In Italy, Asia and New Zealand, long-range earthquake predictions from self-taught forecasters have recently had people on edge. But is it possible to pinpoint when a quake will strike?
It's a quake prediction based on the movements of the moon, the sun and the planets, and made by a self-taught scientist who died in 1979.
But on 11 May 2011, many people planned to stay away from Rome, fearing a quake forecast by the late Raffaele Bendandi - even though his writings contained no geographical location, nor a day or month.
In New Zealand too, the quake predictions of a former magician who specialises in fishing weather forecasts have caused unease.
After a 6.3 quake scored a direct hit on Christchurch in February, Ken Ring forecast another on 20 March, caused by a "moon-shot straight through the centre of the earth". Rattled residents fled the city.
Predicting quakes is highly controversial, says Brian Baptie, head of seismology at the British Geological Survey. Many scientists believe it is impossible because of the quasi-random nature of earthquakes.
"Despite huge efforts and great advances in our understanding of earthquakes, there are no good examples of an earthquake being successfully predicted in terms of where, when and how big," he says.
Many of the methods previously applied to earthquake prediction have been discredited, he says, adding that predictions such as that in Rome "have little basis and merely cause public alarm".
Seismologists do monitor rock movements around fault lines to gauge where pressure is building up, and this can provide a last-minute warning in the literal sense, says BBC science correspondent Jonathan Amos.
But any longer-range prediction is much harder.
"It's like pouring sand on to a pile, and trying to predict which grain of sand on which side of the pile will cause it to collapse. It is a classic non-linear system, and people have been trying to model it for centuries," says Amos.
In Japan, all eyes are on the faults that lace its shaky islands.
On Monday, Trade and Industry Minister Banri Kaieda urged that the Hamaoka nuclear plant near a fault line south-west of Tokyo be shut down, pending the construction of new tsunami defences.
Seismologists have long warned that a major earthquake is overdue in this region.
But overdue earthquakes can be decades, if not centuries, in coming. And this makes it hard to prepare, beyond precautions such as construction standards and urging the populace to lay in emergency supplies that may never be needed.
Later this year, a satellite is due to launch to test the as-yet unproven theory that there is a link between electrical disturbances on the edge of our atmosphere and impending quakes on the ground below.
Then there are the hypotheses that animals may be able to sense impending earthquakes.
Last year, the Journal of Zoology published a study into a population of toads that left their breeding colony three days before a 6.3 quake struck L'Aquila, Italy, in 2009. This was highly unusual behaviour.
But it is hard to objectively and quantifiably study how animals respond to seismic activity, in part because earthquakes are rare and strike without warning.
"At the moment, we know the parts of the world where earthquakes happen and how often they happen on average in these areas," says Dr Baptie.
This allows seismologists to make statistical estimates of probable ground movements that can be use to plan for earthquakes and mitigate their effects. "However, this is still a long way from earthquake prediction," he says.
And what of the "prophets" who claim to predict these natural disasters?
"Many regions, such as Indonesia and Japan, experience large earthquakes on a regular basis, so vague predictions of earthquakes in these places requires no great skill."