Cascadia's Fault

by Jerry Thompson

  His choice of the word plates would allow a new generation of researchers to put the old bogey man of continental drift behind them and move forward into the emerging world of plate tectonics—the new geology. It was a great time to be an earth scientist because there was still so much to figure out about how the system worked, especially along the western coast of North America and around the Pacific Rim.

  New voyages of discovery revealed that the mid-ocean ridge system—that twisty baseball seam of volcanic mountains circling the globe—cut through the wide, seafloor prairie of the Pacific Ocean, fracturing the main plate and pushing smaller pieces off to either side as it spread the sea floor wider. The Cocos plate, for example, had apparently been split off from the larger Pacific plate by a convection cell pushing new magma up through the East Pacific Rise: the segment of the baseball seam running parallel to the coast of South America. Upwelling magma had pushed the smaller Cocos plate eastward underneath Central America.

  Farther south, researchers learned that another broken slab of sea floor was being thrust under the coast of Chile. To the north another was punching its way down beneath the beaches of Alaska. The same was happening under the coast of Japan and in many other places around the Pacific Rim—all because of seafloor spreading.

  Anywhere you looked, broken plates were pushing against one another. At each one of these collision points were large mountain ranges, violent earthquakes, and active volcanoes. The Pacific was circled by a “ring of fire” caused by lumps of the earth’s crust crashing together, melting and erupting.

  While scientists around the world were busy piecing it all together in their minds and on paper, the earth itself was providing physical proof of what was really at stake. In 1960, the broken chunk of ocean crust jammed beneath Chile’s continental shelf finally reached its breaking point and snapped loose in the largest earthquake ever recorded. Scarcely four years later, George Plafker was collecting evidence that the same kind of horizontal fault had caused the 1964 earthquake in Alaska. Even though the old guard had still not accepted the idea of plate tectonics, Plafker was pretty sure he was right.

  One could argue that this should have been the dawning of awareness of the megathrust earthquake threat to British Columbia, the Pacific Northwest, and California as well. A 1965 paper by Tuzo Wilson pointed to the existence of what he called the Juan de Fuca Ridge. The name was chosen because the upper end of the ridge lay due west of the Strait of Juan de Fuca, which runs between Vancouver Island and Washington State. Here was an undersea mountain range that had previously been discovered and then dismissed as an insignificant, amorphous hump of rock running parallel to the coast. But if Wilson and the young turks of plate tectonics were right, the Juan de Fuca Ridge was in fact another part of that fiery seam of volcanic mountains running through the oceans.

  If this ridge turned out to be spreading apart sideways, powered by a cauldron of hot magma, it must also be thrusting a slab of sea floor underneath the edge of British Columbia, Washington, Oregon, and California. Presumably some kind of trench would be located where the two plates met, a “convergent plate boundary” just like the ones off Chile, Alaska, and Mexico. If so, giant earthquakes must surely follow.

  CHAPTER 6

  Nuke on a Fault: Early Clues in Humboldt Bay

  Flying south from the Oregon line in search of tectonic damage, the helicopter finally angled west over the last wall of mountains and down through a hole in patchy clouds to the California shore. We found the Shelter Cove runway, a cracking strip of sun-baked asphalt surrounded on three sides by a golf course on a bench of land just above the sea. My first thought was—how quintessentially West Coast. Fly in for a quick round of golf, go whale watching in the afternoon, have a barbecue on the beach, then fly home at sunset. Just try not to think about that monster earthquake hiding beneath the surf.

  On this particular summer morning in 2007, cameraman Doug Trent and I had planned to shoot aerial pictures for a new documentary called ShockWave on the communities closest to the Juan de Fuca Ridge and fault: the farms, ranches, and small towns from Cape Mendocino north to Humboldt Bay, Eureka, Arcata, and Crescent City, near the Oregon border. There was a reasonable chance the morning fog would burn off by midday, giving us the low-angle light we needed to highlight a half-dozen surface-level fractures where ancient tectonic ruptures had heaved up beaches and hillsides along the foreshore.

