For much of the 20th century, Los Angeles and its surroundings have been the epitome of sunny, sybaritic California. The mountain ranges overlooking the Los Angeles basin that cradles the city, and the Pacific Ocean beyond, have served as the settings for cinematic fantasies ranging from Westerns to the classic detective stories of Raymond Chandler. Hollywood and Beverly Hills are – in the popular imagination at least – home to slim starlets splashing about in pools and frustrated screenwriters pounding word processors. But miles beneath the seedy or splendid glamour of the ‘City of Angels’ lurks an invisible menace: a network of geological faults that, despite having no history of seismic activity, might just end up triggering California’s most destructive quake while seismologists’ backs are turned.
Southern California’s hidden geological faults have already asserted themselves rather violently – most recently, at 4.31 am on 17 January, when one of them triggered the so-called Northridge quake. This killed 61 people, smashed freeways and wrought more than 10 billion dollars’ worth of damage. Geologists are still trying to determine exactly where the fault that caused it lies, but most clues point not to the familiar danger zone of the seismologically active San Andreas Fault but towards a south-dipping fault about 13 kilometres beneath the San Fernando Valley, to the northwest of downtown LA.
DEEP SUSPICION
At the back of some Californians’ minds is the suspicion that the Northridge quake may be a mere preview of a more deadly coming attraction: the Big One, a superquake worthy of a Cecil B. De Mille extravaganza. For decades, seismologists in California assumed that when and if such a cataclysm struck, it would do so somewhere along the San Andreas Fault. This zone, which is visible on satellite photos, runs through much of the state. The fault begins in the Pacific Ocean, and parts of it run onshore in northern California; then, just south of San Francisco, the fault permanently exits the Pacific and cuts overland all the way southeast to the desert just above the border between the US and Mexico. The fault marks the meeting place of two tectonic plates – the Pacific plate to the west, the North American plate to the east. They slide horizontally past each other in opposite directions in what is known as strike-slip motion. But the plates do not move smoothly; they tend to stick, causing strain to build up until it is so great that part of the fault jerks forward, causing a quake.
Advertisement
Some of the strain on the San Andreas Fault is transferred to smaller, subsidiary faults, many of which cause separate and sometimes damaging quakes as movement occurs along them. One example in northern California is the Hayward Fault, which lies parallel to, and about 25 kilometres east of, the San Andreas Fault. The Hayward, running through the densely populated cities of Oakland and Berkeley, east of San Francisco, is one of the most dangerous strike-slip faults in California. Based on its seismic history and rate of ‘slip’, geologists forecast that it is the likeliest of all the faults in the San Francisco Bay area to experience a major quake by the year 2020.
But the San Andreas and its subsidiary faults may not constitute the deadliest threat to Californians. The northern, San Francisco end of the San Andreas is not expected to unleash another quake on the scale of the 8.3 magnitude one that levelled the city in 1906 for a century or more, and the southern branch runs mostly through rural areas. The real danger to southern California’s many city dwellers may come from another source entirely: smaller, mostly invisible faults thought to lurk beneath the sediments of the Los Angeles basin.
Since 1983, southern California has repeatedly been clobbered by a series of quakes occurring on such ‘blind thrust faults’. These subterranean faults are given this name because they involve the vertical motion, or thrust, of one geological block over another. The thrusting motion occurs at a low angle – more horizontal than vertical – leaving the fault partly or totally concealed or ‘blind’, and revealing its presence only by its buckling of the landscape above. This type of fault moves in such a way as to create effects rather like that caused by sticking a knife horizontally just under the surface of a crumbly cake – it generates folds on the surface. The shapes of the hills and mountains created by this process provide geologists with indirect clues about the faults beneath.
HIDDEN THREAT
No one doubts the existence of these subterranean faults, not least because a large number of Californian quakes do not correspond to any known surface fault. Decades ago, seismologists blamed this anomaly on their not having enough of the instruments that detect seismic waves from earthquakes; now they suspect that an abundance of buried thrust faults is really to blame. ‘There are probably hundreds of faults beneath California not known or mapped carefully,’ says Muawia Barazangi, associate director of the Institute for the Study of the Continents at Cornell University in Ithaca, New York. He says this is due in part to their not having produced large quakes during recorded history.
Blind thrust faults are not restricted to California: they have wreaked havoc all over the world. ‘There are damaging thrust-related quakes that occur in all thrust belts,’ says Jay Namson, of Davis and Namson, a prominent geological consulting firm based in Valencia, California. ‘These belts go all the way down through the Andes of South America, through most of Central Asia – the Armenian earthquake of 1988 was probably on a thrust fault, and so are the earthquakes in Iran. In the whole Alpine-Himalayan mountain trend, certainly from Turkey going east to Burma, you’ve got one gigantic active fold and thrust belt.’
There is nothing special about California in this regard, according to Barazangi: ‘We get this type of fault all around the world.’ He says the reason the Californians were taken by surprise in January was that no large thrust-type earthquake had previously been recorded on the Northridge fault. ‘In the Middle East and China we have records of these types of quakes going back 2000 years. In California, we go back 200 years at best.’
