The southernmost part of the fault had its last earthquake sometime around 1680. We know this because it offset the edges of Lake Cahuilla, a prehistoric lake in much of what is now the Coachella Valley, filling with water the flats where the Coachella music festival meets each year. It left behind geologic markers, as did previous earthquakes, so we know that there were six earthquakes between AD 800 and 1700. That means the 330 years since the last earthquake on this part of the San Andreas is about twice the average time between its previous earthquakes. We don't know why we are seeing such a long interval. We just know that plate tectonics keeps on its slow, steady grind, accumulating more offset and energy to be released the next time. Since the last earthquake in Southern California, about twenty-six feet of relative motion has been built up, held in place by friction on the fault, waiting to be released in one great jolt.
Someday, maybe tomorrow, maybe in a decade, probably in the lifetimes of many people reading this book, some point on the fault will lose its frictional grip and start to move. Once it does, the weak fault, with all that stored energy, will have no way of holding it back. The rupture will run down the fault at two miles per second, its passage creating seismic waves that will pass through the earth to shake the megalopolis that is Southern California. Maybe we will be lucky and the fault will hit something that can stop it after only a hundred miles or so—a magnitude 7.5. Given how much energy is already stored, however, many seismologists think it will go at least two hundred miles, and thus register 7.8, or even 350 miles and reach 8.2.
If it ruptures as far as central California, all the way to the section of the fault near Paso Robles and San Luis Obispo, it will hit a part of the San Andreas that behaves differently. This part accumulates a fingernail-growth rate of tectonic offset, just like the rest of the fault. But it's what is known as a "creeping section." Instead of storing energy to release in one big earthquake, the energy here oozes in small motions, sometimes with little earthquakes, sometimes with no seismic energy at all. We think, we hope, that the creeping section will act as a pressure valve of sorts, keeping the earthquake from growing any bigger than 8.2.
In 20078, as science advisor for risk reduction at the U.S. Geological Survey, I led a team of more than three hundred experts in a project we called ShakeOut, to anticipate just what such an earthquake will be like. We created a model of an earthquake that moves across the southernmost two hundred miles of the San Andreas, extending from near the Mexican border to the mountains north of Los Angeles—a likely outcome, though still short of the worst-case scenario.
In the earthquake we modeled, we found that Los Angeles would experience intense shaking for fifty seconds (compare this to the seven seconds of the Northridge earthquake in 1994, which caused $40 billion of damage). A hundred other neighboring cities would as well. Thousands of landslides would cascade down the mountains, blocking our roads, burying houses and lifelines.
In our model, fifteen hundred hundred buildings collapsed and three hundred thousand were severely damaged. We know which ones. They are the types of buildings that have collapsed in other earthquakes in other locations, and which we no longer allow to be built. But we have not forced existing buildings to be retrofitted to accommodate what we know. We might see some high-rise buildings collapse. The 1994 earthquake in Los Angeles and the 1995 earthquake in Kobe, Japan, exposed a flaw in how steel buildings had been constructed, causing cracks in their steel frames. Those buildings are still standing in downtown Los Angeles. We are going to see many brand-new buildings "red-tagged," too dangerous to enter and in need of major repairs or demolition. Our building codes do not require developers to make buildings that can be 'used' after a major earthquake, only buildings that don't kill you. If the code works as it is supposed to, about 10 percent of the new buildings constructed to the latest code will be red-tagged. Maybe 1 percent will have partial collapse. A 99 percent chance of not collapsing is great for one building, but accepting the collapse of 1 percent of the buildings in a city with a million buildings is a different matter. The earthquake will probably not kill you, but it will likely make it impossible for you to get to work—for a very long time.