The global geoscience community was surprised, to say the least, when seven scientists and engineers were each sentenced to six years in jail for their advice several days before a treacherous earthquake struck in L'Aquila in Italy in 2009.
The judge said the seven – who were accused of causing some residents to stay indoors on the night of the quake instead of finding shelter elsewhere – had analysed the risk in a "superficial, approximate and generic" fashion and had participated in a "media operation" to reassure people.
The case has spurred debate as to whether such acts of nature can be accurately forecast.
"Scientists recognise that reliable earthquake prediction is a goal that has not yet been achieved – and may never be achieved to the degree of certainty that the public would like," says Sydney University geoscientist Tom Hubble.
"Making grand statements, even with a covering disclaimer, about the safety or otherwise of an area experiencing increased seismic activity is very brave," he says.
As Charles Richter, of Richter-scale fame, once reportedly remarked: "Bah, no one but fools and charlatans try to predict earthquakes!"
Seismology involves learning about past earthquakes, and most people realise that such events cannot be predicted reliably, says Melbourne University seismologist Gary Gibson. "The earth is complex, and stress levels at earthquake depths are impossible to measure, as are weaknesses in faults that can trigger an initial rupture which develops into an earthquake."
Anticipating earthquakes entails more than just a prediction of where, when and how big a near-future event will be, he adds.
"It includes forecasts that suggest the probability of an event, for example a 10 per cent chance, as well as specifying the area where it will happen, the period during which it will happen and its size or magnitude," Mr Gibson says.
About 90 per cent of such forecasts with 10 per cent probability will not happen.
"Even less precisely, an alert indicates that something unusual has occurred, and there is an unquantified increased possibility that a bigger event may follow," he says.
Seismologists often study earthquakes to develop hazard estimates, which give the average time between quakes of a certain magnitude in a particular area, but with no indication of when the next may occur.
Shallow clusters of earthquakes are common, with one in Australia every year or two. Fortunately, most of them end without a moderate or large quake.
Depending on the number of earthquakes in the cluster and their magnitude, roughly one in 100 clusters leads to a moderate or large quake in the following week or two, Mr Gibson says.
"Before the main L'Aquila earthquake, it would be reasonable to suggest that there would have been a 1 per cent chance, or possibly a little higher, of a larger event in the following two weeks," he says. "This is not a prediction that the event will happen, nor is it a well quantified forecast that it will happen with a justifiable probability," he stresses. But it could be presented as an alert that cluster activity is unusual and may lead to a moderate or large quake.
"The past rate of activity in the region suggests the average probability of a moderate or larger earthquake over a two-week period may be roughly one in 10,000 – or once in 400 years, on average," Mr Gibson says. "Hence the probability of the moderate-to-large earthquake happening was something like 100 times higher than normal."
In some countries, such as the Philippines, public education relating to natural hazards has advanced to the extent that uncertainties can be discussed in the media, Mr Gibson says.
"In other countries, including Australia, the US and Italy, people tend to expect to be told what will happen and what to do about it," he says. This applies not just to earthquakes but also to floods and bushfires. "A forecast or alert regarding an increased probability of an event, that then does not occur, is generally treated as a failed prediction."
Meantime, failed predictions "tend to diminish the reliability of the scientists' advice in the public mind", Associate Professor Hubble adds: "sometimes to the extent that people may entirely stop trusting advice."
South-eastern Iran was severely shaken last month by an earthquake of 7.8 on the Richter scale. Tremors were reported as far off as New Delhi in India. The huge quake came days after a smaller one, measuring 6.3, which hit Bushehr Province in Iran's south-west.
The quakes occurred on the same section of tectonic plate, albeit more than 1000 kilometres apart. Might the events be related?
Perhaps, Mr Gibson says. But predicting the next quake from its predecessor is anything but easy, he points out.
Magnitude 8 and 9 earthquakes – the really big ones – do not, as a rule, directly trigger other large earthquakes at great distances, Associate Professor Hubble says.
Yet aftershocks, along with smaller localised events, are well documented and accepted. "The same applies to foreshocks which are small events preceding a major earthquake – something like the creaks and snaps we hear before a large branch breaks off a tree trunk during high winds," he says.
"Aftershocks occur as the stress regime in the surrounding plate and rocks readjusts after the first large quake," Associate Professor Hubble adds.
Many plate-boundary earthquakes occur because of stress build-up over time owing to plate movements, he says. "This might be towards each other in places, such as the Himalayan belt, or past each other along the famous San Andreas fault."
The earth's crustal plates deform gradually because of these motions, and when the stress within a plate builds to a level greater than its rocks can bear, the rocks break and grind past each other along fault planes.
This, Associate Professor Hubble explains, generates vibrations which move through the rocks as earthquake waves. We notice these as ground-shake.
