The DHS is funding the development of a model that can run simulations based on changing conditions before and during a storm.

By Amy Wold
The Advocate

SHREVEPORT, La. — With the start of another hurricane season today, thoughts should be turning to discussion of preparedness kits and evacuation plans.

However, a group of LSU researchers have made hurricane evacuations — and how to improve them - a year-round pursuit.

For two and a half years, civil engineering professor Brian Wolshon and other researchers have been working on

The city of Houston empties out as Hurricane Rita approaches landfall along the Gulf Coast, Sept. 22, 2005. (AP Photo/Pat Sullivan)

turning a traffic-simulation model into one that can be used for hurricane evacuation planning, which is crucial for residents in south Louisiana and the Gulf Coast region.

Although the evacuation for Hurricane Katrina in 2005 was effective — estimates are that 90 percent of the New Orleans area population left before the storm — there are still some shortfalls.

"What do you do for people who can't evacuate themselves?" Wolshon said.

That's part of the next step in using a new computer modeling system that takes population and land use data and creates a synthetic population, he said. It may not represent a particular person in a particular house, but it's generally representative of people who live in a certain area — how many cars they have, how many family members, where they work and more.

All of that figures in to the time it might take to evacuate a large city, such as New Orleans, during a hurricane, Wolshon said.

Currently, the computer model is being truth-tested, to see if it will provide reliable results. Researchers input conditions that existed during Katrina and then run the model to see if the results reflect what actually happened during that evacuation.

Researchers can then change the conditions and ask what would happen if there was a road closure, major accident or sudden change in storm speed.

During Katrina, there were no washouts of roads, no incidents forcing a major highway closure and no ongoing construction work. In addition, the hurricane evacuation occurred on a weekend and there was enough notice to make it all happen, Wolshon said.

"We were very fortunate with Katrina that there weren't any incidents that closed any roads," he said.

Also, the highway evacuation for people driving their own cars worked pretty well. "Really, there's no real visible outcry to change it," Wolshon said.

But just because the effort worked well last time doesn't mean it will work well the next time, Wolshon said. If one or more of those conditions fails to fall in line, the results could be different, he said.

The new computer model allows researchers to include millions of cars in the system. (AP Photo/Paul Sancya)

The new computer model will help planners anticipate some of those potential problems.

The computer model the LSU researchers are working with was developed in the 1990s at the Los Alamos National Laboratory, Wolshon said. But, it was ahead of its time and so complicated and detailed that when released it "fell on its face" because it couldn't run with available technology, he said.

The U.S. Department of Transportation gave funding to Wolshon's LSU group to adapt the standard traffic model for hurricane evacuation traffic analysis, he said. The federal Department of Homeland Security is funding the development of a model that can run simulations based on changing conditions before and during a storm, Wolshon said.

With the old computer model, researchers could only look at a 19-hour time frame and only include about 300,000 vehicles. Now, they can look at a much longer time frame and include millions of cars in the system, which gives a better picture of what a regional evacuation could look like.

In addition, before this new model, researchers could only look at one travel corridor at a time, but now they can look at every New Orleans street and how traffic patterns might be affected, Wolshon said.

This information will be used as researchers continue to look at ways to improve assisted evacuation for people who may be unable to evacuate themselves, he said.

For example, Wolshon said, plans to use buses and other transportation can be input to see how those modes would interact with other traffic.

Another thing researchers can look at using this computer modeling is how they can minimize the potential of harm to people who may be elderly or ill and where a 12-hour or 14-hour wait in normal evacuation can prove to be deadly, Wolshon said.

One idea is to open one lane of contraflow travel, one way traffic along highways and interstates in southeast and southwest Louisiana, to be used only by people who need special assistance so the travel time can be cut for those whose health is at risk.

The computer model will allow the researchers to test that concept and see if - and how well -it helps.

"Now, we can simulate any condition we want," Wolshon said.

But there are still issues to overcome, since any computer model is only as good as the information that is given.

"One of the problems is identifying who needs transportation," he said.

Looking at the people who use public transportation helps, but knowing who and how many people will need transportation in a storm and their location is more difficult to find.

"That process is enormously difficult," Wolshon said. "And then that changes from year to year. That's a moving target."

Copyright 2008 Capital City Press