When the Wenchuan earthquake killed some 80,000 people in southwest China in May of last year, suspicion immediately fell on the reservoir behind the nearby Zipingpu Dam. Seismologists knew that several hundred million tons of water had filled the reservoir in the preceding few years and that either the water itself or its weight might have weakened a nearby fault and unleashed the quake. A new analysis finds that both scenarios are plausible, but further insight will require the cooperation of the Chinese government.
Last December, an American researcher was the first to prominently report (Science, 16 January, p. 322 ) that the Wenchuan quake may have been triggered by human activity. That study focused on the idea that the sheer weight of the water in the reservoir may have weakened the adjacent Beichuan fault, either by counteracting the stress that was strengthening the fault by squeezing it together or by adding to the stress tending to rupture the fault.
The new analysis finds support for the added-weight hypothesis and also for the idea that the water itself might have seeped kilometers downward into the fault, where quakes get started. Once there, it could have pressurized the water, pushing the fault apart and weakening it. When hydrogeologist Shemin Ge of the University of Colorado, Boulder, and her colleagues put both the water's weight and plausible penetration of the water into their calculations, the amount and direction of stress change on the fault came out to be about what large quakes create when they trigger failure in nearby faults. "We thus suggest," the authors write  in their 28 October Geophysical Research Letters paper, "that the Zipingpu Reservoir potentially hastened the occurrence of the Wenchuan earthquake by tens to hundreds of years."
That "is plausible," says seismologist Ross Stein of the U.S. Geological Survey in Menlo Park, California. But the case isn't yet closed, Stein says. The study "shows a possibility, not a probability."
To make stronger statements, more observations would have to be incorporated into analyses, says Stein. Many such data sets may exist, he says, but the Chinese government has kept them off-limits for most researchers. Meanwhile, he says, China and India continue to build large dams in settings just like Zipingpu's, he notes. If they proceed, such projects demand close monitoring long before the water goes in.