7 April 2011

The anatomy of a megathrust earthquake

Cosmos Online
The recent giant Tohoku earthquake in Japan has rewritten the rulebook on where - and when - we think the megaquakes occur.
Tohoku plate

The subducting Pacific plate east of Japan is responsible for most of Japan's large earthquakes as it descends beneath the over-riding plate. Credit: Craig O'Neill

It was the biggest quake to rock Japan in recorded history. On 11 March 2011, just 130 km off the coast of Sendai, on the main island of Honshu, a section of the massive submarine fault separating the enormous Pacific plate from northern Japan failed.

In a few minutes, over 500 km of the fault catastrophically ruptured, and rocks on either side were thrown over 20 metres along the fault.

The island of Honshu was wrenched 2.5 m towards the east, and the Earth’s tilt axis moved by 16 cm in response to this enormous redistribution of mass. The seafloor upheaval shunted a column of water over 6 km deep upwards – a massive disturbance which culminated in a 10 m tsunami swamping the Japanese coastline.

At magnitude 9, the quake was propelled into another class of monster – a megaquake. The scale used to measure earthquake magnitude, whilst useful, does not do justice to the energy involved in the biggest of the big.

The moment magnitude scale is logarithmic – meaning the energy released by a magnitude 9 is around 32 times greater than a magnitude 8, and almost 1,000 times larger than a magnitude 7.

Bigger than the largest hydrogen bombs ever created, or the most massive volcanic explosions, the energy released by megaquake events have no equal on the planet. They are the most violent events we have had the misfortune to observe.

How are such energetic events generated? The most seismically violent regions of the world today occur at subduction zones, where dense oceanic plates are thrust under overriding crust to be recycled back into the Earth’s mantle.

The process creates some of the most spectacularly deep oceanic trenches in the world, including the Mariana Trench and the Challenger Deep – the deepest point of the world’s oceans.

But the plates are, in the end, billions of tonnes of hard rock scraping against each other, and the stately motion of the plates around the surface of the Earth belies the violence accommodating their motion in subduction zones.

The motion of plates at their boundaries is accommodated by a phenomenon known as ‘stick-slip’. Here friction along a fault prevents smooth sliding of the two plates against each other, and so they stick, allowing stresses to build up.

The motion of plates against each other continues, however, and eventually stresses reach a critical level. At some point the fault cannot handle the stresses being piled upon it and it fails catastrophically. The rocks on either side of fault slip, resulting in an earthquake.

In some cases the level of stress across a fault system can build up to truly incredible levels, and the earthquake, when it occurs releases hundreds of years of pent-up energy. Faults capable of generating such great quakes, of magnitude 8 or above, have been dubbed ‘megathrust systems’.

Fortunately, megaquakes are rare. The geological factors which conspire to allow a fault system to build up such an incredible amount of stress are not common. In fact, nothing of the magnitude of the March 11th Tohuku event has been recorded along that section of the Pacific plate since the records began.

But, unfortunately, our understanding of how the Earth generates such monstrous convulsions is sorely lacking. Geological models did not predict that section of the trench system capable of generating such a powerful event. So why did it occur off Sendai?

As a plate returns to the mantle, the angle which it does so is strongly controlled by its structure and its age. Young, just-formed ocean crust tends to be hot, and not very dense, and so sinks into the mantle at shallower angles.

This generally means it is in contact with the overriding plate for longer, increasing the width of the seismogenic zone. In contrast, old, thick plates are much denser and, on the whole, sink more steeply in the mantle.

As a result they are not in contact with the overriding plate for as long, and have thinner seismogenic zones. On the whole there is an association between subducting young crust, or other seafloor structures such as seamounts, and great megathrust earthquakes.

Speed helps, too. The faster a plate is moving, the faster it accumulates stress at the fault zone, and the greater the frequency at which it can generate large earthquakes.

So the general rule of thumb is that subduction of fast, young plates leads to megaquakes. The biggest recorded quake, the magnitude 9.5 event in Chile in 1960, seems to back this up.

But, as in many areas of science, generalisations can be misleading, and the Tohoku event seems to buck this trend.

Commenting in the 17 March 2011 edition of the journal Nature, Emile Okal, a geophysicist at Northwestern University in Chicago, notes “This earthquake is a lesson in humility.”

The portion of the Pacific plate subducting off the northeast of Japan is 140 million years old, around the oldest subducting crust known. The area has not even been known to have generated a magnitude 8 in historical times, and so was seen as – comparatively – low-risk.

But with historical seismic records only covering the last 100 years, perhaps the key to understanding the potential of the fault was locked away in the geological record.

In 2001 Koji Minoura from Tohoku University in Sendai noted in an article to the Journal of Natural Disaster Science that there was evidence in the local geological record for tsunamis in the Sendai region every 800-1100 years.

