Predicting an earthquake is the ultimate dream for many geologists. Every day the Earth judders and shakes with about 50 earthquakes. Many of these are not felt because they are very small, or because they are in remote areas inhabited only by reindeer and polar bears. But when a strong quake hits a heavily populated area, the effects can be devastating.
The last time this happened was on Dec. 26 last year, when a magnitude 6.3 earthquake razed the city of Bam, in southeastern Iran. At least 26,000 people died and many thousands more became homeless overnight. If the people of Bam had been given just a little warning, most of those lives could have been saved.
For thousands of years, humans have lived at the mercy of earthquakes, but at last things might be about to change. Alasdair Skelton from Stockholm University and colleagues think they may be able to take us one step closer to forecasting when the ground is going to shudder. They are testing their idea in Iceland and soon hope to be able to give Icelanders a few weeks' warning before a big quake. What is more, the scientists are optimistic that their method could eventually be applied to other earthquake-prone countries, bringing peace of mind to millions all over the world.
The theory is straightforward. When an earthquake first begins to flex its muscles, it stresses rocks far underground and opens up little cracks. Eventually, the cracks link up and water that has been trapped deep underground starts to travel upwards, making a break for the surface.
"The cracks may be only a few millimeters wide, but it is enough to allow the water to move," Skelton said.
As the water fights its way up it dissolves minerals in the rocks it passes through. Some minerals dissolve more easily than others and the first water to arrive at the surface carries the minerals that were easiest to dissolve, while the later water carries minerals that took a bit more persuading.
Skelton and his team believe they can use the changes in the water chemistry to estimate when the cracks first opened and to give some idea of how long it will be before the earthquake strikes.
Iceland is an ideal place to study earthquakes because of its simple geology and frequent earthquakes -- about 100-200 a week.
"If you dig a hole at our sampling station on Iceland, the first kilometer will be through sediments and then after that it is mostly a volcanic rock, called basalt, for a very long way," Skelton said.
In the summer of 2002, Lillemor Claesson, Skelton's doctoral student, started sampling water once a week from boreholes, dotted around in five different areas of Iceland. Initially, all her sample sites were going to be in the south of Iceland, but after some discussion with local people she also chose one site in the north, known as the Tjornes Fracture Zone (TFZ).
By extraordinary chance, a magnitude 5.8 earthquake occurred on Sept. 16 2002 in the TFZ, just 10 weeks after she had started taking samples.
"We were really lucky to have an earthquake to study so quickly," she said.
Skelton and Claesson rushed their water samples back to the lab to see if analysis could detect a change in the chemistry in the period leading up to the TFZ quake.
To their delight, they discovered clear spikes in the concentrations of different chemicals: the earliest samples contained iron and chromium, later ones a bit of manganese, next came zinc and finally, samples just before the quake contained copper.