Cascadia's Fault (23 page)

After the conference at Monmouth, conversations between Brian Atwater and Gary Carver moved into new territory. While others were slowly coming to accept the notion that Cascadia's fault was an active subduction zone, Atwater and Carver and a handful of others had jumped ahead to the next big question. Assuming they had proven the existence of past quakes, how could they find out whether the entire subduction zone had ripped loose all at once—as opposed to rupturing in a series of smaller segments, releasing only part of the accumulated strain each time?

“There were two competing hypotheses,” Carver told me. “We were arguing about what
kind
of earthquake it was. Not
whether
it happened.” Looking at how far apart the sunken marshes and forest floors were geographically—all the way from Vancouver Island to Cape Mendocino—and sorting through their calendar of radiocarbon dates, they began to speculate about whether or not Cascadia always ruptured from end to end. Was it a single, Alaska-size event each time, or did some segments of the down-going slab—perhaps the southernmost Gorda plate off Humboldt Bay, for example—rupture separately in smaller events? Meaning magnitude 8 disasters instead of magnitude 9 catastrophes.

“There was the Brian Atwater ‘apocalyptic model,' which was a magnitude 9 which broke the whole subduction zone,” Carver mocked, “and then I had coined the term ‘decades of terror' for a series of earthquakes that occurred very close together in time—so when you radiocarbon dated them, you couldn't tell them apart.” When not trading quips with Carver, Atwater insisted he was “agnostic” about magnitude 9 and that—all kidding aside—they needed some way to prove
or disprove
the worst scenario.

The decay of carbon atoms gave them dates that were accurate to within a few decades, but that wasn't good enough to tell whether all the marshes and estuaries had been buried at the same time by the same great earthquake and wave. Tree-ring dating would eventually help narrow the timeline, but now there was another wrinkle. The build-up
of stress deformation along the coast did not seem to be occurring at the same rate everywhere.

A newer, more accurate releveling survey had been done along U.S. Highway 101 from Crescent City, California, all the way up the coast to Neah Bay, Washington. When Clifton Mitchell and Ray Weldon of the University of Oregon took a closer look at the numbers, they noticed that over the past fifty years the southern end of Oregon and the northwestern tip of Washington had been rising about an inch or more every decade.

The complication was that the area around Newport, about halfway up the Oregon coast, and around Grays Harbor, halfway up the Washington shore, did not appear to be rising at all. To Mitchell and Weldon this suggested that some points along the subduction zone might have
asperities
—rough spots where the two plates tended to hang up or get temporarily locked together. Some parts of the zone—like Newport and Grays Harbor—could be moving without seismic strain build-up while others were locked and loaded for a big rupture. Which sounded like evidence in favor of Carver's “decades of terror” scenario.

Atwater still wasn't sure. He looked at proof of sunken marshes from Vancouver Island all the way down to southern Oregon and possibly including Carver's own tsunami sands near Humboldt Bay in California and thought the magnitude 9 scenario was still viable. He knew he'd have to find some way with tree rings—or whatever—to narrow down the dates. If all those places sank at the same time, it must have been a magnitude 9 or larger and exact dates would prove it.

Carver, on the other hand, could point to the big Nankai Subduction Zone off Japan, which had broken in two magnitude 8 events (1944 and 1946) during and after World War II. If they had been dated with radiocarbon there would have been no way to tell they were separate quakes. Radiocarbon might get you within a decade, but certainly not within two years. Dating the ghost forest's time of death became all the more important.

Here along the west coast, Carver could imagine a sequence similar to the one in Japan. The Gorda plate could break loose first, then the main Juan de Fuca segment a few years later, with the Explorer and Winona segments up north of Vancouver Island after that. Whole decades of terror instead of one big event.

Garry Rogers at the Pacific Geoscience Centre had already written about the possibility of separate ruptures with differing magnitudes. Roy Hyndman and Kelin Wang, also at PGC, were working on a new paper that used temperature variations in the crust to calculate how much of the subduction zone might be stuck and which parts were sliding smoothly due to higher temperatures (near the melting point of rock) deep underground. Their preliminary conclusion was in favor of the magnitude 9 scenario. But soon a new line of offshore evidence from the team at Oregon State seemed to tilt again in favor of the segmented rupture pattern.

With data and new ideas coming together from several directions at once, it was exactly the kind of movable feast that scientists love to sink their teeth into.

 

In the summer of 1989 a new batch of data from Canada was released, adding weight to what Jim Savage had said earlier about the big squeeze in Puget Sound. Herb Dragert and Mike Lisowski had finally amassed enough evidence from laser beams bouncing off mirrors on the mountain peaks of Vancouver Island to release the numbers at a meeting of the International Association of Geodesy held in Edinburgh, Scotland, on August 3, 1989.

A dozen years'worth of repeated geodetic surveys, along with long-term tidal monitoring along the island coast, had revealed exactly the same kind of squeezed mountains and built-up stress that Savage had seen in the hills on either side of Seattle. Dragert's data confirmed “significant regional strain rates,” including a newly detected ridge of uplifted land from Neah Bay, at the northwest tip of the Olympic
Peninsula in Washington, to the mountains northwest of Campbell River on Vancouver Island. The line ran parallel to the subduction zone and was exactly the kind of humping-up you'd expect to see if the ocean floor were stubbed against the continental plate, a big boot kicking it eastward.

The latest seismic data showed all the normal little tremors in the overlying crust and in the down-going plate, but along the subduction interface itself—that eight-hundred-mile (1,300 km) scraping edge where the two plates actually meet deep underground—there was “a distinct absence” of even low-level seismic activity. The same kind of ominous silence that had been noticed before the quake near Adak Island three years earlier.

