Credit: NASA GODDARD SPACE FLIGHT CENTER

Release of data gathered during the Cold War continues to deliver scientific surprises. The latest example emerged at last week's American Geophysical Union meeting in San Francisco, where glaciologists reported new findings about Antarctica based on satellite data. The revelation came from comparisons of modern satellite images of snow dunes (such as the one shown here) with recently declassified pictures originally taken by intelligence satellites in 1963.

The first complete map of Antarctica, produced in 1997 with data from the Canadian satellite Radarsat, revealed many unexpected features, including vast tracts of snow dunes. These megadunes are up to 100 kilometres long, lie 1 or 2 kilometres apart but are only a few metres high. In East Antarctica fields of the dunes cover an area larger than the state of California.

At the meeting Mark Fahnestock (University of Maryland) described how he and Ted Scambos (University of Colorado) compared data from satellites of the US National Oceanic and Atmospheric Administration and from the 1960s military images. It might be thought that the snow dunes would move, even if only slowly, because of the fierce, constant winds that blow across the East Antarctic plateau. It turns out, however, that they have not — at least over the past 30 years. Little is known about how the megadunes formed, but it is unlikely that they grew from drifting snow in the way that sand dunes are built from sand.

Across the other side of the continent, on the West Antarctic ice sheet, vast streams of ice flow into the sea. The source of the ice streams has so far eluded explorers and satellites alike. At another talk at the meeting, Robert Bindschadler (NASA Goddard Space Flight Center) described high-resolution data from Radarsat that reveals several small and slow-flowing tributaries that are feeding the large frozen rivers with snow from the interior. One of the biggest uncertainties in predicting sea-level rises in response to climate change is the uncertain behaviour of the Antarctic ice sheets. So new data (however old) is always welcome.