Cascadia's Fault

by Jerry Thompson

  “I went over to the post office and sent Tom Heaton a postcard,” he laughed. “I thought, ah—he’s the one person in the universe who’d be especially interested in this result.” Instead of the uplifted beach terraces Atwater had expected to find, instead of confirming the Ando and Balazs hypothesis, here he was, mud spattered and dripping, with evidence that Heaton was right.

  In April and May 1986, Atwater took day trips from Seattle to other spots on Washington’s outer coast. “I visited each of the four big estuaries in southern Washington,” he explained. “Copalis, the next one to the south, Grays Harbor, Willapa Bay, and finally the Columbia River down at the Oregon border. And each of these streams had the same signature of abrupt lowering of land and marshes and forests that had been at or above high-tide level, then got abruptly dropped down.”

  During the summer of 1986 Atwater and two co-workers uncovered evidence of at least six different events—presumably six different earthquakes—that had each caused about three feet or so of down-drop. The distant geographic spacing along the Washington shore could be evidence that the quakes were big. If the coastline had slumped in river mouths and bays that were many miles apart, the quakes must have been big. But it would take further digging along the Oregon coast and up the west side of Vancouver Island to say just how big.

  Were they magnitude 8s? Or magnitude 9s? It was too early to tell. Judging by what he’d found thus far in the four widely separated river estuaries, Atwater was pretty sure they were bigger than anything recorded in Washington’s written history. To be on the safe side, he still exercised the normal scientific caution with careful wording in what would soon be considered a breakthrough research paper, published in Science on May 22, 1987.

  “Intertidal mud has buried extensive, well-vegetated lowlands in westernmost Washington at least six times in the past 7,000 years,” he declared in his opening line. “Anomalous sheets of sand atop at least three of the buried lowlands suggest that tsunamis resulted from the same events that caused the subsidence. These events may have been great earthquakes from the subduction zone between the Juan de Fuca and North America plates.”

  May have been ... Until other scientists read the paper, studied the data, and agreed that Atwater’s conclusions were valid, these new discoveries were still not established or accepted as facts beyond a reasonable doubt. The paper would nevertheless create quite a stir. With the possible exception of those turbidite samples recovered by Griggs and Kulm from offshore landslides that also may have been triggered by temblors (a hypothesis still doubted at the time), Atwater’s sunken peat layers were considered the first direct evidence of large subduction earthquakes in the Pacific Northwest. What happened in Alaska and Chile had happened here too. And would probably happen again.

  CHAPTER 12

  Cedars, Peat, and Turbidites: A Tipping Point at Monmouth

  After more than ten years of polite bickering, Bob Yeats brought them all together—the convinced, the doubters, and the fence sitters—for a brainstorming session to consider whether or not there is a major earthquake hazard in the Pacific Northwest. Yeats, who today is an emeritus professor at Oregon State University, had been lured away from the lucrative trenches of economic geology—where he had worked in the 1950s as an “exploitation engineer” and senior staff geologist for the Shell Oil company in Los Angeles—to the dark mysteries of subduction zones and seismic hazard analysis.

  Newly arrived in Oregon, a state with remarkably little in the way of major earthquake history, he counted himself among the fence sitters when the debate about Cascadia’s fault began to heat up in the late 1970s. He even joked about it in a book he wrote called Living with Earthquakes in the Pacific Northwest. In the introduction he shared his skepticism with others who figured giant temblors were a California affliction. “That was certainly my own view in 1977, when I moved to Corvallis, Oregon, even though I had been studying earthquakes for many years—in California, of course. My neighbor said, ‘Earthquakes? Bob, you gotta be kidding!’”

  Yeats, however, was in the audience at the AGU convention in 1983 when John Adams delivered his paper about the eastward tilting of the Coast Range mountains. He knew that Adams was pretty convinced the landscape was being deformed by active subduction of the Juan de Fuca plate. The releveling of the highway survey markers proved that the entire block of coastal mountains was tilting and Adams took this to mean that strain was building up.

