It took three decades to develop Japan's earthquake prediction system, and a mere 20 seconds to tear it apart. When a powerful temblor ripped without warning through the Kobe region in 1995, it left in its wake nearly 6,400 dead, 35,000 injured and 400,000 homeless. At the same time, the 7.2-magnitude shudder dealt a blow to the widely accepted view in Japan that, with the proper technology in place, earthquakes ultimately are predictable. "It would be great if we could accurately predict a big quake," says Tadao Minami, a professor at the University of Tokyo's Earthquake Research Institute. "But Kobe showed us that isn't possible."
While seismologists elsewhere have long been more skeptical about the ability to foretell quakes, Japan is only now starting to concur--threatening the future of prediction research in the world's last bastion. "Optimism about forecasting died out quickly in the U.S. and elsewhere," says American geophysicist Robert Geller, who teaches at the University of Tokyo. "It's finally changing in Japan." Japanese policy makers have started to shift the emphasis and, more importantly, the funding from forecasting to preparation: reinforcing buildings, bridges, expressways and utility lines.
The Education Ministry, the main proponent of prediction research, has seen its earthquake budget decline nearly 60% since pre-Kobe days, while that of the Science and Technology Agency, which supports disaster engineering-related projects, has more than doubled. Scientists are falling in line, or rather following the flow of funds. The Education Ministry announced recently that it's time to focus on research aimed at preventing and reducing damage in a disaster. Until then, the emphasis had been on forecasting, using sophisticated meters and gauges to detect potential precursors, like tremors, fault strains and any uplifting of the earth's crust.
Those who still have faith have gone from mainstream to minority. "Even the term 'prediction' is taboo these days," laments Megumi Mizoue, head of the Meteorological Agency's Earthquake Disaster Prevention Council. "Just because the Kobe quake couldn't be predicted, doesn't mean we won't be able to forecast a 'big one' in the future." Mizoue and others say the government should actually be spending more. "We need to promote all types of earthquake research, including prediction," says Tsuneo Katayama, director of the National Research Center for Disaster Prevention. "Considering how quake-prone Japan is, the government hasn't spent very much."
Since 1965, when Tokyo designated prediction a national project, Japan has invested about $1.4 billion in prediction while downplaying other areas of earthquake research. It was then that the theory of plate tectonics--the shifting of continent-size slabs of the earth's crust--came into acceptance, raising hope that it was possible to detect an upheaval by catching the warning signs. For a country like Japan, crisscrossed by faults and situated over four converging plates, prediction research seemed a good investment. In fact, it was a cheap alternative: quake-proofing bridges, buildings and expressways would run in the trillions of dollars.
The quake warning program may have been economical, but it was hardly a success. After 32 years, it has failed to forecast the time or place of a single major quake. Despite that, Japan still allocates about $100 million of its $180 million earthquake budget for prediction research. Such spending is excessive, argues Geller, the American geophysicist. Like many others, he believes earthquakes are chaotic, nonlinear phenomena. "Scientists said they could predict earthquakes, and the funny pictures government dumped lots of money at their door," he says. "In the long term this has damaged the credibility of the scientific community."
Tokyo also has taken heat for centering its prediction program primarily around Tokai, about 150 km west of Tokyo, which was the site of a large eruption in 1854 that devastated the region. Critics say Tokai research has not helped places like Kobe that have been hammered by quakes. Yet seismologists involved in the project believe that a huge earthquake is overdue in Tokai, a coastal region of 10 million. They have positioned dozens of sensors and gauges in the area capable of detecting the faintest tremors or deformations in the earth's crust--warning signs that might enable them to forecast a quake and warn residents. Says the Prevention Council's Mizoue: "Tokai is one earthquake we may be able to predict."
After Kobe, however, few Japanese are waiting around for a warning. More than 12 million turned out to learn survival skills during last week's anti-disaster drills, held annually to mark the anniversary of the biggest quake in Japan in modern times--the 1923 temblor that killed 143,000 in Tokyo and Yokohama. The Prime Minister and his staff even held a mock crisis meeting in hopes of averting the confusion that marred Tokyo's response last time. "The Kobe disaster made us realize that prediction is not prevention," says Minami of the University of Tokyo. And inaccurate prediction is worth nothing at all.