杏吧原创

BEATRICE MAGNANI spends her days navigating the Mississippi river in a US Army Corps of Engineers vessel that tows an airgun and a hydrophone. 鈥淚t鈥檚 kind of a Mark Twain thing,鈥� she says. Every 7 seconds, the airgun pops, expelling a bubble of pressurised air into the sediments beneath the river bed.

Magnani uses the pressure and timing of the reflected waves to create a picture of what lies beneath the Mississippi鈥檚 murky waters. In a geologically quiet continental interior such as the US Midwest, sediments of different ages should be stacked in layers as neat as those of a Black Forest gateau. Under the Mississippi, however, they are not 鈥� in places, they are broken or folded in on themselves. 鈥淪omething must have deformed them after they were deposited,鈥� says Magnani, .

Something like a huge earthquake. Exactly 200 years ago, between 16 December 1811 and 7 February 1812, a series of four massive quakes ripped through the Mississippi embayment, a low-lying, sediment-filled basin stretching from the Gulf of Mexico northwards to Cairo, Illinois. in present-day Missouri, the quakes measured around magnitude 7 on modern scales, and possibly as much as magnitude 8. In the last of them, the Mississippi river flowed backwards, the riverbanks spewed sand, and Reelfoot Lake 鈥� today a popular hunting and fishing preserve in north-west Tennessee 鈥� formed as the ground opened to swallow displaced water.

That, on the face of it, is rather unexpected. New Madrid lies far from typical arenas of major seismic upheaval, where one of Earth鈥檚 tectonic plates meets another. But the earthquakes there were no unique occurrence. In 1556, the most deadly earthquake on record occurred in Shaanxi province in China鈥檚 northern interior, again nowhere near a plate boundary. Some 800,000 people were killed as, according to a contemporary report, 鈥渕ountains and rivers changed places鈥�. On 23 August last year, a magnitude 5.8 quake struck with an epicentre near Mineral, Virginia. There were no deaths but the incident caused chaos and confusion up and down the US east coast. Earthquakes have struck the interiors of India and Australia in the recent past as well.

These 鈥渋ntraplate鈥� earthquakes have long been a mystery. 鈥淭hey are the last frontier for plate tectonics,鈥� says Magnani. What we are finding out now, though, is giving us pause for thought. It might be that it鈥檚 not just San Francisco and Los Angeles that are susceptible to significant earthquakes, but New York, Sydney and perhaps even London too. Should we be worried?

Earth鈥檚 tectonic plates are the jigsaw-like pieces of its rocky outermost layers, and drift about on more viscous material below. Where plates meet, they move against one another and push each other up and down. Along the San Andreas fault in California, the North American and Pacific plates grind against each other at a rate of 33 to 37 millimetres a year, building up the stress released in earthquakes. Records indicate that California experiences a magnitude 7 or greater quake every 100 to 150 years; the last was the magnitude 7.8 San Francisco earthquake in 1906.

See graphic: 鈥淕reat shakes鈥�

Things might not be much different for intraplate earthquakes. Earth鈥檚 crust is engaged in a slow but constant process of ripping itself apart and crashing back together. At places such as the mid-Atlantic ridge, the nearest plate boundary to the east of New Madrid, this ripping has succeeded, creating a region of volcanism where new material is constantly spewing up from Earth鈥檚 interior. In other places, however, the rip never quite happens. The result is an unstable region that, though often unremarkable at the surface, is more easily stressed than the rock around it. These weak spots in Earth鈥檚 crust are strained by the same geological restlessness that strains faults at plate boundaries; it just takes longer. That, it had been assumed, could explain why intraplate earthquakes occur far less frequently than those at plate boundaries.

In the 1980s, it became clear New Madrid sits atop such a failed rift (). Dubbed the Reelfoot rift, it lies buried beneath the southern and Midwestern US and seems to have shuddered regularly in recent millennia. Magnani鈥檚 colleague digs around New Madrid in search of geological features called sand blows, produced when a powerful earthquake shakes the soil so much that it loses strength and behaves like a liquid, spewing from the ground in a tiny mud volcano. The plains around New Madrid are dotted with sand blows that formed 200 years ago. Underground, Tuttle has found more, suggesting that large tremors racked the area in 300, 900 and 1450 AD.

The United States Geological Survey (USGS) suggests that there is a of a magnitude 6 or larger quake hitting the New Madrid area in the next 50 years, with a 7 to 10 per cent chance of an event as big as the one two centuries ago. Back then, there were hardly any settlers in the region. Today, a quake of that size would displace 7.2 million people in Arkansas, Missouri and Tennessee, and cost at least $300 billion, according to a 2009 funded by the US Federal Emergency Management Agency, FEMA.

New Madrid might not be the only area at risk. Magnani鈥檚 studies of the deformation of Mississippi sediments have uncovered a 45-kilometre-long fault north of Memphis that seems to be part of the Reelfoot system. The 10-kilometre-long Marianna fault in Arkansas, discovered in 2009, could see a magnitude 7 quake, says of the University of Arkansas at Little Rock. 鈥淭he seismogenic potential involves a much larger area than just the active faults we see today,鈥� Magnani says. 鈥淣ew Madrid is just the latest incarnation.鈥�

Clustered and migrating

of Northwestern University in Evanston, Illinois, and his colleagues have come to a further startling conclusion after 20 years of using GPS to map the seismic zone around New Madrid. If the faults in the area are still under strain, they should be moving, just as they are at the San Andreas fault, for instance. But they are not (). In 2009, Stein and his colleague Eric Calais suggested that New Madrid is now in a deep seismic slumber from which it should not be expected to awake for hundreds, if not thousands, of years ().

