Japan also missed precursory motion on the seafloor. After-the-fact analyses showed that the underwater fault had been slipping for several days. Following a large foreshock, that motion seems to have accelerated two days before the main rupture. If sensors had detected the slip early, would the Japanese have been alarmed enough to raise the alert level? It’s hard to say. Most episodes of slow slip don’t trigger big quakes. But more extensive monitoring could have alerted Japan to the fact that dangerous levels of strain were building up offshore in a region where most seismologists didn’t believe a magnitude 9 quake was possible.
“I don’t think there is any magic bullet that will unambiguously identify an upcoming rupture,” said Jeffrey McGuire of Woods Hole Oceanographic Institution. “But I do think there are time periods when large ruptures are more likely than others, and it is a realistic goal for our community to get much better at putting quantitative, probabilistic assessments [on] those risks.”
Japan is determined not to miss any possible red flags the next time. The government pledged $400 million for a cabled network of sensors off the Tohoku coast. More than 150 instrument packages, each containing a seismometer and water pressure gauge, will stand constant watch. A similar network is already in place along the Nankai Trough south of Tokyo where Japanese scientists have long anticipated a major quake. Japan is also boosting its investment in underwater GPS to monitor the way the seafloor deforms as pressure builds. If subduction zones tip their hand at all before they rip, the warning signs are most likely to show up in unusual warping or motion on the ocean bottom, things GPS is well-suited to spot.
In the near term, the Northwest’s best chance of detecting any unusual movement on the Cascadia Fault could be a pair of sensors McGuire and his colleagues plan to plug into NEPTUNE Canada by 2014. Called tiltmeters, the instruments will be able to detect motion of less than half an inch on the fault and track the buildup of strain. Based on the same principle as a carpenter’s level, tiltmeters are often used on land to monitor the way volcanoes inflate and deflate with the movement of magma. Underwater robots will place the instruments directly above the subduction zone in a one-thousand-foot-deep borehole. “The location is great for studying subduction zone earthquakes,” McGuire said. The meters could help determine which parts of the fault are primed to rupture and put better estimates on the size of the next Cascadia tsunami.
Other scientists are working to bring seafloor GPS to the Northwest. The technique was pioneered off Vancouver Island but is so costly and cumbersome only the Japanese have embraced it. Satellite and radio signals can’t penetrate water, so conventional GPS technology doesn’t work under the ocean. The Japanese rely on a two-step process. Acoustic transponders on the seafloor send signals to a ship. Instruments on the ship calculate the transponders’ positions while GPS fixes the ship’s location.
To detect changes in the seafloor, the ship must return to the same spot repeatedly. Setting up a single station on the ocean bottom can cost $120,000. But the real expense is ship time, at up to $50,000 a day. Japan enlisted its Coast Guard to take the readings. American researchers are searching for lower-cost alternatives.
Scientists from the Scripps Institution of Oceanography are exploring the use of buoys and robot gliders to collect data beamed up from the seafloor. Delaney’s dream is a web of interconnected acoustic sensors spaced about a mile apart. The network could continuously monitor conditions along the length of the subduction zone. “That would be a very big deal,” Delaney said. “It would give us the ability to detect the early onset of major deformation events.”
So far, though, intensive GPS monitoring of faults on land hasn’t yielded advance notice of earthquakes. When the massively wired Parkfield section of the San Andreas Fault finally broke in 2004, neither GPS, tiltmeters, nor seismometers dropped any hints. Pessimists say earthquakes may be inherently unpredictable. Optimists, like Harold Tobin, argue that scientists may have been looking in the wrong place. “We’ve had a hundred-plus years of seismology, but all those instruments are on the surface,” he said. “Earthquakes don’t happen on the surface. They happen kilometers down in the Earth.”
