快猫短视频

FOR YEARS, the US Geological Survey had a crisp, easy answer to a perennial question about earthquakes. Whenever one quake followed on the heels of another, reporters would invariably ask whether they were connected. 鈥淎bsolutely not,鈥� would come the brisk reply. 鈥淓arthquakes happen randomly.鈥�

But now, seismologists are not so sure. On 28 June 1992, a major earthquake struck the sleepy town of Landers, California, 200 kilometres from Los Angeles, and within 24 hours, a rash of tremors broke out in four other states, the most distant in northwestern Wyoming, more than 1000 kilometres away. Ever since, researchers have been struggling to understand whether the tremors were triggered by the Landers quake, and if so, how.

Seismic ripples

Their interest is more than purely academic, because triggered tremors-if that is what they are-may warn that a really large quake is on the way in the same place. Last autumn, Lowell Whiteside of the National Geophysical Data Center in Boulder, Colorado, and Yehuda Ben-Zion of Harvard University reported a phenomenon that hinted that an earthquake could trigger tremors right around the Earth.

The shock of a massive quake generates seismic waves. Some, known as body waves, travel directly through the Earth, while others zoom around its surface. The surface waves spread out in all directions from the earthquake鈥檚 focus, like water waves created from a pebble dropped into a pond. But because the Earth is a sphere and not a flat plane, the waves moving in different directions eventually converge-interfering with each other. Where the waves are in phase with one another, they combine to build up in amplitude; those that are out of phase cancel each other out. These interference patterns, called 鈥渇ree oscillations鈥�, cause the Earth to wobble and shake, like a balloon filled with water when it is struck. And it is free oscillations that are the key to Whiteside and Ben-Zion鈥檚 theory.

Free oscillations from a major earthquake cause the Earth to ring like a bell for days or weeks. Its natural frequencies fall far below our hearing range, though. The planet鈥檚 tones, also called its fundamental modes of oscillation, have periods of about 54 and 44 minutes, respectively-basso profundo indeed.

Seismologists have known about free oscillations for decades, and have used them extensively to study the Earth鈥檚 interior. But until Whiteside and Ben-Zion came along, no one seems to have thought of connecting them with earthquake prediction. The team was drawn to free oscillations for a simple reason: they add small, periodic stresses to the Earth鈥檚 crust, and the researchers wondered whether these tiny stresses might be enough to trigger small tremors.

Neither Whiteside nor Ben-Zion believed that free oscillations could cause tremors that were not otherwise on the cards-the stresses they add are simply too small. Yet along highly stressed faults, where small earthquakes are already common, the researchers suspected that free oscillations might nudge the rupture process along- shifting the timing of the inevitable quakes by only a few minutes perhaps, but causing the rocks to slip in synchrony with the oscillation periods.

But why would anyone bother researching such a puny effect? Paradoxically, the very weakness of free oscillations is the source of their strength, at least for earthquake prediction. Since the oscillations add only a small amount of stress, they can only influence the very weakest spots along faults. Those are precisely the regions that earthquake predictors are keen to identify.

These weak spots often start to crack up before a major quake. Seismic records reveal that major quakes are preceded by a number of 鈥渇oreshocks鈥�, often by weeks or months. But tremors occur around the globe all the time, and without the benefit of hindsight, such foreshocks are indistinguishable from seismic activity of other kinds. This is where Whiteside and Ben-Zion鈥檚 ideas come in. They suggested that major quakes-or rather, the free oscillations from major quakes-might selectively trigger foreshocks in highly stressed regions. Thus, if seismologists noticed a high proportion of triggered tremors in a particular area, they鈥檇 be forewarned that a 鈥渂ig one鈥� was in the making.

Of course, it鈥檚 one thing to concoct an elegant theory, and quite another to show that it describes what really happens in nature. The first thing Whiteside and Ben-Zion attempted to show was that free oscillations were indeed capable of triggering tremors. They began by studying records of the aftershocks that followed large earthquakes around the world. When a big quake rearranges the rocks along a fault, it creates some new patches that are highly unstable. The researchers suspected that the free oscillations from the main shock would come back to haunt the fault zone again, triggering aftershocks in these fragile spots.

Free oscillations exert their pulses of maximum stress at predictable moments after the initial earthquake-so if there were any triggering, the aftershocks should coincide with the pulses of maximum stress. After systematically scouring nearly 400 aftershock sequences following major quakes, Whiteside and Ben-Zion found just this pattern. For a statistically significant proportion of aftershocks, the time between shocks neatly coincided with the expected free oscillation periods.

