The Lowell Glacier usually comes to a somewhat abrupt stop at the edge of the Alsek River. Icebergs calving off of the glacier tumble into a wide spot on the river called Lowell Lake. But every once in awhile this glacier surges forward dramatically. At times it has completely blocked the Alsek River, creating a massive glacial lake.

As recently as 1852, Glacial Lake Alsek was 100 kilometres long and about 100 metres deep, making it larger than Kluane Lake. When the ice dam finally broke, it sent a wall of water down the Alsek River. Native stories tell of a group of Tlingits, camped at the confluence of the Alsek and Tatshenshini Rivers, who were drowned in the flood.

If the town of Haines Junction had existed back then, it would have been either underwater or people living there would have had lakefront property. About 15 kilometres of the Alaska Highway would also have been submerged. The Lowell Glacier surges every 15 to 20 years, and in the winter of 1998-99, this river of ice slid forward once again.

The glacier's last major advance was in 1997 when it almost choked Lowell Lake with ice. The glacier receded the following summer, but last winter it surged again, though not quite as far as in '97. These advances have led some people to wonder whether history could ever repeat itself in the Alsek Valley. Could Glacial Lake Alsek ever rise again?

This is not a completely fanciful idea. On hillsides high in the Alsek Valley, long benches show where waves once lapped the shore, and driftwood marks the site of former beaches. No one knows how many times the water of an ice-bound lake has filled the valley, but it is clear that the lake filled to different levels, at different times, over the last few thousand years.

But according to UBC glaciologist Garry Clarke, people in Haines Junction do not need to buy special flood insurance just yet. "I think there is almost no chance that a large lake could form right now because there is not enough ice there to form a high dam. The volume of the glacier is much lower now," he explains.

The last large lake formed during a generally cooler period toward the end of the Little Ice Age in the Yukon. The glaciers in the St. Elias Mountains were much bigger during this time. But Clarke does not completely rule out the possibility that someday a small lake could form again in the Alsek Valley.

"Such a lake would not be visible from Haines Junction and it might only back up the Alsek River a few tens of kilometres. Likewise the outburst floods from such a small lake would not be anywhere near as severe as those released from the former large-volume lakes," he says.

Clarke says there is evidence that a very small lake formed in this century. Crude skis were found on a beach low in the valley, indicating that a lake could have formed as recently as the 1920s.

Surging glaciers seem to go for regular gallops regardless of whether the climate is cooling or warming. When the Steele Glacier last surged, in the 1960s, it advanced more than eight kilometres in two years, sliding forward more than a metre per hour at times.

Clarke has been studying glaciers in the Kluane area since 1962. He usually spends the month of July on the Trapridge Glacier, which flows into the Steele Glacier. He says that almost all of the very large glaciers in the Kluane area are surge-type glaciers that leap forward at fairly predictable times.

But glaciologists are still not sure why some glaciers surge in the first place. There are about 250 surging glaciers in Alaska and the Yukon, but none in glaciated areas such as the Coast Mountains, the Rocky Mountains, the Alps or Scandinavia.

In the St. Elias Mountains, climate does not seem to be the controlling factor as surging glaciers occur both in the temperate marine environment found along the coast and in the dry cold climate of the interior. Clarke says that if neither precipitation nor temperature account for the surging, it is possible that geology plays a major role.

"Glaciers that rest on a sedimentary bed are more likely to surge than ones on hard bedrock. If you have mucked-up material, like till, underneath the ice, that seems to be good for surging," he says.

The key to surging glaciers seems to be a build-up of water underneath the ice. If the water cannot escape, either by running underneath the ice or filtering down through the earth, the glacier can surge forward on the cushion of water.

"What starts the surge is the shutting down of escape routes for the water. It does not need to be bathed in water. You might need only a few millimetres of water that is under pressure," says Clarke.

When the water eventually finds an escape route from underneath the ice, that can end the glacier's advance. "There is a flood burst at the termination of the surge. The glacier burps its water out and that ends the surge," he says.

In the spring and summer, there can be plenty of water in the icefields. When snow on the surface of the glacier melts, the water trickles down through crevasses, or cracks, in the ice. But how does a glacier surge in the dead of a Yukon winter?

"Winter surges are one of the really interesting things because they seem to fly in the face of the simple view that water forces glaciers to surge. In winter there is no natural drainage for water, or it is poorly developed.

"So if the glacier does start to move in winter, it generates heat because of friction. So the glacier could generate its own water through friction, and in winter there is no place for that water to go," says Clarke.

Clarke says that the spring melt could end the Lowell's surge this summer. When more water starts trickling down through the glacier, the increased flow can build a drainage channel underneath the ice and the water can escape.

"That pattern is not unusual. You can get active pulses of surges. On the Lowell you might see a little more surging every winter and eventually it will just stop doing that," says Clarke.

Garry Clarke can be reached through the Department of Earth and Ocean Sciences at the University of British Columbia. In the summer he works out of the Arctic Institute at Kluane Lake.