Some quakes are so powerful they leave the earth speechless. The Great San Francisco Earthquake of 1906, a massive 7.8 quake, shook the region so violently that scientists believe it cast a calming "shadow" over a half-dozen parallel faults, sending the entire Bay Area into a long seismic sleep.
In the seventy years leading up to the catastrophic day, at least fifteen quakes measuring 6.0 to 6.9 slammed the Bay Area. But since the morning of April 18, 1906, when the ground convulsed for nearly a minute, there has been precious little activity -- nothing exceeding a 6.5 on the infamous San Andreas; nor the Calaveras, which flanks I-680 through Pleasanton and Danville. The Hayward Fault, which cuts through the East Bay Hills, has remained essentially idle. Ditto the San Gregorio, Rodgers Creek, Concord-Green Valley, and Mount Diablo faults. These giants have been slumbering in relative peace for the better part of a century -- which is striking given that the region's soaring hills and deep valleys were formed by countless millennia of seismic upheaval. "The Bay Area has the highest density of active faults per square mile of any urban center in the country, and on a long-term basis it has the highest amount of earthquake energy released per square mile of any urban center in the country," says David Schwartz, a geologist with the US Geological Survey. "So we're really kind of living at ground zero."
Loma Prieta was not, in fact, a local quake. The 6.9 that rocked the Bay fifteen years ago last October was centered sixty miles south, in the Santa Cruz Mountains near Watsonville. Although it was within the San Andreas zone, seismologists believe the rupture occurred on a sub-parallel fault, not the San Andreas itself. And because a quake's energy diminishes over distance, Loma Prieta had weakened considerably by the time it reached the East Bay. It was also a short and clean rupture that, in a freak bit of good luck, struck in the early fall when landslide risk is minimal, and during game three of a Giants-Athletics World Series. Hundreds of people who otherwise would have been stuck in rush-hour traffic and crushed by the doomed Cypress freeway were at Candlestick Park or safely at home watching the ball game.
As for the toll -- 63 deaths and an estimated $10 billion in overall damage -- scientists who study the Bay's earthquake potential are unanimous: We ain't seen nothing yet. "[Loma Prieta] was merely a warning shot across our bow," says William Lettis, a Walnut Creek-based consultant in applied earth science. "It was enough to shake everybody up and say hey, earthquakes can happen, but it didn't kill very many people. ... An earthquake in the heart of the Bay Area will be ten to a hundred times worse than Loma Prieta."
It is not an overstatement to say anyone who thinks Loma Prieta was "the big one" is in a dangerous state of denial. On the moment magnitude scale -- the Richter scale successor now used to rate quakes -- each point represents a 33-fold increase in released energy, which means the great quake of 1906 was 2,320 percent stronger than Loma Prieta. If a 6.9 temblor managed to kill scores of locals, pancake an East Bay freeway, close ten bridges, and shut down the Bay Bridge for more than a month, all from a distance of sixty miles, consider the scenario a repeat of 1906 would unleash. Our world would come falling down.
It turns out the protective shadow of the 1906 quake was a mixed blessing. While Bay Area residents experienced fourteen times fewer 6.0-plus quakes in the 75 years following the big one, compared with the 75 years that preceded it, this geological cease-fire came with a price. It has led to decades of public apathy, and has allowed developers and government agencies in this fast-growing region to encroach upon the deceptively quiet fault traces.
Major Bay Area highways and BART structures cross or run dangerously close to the faults. Key gas, water, and sewage lines crisscross them, as do telephone and power cables. Cal State East Bay sits alongside the primed Hayward Fault, and UC Berkeley straddles the fault -- which famously runs between the goalposts at Cal's Memorial Stadium.
Perhaps most alarming is that authorities have placed critical lifeline facilities such as hospitals and power stations on hillsides vulnerable to landslides, or on shoreline landfill that may literally turn to liquid in a powerful earthquake. San Francisco and Oakland airports, Oakland's main sewage-treatment plant, portions of the Oakland port, the Bay Bridge toll plaza -- all are built on this type of land. Despite active retrofitting efforts, some of the East Bay's most critical structures remain vulnerable: the Bay Bridge, BART tracks and tunnels, medical centers, downtown office high-rises, and apartment buildings that house tens of thousands of local residents. As it stands, the Bay Area is seriously pushing its luck.