That leads Stein to make a controversial claim. He doesn鈥檛 buy the idea that intraplate earthquakes are akin to interplate earthquakes, hitting home less frequently but in similarly predictable places. Instead, he characterises them as episodic, clustered and migrating. Seismic energy can jump within a network of small faults that snake their way through the middle of a tectonic plate, he says 鈥� and that is just what is going on beneath the US Midwest. 鈥淚f I had to guess, I would say that over time the motion in New Madrid will be transferred into seismic zones in Indiana and further south into Arkansas,鈥� he says. Whether that will happen on a timescale of decades or centuries, he cannot say.

Work by Stein鈥檚 collaborator of the University of Missouri in Columbia suggests there could be truth in this picture. Last year Liu analysed the occurrence of intraplate earthquakes over 2000 years in the north of China, scene of some of the most devastating historical examples, including the 1556 Shaanxi quake. Liu showed that the epicentres of intraplate earthquakes in China hop around haphazardly. Areas of violent shocks become quiescent; previously docile areas suddenly become active (). 鈥淭he earthquakes appear to be spatially migrating, jumping from one fault to another across long distances,鈥� he says. He thinks that faults in the middle of a plate are mechanically coupled, so that an earthquake along one changes another鈥檚 susceptibility to future movement.

鈥淭he earthquakes appear to be jumping from one fault to another across long distances鈥�

If so, that could have huge ramifications for our understanding of intraplate quakes. Take the Virginia quake of 2011. Its epicentre was in the Central Virginia seismic zone, which has experienced many quakes of around magnitude 3 over the past 120 years, but was not considered particularly at risk of anything bigger. If Stein and Liu鈥檚 ideas are right, the culprit might be seismic energy that roamed into the area from elsewhere. The nearby Western Quebec seismic zone, for example, extends over the northern border of New York State, and was visited by a magnitude 5.6 earthquake in 1944. The Eastern Tennessee seismic zone, stretching from north-east Alabama to south-west Virginia, is also highly active, although most quakes in the region are small. Two magnitude 4.6 earthquakes have occurred there in recent decades: one near Knoxville, Tennessee, in 1973 and another near Fort Payne, Alabama, in 2003.

That amounts to a wake-up call, says Stein鈥檚 colleague . 鈥淓arthquakes like the ones in Virginia and New Madrid could happen anywhere, including in Boston or Chicago,鈥� she says.

鈥淨uakes like the ones in Virginia and New Madrid could also happen in Boston or Chicago鈥�

In 2008, of Columbia University in New York City catalogued all 383 quakes in a 39,000-square-kilometre area around New York City from 1677 to 2007 and estimated the future risk. He concluded that New York can expect a magnitude 5 quake once every century, a magnitude 6 quake every 670 years and a magnitude 7 quake every 3400 years (). That highlights a gulf between perceived and actual risk, says of the British Geological Survey in Edinburgh, UK. 鈥淎n earthquake of magnitude 5.5 to 6 in New York would not come as a surprise to seismologists who have ever studied the area,鈥� he says. 鈥淏ut it would come as a surprise to most people who live there.鈥�

The same goes for other major cities. An earthquake of estimated magnitude 5.7 hit the Dover straits off south-east England in 1580, causing a pinnacle to fall off Westminster Abbey in London some 150 kilometres away. A magnitude 4.3 quake struck the same region in 2007. We should not overstate the risks, Musson says: most modern buildings in these areas could easily withstand a magnitude 5 or 6 quake. Skyscrapers in particular have enough 鈥渟way鈥� in them to counteract the effects, but historical monuments and older buildings such as police stations, schools and fire stations made from unreinforced brick could be vulnerable.

Any larger earthquakes could be more problematic. A magnitude 6.5 quake below Manhattan could cause $1 trillion in damage, according to , a former USGS seismologist who now works for risk-modelling company Risk Management Solutions. She suggests that not just building codes but also critical infrastructure 鈥� such as electrical and telecommunications networks and water and fuel pipelines 鈥� need to be upgraded to reflect the small but real danger.

In the US at least, more information on the vulnerable areas might come soon. , a mobile system of hundreds of seismometers that has been crawling eastwards from California since 2004, is now centred on New Madrid, where it will stay for two years before moving on towards the east coast. Each seismometer records sound waves generated by vertical and horizontal movements in the Earth鈥檚 crust, building up a complete picture of the rocks and the faults that riddle them (New 杏吧原创, 11 April 2009, p 26).

鈥淭he array will help us answer questions about intraplate earthquakes,鈥� says van der Lee. Almost every third US state is thought to have a piece of failed rift in it, she says. Why some, like the Reelfoot, are seismically active but others are not remains a big unanswered question. 鈥淯ntil we find a clear pattern that explains intraplate quakes, we have to expect they could happen anywhere.鈥