The 7.1 Experiment : For Northern California, the Loma Prieta Temblor Was a Natural Disaster. For Scientists, It Was the Opportunity of a Lifetime--and an Earthshaking Surprise.

<i> J.E. Ferrell's last story for this magazine was about the soil poisoning at the Central Valley's Kesterson Reservoir. </i>

GEOLOGIST TOM HOLZER looked through red-rimmed blue eyes at the two dozen rumpled men and women slumped in chairs before him. He was standing in a conference room in the steamy, windowless heart of the U.S. Geological Survey complex in Menlo Park, Calif.

“Has anyone seen faulting?” Holzer asked wearily. He spread his hands, palms to the ceiling. “Anywhere?”

No one in his audience--the exhausted scientists of the USGS branch of engineering seismology and geology--responded.

It had been 50 hours since the 7.1 Loma Prieta earthquake convulsed Northern California a few seconds after 5:04 p.m., Tuesday, Oct. 17, 1989. Holzer, 45, and the engineering group that he directed could count on one hand the hours of sleep they had sneaked since crawling from beneath their desks after the tremors jolted Menlo Park.


Moments after the earth ruptured 10 miles east-northeast of Santa Cruz and 60 miles south of San Francisco, the group had been among a contingent of about 500 earthquake scientists who had swarmed into frenzied action. Riding excitement the way surfers ride waves, researchers in Japan, West Germany, Chile, the Soviet Union and Australia dashed madly to their computers to observe the long undulations coursing through the earth to their seismographs. About 50 others from universities and regional USGS offices in Hawaii, New York, Colorado and Southern California grabbed their instruments and headed for the USGS in Menlo Park, the nation’s premier earthquake research facility, only 30 miles north of the epicenter.

“We think it’s great!” a seismologist gushed at one point in the USGS press room and then attempted to swallow his words. “Well, not great, but . . .”

It wasn’t that he didn’t recognize the sobering destruction, the lives lost, the financial ruin or the paralyzing fear a major earthquake causes. It was just that the same violent shaking that meant disaster to most people meant an opportunity for him and his colleagues to gather data impossible to obtain any other way. Their grand-scale seismic experiments always erupt courtesy of Nature, unplanned and uncontrolled.

Although they can’t put earthquakes into bottles or re-create them, researchers did snare this temblor, and its echoing aftershocks, in their own fashion--they trapped it in a vast, worldwide network of instruments. In the weeks following the earthquake, they set out to identify the beast they had already named after the most prominent geographical feature near the epicenter, 3,791-foot Loma Prieta, the highest point in the Santa Cruz Mountains. They began their pursuit puffed with confidence. They had forecast this particular animal--magnitude, locale and time frame--18 months earlier. They were armed with assumptions and theories built on decades of studying California’s 700-mile San Andreas Fault system, which surely was the locus of Loma Prieta.

Holzer’s group was responsible for measuring, mapping and analyzing the physical effects of the quake, including liquefaction (where usually solid ground had turned to mush), surface cracks and landslides. They had returned to Menlo Park after a day in the field, crawling into sand-filled basements, peering out of low-flying aircraft, bushwhacking through the Santa Cruz Mountains. The bushwhackers had expected to find the surface expression of the rupture: a disjointed but orderly series of cracks running along a 25-mile segment of the San Andreas Fault. Instead, the fault-line team told Holzer in puzzled voices, the hundreds of cracks they found--some large enough for a person to stand in, some only 1/8-inch wide--ran perpendicular to the fault or broke in a direction opposite of what was expected. It just didn’t make sense.

As the frustrated engineering group met, Andy Michael, 30, a round-faced, normally cheerful seismologist, frowned over his computer screen a few doors away. He and the other nine seismologists in the USGS seismology branch, had expected their computers to produce a map showing a fairly straight line of aftershocks echoing and tracing the main rupture along the San Andreas. But the Loma Prieta aftershocks spread in a cloud across two minor faults nearby, the Sargent and the Zayante.

