A New Look at an Old Landslide

Radiocarbon dates indicate the Bonneville Landslide may be far younger than thought

The Oregonian - Portland, Oregon 9/29/99

By Richard L. Hill, staff reporter

A long-dead tree may hold the clues for helping solve a few mysteries about one of the Northwest's most famous landslides.

The Bonneville Landslide in the Columbia Gorge about 30 miles east of Portland covered more than 5.5 square miles, producing a temporary dam and immense lake on the Columbia River that probably led to the Native American legend about the "Bridge of the Gods."

Despite it being one of the region's most studied landslides, questions about exactly when the catastrophic slide occurred — and its precise effect on the landscape and people — remain. Answers could help scientists understand the possible effects from future large landslides in the gorge.

Radiocarbon dates from the core of a Douglas fir buried 150 feet under the massive slide indicate it killed the fir about 400 years ago and perhaps as recently as 250 years ago. That would make the landslide half a millennium younger than a previous estimate, which said the slide occurred about A.D. 1100.

If future work supports the younger date, the slide would be in the same time frame as the last huge offshore earthquake, which rocked the Northwest coast in 1700. Although scientists are confident a quake caused the landslide, they say it's premature to link it with the magnitude 9 earthquake.

From analyzing the radiocarbon dates, "my feeling is that the landslide likely happened between 1550 and 1750," said Patrick T. Pringle, a geologist with the Washington Department of Natural Resources who has been conducting the study with Robert L. Schuster, a landslide expert with the U.S. Geological Survey. "I realize what a big window this is, but that's about all the data allow us to say."

Most recent of four slides The Bonneville Landslide, which tumbled from Table Mountain, has intrigued scientists for decades. It is the youngest and largest of four adjacent slides that make up the 14-square-mile Cascade Landslide Complex north of the Columbia near Cascade Locks and Stevenson, Wash. The area on the Columbia's north side is prone to landslides because of steep terrain made up of formations that tip toward the river. Columbia River basalt overlies the fragile, clay-filled Eagle Creek and Weigle formations. The cliffs exposed when the mountain gave way easily can be seen north of Bonneville Dam.

"What you've got is a deck of cards that is pointing and sliding toward the river," said Alex Bourdeau, an archaeologist with the U.S. Fish and Wildlife Service who is interested in the landslide's effect on the people who lived along the river. "I always said all you had to do is jump up and down on top of Table Mountain and you could have triggered this slide."

The landslide unleashed blocks of rock as large as 800 feet long and 200 feet thick down the mountain, creating a temporary earthen dam more than 200 feet high — three times the height of Bonneville Dam. The slide covered a 3.5 mile stretch of the river, shoving it about a mile off its course. No one knows exactly how long the river was blocked. One estimate is that it took about two years for the water to rise to the top of the dam, creating a huge lake that may have stretched 100 miles east to Arlington. The lake drowned a narrow forest of trees for 35 miles. About 1,800 of the stumps were visible in the river before they were again submerged in 1938 by the reservoir created by the Bonneville Dam.

Eventually, the lake rose high enough to cut through and spill over the barrier, unleashing a catastrophic flood that was nearly 100 feet deep at Troutdale and eroding much of the landslide. Ives, Hamilton and Pierce islands are remnants of the slide, while the uneroded portions produced the famous "Cascades of the Columbia." The cascades, or series of small waterfalls, produced by the slide provided the name for the Cascade Range — perhaps the only time a landslide indirectly led to the naming of a mountain range, said Scott Burns, a geology professor and landslide expert at Portland State University.

Explorers note obstruction Lewis and Clark were the first to document the landslide and its effects. Heading downstream in October 1805, the explorers described the river as being "obstructed by the projection of large rocks, which seem to have fallen promiscuously from the mountains into the bed of the river."

They added "that there are stumps of pine trees scattered for some distance in the river, which has the appearance of being dammed below and forced to encroach on the shore."

When the explorers returned upstream the following spring, they again mentioned the tree trunks standing in the water. They correctly stated that "the passage of the river through the narrow pass at the rapids has been obstructed by the rocks which have fallen from the hills into the channel," although they were off in their estimate that the landslide had occurred "within the last 20 years."

While the Columbia was dammed by the slide, area inhabitants might have been able to cross the river on foot, which probably gave rise to Native American stories about a bridge near Cascade Locks. One version relates how Wy'east (Mount Hood) and Pahto (Mount Adams) were powerful braves, the sons of Old Coyote. They both fell in love with a maiden (Mount St. Helens), and they frequently crossed a bridge over the Columbia to fight each other. Coyote caused the bridge to collapse in an effort to keep the feuding brothers apart.

