tokyo

Tremor terrors: longer forecasts would aim to give more warning of quakes than Kobe had in 1995. Credit: AP/KATSUMI KASAHARA

Earthquake research in Japan should aim at making 10-year forecasts of locations likely to be struck by major tremors, a group of 160 seismologists has recommended.

The proposal, part of a new research plan drawn up by the Japanese seismologists, represents a break from the previous policy, which has focused on short-term prediction of the location and size of earthquakes likely to occur within a few days.

The new plan is expected to serve as the backbone for the latest earthquake-prediction strategy being drafted by the Geodetic Council, an advisory body to the Ministry of Education, Science, Sports and Culture (Monbusho).

According to Yozo Hamano, leader of the earthquake researchers' group and a professor of Earth dynamics at Tokyo University, the ad hoc group was formed last year in response to a government report on earthquake prediction.

Subcommittees of the Geodetic Council had compiled a report last June which admitted that the short-term prediction of earthquakes — a programme that receives an annual budget of more than ¥20 billion (US$149 million) — is extremely difficult, and that there is a huge gap between the public's perception of the ability to predict earthquakes and the actual performance of science (see Nature 388, 4; 1997).

The new plan, with its long-term strategy, would mark the first change in the 30-year-old programme, which has been focusing mainly on predicting imminent earthquakes in the Tokai region, southwest of Tokyo.

According to Hamano, the new approach emphasizes elucidating a quake plan for the whole Japanese archipelago, through specifying the regions that have crustal distortions and are likely to experience an earthquake. “Thirty years of research in earthquake prediction has helped collect data for building new research methods,” says Hamano. “Recent improvement in observational methods, for example global positioning systems and other space-based techniques, are also encouraging for the new earthquake-prediction programme.”

But critics of the prediction programme are sceptical about the scientific basis of the plan, as well as its claim that it would bring changes to the existing programme. The main focus of the programme will see a shift towards long-term forecasts. But research in short-term prediction is still widely supported, and is likely to persist on a smaller scale in the eighth five-year prediction plan, which begins next year.

Robert Geller, a seismologist at Tokyo University, points out that the proposed plan lacks adequate theory or method required for making the reliable forecasts that it claims are possible.

“Accurate earthquake prediction is impossible, and will be for the foreseeable future,” says Geller. “What's needed is basic research on the earthquake-source process, rather than a national project aimed at issuing predictions.”