  Geology professor Lori Dengler at Humboldt State University had supplied Doug and me with a list of sites to photograph, complete with GPS coordinates that made them much easier to find. She referred to this section of California coast as a “fold and thrust belt,” the crumpled edge of a tectonic subduction zone.

  The first thing we photographed was a cleft in the hills directly behind Shelter Cove, the northernmost mapped trace of the San Andreas fault. Through the open door of a JetRanger it looked like just another deep shadow among the redwoods. It was bizarre to think this darkish line was in fact a crack in the earth’s crust, the constantly creeping boundary zone between the Pacific and North American plates.

  In the earthquake of 1906 the San Andreas tore itself apart to the north and south of San Francisco. The northern segment of the rupture ran 250 miles (400 km) from the Golden Gate all the way up the coast to Cape Mendocino—westernmost point of land in the lower 48 states—leaving this visible crack in the hills behind Shelter Cove. Even at such a distance from the ruined city, shockwaves here were strong enough to crack walls, break windows, and topple chimneys in the nearby farming towns of Ferndale and Eureka.

  At Cape Mendocino itself, the San Andreas disappears offshore. In the aftermath of the San Francisco disaster, scientists began to speculate and disagree about where the fault goes from there. Does it continue north toward Alaska? Or does it veer west out to sea? The routing or northern extension of California’s most famous and deadly fault remained a mystery for decades. By the mid-1960s, however, the emerging theory of plate tectonics seemed to promise a better understanding of how this amazingly complex system of cracks worked.

  Geologists and oceanographers knew from mapping the underwater terrain (the bathymetry) that the ocean floor looked like a broken dinner plate. From what they could see with the earliest, relatively low-tech echo-sounding equipment, the bottom of the Pacific Ocean had cracked in several places. The recently “rediscovered” (and newly named) Juan de Fuca Ridge rose from the deeps off the northern California coast, fracturing the sea floor in a northwesterly direction toward Vancouver Island, more or less parallel to the coast.

  A convection cell of hot magma from the earth’s mantle had apparently broken the Pacific plate apart, shoving a slab of oceanic crust (the Juan de Fuca plate) east underneath the oncoming (westward-moving) North America plate. This would eventually become known as the Cascadia Subduction Zone. On closer examination, researchers discovered that the Juan de Fuca plate itself had been fractured. The southern end of it was broken off and appeared to be moving independently.

  It turned out there was a separate, smaller ridge—another seafloor spreading zone called the Gorda Ridge—that looked like a southern extension of the Juan de Fuca system. It appeared to be pushing another chunk of oceanic real estate, the Gorda plate, beneath the California coast. There was also a heaved-up fracture zone running east to west across the larger Pacific plate. All these cracks and broken slabs, including the San Andreas, converged offshore at Cape Mendocino. Geologists decided to call this tectonic wreck the Mendocino Triple Junction.

  The ongoing and extremely slow-motion convergence of plates had fractured, bent, and folded rocks along the shore and hoisted up the beaches in several places, creating terraces we could now see from the air at a place called Singley Flat. They were, on a smaller scale, similar to the sections of heaved-up coastline George Plafker had discovered in Alaska after the 1964 disaster. If I had not been told what to look for, I would never have guessed these grassed-over benches of coastal farmland were the bent fenders of a conti
nental crash, evidence that Cascadia’s fault had caused numerous earthquakes over the years.

  Farther north we shot pictures of the ruptured earth at places called Little Salmon River, Mad River, and McKinleyville. These were even harder for our untrained eyes to notice because today they are camouflaged by a veneer of human civilization, the streets and homes, schoolyards and shopping malls of Eureka and Arcata, California. Who would notice that the nice little house on what looks like a landscaped terrace is in fact perched on the edge of an active, still-moving fault, a fractured wedge of crust that is being shoved upward by the force of plate tectonics?