In southern California, the thrusting results from a ‘bend’ in the San Andreas northwest of Los Angeles. This is where the Pacific plate, moving northwestwards, encounters the North American plate as it drifts southeast. The resulting collision creates folds in the sediments of the Los Angeles basin, like the folds formed when you push a rug against the wall. The folds have made the east-west mountain chain called the Transverse Ranges, including the San Gabriel, Santa Susana and Santa Monica Mountains. The Northridge fault is part of the system of thrust faults in and around the Transverse Ranges.
‘This band of east-west-trending faults is beginning to light up,’ says Thomas Henyey, head of the Southern California Earthquake Center (SCEC), a consortium of institutions based at the University of Southern California in Los Angeles. Yet no one knows why. The possibility that the increase in activity is merely the result of there being more seismometers in use is slim: recent blind thrust quakes – Coalinga in 1983, with a magnitude of 6.7; Kettleman Hills in 1985, magnitude 6.1; Whittier Narrows in 1987, magnitude 5.9; and Northridge in January 1994, magnitude 6.8 – have all been significant. One frightening possibility is that the thrust-fault activity is a harbinger of increased stress on the San Andreas. Whether it also heralds the Big One is still only a matter for speculation, Henyey cautions. But he adds that it’s worth remembering that both the thrust faults and the San Andreas are driven by the same set of forces – the grinding of the Pacific plate past the North American plate.
There are several reasons why thrust faults are scaring so many seismologists in California. Most of these faults are sneaky: they are largely or totally hidden, perhaps under miles of sediments. This makes it hard to tell how long they are – a vital clue to how much seismic energy they could unleash during a quake, as length apparently corresponds to the size of quake. If a fault is hidden, this also makes it more difficult to dig a trench across it, which is an increasingly popular seismological technique for determining how many times the faults have ruptured in the past, and so how likely that a quake will happen there again.
Worse, realistic computer models of thrust faults are extremely difficult to make, as their geometry appears to be more complex than that of strike-slip faults. Geologists discovered this distinction in the 1980s as they located more and more new examples of thrust faults. While the San Andreas runs in roughly a straight line from one end of the state to the other, the collection of buried thrust faults in the Los Angeles basin more likely resembles ‘a cluster or web’, Henyey says. He compares the web to a highway system that branches out from a single road on the outskirts of a town into a network of roads within it. Conceivably, a quake on one fault in the web could spread to another fault, then another, and so on. ‘We may never see all of them, may never know which one will quake next,’ Henyey says. ‘It’s a very complicated scenario.’
MAJOR MENACE
In action, thrust faults are also very destructive. During quakes occurring on such faults, there is often a lot of upward motion that pitches objects into the air. Once there, the objects are free to move horizontally too – and further than they would in strike-slip quakes, where all motion is horizontal but is slowed by friction with the ground. In the Northridge quake, instruments measured exceptionally strong and damaging horizontal accelerations of up to 1.82 times that of gravity.
Seismologists in California still seem a bit dazed by the emergence of thrust faults as a major menace on the quake scene. Until the 1980s, they were concerned largely with strike-slip faults – not surprisingly, as they are so obvious. ‘We were focused on the San Andreas and Newport Inglewood and San Jacinto; these are big strike-slip faults with prominent surface expressions,’ Henyey says. ‘You can see them on satellite photos. You can walk out in the field and stand on the fault. You can see where features have been offset along the faults. And our concept of earthquake faults was pretty much developed by looking at these big strike-slip faults.’
The triumph of plate tectonics theory may also help to explain why thrust faults had been ignored for so long. Back in the 1950s, geologists published the first sensational evidence that the San Andreas Fault had undergone a remarkable amount of horizontal motion: one side appeared to have been offset for more than 500 kilometres since the Late Cretaceous. This observation was explained in the 1960s by plate tectonic theory, which depicted the San Andreas as a transform fault that separates the Pacific and North American plates. ‘Once it was recognised that the San Andreas was such a big strike-slip fault, geologists devoted much of their attention to trying to fit all observations into a strike-slip model,’ says Namson. Anomalous thrust-type quakes were recorded, but received little attention. ‘For some reason (the contradictory observations) just didn’t get into textbooks,’ says Robert Yeats of Oregon State University, an early champion of the importance of thrust faults. ‘I was in graduate school in the 1950s, and there wasn’t much emphasis on quakes. They talked about quakes in terms of strike-slip motion – when they talked about quakes at all.’
UNEXPECTED HIT
Californian seismologists were forced to think again several decades later. Against all expectations, the town of Coalinga, roughly midway between San Francisco and Los Angeles, experienced a major earthquake in 1983. Eight people died and three-quarters of the town’s buildings were destroyed, yet there was no obvious active fault in the area. Then came the Whittier Narrows quake of 1987, which struck the town of Whittier, just east of downtown Los Angeles, and killed eight people. Again, no surface rupture was found. The Whittier Fault lay nearby, but its part in the drama was ruled out when the seismographs indicated that the hypocentre – the point where the quake begins, which tends to lie along the fault – was 14 kilometres underground. The Whittier Fault was too shallow to have been responsible. Researchers concluded that the Whittier quake was caused by a previously unknown, deep, low-angle thrust fault.