Warning signs

Encouraged, Whiteside and Ben-Zion tackled the more provocative parts of their theory. Could the oscillations from a big quake in, say, Baghdad, trigger tremors in Berkeley, California? And, most importantly, could the tremors forecast a fault鈥檚 future? To answer this, they turned again to earthquake catalogues from around the world.

The data they studied most intensively were records for California between 1980 and 1995. They divided each region into 10-kilometre squares, noted the time intervals between all the tremors that happened in those squares each month, and then combined the monthly data into plots covering nine-month periods. Then they compared the time intervals between the earthquakes with the timing of free oscillation periods generated by the 15 to 20 major quakes-generally magnitude 7 or above-that occurred around the world each year. Once again, the results supported the idea that free oscillations were triggering a significant portion of the Californian tremors.

Better still, crescendos of small, triggered quakes and tremors did seem to warn of big quakes that followed. In periods of up to about nine months before a big quake, the number of apparently triggered tremors markedly increased within 10 kilometres of the epicentre of the future big quake. A high incidence of triggering preceded many of the major California quakes over the 15-year period, including the mysterious magnitude 7.3 Landers quake in 1992 and the magnitude 7.1 Loma Prieta earthquake that rocked the San Francisco area in 1989. Whiteside and Ben-Zion later investigated records from Japan, and found a similar triggering pattern before the devastating Kobe quake of January 1995. All told, triggering occurred before 13 of the 15 major shocks they studied.

However, the patterns leave some unanswered questions. For instance, there is no clear relationship between the number of triggered tremors and the precise arrival time of the subsequent big quake. 鈥淥nce triggering reaches a statistically significant level, it means the earthquake could really happen almost any time,鈥� says Whiteside. 鈥淪o obviously, you鈥檙e not going to know that at 8 o鈥檆lock the next morning you鈥檙e going to have an earthquake. The best you can say is, `We know an earthquake is very close. Be prepared over the next several months鈥�.鈥�

When Whiteside presented these findings at a meeting of the American Geophysical Union last autumn, the work met with generous measures of both keen interest and doubt. Terry Tullis, a geophysicist at Brown University who heard the presentation says: 鈥淚 thought, `This is really exciting stuff.鈥� The triggered earthquakes seemed to be pointing out the major earthquakes, like Loma Prieta, Landers and Kobe. The patterns could be random chance, but it was just too tantalising to forget.鈥�

Small stresses

However, Tullis adds, some scientists think the forces involved are simply too weak. Whiteside estimates that the free oscillations trigger small quakes and tremors by subjecting fault zones to stresses of less than 0.1 bar. Many seismologists find it hard to believe that such a small stress could trigger a tremor. They point out that there is at least one other mechanism that causes far stronger stresses but that doesn鈥檛 seem to trigger quakes. 鈥淓arth tides鈥�-the gravitational pull of the Moon and Sun on the Earth鈥檚 crust-subject the crust to about four times the stress caused by free oscillation pulses. And yet, scientists have so far found no convincing link between Earth tides and earthquakes. 鈥淚f Earth tides don鈥檛 trigger earthquakes, how could the free oscillations do it?鈥� Tullis asks.

The question of Earth tides is a problem for Whiteside and Ben-Zion鈥檚 theory. But ironically, it was a study of Earth tides that recently came up with results that strongly support their ideas. The new work, published in the April issue of the Bulletin of the Seismological Society of America, began as a hopeful search for triggering by Earth tides. It ended up as a demonstration of triggering by free oscillations.

Seismologists Lalu Mansinha and his student Kamal, of the University of Western Ontario, Canada, hoped that they might succeed where other studies on Earth tides had failed. No one had managed to find a correlation between the timing of large earthquakes and Earth tides, so Mansinha and Kamal decided to search instead for a link between Earth tides and aftershocks-specifically, the thousands of aftershocks that followed the 1989 Loma Prieta quake. They supposed that the tides, like Whiteside鈥檚 free oscillations, would have a 鈥渓ast straw鈥� effect, triggering the aftershocks on faults that were already close to breaking point.

