Dec. 26, 2004 dawned calm in Southeast Asia, but things didn’t stay that way for long. At 7:59 a.m. local time, deep beneath the seafloor west of Sumatra, two of Earth’s tectonic plates began to slip past one another, releasing stress that had built up between the overlapping slabs for decades. An earthquake, the largest anywhere on Earth in more than 40 years, had begun.
The tip of the rupturing area quickly spread upward to the ocean bottom and then raced along the seafloor faster than a bullet shot from a rifle. Vast expanses of seafloor, along with the ocean above it, were thrust upward as much as 5 meters. Tsunamis raced away from the sudden bump in the ocean at jetliner speeds and crashed into coasts.
According to official tallies by the 11 countries struck by tsunamis that day, the huge waves killed nearly 180,000 people. Eight months later, another 50,000 or so are still missing. The deaths of so many people, including tourists from dozens of nations, made the event a global tragedy.
The physical effects of the massive quake were felt worldwide as well. Beyond the area of high-frequency rumbling, the surface of the ground slowly heaved up and down at least 1 centimeter. That motion, though imperceptible, triggered a swarm of small earthquakes in Alaska, a quarter of the way around the globe from the epicenter of the quake.
The tsunamis, which towered a dozen or more meters high along the western coast of Sumatra, were detected at heights of at least 1 cm at many tide gauges around the globe.
Global Positioning System (GPS) equipment in Southeast Asia chronicled the ground’s sudden horizontal surge of several centimeters in a matter of minutes during and after the quake.
Ever since Dec. 26, scientists examining such data have been discovering how Earth behaves during and after a massive earthquake. Postquake analyses indicate that constant monitoring of a dense network of GPS instruments near earthquake-prone regions could play a critical role in future tsunami-warning systems.
In the early 1960s, the last time there was an earthquake as big as December’s, seismic instruments traced ground motions on paper charts with colored pens. Today, however, many seismometers are digital, simultaneously recording in high fidelity at many different frequencies the tiniest of ground motions. To log the same amount of data with 1960s-era technology, a strip-chart recorder would require paper 180 m wide.
Furthermore, the modern seismometers are deployed worldwide in broad regional networks—a configuration that enables scientists to analyze what happened during the Sumatran quake from many different angles, both literally and figuratively.
Even modern seismometers, however, offer mostly a look back at an event. Because detailed analysis of their data takes hours—or even days—the machines make a poor early-warning system for tsunamis from earthquakes.
The fastest vibrations to emanate from December’s temblor—and therefore the first to arrive at seismometers worldwide—were the so-called P waves, which travel through rock as pressure waves. By analyzing when those waves reached seismometers, scientists have pinned down the quake’s epicenter as about 30 kilometers below Earth’s surface and 150 km off the western coast of Sumatra (SN: 1/8/05, p. 19: Tsunami Disaster: Scientists model the big quake and its consequences). That spot lies along the line where the India tectonic plate is forcing its way beneath the Burma plate at an average rate of about 6 cm per year.
During the first minute or so after the forces holding the two tectonic plates in place began to give, the tip of the rupture spread up to the ocean floor and then raced northwest at a speed of about 500 m per second, says Thorne Lay, a seismologist at the University of California, Santa Cruz. If the slippage had stopped after that first minute, the quake would have registered a respectable magnitude of around 7.
However, the rupture next accelerated to a speed of about 3 km/sec for 4 minutes. Then, it slowed somewhat but continued north for another 6 minutes.
In all, slippage during the Dec. 26 quake occurred along a 1,200-to-1,300-km-long stretch of the two plates’ intersection, or subduction zone. That’s the longest rupture ever recorded for an earthquake, say Lay and his colleagues. At the northernmost reaches of the rupture zone, the tectonic plates probably slipped only 1 or 2 m during the quake. But deep within Earth, at the southern end of the Sumatran temblor’s rupture zone, the tectonic plates probably slipped past each other between 15 and 30 m in less than a minute, says Lay.
Magnitude estimates, based on both the size of the fault zone and amount of slippage, range as high as 9.3, Lay notes. He and his colleagues described the Sumatran quake in the May 20 Science.
In just 11 minutes, the temblor released energy equivalent to that stored in 250 million tons of TNT, says Roger Bilham, a seismologist at the University of Colorado in Boulder. That’s approximately the same amount of energy released by all earthquakes worldwide during the previous 10 years, notes Lay.
The quake initiated, in addition to P waves, a rolling motion called Rayleigh waves that spread through the ocean floor and dry land. These waves, for instance, caused the ground to move up and down 9 cm in Sri Lanka, more than 1,500 km away.
A blast of Rayleigh waves spaced 20 to 30 seconds apart went around the world repeatedly for days after the quake. In most spots, they passed without fanfare. In Alaska, though, the first pass of the ground motions triggered a swarm of small earthquakes near Mount Wrangell, a volcano about 330 km east-northeast of Anchorage and more than 11,000 km from Sumatra.
The Rayleigh waves reached the network of six seismometers around Mount Wrangell about an hour after the quake occurred in Sumatra, says Michael West, a seismologist at the University of Alaska in Fairbanks. Data from those instruments suggest that each vibration moved the ground in the region up and down about 1.5 cm. During the 11 minutes in which the first Rayleigh waves passed through, the seismometers recorded a swarm of 14 quakes. Scientists located six of these quakes with precision.
The Alaskan quakes were spaced about 30 seconds apart, the same interval at which the peaks of Rayleigh waves were sweeping through the region, says West. Furthermore, a tally of the seismic activity around Mount Wrangell in the 2 days before and the 2 days after the Sumatran quake suggests less than a 1 percent chance of having six random earthquakes in the span of 10 minutes.
