Weather - Scientists predict California earthquake for 2004

A group of scientists from the University of California in Los Angeles (UCLA) has provoked controversy by predicting that there will be a magnitude 6.4 or higher earthquake in the south of the state by September 2004.

According to the scientists, the earthquake will hit a region that includes the south-eastern part of the Mojave Desert and an area to the south of it. Furthermore, the group also claims to have predicted the San Simeon and Hokkaido tremors in the last half of 2003.

Although the team, led by veteran seismologist Vladimir Keilis-Borok, does not claim to be able to deliver an exact estimate of the date and time of a future earthquake, they are confident of the rough area as well as predicting a nine-month period in which it will occur.

The UCLA team includes experts in seismology, geodynamics, pattern recognition, chaos theory, public safety and statistical physics. It has been able to develop algorithms to detect precursory earthquake patterns in the form of the small chains of tremors that occur before a major event.

If such a chain is detected then the team analyses the history of the area, looking for similar events and patterns.

Successful earthquake prediction has proved almost impossible until this point, with scientists only being able to issue vague warnings over certain areas where there have been major earthquakes in the past.

Two of the most notorious of these areas include the San Andreas Fault in California and the area around Tokyo. Both of these areas are expecting major earthquakes in the future that could cause a massive amount of damage and a possible widespread loss of life.

Professor Keilis-Borok acknowledges that it will take time for other scientists to be convinced about the methods used by the team. However, he added: "Application of non-linear dynamics and chaos theory is often counter-intuitive, so acceptance by some research teams will take time. Other teams, however, accepted it easily."

Commenting on the system, Bill McGuire, director of the Benfield Hazard Research Centre, said: "Should the prediction fail, then the method can be discarded - any useful predictive tool must work all the time. If it fails even once then it is not workable. If the prediction should come to fruition then seismology may at least have taken a step along the way to true earthquake prediction."