Science@Berkeley Lab nameplate Berkeley Lab logo
February 18, 2005
An Inside Look at the San Andreas Fault

Berkeley Lab scientists have developed a better way to eavesdrop on one of California's crankiest and least understood inhabitants. They've constructed CT-scan-like images of the San Andreas Fault that reveal nucleation zones of past quakes and areas where stress continues to build — at a resolution that is ten times greater than other images.

"Our high-resolution images could allow us to watch stress evolve deep within faults like never before," says Valeri Korneev of the Earth Sciences Division (ESD), who developed the imaging method with fellow Earth Sciences Division scientists Robert Nadeau and the renowned seismologist Thomas McEvilly, who died in 2002.

Their newly developed technique is among the bonanza of earthquake research emerging from the small town of Parkfield, California, known as the seismology capital of the world. Here, where the San Andreas Fault cuts through central California, a magnitude-6 earthquake has occurred on the average of every 22 years for about the last 100 years. (The last magnitude-6 quake struck several years late, rattling the town on Sept. 28, 2004 after a 38-year lull). In addition, hundreds of tiny earthquakes shake the fault yearly. To learn more about these quakes, big and small, the U.S. Geological Survey has placed a network of instruments near the town that records every seismic event.

The 1966 magnitude-6 quake opened a crack in the highway near Parkfield, the "earthquake capital of the world."

After sifting through data relating to thousands of Parkfield microearthquakes dating back to 1987, Korneev and colleagues became interested in a type of low-velocity wave that is detected at the tail end of S and P waves, the strong, far-reaching waves that shake buildings and enable seismologists to determine the strength and epicenter of quakes. (P stands for primary, pressure, or push-pull; S stands for secondary, shear, or shake.) The researchers determined that the slower waves they had identified are fault-zone guided waves, meaning they're trapped within the fault's weak and fractured rock as they travel, much the way sound waves are trapped within the brass piping of a trumpet.

Korneev had a hunch that because these waves are locked inside the fault during their journey from a quake's center, they could offer an unparalleled window into the fault's structure. Until now, seismologists used P and S waves to paint a picture of a fault's makeup. But these waves often travel miles from the fault through bedrock and sediment before they're detected, which clouds any information they might contain about the fault's structure.

Because of this, images constructed from P and S waves can only depict fault zones at a lateral resolution of about five kilometers. Unfortunately, the width of a fault's most active region only spans a few hundred meters or less, meaning images derived from S and P waves are too coarse to yield highly detailed pictures of the fault. The Berkeley Lab team believed guided waves offered a better way.

"We are interested in what happens in the narrow zone of the fault where earthquakes originate," says Korneev. "And guided waves are trapped inside the fault, precisely where we are trying to find structural information."

To test their theory, the team obtained guided wave data collected at two Parkfield borehole monitoring stations, and turned to ESD's Center for Computational Seismology, which serves as the core data-processing, computation, and visualization facility for seismology-related research at Berkeley Lab. Using computer algorithms similar to those that transform x-rays into CT scans of the brain or heart, they developed a two-dimensional image of the San Andreas Fault that portrays a 35-kilometer stretch of the fault, plunging from the Earth's surface to 10 kilometers underground, at a resolution of about 500 meters.

In this tomographic reconstruction, small red stars indicate the hypocenters of earthquakes that are greater than magnitude 4. The large star indicates the hypocenter of the 1966 magnitude-6 quake.

The image has helped them identify a zone that could mark where the creeping part of the fault abuts the locked part of the fault, an area that may correlate to the hypocenters of four earthquakes of greater than magnitude 4 that shook Parkfield in the past century.

"With images like these, the next logical step is to use guided waves to actively image a fault and track stress as it accumulates," says Korneev.

As Korneev explains, seismologists today rely on passive monitoring when it comes to tracking stress deep within faults. They wait for seismic events and collect the data as it rolls in. Korneev, however, foresees a time when seismologists use specialized machinery to create their own guided waves, and then use these waves to image areas where stress is accumulating — rather than areas where stress is released.

"If we had a source that can generate guided waves, then we could repeatedly send signals into the fault and look for the evolution of the zone," says Korneev. "We could see how critical stress builds up, and learn how these changes relate to the seismic life of the fault."

Additional information