Election Section

Stanford scientists find faults in earthquake prediction model

By Angela Watercutter
Saturday September 21, 2002

SAN FRANCISCO — A new study by two Stanford University scientists could shake up a long-held theory that helps geologists forecast earthquakes. 

The theory holds that earthquakes are “time-predictable,” meaning the energy buildup that causes them happens on a somewhat regular basis. But the scientists found that wasn’t the case in a rather geologically simple area of the San Andreas fault near Parkfield. 

Stanford geophysicists Jessica Murray and Paul Segall say in the Sept. 12 issue of Nature that their research in Parkfield shows the time prediction model failed in what should have been an ideal locale for its application. 

“I think we’ve really shown that quantitatively this model doesn’t work at this location,” Murray said. “I think that this will lead people to realize how uncertain this model is.” 

The Stanford team’s findings make the successful application of the theory even harder in complex fault regions such as the San Francisco Bay area. 

Scientists in Parkfield have been waiting for a substantial earthquake since 1988, when time prediction said the area was due for a quake. Midsize earthquakes have hit the area roughly every 22 years since 1857. 

The Stanford researchers used a long history of data from the Parkfield site, as well as data on how much the earth had moved. They found there was enough energy for an average-sized quake, and showed that effects from nearby temblors had passed. 

No quake came, leading them to believe the time prediction theory didn’t work. And the current lapse of time between earthquakes is one of the longest in recorded history, according to Murray. 

However, it could just be that Parkfield isn’t as timely as scientists had originally thought, and now the site is showing it can be as irregular as anything else in nature, according to Ross Stein, of the U.S. Geological Survey in Menlo Park. 

“On the one hand it’s almost one complete cycle late, yet we have to acknowledge we see that kind of variability everywhere,” Stein said. 

Earthquakes are caused by the constant scraping of the North American and Pacific continental plates. When the ground gets stuck it jerks loose in sudden bursts that shake the ground. What scientists still hope to determine is how to predict when they’ll occur. 

Even if the Stanford team’s findings lead to dismissal of the time-prediction theory, there are still other theories scientists can investigate and use. Some attempt to predict the size of the shake, but not when it will happen. Others look at how earthquakes interact to anticipate where the next section will break. 

“It’s an important study, and every time we can test something we assume we’re ahead of the game no matter what the consequences,” Stein said.