Large prehistoric earthquake ripped Seattle.

On a wintry day about 1,000 years ago, a fault running beneath the site of Seattle's Kingdome let loose a whopper of an earthquake that sent a tsunami wave sloshing through Puget Sound. The ground shook with such fury that avalanches tumbled from the Olympic mountains and landslides ran into Lake Washington, near where some of Seattle's tonier neighborhoods now stand.

No written documents record the quake. But through a remarkable series of complementary studies, several teams of geologists have pieced together enough evidence to show that a strong quake did strike Seattle -- a realization that will force hazard planners there to reconsider the seismic risk to their city. Until now, seismologists had not considered a threat so close to Seattle.

The separate research groups present evidence of the prehistoric quake in five papers published in the Dec. 4 SCIENCE.

"It's a combination of all these diverse lines of evidence that suggests there was a large earthquake in the Seattle area about 1,000 years ago. What this has shown is that something like this can happen, which wasn't known before," says geologist Robert C. Bucknam of the U.S. Geological Survey in Denver.

The convergence of geologic techniques used in this case study can also help researchers study the seismic hazard of other regions around the globe.

Bucknam and his colleagues discovered evidence of the prehistoric Seattle earthquake while studying a raised terrace at Restoration Point on Puget Sound. Cut by ocean waves, the terrace once lay at sea level. But the former shoreline now sits some 7 meters above high-tide level. Bucknam's group believes the land must have risen quite abruptly, because the ocean did not cut any terraces between the current and former shorelines. Carbon-14 dating of organic material on the terrace indicates that the uplift occurred between 500 and 1,700 years ago.

This raised shoreline at Restoration Point and another one in Seattle lie directly south of a geologic structure that parallels interstate 90, running in an east-west direction through the city. Geologists have suspected the structure was a fault and this year named it the Seattle fault. But until Bucknam's group documented the evidence of uplift, it remained unclear whether the fault was active.

While land south of the fault went up. Bucknam's group sees a different story on the fault's northern side. At Winslow, just 5 kilometers north of Restoration Point, the land hasn't gone through any sudden change in level during the last 2,000 years. This finding helps provide a measure of the prehistoric earthquake's size. In 1980, a magnitude 7.3 earthquake in Egypt jacked up land on one side of a fault by 5 meters -- an amount comparable to the uplift in Seattle.

Because such rapid shifts in land height often cause tsunami waves, Brian F. Atwater and Andrew L. Moore of the University of Washington searched the region for evidence of past surges. At two sites, they found signs that a tsunami flooded tidal marshes, blanketing them in centimeters of sand. Using especially precise carbon-14 dating techniques, they found that the tsunami hit between 1,000 and 1,100 years ago.

In previous work, Atwater had found tsunami deposits along the Washington coast. This and other evidence convinced many scientists that the Pacific Northwest coast has produced great quakes of magnitude 8 or larger, caused by a piece of ocean floor subducting beneath the North American continent (SN: 2/17/90, p.104). The discovery of these tsunami deposits led many researchers to look within the Puget Sound area for signs of shaking caused by great coastal quakes. While the search has turned up hints of coastal shocks, some of the emerging evidence fits the idea of a jolt on the much closer Seattle fault, Atwater says.

In particular, tree-ring specialists have studied deposits left from landslides that carried trees and rock into Lake Washington. The researchers, from Lamont-Doherty Geological Observatory in Palisades, N.Y., and from the University of California, Berkeley, recovered trees from the lake and used carbon-14 dating to determine that the landslides occurred between 1,000 and 1,300 years ago.

With a bit of luck, they established an even more precise connection by comparing rings in Douglas firs found in Lake Washington with rings in a Douglas fir found in the tsunami deposit discovered by Atwater. Though these sites lie 23 km apart, the ring analysis shows that the trees all died within the same year and season -- sometime in the fall, winter, or early spring.

Other researchers studying sediments in Lake Washington found evidence ents in Lake Washington found evidence of a major disturbance about 1,100 years ago. In the Olympic mountains, geologists dated six prehistoric rock avalanches to between 1,000 and 1,300 years ago.

In a commentary in SCIENCE, geophysicist John Adams of the Geological Survey of Canada in Ottawa says that a repeat of the Seattle quake would shake the city far more than would a larger subduction quake along the coast. Damage from a shallow Seattle quake would also exceed that from deep quakes that hit Puget Sound in 1949 and 1965, Adams says.

Bucknam's group has also found evidence of uplift between 1,000 and 1,500 years ago in an area southwest of Seattle. He suspects a different fault may have caused the prehistoric uplift there, raising the possibility that the Puget Sound area has several active faults. While these quakes could cause considerable damage, geologists have yet to find enough evidence to determine how often they occur. "It might be thousands of years [from now] or it might be tomorrow," Bucknam says.