Another anomaly was showing up on his screen as well: The epicenter--the point on the earth’s surface directly over the buried hypocenter where the rupture began--was nearly 4 miles west of the San Andreas instead of right on top of it.

And what was this speculation coming from the USGS regional office in Pasadena? The land mass known as the Pacific Plate, the scientists in Pasadena suggested, had shoved into and over the North American Plate. C’mon, Michael thought. Everyone knows the San Andreas is a strike-slip fault: The Pacific Plate creeps north and the North American creeps south along a vertical plane, which is the fault itself. One day, all this creeping will make Los Angeles a suburb due west of San Francisco. But no one expected the San Andreas to behave like a thrust fault, an entirely different beast.

In fact, so much of the seismic information was unexpected that Michael and his co-workers were wondering whether the computers had been knocked silly by the earthquake. They had spent hours double-checking everything. The night was dragging on. Nothing was changing.

Back in the conference room, a geophysicist named Will Prescott, head of the USGS tectonophysics branch, quietly slipped in to hear the latest from the engineering group. A slim, bearded 44-year-old, with brown eyes hidden behind thick glasses, he had already heard about the data that had Michael perplexed. It would be at least another day before his team of geodetic scientists--geophysicists, geologists and technicians who measure and study changes in the earth’s crust--could analyze the data it had collected so far and offer a good picture of exactly how the area around Loma Prieta had shifted. But even the geodetic group’s preliminary data indicated something strange was going on. Just hours before, a geophysicist reported that the quake had moved a mountain only a few inches north. An earthquake the size of the one Prescott had felt on Oct. 17 should have generated displacement that was, well, a little more earthshaking.

Holzer ended the meeting after assigning the next day’s tasks and settled down to paper work. Fifty-two hours gone, and still the scientists hadn’t grasped the essence of Loma Prieta. If their assumptions were correct, they would have known what the beast they had trapped looked like. It was becoming clear that the San Andreas wasn’t what they thought it was. “It wasn’t fitting our model,” Holzer said simply.

In other words, the Earth had fooled them again.

ON FRIDAY MORNING, Holzer’s fault team trudged back into the mountains to look for the fault-line cracks they thought had to be somewhere. Along the way, they set out instruments to measure the creeping that usually occurs after an earthquake. Holzer had asked the geologists to pinpoint and study as many effects of the quake as possible before they were plowed under or washed away by expected weekend rains.

Sue Hough, a 28-year-old seismologist at Columbia University’s Lamont-Doherty Geological Observatory in Palisades, N.Y., had left her husband in charge of their two young children and flown west on Thursday, Oct. 19, with five colleagues. They came bearing the latest high-tech portable seismographs that measure how different types of soils shake during an earthquake. Volunteering their services and their equipment, they were sent by Holzer to a spot his group had yet to inspect on foot: the area around the collapsed section of Interstate 880, the Nimitz Freeway, in Oakland.

Early Friday, Hough--a soft-spoken woman with deep-set hazel eyes--walked along side streets near the pancaked freeway and looked for surface signs of the quake. In the warm morning sun, she saw rescue workers frantically digging for survivors in the concrete rubble. “I used to live in this area,” she thought. “I used to drive this freeway and think about earthquakes. I used to think it was a coffin.”

Because of the detours set up by rescuers, Hough and two colleagues zigged and zagged around the collapsed section, comparing an Oakland street map against a soils map drawn by USGS engineering seismologist Roger Borcherdt in 1976, nearly 20 years after the double-decker structure was built. The soils map was the first produced to identify areas that would be particularly unstable during an earthquake.

Hough was astonished to see just how closely the boundaries of the destruction coincided with the change in soil types. The collapse began where the freeway crossed from relatively solid sediments to fill that covered San Francisco Bay mud. It looked like a textbook example of why the map had been drawn in the first place. Now all she had to do was prove that the collapse was related to the unstable ground below.

