WASHINGTON — No one had any prior notice when earthquakes shook Hawai'i last Sunday morning, leaving swaths of damage throughout most of the state.

That's because the science of predicting earthquakes, especially large ones, is still in its infancy, and that's putting it charitably. Experts can't even agree if it's possible to predict when big earthquakes will hit.

"I really don't know" whether it's possible to make earthquake predictions, said Lucy Jones, a seismologist and former scientist-in-charge for the U.S. Geological Survey in Southern California. "Fifteen years ago, I actually would have said I don't think so. I'm not so sure of that anymore. There may be real differences in how a big one gets started."

Paul Segall, a geophysics scientist at Stanford University, said it's unlikely that scientists will be able to make reliable earthquake predictions in the next decade. "I certainly wouldn't bet a lot of money on it," he said.

For now, seismologists studying faults, tremors and past earthquakes can estimate the probability of earthquakes, and how strong they'll be, in a specific area over the next 30 to 50 years, said David Oppenheimer, a USGS seismologist in Northern California.

While that might be good for determining building codes, it doesn't do much to warn people that they should prepare for a disaster.

Cree Tabac, who was in her daughter's house in Kawaihae Village on the Big Island when the quake struck near there last Sunday, said she would have appreciated a warning.

Some kind of warning would have helped, because she would have made sure that family members were accounted for and that supplies were adequate, she said.

"You are going to make sure that you have water, milk for the babies, that kind of thing."

In California, one of the most active earthquake regions in the country, earthquake forecasting has taken another step with USGS scientists now producing a continuously updated forecast map.

"That's an earthquake prediction," Jones said. "That is telling you the increased chance of an earthquake in the next 24 hours."

But that prediction is based on 50-year earthquake probability maps combined with statistics of how earthquakes trigger other earthquakes, which are mostly aftershocks, Jones said.

Oppenheimer said Japan, Mexico and other countries have early-warning systems that use sensors and computers to sound a warning once an earthquake starts. It sends a warning ahead of the damaging ground waves and gives an alert, measured in tens of seconds, before they hit.

"The amount of warning is limited, but there are still things you can do," Oppenheimer said.

The Japanese use the system to stop high-speed trains. Other possibilities include changing traffic lights to red to close bridges and having elevators in tall buildings stop at the next floor to let people off, he said.

The warning system idea is being studied in California now in a three-year project, Oppenheimer said.

"We don't know what the verdict will be on whether it's feasible in California, mainly because the earthquake faults are so close to urban areas," he said.

Jones said despite the past failure of earthquake predictions, study continues on the underlying forces that control how earthquakes happen.

"The hope is that this will eventually lead to usable earthquake predictions," Jones said.

She said a new Center for the Study of Earthquake Predictability has been established to create standards for such claims.

One problem is, somebody is always predicting an earthquake and sometimes they are lucky and get it right, Jones said.

"Yet how do you discriminate between somebody who is just lucky and somebody who actually has information about the time a earthquake will occur. We are going to at least know within five years whether or not to give up on earthquake predictions. That's a big step forward."

Reach Dennis Camire at dcamire@gns.gannett.com.