Tobin is going deep. The University of Wisconsin geoscience professor is leader of the first project to drill into a subduction zone. His earliest experience with seafloor drilling was off the coast of Oregon. As a graduate student, Tobin participated in expeditions that augered holes up to 1,500 feet into the Cascadia fault system. But those projects barely scratched the surface. Tobin wanted to burrow down to the zone where earthquakes are born. The only nation willing to finance an undertaking of that magnitude was Japan. “Earthquakes are their greatest national security threat,” he said. “The way the U.S. thought of terrorists after 9/11 is how Japan thinks of earthquakes and tsunamis all the time.”
No existing research ship was powerful enough to handle the job, so Japan built its own at a cost of half a billion dollars. Called Chikyu, which means “Earth,” the vessel features the same kind of drilling technology used by the oil industry. Daily operating expenses range between $200,000 and $400,000. “It’s not an exaggeration to say that every few days is a million dollars,” Tobin said. Since the ship started drilling off Japan’s coast in 2007, it has set several depth records. The goal is to reach 3.5 miles down, into the zone where one plate grinds past another. It will be the deepest research borehole in the ocean and the first time science has reached into the living heart of a subduction zone.
The target is the Nankai Trough, which has so much in common with Cascadia that Tobin calls them sisters. “We’re going to learn a lot from this, and much of that will translate directly to understanding the hazard and the science of the Cascadia Subduction Zone.” Tobin and an international team of collaborators will pack the hole with instruments to measure temperature, strain, creep, and fluid behavior. They’re already poring over miles of core to better understand the way rocks change and break during an earthquake. In one set of cores from the fault zone, researchers found a thin band where rock had been heated to nearly six hundred degrees by the frictional heat of a past quake.
Theory predicts that earthquakes start with small ruptures that propagate and cascade, but no one knows the steps or how long the process takes. “We don’t know enough yet to even know if precursory signals exist, but it’s worth looking,” Tobin said. “We need to have long-term monitoring before an event, then be there during the event.” He doesn’t expect the instruments or boreholes to survive the next great Nankai quake, but he hopes to collect data up until that instant. Then he would be happy to move on. “I would love to do this in Cascadia, too.”
While subduction zone studies in Japan promise insights for the Pacific Northwest, the Italian city of L’Aquila provides an example of what can go wrong when it comes to sharing those insights with the public. In a case that made seismologists around the world shudder, six scientists and a government official were convicted of manslaughter there in 2012.
Most news coverage of the trial cast the scientists as scapegoats, prosecuted because they couldn’t predict an earthquake that killed 309 people and left the lovely, medieval city in ruins. But the charges had nothing to do with earthquake prediction. The scientists were hauled into the dock because they failed to adequately communicate seismic risks to frightened residents in a community that had been rattled by hundreds of small tremors.
The accused were members of a science advisory panel convened in March 2009 to assess the likelihood of a damaging quake. Wiretapped conversations later revealed that the meeting was largely a public relations ploy to quell panic over the earthquake swarms and a local man’s prediction that a massive shock was imminent.
After less than an hour’s discussion, the panel members dismissed the prediction as bogus and agreed among themselves that small quakes were not necessarily a precursor to anything bigger. Then the scientists left town without issuing a statement. It fell to the bureaucrat on the panel to answer questions from the pr
ess.
Bernardo De Bernardinis, deputy chief of Italy’s Civil Protection Department, told journalists there was no danger because “there is an ongoing discharge of energy.” De Bernardinis even endorsed a television reporter’s suggestion that residents should relax with a glass of wine. The deputy chief suggested a Montepulciano. When the magnitude 6.3 quake struck a week later at 3:32 AM, most residents were in bed. Many told the court De Bernardinis’s reassuring statement convinced them not to sleep outdoors, as was traditional when small quakes rumbled the region. The victims were crushed or buried in rubble. After a lengthy trial, the court sentenced all seven defendants to six years in prison. (The Italian justice system allows for multiple appeals, and it’s not clear whether any of the men will serve time.)