The next day, Hough and her colleagues placed five seismographs throughout the area, including one on the fill above Bay mud and one on the more solid sediments. For the next six days, the seismographs recorded all strong aftershocks. Later, Hough would transfer the data to a computer disk and run it through a computer program that translated the bytes into pictures of jagged lines showing horizontal motion (side to side and forward and back) and vertical motion (up and down). On Saturday, Hough put a sixth seismograph on an uncollapsed section of freeway for 15 hours.

When the Loma Prieta earthquake ruptured nearly 12 miles underground, the earth was responding to the still-mysterious forces in the asthenosphere where rock flows like warm plastic. Riding this subterranean goo are continent-sized jigsaw pieces of crust. Four miles to 40 miles thick, these tectonic plates do-si-do past each other in a continuous square dance around the Earth. The Pacific Plate moves past the North American Plate at an average rate of nearly 4 centimeters each year. In some segments, tons of compressed rock on the edges of the plates get caught on some subterranean snag for years. When enough pressure builds, they suddenly rip free and slam into adjacent rock, setting off waves of different sizes and shapes that ripple, tear, undulate and ricochet through the earth.

The first waves emanating from an earthquake, called the primary waves, act like the waves coursing through a Slinky that is stretched out on the ground and jerked at one end. The wave doesn’t move up and down; it contracts and expands the material through which it moves. These are the “What was that?” waves, Holzer says. These waves break through the surface at about 4 miles per second to disturb the air and create the terrible primeval roar heard as the earthquake’s slower, but more damaging, second waves come careering along at about 2 miles a second. These secondary waves, called shear waves, wrench buildings from side to side, toss desks into the air and fling refrigerators across kitchens.

It was the shear waves from Loma Prieta that Hough wanted to measure. As the waves traveled from the hypocenter, they weakened, until they encountered loosely packed soils, which have room to shake more intensely than compacted soils and bedrock. Bay mud was one of those soil types that could be expected to amplify waves. Not only did Hough want to ascertain how much the mud amplified the shear waves in contrast with the more settled sediments nearby, she also wanted to see if the freeway resonated to those amplified waves. That is, she wanted to see whether, in an aftershock, the freeway’s normal vibration pattern or frequency matched that of the amplified earthquake waves. If it did, the increased frequency would cause the freeway to shake harder and harder.

While Hough was working around the collapsed freeway, dozens of others in Holzer’s group were chasing Loma Prieta’s shear waves elsewhere. Seismologist Paul Spudich headed a group placing seismographs along the tops, sides and bottoms of ridges in the Santa Cruz Mountains. The destruction had been incredible there--houses twisted and thrown off their foundations, redwoods snapped in half. David Boore, another seismologist, was awaiting results--gathered by hand from old-fashioned, non-computerized instruments--from another type of seismograph called a strong-motion detector. Designed to record jolts that would knock other seismographs off the scale, more than 120 detectors had been tucked into structures all over the Bay Area. When he obtained the records from each one, Boore would see the pattern of the strong jolts and how the various structures housing the instruments had responded.

Geologist John Tinsley coordinated the search for liquefaction in the Monterey Bay area. On Friday, and every day for four weeks, he recruited a dozen people to hunt down telltale miniature gray “volcanoes"--sand boils. Wet and loose sandy soils or fill not only amplify a shear wave, they also can be shaken so severely that they liquefy. Sand particles collapse into each other, lose the strength to support themselves and anything above them, and behave as if they were fluids. The sand boils form when water, squeezed from between collapsing sand particles, shoots to the surface, taking sand with it. Hundreds of sand boils erupted during the Loma Prieta earthquake--between Oakland International Airport runways, at the Alameda Naval Air Station, at the east end of the Oakland-San Francisco Bay Bridge and in San Francisco’s Marina District.

In the Marina, another liquefaction team headed by geologist Tom Hanks found bits of tar paper and pieces of old glass--ironically, detritus from the 1906 quake that had been used for landfill--that also spurted to the surface. In some cases, sand boils erupted with enough force to break through thin concrete floors of basements and fill them with sand.

The liquefaction teams, once they found the sand boils, began taking aerial photos, measuring displacement, drilling for soil samples. They also placed seismographs to measure the aftershocks to determine how Loma Prieta had transformed terra firma into melting Jell-O.

