Credit: NASA/GSFC/LARC/JPL, MISR TEAM

When a huge chunk of Antarctic ice shelf broke up in 2002, it provided dramatic pictures (see right) for the world's press and a control experiment for researchers. The ice shelf, Larsen B, is a floating extension of the ice of the Antarctic peninsula. The collapse of a substantial part of it — more than 3,000 km2 — was attributed to increasing temperatures and released shoals of icebergs into the Weddell Sea. But a southerly remnant remained in place, enabling ideas to be tested about how ice shelves might affect glacier flow from the continental interior.

Two groups now report their results of satellite-tracking glacier behaviour in the region (E. Rignot et al. and T. A. Scambos et al. Geophys. Res. Lett. 10.1029/2004GL020697; 10.1029/2004GL020670). They found that five glaciers flowing into the area formerly buttressed by the ice shelf all accelerated at various times, whereas two farther south, which ran into the remnant ice shelf, did not. Speed of glacier flow is also reflected in their thickness: higher flow rates stretch and thin the ice, in these cases yielding estimated rates of thinning of tens of metres per year.

The main implication is that ice shelves act as a restraint on glacier flow. This conclusion was by no means obvious. Earlier, theoretical studies gave conflicting results; and there are also possible confounding factors, such as water, produced by seasonal melting of surface ice, acting as a lubricant at the glacier base.

A prospect for the future — and a worrying one as far as larger ice shelves and glaciers are concerned — is that a feedback system could kick in, accelerating glacier melting and producing significant rises in sea level.