San Andreas May Have Had Help Triggering a Historic Earthquake

New evidence about a devastating quake in 1812 hints that the San Jacinto fault may be a bigger seismic risk than anyone thought

Great Stone Church
A woman sits amid the ruins of the Great Stone Church, which collapsed in an 1812 earthquake, at Mission San Juan Capistrano. courtesy Flickr user J Jakobson, CC BY-NC 2.0

On December 8, 1812, an earthquake shook the Spanish mission of San Juan Capistrano in southern California, toppling buildings and killing 40 people attending mass at the mission. That doesn't come as a big surprise in the history of a region known for plentiful and powerful temblors, and this particular event, estimated as a magnitude 7.5, was long thought to be yet another product of the infamous San Andreas fault.

But now scientific detective work has revealed that the 1812 earthquake might have been the result of two faults acting together—and that means the people of southern California are on shakier ground than anyone thought.

Julian Lozos, an assistant professor of geophysics at California State University, Northridge, built a computer model of the San Andreas fault and the adjacent San Jacinto fault, centered on the region around San Bernardino. Coupled with geologic signs of past earthquakes and historical records, his model shows there's a good chance the 1812 quake started along the San Jacinto fault, and the energy from that initial shaking caused the nearby San Andreas to rupture as well in a kind of cascade effect.

"The implications stretch just beyond this one damaging earthquake," he says. "The fact that the effects of this historic earthquake can be explained by the San Andreas and San Jacinto working together means that this is, at the very least, a physically plausible thing"—and that it could happen again.

That would be a disaster for the cities of San Bernardino and Riverside, which sit right on top of the area where the two faults come close together. The San Jacinto comes within a mile of the San Andreas at Cajon Pass, where a major highway, Interstate 15, passes through. Hazard maps from the city and county show that I-15 runs right over a region at high risk of liquefaction, when the ground essentially turns to mush during a quake.

Combination earthquakes aren't necessarily more powerful than single-fault ones, but they do travel in different ways. Instead of zipping relatively neatly along the fault line under San Bernardino, a multi-fault earthquake—even a less powerful one than the 1812 temblor—could jump right across a very densely populated region, causing even more damage than anything the San Andreas could produce alone.

"A San Andreas-San Jacinto joint 7.5 rupture is scarier, because more of the fault goes through a more densely populated area than the southernmost San Andreas does," Lozos says.

Earthquakes in that part of California are mostly caused by strike-slip faults, where two big chunks of Earth's crust are sliding past one another. In this case, the Pacific plate is moving roughly north past the North American plate. Since the faults aren't perfectly smooth, the two pieces of crust catch on each other (the strike) and once enough tension builds, they release suddenly (the slip). That release is what we feel as an earthquake.

If two faults are close enough, a rupture in one can trigger a rupture in another. This observation isn't new—the 1992 Landers earthquake reached magnitude 7.3 after multiple faults ruptured.

"It zigzagged between six or seven different faults," Lozos says. Luckily that quake was centered in the Mojave Desert, and the nearby towns of Yucca Valley, Joshua Tree and Twentynine Palms are relatively small. The question was whether the same thing could happen along the San Andreas and its subsidiary faults, such as the San Jacinto, affecting much more populated areas.

Earthquake Records Map
A map indicates where people made historical records of the December 1812 earthquake in southern California. Lozos Sci. Adv. 2016; 2 : e1500621

Lozos started by looking at existing data about earthquakes in the past, including changes in geologic layering that indicate when and where older quakes happened. Faults aren't always continuous; they can be made up of several sections, called strands, that are separated by short bits of intact crust. Studying how sediment layers have shifted around these strands can reveal whether they were involved in an earthquake.

Critically, Lozos found geologic data for three strands—two on the San Jacinto and one on the San Andreas—that showed evidence of movement in the 19th century. However, accounts from the period only tell of two major quakes, the one in December 1812 and another on November 22, 1800. That suggests one of those quakes had "jumped" between the fault strands.

Lozos also looked at earlier studies of precariously balanced rocks conducted by Jim Brune of the University of Nevada Reno and Lisa Grant Ludwig at UC Irvine. Factoring in the shape of the rocks and structure of the pile, certain kinds of shaking will topple these natural structures. Looking for balanced rocks that are still standing shows where past earthquakes didn't happen, helping to narrow down the regions where the two 19th century quakes occurred.  

Lozos then created a computer model based on the physics of the faults around San Bernardino, incorporating data such as the characteristics of the rock. He input various initial conditions until he got a simulated earthquake that produced the same effects as the ones he observed in the gathered data. The most plausible way to produce a magnitude 7.5 earthquake that damages buildings in the right pattern was if the San Andreas and San Jacinto ruptured together, he reports this week in Science Advances.

One reason nobody has really studied this phenomenon in the San Andreas is that it's so big relative to all the other faults in the state, Lozos says. Generally the assumption has been that big temblors are coming from the big fault.

David Oglesby, a professor of geophysics at the University of California, Riverside, says the model Lozos designed is believable because it works under many different scenarios. "You could make a model do anything if you put in the right assumptions" Oglesby says. "But this one works without too much fine tuning."

The model also agrees with geologic data spanning centuries, says Nate Onderdonk, associate professor of geoscience at California State University, Long Beach. In his studies of the northern part of the San Jacinto fault, data shows that there was not only a seismic event there in the right time frame, the early 19th century, but that it was bigger than what could have been contained in one section of the San Jacinto by itself.

Onderdonk adds that he is submitting an independent study that shows this has happened several times in the past two millennia—adding to evidence that a devastating joint quake could happen again in the future.

Learn more about this research and more at the Deep Carbon Observatory.

Get the latest Science stories in your inbox.