In the late 1830s, Daniel Lee wrote in an account of the region's geology that "the Indians say these falls were not ancient and that their fathers voyaged without obstruction in their canoes as far as The Dalles. They also assert that the river was dammed up at this place, which caused the water to rise to a great height far above, and that after cutting a passage through the impending mass to its present bed, these rapids first made their appearance."

Signs of flooding Researchers with the U.S. Geological Survey have been examining sites downstream from the slide area to determine the effects of the flood unleashed by the river's breakthrough of the natural dam.

"We've been looking at backwater deposits in the Sandy River from a big flood that dumped a lot of Columbia River sand," said Thomas C. Pierson, a hydrologist at the agency's Cascades Volcano Observatory in Vancouver, Wash. "The sand is nearly 100 feet above sea level, which would make it a flood of about 80 feet deep at the mouth of the Sandy River."

Pierson said research suggests that two floods caused by the breaching of landslide-dammed lakes have occurred — one about 450 years ago and the other 1,600 years ago.

Pringle said the emerging evidence and concerns about future slides make it important to study other large slides in the Columbia. "There is plenty of material still present in the gorge to pose future threats," he said. "In fact, these types of landslides commonly leave steep scarps that may themselves be susceptible to failure."

Research on the landslide intensified in the 1930s when the Bonneville Dam was being built and in the 1970s with the construction of the second Bonneville powerhouse, which was completed in 1978.

Looking upstream Archaeologists say the landslide had a significant impact on the native inhabitants. No evidence exists that a village or seasonal camp site was destroyed by the slide itself, although the filling of the lake and the later "outburst" flood would have inundated any dwellings.

"One of the problems has been that everyone's attention primarily has been focused downstream from the event," said Bourdeau, who has been studying the slide for 20 years. "What people haven't done is go upstream and look for villages that would have been drowned by this big lake filling behind the landslide. They probably exist, but unfortunately they're now all drowned again by all the dams, so it makes them difficult to look for."

Bourdeau said the slide would have had a serious effect on migrating salmon if the earthen dam had been there a couple of years.

The eventual erosion of the dam and the creation of the rapids, however, led to a boom in the native population along the river. The cascades formed the narrowest constriction in the gorge, obstructing anadromous fish runs and providing an ideal place to harvest the fish.

In addition, because the rapids formed an obstruction to river transportation, travelers had to portage around the barrier. Bourdeau said the Chinook placed villages on each end and at the center of a 4-mile-long trail that went around the rapids, enabling them to control river trade and travel. By collecting tolls, they were able to increase their wealth and power.

A report in 1984 by Rick Minor, an archaeologist with Heritage Research Associates in Eugene, for the U.S. Army Corps of Engineers said about a half-dozen village and fishing-camp sites on or at the edge of the landslide deposits had been studied in previous years.

Radiocarbon dates from drowned trees reported in 1958 indicated that the landslide occurred between A.D. 1250 and 1280. Minor compared radiocarbon dates of wood samples taken in 1978 from within and below landslide deposits with radiocarbon dates obtained from archaeological sites in the landslide and flood area. He determined that the landslide took place about A.D. 1100 and that the earliest occupation of a village on the site occurred about 100 years later.

The date puzzled Pringle and Schuster, however, because they thought the submerged trees visible in the Columbia until the 1930s would have rotted away had they been that old. "There should have been nothing left of those trees if they were 800 or 900 years old," Pringle said. "And that kept bugging me."

Pursuing the old tree Schuster recalled that a Douglas fir buried by the landslide had been recovered during the building of the second powerhouse in 1978. Pringle tracked down the tree — which died when it was about 140 years old — in a storage area at the Columbia Gorge Interpretive Center near Stevenson.

They had radiocarbon tests conducted on two small segments of the tree, one about 120 annual growth rings from the bark and the other 20 rings from the bark. The deeper segment had a radiocarbon date of 410 years, and the portion closer to the bark had a date of 360 years. Both dates have a margin of error of 80 years.

Pringle is hoping a tree-ring study will help pinpoint a more precise date. "I'm confident we'll nail it with more work," he said. "We are exploring more options and may try to get a higher precision radiocarbon date to help us narrow the gap a bit.

Bourdeau said the new dates pose a challenge for archaeologists who have been using the older dates in studying the effects of the landslide and flood downstream. "You have to start over on your analysis if indeed the younger dates turn out to be right.

"I'm pushing to get the archaeologists and the geologists out in the field together to look at the same things," Bourdeau said. "They can learn from us, and we can learn from them. Landslides in the Northwest have become a big topic in the past few years, and this event can teach us a lot about what effects it had on people, the geology, the river and the wildlife."