Cascadia's Fault

by Jerry Thompson

  The significant point was that the two plates had to be locked together for this kind of strain to accumulate. Savage and his team were saying as politely as they could that Ando and Balazs must be wrong. “The implication is clearly that the Washington and Vancouver Island coasts are subject to great, shallow, thrust earthquakes,” they wrote in June 1981.

  A year later, when John Adams was drafting his conclusions for the paper on Coast Range tilting based on highway survey data, he spoke to Savage on the phone and heard his idea about why the mountains might be tilting to the east. In Japan and several other places, subducting plates were bending the outer coast downward. Why would the Pacific Northwest coast be different? Why would it tilt eastward? Savage had a hunch that if the two tectonic plates were locked at a point only slightly inland from the coast—if the point of impact was nearly head on and close to the beach instead of deeper down and farther inland, beneath North America—then maybe the Juan de Fuca plate would push the mountains and beaches upward (like a crumpled fender) instead of dragging the coast down and curling it under, like in Japan.

  Adams latched on to this as a possible explanation for the eastward tilting and thought he’d made a pretty convincing case. He was not completely successful, however, in exorcising the demon of aseismic subduction. At the request of his more senior coauthor, Robert Reilinger, he wrote a concluding paragraph that equivocated enough to dull the edge considerably. Taken as a whole, the tilting data and the lack of large quakes “suggests that subduction is occurring aseismically, although alternative interpretations are possible,” they wrote cautiously. Thinking about it more than thirty years later, Adams had to laugh. “That phrase was largely put in to—shall we say—to take the controversy out of the paper, to make sure it got through the peer-review process.”

  But Adams was already hard at work on another study that would pull fewer punches. He was like a dog with a bone. He knew the mountains were tilting, he knew Puget Sound was getting squeezed, and he intended to follow up on those beaches that had been shoved up into marine terraces along the Oregon coast. He knew from his New Zealand studies that it took hundreds of years to build up enough strain for a giant subduction quake. And he knew about those turbidite mud cores from offshore landslides that also were roughly five hundred years apart. Were the deep-sea landslides physical proof of Cascadia’s violent past? He was determined to find out.

  CHAPTER 10

  The Whoops Factor: Cascadia’s True Nature Revealed

  While John Adams watched the Coast Range tilt and Jim Savage tracked the squeezing together of mountain peaks in Puget Sound, Mike Schmidt was learning about a new technology that could make the measurements far more precise. He would eventually join a team of researchers on Vancouver Island, where the distance between several mountain peaks was being resurveyed to find out whether continental drift was shoving them closer together as well.

  My first impression of Schmidt was that he’d rather climb a big rock pile than stand there and look at it. He’s a bearded bear of a guy who seems to have chosen the right career. In 1992 he led a team of Canadian scientists to the top of Mount Logan, Canada’s highest peak and its fastest rising mountain. Fast in geologic terms, it grows by several fractions of an inch each year.

  As a mountain climber, geophysicist, and surveying engineer, Schmidt wanted to establish new geodetic markers near the summit and try out some brand new and allegedly portable GPS technology, which was still in the experimental stage at the time, to trace the peak’s constant movement. Logan, which occupies a big chunk of the southwestern corner of the Yukon, is poking up and creeping horizontally for the same reason that mountains in Puget Sound near Seattle are getting squeezed together. In the case of Logan, the floor of the Pacific Ocean is jamming itself underneath North America from the Gulf of Alaska.

  This is the same tectonic force that caused the 1964 Alaska earthquake. The Pacific plate is thrusting the entire St. Elias Range a tiny bit higher and shoving it slowly inland at the same time. Because 1992 was Canada’s 125th birthday and the 150th anniversary of the Geological Survey, Schmidt came up with the idea of putting together an expedition to climb the mountain and settle a long-standing debate about how high Mount Logan really was. With sponsorship from the Royal Canadian Geographical Society, he and his team did exactly that.

