Sketchy science of quake prediction

Think scientists are close to being able to predict earthquakes? Well think again, US experts said this week.

The magnitude 9 quake off Sumatra that triggered the devastating Boxing Day tsunami was a potent reminder that while it is easy to say where big quakes will happen, pinning one down to a day, a week or even a decade has proved impossible, they said.

That quake apparently gave no warning whatsoever, researchers told the American Association for the Advancement of Science.

Quake experts do not even agree how to measure a tremor's magnitude to see if an experiment for predicting activity has worked, said Thomas Jordan of the Southern California Earthquake Centre.

"Most seismologists, including myself, are pessimistic that in the next five years to 10 years that we will ever be able to come up with a silver bullet earthquake solution," Jordan said. "But never say never."

There are some suggestions that very low-frequency seismic activity might predict some quakes.

US Geological Survey researchers have measured low-pitched rumbles from deep under California's San Andreas fault, about 22km southeast of Parkfield.

They resemble measurements made at subduction zones in Japan and the Pacific Northwest.

A subduction zone is where one of the Earth's tectonic plates is slipping under another and the Sumatra quake occurred at one such zone.

"We just don't understand what those signals mean," Jordan said. "As best we can tell, there are no reliable short-term precursors to earthquakes."

David Applegate of the US Geological Survey said Parkfield, which has been covered with instruments for 20 years in the hope of catching a tremor in the act, is providing useful information, but nothing as yet that can be used to begin predicting an earthquake in a useful way.

More important will be finding out the highest-risk areas so building codes can be brought up to date and emergency services told where the ground will shake the most, so they can respond quickly.

- REUTERS