Lessons from the San Francisco Earthquake

From the local vista point known as Twin Peaks, Mary Lou Zoback, a senior scientist with the U.S. Geological Survey (USGS), looks out on a breathtaking view of San Francisco–the gilded dome of City Hall, the diagonal stripe of Market Street, the little neighborhoods marching up and down steep hillsides. Slowly she pivots, taking in the sailboats on the bay, the Golden Gate Bridge, the shimmering surface of the Pacific Ocean. Just out there–she points–a couple of miles offshore, lies the place where, early in the morning of April 18, 1906, the earth’s crust cracked like an eggshell, unleashing what–even in the aftermath of 9/11 and Hurricane Katrina–stands as one of the greatest disasters in U.S. history.

Trees were “lash’t as tho by a gale,” bystanders reported, and fields undulated “like the waves of an ocean.” Buildings swayed, clocks stopped, church bells rang, water mains burst, gas lines broke, electrical wires snapped and sparked. Then came the flames, which for three days burned out of control as firefighters stood helplessly by.

Today, 100 years later, the damage that resulted from the great quake seems nearly as shocking as it did then: some 28,000 buildings destroyed, more than 3,000 people killed, at least 225,000 more–roughly half the population of the city of San Francisco at the time–left homeless. But, more shocking still, was the fact that no one, not even scientists, could explain why, without warning, such fury had erupted from the earth below.

That quickly changed, however, as geologists, led by Andrew Lawson of the University of California, Berkeley, raced into the field, making observations that established the existence of a fault line that parallels the California coast for more than 700 miles. They named the fault the San Andreas, after a jewel-like lake that lay within the rift zone less than 10 miles south of what was then America’s largest and richest Western city.

In a two-volume report published in 1908, Lawson and his team went on to elaborate a new model of earthquake formation–the elastic-rebound theory–that holds up to this day. For years, they correctly surmised, stress had been ratcheting up along the San Andreas until finally it became so overwhelming that the earth’s crust snapped like an overextended rubber band. Moreover, the buildup and release of strain appeared to be recurrent, resulting over time in a succession of earthquakes “of greater or less violence.” These pioneering researchers provided the first big clue that earthquakes occur in cycles–that in the area around San Francisco Bay, earthquakes are as certain, if not as regular, as the seasons.

It’s this certainty that lends urgency to the efforts by Zoback and her colleagues to remap the San Andreas and its subsidiary faults, to amass new clues to its murky prehistory and to re-create in cyberspace the primordial violence of the 1906 quake. In addition to being the centennial of the last Big One, April 18, 2006, marks the approximate midway point in the countdown to the next Big One–100 years of stress accumulation in one of the world’s most earthquake-prone regions. The more scientists learn about the ways in which that stress may be released, the more ominous the next earthquake cycle seems.

From the air, the San Andreas stands out as a linear gash in the earth’s surface that is easy to spot. On the ground, however, it is often hard to read, particularly north and south of San Francisco where it strays offshore, runs through dense redwood forests and even disappears beneath houses and streets. In many populated areas, it’s impossible to tell just where the active strands of the fault lie because so many features have been filled in or bulldozed away.

Thanks to the efforts of USGS paleoseismologist Carol Prentice and her colleagues, however, residents of the Bay Area will have a much better sense of the precise path the earthquake took. Working with old photographs, Prentice has found a number of the missing signs of 1906–abrupt jogs in fences that once straddled the rupture zone, for example–and located them on aerial photos. Among the communities bisected by the fault break is San Bruno, a city of 40,000 that borders San Francisco international airport.

Luckily there are not too many structures located within the strip, about 1,000 ft. wide, that defines the San Andreas Fault zone. The same cannot be said about the nearby Hayward Fault. Along with the Calaveras, San Gregorio and Rogers Creek faults, the Hayward forms part of what scientists refer to as the San Andreas system, and it runs for 60 miles along the hills of the East Bay, cutting through the University of California, Berkeley, football stadium and skimming uncomfortably close to the Caldecott Tunnel, through which 153,000 cars pass daily. Major highways, including Interstate 80, cross the Hayward Fault, as do the pipelines that bring water down from the snow-clad Sierra. There are hundreds of privately owned structures in the fault zone, virtually all built before the state passed a tough earthquake-zoning law in 1972.

The hazards posed by earthquakes do not stop at the fault zone. Most of the damage caused by a quake comes not from the rupturing of the ground underfoot but from seismic waves that propagate out from the fault at 8,000 or more m.p.h. While the punch packed by these waves tends to diminish as the distance from the fault increases, that’s not always the case. From historical accounts, USGS seismologist Jack Boatwright has assembled a ShakeMap for 1906–a map that displays the intensity of shaking in different areas. For San Francisco and other communities close to the San Andreas, it was quite severe. But even more severe was the shaking that occurred in the city of Santa Rosa, more than 15 miles away from the fault. On a scale of 1 to 10, Santa Rosa stands out as a 9-plus, somewhere between “violent” and “extreme.”

Why did this area get slammed so hard? At least part of the answer lies in the loosely consolidated sediment that sits below the surface. Seismic waves pass quickly through bedrock, but they become trapped in sediment-filled basins. “It’s sort of like being in a bathtub filled with water,” says USGS seismologist Thomas Brocher. “When you start splashing, the waves keep bouncing up and down and from side to side.” The basin effect amplifies not only the intensity of the shaking but also its duration, which is no doubt why buildings collapsed in Santa Rosa in 1906, killing some 100 people. There are similar sedimentary basins throughout the Bay Area–around the Silicon Valley city of Cupertino, for example, and the expanding subdivisions that surround the Lawrence Livermore National Laboratory.