  They couldn’t simply fly to the top because the summit was too high and the air too thin for ordinary helicopters. Plus the researchers themselves would need time to acclimatize to the altitude before starting the hardest part of their work. So instead of an easy ride in a chopper, they made nine trips over three days in a single-engine, ski-equipped Helio Courier airplane to airlift the fifteen members and all their gear to a base camp on the Quintino Sella Glacier, 9,055 feet (2,760 m) up the mountain. From there they had only another 10,495 feet (3,199 m) to go, Schmidt told me, slogging steeply uphill the hard way. It was the only “relatively safe” option.

  For thirty days starting in early May they skied and climbed and packed loads of food, tents, sleeping bags, clothing, climbing gear, and heavy cases of the new high-tech survey equipment—satellite receivers, antennas, and heavy batteries—steadily upward through spectacular spring sunshine and howling late-winter snowstorms that nearly forced them back. Being so close to the Gulf of Alaska meant that nasty weather could blast across the slopes with almost no warning. And it did, several times.

  By Schmidt’s account, though, the expedition was a complete success. They nailed a new brass survey marker at 18,044 feet (5,500 m), on the edge of the Logan plateau; when they finally reached the summit, the portable GPS system worked perfectly and the official height of the mountain was confirmed at 19,550 feet (5,959 m). But they also proved under extremely harsh field conditions that this new, extremely precise technology could be used to help figure out what was really happening along Cascadia’s fault.

  Measuring mountains and the drift of continents using satellites, sophisticated antennas, and software to track the warping and bending and horizontal migration of land caused by plate tectonics would become the focus of Mike Schmidt’s working life. At the Pacific Geoscience Centre on southern Vancouver Island, he helped develop the technology and methodology for tracking the minute and ongoing deformation of the earth’s crust. But before GPS, there were lasers and Geodolites and each step along the way was a huge improvement.

  I heard the story of how it all began from Schmidt’s senior colleague, Herb Dragert, who was there at PGC when the study of migrating mountains began in 1976. Apparently a burning desire to prove that Tuzo Wilson at the University of Toronto was right all along about plate tectonics had been Dragert’s personal motivation. As an eager young student, his imagination had been fired by Wilson and those big ideas about drifting continents. “He kind of said to us, ‘Okay—we do get plate convergence on the west coast and we should be able to measure the actual motions of the earth’s surface,’” Dragert recalled with gusto. “These mountains should be squeezed!” Which is precisely what he, Schmidt, and a team of others from the Geological Survey would try to confirm.

  On Vancouver Island the idea was to locate the original stone cairns and brass markers on mountain tops that had been surveyed back in the late 1930s to remeasure the distances and the angles between the peaks and thus find out whether—or indeed, how much—they had moved by the late ’70s. Decades later I wanted to see how the work had been done. In 2007 I needed a way to illustrate the process for a television documentary, so Schmidt was going to show me by doing it again.

  Thus I found myself in a helicopter once more, flying this time toward the lumpy shoulder of Mount Landalt, a few miles north of Lake Cowichan on southern Vancouver Island. The skids of the JetRanger touched down gently on a bed of gray lichens and green moss sprinkled with tiny, bright fuchsia-colored flowers. Schmidt pulled out the first of two steel cases, each a little larger than your standard, full-size suitcase. Then he carried a set of heavy-duty tripod legs to the summit and set t
hem up directly above the control point, the brass marker that had been established by the original survey crew back in 1937.

  The instrument he hauled out of its foam-padded metal case was about twice the size of that old breadbox your grandmother used to have. It was a Rangemaster III, with a bright orange housing, knurled brass knobs, a black instrument panel, and a self-contained digital computer that flashed the distance calculation to an LED readout. State of the art in 1976, it still appeared in perfect working order after many years in a storage locker. The helicopter pilot then flew across to Mount Whymper, the next nearest mountain in the hazy distance, where he landed and set up a reflector box on a tripod directly above the survey marker on that peak.