Herb Dragert's evidence that the subduction zone was locked and loading the kind of strain energy that could only be released in a big shock—the proof that his mentor, Tuzo Wilson, was right about tectonic plates squeezing Vancouver Island against the mainland—was now a matter of public record. The implication was, he wrote, that “a definite potential exists for a future megathrust earthquake.”

He and colleague Garry Rogers had also decided that a version of Cascadia's emerging story needed to reach a wider public, so they wrote a short, provocative article entitled, “Could a Megathrust Earthquake Strike Southwestern British Columbia?” which was published in a quarterly government magazine called
Geos.
Hoping, perhaps, that the headline might capture the attention of editors and journalists in the mainstream media, they borrowed the language of a crime novel, talking of “smoking guns” and “lethal weapons.”

Piece by piece they laid out the clues unearthed by scientists on both sides of the border. In answering their own provocative questions, they did their best to remain scrupulously neutral while leading readers to the obvious conclusion: “We see no evidence to preclude the occurrence of a megathrust earthquake . . . Geological evidence suggests that large earthquakes may have occurred.” On one point, however, they did
not equivocate; they stated flatly that “crustal deformation is currently taking place.”

As for predicting the next rupture, they wrote, “Unfortunately, our present data are too sparse to provide information on the likely timing of such an earthquake in this region.” In the closing paragraphs they got serious about the significance of their mystery tale. The consequences of a great subduction jolt and the tsunami it generated would be devastating because so many communities would be affected at the same time. “If such an earthquake occurred, it would likely be the largest economic and social catastrophe due to nature ever to affect Canada,” Dragert and Rogers wrote. They carefully noted that much of the evidence was still circumstantial. “As with the pursuit of the smoking gun in the mystery novel, we still need more evidence to prove that the gun has fired or is likely to fire again. But our suspicions are mounting.”

The
Geos
article did have the desired effect. It was picked up and rewritten by several newspapers in Canada, making Dragert and Rogers briefly famous as young, intrepid scientific sleuths. Their moment in the limelight was no doubt less painful than the cluster-attack on Gary Carver by reporters at the convention in Phoenix, though the message was essentially the same.

 

On October 17, 1989, I was back in Toronto, jetlagged from Europe and completely immersed in interview transcripts, trying to write a script for a documentary on Poland's new Solidarity government. I had a quick dinner and went to bed early, feeling exhausted. I had just drifted off to sleep when the phone beside the bed rang loudly and jolted me into a semi-conscious stupor. It was Sally Reardon, one of the senior producers on the desk at
The Journal,
the CBC program I was working for.

“Major earthquake in San Francisco!” she proclaimed, breathlessly.

At 5:04 p.m. Pacific time, during the warm-up for a baseball game in Candlestick Park, a segment of the San Andreas fault broke near the
coast deep under a mountain called Loma Prieta, about nine miles (15 km) northeast of the seaside resort of Santa Cruz. For about fifteen seconds the ground shook violently. The crowd in Candlestick Park knew immediately what it was and started moving toward the exits, their barely suppressed panic captured live on network television.

It was game three of the World Series, featuring, ironically, two local teams, the San Francisco Giants and the Oakland Athletics. The first major California temblor covered live on television became known briefly as the World Series quake; geologists would later officially name it the Loma Prieta earthquake. When contact with San Francisco was reestablished, the pictures were disturbing.

Freeways and bridges had collapsed, killing dozens of people. Because the quake happened at rush hour, a traffic helicopter was already in the air and over the water with its camera rolling when a car drifted helplessly in what looked like slow motion over the edge of a fifty-foot (15 m) section of the double-decker Bay Bridge that had dropped like a trap door onto the span below. Because two local teams were playing in the Series, thousands of people had either gone to the stadium or stayed downtown after work to watch the game in bars with friends. As a result, rush hour traffic was lighter than normal. The death toll could have been much higher.

In Oakland things were worse. A mile-long (1.6 km) section of the Nimitz Freeway—built on former marshland—collapsed, crushing cars on the lower level, instantly killing forty-two people and injuring many more. Some cars on the upper deck were tossed around and flipped; others were left dangling over the side as the freeway bucked and twisted during the temblor. Nearby residents and factory workers rushed onto the bridge with ladders, tools, and forklifts and began digging survivors from the rubble. The volunteer rescuers were at it round the clock for the next four days. It would take eleven years to rebuild the freeway.

High-priced houses and condos built on landfill in San Francisco's upscale Marina District shifted on their liquefied foundations. Seven
buildings collapsed outright, another sixty-three were damaged beyond repair. Water and gas mains broke, creating hellacious fires. The Embarcadero Freeway, another two-level roadway, was damaged and would have to be torn down. Golden Gate Bridge survived but would need a $75 million retrofit. A six-block stretch of downtown Santa Cruz's historic business district was reduced to rubble.

At magnitude 7.1 this was the largest and most damaging rupture on the San Andreas since 1906. All told, 63 people died that night, 3,757 were injured, and more than 38,000 needed emergency housing. The damage toll topped $6 billion. Once again the world's attention was focused on California's most famous fault.

Meantime, although hardly anyone paid them heed, a small squad of quake hunters continued digging in the mud and turning over rocks from Cape Mendocino to Vancouver Island. While San Andreas dominated the headlines, Cascadia's smoking gun—final convincing evidence of a much larger quake—would soon be found. Quietly and with little fanfare. In tight little rings of western red cedar, in the voices of Native elders who told of a dark and violent night many generations ago, and in sacks of rice destroyed in a shogun's warehouse on the far side of the Pacific.