On the morning of September 29, 2009, a violent temblor shook the South Pacific seafloor between the islands of Tonga, Samoa and American Samoa. The earthquake stirred up tsunami waves that quickly pummeled the islands; nearly 200 people were killed, most of them in Samoa.

The deadly quake was the largest anywhere in the world that year. But new research indicates that the actual event was even more violent than initial estimates had shown and comprised three temblors: a magnitude 8.1 quake followed within minutes by another pair that combined to make an event of nearly equal size about 50 kilometers away, according to a study published in the August 19 issue of Nature. (Scientific American is part of Nature Publishing Group.) Another paper in the same issue of the journal draws similar findings from different data, concluding that the primary earthquake obscured a second quake of equal or slightly greater magnitude. If the analyses are correct, then the two largest earth-shaking events of 2009 occurred within minutes of each other in extremely close proximity.

The originally registered earthquake was a rupture within the Pacific tectonic plate, which meets the Indo-Australian Plate between the Samoan and Tongan island chains, but that event failed to account for all of the measured seismic and ocean waves. "Within a day it was recognized that when people predicted what that earthquake would be for normal faulting, there were some discrepancies with observations," says Thorne Lay, a seismologist at the University of California, Santa Cruz, and lead author of one of the studies. The possibility of a hidden quake was soon raised by Chen Ji of the University of California, Santa Barbara, at a meeting of the American Geophysical Union in December.

Lay's group used seismology data to uncover the second temblor, itself composed of two smaller events that together amounted to a magnitude 8 earthquake. The initial earthquake appears to have preceded the start of the secondary quakes by less than a minute, largely masking their presence. "There was in effect a stealth earthquake that was obscured by the first one," Lay says. "This is the first time I've ever seen a magnitude 8 earthquake that no one has in their earthquake catalogue."

A second team relied on displacements of GPS stations after the earthquake and measurements of ocean waves from tsunami-monitoring buoys to come to a similar conclusion—that a nearly simultaneous strong earthquake accompanied the initially recognized temblor. Although initial seismic analyses clearly identified the primary quake, some of the buoy data hinted at a different sort of temblor. The waves were consistent with slippage at the subduction zone between the Pacific and Indo-Australian plates rather than with a rupture on the Pacific Plate alone.

Corroboration came from GPS stations on the small Tongan island of Niuatoputapu, which had registered a dramatic shift. Instead of moving 80 millimeters southwest as would be predicted from the initial earthquake, Niuatoputapu had shifted 350 millimeters to the east. Those GPS measurements were an indication that "wow, there must have been movement from the subduction as well," says lead study author John Beavan, a geophysicist at the government-owned GNS Science in Lower Hutt, New Zealand.

"Two such large earthquakes so close together in time and space is unusual," Beavan says. In 2006 and 2007 the Russian Kuril Islands witnessed two earthquakes, each greater than magnitude 8, but the events were separated by about 60 days. "Typically there is a time delay of weeks or months," Beavan adds. "Having it happen on a timescale of seconds or minutes is pretty remarkable."

The Kuril Islands sequence followed a historical pattern for earthquake pairs—the subduction event preceded the rupture within the plate. Typically, slippage at the plate boundary tugs on the subducting plate, Lay notes, producing a smaller-magnitude rupture within the submerging plate later on. But the South Pacific earthquakes of 2009 reversed that trend: a violent plate rupture appears to have triggered rapid slippage between tectonic plates. "It's an unprecedented scenario," he says.

The surprising twists in the 2009 earthquakes cloud an already difficult process of predicting earthquakes and generating timely warnings of impending tsunamis after a tremor hits. "That's yet an additional level of complexity that most of the tsunami warnings don't account for," Lay says. "I don't have a solution for that. It's just the Earth being pernicious."