By all rights, the 8.8-magnitude earthquake that struck

Chile in February 2010 should've

caused a tsunami that wreaked havoc on Hawaii. The quake was massive — tied for the

fifth most powerful since 1900. It originated in a subduction zone along the

Chile margin, a unique geological area where three tectonic plates are being

subsumed by the South American continent.

If that wasn't enough of an omen, a

9.5-magnitude quake in Chile 50 years earlier generated a tsunami that bent Hawaiian

parking meters like paperclips.

Yet, when the 2010 quake occurred, the resulting tsunami

arrived in Hawaii with a whimper. In fact, it barely warranted evacuation

warnings issued by the National Oceanic and Atmospheric Administration's (NOAA) Pacific

Tsunami Warning Center in Hawaii. Those warnings were based on a tsunami model that did not accurately predict the effects of Chilean tsunami.

The failure of that model — and the accuracy of another created by NOAA's Center for Tsunami

Research in Seattle, Washington — will prove invaluable at predicting future tsunamis. After all, each

prediction failure is an opportunity to figure out why the system failed in the

first place.

But the failure also underscores another question: Are

tsunamis predictable?

Researchers have become increasingly adept at predicting

what a tsunami will look like once it occurs, based

on emerging technology and information culled from past quakes and tsunamis.

 "We're still building models from

the 1960 Chilean earthquake," said Chris Goldfinger of Oregon State University.

"Twenty years from now, people will be sorting out the models for the 2010

quake. It's really a never-ending process."

Models that forecast the effects of tsunamis on Hawaii,

for example, must consider an underwater ridge that protects it from most


Forecast models depict the size and number of tsunami

waves, approximate arrival time, where the waves will make landfall, and how far

inland they will reach. They are compiled by tsunami researchers for faults

around the world; like an 8.8-magnitude quake along the Chile margin, for

example. When this quake occurred, researchers pulled up models based on this

very event; responders based their reactions on the predictions, which led to evacuation


Because of the factors involved in compiling an accurate depiction

of a tsunami and the relatively sluggish computing power currently available,

researchers compile models before they occur.

"It just takes too long to do them in real

time," Goldfinger said.

Time is an essential factor following a tsunami, and

quickly detecting tsunamis when they occur is as much a part of the field as

predicting the course and form they will take. Even a few extra minutes of warning can give

people in the tsunami's path time to reach higher ground, which unequivocally saves lives. The

Indian Ocean tsunami of 2004 struck land within 20 minutes and

killed around a quarter-million people because the area lacked an early warning


Since around 85 percent of tsunamis come from the Pacific

Ocean, American researchers have focused  largely on the Pacific Northwest in

recent years. Geologists have predicted that the Cascadia subduction zone that

runs off the coast from Mendocino, Calif. to Vancouver, British Columbia may be due for

a massive quake. It's one of only two fault lines in the U.S. capable of

producing a major tsunami.

Since the mid-1990s, NOAA

has placed buoys in the Pacific between known faults and land. These buoys

are capable of sensing and recording tsunami activity and relay the information

back to land as an early warning system. As a tsunami is generated, these

tsunameter buoys can provide hours of additional time to reach safety.

The profile of a predictable tsunami is still being developed.

Most recently, researchers have begun to investigate how to detect faint

electrical currents created by tsunamis. As wave action drags ions through

saltwater and along the Earth's magnetic field, it generates an electrical

field that Manoj Nair of the University of Colorado has proposed may

be detected as an early warning system.

Tsunami researchers are taking this science

as far as they can, but they consistently run into a roadblock. Most tsunamis

are generated by earthquakes, and to predict one before it happens requires the

ability to predict a quake. Scientists have thus far been unable to make narrow


"We can't currently predict earthquakes over

timescales of days, weeks, months or even a few years," said University of

Washington seismologist Heidi Houston. "We have mostly moved to

forecasting earthquakes in a region over the next 30 to 50 years."

Still, the work tsunami researchers are compiling would

prove adequate in predicting tsunamis should earthquakes ever become


"If we

could predict an earthquake in sufficient detail, then scientists could do a

pretty good job predicting the resulting tsunami," Houston added.

Image: Indian Ocean tsunami, 2004 (NOAA); boat flung ashore in Indonesia by Indian Ocean tsunami (Chuck Simmins, Flickr)