It would be a few more weeks before the engineering group slowed down enough to begin analyzing the data. But a few things were becoming clear.

The assumptions they had made about the way soils would respond in a major quake had been confirmed, with some minor surprises centering on how vigorously the loose soils had shaken. What they were turning up now was enough data to help create more and better maps and to tweak their early conclusions on the way land around Loma Prieta responded when the San Andreas snapped. They were already telling the news media and state and county engineers a mild version of “I told you so.”

But there was still the matter of those cracks. A week after the quake, the fault-line search had petered out. There were simply no cracks that demarcated the rupture of the San Andreas Fault. Unlike every other major San Andreas quake ever studied, there was no place where you could stand on one side of the fault line and see displacement on the other side. What emerged indicated a new phenomenon. Although many of the cracks were landslide cracks indicating shifting topsoil commonly seen in any earthquake, others were inexplicable. The geologists were stumped.

At first Dan Ponti and his team of researchers theorized that the cracks outlined huge blocks of ground that had rotated during the temblor. Later, they proposed that perhaps the mountain ridges were spreading the way a loaf of French bread splits in the oven.

Although they expected landslide cracks from shifting topsoil and rifts where a fault ruptures to the surface, they didn’t expect such dangerous fissures--3 feet wide and 10 feet deep. Nor did the researchers expect to see them so far away from the fault line. Whatever the mechanics and the ultimate meaning of the cracks, it looked as if they had discovered a new effect of earthquakes, and a new hazard as well.

WHILE THE engineering team measured cracks and set out seismographs Friday, a jubilant Andy Michael, parked once again in front of a blinking computer screen, cheered silently. There was no doubt about it. The data were clean. Something new had occurred on the San Andreas Fault, and he was part of the team that discovered it.

What he was looking at was massaged data from a network of seismographs all over the area that painted a cross-section of the aftershocks. On his screen, the aftershocks appeared as a smattering of tiny octagonal dots whose diameters increased with magnitude. The pattern showed plainly that the earth didn’t rupture along a neat vertical slice of the San Andreas as expected. Instead, the dots showed a fault plane that tilted at a 70-degree angle. The Pacific Plate, as it had moved north, pushed up and over the North American on that angle; “it shows thrust,” Michael said. This segment was an oblique fault--a combination of strike-slip and thrust faulting--something unheard of on the San Andreas before the Loma Prieta earthquake.

An angled San Andreas Fault plane also would explain the unexpected location of the epicenter. Most of the San Andreas is like a vertical wall plunging miles beneath the surface of the earth. Earthquakes begin to rupture somewhere along the bottom of that imaginary wall. That rupture point, the hypocenter, is usually found directly below the epicenter. In fact, before Loma Prieta, the hypocenters of San Andreas Fault earthquakes were always located directly below the epicenters. But Loma Prieta’s epicenter was 4 miles from where the top of the San Andreas wall ran through the Santa Cruz Mountains. That meant that the wall in this case was tilted, as if the Pacific Plate was leaning against the North American. And when seismologists drew a straight line from the hypocenter to the surface, the epicenter was 4 miles away from the top of the wall.

Creating a detailed picture of what happened on the Loma Prieta segment of the San Andreas Fault was what the seismology branch was all about. The puzzle pieces these scientists were looking for were location (epicenter and hypocenter), length of rupture, magnitude and the mechanics of the faulting itself. Like the seismologists in the engineering group, they relied on seismographs to provide them with wave data to analyze the earthquake. But unlike the engineering group, their seismographs recorded only the primary waves--the first clean shivers that look so familiar when they’re depicted as jagged lines on a rolling drum of paper.

Today, those drums are old technology, and seismologists refer to them only occasionally. Computers, with their ability to crunch massive amounts of data quickly, collect and process the seismic data now.

In the basement of the USGS complex, about 100 technicians maintain the seismology branch computers. They are hooked to a network of about 500 seismographs at 400 California sites. When Loma Prieta struck, the computers automatically began processing the stream of data to change the bytes into dots on the computer screen. The biggest dot was the epicenter.