And now for the really bad news: We're coming out of the shadow. US Geological Survey geophysicist Tom Parsons estimates that the 1906 quake reset the Bay Area's faults, delaying their next ruptures by 17 to 74 years, depending on the fault. But he also believes the relaxation effect wore off by 1990 -- and that local faults are now busy accumulating the stress that will ultimately burst free in a deadly new barrage of seismic activity.
How likely is a killer quake? In 2002, a working group composed of researchers from the USGS, universities, and private institutions calculated a 62 percent chance that a 6.7-or-larger quake will strike on a local fault by 2032. In addition, they came up with an 80 percent probability that we'll see one or more local shakers in the 6.0 to 6.6 range over the next 28 years.
These numbers are truly worrisome, given that even a Loma Prieta-sized rupture in an urban fault zone could devastate the East Bay. If the weatherman gave a 62 percent chance of rain, Lettis points out, most people probably wouldn't leave the house without an umbrella. "I've heard people say, and seen letters to the editors in newspapers, that when we come out with these earthquake forecasts and these probabilities that we're trying to scare people and raise money," Schwartz says. "You have to realize that we're scientists and we're doing this as part of our job, but we also live here. We have our families and our homes and our investments, so this is more than just calculating numbers. It really impacts us on a very personal level, too.
"People say to me, 'Oh yeah, I've lived in the Bay Area all my life, I know what earthquakes are like,'" he continues. "The reality is they don't know what a strong earthquake really is. And they are going to find out."
Location, Location, Location!
How much time do we have? Nobody really knows. The 62 percent figure gives no inkling as to whether the next deadly quake will happen tomorrow or twenty years from now. The forecasters chose their thirty-year window because it's the length of the average home mortgage -- something people can relate to. They used a threshold magnitude of 6.7 because that was the size of the 1994 Northridge quake, which killed 57 people and did more than $40 billion in total damages, demonstrating the sort of toll a well-built urban region might expect from a moderate quake.
Long-term forecasts, however, are vague by nature. Pinpointing the time, location, and magnitude of any given quake is well beyond the reach of today's science. Although old wives' tales regarding earthquake prediction are legion -- that we should look for changes in animal behavior, in well-water levels, in the weather -- seismologists have yet to identify any reliable early-warning sign. The probability game is further complicated by several X-factors that determine a quake's size and damage potential. Among them: how long the shaking lasts, the direction of rupture along the fault, the sort of soil the quake propagates through, and how it is influenced by nearby faults.
To comprehend the complexity quake forecasters face, it helps to understand the layout of the Bay Area fault system. The majority of local faults run roughly parallel, bending in a slight rainbow-like arc that mirrors the coastline. Pressure on the faults builds up continuously as the region's tectonic plates -- impossibly massive slabs of rock that drift about on the earth's surface like slow-moving bumper cars -- grind against one another. In the Bay Area, the Pacific Plate scrapes northwest past the North American Plate at an average distance of 45 millimeters per year. That doesn't sound like much, and some fault segments can slowly "creep" to absorb the movement. Over time, though, most faults simply accumulate underground strain until they snap. At that point, they can move feet or meters on the surface all at once, and that's when the serious shaking starts. The farther a fault slips, the bigger the quake. And the longer and harder it shakes an urban area, the greater the damage, and the death toll.
The damage to a given community depends in part on the direction a fault breaks. It can rupture from one end to the other, or from the middle toward both ends -- there's no predicting it. In general, structures at the far end of the break bear the brunt of the force. "If you're close to the fault ahead of the direction of rupture, all that energy is piling up to get to your building," says Gregory Fenves, chair of UC Berkeley's department of civil and environmental engineering. "It creates very large pulses." Anything standing in the way of those compounded energy waves takes a beating. For example, if the Hayward Fault ruptures from the Fremont end and propagates north, that's worse news for Berkeley than if the rupture starts there and heads south along the fault.
Faults tend to rupture in cycles, but there are two theories on how predictable the cycles are. According to one view, a fault can handle only a fixed amount of strain before it ruptures, shedding its entire load like a stretched rubber band snapping back into shape. By calculating the rate at which the fault builds up stress, scientists could in theory predict the number of years between earthquakes. In the opposing view, a fault releases only part of its built-up tension during a quake, a scenario that would make prediction far more problematic. "It's not clear yet whether earthquakes occur completely episodically, like every two hundred years, or whether you have [one] now and you wait one hundred years and for the next one you wait five hundred years," earth science consultant Lettis sighs. "It might be very unpredictable. If it were real periodic we'd have it nailed, but we just don't know."