  Back on Mount Landalt, Schmidt peered through his viewfinder on the Rangemaster and used a portable radio to call the pilot, who then tilted the bank of mirrors on Mount Whymper until Schmidt got a return signal—a reflection of the laser beam—on Mount Landalt. Scintillating shafts of cherry-red light bounced off a dozen mirrored prisms in the reflector box. The Rangemaster then performed its magic: a quick calculation of the time it had taken the beam to shoot across the valley from one peak to the other and bounce back. With laser gear like this they could measure the distance between peaks up to twenty-five miles (40 km) apart and be accurate within fractions of an inch. A significant improvement, but there was still another problem to solve.

  With nothing more sophisticated than pack horses, climbing gear, and transits (telescopic instruments mounted on tripods for measuring precise, horizontal angles between objects that are far away) the British Columbia survey crew back in 1937 could accurately plot the geometry between a series of peaks. But computing the exact distances between mountains was extremely difficult. In those days most surveyors still used sixty-six-foot chains (eighty chains to the mile, or about fifty to the kilometer) to establish a baseline measurement. If you know the angles and the length of one side of a triangle (the baseline), you can calculate the lengths of the other two sides. But because of jagged mountain terrain and dense bush, the distances calculated and printed on the old maps of Vancouver Island were too imprecise to work as reference points in a modern-day study of minute tectonic creep along a fault. The new laser equipment that became available in the 1970s changed everything.

  Dragert and Schmidt and their team used the laser Rangemaster to redraw the original triangles between the peaks and see whether they had changed. With the Rangemaster’s new measurements they knew exactly how long each leg of the triangle was and could then calculate the precise angles between the peaks. Then they compared the new triangles to the old ones from 1937. When they did—bingo!—they saw that the angles had changed, which meant that at least some of the peaks had moved since 1937. “We proved that the margin was deforming,” said Dragert. “The mountains were indeed being squeezed landward,” ever closer to the continental mainland.

  He pointed to a specific example on the old map, a triangle of dark lines drawn by the original surveyors. The new laser triangle clearly did not match the old one. It was bent out of shape because the mountain closest to the west coast—Mount Grey, a 4,570-foot (1,390 m) peak about halfway up the Alberni Inlet—had been shoved eastward nearly eight inches (20 cm) in less than forty years. Not a huge amount, by the sound of it, but imagine the entire island coastline, hundreds of miles’ worth of mountain rock, being pushed horizontally like that. When I thought of another few inches of horizontal movement each year—for 310 years since the last Cascadia earthquake—the total amount of accumulated strain built up along the fault was mind-boggling.

  “And that was totally consistent with our expectation,” said Dragert, “that if the subduction zone is locked, we have to see deformation.” He poked his finger emphatically at the map again. “And indeed we saw the deformation which was the final nail saying, ‘Look—this is not slipping smoothly. This subduction zone is locked!’” One might think that should have been the end of the aseismic subduction idea. But it was not.

  For those who knew what the data were saying, the late 1970s and early’80s must have been an exciting time. To Herb Dragert’s great satisfaction he had proven his mentor, Tuzo Wilson, right after all. The general public, however, knew almost nothing about it because the latest evidence and the debate it spawned were confined primarily to academic journals, some so specialized that only a handful knew where the cutting edge of this new science could be found.

  Even a lot of scientists were unaware of the variety, the volume, and the geographical extent of the data that were piling up. At the annual meeting of the American Geophysical Union in December 1981, more than a few heads turned when John Adams spoke about mountains of the Coast Range tilting to the east. Like Herb Dragert and Jim Savage, Adams was convinced this could only be happening if the two plates were locked together. His presentation at the AGU had an impact on several other scientists.

  “I guess something of it caught the eye of the people in San Francisco who were doing the WPPSS project,” Adams told me. When he uttered the unfortunate acronym, it sounded like whoops, which is the expression almost everyone would come to see as appropriate for the infamous project eventually. The Washington Public Power Supply System was a megaproject involving construction of five nuclear power plants, two of which were to be built in the small town of Satsop, west of Olympia along the mountain highway leading out to Grays Harbor on Washington’s west coast.