  Native people who have lived in beachside villages along this coast for thousands of years tell stories they learned from their elders of horrific ground shaking on a winter’s night long ago, followed by a killer wave that wiped out entire communities. For the most part, though, the tide of white settlers who began arriving here in the 1850s to homestead and log the redwood forest were unaware of, or simply uninterested in, the local knowledge of Aboriginal people. After 1906, they knew—the whole world knew—about the deadly San Andreas, but the concept of plate tectonics and the fact that a moving block of ocean floor could cause an even larger shock was still unknown.

  The scientists and engineers who in the early 1960s drew up plans for an atomic power plant—California’s first commercial reactor—to be built on the shore of Humboldt Bay assumed there was no major seismic threat to worry about. Ironically, building the reactor would help scientists discover the reality of Cascadia’s web of faults.

  Around the corner and up the coast from Cape Mendocino, the side-by-side beach towns of Eureka and Arcata were inhabited in 1963 by an uneasy mix of loggers, commercial fishermen, and back-to-the-land idealists who would soon be labeled hippies and environmentalists. The biggest construction project in decades—the atomic power station at Humboldt Bay—was coming to an end and the Pacific Gas and Electric Company (PG&E) was about to deliver a sixty-thousand-kilowatt jolt to the local economy.

  Even though terms like meltdown and China syndrome had not yet colored the vocabulary, a tide of negative opinion had already derailed another nuclear plant downstate, so the residents of Humboldt County were aware of the potential for controversy and had mixed feelings about the promise of “power too cheap to meter.” Work at Bodega Head, about fifty miles (80 km) north of the Golden Gate, had been delayed by vigorous opposition for six years. Construction workers had managed to dig a deep hole for the foundations when geologists confirmed that the San Andreas fault ran right beside (some said directly underneath) the reactor site. Eventually, PG&E decided to abandon the project.

  Perhaps because the San Andreas veered out to sea at Cape Mendocino, state officials agreed with PG&E that seismic risk would not be an issue on the north coast at Humboldt Bay. With no large earthquakes in the area—at least not since the 1850s, when white settlers started keeping written records of local history—project managers at the utility and engineers who were designing the reactor were convinced the level of risk was within acceptable limits.

  The federal Atomic Energy Commission, which would eventually license the plant, defined an active fault as having had one “event” (earthquake) in the past 35,000 years. Or more than one in 500,000 years. AEC regulations in effect at the time specified reinforced, anti-rupture reactor vessels only when an active fault came within a quarter of a mile (0.4 km) of a plant. But with no detailed information available about the quake history of nearby faults, and with little understanding of the newly discovered Mendocino Triple Junction or the implications of plate tectonics, any seismic threat seemed distant and hypothetical. Not a problem at Humboldt Bay.

  “The plans for it were being developed as Tuzo Wilson’s first papers on plate tectonics were coming out,” Lori Dengler told me. As a professor of geology and then department chair at Humboldt State University, she studied the plant’s history and the simultaneous dawning of awareness of Cascadia’s fault. “There was absolutely no inkling of a subduction zone or great earthquakes,” she said. PG&E, trusting the best science available at the time, signed a contract to build the reactor six miles (10 km) south of Eureka.

  After the showdown at Bodega Head, the Humboldt plant became a test case for PG&E. Company officials and state and federal politicians no doubt wanted to prove that atomic power could be harnessed safely. The reactor started boiling water and generating electricity for the northern California grid in August 1963. Less than a year later, however, two plates shifting on a fault in Alaska focused worldwide attention on tectonic theory and sparked intense debates about seismic risk on the West Coast. Indirectly this heightened awareness would ultimately shorten the lifespan of the power plant near Eureka.

  Ironically, it was the tsunami rather than the earthquake that hit home first.

  When the Good Friday earthquake of 1964 wrecked the south coast of Alaska, it sent a train of waves crashing down the coast all the way to California and into the streets of Crescent City, where more than a dozen people were killed. Crescent City is only an hour’s drive north of the reactor on Humboldt Bay. Yet the vulnerability of the power plant to a similar wave or to a massive seismic rupture (or even to a local earthquake) did not occur to people living in the area. Not right away.

