The massive recent earthquakes off Indonesia might not have killed so many if they could have been predicted. But earthquake forecasting is still an inexact science, although researchers are trying to improve its precision.
Unlike bad weather, earthquakes still take people by surprise. When hurricanes advance, meteorologists can warn people to evacuate areas they expect to be hit, but seismologists cannot provide similar alerts.
"We just don't have the knowledge that would enable us to issue a short term warning that people could act on," says U.S. Geological Survey expert James Dewey.
He explains that the best that geologists can do is provide the statistical probability of a major ground tremor based on past history of such activity and on the occurrence of smaller quakes in a region, which might be precursors to a big one.
"We prepare maps showing where the probability is highest based on both the geographic frequency and the frequency with time," he adds.
Although such probability forecasts cannot save lives when a big shock strikes, they are not useless. At the University of California branch in Davis, seismologist John Rundle says long-term predictions are important for civil engineers and disaster planners in areas prone to earthquakes.
"We can stockpile supplies near these hotspot locations," he says. "We can improve the roads and infrastructure in these locations. We can do a lot of things to mitigate the damage that you could not do if I were to come to you and say there is going to be a large earthquake next week."
Although short-term earthquake forecasts are not possible, scientists are working to make them a reality. The director of the Southern California Earthquake Center in Los Angeles, Thomas Jordan, says they face a big challenge.
"The problem is that what's going on down there is very complex and it's very difficult to predict when a small earthquake is going to turn into a big one," explains Mr. Jordan.
The complexity relates partly to fault lines. These are where large floating slabs of Earth's rocky crust abruptly bump against each other or overlap, creating the quakes. Mr. Jordan says it would be easier to predict major movements if they occurred along only one fault line. But they take place in a complex network of faults, which scientists don't quite know, yet.
"We need to understand how those complex fault systems actually work, and we've made great progress there," he notes. "We, in particular, are beginning to understand how faults interact with each other, how one earthquake affects the forces acting on another fault and changes the clock [timing] for that particular fault."
The University of Southern California geophysicist says better instruments buried underground or at sea bottom are being used in this research and are helping scientists get better at predicting small quakes, particularly aftershocks of large ones.
Scientists are also employing such instruments in their long-standing search for signals that might precede an earthquake. Mr. Jordan says some interesting data are already coming in.
"We now have detected signals in places like the [U.S.] Pacific Northwest and in Japan that are slow, rhythmic signals that may be associated with the buildup of stress and release of stress at the bottom of these seismogenic zones that may in the future help us anticipate when large earthquakes will occur," he adds.
A fault running lengthwise through California is the site of a 10-year project to determine if quake warning signals occur. Beginning later this year, a team led by Stanford University geophysicist Mark Zoback will drill along the fault and bury detectors to measure geophysical movements. They will also retrieve rocks and fluids for analysis.
"Does the surrounding section of the fault slip slowly before the earthquake? Does the fault itself initially begin to slip slowly and then accelerate into an earthquake slip? Does the pressure of the water and gasses in the fault change before an earthquake event? Or is there some other signal that might be used as a way of predicting its future occurrence?" asks Mr. Zoback.
Satellites in space and strain meters on the ground are being used to measure slight terrain movements or deformations in fault zones to see if they are precursors to earthquakes.
Research like this will take a long time, offering people living in seismologically active regions no immediate promise of short-term earthquake prediction. But James Dewey of the U.S. Geological Survey says there is hope.
"It's possible that with the additional data of the sort that are now being collected that these will come together and point us to a methodology for accurately predicting the time and place of an earthquake," says Mr. Dewey.