Cracking the Tsunami Code

It's a tale of two earthquakes. One, the second-biggest ever recorded, ripped the seafloor just west of Sumatra, Indonesia, in December 2004 and unleashed a tsunami that caused destruction across the entire Indian Ocean basin. The other, nearly as large, struck the same general area 3 months later but caused only a localized tsunami and much less coastal devastation. Published tomorrow in Science, an international team reports finding what could be the critical difference between these disasters: the density of the sedimentary rock overlying the quake zone. Measuring that variable in other, geologically similar areas prone to major earthquakes could give seismologists and emergency planners a better understanding of where to expect the worst.

The giant Sumatra earthquake that hit on 26 December 2004 triggered a fast-moving tsunami that drowned nearly a quarter of a million people and left millions homeless in coastal areas as far away as East Africa. But that was only a part of the nightmare. The big quake, which registered magnitude 9.2, generated thousands of aftershocks in the ensuing months, including a magnitude-8.7 temblor in March 2005.

The second quake caused plenty of damage in the Andaman Islands and the adjoining coastal areas of Sumatra—but only a relatively minor tsunami. Immediately afterward, scientists began wondering why two large earthquakes, both hitting the Sumatra fault only a few months apart, could produce such divergent results. So the researchers set out to compare the seismic characteristics of the quake locations. During the summer of 2008, they sailed along the Sumatra fault in the research vessel Sonne and used blasts of compressed air to bounce sonic echoes off the seafloor, creating density profiles of the underlying rock. Each quake had occurred when one huge chunk of Earth's crust suddenly slipped under, or subducted, another. Many such subduction zones exist across the globe, such as along the entire west coasts of North and South America. In the Sumatran earthquakes, the Indian Ocean plate subducted the Indonesian plate. The researchers analyzed the resulting profiles for distinctions in the two subduction zones, also called décollements.

The scientists found that the décollement of the 2004 quake contained a stretch of sedimentary rock under the sea floor that was considerably less dense, and therefore weaker, than the surrounding rock. When the temblor struck, that weakness probably caused a bigger portion of the sea bottom in the area to deform suddenly, thereby displacing more seawater and creating bigger tsunamis. The décollement of the 2005 quake, in contrast, consisted of denser rock, which slipped less during the event and displaced far less water.

It's a distinction that could have implications for other locations prone to subduction-related earthquakes, says marine geophysicist and lead author Simon Dean of the University of Southampton in the United Kingdom. Wherever a fault zone contains a well-defined patch of weakness, he says, it could lead to "a very different type of earthquake rupture, one which is more likely to cause a large tsunami."

The paper "identifies important clues in the sediments [that reveal] why the 2004 Sumatran earthquake generated a deadly tsunami and the adjacent 2005 earthquake did not," says seismologist Arthur Frankel of the U.S. Geological Survey in Seattle, Washington. Those clues could indeed help to predict "whether great earthquakes in other subduction zones will produce large tsunamis," he says.

Marine geophysicist Jian Lin of the Woods Hole Oceanographic Institution in Massachusetts agrees. The paper provides "strong evidence" that sedimentary layers can "have a major influence on the behavior of [earthquakes] tsunamis," he says.