The San Andreas Fault Is Primed for Central California's Next Big One

It remains impossible to "predict" earthquakes, but we can increasingly make estimations about how well primed a given patch of geography is for an earthquake to happen.

A study out this week from researchers at the United States Geological Survey cautions that the San Andreas fault, perhaps the most notorious or at least well-recognized fault in the United States, has built up a dangerous level of earthquake potential, accumulating a large "seismic moment" that has gone unrelieved by the normal process of fault slippage.

Fault slippage is more properly known as fault creep. Faults, of course, are cracks in the giant slabs of rock that make up the Earth's crust—sometimes they're just cracks, but sometimes they form the boundaries between the giant, distinct plates that make up the different pieces of the puzzle that is Earth's crust.

The San Andreas fault is in the latter category, dividing the Pacific Plate from the North American Plate. These plates "float" on top of the much softer, highly viscous geological layer below, called the asthenosphere. As these plates float, they also move in relation to each other; the Pacific and North American plates are moving in opposite directions from each other at about twice the speed at which your fingernails grow.

As these plates slip by each other, a couple of things happen. For much of the time, it's just a slow grinding, known as aseismic creep. This is where the plates move around without causing an earthquake.

Along the San Andreas fault, which is really more of a network of cracks shooting off from one big one, usually around 21 percent of plate movement occurs aseismically, according to the study. Where the fault has big bends and kinks, it's a whole lot harder for creep to occur; suddenly the plate isn't just pushing along its neighboring plate, it's pushing against it as well, like a locked puzzle piece.

The creep alternative is indeed known as "locking." Rather than push along, the plates get hung up on each other, snagged. The plates want to keep moving, however, so where the two are hung up against each other, pressure builds. This building pressure is what's known as a seismic moment. You can probably guess the outcome of this situation: an earthquake. As such, assessing locking relative to creeping is a crucial way in which geologists can make seismic hazard assessments.

For a fault line, there are few other options besides everyday creep and full-on earthquakes; mostly, a locked plate is a locked plate. "It is possible that creep or another form of aseismic slip can occur to release strain in a way that is in addition to our model, which accounts for faults that are creeping more or less steadily," James Lienkaemper, the current study's lead author, told me. "Those means include afterslip following earthquakes and large anomalous creep events."

"We have a number of examples of anomalous creep events, which usually are not very significant in terms of strain accumulation," Lienkaemper said. "However, for the central Calaveras Fault [a branch of the San Andreas fault], and maybe the central creeping section of the San Andreas, such events may be quite important."

Right now it's a bad scene along the San Andreas, which extends along nearly the entire coastline of California. The current study, which is based on data taken from 168 different seismic monitoring and GPS instruments spread around the San Andreas, mostly looked at the fault as it passes through the densely populated areas around San Francisco Bay. While some less bendy parts of the fault are minimally locked, other portions are nearly 100 percent locked.

"From its mean creep rates, we infer the main branch (the San Andreas fault) ranges from only 20 to 10 percent locked on its central creeping section to 99 to 100 percent on the north coast," the study reports. "From mean accumulation rates, we infer that four urban faults appear to have accumulated enough seismic moment to produce major earthquakes: the northern Calaveras, Hayward, Rodgers Creek, and Green Valley. The latter three faults are nearing or past their mean recurrence interval."

It's important to note that this isn't an earthquake forecast, but a stern warning (and hardly the first of its kind for the San Andreas fault). "There are no reliable methods to predict earthquakes yet," Lienkaemper cautioned. "Creep doesn't help to predict them either. Modeling the amount of strain that is relieved by creep is a practical way to estimate how much strain is stored that may be released seismically." Figure it's a countdown clock without numbers. The inevitability is there, growing stronger everyday, but when it'll actually hit zero remains unknown.