Location is one of the first things the seismologists look to the computers to discover. They used to calculate it off the printed jagged lines of the primary and shear waves, using a formula developed in the 1890s. It’s so simple that it can be computed in your head, as long as you live far enough away from the epicenter to feel the primary waves and the stronger shear waves separately. (Close to the epicenter, the waves erupt right on top of each other and feel like a jumble of violent tossing and turning.) All you have to do is multiply the seconds between the first shake and the second by five. During the Loma Prieta quake, a Sacramento resident, for example, could have multiplied the 20 seconds between the first jolt and the second by five and come up with an epicenter that was about 100 miles away.

At Menlo Park, only 30 miles from the epicenter, so few seconds lapsed between the first rumble and the second, that the researchers in the seismology branch barely had time to crawl under their desks to avoid flying books and bookcases. They raced to the computer room seconds later when they discovered that the quake had knocked out their personal computers. Although backup power kept most of the main computers in the basement running, the quake had also cut off about 20 seismographs near the epicenter.

That’s one reason they believed their initial data was suspect, Michael said. That first night, he and his colleagues in the seismology group, who were accustomed to working on mostly self-assigned projects in their own offices, were thrown together in the basement computer room, watching dot patterns of aftershocks till dawn.

David Oppenheimer, a seismologist who specializes in calculating probabilities of earthquakes, watched with some concern the computerized pictures of the aftershocks creeping up the San Andreas Fault toward the San Francisco Peninsula. For a while, he and the others thought another big earthquake might be getting ready to occur.

They determined later that they used 38 seismographic stations to arrive at the location of Loma Prieta. They relied on 267 to figure out the focal mechanism--exactly how the earth moved. But because their instruments were too sensitive to record the enormous impulses coming from such a large quake, they let the USGS National Earthquake Information Service in Golden, Colo., determine the magnitude.

The method of determining relative magnitudes of earthquakes was developed by seismologist Charles Richter in the 1920s. Richter’s scale compared the size of the largest shear wave--the highest crest on a seismogram--and the distance from the epicenter the wave was recorded. His scale is logarithmic: Each increase of one in the rating indicates a tenfold increase in the amplitude of the seismic wave, and about a 32-fold increase in total energy released.

The first Loma Prieta magnitude issued by the USGS seismologists in Golden was 6.9, based on four seismographs in Alaska. For the next week, however, magnitudes poured into Golden from many of the 2,000 seismographic sites around the world, including those of the World Wide Standardized Seismograph Network the U.S. Department of Defense installed in the 1960s during the height of the Cold War to monitor Soviet underground nuclear tests. Since the earth’s crust isn’t uniform--the speed and size of the earthquake waves vary at each station--not every seismograph calculates the same magnitude. The USGS averaged hundreds of magnitudes, which ranged from 6.5 in Tahiti to 7.5 in Eastern Europe, and one week after the quake, revised the magnitude to 7.1.

But the puzzle of Loma Prieta was missing one important piece: How close to the surface did it come? For most earthquakes of this size, it would have been easy to determine where the rupture ended because there would be a visible rift in the ground. But because scientists weren’t finding any evidence of a surface break, Michael looked to the geodetic scientists in the USGS tectonophysics branch for the answer.

PEOPLE THINK we have a great job,” said geologist Karl Gross, 38, field operations manager for Will Prescott’s group of global surveyors, the geodetic scientists. Gross, a former U.S. Marine who did embassy guard duty in the Middle East, spends most of his time now measuring the great outdoors to a tolerance of an eyelash. “They see us outside on top of mountains, in beautiful places with great views. Of course, they’re not up here when there’s 30 knots of wind and a temperature of 5 degrees.” Or hunched over computers at 2 a.m. fitting the measurements into a model that explains what happened 12 miles under the earth’s surface.

