³ÉÈËÂÛ̳

Why are megathrust earthquakes so powerful?
There are several kinds of geological fault, which differ in how plates move relative to each other.

At a strike-slip fault like the San Andreas in California, the plates are moving sideways past each other. The area where the plates meet is roughly vertical. This is important because below a certain depth, the crust gets so hot that it can flow and therefore doesn't build up the strain needed for earthquakes. This depth effect limits the size of the fault and hence its power.

In a megathrust earthquake area, one plate pushes under another in a so-called subduction zone. The fault interface is almost horizontal. Typically the fault descends at 10–20° from the horizontal. This means that it stays in the cool, brittle, shallow zone for longer, resulting in a much larger stress zone.

In addition, these faults are often very long – 1,000km in the case of the Cascadia Subduction Zone. Together, these factors make a megathrust earthquake the most powerful in existence.

How powerful will a Cascadia megathrust earthquake be?
Scientists now use the Moment Magnitude scale to measure the amount of energy released by an earthquake. This has replaced the famous Richter Scale which doesn't work very well for large earthquakes.

• Work out the area of the fault in square kilometres
• Take its logarithm (or, to a crude approximation, count the number of zeroes)
• Add 4

For Cascadia, the locked part of the fault is 1,000km long and about 100km wide. This gives an area of 100,000km² and an estimated magnitude of 9.

This is backed up by other megathrust earthquakes. The Sumatra earthquake of 26 December 2004 was a magnitude 9 and was caused by a fault very similar in size to Cascadia.

Kenji Satake has also used a tsunami simulation to work out how big the AD1700 earthquake was. He calculated that only an earthquake of around magnitude 9 could have created a tsunami to match the descriptions in the Japanese records.

So all the evidence points to the Cascadia Subduction Zone being capable of producing at least a magnitude 9 earthquake. It is possible that next time the fault might rupture in sections, leading to a series of smaller earthquakes. In many ways, this scenario is just as worrying. It would take 30 magnitude 8 earthquakes to release the same amount of energy as a magnitude 9.

How likely is it that the next Cascadia earthquake will occur this century?
The historical record shows that the average gap between large Cascadia earthquakes is about 550 years. Since there have been 300 years since the last event this might seem reassuring for now. This figure masks a huge variation - from 200 to over 1,000 years. The vital question is therefore whether the next gap will be large or small.

Dr Chris Goldfinger at the University of Oregon has been using deep sea sediments to look even further back into the past. He's looking at turbidites – traces of past underwater landslides. These landslides involve the collapse of huge amounts of sediment, dislodged by the earthquake. This has enabled him to get dates for 18 past earthquakes going back 10,000 years. The intriguing thing about these data is that they hint at a pattern.

Whenever there has been a long interval (more than 500 years) this has been followed by a short interval. At the moment scientists aren't sure if this pattern is real but if it is, it has worrying implications. The last gap, prior to the 1700 earthquake, was a long one. This might imply that the next gap will be a short one, of 2-4 centuries. Since we have already had 3 centuries, it's very possible that the next earthquake may occur within the next 100 years.

Further reading:
Atwater Brian F, Musumi-Rokkaku Satako, Satake Kenji, Tsuji Yoshinobu, Ueda Kazue, Yamaguchi David K, "The orphan tsunami of 1700, Japanese clues to a parent earthquake in North America", University of Washington Press
On sale 1 October 2005

Back to top

Back to the Horizon homepage