Stanford University

News Service



CONTACT: Stanford University News Service (415) 723-2558

Stanford geophysicists find new evidence for a horizontal fault below San Francisco Bay

STANFORD -- "San Francisco has its faults," observe tourist T-shirts in Bay Area airports. Despite the fame of earthquake faults like the San Andreas, surprisingly little is known about their actual structure. Conventional wisdom says that the cracks we see at the surface continue unimpeded through the earth's crust. But this notion has been overturned by a study performed by geophysicists at the U.S. Geological Survey, Stanford and other research institutions.

Using sound waves to make a three-dimensional image of the crust, for the first time they have mapped the three-dimensional structure of the Bay Area to a depth of 20 miles. The result, published in the Sept. 2 issue of the journal Science, shows that the familiar San Andreas and Hayward vertical faults are undercut by a fundamentally different type of fault - a horizontal break that slashes across the earth's crust 9 miles below San Francisco Bay.

That slash may directly link the vertical strike-slip faults on either side of the Bay and transfer stress between them, meaning that a quake on the San Andreas fault could influence the risk of a quake on the Hayward fault, and vice versa. In addition, the horizontal break may be a prime breeding ground for thrust faults similar to those that caused the 1994 Northridge earthquake in Los Angeles.

The finding shows that an earthquake is not an isolated event that occurs on one fault, at least not in a region like the Bay Area, said Simon Klemperer, Stanford associate professor of geophysics, and a co- author of the study. "We live in a plate boundary zone, where earthquakes occur on a system of faults which are interlinked," Klemperer said.

Although it was dubbed a "superfault" in news headlines, scientists are wary about calling the newly mapped horizontal fracture a fault, Klemperer said - at least until they have evidence that it moves. Regardless of what the break is called, this new proof of its presence has updated scientists' assumptions about what earthquake faults look like below the surface.

"The result is interesting because it hints that the horizontal structure goes underneath the San Andreas and Hayward faults," said geophysicist Tom Brocher of the U.S. Geological Survey, who is principal author of the Science article. "This means that the San Andreas and Hayward faults are limited to the upper crust. They are superficial features in that they only cut the top 9 miles of a 60-mile-thick lithosphere." Previously it had been assumed that major faults cut all the way through the crust.

"It means that the ground we walk around on is not attached deep down - the earth is really mobile," Brocher said.

Klemperer said the finding also confirms evidence that seismic imagers have been collecting in similar regions where the massive plates of the earth's crust are sliding past each other. In the Aleutian Islands and Tibet, for example, Klemperer and his colleagues have seen other strike-slip faults that stop in the mid-crust.

The fieldwork for this study, known as the Bay Area Seismic Imaging Experiment (BASIX), was carried out in September 1991. A U.S. Geological Survey ship traversed the major East Bay faults and the San Andreas fault as it sailed from the mouth of the Sacramento River through the Bay and out beyond the Golden Gate Bridge. The ship generated controlled explosions by releasing a compressed air charge each minute during its travel. Sound waves generated by the underwater explosions were reflected off rock layers in the earth's crust and collected by receivers, called geophones, both in the ship's path and at remote land sites.

It took teams of scientists, graduate students and volunteers from five research institutions to collect all the BASIX field data. The Stanford team's task, for example, was to tramp over the Bay Area hills and place the land geophones in a carefully chosen array 40 and 60 miles from the ship's path. In the three years since the ship's voyage, researchers from the five institutions - the U.S. Geological Survey, the Woods Hole Oceanographic Institution, Stanford, the University of California-Berkeley, and Pennsylvania State University - have coordinated their efforts to analyze the complex patterns of acoustical signals and interpret what they see. The discovery of the horizontal break beneath Bay Area faults is one of several new findings expected to come from that data.

The scientists use the sound waves that were reflected vertically back to the ship's path to determine the echo time to the deep horizontal break, which acts as a reflector to such waves. They use the wide-angle data collected at the land stations as complementary information to calculate the depth of the break and determine the composition of the crust at various depths over a wide geographical area.

Stanford's contribution to the project included analyzing that wide-angle data. Klemperer and John Hole, a postdoctoral fellow, led one of several geophysical teams that worked to identify the underground change in rock composition, and therefore the presence of the potential fault. In addition, the measurements made by the Stanford team were critical in defining the geographical extent of the deep horizontal break. It extends beneath San Francisco Bay, north to San Pablo Bay and perhaps beyond.

The wide-angle data are recordings of sound waves that are refracted or bent as they pass through the earth's crust and eventually travel back to the earth's surface. Since the sound waves travel at different speeds in different rock types, the travel time of the sound waves to the geophones enabled Klemperer, Hole and their colleagues to construct a velocity profile of the rock layers.

They saw a sharp contrast in sound wave velocity above and below the break. The waves traveled at 5.5 to 6.2 kilometers per second in the upper rock and 6.5 kilometers per second in the lower rock. The wave speeds suggest that the rock above the apparent fault is silicon-rich Franciscan rock, which is exposed at the earth's surface throughout the Bay Area. The rock below appears to be iron- and magnesium-rich rock, similar to that found in the ocean's crust.

Although there are competing explanations for the origin of the horizontal break and the presence of the oceanic rock below it, each has similar implications for earthquake hazard.

One model suggests that the lower crust below the break was formed when the oceanic crust dove under the continental plate 10 million years ago as the North American and Pacific tectonic plates slid past each other. If this occurred, the actual deep boundary between the two plates would have shifted eastward and the horizontal fault would represent a physical connection between the old plate boundary, marked by the San Andreas fault system, and the new plate boundary, marked by the newer, inshore faults on the East Bay.

Stress may be transferred more readily between the San Andreas and the Hayward fault systems if such a horizontal connection exists. This would explain the phenomenon of earthquake "pairing" observed over the last 160 years, in which three pairs of earthquakes of magnitude 6.5 or greater occurred on opposite sides of the Bay within two to six years of each other.

Whether the horizontal surface connects the vertical faults or not, it is under some degree of compression that can lead to buckling of the crust above the fault. This force could cause the deep break to sprout new thrust faults similar to those which were responsible for the 1994 Northridge earthquake in Los Angeles. While quakes generated by thrust faults in the Bay area have been smaller in magnitude and less frequent than quakes on strike-slip faults, they still pose a substantial hazard because a thrust fault is often unknown until it moves.

The horizontal fault is just one structural feature to be identified by the BASIX team. Klemperer and Hole plan to look at their deep crustal reflection data again to elucidate other aspects of the Bay Area's three-dimensional fault structures.

One project aims to look at the reflections of the sound waves off the vertical San Andreas fault in an attempt to image it more accurately. Currently, knowledge of the San Andreas and other vertical faults is built up largely from data collected at the time of earthquakes. Seismically inactive parts of the fault have not been extensively mapped, particularly at depth.

The BASIX study already has confirmed that it is very hard to look at features on the surface of the earth and extrapolate to guess what they look like at depth. "That's an important lesson for the field geologist," Klemperer said.

Brocher said the researchers cannot be sure that the new data will someday make it easier to predict earthquakes. "It does help us understand the framework that earthquakes work in," he said. "It is another piece of the puzzle."


Jennifer Howard is a science writing intern with Stanford News Service


This is an archived release.

This release is not available in any other form. Images mentioned in this release are not available online.
Stanford News Service has an extensive library of images, some of which may be available to you online. Direct your request by EMail to

© Stanford University. All Rights Reserved. Stanford, CA 94305. (650) 723-2300. Terms of Use | Copyright Complaints