When the Loma Prieta earthquake hit, Gross and most of the other members of the Menlo Park geodetic group were measuring other parts of California in a continuing project to determine how the surface of the western United States is changing shape. After hurrying back to the Bay Area and checking on families and friends, the nine geodetic scientists and technicians began a schedule that could have been a study in sleep deprivation. They arose at 4 a.m. to drive to mountaintops as far as three hours away, took measurements until sunset, drove back to Menlo Park to rest, changed equipment, went back out at midnight for nine hours and returned to process data.

The plan was to remeasure the distances from Loma Prieta to several other mountains in the San Francisco Bay and Monterey Bay areas. Placed on a map, the measurements look like a black spider web with Loma Prieta in the center. After gathering figures, with two types of instruments, on how much the earth had moved during the quake, the scientists would begin to monitor the smaller increments of post-quake creep.

Geophysicist Mike Lisowski and physical science technician Gary Hamilton, bundled against the cold, dry air atop Loma Prieta, aimed a laser beam from a contraption called a geodolite toward Mt. Allison 26 miles away through holes they had cut long ago in a chain-link fence at a television station antenna facility.

“We’ve measured this line every month for the last five years,” said Lisowski, who gulped a bite of cold oatmeal as he waited for Gross to find the benchmark on the other mountaintop. The USGS or the National Geodetic Survey has been planting benchmarks--round brass plates--on virtually every prominent spot in the nation since the 1800s.

Gross set up a panel of reflectors that looked like numerous car tail lights on a stiff board, radioed that he was in position and flashed a mirror against the sun to help Lisowski aim the laser.

Lisowski peered through a telescopic sight on the side of the 100-pound white geodolite, which was perched on a tripod above the Loma Prieta benchmark. He made gentle adjustments until he could see a tiny red dot--the laser beam’s reflection--on the far mountain. He looked in the sky and spoke into a walkie-talkie: “Tom, we’re ready when you are.” A Cessna 182 swooped overhead, banked and flew from the top of Loma Prieta toward the mountain where Gross, walkie-talkie in one hand, binoculars in the other, watched carefully.

“The airplane flies along the laser path to measure the temperature and humidity,” said the tall, hefty Hamilton. He munched on a bag of Corn Nuts. “There’s a delay in the signal that’s caused by the density of the atmosphere.” They needed to know that delay to make accurate distance readings.

Lisowski, binoculars to eyes, guided the plane as a computer recorded the time it took for the laser pulses to travel from Loma Prieta to the other mountain and back. “A little left. A little up. Looking good.” Soon the plane was out of sight, and control passed to Gross. After the plane flew over Gross, it banked and flew back to Loma Prieta on a second pass.

The work continued throughout the day as crews dragged reflectors to six sites, from Mt. Allison in the East Bay to a low hill in Watsonville. At one point, a woman with a benchmark on her property asked Gross how much her land had moved during the earthquake. Lisowski consulted his computer and radioed back: “Tell her she’s 7 inches closer to Loma Prieta than she was two weeks ago.”

At 2 the next morning, yellow and white lights glimmered brightly in the cold, windless air from San Jose to Oakland. The soft machine hum of civilization rose 2,659 feet to Mt. Allison, where geologist Jerry Svarc set up the latest in surveying equipment: a small, white satellite receiver--"the conehead,” he said.

The receiver works with the U.S. Global Positioning Satellite System, which was designed as a military navigational aid. The essence of the system is that it can tell people instantly where they are on the planet, to within a couple of feet. With solar panels extended and plump bodies canted toward Earth, the satellites resemble 1,500-pound fruit flies the size of Volkswagen bugs. Svarc had to track four of them as they orbited 12,000 miles overhead, from 3 a.m. to 9 a.m., continually broadcasting radio signals indicating the time and their location.

At 3 a.m., Svarc’s conehead picked up the satellites’ signals and transmitted them to a computer in the back of a USGS truck. By pinpointing its position relative to all four satellites, the computer calculated where it was--longitude, latitude and altitude.

