How Sleuths Of Science Uncovered Seattle Fault

"You see, but you do not observe." Sherlock Holmes

Call it the Mystery of the Drowned Forest.

The final clue that nailed the earthquake culprit that sent stands of old-growth forest sliding into Lake Washington 1,100 years ago was not a smoking gun, but a Douglas fir tree entombed in mud at Seattle's West Point.

Dating the well-preserved tree rings on the buried fir provided the final proof that the same disaster that caused the Lake Washington landslide about 900 A.D. also caused the West Point tree to be buried by a Puget Sound tsunami.

Clearly, something big had shaken the entire region. Scientists now propose that it can be traced to an earthquake fault that runs from Winslow on Bainbridge Island under downtown Seattle and eastward to the northern tip of Mercer Island and on to Lake Sammamish.

Science magazine last Friday called the research the best case of circumstantial evidence ever assembled by scientists to prove a historical earthquake.

Like detectives assembling a criminal case, more than a dozen scientists pulled together clues at sites ranging from the Olympic Mountains to the bottom of Lake Washington.

"People ought to see this as another opportunity to remind themselves that this is big-time earthquake country," said Craig Weaver, coordinator here of the U.S. Geological Survey (USGS) earthquake program.

The quake caused rock avalanches in the Olympics, uplifted southeastern Bainbridge Island 20 feet and Alki Point 13 feet, produced a tsunami that hit the southern tip of Whidbey Island, drowned West Point with a monster wave, and sent trees sliding 200 to 400 yards into Lake Washington.

About 1,100 years after the event, the evidence is still there. But it took scientists who, like Sherlock Holmes, observed - and not just saw - to recognize it.

The case has been developing for some time. Settlers on Lake Washington had long known of a bizarre submerged forest of huge old-growth firs that reared up from the bottom and posed a menace to navigation.

In 1919, 186 remarkably preserved trees were removed from three spots in the lake at depths that ranged from 65 to 132 feet. Certainly the trees had not grown underwater. The best guess was that a landslide had carried them upright into the lake where the cold water preserved them; some were still being salvaged for wood as recently as last year. Radiocarbon dating suggested they had been in the water between 800 and 1,400 years.

There was another mystery in Seattle. The gravitational pull of the Earth varies on its surface depending on the thickness and density of the crust. The second-sharpest gravitational field variation in the United States, after a point in the Sierra Nevada mountains, occurs in the middle of this city. South of that line, gravity's pull is stronger than north, suggesting some kind of division in the underlying crust.

In fact, the wishful weight watcher can find comfort that he or she actually weighs less (imperceptibly so) at Northgate than at Southcenter. Bedrock is thicker and a half-mile closer to the surface in the south than in the north.

Two geologists, James Yount and Mark Holmes, proposed that such a sharp division suggested a shear in the crust, what they dubbed the Seattle Fault.

Scientists also knew earthquakes had rocked Puget Sound in 1909, 1939, 1946, 1949 and 1965. The 1949 quake measured magnitude 7.1 and the 1965 quake measured magnitude 6.5, causing more than $12 million damage. Is there a potential here for something worse?

A USGS scientist based at the University of Washington, Brian Atwater, was finding evidence on Washington's coast that indeed there is. After studying land that had abruptly subsided, killing trees and creating a salt marsh, he concluded that a magnitude 9 earthquake - considerably bigger than the 8.1 quake that destroyed San Francisco in 1906 - had shaken Washington's coast about 300 years ago.

Atwater explained last year that there was probably a subduction zone quake, caused when the Juan de Fuca plate under the Pacific Ocean slips beneath the continental plate of North America.

Similar quakes had occurred under Puget Sound as well, but this far eastward the junction of the two plates is 30 miles deep, reducing the potential damage.

But what if there was a quake closer to the surface, in a fault in the surface crust? That would be close enough to cause immense damage.

Another USGS geologist from Colorado, Robert Bucknam, decided to apply Atwater's methods to Puget Sound. Geologists had already mapped places on the Sound that gave evidence of past earthquakes. Bucknam, along with botanist Estella Leopold of the University of Washington and geologist Eileen Hemphill-Haley of the USGS in California, decided to take a closer look and try to date the damage.

What they found was a clear pattern. Southwest of Seattle, land had risen abruptly about 1,100 years ago. At Restoration Point on Bainbridge Island, sandstone had abruptly been shoved 20 feet high. Study of aerial photographs showed a beach bar had been marooned inland.

Salt marshes in Southern Puget Sound and the tip of Hood Canal had also been left high and dry. At Alki Point, excavation of a building lot indicated the old beach had been heaved upward about 13 feet.

