Cascadia's Fault

by Jerry Thompson

  In the immediate aftermath of the Mexico City disaster, seismologists, marine geologists, and engineers tried to draw conclusions about the underlying cause and what it might mean for other supposedly aseismic zones around the world. Perhaps these monster quakes had happened before but researchers had not looked far enough into the past to find the evidence. Perhaps “all of recorded history” was simply too brief, in geologic terms, to see a repetition of these enormous undersea earthquakes. If it takes several centuries to build up enough stress for a quake this big, perhaps the last one happened so long ago there was nobody around to write it down.

  Some scientists thought the Michoacán zone was a “seismic gap” where strong earthquakes (in 1939, in 1973, and again in 1979) had relieved stress on either side of the main segment, but the part in the middle—the Michoacán zone itself—had remained stuck for nearly two centuries. The zone was the only part of this tectonic plate that had not snapped free of the continental crust that had drifted over it. It was a holdout—a ninety-mile (145 km) slab of sub-sea rock that was bound to rip loose sooner or later. And when it did, the amount of strain released was unprecedented.

  For half a minute that must have felt like a lifetime, 320,000 square miles (825,000 km2) of Central and North America shuddered and rumbled up and down and from side to side. More than twenty million people, some as far away as Los Angeles, Guatemala City, and Corpus Christi, Texas, felt the earthquake. Even though the rupture zone was thirty miles (50 km) offshore, south and west of Mexico City, the quake might as well have been directly underneath the downtown core.

  Like the lowest notes of an upright bass, this fractured slab of sea floor played fatal music, a throbbing rhythm that pulsed with stunning efficiency through 190 miles (300 km) of continental crust to reach the capital city. The first burst of notes lasted roughly thirty seconds—about the duration of most normal earthquakes that occur closer to the surface. But then another segment of the plate broke loose and the vibrations started again. Many towers that had survived the initial attack were too crippled to endure the second.

  As if that weren’t enough, a significant part of the center of the city was constructed upon a layer cake thirty feet (9 m) thick of sand and gravel and clay, washed into the Mexico Valley from a ring of volcanic mountains. The modern city occupies the same site as Tenochtitlan, capital of the Aztec empire, built on the shore of an ancient lake called Texcoco. After Hernán Cortés conquered the Aztecs in 1521, Spanish engineers drained the lake to make room for a larger settlement. They knew nothing at all about the risk of massive earthquakes from the sea.

  Four centuries later, eighteen million Mexicans had stacked a sprawling metropolis around and on top of this mudflat basin. When deep bass notes from the Michoacán gap hit the loosely packed soil, it throbbed and resonated. Apartment blocks on the dried-up lake moaned and swayed to the rhythm of the quake. Swinging side to side like inverted pendulums, each tower rang at its own frequency, depending on how tall it was. High-rise towers on the sands of Texcoco were in big trouble.

  Every two seconds another shockwave thundered in from the coast like a horizontal pile driver hitting the foundations of the ancient city. Concrete and steel joints were cracking, walls were pulling away from floors. Tall structures standing too close together smashed into each other at the top and tore themselves apart from the roof down. Multi-wing complexes that were not perfectly square or rectangular—those with T-shapes or with a number of sections connected to a central hub—yanked themselves apart at the joints because each wing rattled and snapped to its own beat in syncopated self-destruction.

  Anything between six and seventeen stories tall had a special problem—an internal rhythm in near-perfect synch with the earthquake itself. Shockwaves from the quake caused these midsize towers to hum like a tuning fork. This “resonant oscillation” amplified the amount of energy that pulsed through their vertical frames. With the arrival of each new wave, the foundations would be slammed to the side yet again before the rooftops had flexed back to a vertical position. Like rocks along the fault itself, the urban bricks, mortar, and steel inevitably failed.

  As stress was relieved along the fault, the continental shelf off the west coast came unstuck from the oceanic plate that had snagged it more than ten miles (16 km) below the surface. The overlying crustal plate broke free, rebounded sideways as far as eight feet (2.5 m) and then sagged. Western beaches slumped as much as thirty inches (80 cm), causing the local sea level to rise.