Once the computer locked into the satellites, Svarc sat in his truck and waited, under orders not to abandon the $150,000 equipment. Just before dawn, he watched car lights multiply on the freeways far below. This was the future. In the past few months, geodetic scientists have shown that the satellite system is in many ways as accurate as geodolites, and it doesn’t require two teams in the mountains or clear days. The satellites can be used day or night in any weather.

With the geodolite and global positioning system, Lisowski and Prescott took the data that they and the other geodetic scientists had collected and provided the missing piece of the puzzle: how far below the surface of the earth the earthquake stopped. They did this by working backward--putting all the surface changes in a computer and determining which model of a rupture fit best. The answer they came up with was 3.7 miles.

So, two weeks after the earthquake, the seismologists and geodetic scientists offered this description of the Loma Prieta earthquake: Eleven-and-a-half miles below the surface of the earth, a 25-mile-long segment of the Pacific Plate lumbered 6.2 feet north and pushed 4.3 feet up and over the North American Plate. Even though the rupture stopped 3.7 miles from the surface of the earth, the ground level uplifted as much as 14 inches and, in some spots, moved as much as 7 inches north.

THE DECEMBER MEETING of the American Geophysical Union in San Francisco was the place in 1989 for geologists, geophysicists, seismologists, volcanologists, meteorologists, physicists, oceanographers, hydrologists, geochemists and engineers to swap theories and data. The gathering was especially exciting because the earthquake scientists had a new temblor to show off: Loma Prieta.

Thirteen Loma Prieta scientists delivered 15-minute synopses of their observations to colleagues packed into San Francisco’s Civic Auditorium. Andy Michael presented the seismology branch results. Will Prescott delivered the geodetic scientists’ conclusions. David Boore discussed preliminary data from the engineering group. Sixty additional presentations were relegated to a packed afternoon “poster” session, where researchers tacked their papers to giant bulletin boards in a hangar-sized hall. Hundreds of scientists jammed the aisles to read about liquefaction, aftershock warning systems and the way helium in soil acted before the quake. Some scientists decorated their bulletin boards with photos, elaborate illustrations and bold headlines; others augmented their presentations with videotapes or slide shows.

Sue Hough preferred the subtle approach. Among the paper and photos tacked to her bulletin board was a small plastic bag full of Bay mud. Her seismographs had indeed shown that aftershocks had shaken the fill over the Bay mud harder than the older, relatively solid sediments. It also looked as if the pancaked section of the freeway resonated to the amplified earthquake waves, and thus may have shaken even harder.

Like all the information coming out from the engineering group’s research, Hough’s has a direct connection to life in the fault zone. She had already sent her findings to Caltrans, which had to decide how and if it would rebuild the collapsed freeway sections. “Last I heard, they were putting in a proposal for a single-deck freeway,” Hough said with a sigh. “As a seismologist, I’d like to see them go in a little further onto stronger soils, but there are problems in displacing communities.”

The USGS liquefaction teams had their data posted in the hall, too: pictures of sand boils and slumped buildings, and preliminary maps pinpointing problem areas. “A sizable contingent of practicing engineers believes that once a (soil) deposit liquefies, it doesn’t do it again,” drawled John Tinsley. “We were eager to see if the same areas that (had problems in 1906) suffered liquefaction. We found that generally to be true.”

Dan Ponti stood around talking about cracks. The state, he said, may need to zone some areas for what should be considered an earthquake engineering hazard as serious as liquefaction and amplification. It was a hazard with a silver lining, though. There’s pretty good evidence, he said, that the ground had cracked this way before, during previous earthquakes. That means, Ponti said, that by “trenching” the cracks--conducting an archeological dig in the area--geologists may be able to obtain a historical record of when and how other earthquakes had occurred. And that is one step in predicting the next one.

In fact, the question of prediction was a hot topic at the poster session. David Oppenheimer stood amid the billboards discussing probability studies. His own research into the effects of Loma Prieta on other segments of the San Andreas was just beginning.