One reason we don't know is that scientists have been able to peer back only a few centuries into the histories of local faults -- Schwartz thinks the seismic record for Northern California is complete back to about 1850 for quakes 5.5 or greater, thanks to estimates based on old newspaper reports. But in dealing with a subject as old as the Earth itself, 150 years is a blink of the eye. Scientists are able to go back in time a bit farther by digging trenches along faults, where the Earth has preserved a sort of fossil record of past ruptures. By carbon-dating bits of bone, ash, or plant material embedded in the soil, they can estimate the date of each break. "Usually you can get back a few thousand years if you're lucky, but it really depends on the site," says Lind Gee, a professor and researcher with UC Berkeley's Seismological Laboratory. "In the Bay Area it's really hard to find good sites. Most of them have been built over."
It's difficult to gather data going back much earlier -- the trenches get too deep and too close to the water table, causing them to flood or collapse, and scientists often can't find enough materials to carbon-date. Even several thousand years of a fault's history, moreover, may reveal but a small part of its mysteries. "Even if it has behaved like this for the last few thousand years, that doesn't mean that that's the way it's behaved for the last thirty thousand," Gee points out.
What trenching reveals is nevertheless alarming. For example, it has shown that the Hayward Fault has ruptured at intervals of roughly 175 years over the last two millennia -- although the four most recent breaks were only about 130 years apart. The last quake on the Southern Hayward Fault, which until 1906 was considered the "great" San Francisco earthquake, was a 6.9 shaker in 1868 -- 137 years ago. There hasn't been a quake on the Hayward's northern segment since the American Revolution.
Hayward is the fault local geologists are most concerned about. The forecasters have calculated a greater than one-in-four probability that a large quake will hit on either the northern or southern segment by 2032. "These two earthquakes are likely to produce the greatest amount of social disruption and economic loss of any earthquake in the United States, potentially in the world, because it's such a densely urbanized corridor and it affects such a rich economic environment," Lettis says.
As if that weren't bad enough, here's another stomach-churning proposition: At close range, one big shaker can trigger another -- although massive quakes tend to relax faults that run parallel, they typically increase the strain on faults they butt up against. USGS geophysicist Ross Stein, an expert on fault interactions, says nearby faults are in a kind of "conversation" with one another. "It's possible for one fault to relieve stress on the other and inhibit earthquakes, but also for a fault to increase stress on surrounding areas and increase the likelihood of a large earthquake," he says.
A rupture on one fault section can likewise trigger later quakes on the same fault. Stein cites the North Anatolian Fault in Turkey, which is considered a "twin" to the San Andreas because it is of the same type, the same length, slips at a comparable rate, and can unleash quakes of similar magnitude. Given the North Anatolian Fault's slip rate, it should have an earthquake roughly every 250 years. Instead, Stein says, it has had twelve large ones in the past sixty years, with the ruptures generally progressing along the fault, each one touching off where the last one ended, like a string of firecrackers. "They're happening much more closely spaced in time than we would expect if there weren't some kind of link between them," he says.
In the case of 1906, he says, the Bay Area was actually lucky in that the stressed endpoints of the San Andreas are quite far away -- the northern end is offshore and the southern end is along a section of the fault that, according to Stein, "probably just creeps to dump that stress and doesn't produce large earthquakes."
Another rupture on the San Andreas, however, could stress the San Gregorio Fault. The Calaveras and Concord-Green Valley faults, meanwhile, dead-end into one another. Of particular concern is the relationship between the Hayward Fault and the Rodgers Creek Fault directly to the north -- where scientists forecast a 15 percent chance of a 7.0 shaker by 2032. "We have the possibility here that we could have an earthquake on the Northern Hayward Fault that would stress the Rodgers Creek, and a month or two later we might have a second big earthquake," Schwartz says. "While those might sound like doomsday scenarios, they are based on what we are looking at around the world. These are very real possibilities."