  Offshore, while the continental shelf was rattling loose and rebounding to the west, it also heaved upward. Roughly three thousand square miles (7,500 km2) of the shallow sea floor lifted a huge mound of salt water. When gravity broke this hump of hoisted water into elliptical waves—a train of tsunamis raced across the Pacific.

  Given the magnitude of the earthquake, the size of the tsunami was relatively small. From Manzanillo to Acapulco and Zihuatanejo, the waves measured from three to ten feet (1–3 m) but caused relatively little damage. When the tsunami hit Hilo, Hawaii, hours later the wave was less than nine inches (22 cm) high. One possible explanation is that the ocean floor had been shoved underneath the continental landmass at such a shallow angle that the volume of seawater lifted by the earthquake did not amount to much, compared to other tsunamis generated by large earthquakes.

  But in September 1985 scientists were still struggling to understand what had happened. The intricate details of tectonic motion were still pretty sketchy. One thing, however, was clear: a seemingly logical explanation for the lack of major earthquakes off Mexico’s west coast—if we haven’t seen any in all of recorded history, they must not happen here—was wrong.

  Since a great earthquake had just happened in a supposedly aseismic zone like Mexico’s, where else might the same thing occur? What about the coast of northern California, Oregon, Washington, and British Columbia, also thought to be aseismic? A major disaster like Mexico’s had not happened in the Pacific Northwest in all of recorded history either. By the same logic, if mega-quakes were likely to happen up there, surely we would have seen one by now.

  On the other hand, maybe not. Overnight it seemed anything was possible. Perhaps a megathrust quake could happen on any offshore “subduction zone.” Teams of scientists scrambled to Mexico as quickly as they could to examine this newest twist in the young science of plate tectonics. Ideas about how the world’s largest earthquakes are created would change in the months and years ahead. For some scientists, the Mexico City disaster would be a tipping point.

  As I fell asleep that night in New York, I too faced a knowledge gap. I did not know the Mexico City earthquake had become the opening chapter of a story I would soon be covering myself—a plot line I would follow for the next twenty-five years. It’s a mystery that continues to unfold like a dimestore thriller, one that probably will not end for me until my home on British Columbia’s coast has faced its own seismic nightmare.

  CHAPTER 2

  Lessons from the Rubble: A Front-Page Story

  From the parking lot my first impression of the Pacific Geoscience Centre on southern Vancouver Island was that this sprawling complex of government laboratories belonged on the cover of some architectural magazine. The atrium was covered by a steep, angular skylight of tubular white steel and a grid of glass panels. Against a fetching backdrop of West Coast cedars and ferns, the mature landscaping included a small but brilliant reflecting pool. Nice digs for such a muddy-boot kind of science.

  The downside to this idyllic locale was that it nestled cheek by jowl against the main runway at Victoria’s international airport. Walking in to PGC (as it’s known locally), I glanced at David Kaufman, the Toronto-based CBC producer working with me on a new documentary assignment to find out what scientists had learned from the Mexico City disaster. We both winced at the thought of random bursts of ruinous noise from turboprops, passenger jets, and helicopters that would no doubt make on-camera interviews here a challenge.

  We were met
at the door by seismologist Garry Rogers, who gave us the fifty-cent tour along a convoluted route past a museum-like display of earth science exhibits, presumably designed for public and school tours, to his office deep inside the complex. Once past the main public areas, the work space began to look more familiar, like a thousand other labs and universities.

  Arriving at the top of a vast open stairwell, on a landing behind a plate glass wall, I saw a rack-mounted array of fifteen seismographs, each paper drum a furrowed field of squiggly parallel lines, pen strokes that looked like several days’ worth of data—every little twitch of the ground under this mountainous corner of the world. Pinned to corkboards, stapled onto beige walls, and taped to nearly every office door were charts and diagrams, cutaway illustrations of subduction zones and volcanoes, aerial photographs of the coastline, long strips of seismograph paper smeared with erratic ink blots, and reprints of science posters.

