By Sid Perkins for Nature magazine
Beginning in late 2004, a flurry of massive, tsunami-spawning earthquakes have rocked the world, first slamming Indonesia, then Chile and most recently Japan. Temblors that size are rare indeed: only 7 quakes as large or larger than 8.8 -- the magnitude of last February's Chilean event -- have occurred since 1900.
So what does it mean that three of those seven shocks have happened almost within the span of six years? While some scientists argue that these 'megaquakes' could be the vanguard of an extended outburst of strong seismic events, many others suggest that the apparent cluster of recent temblors is nothing more than a statistical fluke.
The recent spate of far-flung quakes is remarkably similar to a cluster that occurred in the middle of the last century, says Charles Bufe, a seismologist retired from the US Geological Survey (USGS) in Denver, Colorado. The seismic events in that supposed grouping, consisting of 3 magnitude 9 or higher temblors, struck Kamchatka, then Chile and then Alaska within a 12-year interval. The odds of quakes that large occurring randomly within such a short time span is only four per cent, Bufe noted today at the annual meeting of the Seismological Society of America in Memphis, Tennessee.
In an update to an analysis first published in June 2005, Bufe and colleague David Perkins, a USGS geophysicist also in Denver, argue that the most recent round of large temblors may mark the beginning of a new global outbreak of megaquakes. According to their model, Bufe says, the probability of another quake of magnitude 9 or larger striking in the next 6 years is about 63 per cent. "There's now an increased hazard situation for these very large earthquakes," he notes.
It's an ominous warning considering what scientists are now reporting about the monumental forces behind the 9.0 Tohoku quake, which wreaked devastation on Japan on 11 March. At the meeting, researchers revealed that the main shock ruptured a previously locked seismic interface more than 250 kilometres long and 175 kilometres wide. Although most of the quake's energy was released in the first 2 minutes, several aftershocks -- many of them magnitude 6.4 or larger -- occurred in the 20 minutes or so that followed. Altogether, that temblor and its aftershocks ruptured areas that had previously slipped in five separate quakes, the researchers say. As a result, the entire northern portion of Honshu, Japan's largest island, moved about 1 meter toward the east, with one site near the temblor's epicentre sliding 5.4 meters horizontally and sinking 1.1 meters -- a sudden subsidence that aggravated the damage from the tsunami that slammed the shore minutes later.
Yet the apparent clustering of such megaquakes, including the recent Indonesian, Chilean and Japanese events can be accounted for without a direct link, several scientists say. "When you run statistical tests, you can often get numbers that sound interesting," says Richard Aster, a geophysicist at the New Mexico Institute of Mining and Technology in Socorro. In this case, he suggests, the clumping could come down to the statistics of small sample sizes. Since 1900, there have been only 14 quakes larger than magnitude 8.5. And whereas modern seismology goes back only a little more than a century, the tectonic processes that generate major earthquakes unfold over hundreds or thousands of years, he adds.
Looking for patterns
In a separate analysis, Andrew Michael, a seismologist with USGS in Menlo Park, California, scrutinized databases of major quakes for evidence of clustering. Rather than use a single threshold for earthquake magnitude, he ran several statistical analyses using different magnitude thresholds, looking for patterns in quakes over various intervals ranging in length from months to years -- and found nothing. "I've run a large number of tests and can't find any reason to reject the idea that clustering is random," he says.
That's not to say that major quakes don't stimulate further seismic activity. Barely 4 months after the December 2004 quake struck Indonesia, a magnitude 8.6 temblor occurred just down the coast -- the result, scientists say, of the first quake's redistribution of stress in Earth's crust. Usually, the extent of such stress shifts are limited to the immediate region, says Aster. Although there is evidence that the ground motions induced by major temblors trigger small quakes thousands of kilometres away, there's no sign that such triggering occurs for large quakes, he adds.
Research published online 27 March in Nature Geoscience bolsters those notions. Tom Parsons, a seismologist also with the USGS in Menlo Park, and colleague Aaron Velasco, of the University of Texas at El Paso, analyzed the USGS earthquake database to see if temblors of magnitude 7 and higher might have triggered midsized quakes elsewhere in the world. Between 1979 and 2009, seismometers recorded 205 quakes with magnitudes above 7, Parsons notes. Although many of those quakes triggered local aftershocks in the day or so after the initial event, Parsons and Velasco found no corresponding increase in the frequency of distant quakes with magnitudes ranging between 5 and 7.
The team's analysis also suggests that stress redistribution to nearby faults after a major quake is limited to distances from the epicentre no more than two or three times the length ruptured by the original quake. That, says Parsons, means that even megaquakes shouldn't trigger large quakes more than a couple of thousand kilometres away.
This article is reproduced with permission from Nature magazine. It was first published on April 14, 2011.