The clincher, says West, is that every one of the Alaskan quakes occurred at a point in the cycle when Rayleigh waves would have caused horizontal ground motions that weaken the forces holding together a fault, making an earthquake more likely. West’s team reports its analyses in the May 20 Science.
Rayleigh and other seismic waves from the Sumatran quake circled the globe many times, growing weaker with each circuit. The planet vibrated as if it were a bell struck by a giant hammer, Jeffrey Park of Yale University reported in May at the meeting of the American Geophysical Union (AGU) in New Orleans.
Some vibrations match the frequency of Earth’s so-called breathing mode, in which the entire surface of the planet rises in unison and then falls. As of the beginning of July, Earth’s surface was still vibrating from the 2004 quake, each cycle moving the ground up and down about 0.5 micrometer every 20 minutes or so.
On Dec. 26, seafloor upheavals triggered tsunamis that scoured coasts around the Indian Ocean. In the decades before the quake, the Burma tectonic plate had been slowly flexing, absorbing the energy of its collision with the India plate.
When plate-to-plate friction gave way, the Burma plate snapped back toward its relaxed shape. In minutes, a broad expanse of seafloor about 300 km west of Sumatra rose as much as 5 m. Closer to shore, some regions of ocean bottom dropped about 2 m during the same interval. The resulting bump and trough in the surface of the Indian Ocean spawned that day’s deadly tsunamis. Overall, shifts in the seafloor during the quake displaced about 160 cubic km of water, says Bilham.
The first wave, possibly with a height of 15 m or more, slammed into the Sumatran coast within minutes of the quake. Thailand was struck about 75 minutes after the temblor, and water levels began to rise at Colombo, on the southwestern shore of Sri Lanka, almost 3 hours after the quake. The waves killed people in several East Africa nations, more than 5,000 km away from the quake. On the coasts of North and South America, Antarctica, and the Arctic Ocean, tsunami waves washed in but measured just a few centimeters.
For the first time, instruments on several Earth-orbiting craft caught glimpses of tsunamis as they spread. The Jason satellite, which is jointly operated by the United States and France and is equipped with a radar altimeter, happened to pass over the eastern Indian Ocean about 2 hours after the quake occurred. The satellite spotted a tsunami, which measured only 90 cm tall in the open ocean, speeding along at speeds of about 700 km per hour. Although this sighting was more than an hour before the first waves struck Sri Lanka, the Jason-satellite data couldn’t have been used to warn tsunami victims because scientists needed several hours to analyze the information.
About 4 hours after the quake, NASA’s Terra satellite passed over the strait between Sri Lanka and southeastern India, says Michael J. Garay of the University of California, Los Angeles. Its cameras captured an unprecedented series of orbital images showing the tsunamis in action. As the tsunamis approached the Godavari River delta in India, they encountered shallow water, slowed down, and grew to several meters high.
Analyses of the Jason-satellite photos of the same area suggest that the tsunamis were traveling about 45 km/hr when they washed ashore. The handful of deadly waves that inundated the river delta that day were spaced about 9.3 km apart and struck land every 12 minutes, says Garay, who described the images at the AGU meeting in New Orleans.
Seismometers and satellites weren’t the only sentries on duty during the December quake and tsunamis. GPS equipment installed throughout Southeast Asia monitored the movements of Earth’s crust. These devices constantly receive signals from orbiting satellites, which enables scientists to track ground movements within centimeters.
Although GPS equipment isn’t currently part of any earthquake monitoring system, postquake analyses of data gathered by the devices in Southeast Asia suggest that it could be.
GPS equipment across a broad region, including sites in China, recorded a shift as a result of the Sumatran quake. Most of this slippage occurred along the southern portions of the rupture zone.
Some spots in the Andaman Islands, near the northern end of rupture, moved more than 4 m to the west during the first few hours after the quake, and the tectonic plates deep within Earth scraped past each other another by 7 to 20 m during this slow-slip phase.
In Medan, Indonesia, about 300 km east of the quake’s epicenter, the ground surged westward 14 cm in the 15 minutes or so after the quake began, says Jeffrey T. Freymueller of the University of Alaska in Fairbanks. Equipment at many other sites in Southeast Asia observed sudden shifts of a similar magnitude—a sure sign that the earthquake was huge, he noted at the AGU meeting.
In the 4 hours after the Sumatran quake, seismologists updated their estimates of the temblor’s magnitude—always upward—at least four times. That’s because the traditional method of estimating magnitude depends on the slow-moving seismic waves produced by a quake.
However, scientists have shown that analyses of GPS data could have revealed the approximate energy released by the earthquake in less than 10 minutes, says Freymueller. That’s quick enough and accurate enough for a system that would issue tsunami warnings on the basis of undersea seismic activity, he adds.
The tsunami-warning network now in place in the Pacific Ocean depends on seafloor sensors that relay signals from tethered buoys to satellites (SN: 3/6/04, p. 152: Killer Waves). These instruments are geared to gather data about a spreading tsunami and to alert shoreline civil defense authorities within minutes. However, no such warning system exists in many tsunami danger zones, including the Indian Ocean.
The current GPS system, by reading land movements on the shores of oceans, could supply warnings almost as quickly as does the Pacific Ocean system, with its seafloor sensors and buoys, says Freymueller.
If GPS equipment that relayed its data round-the-clock was positioned every 200 km or so along subduction zones where large earthquakes occur, scientists would have much of what they need to provide some measure of tsunami warning, he notes. GPS systems would be cheaper to install and maintain in some vulnerable areas than a sensor-buoy system would be, yet they could be as rapid in their response.
Freymueller says, “When a tsunami’s on the move, you don’t measure time in weeks or in days, but in minutes.”