  Adams had also mentioned to anyone who cared to listen his fascination with the turbidite cores found off the Oregon coast in 1971 by Gary Griggs and Hans Nelson, two graduate students working under the direction of OSU professor LaVerne Kulm. Yeats knew Kulm and his students personally, of course, and was well aware of their discovery. Now, here was this young Adams fellow from Cornell, recently transplanted to the Geological Survey of Canada, suggesting the cores might be evidence of very large prehistoric quakes.

  Yeats’ reactions were doubt and caution. “As a student of natural disasters, I worry about needlessly alarming the public,” he wrote in a book aimed primarily at the general reader. “What would be the reaction of people in major cities like Seattle, Tacoma, and Portland to such bad news? ‘Cool it, John,’ I said. Good scientist that he is, John Adams ignored my advice and published his results anyway.”

  Not surprisingly Adams’ new information made more or less zero impact on the public at large because it was read mostly by other scientists. The media still had not picked up on the story. That was about to change.

  Yeats was concerned enough about the implications of a growing body of evidence that he got together with Oregon state geologist Don Hull to organize a special seminar to be held in Monmouth in February 1987, just ahead of the regular meeting of the Oregon Academy of Sciences. The agenda featured John Adams, Tom Heaton, and Brian Atwater, as well as “skeptics who had previously advocated the idea that no earthquake hazard exists on the Cascadia Subduction Zone.”

  Somehow The Oregonian in Portland got wind of the meeting and wanted to send science writer Linda Monroe to cover the story. Yeats was reluctant because press coverage might cause the lead investigators to pull their punches. “I wanted the scientists to be completely candid, not worrying about a front-page doomsday quote in a major newspaper,” he explained. Monroe asked him to trust her and he did, so the conference went ahead as planned.

  Yeats described the atmosphere as electric going in to the meeting, yet in the end ironically, and perhaps surprisingly, hardly any sparks flew. The presentations were so solid the doubters were either convinced or decided to keep their thoughts to themselves. “There was no argument,” Yeats wrote, “no controversy! Most of the scientists at the meeting were so impressed with the results presented by Adams, Heaton, and Atwater that the no-earthquake opposition retreated to the sidelines. The meeting marked a paradigm change, a fundamental change in our thinking about earthquakes in the Northwest.”

  Linda Monroe came away with a scoop. Readers of The Oregonian were informed that, in Yeats’ words, “Oregon, as well as the rest of the Pacific Northwest, is indeed Earthquake Country! None of us felt as safe after that day as we thought we had been the day before.” Oregon had joined the club with California, Alaska, Chile, and the rest.

  Before leaving Monmouth, John Adams asked if he could take a first-hand look at those turbidite samples in the core lab on the Oregon State campus in nearby Corvallis. Up to that point he had only read the published reports. What he found when he studied the actual mud cores and data logs apparently strengthened his conviction that a very large series of seismic shocks was the only logical explanation for so many landslides happening simultaneously so far apart along the continental margin. He returned to Ottawa and went to work on a new paper that would put the turbidites front and center as physical evidence of past ruptures on Cascadia’s fault.

  Brian Atwater, meantime, received word in March 1987 that the final draft of his buried-peat story had been accepted at Science. After his talk at Mo
nmouth, excitement about what he’d found at Neah Bay and points south quickly spread. Gary Carver of Humboldt State, who had come to the conference to report his evidence of active thrust faults along the coast of northern California, was one of those impressed by Atwater’s presentation at Monmouth.

  “I saw it for the first time and it made sense,” Carver told me. “So when I came home from that meeting, I flew into the airport there at Humboldt and instead of home, I drove out to the Mad River slough on the north end of Humboldt Bay and walked. The tide was out and I walked onto the tide flat and leaned over the tide channel and reached down there and scraped the bank with my hand and saw the buried peats that were identical to the ones that Brian had been working on. My first thought was—oh, this is part of the subduction zone! That was an awakening moment.” Carver chuckled.

