In an article originally published March 14th, New Scientist explains why earthquakes are so hard to predict, how seismologists have tried to foretell quakes in the past, and what promising approaches may lead to successful prediction in the future.
Can earthquakes be predicted?
No—at least not so that we could issue an advanced warning for a specific time at a specific location that would allow for an orderly evacuation. Most quakes do, however, occur in predictable locations along well known fault zones, as was the case with last [month]'s megathrust off the coast of Japan.
How close can we come to predicting earthquakes?
For places with a high rate of historic activity, the chance that a quake will hit in a future period of several decades can be quite high. "We have models that say in southern California over the next 30 years, the odds of having an earthquake of magnitude 7.5 or greater is 38 per cent," says Thomas Jordan, director of the Southern California Earthquake Center and a member of the Collaboratory for the Study of Earthquake Predictability. If the same models are used to calculate the chance of an earthquake occurring in southern California within the next week, the odds drop to roughly 0.02 percent, Jordan says.
Why are big earthquakes so hard to predict?
Reliable predictions require precursors – some kind of signal in the earth that indicates a big quake is on the way. The signal has to happen only before large earthquakes and it has to occur before all big quakes. At the moment seismologists have failed to find those precursors—if they even exist.
What sorts of signals have seismologists considered for predicting large earthquakes?
A wide range of potential signals have been studied, ranging from increases in radon gas concentrations, changes in electromagnetic activity, foreshocks signalling a larger quake to come, warping or deformation of the Earth's surface, geochemical changes in groundwater – and even unusual animal behaviour in the moments leading up to a major quake.
Have any of these approaches worked?
For each of the signals listed above, we have evidence that they may behave erratically leading up to a large earthquake. Unfortunately, these irregularities also occur when no large quake follows.
"When you really bring the whole weight of statistical rigour to it, nothing stands up," says Susan Hough, a geophysicist at California Institute of Technology in Pasadena.
What promising approaches may lead to successful prediction in the future?
A number of researchers are continuing to look at changes in electromagnetic signals preceding major quakes. The approach is buoyed by the work of Friedemann Freund at the NASA Ames Research Center in California. Freund showed that compressing a rock can lead to the formation of positive electrical charges in the earth that could account for unusual electromagnetic signals prior to an earthquake.
Photo: Ishinomaki, Japan in 2010 and again one day after last month's tsunami, via NY Times