Residents of Tokyo likely had about 80 seconds of warning before a devastating quake rumbled through the city after striking 373 kilometers away, off Japan’s northeast coast, thanks to a new early warning system. But tsunami alerts take longer to generate, giving just minutes of warning before the waves first struck the coast—a reflection of the differing technologies needed to detect earthquakes and calculate their impacts, researchers say.
Japan has the world’s most advanced earthquake early-warning system, with more than 1,000 seismographs scattered over the country. Collectively, they detect tremors and allow for brief advance warnings not only to vulnerable sectors like railroads and utilities—so they can slow down high-speed trains and shut off gas lines—but also to the public via television, Internet and text-message. “This hits probably what I would consider the best prepared country in the world for earth quake preparedness,” says Stephane Rondenay, a geophysicist at MIT.
The seismographs detect the first evidence of a quake by the arrival of
P-waves, which have long wavelengths and generally do little damage. The
shorter-wavelength and slower–but far more damaging–S-waves come next,
usually seconds later. The time difference in the arrival of these two
type of waves can be used to estimate distance to the epicenter.
The quake struck at 5:46 GMT. It would have taken about ten seconds for sensors to detect enough signals to conclude the quake was serious and issue the alerts. Since the more damaging secondary waves travel at 4 kilometers per second, it would have taken them about 90 seconds to travel the 373 kilometers to Tokyo, Rondenay says.
While the systems can only furnish warnings from a few seconds to a minute or two before serious shaking starts, this can be enough time for people to take cover, stop performing surgery in a hospital, exit an elevator, or pull over to the side of the road. Countries including Taiwan and Mexico are implementing similar systems, and California has been researching a warning system. A Stanford University research effort, called Quake-Catcher Network, aimed at using data from accelerometers in laptop computers to detect shaking.
Tsunami warnings take longer to generate. The Japanese government issued a local tsunami warning three minutes after the quake struck, and the Pacific Tsunami Warning Center, run by NOAA, issued its regional warning nine minutes after the quake struck, or at 5:55 GMT. These longer response times were a reflection of heavier computation requirements. NOAA must determine whether the quake happened in an ocean basin, the likely deformation of the ocean floor, and what kind of motion was created by the quake, Rondenay says. Too many false alarms would weaken faith in the system, and warnings issued too quickly would not have good data about arrival times in various regions, he added. “A lot more parameters have to be taken into account,” he says.
Fortunately, while tsunamis are fast, they are not as fast-moving as the earthquakes themselves. Tsunamis travel as fast as jet aircraft in deep water, where they are spread over a far deeper water column and make for smaller and very gradual waves at the surface. They slow down and produce higher waves in shallower water. Residents of the hardest-hit areas likely only had 15 minutes of warning, though Tokyo would have had at least 40 minutes of warning.
Japan may also have benefited from structural improvements made since the devastating Kobe earthquake in 1995. “My sense is that buildings behaved pretty well. More and more buildings now are being made with dissipation devices,” that absorb tremors, said Eduardo Kausel, an engineering professor at MIT who specializes in seismic engineering. “The Japanese learned a hard lesson in Kobe and were embarrassed by the level of damage, so they have tried to correct most of the known defects in the intervening years. But no matter what you do, you will have damage. There is uncertainty in earthquake engineering.”
When designing an embedded system choosing which tools to use often comes down to building a custom solution or buying off-the-shelf tools.