They were, however, only confirming an existing hypothesis. Less than a year earlier an oil geologist and colleague of Namson’s, Thom Davis, had inferred the existence of a thrust fault below Whittier. Davis suspected that southern California was full of such buried mysteries; and his views, regarded as offbeat in the early 1980s, are close to becoming orthodoxy in the wake of the quakes at Coalinga and Whittier Narrows. The Northridge quake merely ‘drove the nail into the coffin’, says Henyey. Now geologists are trying to map thrust faults all over southern California.
Like Davis, some geologists, especially those with oil companies, already knew about these faults and have even studied them for years in the course of their work. Yet seismologists did little about them until the Coalinga quake had shown that they were dangerous. The search is now gathering speed. In mid-February geologists funded by the US Geological Survey (USGS) and working for William Lettis & Associates, a geological consulting firm in Oakland, reported the possible existence of a band of blind thrust faults extending beneath about 150 square kilometres of the Santa Clara Valley in northern California, which includes Silicon Valley and the city of San Jose. The geologists’ analysis was based on a variety of clues, such as evidence of geological features being displaced in local creeks and aerial photographs that show lines of vegetation which may have been nurtured by water seeping from a fault.
Davis and Namson look for buried thrust faults by using a mathematical and geometrical technique called the ‘balanced cross section’. Oil companies have long used this technique to look for subsurface folds shaped like an upturned U, which can trap oil floating on top of the water table. The method involves comparing different surface folds across a region. From their knowledge of what size and shape of buried fault would have produced each type of fold, and by tracing common patterns among many folds, they can infer the existence of a single, large blind thrust fault. Oil drilling, which gives access to strata deep in the crust, provides additional information.
High-tech equipment is also being used in the search for buried thrust faults. One device detects hidden faults from the way they reflect sound waves generated by dynamite explosions or by trucks with ground-thumping machines. Satellites and surveying equipment are used to detect subtle differences in height, often a matter of centimetres, over the surface of the ground, as these may be a result of thrust movements far beneath the surface. Oil seeps are also investigated to determine if they’re linked to subsurface folds resulting from thrust faults.
The oil companies may have a head start in the hunt for thrust faults; but they haven’t yet agreed to share much of their data with outside scientists. Firms that have spent millions of dollars gathering such data over the decades are reluctant to share their investment with the public – which would inevitably include their commercial adversaries. The firms may also simply have a lot of trouble collecting and organising the data: ‘We know they have large amounts of data, (but) we don’t know how well the data has been archived,’ Henyey says. ‘Much of this data is in file cabinets somewhere or filed on (magnetic) tape somewhere.’
The work of Davis and Namson is part of a larger project, currently run by the SCEC and funded by the National Science Foundation in Washington DC, to develop a ‘master model’ of quake activity in southern California. The model will be constantly updated with new information gleaned from quakes as they occur. Even with this model ‘we won’t be able to tell you exactly where the next earthquake will occur,’ Henyey acknowledges. ‘But we will tell you which areas are most likely to be earthquake-prone.’
Keay Davidson is science writer for the San Francisco Examiner.
* * *
The subtle art of quake-casting
No one saw the Northridge quake coming. No unusual swarms of quakes, peculiar emissions of gases from the ground, puzzling fluctuations in local electromagnetic activity or odd animal behaviour preceded it. There were none of the ‘precursors’ that, according to seismological observation and popular legend, have forewarned of other quakes. Any hope of short-term prediction of earthquakes may have gone for good.
However, US Geological Survey scientists such as Allan Lindh still have high hopes for long-term quake forecasts. These estimate the probability of a major quake on a given fault over a long period of time, say 30 years. The Loma Prieta quake in the San Francisco Bay area in 1989 may have been an example of a successful probability cast. Before it struck, Lindh and others had said the Santa Cruz Mountains – a range which includes Mount Loma Prieta – was the likeliest spot for a major quake in northern California. Sure enough, a quake with a magnitude of 7.1 hit there in October 1989, killing 65 people and causing billions of dollars’ worth of damage.
The most famous of all quake-casts was not quite so successful. In the 1980s, USGS experts predicted that the area around the village of Parkfield in central California would experience a quake of about magnitude 6.0 sometime around 1988, and no later than 1 January 1993. They based the forecast on the village’s unusual history: it has experienced a significant quake once every 22 years on average since the middle of the 19th century. But New Year’s Day 1993 came and went . . . and still no quake. Nevertheless, they are leaving their expensive seismometers and other equipment in place in the belief that it will happen sooner or later, probably before 2000.
Might the Loma Prieta quake have given other hints of its coming? In late 1993, a report edited by Malcolm Johnston, a seismologist with the USGS, discussed possible ‘precursors’ of Loma Prieta. These included unusual pre-quake activity at a geyser in Calistoga and a surge in the natural electromagnetic field south of San Francisco 12 days before the quake. Coincidences? Perhaps. But for now, they rank among the slender reeds that support the diehard advocates of quake forecasting.