Unlocking the secrets of earthquakes

Scientists are trying to get to grips with what causes big tremors.

Earthquakes are among the worst natural disasters, causing death and devastation. For centuries it was thought they strike randomly and without warning. But do they?

Scientists have been searching for ways that could help to predict where and when future tremors may occur. Some of the proposed methods range from weird to wacky. They include detecting spikes in radon gas emissions to the behaviour of toads (a study being carried out on them near to L’Aquila in Italy discovered they nearly all disappeared from the area ahead of the big earthquake that struck in 2009). There are also old wives’ tales of how snakes in the US predict when tremors are about to strike too.

A new theory hit the headlines last year, as it raised the prospect of a string of big tremors occurring in 2018. It suggests that the frequency of major tremors over a magnitude seven could be linked to the regular fractional slowing of the Earth’s rotation. Researchers found clusters of large earthquakes every 30 years or so, usually about five to ten years after the Earth slows. As the last time this happened was in 2013, so, the theory goes, this could lead to more earthquakes this year.

An earthquake can be triggered by an increase in stress equivalent to much less pressure than it takes to inflate a car tyre.

But Hiscox’s Stephen Gibson, an Oxford-trained geologist who for seven years was Hiscox’s seismic risk analyst, is skeptical of the link. “The periods when Earth’s rotation decelerated the most don’t fit neatly with the peak periods of large-magnitude earthquakes occurring. So the coupling seems to be relatively loose at best.”

One theory gains ground

But one theory of how to predict earthquakes has gained increasing credence over the past 25 years. It began by mapping the location of aftershocks (the tremors that occur after a big earthquake), which had been largely overlooked by seismologists, as a way of trying to unlock the secrets of earthquakes. 

Stress builds up as Earth’s tectonic plates grind past each other. As the plates on either side of the fault move in opposite directions and as the rocks on each side press against each other they exert what are known as Coulomb stress.

Geoscientists, including Ross Stein (whose earthquake predictions have garnered a lot of media attention), spotted there were a string of subsequent tremors in areas where Coulomb stress had risen after major earthquakes. This is because the stress that is relieved during an earthquake doesn’t just disappear but moves along the fault and even to neighbouring faults. Stein suggests that an earthquake can be triggered by an increase in stress equivalent to much less pressure than it takes to inflate a car tyre.

Stein and his colleagues began to calculate the increasing Coulomb stress in a region following a big tremor to see if it could help to predict where future earthquakes could occur.  

But just as a major earthquake can trigger tremors elsewhere, it can reduce the likelihood of another one occurring in the same place.

In 1997, Stein and two other scientists predicted there would be a 12% chance that a major earthquake, of magnitude 7 or higher, would occur around the Turkish city of Izmit over the next 30 years. In August 1999, a magnitude 7.4 quake destroyed Izmit, killing 25,000 people and costing over $6.5 billion. Three months later, a magnitude 7.1 tremor hit Düzce, 100 km east of Izmit in a fault section the scientists had predicted would experience an increase in stress resulting from the earlier earthquake.

Since the devastating 2011 Tohoku earthquake, several seismologists, including Stein and Shinji Toda among others, have warned of an increased threat for Tokyo and the easterly Chiba region. Although increased low-level seismicity has been observed no significant event has occurred – so far.

The stress triggering theory makes “some pretty big, assumptions,” says Gibson. “They effectively put a grid over the region around an earthquake location and work out whether the squares in that grid have experienced an increase in Coulomb stress to assess whether they are more likely to suffer an earthquake.”

But, he adds, “there’s a definite trend towards understanding areas that are stressed and therefore due a rupture.” Research following the magnitude 8.8 earthquake that hit Maule in Chile in February 2010 and the magnitude 8.2 tremor that affected Iquique in April 2014 showed a correlation between where the ruptures occurred and areas that had seen a build-up in Coulomb stress. 

Dynamic earthquake triggering suggests the aftershock zone could extend across the entire surface of the Earth.

But just as a major earthquake can trigger tremors elsewhere, it can reduce the likelihood of another big one occurring in the same place. One theory for why the San Andreas fault near San Francisco has not suffered other big ‘quakes since 1906 is because the region was in a ‘shadow’ where stress had been relieved by that big tremor. That stress is, however, predicted to slowly build up again in that area of the San Andreas Fault.

Butterfly theory of quakes?

Another theory suggests that a major earthquake could increase the likelihood of other tremors much further away than the same or neighbouring fault. Dynamic earthquake triggering suggests the aftershock zone could extend across the entire surface of the Earth. It suggests that when an earthquake happens, fluid in Earth’s crust is displaced and may move into faults, making them more pressurised – and more susceptible to rupturing. 

It’s an interesting, but embryonic theory says Gibson. Its advocates admit it can be difficult to sift coincidental seismic activity that occur across Earth from dynamically-triggered tremors, but a number of geophysicists are investigating it.

Still a lot to learn

Although scientists’ understanding of earthquakes has come on in leaps and bounds they can still be taken by surprise. At 2.46pm on March 2011 one of the most powerful earthquakes ever recorded struck off the coast of Tohoku in northeastern Japan. The magnitude 9.0 tremor created a tsunami that destroyed 300,000 homes and left 22,000 people dead or missing. But it came as a complete shock to most seismologists, who had not expected such a powerful tremor to occur in that location or to trigger such a large tsunami.

 “Seismology is a relatively young science,” says Gibson. “A reliable catalogue of events only really stretches back to the 1920s, so that’s less than 100 years of data to play with. Some earthquake zones, such as Canada and parts of Japan, have only experienced one tremor during that period, so it’s very difficult to pick apart the numbers,” explains Gibson.

Earth is 4 billion years old, so seismic activity in the past 50, even 100 years, is like the blink of an eye.

It can also be difficult to get the data necessary to help predict a region’s earthquake risk. “The US Geological Survey is the authoritative voice on seismic risk in the US, but Japan has several agencies, while there isn’t really any single source of data in Chile,” says Gibson.

Scientists estimate when another tremor may occur in a seismically active region, based on their calculation of how frequently earthquakes strike there. So for example, some parts of California, such as the Hayward Fault east of San Francisco, are deemed to be overdue for another earthquake, while the risk of one in the Nankai Trough, which runs off the southern Japanese coast, is forecast to increase over the next 30 or 40 years having last ruptured in the 1940s.

But scientists are some way from being able to predict when an earthquake might occur in a particular region any more accurately than years or even decades.

So what is the Holy Grail for earthquake forecasting? 

The most realistic aim, says Gibson, would be to develop a method that enables the authorities in those countries and regions most at risk of tremors to draw up better-informed crisis plans and could help insurers and reinsurers to get a better understanding of where are their biggest exposures.

But, he warns, there will continue to be elements scientists would never be able to predict, until they occur – the “unknown unknowns” of seismic risk, if you like. Scientists are expanding their knowledge of earthquakes very quickly, but they still have a long way to go. “Earth is 4 billion years old, so seismic activity in the past 50, even 100 years, is like the blink of an eye. More people are thinking about earthquake risk and are thinking about it in more ways. That can only be good,” concludes Gibson.

All comments