  We found an empty square table in the coffee nook that looked like the only space big enough to fit several scientists in a group interview. Three of the men we were about to meet were out-of-towners with no office space in the building, visiting members of a team of seismologists and engineers who had just returned from a fact-finding mission to Mexico City. Over the coming days and weeks they would jointly write a science paper for the Canadian government’s committee on earthquake engineering, the panel of experts who make changes to the national building code.

  “The most surprising thing for me,” said Dieter Weichert, head of seismology at PGC, “was this two-second period of earth motion, which for an earthquake of this size should have been longer.” He described the tectonic origin of those strong pulses of energy that had pounded through the earth with clockwork regularity every two seconds. Think of a metronome clicking back and forth or a relentless drumbeat.

  It seemed to Weichert that a broken plate of oceanic crust ninety miles (145 km) long should have vibrated more slowly, perhaps once every six seconds. One might think the larger the slab of rock the slower the drumbeat, yet for some reason that’s not what happened. The shockwaves in Mexico had come bang, bang, bang—every two seconds.

  More fascinating to me, since I’d not heard any of this before, was the discovery that the deep soil of the dried-up lakebed underneath the city’s downtown core had also vibrated every two seconds, which amplified the force of the incoming seismic waves. Who knew that soil had its own frequency and would quiver like a bowl of jelly? I tried to imagine evenly spaced ripples rolling through dirt like waves on the ocean.

  It turns out the speed of vibration is determined by the type of soil (what it’s made of) and how hard or loosely packed it is. Weichert explained that structures of a certain height (six to seventeen stories tall) can also have a natural vibration period of roughly two seconds—which amplify the shockwaves yet again. The image of a seventeen-story tuning fork came to mind.

  The seemingly random damage pattern Robb Douglas had told me about on the phone from Mexico City back in September began to make sense. Before leaving on this shoot I had studied his news footage from the disaster zone: one condo block in ruins while another right beside it, maybe only a few stories taller or shorter, was left standing.

  To my surprise, the Mexico City building code already included specific and more stringent regulations for anything constructed on the lakebed because there had been damaging quakes there in the past and local officials thought they knew what to expect. But Ron DeVall, a civil engineer in his midthirties who sat across from Weichert, pointed out that this bizarre, unlucky, two-second rhythm had never been seen before.

  “They knew that the effect of the soil was pronounced in these areas and their code did account for it,” confirmed DeVall. “On firm ground they were talking about basic acceleration of 4 to 5 percent of gravity.” In other words, to survive the kinds of shaking that had hit Mexico City in 1957 and 1979, any new project constructed on bedrock since that time was in theory designed to withstand a lateral jolt against its foundations equal to 4 or 5 percent of gravity—meaning 4 or 5 percent of the building’s own weight. That much lateral reinforcement was a fundamental requirement of the local code.

  Anything constructed on the lakebed zone of the city was supposed to be even stronger, able to withstand twice as much lateral force—up to 10 percent of gravity. The shockwaves of September 19, unfortunately, had been much more powerful than that. “In actual fact,” said DeVall, “some of the instruments were measuring 20 percent gravity. So they were getting a very large shake, much higher than they had anticipated.”

  DeVall compared the crumbled towers to a child on a swing. “If you can time the rhythm of your pushing, you can just drive that swing higher and higher and higher. In essence, that’s what happens to a building sitting on the lakebed. The soil and the base rock were almost in resonance and the vibration was amplified to the building. The forces generated were much higher than what the building codes predicted. And the effect on the buildings was dramatic. It produced a much wider range of major damage than the previous earthquakes did.”

  Two other factors added to the toll of damage and destruction as well. Most of the Mexico City seismic regulations were written and passed into law as a result of the 1957 event and did not apply retroactively to thousands of city blocks constructed before that. Tragically, this included many schools and hospitals built decades earlier. “Seeing a lot of hospitals and schools damaged, which are traditionally thought to be areas of refuge in a disaster,” said DeVall, “to find them gone is, to my mind, a serious problem.”

