New method for predicting earthquakes

The Swedish Defense Research Agency (FOI) and Uppsala University have shown that a method previously used to warn about mining quakes can be used to predict where and when earthquakes are going to take place.

“Using this method, major quakes like the one that caused the tsunami could be predicted better, both in terms of time and geographic area,” says Leif Persson, a researcher at FOI.

Seismology researchers at FOI and Uppsala University have retrospectively examined the occurrence of earthquakes in the area around Sumatra and the Andaman and Nicobar Islands during the five years preceding the catastrophic earthquake off the coast of Sumatra the day after Easter 2004. This earthquake caused the tsunami that claimed the lives of a great many people around the Indian Ocean.

“We wanted to see whether the methods and analytical tools we use to warn about tremors in mines could be used to predict earthquakes,” explains Leif Persson, who pursues research at FOI and Uppsala University.

In the area around Sumatra and the Andaman and Nicobar Islands earthquakes are common. The tension created in the area when continental plates rub against each other produces a number of minor earthquakes several years before the big earthquake comes. During the five years analyzed by the scientists there were 624 quakes in the area. On October 24, 2002, a powerful earthquake was registered at 7.1 on the Richter scale, and on September 13 the same year there was a 6.7 quake. The earthquake the day after Easter Sunday 2004 that caused the tsunami disaster measured 9.0 on the Richter scale.

Scientists from FOI have studied how the activity in the crust of the earth changed ahead of these tremors. They analyzed how the quakes were distributed in time and space within the area. The researchers used the relation between major and minor quakes, the so-called b-value. The lower the b-value, the greater the increase in tension in the earth’s crust, which entails a greater risk of major quakes. By plotting a ‘window’ consisting of 50 quakes and then moving this window in time from January 2000 to December 2004, they made a revolutionary discovery.

“We found that all of the major tremors were clearly visible in a time perspective. The b-value dropped drastically before the big quakes,” says Leif Persson.

To find out where the b-value was geographically at its lowest, the scientists plotted circles containing 50 quakes each on a map. The result of the analysis was extremely clear. The epicenters of the major earthquakes were clearly marked on the map. A new quake that occurred outside Sumatra on March 28 this year also confirmed the researchers’ analyses.

“In other words it is possible to predict where and when an earthquake is going to occur by using methods previously used only to provide warnings about mining quakes. There is a tremendous amount of interest in our conclusions in the scientific community,” says Leif Persson.