  Even though the nuclear plant at Humboldt Bay in California was already known to be in trouble because of crustal fractures caused by Cascadia’s tectonic motion, another pair of reactors were going to be installed pretty much on top of the same subduction zone in Washington State. As luck would have it, several of the geo-engineers from Woodward-Clyde Consultants, the group doing the seismic risk analysis for WPPSS (the same company that had done the assessment of Humboldt Bay), happened to be in the AGU convention hall that December.

  As Adams remembered it, “One of them basically said, ‘We’re interested in what you’re doing. Would you like to come down and talk to us?’” The Woodward-Clyde geologist who extended the invitation was David Schwartz. He had met Adams a few years earlier and considered him “one of the more interesting guys in paleoseismology,” with a new take on active faults.

  The presentation was “for our enlightenment,” Schwartz explained. He wanted to stimulate discussion within the consulting team about “what was going on up there” in Washington State in terms of the seismic risk factors that might affect the WPPSS project. Adams arrived at the San Francisco offices of Woodward-Clyde a few weeks later and quickly glanced at a preamble document prepared for the meeting. “I guess you could say it was open-minded,” said Adams. “It certainly wasn’t coming down very strongly one way or the other.” Meaning Schwartz and his colleagues appeared to be scientifically neutral when it came to the question of Cascadia’s fault.

  The day’s agenda included two presentations. Masataka Ando went first and laid out the argument that most geologists at the time believed to be true. As David Schwartz distilled it, “You knew the plate was going down but the question was—how was it going down? Was it going down aseismically—in which case you can make the argument you aren’t going to have large earthquakes? Or was it locked? That was sort of the crux of the issue.” Ando, of course, thought it was aseismic.

  John Adams delivered the second presentation, his summary of the Coast Range tilting data. He saved the kicker for the end, wrapping up his talk with an overhead slide of the Griggs and Kulm turbidite landslide data from the Oregon coast. Huge offshore mudslides in deep-sea channels from river systems hundreds of miles apart could have been caused only by very large earthquakes, in his view.

  He told the team from Woodward-Clyde, “This is symptomatic of an active subduction zone.” He then tried to encourage follow-up research by suggesting, “The earthquakes don’t happen more often than every four hundred years or so, but here’s the sort of evidence you could use to tie it down.” Me
aning those mud cores.

  David Schwartz felt himself being swayed by the data. “From my perspective even back then, it was hard to get all of the secondary deformation if things were just sliding aseismically,” he said. The folding and fracturing of rocks along the shore, the compression in Puget Sound, the tilting of mountains—how could all that deformation happen on the surface if the oceanic plate were sliding smoothly underneath?

  Schwartz explained that his company had been hired to perform the FSAR—the Final Safety Analysis Report—on the Satsop power plants, which would soon be presented to the Nuclear Regulatory Commission. Their preliminary draft back in 1974 had been written in a completely different atmosphere, when “the Cascadia Subduction Zone was never a consideration.” At that time the official U.S. seismic hazard maps “did not identify it as an earthquake source,” explained Schwartz.

  The science was changing quickly, however, and so was the political environment surrounding nuclear power plants. In the next draft of the FSAR, circulated internally to the consulting board at WPPSS, a new, more cautionary tone had replaced the earlier optimism. “In that document we were definitely opening the possibility that the subduction zone should be considered as a source of strong ground motion,” he told me. This was a very different view of the situation and, not surprisingly, it didn’t go over well with the WPPSS officials. They were “not delighted by this turn of events,” Schwartz remarked dryly, and things slid downhill from there.

  A different firm of consulting engineers had been hired to design the reactor, and they clearly disagreed with this alarmist talk from Woodward-Clyde of potential seismic shocks—Alaska-size jumbo quakes, backed up by a handful of mostly theoretical papers with equivocal conclusions and very few hard facts—which threatened to wreck their chances of building a pair of billion-dollar atomic power stations. They insisted on a face-to-face meeting to put things back into some kind of perspective.