The pattern was close to that of a 1980 earthquake of magnitude 7.3 at El Asnam, Algeria, where the land lifted more than 30 feet on one side of a fault.

Meanwhile, another Denver USGS geologist, Robert Schuster, and two state Department of Natural Resources geologists, Robert Logan and Patrick Pringle, were finding evidence of past catastrophe in the Olympic Mountains.

They found 11 prehistoric rock avalanches that in five cases dammed streams and created lakes. In six cases, they used buried wood or drowned trees to date the slides and found three or four occurred at the same time: about 1,100 years ago. The plot thickened.

While those geologists were up in the mountains, other scientists were down in lake-bed mud. Sally Abella of the University of Washington, a limnologist, or lake expert, and Robert Karlin of the Nevada School of Mines, a magnetism expert, took cores of Lake Washington sediment. The idea was to see if earthquake-caused landslides into the lake left a record in the silt.

It did. The pair measured the susceptibility of the mud to become magnetic. Sediment with the least dead organic matter, they knew, was more likely to magnetize. And if there were no remains of plants or animals, such sediment probably came from a flood or landslide. Sure enough, there was a "spike" in magnetic susceptibility from the sediment of 1,100 years ago.

Earlier studies by Leopold had already ruled out one possibility to explain this spike: that there was some dramatic change in the surrounding watershed. Leopold's examination of old pollen showed no significant vegetation change the past 6,000 years. Abella said the pair also knew the Cedar River at that time did not flow into Lake Washington, and the Sammamish River was incapable of bearing enough flood sediment.

As Sherlock Holmes said, when you've eliminated the impossible, that which remains, however improbable, must be the truth. The earthquake that carried old-growth forests into the lake must also have carried in the fresh sediment that caused the spike. Chalk up another clue.

The science sleuths still weren't done. If land under Southern Puget Sound had reared up 1,100 years ago, the displaced water should have formed a tsunami, or tidal wave, that would have rolled northward. Was there any evidence?

Atwater's graduate student, Andrew Moore, waded into a peat bog at an obvious spot, the head of Cultus Bay at the south end of Whidbey Island. There, forming a bright stripe against the dark peat, was a layer of sand laid down by the wave. "Ooh, this was interesting," Atwater recalled. "So we sent in the layer for dating and it came back at 1,100 years."

Shortly afterward, excavation for the construction of the West Point secondary sewage treatment plant began. Curious things began to turn up. First there was a midden of shells left by Native Americans, below the low tide mark. Had West Point subsided as land to the southwest had risen?

Moreover, that same telltale layer of sand showed up, deposited so suddenly that it shut out oxygen that would have allowed decomposition of marsh bullrushes. Those stalks remained preserved when uncovered 1,100 years later. "You could pull these 1,000-year-old sticks out and snap them," Atwater said.

Even more astounding, construction crews uncovered a big Douglas fir log virtually undecayed except for its top layer of bark. Did this tree and the drowned forests of Lake Washington die at the same time?

Gordon Jacoby and Patrick Williams of the Tree-Ring Laboratory at Lamont-Doherty Geological Observatory in Palisades, N.Y., were called in to find out. In return for showing a wood salvor where the underwater trees of Lake Washington were, they got cross-sections to examine. They took another cross-section from the West Point tree.

"The results reveal that the West Point tree and all the trees from the landslides south of Mercer Island died in the same season of the same year," they concluded in their research paper.

There you have it, Watson. Uplift. A sharp break in the gravitational field. Olympic rock slides. Drowned forests. Lake-bed mud. A layer of sand in a peat bog. A sunken midden. Preserved trees 15 miles apart that died in the same catastrophe. Elementary, old chap. An earthquake, estimated at magnitude 7 to 7.5 judging from the damage and the limited length of the Seattle Fault.

Case closed? Not quite.

The obvious question is could it happen again, and how soon? The Sherlock scientists don't know. It is a big event that hasn't repeated in more than 1,000 years. "I have a hard time striking a middle ground, viewing this as either very sensational or very trivial," admitted Atwater. It is hard to express a level of risk.

Another puzzle is why does the Seattle Fault run east to west, when most West Coast faults run north to south because of the colliding ocean and continental plates. "Why is it perpendicular to the plate boundary?" Atwater pondered.

Even more intriguing is whether such quakes have happened many times. The lake-bed cores taken by Abella and Karlin suggest similar disturbances on four other occasions. One, about 300 years ago, may coincide with Atwater's coastal quake. Others are about 1,650 years ago, 2,300 years ago, and 3,000 years ago.

"I look at my computer and wonder if I should move it another inch back from the edge," Abella joked. "But if we had an earthquake that big, I don't think it would matter where my computer was."