Russians and Chinese
Seismic soothsaying is the preserve of Chinese and Russian seismologists, whose forecasts and alerts seem to be improving.
The Chinese have forecast almost half of all earthquakes of seven or more on the Richter scale since 1978. Using the level of water in wells to make their predictions, they realised that levels rise or fall according to whether water is being absorbed by or streaming out of aquifers.
This, in turn, depends on motions within the earth. In keeping with eastern notions of cyclicity and the interplay between opposites, the Chinese have found periodic changes in seismic activity in western China. One area might be active for about 150 years while another is dormant; the dormant area then becomes active for the next 150 years, and so on.
"The approach of the Chinese could work for continental regions," says University of Tasmania earth scientist Jo Whittaker. "But it isn't going to be useful for regions where earthquakes occur offshore and cause tsunamis – such as Indonesia and Japan. In cases like that, a more statistical or modelling approach seems to be the only viable option."
Russian seismologists, meanwhile, have tried to identify seismic patterns that might help predict the time, area and size of a quake. Their computer models, based on statistical analyses, occasionally raise false alarms, while many earthquakes go unpredicted. Yet the Russians' success rate is better than that of chance alone.
Western experts remain largely sceptical of these efforts, although a few, such as Mr Gibson, are open-minded. "There are, undoubtedly, tell-tale patterns but, as yet, we can't be sure how reliably they can be applied in different places. The Russians are probably on the right track, although, frankly, I don't understand the physical mechanisms involved. Perhaps we Westerners are too bogged down in the technicalities and should stand back to appreciate that the whole is more than the sum of its parts".
Plates in action
The geological forces that shape continents are as active today as they were 4.5 billion years ago when our planet was in the throes of being formed.
The earth comprises a series of layers. The outermost layer is called the crust and the innermost the core. In between lies an area known as the mantle, which is the most important in understanding where big earthquakes occur.
Beneath the relatively thin, rigid crust sits a mantle of solid crystalline rock. At a depth of more than 100 kilometres or so, the rock is so hot that it's malleable and squishy and can flow like plasticine.
Over eons, the covering crust has broken into massive brittle sections called plates. There are seven main ones and several smaller chips off the old blocks, all moving and shoving against each other as they slide across the hot mantle – not unlike boats being heaved across a sandy beach.
The plates are continuously colliding and being forced under one another, a process called subduction. Occasionally, they fracture, causing earthquakes and volcanoes. Their effects can be felt soon afterwards from hundreds of kilometres away because the elastic waves produced by an earthquake travel at between four to 10 kilometres a second.
The fastest seismic waves take less than 20 minutes to reach the other side of the earth, a distance of almost 13,000 kilometres.
No part of the earth's surface is free from earthquakes – although some regions experience them more frequently. They are most common at boundaries where different plates meet. The largest events usually happen where two plates are colliding, particularly around the edge of the Pacific Plate in New Zealand, Papua New Guinea, Japan and the Americas.
Intra-plate earthquakes occur in the relatively stable interior of continents away from plate boundaries. These are less common and generally originate at shallower depths but seldom follow recognisable patterns.
Although Australia is not on the edge of a plate, it experiences earthquakes because the Indo-Australian plate is being pushed north towards the Eurasian, Philippine and Pacific plates. This causes stresses to build in the plate's interior, which are released during earthquakes.
What to do
"The best advice for people living in earthquake-prone areas is to live in earthquake-resistant buildings," Associate Professor Hubble suggests.
Mr Gibson agrees. "Do not place buildings where surface ruptures may occur – namely along known active faults – or in landslide-prone regions, such as in steep valleys. The same goes for tsunami-prone regions along coasts, or liquefaction-prone regions with saturated near-surface sands and silts." Some of these restrictions, he concedes, may be hard to accept.
Earthquake-resistant buildings need not be expensive or of high-quality, Mr Gibson adds. "Lightweight village houses in many Pacific areas do not cause widespread casualties due to collapsing after an earthquake."
Learn about earthquake predictions in Russia at: http://english.ruvr.ru/2013-03-31/Earthquake-predicting-device-launched-in-Russia/
And in China at: http://rense.com/general61/use.htm
Find out more about the controversial Italian trial in the British journal Nature Geoscience at: www.nature.com/news/2009/090910/full/news.2009.899.html
Delve deeper into the nature of plate tectonics and continental drift in the Australian curriculum textbook Oxford Big Ideas, science 9 (Oxford University Press, 2012)
VCE Physics: www.vcaa.vic.edu.au/Documents/vce/physics/PhysicsSD-2013.pdf
F-10 Physics (under sub-strand 'physical sciences' in AusVELS): http://ausvels.vcaa.vic.edu.au/Science/Curriculum/F-10
Please send bright ideas for new topics to email@example.com