Professor Rohling, who is currently based at the National Oceanography Centre Southampton but will join the Australian National University next year, says: "This is the first time that these rates could be measured for any other period than the end-of-ice age 'terminations/deglaciations'. Although it is always hard to step from palaeo reconstructions to future projections, it suggests that when significant ice-volume adjustments happen, they are rarely slow.

"Ice sheet responses to a change in climate forcing are like the responses of heavy freight trains to firing up the locomotive. They are hard to set in motion (slow to 'spin up'), but once they are reacting, they will be equally slow to 'spin down'. So a lag of a few centuries is worrisome, because we have been warming up the climate for 150-160 years now. If the natural relationship (when changes in climate were slower than today) also holds for the very fast changes in climate today, then we are coming into that 'window' of time where we may expect to start seeing some unprecedented responses in the large ice sheets. This then may tie in with observations of the past decade or so of large ice-shelf collapses around Antarctica and Greenland, the major melt-area expansion over Greenland, changes in the flow speed of major ice streams (both Antarctica and Greenland), and increasing ice-mass loss over West Antarctica/the Antarctic Peninsula and Greenland.