  If there was any encouraging news from Mexico, it was that the damage was not evenly spread across the entire urban area, according to Robert Lo, a civil engineer who specializes in soil conditions. Lo even joked that on the day they had arrived in Mexico, roughly three weeks after the temblor, they hadn’t seen any significant damage. “We thought we were in the wrong city,” he smiled.

  It soon became obvious that the majority of the urban area had been built on solid ground and that well-engineered buildings were able to ride out the shockwaves with relatively little damage. Ironically, a lot of older buildings not expected to survive—built of masonry and other materials considered quite brittle and susceptible to shockwaves—were still standing because they weren’t tall and hadn’t resonated with the vibrations coming from this particular quake. The airport remained open throughout the disaster, the metro subway kept running, nearby dams did not collapse, and most of the main waterlines survived the shaking as well. Even the electricity remained on in many neighborhoods.

  But those ruined hospitals and schools had convinced DeVall that we in North America ought to learn something from Mexico’s misery. “It might be prudent on our part to review these buildings to make sure that they will sustain an earthquake,” he said. “You definitely want that hospital not only to be standing up afterwards, but functioning afterwards.” He pointed out that New Zealand, another seismically active area, had already begun this kind of survey of essential infrastructure. The bad news was that, as far as he knew, no seismic inspection was planned for urban regions of the Pacific Northwest.

  News reports at the time said the Mexico City rupture had been unexpected. What I wanted to know was why. Aren’t all earthquakes unexpected? “Well,” explained Dieter Weichert, the Mexico disaster “was a bit of a surprise because they had a historic record of two hundred years without a large earthquake. And there was reason for thinking there might not be one.” True, several smaller shocks had damaged Mexico City in the past and yes, there had been ruptures to the north and south of the plate that slipped this time. Through all the previous rumbling and lurching, however, this one segment of the ocean floor had remained quiet and nobody knew why. Could one subsection of the ocean floor be so different from the broken fragments on either side? To me this sounded like the missing clue at the heart of a great mystery.

  The unexplained absence of large quakes over a long time seemed to be the soft underbelly of
the “aseismic subduction” hypothesis—the conventional wisdom that the jagged edges between some tectonic plates are “not like the others” and don’t get stuck together. Weichert told me that some parts of the ocean floor were younger and hotter rock than the overlying continental crust and perhaps that was why this slab had been aseismic. The interesting twist was that in the aftermath of Mexico City, at least for the scientists gathered in this room, there was reason to believe something might be wrong with conventional wisdom—at least the aseismic part of it.

  “People thought that the plate was subducting smoothly,” said Weichert. And yet, quite obviously, it had been stuck and building up strain for two centuries. Seismologist John Adams from the Geological Survey of Canada’s earth sciences lab in Ottawa told us that a tectonic plate very much like the one that wrecked Mexico City—only three times larger—lies just offshore west of Vancouver Island, extending down the outer coastline of Washington, Oregon, and northern California.

  “It’s been demonstrated that the Juan de Fuca plate is subducting under North America. But we see no earthquakes along it. And we have a historic record of 150 or 200 years without large earthquakes. Therefore, there are two possibilities: either it is smoothly sliding under North America, doing it continually and without large earthquakes, or strain is building up for a large earthquake—of a type that would only happen every 200 to 500 years.”

  Was there an echo in this story? I was hearing for the first time about a slab of Pacific Ocean floor called the Juan de Fuca plate, which runs from Vancouver Island to northern California and is sliding eastward underneath the continental plate of North America. This eight-hundred-mile (1,300 km) fault, the boundary line between two tectonic plates, will later be named the Cascadia Subduction Zone. As I listened to John Adams, it sounded exactly like Dieter Weichert’s description of the plate that got stuck sliding underneath Mexico. Until September 19, both fault systems had been thought of as aseismic.