A common misconception is that several small earthquakes can somehow defuse or postpone a more devastating one. Alas, it doesn't work that way. Since every step up in magnitude represents a 33-fold increase in a quake's power, a 7.0 actually releases 1,089 times (33x33) as much energy as a 5.0. At the same time, each step up the scale represents a tenfold drop in frequency of occurrence -- 5.0 quakes, in other words, are a hundred times more common than 7.0 quakes, but more than a thousand times weaker. "If I wanted to avoid a Loma Prieta magnitude 7 by knocking off magnitude 5s, I'd need a thousand of them," Stein says. "If I wanted to do it with magnitude 3s, I'd need a million of them. We don't get that many small ones. You wouldn't be able to drive your car if we had a system in which small earthquakes did the job of a big one, because you'd be thrown over the road all the time."
So it comes down to this: We're in big trouble. "You can't hide," Lettis warns. "If you live in the Bay Area, you will experience the affects of a major earthquake in your lifetime."
And since we can't avoid it, the biggest personal survival factor for Bay Area residents may simply come down to location: Where will they be when the ground starts to shake?
All Fall Down
"There's an old saying that earthquakes don't kill people -- buildings do," says Keith Knudsen, senior engineering geologist for the California Geological Survey. He has a point -- back when California was mostly open space with few residents, quakes didn't do much damage. As of the 2000 Census, the Bay Area had a population of nearly 6.8 million, and it takes a complicated web of bridges, roads, and buildings to transport and house all of those people. Unfortunately, during the decades when much of the Bay Area's current infrastructure was built, there was scant earthquake data available, and most buildings weren't designed for the larger temblors scientists now expect. "There was somewhat of a resistance in the engineering community that, well, earthquakes haven't been bigger than a certain level, and since we've designed buildings to reach a certain level, that's good enough because we've seen good performance," says Cal's Fenves. "With Loma Prieta in '89, Northridge in '94, and then the Kobe earthquake in '95, we started to see the effects that even moderate-sized earthquakes in urban areas had. ... I think the message started getting across to the engineering community that our design criteria were not based on realistic estimates of ground motion."
The 6.9 quake that hit Kobe, Japan, in 1995 proved a striking local object lesson because the city resembles Oakland in so many ways -- both are highly industrialized port cities that are partially built on soils vulnerable to liquefaction. The Kobe quake killed nearly 6,500 people and caused some $150 billion in damage, most notably to its port, leaving its local shipping industry crippled.
Since Loma Prieta, state agencies, most public utilities, and many big property owners have begun retrofitting programs to counteract known vulnerabilities before it's too late. But much of the work is yet to be done, and critical structures and lifeline services would be devastated were a large quake to strike tomorrow.
The quake itself poses five main hazards, and the East Bay is at risk for all of them. In descending order of damage potential, they are: 1) Buildings and structures will topple; 2) Structures will be torn apart by the rupturing fault; 3) Soil that appears firm will turn to mud; 4) Landslides will push everything downhill; and 5) In a peculiar phenomenon known as a seiche, the water in the San Francisco Bay could slosh back and forth like a wave trapped in a bowl. Although this is the least probable, the other hazards will likely cause extensive damage over a wide area. Here's a preview:
The Earth Shakes
The violent ground motion produced by a local 6.7-plus earthquake will affect the Bay Area profoundly. The people who monitor earthquakes are keeping their fingers crossed that it will strike in the wee hours. That's because single-story wood-frame houses, which are very common in this region, are expected to fare well provided they've been reinforced and bolted to their foundations. "Typically our housing is far better than most housing in the world," says Jeanne Perkins, earthquake program manager for the Association of Bay Area Governments. "Just our run-of-the-mill wood-frame houses, even if they slide off their foundations, they don't kill people."
That doesn't mean homeowners are in the clear, says Mary Comerio, a Cal architecture professor who specializes in earthquake engineering research. The Northridge quake, for example, damaged the average family dwelling to the tune of $30,000 to $40,000, a troublesome figure given that less than 15 percent of California homeowners have earthquake insurance.
The hardiness of these wood-frame homes is about the only good news out there. Renters will have a much bigger problem on their hands. The Bay Area is riddled with "soft-story" apartment buildings, which contain large, open areas such as parking garages or retail space on the ground floor. They are cheap to build and rent, making them the housing of choice for students and working-class families. "It was a way to build multifamily buildings in the 1960s and accommodate parking," Comerio says. "In an urban setting, it seemed like a terribly reasonable solution to parking."