The biggest basin lies well east of the Bay, in the broad delta formed by the convergence of the Sacramento and San Joaquin rivers. Among the most catastrophic consequences of a big earthquake in the Bay Area, says University of California at Davis geologist Jeffrey Mount, would be the failure of the delta’s aging levee system, which protects not just farmland and residential areas but also the water supply for some 23 million people. Shaken hard enough, the foundations of the levees would crumple, and in a kind of hydrological chain reaction, brackish water from the Bay would surge inland, contaminating the freshwater that aqueducts carry all the way to Los Angeles.

Earthquakes, scientists now know, occur along the San Andreas because the immense slabs of rock that make up the earth’s crust are ever so slowly sliding past one another, borne by poorly understood currents that roil through a sea of semimolten rock. By keeping tabs on the position of key landmarks on either side of the fault, scientists can measure the speed at which the plates are traveling, in this case about 2 in. a year. The problem for the Bay Area boils down to this: except for one short section, the plates on either side of the San Andreas are tightly locked together. It’s only when the stress becomes overwhelming that the San Andreas breaks apart, allowing the plates to lurch forward, 10 ft. to 20 ft. at a time.

In principle, this cycle of stress accumulation and release should be fairly regular, but scientists are finding it is not. Paleoseismologist Tina Niemi of the University of Missouri–Kansas City, for example, is studying a stream-fed marsh near Tomales Bay that has preserved evidence of past earthquakes in its sedimentary layers. By trenching through those layers to a depth of 15 ft., she has uncovered buried fissures formed by recurrent earth movements along the San Andreas. On average, that pattern repeats every 250 or so years, but “average” in this case covers a wide range. In one instance there appears to be a 600-year interval between quakes, in another just five decades.

Already, scientists say, there is a greater than 60% probability that one or more of the faults in the San Andreas system will unleash an earthquake of magnitude 6.7 or higher over the next three decades, and among the most likely candidates is the Hayward Fault. The last big earthquake on the Hayward occurred in 1868; it caused so much damage that it was known as the great San Francisco earthquake until 1906 displaced it. “The Hayward Fault is locked and loaded,” says Brocher, “and it could fire at any time.”

What will happen when the Hayward Fault–or the San Andreas–goes off? Scientists who study ancient quakes cannot answer that question because it depends on details that sediments do not preserve. But using a new 3-D model of the earth’s crust in the Bay Area, USGS geophysicist Brad Aagaard and his colleagues can run simulations that tweak different parameters for earthquakes that have already occurred and for those still to come. The results range from the expected to the quite surprising.

At his computer, Aagaard first conjures up the 1989 Loma Prieta earthquake, which started, many scientists think, along a spur of the San Andreas some 60 miles south of San Francisco. Across a Landsat image of the Bay Area, Aagaard’s simulation takes the form of a spreading blob of mixed colors that indicate shaking intensities, from low-intensity blue to medium-intensity yellow and high-intensity red. Then Aagaard calls up 1906. The difference is immediately apparent. This time red flows across the landscape like a river of lava, and among the places that glow the brightest is the area around Santa Rosa, just as the ShakeMap says it should.

Aagaard and his colleagues have started using their earthquake simulator to try to answer the most tantalizing questions of all: What if the rupture of the fault had not started directly off the San Francisco coastline? What if it had started farther south, so that instead of breaking away from the city it had aimed right toward it? What if it had started farther north and broken south? In the first instance, the tentative answer is that San Francisco gets shaken even harder; in the second, it’s Silicon Valley and the Livermore Valley that find themselves clamped in the lion’s jaws. “1906 is the most powerful earthquake we can imagine hitting Northern California,” says Mary Lou Zoback, head of the USGS Northern California Earthquake Hazards Program. “But it may not have been the worst-case scenario.”

Concerned about what their research is showing, Zoback and her colleagues have redoubled their efforts to raise public awareness of the hazard that lurks below. Later this month their voices will be reinforced by the more than 2,000 scientists, engineers and emergency managers gathering in San Francisco for a special 100th Anniversary Earthquake Conference.

The question is: Will Bay Area residents pay attention to what these public-spirited researchers say? The ghost of Hurricane Katrina, no less than that of 1906, will haunt the centennial as it gets under way. “Katrina has shown us what a $100 billion–plus disaster looks like, the kind of disaster no one wanted to talk about before,” says Chris Poland, chief executive of Degenkolb Engineers and chairman of the conference. “It’s shown us what happens when you damage a community so much that its economy stops.”

While it is true that the communities in and around San Francisco have taken a number of laudable steps–constructing a whole new span for the San Francisco–Oakland Bay Bridge, for example–it is also plain that they need to do more. There are tens of thousands of older buildings in the Bay Area that do not meet modern earthquake standards, among them office and apartment buildings whose upper floors rest atop an unreinforced storefront or garage. In an earthquake, such “soft-story” buildings are likely to collapse or sustain damage so severe that no one will be able to live or work in them.

That is what the flooding from Katrina did to New Orleans, and the vividness of what it means to a modern city to lose so much housing and so many jobs has given the 1906 centennial a somber emotional edge. At risk in this case is not just a very large metropolitan population–the Bay Area now has about 7 million residents versus perhaps 800,000 in 1906–but also a vibrant $350 billion economy that includes one of the nation’s largest financial hubs, one of its busiest ports and one of the world’s densest concentrations of technical and scientific talent.

Time, unfortunately, is not on the Bay Area’s side. Scientists say the “shadow” of the 1906 earthquake–a kind of protective umbra generated by the enormous release of stress 10 decades ago–is already beginning to dissipate. That means the Bay Area will soon be rocked by the next cycle of seismic unrest, with smaller but still damaging earthquakes signaling the start of a new era of danger for a city that’s had more than its share.