Amos Nur has always been fascinated with earthquakes. Since joining the geophysics faculty at Stanford in 1970, he has witnessed tremendous technological advances in the field of seismology.

But Nur remains "disheartened" by the simple fact that he and his colleagues are still unable to predict when an earthquake will hit.

"I am struck by the number of outstanding questions we are still asking today," said Nur, one of more than 100 international earthquake experts who attended the Third Conference on Tectonic Problems of the San Andreas Fault System at Stanford.

The event - sponsored by Stanford's School of Earth Sciences and the U.S. Geological Survey (USGS) - was held on campus and at the USGS regional headquarters in nearby Menlo Park from Sept. 6 to 8.

"We were hoping to be able to predict earthquakes after the last conference almost 30 years ago, but we still can't," said Nur, the Wayne Loel Professor of Earth Sciences. "It's kind of disappointing."

During the conference, scientists from around the world presented a wide range of research efforts designed to explain the San Andreas fault system - a complex series of ruptures in the ground that runs nearly the entire length of California.

The fault marks the boundary between the North American and Pacific plates - two enormous land formations that float atop the Earth's mantle, constantly grinding together in a process called plate tectonics. As the plates push against each other, stress builds up along the fault and triggers large and small earthquakes.

The first San Andreas fault conference took place at Stanford in 1968, the second in 1973, so a third conference was long overdue, according to Robert Kovach, associate chair of the geophysics department who participated in all three events.

"We've made a lot of progress," maintained Kovach, "but in retrospect, we're still facing some of the same questions."

Heat-Flow Paradox

One of the unresolved controversies is a phenomenon that geophysicists call the Heat-Flow Paradox.

"The simple-minded idea," said Kovach, "is that when you have two blocks of earth grinding against one another, there should be friction, and that should produce some sort of heat."

However, scientists probing the San Andreas fault since the 1960s have been unable to detect the amount of heat predicted by simple friction laws.

"There obviously is heat generated," Kovach pointed out, "but it seems to be dissipated in a mysterious fashion that we don't fully understand."

One explanation offered for the missing heat is that the San Andreas fault is very weak compared to the crust that surrounds it. While a strong fault can accumulate a great deal of stress before triggering an earthquake, a weak fault requires a lot less stress before rupturing - and less stress means that less energy is released in the form of heat.

Experiments have been conducted on a variety of weak rocks, which emit very little heat when rubbed together under laboratory conditions. But so far, no rock samples have been found that are chemically stable under fault zone conditions.

So the paradox remains - and so does the debate, according to conference organizer Goetz Bokelmann, a visiting associate professor of geophysics.

"The main controversy has to do with the strength of the San Andreas fault," said Bokelmann. "Is it weak, or is it strong? It is an important distinction for our understanding of the mechanics of the fault system and earthquakes.

"Probably we have lots of variation along the San Andreas," he observed. "Parts of the fault, including Northern California, are locked and building up large stresses, while other parts, such as Central California, are creeping slowly with relatively low stress levels."

The fault east of Los Angeles is an even more complicated story - a point driven home during the conference by Jeanne L. Hardebeck of Caltech. She presented new evidence showing that the portion of the San Andreas that runs through the Southern California desert is probably a weak fault surrounded by a weak crust. The implication of her findings is that earthquakes are likely to happen anywhere along the southern fault, not just in localized areas.

But John Townend, a Stanford graduate student in geophysics, disputed the Caltech study, which resulted in a long - but cordial - debate between Hardebeck and Townend.

"What's really exciting," commented geophysics associate Professor Gregory Beroza, "is to get a group of people together who are thinking about the same sorts of problems from different perspectives, and see where out understanding matches and where it doesn't, and try to resolve that. That's how we make progress."

Breaking new ground

On Sept. 3, just three days before the conference convened, California's Napa Valley wine country was shaken by a magnitude 5.2 earthquake. The temblor, which caused injuries and millions of dollars in damage, erupted on a previously unknown section of the fault, according to USGS geologists.

"I don't think that we should be so surprised that earthquakes are breaking new ground, so to speak, in unexpected places," commented Kovach. "Even if we didn't recognize it before as being visible on the surface, it doesn't preclude the fact that we could get earthquakes in strange places like Napa."

"We had no role in planning that earthquake," quipped Beroza.

He and Bokelmann already had prepared another report suggesting that there is a weak zone in the Earth's lower crust that could help explain how earthquakes work - at least along the San Andreas.

"In Northern California, most seismicity occurs between 0 and 18 kilometers [0 and 10.8 miles] below the surface," Bokelmann told the conference. "But one thing which we know very little about is how the deeper part of the Earth's crust deforms."

Bokelmann argued that the lower crust (below 10 miles) actually weakens over time, which he said affects the evolution of the fault system.

New technologies

Other presentations during the conference focused on newer technologies that have advanced earthquake study.

Graduate geophysics student Jessica Murray showed how data gathered by satellite global positioning systems (GPS) are being used to analyze deformations in the Earth's crust.

Paul Vincent of the Lawrence Livermore National Laboratory in California demonstrated how tiny movements along the fault can be detected using radar images collected from specially equipped satellites - a method known as InSAR (Interferometric Synthetic Aperture Radar) that was pioneered at Stanford in the past decade.

Allan Rubin of Princeton University and Robert Nadeau of the Berkeley Seismological Laboratory described new techniques for determining the precise locations of microearthquakes - tiny aftershocks that occur by the thousands following a major temblor.

"Thanks to microearthquake analysis," said Beroza, "we know a lot more about the structure of the Earth, how complicated plate boundaries are and how complicated the earthquake process is."

Nur agreed, although one of his own theories received a lukewarm reception during the conference.

Nur maintains that fracture lines running perpendicular or at sharp angles to the San Andreas fault system are caused by blocks of earth that slowly rotate on their axes - like a stack of books rotated together on a shelf.

"The controversy for me," said Nur, "is that most of my colleagues still don't buy it!"

Although he made few converts to his theory of fault rotation, Nur left the three-day conference feeling optimistic.

"I believe that we are on the verge of a revolution," he predicted. "We have more seismic data, more powerful computers, as well as GPS and satellite imaging techniques that didn't exist 10 years ago even.

"And so, despite my downhearted feeling about what we haven't accomplished in the last 27 years, I get the feeling that we are entering a whole new era in earthquake studies and, ultimately, in earthquake prediction."

Organizers plan to convene a fourth San Andreas fault conference at Stanford in two or three years.

-By Mark Shwartz-