Menu Close

Revealing cracks in seismology

Is earthquake prediction even possible? Soe Than WIN/AFP

Why have so many lives been lost in Japan and New Zealand recently? And why have so many survivors – the so-called “lucky ones” – had their livelihoods and homes destroyed?

As a seismologist, I ask myself what has gone wrong, and keep coming back to three basic problems:

1) Seismological research has not been able to produce the kind of earthquake warnings people want (i.e. predicting when, where and how large the next quake will be).

2) The warnings seismologists have been able to produce are unpalatable to some sectors of society, and perhaps difficult to understand; and, as a consequence, protective measures have not been implemented.

3) Seismologists tend to underestimate how severe shaking, or indeed a tsunami, could be.

What seismologists can and can’t do

Though some people claim otherwise, seismologists cannot predict the exact time and location of an earthquake; what they can do is predict which geographical areas are most likely to experience large earthquakes.

The largest earthquakes occur on so-called “subduction zones” (where one large tectonic plate slips under another and gets stuck until the pressure is too great to bear).

This type of earthquakes is often followed by a large tsunami.

The so-called Pacific Ring of Fire (which borders Japan and New Zealand, as well as the west coast of North and South America, eastern Asia and many Pacific Islands such as Tonga and Fiji) is prone to this type of earthquake. Countries on the eastern edge of the Indian Ocean, (including India, Indonesia and Thailand) are also particularly susceptible.


So-called “clusters” of large subduction zone earthquakes occurred between 1952 and 1965 along the pacific rim in Russia, Chile and Alaska, with more than six earthquakes of a magnitude greater than 8.5 on the Richter scale.

Another such cluster seems to have started with the 2004 (magnitude 9.4) Boxing Day earthquake in Sumatra, which was followed by an aftershock and two more earthquakes with magnitude greater than 8.5 (Chile in 2010 and this month’s Japan earthquake).

But whether we are at the end or middle of this cluster is anybody’s guess.

Just today, we’ve heard news of a fresh earthquake in Japan, near the east coast of Honshu and reportedly of magnitude 6.5, although its strength has since been downgraded.

Soil and water

Seismologists and geotechnical engineers can advise which types of soil and buildings are most likely to fail in an earthquake.

Soft soils, especially when water-saturated, can amplify seismic waves and cause large shaking in buildings. When this soil is shaken strongly, liquefaction occurs, causing it to act like a liquid, and causing the structures built on it to fall over.

Maps of the highest areas for liquefaction potential were available well before last month’s Christchurch earthquake (e.g. the November 2004 Draft emergency plan)

Areas with steep slopes are prone to landslides during earthquakes, making buildings at the top susceptible to being carried down with the land, and those at the bottom susceptible to being squashed.

Maps with information about all these types of hazards are freely available on the internet to anybody (e.g. the hazard map for Wellington).

Mapping faults

Geologists map faults that reach the earth’s surface and dig trenches on those faults to determine how often in the past that section of the fault has broken, causing an earthquake. While this information is extremely valuable, some earthquakes occur on hidden faults.

The seismic hazard maps in New Zealand try to take this into account. As a result, in New Zealand it is considered that an earthquake up to magnitude 7.2 could occur anywhere, at any time, even if there is no mapped fault.

Earthquake relationships

Because earthquakes tend to obey simple relationships, seismologists can give probabilistic figures based on what has happened in the past.

The “Gutenberg-Richter Relationship” states that for every magnitude increase there are about ten times fewer earthquakes. So, on average, there is about one magnitude-8 earthquake in the world every year, ten of magnitude 7, 100 of magnitude 6, and one of magnitude 9 roughly every ten years.

Similarly, “Bath’s Law” tells us that about half of all earthquakes will precede a large aftershock only one magnitude unit or less below the main shock. So far in Japan, the biggest aftershock has been magnitude 6.1. The Christchurch aftershock, at 6.3, occurred later than expected but within the expected magnitude range.

Misjudged sizes

No-one expected a magnitude 6.3 earthquake could cause as much acceleration as the Christchurch earthquake did (more than twice the acceleration of gravity), and therefore buildings were not built to withstand such high accelerations.

Likewise, the Japanese nuclear power plant at Fukushima had a sea wall to halt any potential tsunami, but the wall was too small, so seawater flooded and destroyed the back-up generators meant to keep the fuel cool.

Finding solutions

Even though buildings collapsed in New Zealand and Japan, most of these were older ones built before modern building codes, and those collapses did not cause the largest property loss.

In New Zealand, liquefaction has taken the biggest economic toll. In Japan, water in the form of a tsunami caused the biggest loss of life and is one of the main causes of the continuing nuclear crisis.

Land-use planning is therefore high on my list of things that should be done.

Buildings should not be built in areas that are prone to liquefaction. Similarly, nuclear power plants should not be built in areas where there is any possibility that they would be flooded by tsunamis or rivers. They also need to be able to withstand large earthquakes (and also bomb and airplane attacks).

Ultimately, these land-use planning questions are a political issue; but seismologists and engineers can, and should, be included in discussions about how and where the recently devastated zones are rebuilt.

Want to write?

Write an article and join a growing community of more than 187,000 academics and researchers from 4,998 institutions.

Register now