Earthquake probabilities are based on the theory that the longer a segment of the San Andreas goes without an earthquake, the more likely it is to break and break big. There had been three forecasts--prediction is too strong a word--for the segment where Loma Prieta ruptured. They varied in probability from 30% to 47%, over 30 years, and in segment length, from 18 miles to 45 miles. They all agreed on the magnitude: 6.5. With the 7.1 temblor (six times larger than the forecasts), seismologists consider the chance of another quake in that area small. But what happens on the northern end of the segment? In the days following Loma Prieta, while Oppenheimer was watching aftershock patterns, he noticed a strong jolt under Daly City, on the next segment north from the Santa Cruz area. He began a new research project, arranging for more instruments to be placed on the peninsula segment. “Loma Prieta clearly loaded that section of the fault,” he said. But when will an earthquake occur there?

“Some models predict an earthquake in five years; some models say 50 years,” Oppenheimer said. “It’s really speculative stuff. We just don’t understand the process.”

As the poster session drew to a close, Andy Michael looked forward to pursuing his science at a slower pace, perusing some of the 100,000 individual records of aftershocks. And he planned to finish the report he was writing when Loma Prieta hit--an analysis of the Whittier Narrows earthquake in Los Angeles in 1987.

The geodetics group was heading back into the field to finish its interrupted terrain measurements in Mammoth and the Tehachapi Mountains while it continues to monitor the Loma Prieta segment of the San Andreas.

Tom Holzer and the USGS engineering seismologists and geologists planned to finish their analysis in San Francisco’s Marina District and the Santa Cruz Mountains and present their reports to city and county officials early in the spring. Holzer anticipated some controversy. The earthquake threw the engineering group into the middle of discussions about whether residents in the mountains would be allowed to repair or rebuild their homes, and the Marina soils report promised to raise a political brouhaha as well. Many of the repairs to homes and underground gas and water pipes are just “quick fixes,” said USGS structural engineer Mehmet Celebi.

While the poster researchers packed up, dozens of scientists and engineers swapped business cards and arranged to obtain raw data from the instruments that had snared Loma Prieta’s primary and secondary waves to do further research. Researchers would be studying the effects of Loma Prieta and producing scientific papers for years.

“There are probably things about this earthquake that we’ll never understand,” Ponti said. “The Earth doesn’t want to tell us all of its secrets. But that’s the nature of the beast.”



The Loma Prieta quake was both a strike-slip, (horizontal) rupture and a thrust (vertical) rupture. Starting about 11.5 miles underground and extending to within 3.7 miles of the Earth’s surface), an angled segment of the San Andreas fault broke as the Pacific Plate moved northwest in relation to the North American plate and thrust up and over it. The angled, thrust motion never before associated with the San Andreas, explains why the epicenter was miles off the San Andreas Fault line. At the surface, the ground was pushed up 14 inches and north about 7 inches. At the rupture point or hypocenter, the Earth moved up about 4 feet and north about 6 feet.


Liquefaction and amplification which increase a quake’s destructive power, occurred widely during the Loma Prieta quake. In San Francisco’s Marina District, for instance, surface eruptions of liquefied deposits-sand boils-told the tale. Amplification contributed to the collapse of Oakland’s Nimitz Freeway, which killed 42. Other physical effects included large cracks and landslides near the epicenter.

Liquefaction: When seismic waves hit loosely packed deposits that contain water, the waves increase in strength and cause the deposits to liquefy. Liquefied deposits can erupt in a sand boil.

Amplification: When seismic waves hit loosly packed deposits-which have room to shake more intensely than compacted deposits-the waves increase in strength.

The Probability of Earthquakes on the San Andreas Fault, 1988-2018, by Segment I: North Coast: Less than 10% II: S.F.Peninsula: 30% III: Santa Cruz Mtns.: 30% IV: Central Creeping: 10% V: Parkfield: Greater than 90% VI: Cholame: 30% VIII: Carrizo: 10% VIII: Mojave: 30% IX: San Bernardino Mtns.: 20%

X: Coachella Valley: 40%

These probabilities were calculated 18 months before the Loma Prieta quake hit. The quake fulfilled the forecast for Segment III; it also may have “loaded” Segment II, increasing the chances for a major earthquake there.