The Mojave Desert, 65 000 square kilometres of parched wilderness, is Los Angeles’s back yard. Its wide alluvial basins and black mountain ridges are a haven of tranquillity to the northeast of California’s most frenetic city. But on 28 June 1992, the peace was broken by an earthquake measuring 7.3 on the Richter scale that shattered the desert surface like a porcelain plate. Beginning near the tiny town of Landers, about 100 kilometres east of the city of San Bernardino, the earthquake fissured a zone 85 kilometres long and 2 kilometres wide. Since then it has generated around 60 000 aftershocks and ratcheted up the stress on nearby faults.
At the time the Landers quake hit, seismologists had been expecting a large quake somewhere in southern California but most predicted that this would be 50 kilometres further east, on the infamous San Andreas Fault. No one had imagined that the largest earthquake to strike the area in 40 years would tear through a jumble of minor, seemingly unconnected faults, several of which were unmapped.
Within days, a small army of geologists had descended on the Mojave to tackle the mystery beneath the sand. Meanwhile, other scientists were turning to theoretical models. Now, researchers working in different disciplines have come up with opposing explanations of the earthquake. The controversial conclusion from a group of geophysicists from Stanford University is that the Landers rupture is part of a new zone of increased earthquake activity that will ulti-mately recarve the continental margin leaving it some 400 kilometres east of its present location. But these findings are based more on theory than on field observations. Many geologists familiar with the desert dispute them.
Advertisement
Researchers in the field maintain that the faults around Landers are millions of years old and are unlikely to rupture again for several thousand years. They argue that the hypothesis of the Stanford team ignores the geological evidence on the ground, misuses data and is inconsistent with geological features of the desert itself. These field geologists view the quake at Landers as just another seismological event in an area that has been prone to them for at least five million years.
QUAKE EXPECTATIONS
The dispute is more than academic. The Landers earthquake shattered not only the desert, it also rocked the reassuring notion that the most dangerous faults in this state of 30 million people were already mapped and under vigilant scrutiny. Some geologists, however, warn that a misreading of the significance of Landers could divert attention from California’s greatest seismic menace – the ancient, yet active, San Andreas Fault.
The area that is now California has been chiselled into shape over the past 30 million years by the impinging motion of two huge crustal plates: the Pacific plate, which grinds northwestward, and the North American plate, moving in the opposite direction. Most of this relative motion is concentrated in the San Andreas Fault, which runs the length of the state.
On parts of the fault, the plates creep at an almost imperceptible 35 millimetres a year. But other segments move in violent lurches, rupturing every 40 to 100 years. During these so-called ‘strike-slip’ earthquakes, the fault slips up to 10 metres horizontally. In the past two centuries, there have been 25 major earthquakes along the fault measuring 6 or more on the Richter scale. The infamous San Francisco earthquake of 1906, with a magnitude of 8.3, demonstrated the deadly power of the San Andreas. Around 2500 died in this quake and the fires that followed.
In its early life, about 10 million years ago, the San Andreas Fault lay offshore, well to the west of its present location. It is still moving today, and many seismologists fear that this is slowly causing it to lock up – crippled like an old man with a crooked spine. The problem lies north of Los Angeles, in the Transverse Ranges, where the fault swerves westward. Experts believe that this bend impedes the easy slipping motion of the fault. Opinion is divided on just what this means, however. Most seismologists believe that the San Andreas Fault is still alive and kicking, and likely to continue moving fast for millions of years. Others believe the bend is indirectly causing new faulting, much of it still unidentified. Kevin Furlong, a specialist in plate tectonics from Pennsylvania State University, thinks it may be causing new faults to form just south of the Mojave Desert, near the Salton Sea. And Stanford geophysicist Amos Nur believes the bend was implicated in the Landers quake.
But this is not the whole picture. To understand what happened at Landers it is necessary to consider the broader geology of southern California. Research over the past decade has shown that about a quarter of the motion between the two plates does not occur along the San Andreas Fault at all. It is spread out over a region that extends across the state to its borders with Nevada and Arizona, forming the Eastern California Shear Zone. Here, the migration of the plates has created a network of smaller faults, carving the land into adjacent crustal blocks that thrust and stretch the surface.
The Mojave Desert sits smack in the middle of the shear zone and the accumulated effects of its motion are evident. It is made up of a mosaic of mountains, valleys and wide basins. The mountains, that have uplifted in successive stages, may be more than 100 million years old. The alluvial plains between them are very young – no more than 12 000 years old – but they are already subtly contoured into mounds and shallow depressions by earthquake fissures now buried in the sand.
After 20 years of field studies, Roy Dokka, a geologist at Louisiana State University in Baton Rouge,has pieced together a geological history of the desertthat spans 24 million years. During that time, each of the three major geological events has left its mark: a period of explosive volcanism which, starting about 24 million years ago, was followed by a monumental folding of the desert crust 19 million years ago, and finally the creation of the Eastern California Shear Zone over the past 5 to 10 million years.