Watching the case unfold from Seattle, John Vidale felt uneasy. One of these days, the Cascadia Subduction Zone could start acting up. Maybe a magnitude 6 quake will pop on the plate interface. Perhaps an episode of slow slip under the Olympic Peninsula will continue for months instead of weeks, potentially raising the risk of a major quake. As director of the Pacific Northwest Seismic Network at the University of Washington, Vidale will be among the first to know if anything unusual is afoot. So he finds it easy to put himself in the shoes of the Italian scientists.
Vidale’s not worried about going to jail. Scientists who advise the U.S. government are generally shielded from liability. But the rest of the L’Aquila scenario could play out in the Northwest, with a spooked public demanding answers from seismologists who have no real way of knowing what the Earth is going to do next.
“If strange things are happening, we will just have to be transparent and try not to say more than we know,” Vidale said. He’s also determined to avoid the missteps of the Italian experts, who were accused of providing “inexact, incomplete, and contradictory information.”
There was no basis for De Bernardinis’s claim that the city was safe because small quakes were bleeding off energy. “That was just stupid,” said one American seismologist. “You never say there’s no danger.” The scientists on the panel knew De Bernardinis was wrong but didn’t step forward to correct him. Nor did the experts remind residents about the region’s history of deadly quakes or caution people who lived in ancient stone buildings. And although the scientists were technically correct that swarms of small quakes do not usually presage anything bigger, they failed to provide the one piece of useful information seismology can offer in a situation like that: earthquakes can trigger other earthquakes. The odds of a big quake striking L’Aquila were still low in the spring of 2009, but they were higher than usual because of the recent tremors.
The question of what to tell the public if Cascadia starts to rumble is already on the agenda of the National Earthquake Prediction and Evaluation Council (NEPEC); yes, there really is such a thing. NEPEC was established in 1976 to advise the USGS at a time when routine quake predictions seemed just around the corner. They weren’t, and the council went dormant for more than a decade until the USGS revived it in 2006. Much of what the council does now is debunk earthquake prediction claims, whether from spiritual channelers in Taos or retired engineers in Pomona. NEPEC’s other job is to help the Survey grapple with questions of risk, communication, and forecasting, the very issues at center stage in L’Aquila.
Vidale serves on the council, which has drafted statements that could be issued after a moderate quake on the subduction zone. “The likelihood of an earthquake being followed by a larger event is elevated over the background rate for a period of three to ten days,” reads one version. Beyond that, there won’t be much more scientists can say, except to remind people what a Cascadia megaquake could do and offer generic preparedness advice. It will be up to political leaders and emergency managers to decide whether any action is warranted, Vidale said.
Tom Jordan, director of the Southern California Earthquake Center, is pushing for a more systematic approach. He calls it operational earthquake forecasting and says the goal should be to emulate meteorologists. “We need to be producing something like seismic weather reports every day,” Jordan said. He’s convinced seismologists already have many of the tools they need to calculate the odds that an earthquake will strike in the near future. The main risk factors are the occurrence of other earthquakes in the vicinity and the level of stress on the target fault.
Jordan chaired a commission that examined the L’Aquila incident and advised the Italian government on how to respond the next time earthquake swarms or small quakes strike in seismic hot spots. A version of what the commission proposed is already in place in California, Jordan said. The state’s earthquake prediction council, on which he serves, advises the governor when tremors pop up near faults that may be primed to rupture.
For example, at about the same time foreshocks started rattling L’Aquila, a swarm of fifty small earthquakes shook the shores of the Salton Sea east of San Diego. The area lies near a portion of the San Andreas Fault that hasn’t ruptured since 1680. Residents were awakened before dawn on March 24 by a magnitude 4.8 shock, the region’s biggest since record keeping began. Within a few hours, Jordan and the other members of the California Earthquake Prediction and Evaluation Council (CEPEC) met by teleconference and crafted a statement: “CEPEC believes that stresses associated with this earthquake swarm may increase the probability of a major earthquake on the San Andreas Fault to values between 1 and 5 percent over the next several days.” Numbers that low would never justify an evacuation, Jordan said. Nor did the alert cause an exodus from the lightly populated valley. Emergency managers weren’t impressed by a 5 percent risk, but some took advantage of the incident to emphasize the value of earthquake kits. The San Andreas didn’t rip.