Since Earth tides exert their maximum stresses roughly every 24, 12 and 8 hours, the team expected to see a higher concentration of aftershocks at those intervals. They didn鈥檛. 鈥淲hen we failed to see the tidal periods, we were initially quite disappointed,鈥� says Mansinha. 鈥淚n fact, Kamal and I had long discussions. He said, `Obviously it鈥檚 not working,鈥� and I said, `Well, let鈥檚 look at the remainder of the record.鈥� He did this with great reluctance because he was convinced there was nothing there. Then he saw it, and I saw it, and we both said, `My God!鈥�.鈥�

What they saw were strong peaks-a clustering of aftershocks-at intervals of 55.4 minutes, 43.2 minutes and 27.7 minutes, corresponding with three free oscillation periods. They recognised the peaks immediately, Mansinha says, because they just happened to have conducted unrelated work on free oscillations the year before. Intrigued, they used a statistical program to analyse the data in more detail. To their amazement, they found that free oscillations triggered about 10 per cent of the aftershocks that occurred within 6 days of the main Loma Prieta earthquake.

Unfortunately, the data still do not shed much light on exactly how free oscillations could have caused the aftershocks. But as Mansinha says, this is not the first time that nature has challenged seismologists with this sort of problem. The strains generated by the Landers quake, for example, must have been much weaker than Earth tides, and yet they somehow triggered quakes at astonishing distances. Stranger still, whatever the forces were that ruptured the distant faults, they left the fault next door to Landers, the highly stressed San Andreas, untouched.

Indeed, the strange pattern of events following the Landers quake suggests that a number of factors-not just the magnitude of stress-must determine whether a fault ruptures or holds. Joan Gomberg, a US Geological Survey seismologist at the University of Memphis, Tennessee, believes that the rate at which the stress is applied may be critical. Some earthquake models show that even a small stress, if applied suddenly, can set off a kind of chain reaction that triggers an earthquake, she says.

The finding that free oscillations trigger tremors-while decidedly controversial-at least fits with this basic idea. Free oscillations do apply their stresses far more abruptly than Earth tides (with oscillation periods of less than an hour compared with Earth tides鈥� 8, 12 or 24 hours), and this could make them more effective in setting a rupture process in motion. And since free oscillations repeat so many more times each day, they also have many more opportunities to jiggle a highly stressed fault into failure.

Twisting rocks

Mansinha also believes that some free oscillations may apply a special pattern of stress that encourages faults to fail, while Earth tides may not. Earth tides alternately pull and push, stretching and then compressing the planet towards the shape of a rugby ball. While some free oscillation modes deform the Earth in a similar fashion, others start the globe twisting, with the top and bottom hemispheres going in different directions. Mansinha argues that faults, where two bodies of rock slip past one another, might be especially vulnerable to the effects of such twisting motion.

鈥淚t鈥檚 like when you dance the twist,鈥� says Mansinha. 鈥淭he top part of your body is going one way, and the bottom is going another way. Your waist isn鈥檛 moving, but it has a good deal of stress. That鈥檚 why if you twist too much, you鈥檒l feel the pain.鈥� The fault zones most vulnerable to triggering by free oscillations may turn out to lie within the waist, where the twisting motions nudge the two sides of the fault to slip in opposite directions.

The nature of faults themselves might also enhance the impact of free oscillations, Mansinha speculates. Like slits in a piece of paper, the faults open the Earth up, creating more surfaces than occur in unbroken pieces of crust. The repeated sharp vibrations from free oscillations are especially violent and powerful in these breaks. Earth tides, by contrast, apply slow, steady stresses to broad areas of crust, so they don鈥檛 have any especially powerful effects on faults.

Of course, even if free oscillations do trigger some small quakes and tremors, as Whiteside, Mansinha and their collaborators suspect, this doesn鈥檛 mean that they will necessarily turn out to be the key to earthquake prediction that seismologists are seeking. Many other suspected precursors have appeared promising in the past, only to fail under closer scrutiny. But Tullis is sufficiently intrigued by the work of Whiteside and Ben-Zion to give it that careful look. He hopes first simply to replicate their results, examining their data more closely to make sure their statistics are watertight. If possible, he鈥檇 also like to test their method for 鈥減redicting鈥� small to moderate quakes that occurred in 1993 and 1994 in Parkfield, California.

But it is early days yet, and not even Whiteside is prepared to guess the outcome. 鈥淚t鈥檚 like any other idea in science,鈥� he says, 鈥渋t may turn out that it isn鈥檛 anywhere near as good as it might look on the surface. Or it may have possibilities that nobody has even thought of.鈥