  He and colleague Bud Burke quickly found and mapped more of the same: layers of peat buried under gray bay mud, some with layers of tsunami sand and the same kinds of buried stumps and forest debris that Atwater had discovered up north. Seven buried marshes would soon be found at Netarts Bay, along Oregon’s north coast, by oceanographer Curt Peterson and geologist Mark Darienzo. As many as eight more would soon be found in and around Coos Bay in southern Oregon by Alan Nelson of the USGS and his colleagues. Up in Canada evidence for ten possible tsunamis would be found at the head of the Alberni Inlet on Vancouver Island by John Clague of Simon Fraser University and Peter Bobrowsky of the British Columbia Geological Survey.

  While Atwater’s 1987 buried-peat paper was still being readied for publication, Tom Heaton saw a preliminary draft and decided to cite the breakthrough in a review article he and Stephen Hartzell, of the USGS, were writing to underline the similarities between Cascadia and other deadly subduction zones. The work begun with Hiroo Kanamori continued with an update in Science that in turn got picked up by Walter Sullivan of the New York Times.

  Sullivan’s distillation of this carefully worded warning from one of America’s top research labs produced a story with the power to shock any who paid even scant attention. For probably the first time in a nationwide mass-circulation newspaper, the threat posed by Cascadia’s fault was given the same kind of serious and sobering treatment as the San Andreas. The opening line was among the least inflammatory yet nonetheless cautionary proclamations written about seismic mayhem.

  “Analysis of the geology along the coasts of Washington and Oregon has raised the possibility of an earthquake there as severe as any recorded elsewhere in this century,” Sullivan wrote. The meat of Heaton and Hartzell’s work was a detailed and specific comparison between Cascadia and the subduction zones of southern Chile, Colombia, and southwestern Japan, which had “repeatedly experienced severe earthquakes.” The effects of ruptures like these on the cities of Portland and Seattle would be “difficult to predict, since no modern city has ever experienced such shaking.”

  In other words, the high impact and long duration of shaking felt in Chile’s 1960 magnitude 9.5 disaster has never happened to a large city with a forest of tall buildings. None of the San Andreas temblors has ever hit magnitude 9; San Francisco had only a few tall buildings in 1906; the high-rise core of downtown Los Angeles has never been shaken by a force as strong as a megathrust subduction quake. Even those devastating shockwaves in Japan occurred before most of the high-rise skyline of Tokyo was built. Now two respected seismologists from Caltech and the USGS were warning that Cascadia’s fault might do a kind of urban damage never before seen in the modern world.

  Sullivan decided to underscore the point that Cascadia could be every bit the menace of San Andreas and then some: “One cause for concern, the authors wrote, is a tendency of earthquake tremors from a descending plate to be far more damaging, at distances beyond 30 miles [50 km], than those from horizontal slippage such as that along California faults. Furthermore, they say, oscillation rates are of a nature especially damaging to large buildings.”

  Although Sullivan did not mention Mexico City as proof, the images of collapsed apartment blocks certainly flashed into my mind when I read the story. Mexico City may not have been thought of as particularly modern, but its high-rise towers were certainly the focus of harmonic amplification and extreme damage caused by the long wavelength of shockwaves traveling 190 miles (300 km) inland from the 1985 rupture. What Heaton and Hartzell were telling us was that we should imagine the same kind of amplification happening to younger cities with dense clusters of tall buildings that have never been tested in a subduction event—namely Vancouver, Seattle, and Portland.

  When the eastward motion of Cascadia’s sea floor is added to the westward movement of the North America plate, the “convergence rate is about 13 feet [4 m] per century,” Sullivan continued. “According to Dr. Heaton and Dr. Hartzell, the key question is whether the sea floor has been smoothly slipping under the continent or is ‘locked’ and accumulating strain. If it has been storing elastic energy for a long time, a sequence of several great earthquakes or a single giant one, comparable to that in Chile, would be necessary to relieve the tension.”

