ONCE AGAIN, IT WAS A FAULT THEY couldn't see that slammed Los Angeles, crumpled their freeways and toppled their buildings like Tinkertoys. Although the San Andreas fault gets all the notoriety, geophysicists are now realizing that dozens of invisible, even unsuspected, fractures in the rocks beneath the Los Angeles basin cause most of the earthquakes, large and small. And that has radically changed the seismic calculus. Since these hidden faults do not have wires and meters attached to them like an intensive-care patient. scientists will not get any tip-off that a big temblor is coming--as they hope to if, say, the heavily instrumented San Andreas gives way. Worse, "if a quake jumped from fault to fault and ripped the full 100-mile fault zone," says seismologist Tom Henyey of the Southern California Earthquake Center, "there could be an earthquake that registers close to an 8" on the Richter scale-about 125 times more powerful than last week's jolt.
In the reigning theory of geology, the earth is likened to a chocolate-covered cherry that's been roughed up. The crust--the chocolate shell--has been smashed into seven big chunks called tectonic plates. The plates float over the molten magma (the gooey syrup), slipping and sliding laterally along each other's edge. But the jagged, rocky plates do not slide past each other smoothly. They stick. When the rock can no longer withstand the stress, the plates surge past each other with all of the accumulated strain --an earthquake. It turns out, however, that faults lie not only at the junctures of plates but within them as well. Just east of L.A., the San Andreas fault, which separates the Pacific and North American plates and usually runs north-south, turns abruptly west. Here, the Pacific plate, drifting northwest, rams the North American. The battering shatters the continent's crust, creating a web of "blind thrust faults" that resembles a windshield after a car crash. (These faults are "blind" because they do not break the surface and so are unseen. In contrast, the San Andreas pierces the ground hundreds of times.) These buried thrust faults, which scientists began to take seriously as quake hazards in the 1980s, run horizontally, dividing a plate into a top and bottom. But they are not perfectly flat. Instead, they angle up like a ramp. When the rocks can no longer withstand the terrific force of the ramming Pacific plate, the top chunk of real estate surges over the slab below.
That's what happened in Northridge. Nine miles underground, one of the thrust faults--the culprit hasn't been identified yet--gave way. The seismic wave traveled outward at two or three miles per second along the tilted fault line, but never broke the surface. "There's this whole seismic hazard from buried thrust faults that we didn't even appreciate until six years ago," says seismologist James Dolan of the California Institute of Technology in Pasadena. But since then southern California has been shaken by seven thrust-fault quakes registering 4.5 or above. Thrust faults also lie under the Alps and the Himalayas.
This activity has forced researchers to consider a horrendous possibility: that the hidden thrust faults are connected deep underground. If they are, they could break at the same time. If several thrust faults let go simultaneously or in close sequence, Los Angeles could suffer a quake in excess of magnitude 7.5 on the open-ended Richter scale, far more powerful than last week's 6.6 shock. "The lesson here," says Dolan, "is we're not just concerned about The Big One"--usually defined as an 8.0 quake along the fragile San Andreas, which comes no closer to Los Angeles than 25 miles. "There are now several plausible sources for other Big Ones."
To gauge the danger of blind thrust faults, Caltech seismologists were checking five of them under the Santa Monica Mountains last week. The goal: to determine whether the faults last caused an earthquake at about the same time, or centuries apart. If they all surged at once, that would be evidence that the faults are connected. The Northridge quake offered evidence of just such a link: aftershocks shot along several faults to the north beneath the San Gabriel Mountains. But so far, exactly how the thrust faults interact remains a mystery. "It's an incredibly complex jigsaw puzzle," says Dolan.
No wonder, then, that scientists have no idea how often a thrust fault might give way. When it comes to predicting such a disaster, seismologists turn humble. "We prefer the word 'forecasting'," allows geologist John Sims of the U.S. Geologic Survey in Reston, Va. "It sounds like weather, which allows us to be wrong." Even with such a seemingly powerful crystal ball as foreshocks--the rumbles that precede many true quakes-- scientists have only 20-20 hindsight. A couple of weeks before the Northridge quake, temblors shook the shore off Santa Monica; quakes measuring as high as 3.7 rattled the area around Northridge in early January. Henyey suspects that all three sets of quakes were related, but others aren't so sure. "We have earthquakes in California all the time," says geophysicist Willis Jacobs of the USGS in Golden, Colo. "It's really difficult to say what they mean." That's true even when foreshocks act predictably. A 4.8 quake shook the San Andreas near Parkfield on Nov. 14; according to the best quake model, that size temblor at that spot presages a bigger jolt. Scientists predicted that a 6.0 San Andreas quake would follow within days. It didn't. But that failure hasn't changed the overall seismic forecast: scientists still expect a quake to strike a densely populated area of California by 2018 at the latest. It would cause tens of billions of dollars in property damage and kill hundreds, and perhaps thousands, of people. And that's the optimistic prediction. It doesn't even include the threat from thrust-fault quakes.
..MR.-
A previously unknown fauld nine miles underneath Northridge probably caused the 6.6 quake that rocked the Los Angeles basin last Monday. ..MR0-
..MR.-
In strike-slip faults, such as the San Andreas, two chunks of earth grind against each other sideways.
One side of a thrust fault slides up and over the other.
A buried thrust fault is one that doesn't break the surface. ..MR0-