Here's the problem, Perkins says: "These have been known to be hazards for years. In earthquakes, the first floor pancakes. It happened in San Fernando and in Northridge and every big California quake in between, and a minuscule percentage of these have been retrofitted."
Largely due to the prevalence of such buildings, the Association of Bay Area Governments estimates that a 6.9 quake along the Hayward Fault would leave 156,000 housing units uninhabitable; a 7.3 on the San Andreas would wreck 110,000 housing units. Following a 7.0 Hayward quake, Comerio estimates that the occupants of 150,000 to 200,000 housing units will be looking for a new place to stay, some simply because they are afraid to go home. But there will be nowhere to go. "How do you rehouse low- and moderate-income populations in a building stock for which there are not going to be many vacancies or options?" she asks. "Where is everybody going to go? Are they all going to move to Manteca? Because there isn't any affordable housing in the Bay Area already."
The researchers' nightmare scenario is a big quake that strikes in the middle of a workday. Gregory Fenves is especially worried about nonductile reinforced concrete -- a material used in older Bay Area commercial office buildings, dorms, and other high-rise structures. "Any concrete building in the Bay Area that was built before about 1975 is likely a hazard," Fenves says, because with strong shaking, such buildings are prone to what he describes as a very fast, brittle collapse. "Typically, lower floor columns will sort of shear and once they move sideways ... a building just loses a story and it sort of drops down."
Strong shaking also is expected to devastate hospitals and the older concrete office buildings that house clinics and other medical services -- among the services that will be most critical following a major quake. "Most of the major hospitals in the East Bay seem to be within a mile of the Hayward Fault, so we're very concerned about how they would survive," says Susan Tubbesing, executive director of the Earthquake Engineering Research Institute in Oakland.
A report produced by California's Office of Statewide Health Planning and Development offers a shocking look at the hospital industry's lack of readiness. For example, at the Alameda County Medical Center's Highland campus, four of the five buildings were given the lowest safety rating, indicating "significant risk of collapse and danger to the public." Eleven of the twelve buildings on Kaiser's Oakland campus got the same low rating, as did six of ten buildings that make up Eden Medical Center in Castro Valley and four of the seventeen buildings at the Mount Diablo Medical Center in Concord. Of the 163 buildings that make up the major hospital facilities in Alameda and Contra Costa counties, only 21 were deemed "reasonably capable" of providing services to the public after a large quake. The rest will likely be overwhelmed, at best able to handle the most critically injured patients. "There is nothing worse than seeing a hospital in an earthquake zone that can't be used because it's collapsed or partially collapsed," Fenves says.
These hospitals will certainly be in demand if the eastern span of the Bay Bridge repeats its Loma Prieta performance, when a fifty-foot section of the upper deck flopped onto the lower deck. The eastern span is considered a significant collapse risk, yet despite its critical economic role -- an average of 280,000 vehicles pass over it daily -- efforts to replace it recently ground to a halt over funding issues. Despite some interim retrofitting, engineers widely consider the span not much better off than it was prior to Loma Prieta. It is simply an old bridge, they say, built in the 1930s before computer modeling allowed engineers to simulate the effects of strong ground motion. It's a riveted steel structure with a complex design and many small parts, without much built-in redundancy were some of these parts to fail.
"The way that the steel is fabricated and placed in an intricate geometrical design, you would never see in modern bridges," says Bruce Bolt, an emeritus professor at Cal's civil and environmental engineering department who has served as an ad hoc adviser to Caltrans. "Under earthquake loads, each of these small elements that make up the complicated structure you see now are liable to failure. This is why much simpler structures such as suspension bridges are now used around the world."
Engineers emphasize that Loma Prieta, even from a considerable distance, had the power to dislodge some of the four-inch supports that held the eastern span's deck sections in place. Longer supports have since been installed, Bolt says, but only as a stopgap measure; the fixes are not guaranteed to prevent collapse in a large local quake. The bridge's single biggest vulnerability, Fenves adds, is likely its foundations, which he calls "very, very small." Were the foundations to be damaged in a 1906-scale quake, Bolt concurs, it could easily take six months to reopen the bridge.
BART may have similar problems with the foundations of the nearly two thousand columns that support its aerial tracks. "The primary concern is that the foundations are just small enough and weak enough that they'll rock and they'll break," says Thomas Horton, group manager of BART's earthquake safety program. "None of our scenarios show structures collapsing, but they end up leaning quite a bit. It's not safe for our patrons."
