'Route PM': Building a Better Evacuation Plan

A new geographic information system-based software tool, developed under the direction of the Department of Homeland Security (DHS), gives emergency managers an unprecedented ability both to customize evacuation plans for the future and to create new plans as circumstances change.  Beginning with a concept in early 2009, the Real Time Evacuation Planning Model (RtePM, pronounced “Route PM”) was recently created by the Johns Hopkins University Applied Physics Laboratory (APL).

Richard Waddell, RtePM’s Program Manager of the Homeland Protection Business Area in APL’s Asymmetric Operations Department, discussed the start of the project as follows: “We were tasked and funded by DHS’s Science and Technology [S&T] Directorate – the Program Manager is Herb Engle – to find out what the emergency management community in the southeastern United States needed in terms of technology … to help them plan for hurricane response.”

During a conference call between APL and a State Emergency Manager Focus Group comprising representatives from 11 states, five primary needs were identified for evacuation planners. The number one need, they all agreed, sounded simple: “Give us a way to draw a polygon on a map, push a button, and get an evacuation clearance time for that area.”

Although state emergency planners already knew several ways to generate the information needed, in many cases those “ways” were based on census and infrastructure data anywhere from five to ten years old. Information that is not current, though, is not useful, either, because most states, particularly those in the Southeastern area of the country, have experienced major population and infrastructure changes during the past decade. There was no way, therefore, said APL’s Russell Strickland (APL’s Project Lead for RtePM, and former Deputy Director of the Maryland Emergency Management Agency) “to change the parameters and develop a new clearance time estimate.

“Our conversation with users,” he added, “also showed us the need to let them allow for seasonal populations, such as during summer beach seasons, and for single-event population increases, like racing at the Talladega Superspeedway [in Alabama].”

Three Drivers, an Engine, an Interface, and a Prototype

There are three fundamental “drivers” in the art and science of evacuation planning, Strickland also explained. “At the top end, there are the mandatory or regulation-required plans, such as for nuclear power plants, chemical releases, and dams. Next, plans for hurricanes – which may almost be considered mandatory in hurricane-prone states. Finally, there are those plans that a jurisdiction decides are particularly important for its region.” Included in the latter group, he continued, are incidents such as wildland fires – ” … which in the United States are the most frequent reasons for evacuations of 1,000 people or more. And there are the daily evacuations for hazardous materials releases that occur anywhere in the country and are our most frequent cause for evacuations.”

Creating a tool that could prove useful for all three of these fundamental planning needs was not the original goal of the DHS request. However, as work progressed, the RtePM team realized they could create both a simulation engine and a user interface that would allow planners to handle almost any type of evacuation imaginable.

After a prototype had been developed, Strickland said, “We did initial field testing in Mobile [Alabama] and got great feedback. They really understood what we were trying to do, and they provided outstanding guidance to make sure it [the prototype] actually did those things. That’s also when DHS told us, ‘Listen to the people. If they want something in there, put it in there.’” Strickland said that the team has already presented interactive demonstrations of RtePM for representatives from all 50 states, and has incorporated their individual and collective feedback to further improve the program.

“We view RtePM as a critical component of the evacuation planning and crisis response toolset that DHS S&T will start to transition to local, state, and federal users over the next year,” said DHS’s Joseph Kielman, Chief Scientist for Disaster Management and Chief of the Visual Analytics Technologies Branch in the department’s Infrastructure Protection and Disaster Management Division. “A common suite of integrated tools usable on multiple levels – that is, on a smaller scale for individual buildings, ranging to a medium scale for large sports or entertainment venues, and ultimately to the large scale required for cities and even multi-state regions – is one of the objectives [that were established] for this work.”

The Heart of the Program: How Many People & How Fast?

At the heart of RtePM is what is called a dynamic clearance time calculator, which uses two sets of data to estimate the length of time needed to evacuate/clear a specific geographic area. By combining data on roadway capacity with demographic information, and providing an easy-to-use graphical interface to set various parameters for certain aspects of human behavior, RtePM not only offers considerable flexibility but also requires little user training.

An individual user can, in fact, generate new simulations simply by drawing a line around an area, selecting from such variables as side streets and specific neighborhoods, and re-running the dynamic clearance time calculator. Additional improvements in estimating the clearance time have been achieved in the newest version of RtePM through the incorporation of daytime population data sets, thanks primarily to use of the LandScan High Resolution Global Population DataSet developed by the Department of Energy’s Oak Ridge National Laboratory.

