land and enhancing mass loss through ice calving.

Estimating the likelihood that climate change will cause widespread Antarctic ice-shelf loss, and the duration of ice-flow acceleration in response to such a loss, is crucial for constraining future sea-level rise. However, accurately modelling ice sheets, ice shelves and their interactions with the ocean and atmosphere is computationally intensive. Consequently, most sea-level projections from models that explicitly include ice-sheet flow and ice-shelf melting extend only a few centuries into the future, or include only part of the Antarctic ice sheet. Existing multi-millennial projections of the ice sheet's contribution to sea-level rise are based on relatively simple statistical relationships between global temperature and sea level derived from palaeoclimatic and instrumental data6,7.

In contrast to earlier studies, Golledge and colleagues use a comprehensive ice-sheet model, with forcing of precipitation, ocean and air temperature from global climate models, to simulate mass loss from the Antarctic ice sheet from the present to the year 5000 under various scenarios of climate change. They find that ocean warming, rather than atmospheric warming or changes in precipitation, is the dominant driver of mass loss. Ocean warming causes loss of large Antarctic ice shelves and a sustained acceleration in ice-sheet discharge into the ocean in all of the scenarios, except for one that includes the most extreme reduction in greenhouse-gas emissions compared with 1990 emission levels. In the scenarios in which ice shelves are lost, the long-term contribution of the Antarctic ice sheet to global sea-level rise ranges from roughly 3 to 9 metres. The majority of that contribution comes after 2300, with enhanced rates of sea-level rise lasting until at least 3000 — long after ocean temperatures have stabilized.

A previous theoretical analysis8 found that high model resolution is needed around the grounding line — which marks the transition from a grounded ice sheet to a floating ice shelf — to simulate rapid grounding-line migration accurately. Golledge and co-workers instead used a fairly coarse model resolution near the grounding line, because of the computational constraints of performing multi-millennial simulations of the entire Antarctic ice sheet.

For each climate-change scenario, they simulated an upper bound of sea-level rise using a scheme for correcting errors arising as a result of the coarse resolution near the grounding line, and a lower bound that was calculated by not applying such corrections. The sensitive dependence of projections of sea-level rise on such correction schemes is vividly illustrated by the gap of several metres between the simulated lower and upper bounds. Efforts to develop more-accurate ways of representing the grounding line4 that can be incorporated into multi-millennial ice-sheet simulations should continue, and will lead to a considerable reduction in uncertainties in long-term sea-level projections.

Golledge et al. ultimately confirm the suspicions of earlier glaciologists that the fate of ice shelves largely determines whether Antarctica contributes less than 1 metre or up to 9 metres to long-term sea-level rise. Although ocean warming is responsible for most ice-shelf melting in these simulations, other studies9,10 have suggested that warm air temperatures lead to water ponding on the surface of ice shelves. This eventually causes hydrofracture (deepening of fractures by the drainage of meltwater ponds) and the rapid break-up of ice shelves. Capturing such complex processes in models is difficult, but, in one study that included these effects11, Antarctic ice-shelf loss is even more rapid and widespread than in the current simulations. If such a rapid ice-shelf break-up does occur, then Golledge and colleagues' simulations might represent a best-case scenario for future sea-level rise.



  1. 1.

    et al. Nature 526, 421–425 (2015).

  2. 2.

    & J. Earth Planet. Sci. Lett. 415, 134–141 (2015).

  3. 3.

    IAHS Publ. 79, 217–225 (1968).

  4. 4.

    , & J. Geophys. Res. 114, F04026 (2009).

  5. 5.

    , , , & Cryosphere 6, 1497–1505 (2012).

  6. 6.

    , , & Nature Clim. Change 2, 867–870 (2012).

  7. 7.

    et al. Proc. Natl Acad. Sci. USA 110, 13745–13750 (2013).

  8. 8.

    J. Fluid Mech. 573, 27–55 (2007).

  9. 9.

    , & Antarctic Res. Ser. 79, 79–92 (2003).

  10. 10.

    , & Geophys. Res. Lett. 40, 5872–5876 (2013).

  11. 11.

    , & Earth Planet. Sci. Lett. 412, 112–121 (2015).

Download references

Author information


  1. Alexander Robel is in the Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, California 91125, USA, and in the Department of Geophysical Sciences, University of Chicago.

    • Alexander Robel


  1. Search for Alexander Robel in:

Corresponding author

Correspondence to Alexander Robel.


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.