Floodwater took just 16 minutes to scour
2 miles

A model by California researchers raises new questions about how to plan for future breaks

After Ka Loko Dam burst about 5:30 a.m. Tuesday, it took 16 minutes for a wave of water -- as high as 20 feet -- to move over two miles of terrain and wash over Kuhio Highway, according to a state-of-the-art model produced by a University of California research team.

The only warning residents of a small Kilauea community got that the monstrous wave was headed their way was its thunderous, rumbling roar -- the sound of 330 million gallons of water, littered with debris, decimating everything in its path to the sea.

It came as some people were eating breakfast, state Civil Defense Vice Director Ed Teixeira said. Others were perhaps awakened from their sleep. Moving at between 5 and 10 mph, after slowing down from up to 20 mph in steep, narrow canyons higher in Kilauea, the waters tore through homes around Wailapa Road, sweeping seven people and two homes away.

Based on the model, officials say that even if authorities had been alerted as soon as the dam broke, there's little chance an evacuation could have been completed in time to save those who perished.

"Once the dam breaks, there's just so little time to do anything," said Brett Sanders, the University of California at Irvine associate professor of civil engineering who put together the Ka Loko computer simulation.

"Do you even know which way to run?"

The model, which Sanders provided to state Civil Defense, is raising new questions about how to successfully plan for future dam breaks, especially when most earthen dams in the islands don't have electronic monitoring devices, which warn officials when water levels are too high.

For many dams in the state, residents wouldn't know about a break until floodwaters were headed their way. And even if they did have a few minutes warning, most wouldn't know which way to flee.

In fact, officials say, the bulk of those who live in a dam's flood zone don't know it. "I think that now will become a question when you're buying property," said Ray Lovell, spokesman for state Civil Defense.

The agency has lauded Sanders' Ka Loko model and is already thinking about its potential uses in mapping out where floodwaters would likely go if a particular dam broke. Such models do exist, officials said, but they are based on older data and simpler calculations.

Also, high-risk dams are required under state law to have emergency action plans, which include evacuation areas. Though Ka Loko Dam was not "high risk" -- because there was only a small population downstream -- such plans could be required of all dams in the future, state Department of Land and Natural Resources Director Peter Young said.

And a computer modeling program would greatly ease the work of figuring out the path of floodwaters from a breached dam.

"Imagine that every reservoir could have a dam break simulation done," Sanders said, adding that dam flood zones could be distributed as widely as tsunami inundation areas are.

"If planning was done ahead of time, then folks could see ... that going up may not help you if you're going in the path of the flood wave."

It's still unclear whether Sanders' model of the Kilauea flood exactly matches what happened on Tuesday, as officials have not yet been able to make their own map of the floodwaters.

But both Teixeira and Maj. Gen. Robert Lee, head of the Hawaii National Guard, said the model appears accurate. Both have seen the flood damage from the air.

Sanders said his simulation program, funded by a National Science Foundation grant, isn't yet available to planners. But he's willing to provide it for some dams in Hawaii given Tuesday's tragedy. His model takes into account topography along with the likely speed of floodwaters.

Sanders explains the program that produced the Ka Loko model in an article set to appear next month in the Journal of Hydraulic Engineering.

"I would argue that what we have is probably the next-generation-type model. We have developed a fair amount of confidence in terms of what the model is able to do," Sanders said, adding he decided to run his program on the Ka Loko flood after hearing about it on the news.

By coincidence, a member of his research team is from Kauai.

When Stephen Esaki, who was born and raised in Kapaa, heard about the dam failure, he called home right away.

Then, he set to work researching the dam break to get figures, which were plugged into the program. "I was just shocked it would apply so close to home," said the University of California at Irvine senior.

"Since Tuesday, I've been thinking about how we could use the model to help communities to understand, if it (another break) were to happen, where we could evacuate."