If the Italians had offered a similar assessment, coupled with a cautionary statement about central Italy’s seismic danger, criminal charges would probably have never been filed, Jordan said. “In some sense you have to empower people to do what they think is the right thing to do.” By laying out what they knew and didn’t know, the California scientists let residents decide whether or not to act. “The anger you see in L’Aquila is over the fact that people felt they were deprived of important information.”
The public appetite for seismic information is growing. More than twenty thousand people a day visit the Pacific Northwest Seismic Network’s website and a community of enthusiasts dissects the data posted online. When the network sought 90 homeowners to install and maintain seismometers in their basements and garages, nearly two thousand people volunteered. Vidale and his colleagues started a blog and Facebook page in 2011, and they’ve been fielding a steady stream of questions and comments ever since.
Jordan sees regular earthquake forecasts as another way to communicate with the public and help counter misinformation floating around in cyberspace. But his approach hasn’t gained much traction. Weather forecasts can help you decide whether to carry an umbrella or plan a hike. Most people have no idea how to react to a forecast that calls for a 1 percent chance of an earthquake. Some seismologists question the probability calculations. Compared to weather forecasters, scientists who study earthquakes remain virtually in the dark about what the next week—or the next ten years—will bring. It’s easy to crank out numbers, but it’s impossible to validate them.
Vidale puts more stock in a type of alert that can guarantee a quake is on its way. There’s a catch, of course. The earthquake early warning system he and others hope to implement along the West Coast would offer at most five minutes notice that the ground is about to shake. In some areas the window could be as slim as fifteen to thirty seconds. That’s because the technology doesn’t kick into action until the earthquake is already under way. The system is designed to detect the first seismic waves to arrive and sound the alarm before more destructive pulses hit.
Half a minute isn’t enough time to jump in your car and evacuate to Idaho. But it would allow you to climb off a ladder, duck under a table, or put down your scalpel if you happen t
o be performing surgery. Japan completed a $1 billion early warning system in 2007 and has issued alerts for more than a dozen strong earthquakes since then. Many Japanese businesses and industries are hardwired to automatically shut down machinery, recall elevators, and switch to backup power the moment the warnings arrive. Schools broadcast the alerts over their public address systems and students dive for cover. The system’s biggest test came on 3/11. Before the worst shaking hit, twenty-seven bullet trains slowed to a stop without a single derailment. Warnings went out to people all along the coast and, coupled with tsunami warnings, saved thousands of lives. “Japan showed that early warning has a lot of value,” Vidale said. Mexico, Taiwan, and Romania have similar systems in quake-prone urban areas.
The approach works for two reasons. Seismometers closest to an earthquake’s source detect the motion first and can relay information faster than the seismic waves travel through the ground. The fact that earthquakes send out several different types of seismic waves can also provide a short warning window. The fastest waves to spread are called p-waves. They zip through the ground at up to 17,000 mph and are responsible for the first, small spike that appears on a seismogram. P-waves are less destructive than the slower-moving secondary waves and surface waves that follow on their heels.
At sites distant from a quake’s origin, the lag between the arrival of p-waves and the most damaging motion can stretch to minutes. In a Cascadia megaquake, the strongest shaking won’t hit Seattle until two to five minutes after the start of the rupture. In California, where cities are riddled with fault lines, warning times would be much shorter. The West Coast warning system that Vidale and his colleagues hope to build would also be tied into GPS stations. With instantaneous measurements of how far the ground has moved, GPS can judge the size of big earthquakes much less ambiguously than seismometers can.
Full-Rip 9.0: The Next Big Earthquake in the Pacific Northwest Page 26