The transit agency has itself a bit of a dilemma -- a single quake is unlikely to break more than a hundred or so of those columns, but since its track network is so large and comes close to so many different faults, there's no way to predict which columns are most at risk. As a result, BART plans to retrofit them all.
The Earth Ruptures
Ordinary Bay Area roadways may not fare much better thanks to another earthquake hazard called surface fault rupture. Most local faults are known as "right lateral strike-slip faults," which, to a driver passing over the fault as it ruptures, means the road ahead will seem to move abruptly to the right. Scientists estimate a quake on the Hayward Fault could cause three to nine feet of displacement, enough that on some roads the lanes would no longer line up at all.
The Association of Bay Area Governments estimates that some 1,700 roads would be closed following a 6.9 quake on the Hayward Fault. "Every road that crosses the fault would be closed -- you'd have every street from Milpitas to San Pablo Bay," Perkins says.
BART tracks also are vulnerable to this phenomenon -- particularly the system's Berkeley Hills tunnels, which span the fault. Any significant ground displacement would put a kink in the tracks that would make the two single-train tunnels impassable. The normal solution in a case like this, says BART's Horton, would be to widen the tunnels before a quake hits. That way, one tunnel could be made to accommodate a slow-moving train while the other underwent permanent repairs. But BART ran into a sticky situation with the Berkeley Hills bore, Horton says. While each tunnel has only six to eight inches to spare, the soil within them is so unstable from thousands of years of seismic shifting that it would be unsafe to attempt a widening while trains are in operation. As an alternative, BART considered rerouting the tracks around the tunnel, but that idea was a bust. "We couldn't maintain a slope that BART could accept," Horton explains. "It was hugely expensive, of course, and by the time we were done we just concluded that there was no practical way to retrofit. So what we have concentrated on is a response plan to rapidly get the tunnel back into service."
Rapid, in post-earthquake parlance, is a relative term. According to a study conducted for BART by the Bechtel Corporation, it would take up to five months just to reopen one of the tunnels for slow-moving trains postquake, and some 28 months to get both tunnels up to speed. In the meantime, many of the 45,000 people who ride BART through those tunnels daily would need a new way to get around.
Ground displacement could create particular chaos with the countless underground gas, sewer, oil, and water lines that span local faults. Simply put, they'll all be ripped in half, leaving thousands without water, gas, and other services. Tubbesing of the Earthquake Engineering Research Institute says people should expect to cope without power and running water for at least 72 hours after a major quake. Even worse, ruptured gas lines have been known to spark fires after earthquakes.
The sudden loss of basic utilities and the prospect of gas fires may cause some panic. Perkins worries residents will desperately try to escape the East Bay by forging the broken roads in their SUVs. "All the water lines are going to be damaged, and the gas lines," she says, "and when you have somebody in their Hummer trying to go across this muck they're probably going to break a gas, water, or sewer line and they're going to make it worse. It's going to be harder for the crews to repair them. It's going to be a mess."
The Earth Turns to Mud
If it's hard to imagine solid ground turning to liquid, Keith Knudsen of the California Geological Survey has a familiar metaphor: "Have you ever jumped up and down on the sand at the beach just after a wave comes?" he asks. "You do that and the sand begins to flow. It loses its ability to support you and flows like a liquid, and all you're doing is adding vibrations or pressure waves to that mix of sand and water."
The same thing happens when an earthquake pounds on relatively loose soils close to the water table -- ground that seemed solid prior to the quake begins to flow, and can sink and pull apart. It's a danger intrinsic to riverbanks and shorelines, but is likely to cause problems wherever the San Francisco Bay has been artificially filled in to create new land. Some 77 square miles of Bay Area real estate are built on fill, including Treasure Island and half of Alameda, San Francisco's Marina and Mission districts, and many of the waterfront commercial warehouse districts on the East Bay side from Richmond down through the South Bay. "Back in the early 1900s they would build a wall out into the bay a little bit, a dike or something, and then they would just take a suction hose and put it down on the sea floor on the bay side of the wall and pump sand to the other side," Knudsen explains. "That sand they would let dry out over time. They didn't compact it or put drains in, and they would build on top of it after a few years."