The force and complexity of the 1992 Landers earthquake surprised earth scientists, despite their detailed knowledge of the area. It was also unusual in that it hopscotched over five known faults – Johnson Valley, Homestead Valley, Emerson, Camp Rock and Galway Lake – as well as carving out previously unknown faults, including the Kickapoo. Even today, traversing the faults, every other step straddles a partially filled fissure, one of thousands that splay off the main breaks. The ramified array astonished scientists accustomed to thinking of earthquakes occurring along a single fault line.
Nur believes there is a simple explanation to this surface complexity. Together with Gregory Beroza, also at Stanford, and Haigon Ron at the Institute of Petroleum Research and Geophysics in Holon, Israel, he has divided the fault lines into two categories, depending on their age. The more ancient lines, which are several million years old, tend to run northwestwards and are losing their ability to release tectonic stress. It is the younger north-south trending fault lines, possibly just a few tens of thousands of years old, that interest Nur and his colleagues.
Nur believes that the north-south Landers ruptures, and five similar earthquakes that occurred in the region over the past 50 years, may be the first evidence of a major fault that could realign the boundary between the Pacific and North American plates. He predicts that the 120-kilometre-long Landers-Mojave line may one day grow to replace the San Andreas Fault as the main plate boundary. This emerging fault, says Nur, extends up the east side of the Sierra Nevada mountains, and somewhere near Oregon will step back over the Pacific Coast and plunge into the oceanic crust to rejoin its northern counterpart. ‘It’s already happening,’ he says.
According to Nur’s model, blocks of crust sandwiched between the northwest trending desert faults have slowly rotated out of a position that allows them sufficiently to release tectonic stress. This stress builds up as crustal masses press against each other. The direction of compression is determined by the relative motion between abutting plates. Nur explains that when rock is compressed, it typically fractures at an angle of 30 degrees away from the direction of compression. So, active faults form at a predictable angle in relation to the direction of tectonic stress.
But this picture is complicated by the rotation of blocks of crust within the plates. As plates rarely move directly parallel to each other, rotations occur when the jagged edges of smaller blocks get snagged and swept up by larger plate motions. Nur contends that such block rotations are causing the old desert faults to lock up and new ones to open.
But why are these new faults not oriented in a northwesterly direction? Nur’s answer is that the direction of compression itself has changed as a result of large-scale interactions between the North American and Pacific plates. From the angles of fracture of the younger faults in the Landers quake, Nur calculates that this direction of compression has rotated about 15 degrees clockwise over the past few million years.
PETRIFIED COMPASSES
Nur’s evidence for rotating blocks comes primarily from data embedded in rocks. As molten rocks solidify, iron-rich minerals within them align with the Earth’s magnetic field, which runs approximately north-south. The angle of these minute compasses varies predictably with latitude. If the crust in which the rocks reside rotates, the compasses will be thrown out of alignment. Nur interprets the palaeo-magnetic data as showing that since the older faults formed, blocks of crust rotated up to 15 degrees anticlockwise in the central Mojave and between 40 degreees and 50 degrees clockwise in the eastern Mojave.
Nur’s work combines palaeomagnetic field evidence, measurements of the stress field on site, and theories of rock physics extrapolated from laboratory experiments. He is convinced that the only way to advance the understanding of how earthquakes occur and plates break is to combine the disciplines of rock mechanics and field geology.
But not everyone is convinced. Many geologists who know the Mojave well have challenged Nur’s figures. They include geologist Ray Wells from the US Geological Survey (USGS), who spent four years in the desert chasing down the Peach Spring tuff – an ash flow deposited from a glowing volcanic avalanche about 18 million years ago. Its remnants are chunks of basaltic rock strewn across the desert, which are embedded in sand or dug out of road cuts. From its origin near the lower Colorado River, the lava stream flowed 400 kilometres, dissecting the central Mojave area. Palaeomagnetic data from this flow show inconsistent rotations of less than 13 degrees, going both clockwise and anticlockwise. The net rotation was generally zero, he says. ‘This doesn’t negate Amos’ story,’ Wells adds, but it does place more importance on the change in the direction of stress.
Some experienced field workers are less forgiving. Dokka claims that Nur’s team misuses the palaeomagnetic data, by lumping together information that spans 22 million years. The northwest trending faults are less than five million years old, so crustal rotations before that time are immaterial, says Dokka. Another critic is David Schwartz of the USGS, who also has decades of experience in field research. He believes that geologists cannot rely on data alone, they must also be directed by subtle intuitions to guide their interpretations of the Earth. ‘In field geology,’ he says, ‘you have to see something in three spatial dimensions and in time, with an intuitive understanding of the process.’
PIECING IT TOGETHER
Although the palaeomagnetic data appear contradictory, some scientists believe that indirect evidence supports Nur’s hypothesis. Among them is Ross Stein, a geophysicist with the USGS. He notes that one of the largest earthquakes in California’s history ruptured the entire 100-kilometre length of the Owen’s Valley, east of the Sierra Nevada. On 26 March 1872, this 7.6 magnitude earthquake occurred not on the San Andreas, but a couple of hundred kilometres north of Landers, in the middle of Nur’s proposed new plate boundary.
