The strong earthquake that sent panicked Alaskans, fearing a tsunami, to high ground early Tuesday is a reminder that, at least around the Pacific Rim, the biggest danger from quakes often comes not from the land but from the sea.
Given the need for speed — deadly waves can strike within minutes in some cases — tsunami forecasters tend to shoot first and ask questions later, sending out broad alerts that are often scaled back or canceled within hours.
In the case of Tuesday’s quake, waves could have struck the nearest land, Kodiak Island, within half an hour. The U.S. Tsunami Warning Center, one of two operated by NOAA, sent out its first alerts within 3 minutes of the event, a typical time frame, said Michael Angove, who manages the tsunami program at the National Oceanic and Atmospheric Administration.
The quake Tuesday turned out to be largely harmless, a function of its 7.9 magnitude, which made it about 100 times less powerful than the world’s strongest earthquakes, and its location in the Gulf of Alaska about 170 miles from Kodiak Island.
But being centered offshore is no guarantee against destruction. Some of history’s most catastrophic quakes have occurred under the sea, with most of the deaths and damage caused by water rather than falling buildings or crumbling infrastructure.
In the 2011 Tohoku earthquake in Japan, for instance, the Fukushima nuclear reactors survived the magnitude-9.1 quake’s initial shaking. It was the larger-than-anticipated tsunami that struck minutes later that doomed the plant by knocking out emergency power generators.
In the 2004 Boxing Day earthquake near Aceh, Indonesia, also of magnitude 9.1, almost all of the quarter-million deaths were from tsunami waves that inundated coastal zones around the Indian Ocean.
And although the 9.2 magnitude 1964 Alaska earthquake — the second largest quake ever recorded — was technically centered on land, on a fjord in Prince William Sound, about 90 percent of the 131 deaths were from waves. In one native coastal village, the waterarrived while the ground was still shaking, killing one-third of the residents.
Those quakes occurred around the Pacific Rim, as do thousands of others every year.The margins of the Pacific are where the tectonic process known as subduction is readily apparent. In subduction, one large section of the earth — a plate — slides slowly under another, as the earth’s surface recycles itself over geologic time.
Sometimes, parts of a plate get stuck until they break in an earthquake. If this happens near or under water, then the movement of blocks of rock displaces massive amounts of water, spawning a tsunami.
Some kinds of quakes are much more likely to cause tsunamis than others; so-called megathrust earthquakes, in which one side of the fault rides up over the other, are especially likely to generate waves. The Tohoku, Aceh and 1964 Alaska quakes all involved megathrust faults.
But tsunami forecasters seldom have the luxury of time to wait for an analysis of the quake’s mechanism, Mr. Angove said.
“When we have an event that’s near our coast, it’s a very challenging problem for us,”he said. “There is a lot of sophisticated seismic waveform analysis. But we don’t have time to do that level of analysis when you’re talking about eight to 10 minutes of travel time.”
Instead, he said, the decision to send outthe initial alerts is based on a rapid analysis of two factors: magnitude of the quake and its location.
On Tuesday, the advisories continued in the hours after the quake, with some cancellations announced for parts of the west coast of the continental United States. Small tsunamis of about half a foot were reported in several Alaskan communities on Kodiak Island and elsewhere before the last alerts were canceled about 3 and a half hours after the earthquake.
The decisions to maintain or cancel alerts were based on actual readings from buoys showing water levels in key locations, Mr. Angove said, rather than from analysis of the quake motion.
As it turned out, within a few hours the United States Geological Survey had determined that the quake was not a megathrust one. Rather, it had more in common with the San Andreas fault, and its horizontal motion would have been unlikely to displace much water.