EMBARGOED BY NATURE FOR RELEASE AT 2 P.M. EST MARCH 24

Nature has a way of throwing curves at glacial scientists trying to diagram climate history and find evidence of cycles that might have altered weather repeatedly during hundreds of thousands of years.

Like tree rings, each glacial layer corresponds with a particular year and contains clues about what was happening in the atmosphere. But as researchers drill cores through Antarctic ice sheets, it is difficult to be sure each layer contains snow that actually fell at that spot and that the layer is at the right level to correspond with a particular time frame.

They can surmount the first problem by drilling at an ice divide, essentially a peak on the glacier that forces ice to flow downhill one way or another, the same as an alpine continental divide does with water.

Now they have a new tool to overcome the second difficulty - ground-penetrating radar that shows precisely at which level each ice layer lies. University of Washington researcher Edwin Waddington, an acting geophysics professor, and three colleagues from the British Antarctic Survey describe the radar's effectiveness in Thursday's (March 25) issue of the journal Nature.

British researchers David Vaughan, Hugh Corr and Christopher Doake used radar to produce images to a depth of about 100 meters at an ice divide at Fletcher Promontory in Antarctica. At the divide's core, the images show an arch formed as each new layer settles lower on either side of the core. The authors call such arches "Raymond Bumps" for UW geophysics professor Charles Raymond, who postulated their existence in 1983. Raymond showed that ice deep under an ice divide should be very hard and slow to flow, so the upper layers would tend to drape themselves over it, much as a blanket over a sleeping person settles lower to either side. The radar images also show other distortions, and indicate the ice is behaving in a way inconsistent with researchers' normal assumptions when modeling flow within ice sheets.

Ice cores could produce skewed findings if scientists are unaware of the bumps' existence. Waddington, who helped interpret the radar data, notes that some cores are so well stratified that precisely dating each layer is no problem. "What is a problem is that you know these layers are getting thinner and thinner toward the bottom because they're being compressed from above and stretched to the side by the force of downhill flow," he says.

Computer models are able to calculate the forces that compress and stretch the layers, Waddington says. But they must accurately predict the size of the subsurface bump or they won't precisely reflect how thick each layer actually is and the research will yield false results about snowfall rates in the past. The findings emphasize the need for glacial geophysicists to help interpret ice core data, he says.

The models use information about the thickness of a given ice layer at various spots. From that, a snowfall average for a given year can be deduced. Then scientists studying the corresponding layer of an ice core can determine, by its thickness, whether the core site had more or less snow than nearby areas, and whether the amount might have been influenced by climate phenomena. The new research shows radar can be used to map a large area to determine snowfall patterns and average thickness of individual layers, Waddington says. That will give scientists an idea of the best places to drill, and also will provide a record for comparison with predictions from climate models.

The British scientists used commercial radar in their work, but Waddington notes that the UW geophysics department has developed its own complementary radar for glacial study. That radar, pulled behind a snowmobile, can't image the top 50 meters "but it's very good at seeing all the way to the bottom in places where the ice is one to two kilometers thick," he says.

The radar finding is important as scientists use Antarctic ice cores to understand changes in atmospheric chemistry, temperature and precipitation through long periods of history. The most complete ice core goes back to about 500,000 years ago, through four ice ages. Researchers are looking in more recent ice layers for signals that tell of cycles such as the El Nino Southern Oscillation, which produces the El Nino and La Nina climate phenomena. They also are seeking evidence of very large and very fast climate changes.

"They're looking for ENSO signals in the ice cores through the past several hundred years, and rapid climate shifts through ice ages of 100,000 years apiece," Waddington said.

###

For more information, contact Waddington at (206) 543-4585 or [email protected];
Vaughan at 44 01223 221481 or [email protected];
Corr at 44 01223 221496 or [email protected];
or Doake at 44 01221 221488 or [email protected].