Furthermore, the Landers rupture has the same appearance as other emerging young faults, Stein says. New faults may initially rupture on a small scale, with rough broken fractures, and then coalesce to form a well-worn line like the San Andreas Fault. This was seen, for instance, in the 7.5 magnitude Hebgen earthquake in Montana which occurred on faults Stein describes as ‘essentially brand-new – around 300 000 years old’.
In order to build up a picture of how the Landers faults will develop, a team of USGS geologists led by Schwartz is trying to piece together a complete history of their past. If they can prove that the north-south oriented faults have a long history, with consistent patterns of rupture, Nur’s theory could be scuppered.
Since July 1992, Schwartz’s team has excavated six trenches on the Homestead Valley Fault, work that resembles a painstaking archaeological dig. After opening each trench and bolstering its walls, the geologists scrape them clean, put up a grid system and draw detailed maps of the earthen sides. They examine the rock layers for places where faulting once ruptured to the surface and has since been buried.
Signatures of ancient earthquakes include features that at one time spanned the fault but which are now fractured. By measuring the distance separating two rock segments that were once continuous, geologists can determine how far the opposing sides of a fault slipped. Earthquakes can be carbon-dated if organic material such as charcoal is present.
Schwartz has dated three earthquakes that occurred before the 1992 event. The trench that revealed the most data was cut across a dry lake bed that was once fed by the runoff from surrounding mountains, building up layer upon layer of sediment. Fractures in these layers provide a vivid picture of ancient earthquakes. Schwartz found one earthquake that occurred between 6000 and 8500 years ago, one dated to about 12 500 years ago, and an even earlier one which he has not yet dated. The geologists identified similar earthquake features in their other digs. ‘The work shows that for the last 15 000 to 30 000 years, very clearly, things have ruptured in essentially the same place, with the same sense of movement,’ Schwartz says.
Surface features suggest that earthquakes were occurring even earlier, he adds. One such feature is a sloping hillside that runs down to the Emerson Fault. This so-called ‘alluvial fan’, formed by rain-washed deposits, has been displaced by about 300 metres on either side of the fault. Schwartz estimates that if the Emerson Fault ruptured at 6000-year intervals, with the ground slipping horizontally three metres each time, then the total offset would have accumulated over the course of 600 000 years.
‘So, on a timescale of hundreds of thousands, perhaps even several million years, these faults have been very stable in their behaviour,’ Schwartz concludes. He is convinced that the north-south trending faults, which Nur claims are new, are ancient minor faults that were reactivated. But, says Nur, there are ancient faults in all directions. ‘Why weren’t they all activated? Why only these two distinct directions relative to the state of stress?’
Schwartz plans another season of field work, during which he will study more sites to get a wider picture. Meanwhile, two other teams are excavating other faults that ruptured during Landers. Thomas Rockwell at San Diego State University leads one team, Kerry Sieh at the California Institute of Technology in Pasadena leads the second. Their work shows a large degree of consistency between faults. Several faults, for example, record either a cluster of small earthquakes or one big earthquake around 9000 years ago, says Sieh. ‘It could be that the Eastern California Shear Zone has a flurry of activity over a few hundred years and then is quiescent for thousands,’ he says. ‘It’s possible we’re in a period of activity.’
UNCERTAIN TIMES
Until more data are gathered, uncertainties persist. Even Schwartz admits that his methods are not beyond question. It is not always clear, for example, that the age of material gathered from layers in an excavated trench represents the age of earthquakes themselves. A piece of charcoal could have formed at one time and been transported many years later to the site where the earthquake occurred, and where it is now buried.
Nur is unruffled by the accumulating palaeoseismic evidence, and continues to defend his more theoretical approach. ‘Faults just don’t form randomly by waiting a long time,’ he says. ‘The process is constrained by the mechanical behaviour of rocks and stress.’ Nur believes that Landers was devastating to the science of earthquake prediction. ‘Forget about time and magnitude,’ he says. ‘Not even the location was predicted . . . Landers happened on a fault that had never been mapped, at a magnitude that’s completely inconsistent with an unmapped fault. If we don’t figure out what happened here, it’s a major setback – and it’s sheer luck that Landers was not densely populated.’
Around 200 people live within a 30-kilometre radius of the Landers earthquake, which killed one person. The nearby towns of Victorville and Barstow boast populations of 20 000 and 50 000. And five million people live less than a two-hour drive away, in the San Bernardino County. That population has sprawled east from Los Angeles, crowding right up to the San Bernardino Mountains, which are crowned by the San Andreas.
We know now that the Landers earthquake increased the stress on that part of the San Andreas. Stein calculates that while Landers eased stress on some faults, it brought the section that borders San Bernardino County some ten years closer to failure. Nobody, however, knows exactly when that earthquake will occur – just that it will be sooner than would have been the case. Stein thinks this information alone merits a shift in priorities on the part of organisations involved in seismic studies.
Although Nur’s hypothesis may be unprovable on human timescales, unearthing the secrets under the desert could clarify tectonic processes still in the making. Since the Landers earthquake, organisations such as the Southern California Earthquake Center, have gradually turned more of their energies eastward. The recent Northridge quake, however, seems likely to refocus research efforts on the Los Angeles basin.
Bernice Wuethrich is a freelance writer based in Baltimore.