RtePM is able to run an evacuation simulation for a relatively small geographic area in the time it takes to refresh the screen; simulation of a larger – i.e., densely populated – area takes up to two hours. “We ran a large scenario for a Category 4 hurricane in the Houston/Galveston, Texas area,” Waddell pointed out, “with a population of 1.6 million, and 922,000 of them evacuating, in a bit under two hours.” Similarly impressive results were obtained for the Hampton Roads region of Virginia. (For the nation’s largest metropolitan areas, Waddell says, the team would have to include models of public transportation and pedestrian evacuations.)

Listening to the Users – And the Biggest Challenge Ahead

Hearing and meeting the needs of numerous evacuation planning professionals – and incorporating into the program their individual and collective knowledge of regional differences, expectations, and experiences – has helped create a set of tools and options in RtePM that reflect real-life behaviors. “In speaking to hurricane planners in Florida,” Strickland says, “we learned that if a hurricane is heading for Miami, many people don’t head north. Instead, they head south, down to the Keys, to retrieve their boats, put them on trailers, and only then [do they] drive north. That creates much longer vehicles that move much more slowly. Sitting here in our ivory tower, we would have never thought about that possibility.”

Thousands of miles away, in the West and Southwest, there is a similar behavior pattern, Strickland pointed out: “Working with the National Wildfire Coordinating Group, we learned that, during wildland fires in Western states, people will not leave without their horses, which means [they need] longer, slower vehicles. Because of this feedback, we added a toggle button to the screen that allows the user to adjust for greater-than-normal vehicle length, which changes the algorithm’s calculation of evacuation times.”

“We also learned that we can’t always leave off small access roads in our mapping,” Waddell added. “On the populated and developed coasts, we can do that, but in rural areas, all the roads are potentially critical evacuation routes, so we have to use the data available for all the roads [in any given area of the country].”

“Even North Carolina’s Route 12, which is the only road that connects the Outer Banks towns, doesn’t show up on mapping unless you go deep into the data,” Strickland commented.

The RtePM team is still working on ways to more effectively address what is perhaps the most unpredictable variable: human behavior. “What will people’s evacuation behavior be? It’s our biggest challenge,” Strickland said. “There is no absolute.” Human behavior is, of course, an important issue that has always affected the nation’s emergency planning community – but has only recently led to serious academic research. The team uses survey information from different areas of the country to develop informed judgments about different types of evacuation events (immediate and planned). “We look for the numbers of people who say they will evacuate,” Strickland continued, “and when they will evacuate, to create our curves for volume and capacities.”

“A lot of data needs to be collected that is not [presently] being collected,” Waddell said. “But,” he immediately added, “There is a good reason for that: Data collection in the middle of a disaster evacuation is understandably not a high priority.” In fact, the team has designed RtePM to help solve its own problems by, among other actions, encouraging users to help create larger and more relevant databases. “Users understand that, if they use RtePM to build an evacuation scenario, they would want to collect data to see how well it worked.”

Planning for the Future – Transition Version in Three Months

Although the system is not yet fully real-time, owing in part to a dearth of real-time traffic monitoring data, RtePM is already able to generate new simulations quickly enough to be very effective in most cases, and has drawn praise from evacuation analysts and government agencies for its current capabilities. “It gives us confidence in the model,” Waddell says. “DHS wants to get a planning tool into the hands of planners that they can use day-to-day,” Strickland adds. “For now, that is more important than having real-time data.”

Requests for RtePM information and trial use have also come from government, academic, and private agencies and facilities across the United States. RtePM is still in development for DHS, but a target delivery date of April 2012 has been set for a transition version that will be able to make the jump to a real-world tool for jurisdictions ranging in size from large metropolitan regions to rural counties.

“The idea is for this to be affordable,” says Waddell. “It’s all done with open source software, it’s web-based, and it uses road-network data sets that local agencies can access free, thanks to the DHS initiatives. FEMA [DHS’s Federal Emergency Management Agency] and the U.S. Army Corps of Engineers have worked closely with us during RtePM development and testing. Hurricane evacuation planners update their plans every five years or more based on data provided by the Corps; with our tool, they could do it monthly, and for less cost.”

Geoff Brown

Geoff Brown is a science writer/public affairs officer at the Johns Hopkins University Applied Physics Laboratory (APL) in Laurel, Maryland. He provides communications support for APL projects ranging from national security programs to space exploration missions. Previously, he wrote about science and technology for a variety of publications as a freelance journalist; was executive producer for a live public radio talk show; and served as managing editor for Baltimore magazine.



No tags to display


Translate »