  All eyes were on the immediate tragedy and its aftermath. The notion of an Alaska-size quake from a subduction zone just twenty or thirty miles (30–50 km) off the California coast struck no fear in the hearts of Eureka, Arcata, or the other small communities of northern California. Only those scientists who’d read—and were convinced by—the latest research realized what the Alaska story meant to the rest of the coast.

  Many frontline scientists themselves were still at odds over what to make of plate tectonics. In February 1965, almost a year after the biggest known earthquake in North American history, Frank Press, the distinguished seismologist at Caltech, published a paper describing the fault that had wrecked the Alaska coast as a steep, vertical crack in the ground, similar to the San Andreas. Four months later George Plafker, the young geologist who did the muddy-boot work of measuring and documenting the aftermath, published his own paper saying just the opposite: that the fault ran at a shallow angle from the sea floor underneath the continental landmass. Plafker thus cast his vote in favor of tectonic theory while Press, the established authority, took a more cautious position—and his was still the majority view.

  A nearly identical disagreement about the Mendocino Triple Junction arose three years later. A team of scientists at the University of California seismology lab in Berkeley released a paper in 1968 suggesting the San Andreas did indeed extend out to sea from Cape Mendocino and was in the process of cracking its way toward Alaska. Searching through seismograms for clusters of quakes, they discovered that Cape Mendocino was trembling at a rate “two to three times that of the combined central California clusters,” which seemed to confirm that those broken slabs of sea floor were on the move, just as Tuzo Wilson had predicted they would be.

  Bruce Bolt and his colleagues at Berkeley calculated a series of fault plane solutions and decided that many of the smaller tremors were generated by cracks running northwesterly through the Gorda plate toward Alaska—not northeasterly, as you might expect if the Gorda plate were being pushed underneath the California coast by seafloor spreading. Essentially Bolt and his team believed the region was “dominated tectonically by the San Andreas.” No mention in their paper of spreading ridges offshore, low-angle faults, or pieces of oceanic crust getting stuck underneath the continent. The Alaska threat did not apply to California in their view.

  Then along came Eli Silver from the U.S. Geological Survey in Menlo Park. In 1971 he wrote a new paper on the tectonics of the Mendocino Triple Junction and came to the opposite conclusion from Bolt. Recall that there are always two possible answers to a fault plane calculation—one being ninety degrees different from the other. While Bolt chose the northwesterly solution, Silver was sure the northeastern vector
(and tectonic theory) was correct.

  Meantime, as if anyone needed a reminder, the San Andreas showed once again that it still had the power to shock, wreak havoc, and dominate the attention of scientists and the general public. On February 9, 1971, the magnitude 6.6 Sylmar earthquake struck the San Fernando Valley, killing sixty-five people and rattling Los Angeles emergency planners, police, and firefighters like nothing had in years. Seeing two hospitals (one of them brand new) and a dozen freeway bridges collapse prompted state politicians to pass a new law (the Alquist-Priolo Earthquake Fault Zoning Act) that would tighten building codes and severely restrict construction of residential buildings and critical infrastructure on or near any active fracture zone in California.

  Only a few years later, two geologists hiking the north woods of the California coast would quietly shift the focus back to the atomic power plant at Eureka. Gary Carver, a freshly minted professor at nearby Humboldt State University, and Tom Stephens, one of his senior thesis students, were conducting field research on a forest company’s land in the upper Mad River drainage basin east of Arcata when they discovered some “very large and previously unknown and unmapped faults.”

  The Simpson Timber Company had kept hundreds of miles of access roads closed to the public during the 1950s, ’60s, and ’70s while they logged off huge stands of old-growth redwood forest. None of the land had been geologically mapped and Carver and Stephens didn’t quite know what to expect. “We were able to go back in where geologists hadn’t been for a very long time,” Carver told me.