  Even though Cascadia has not produced a major jolt in the Northwest since it was “permanently settled by Europeans in about 1810,” Sullivan reported, “there are indications of periodic seafloor landslides and coastal subsidences that could have been triggered by such events in the more distant past.” So finally, there it was in print for a general audience to digest: reference in the New York Times to the turbidite landslide cores from Griggs and Kulm via John Adams, and to the sunken coastal meadows that Brian Atwater had found. The so-called smoking gun evidence of Cascadia’s violent past was now a matter of very public record.

  For Brian Atwater the next step was to ask two questions a lot of people were asking him: how big and how often? To find the answers he packed his kit and returned to the coast in the summer of 1987 to conduct a systematic survey by canoe of those three southern Washington estuaries: Copalis River, Grays Harbor, and Willapa Bay. He paddled miles and miles of shoreline and hiked through marsh, muck, and greasy river mud until persistence, and serendipity, paid off again.

  In his initial reconnaissance at the Copalis River, he had walked in, venturing only a short distance from the road. “I missed the ghost forest,” he smiled, pointing across the lagoon toward the grove of weathered hulks bathed in sea mist and drifting fog. “It continues on upstream for another mile or two. It’s spectacular all the way up.”

  What made the dead trees important was the possibility that they could help pinpoint the year and season of the earthquake that presumably had killed them. The first scientist to try this tactic was David Yamaguchi, who had earned a PhD in forestry from the University of Washington and was working on a project for the USGS to use tree-ring dating as a way of figuring out when Mount St. Helens had erupted prior to 1980. He offered to help Atwater by trying this same technique to date the coastal earthquakes.

  In May 1987 they took their first trip together to Willapa Bay. Atwater showed Yamaguchi the stumps of Sitka spruce, the main arboreal victims of great Cascadia ruptures. Yamaguchi chainsawed a few samples, but they didn’t look very promising because tree-ring scientists usually prefer to sample from tree trunks—not stumps. Unfortunately, the spruce trunks had all but rotted away.

  The great moment of good fortune came a few months later when they worked their way through the mist and saw for the first time the weather-beaten and moss-draped trunks of western red cedar—what would become known as the Ghost Forest of the Copalis River. “When Dave and I first started working together, we didn’t know that big forests of dead cedar trees existed,” Atwater told me. “Red cedar is more durable. The trunks are still here, standing dead three hundred years after they were killed.”

  They figured similar trunks could probably be found along other tidal streams as well, and the more evidence, the better. Yamaguchi came up with the clever idea of placing ads in coastal newspapers, asking local residents if they knew about any more of these ancient b
eauties. And they did. Cards and letters arrived pointing them toward ghost forests near Grays Harbor, Willapa Bay, and along the Columbia River, a stretch of the Washington coast nearly sixty miles (100 km) long.

  Did they all die during the same year and season? They should have if that entire segment of the coast had broken all at once in a single earthquake. Or did they instead die in different years at difference places as a result of a series of smaller earthquakes? Timing was everything.

  Yamaguchi’s first effort to establish a time of death for the spruce stumps had failed because, with all the rot, there were not enough rings left to count. But working with red cedars would be different. Step one of the ring-matching process involved finding a group of same-age trees that were at least as old as the ghost forest—and still alive—to establish a baseline growth pattern up to the current date. Wide rings that grow during good years with plenty of rain, for example, should be found in all the trees in the area. The same with narrow rings that grow in years of drought or fires or other kinds of trauma. The patterns should all match year by year, almost like fingerprints of the local climate.

  Once this ring pattern was established, Yamaguchi would be able to work backward from the current year’s growth ring and assign specific dates to individual rings in the past to determine in which year the ghost forest cedars died. Later in the summer of 1987 he and Atwater found the live trees they needed for comparison. At the time Weyerhaeuser was harvesting the fringes of a stand of old-growth red cedars that had witnessed—and survived—the great earthquake by inhabiting a hillside above tides, on an island in the middle of Willapa Bay.

  “It’s a shame that these trees were being cut,” Yamaguchi commented. “They’re beautiful trees. But we recognized that that was a place where we could gather modern reference samples.”