Liquefaction causes damage in several ways: Part, but not all, of the ground beneath a building can liquefy, leaving it askew. Or the liquefied earth can ooze outward, buckling structures such as roads or sidewalks that were anchored in the soil. The ground can also simply lose its bearing strength, causing structures to sink. During the Loma Prieta quake, liquefaction cracked three thousand linear feet of runway at Oakland International Airport; some of the cracks were a foot wide and a foot deep. And sand boils -- much like the sand bubbles that form on a beach when the surf recedes -- up to forty feet across surged up beside the runways. In addition, parts of the Port of Oakland sank by up to a foot.
Liquefaction changes the density of the soil surrounding buried pipelines and tanks, causing these structures to float upward. This can be disastrous: sewer, water, or natural gas lines may snap as they are pulled toward the surface, and gas station fuel tanks can bubble up from the ground, creating an imminent fire hazard.
The most dramatic and immediately deadly consequence of liquefaction could involve the Transbay Tube that carries BART trains under the bay between Oakland and San Francisco. If the soil around it were to liquefy in a major quake, the tunnel -- essentially a huge pipe -- could very well try to float. BART's Horton points out that the tube consists of 57 welded sections; the junctions between them are slightly weaker than the rest of the tube and could potentially crack or break during an uplift. In another scenario, the tube could pull free of the seismic joints that anchor it at either end, breaking the gasket that seals the tube from the outside. In either case water could rush in, swallowing entire trains along with their riders.
At best, such a rupture would seriously endanger anyone stuck in a BART train under the bay -- during peak hours, when several trains navigate the tube at any given time, that can mean thousands of people. Also, since the first few stations on the San Francisco side of the bay are below sea level, water could inundate the entire system as far as the 16th Street or 24th Street stations. "The entire downtown Market Street area would be flooded and the MUNI level would probably also be flooded," Horton says. "It wouldn't happen instantaneously, but it's not at all clear how fast it would happen."
While liquefaction is the bane of the flatlands, East Bay hill dwellers will have their own private nightmare. Already, the hills are undergoing a slow, perpetual landslide that has gradually pushed some people's houses across their neighbors' property lines. All it will take is a big seismic nudge to radically speed up that process. "With every wet season they may move a few inches, but in an earthquake they might move twenty or thirty centimeters," Knudsen says. "Those are not catastrophic things; they don't kill a lot of people, but they do a lot of damage. They rack and tilt houses and break pipelines and damage roads."
If a quake strikes during the wet season, however, it can lead to something more disastrous called a "debris flow." This is a highly fluid and faster-moving downhill surge strong enough to pick up heavy objects. "It's like a very muddy river," Knudsen explains. "It moves rapidly, and because it moves fast there are rocks and things that get caught up in it. They can damage things, and it can fill up the lower levels of houses and engulf automobiles."
The slides also can tear chunks out of roadways that run along the coastline, rendering them impassable -- think Devil's Slide -- and destroy pipes or conduits running beneath the roadway.
Timing Is Everything
With the prospect of death and destruction underscoring their work, people like Lettis and Schwartz are surprisingly cheerful when they talk about the future -- they have great faith that, given advancements in earth science and construction techniques, we will eventually be able to out-engineer the worst effects of anything nature might throw our way.
As evidence that modern science really does make a difference, GeoHazards International, based in Palo Alto, recently estimated that better engineering and enforcement of building codes has reduced the earthquake death toll by a factor of ten in industrialized nations over the last half century. The 2003 magnitude 6.6 quake in Bam, Iran, for instance, killed more than thirty thousand people and damaged or destroyed 85 percent of the area's buildings. The following year in Northridge, California, a larger quake killed only dozens. "One hundred years from now as society moves forward, we still probably won't be able to predict when an earthquake will be," Lettis says. "But we will have improved our building response and emergency response such that it doesn't matter when the earthquake happens; we're ready for it and we'll minimize loss from it."
The big caveat, as Schwartz points out, is we don't have a hundred years to spare -- the seismic time-out provided by the 1906 quake could already be over. "Even though we may be in the stress shadow, the sun is coming out tomorrow," Lettis says.
"I think the question," Schwartz concludes, "is do we have enough time to do what has to be done?"
Seven Days - December 9, 6:10 PM
Seven Days - December 8, 2:33 PM
Seven Days - December 6, 9:52 AM
Seven Days - December 4, 9:20 PM
Seven Days - December 4, 5:31 PM