The last tsunami occurred in 869 AD, and Minoura prophetically noted that Sendai was overdue. Even more disconcerting is the growing suspicion that a recent spat of seismicity seen around the Indo-Pacific rim might be related.

Since Boxing Day 2004, there have been four megaquakes of magnitudes greater than 8.5 around the Indo-Pacific hemisphere. The devastating Sumatran quake and tsunami of 2004, which clocked in at a magnitude of 9.2, was quickly followed by another 8.6 event on 28 March 2005, off western Sumatra.

Again on 13 September 2007, a 8.4 magnitude quake hit the Mentawie Islands. In addition to Sumatra, the trend continued with the 8.8 Maule quake off Chile in February 2010, and the 2011 magnitude 9.0 Tohoku quake off Japan.

There were seven great quakes larger than magnitude 8.5 in the century preceding Boxing Day 2004, and yet there have been four in the last seven years. Is this just a blip in the numbers? Some random fluctuation in the rate of these events? Or is there a physical reason for this recent spat of megaquakes?

The Tohoku quake itself was preceded by smaller but still deadly 6.3 magnitude quake in Christchurch, New Zealand, which devastated the city on the 22 February 2011. That event followed on the coattails of a magnitude 7.1 quake nearby in Canterbury in September 2010.

Despite being many thousands of kilometres apart, these events all occurred on the boundary of the enormous Pacific plate. Is there some way in which these distant events could be related?

It turns out there is. Writing in the 2009 in the journal Nature, geophysicist Thorne Lay and colleagues pointed out a pair of earthquakes that rocked the Kuril Islands north of Japan starting in 2006.

Whilst the behaviour of aftershocks has been well studied, less well known are events called ‘doublets’, where two large earthquakes can occur months after each other, on very different fault zones.

The Kuril Islands experienced such an event beginning with an 8.3 event on November 2006. The motion caused by that earthquake stressed the plate and culminated in another 8.1 quake – a direct response to the first event.

But while mechanism is able to explain locally related quakes, whether it is able to scale up to explain the Pacific plate’s behaviour is another question.

The issue is that with only 11 quakes greater than magnitude 8.5 recorded since 1900, there simply isn’t enough data to draw any meaningful correlation on megathrust earthquakes.

Tom Parsons from the U.S. Geological Survey and colleagues presented work at the 2009 American Geophysical Union autumn meeting on whether there was any association of earthquakes greater than magnitude 7 generating seismicity above magnitude 5 around the world.

Their conclusions – that there is no evidence that large earthquakes generate other quakes more than 1,000 km away – would seem to argue against an association of these giant quakes.

But given the difficulties with large earthquake statistics, the jury is still out.

For now, earthquakes remain the only natural disaster for which we have no early warning system. They can hit at any time, and though we have some rough idea of where the risk is, these events continue to surprise us.

The first inkling geologists may have of a subsurface fault system is when it ruptures during an earthquake – as was the case in Christchurch recently.

The two major advances sorely needed in earthquake science is a better characterisation of the ground beneath our feet – from understanding and imaging hidden fault systems, to identifying rock types and understanding how they behave under stress.

Such efforts are underway, from the recent Scientific Drilling into the San Andreas Fault deep drilling project to the Japanese offshore drilling vessel the Chikyu, which has recently focussed its efforts in the highly seismic Nankai trough.

The second major advance is understanding the processes involved. Since the Earth, and earthquake physics, is highly complicated at all scales, the only way to do the problem justice is by massive computer simulations, and techniques to achieve this are just now coming online.

The Japanese scientific establishment was hit hard by the March 2011 quake, and much critical infrastructure has been damaged. Fortunately, the science will continue.

The Japanese scientific drilling vessel, the Chikyu, was docked north of Sendai during the Tohoku quake hit. Within half an hour, it had steamed out to sea with everyone touring still onboard – including 48 school children – to avoid the tsunami. It was just in time.

The incoming tsunami spun the ship two and a half times, and damaged one of the thrusters, but no one was hurt – a testament to the tsunami early warning system.

The co-chief scientist on board at the time, Fumio Inagaki, commented in the journal Nature, “It was almost a miracle no life was lost.” Over 28,000 people in Japan weren’t so lucky.

Craig O’Neill is a Future Fellow and senior lecturer in geophysics at Macquarie University in Sydney, currently working on a project looking at megathrust earthquakes around the Australian plate boundary.

Sign up to our free newsletter and have "This Week in Cosmos" delivered to your inbox every Monday.

>> More information
Like us on Facebook
Follow @CosmosMagazine
Add Cosmos to your Google+ circles

Get a weekly dose of Cosmos delivered straight to your inbox!

  • The latest in science each week
  • All the updates on our new website launch
  • Exclusive offers and competitions

Enter your name and email address below: