The interdisciplinary project team--consisting of atmospheric scientists, oceanographers, engineers, and social scientists from the University of Delaware, Cornell University, the State University of New York-Stony Brook, the Renaissance Computing Institute at the University of North Carolina-Chapel Hill, and the University of Oklahoma--have developed a new decision support tool called the Integrated Scenario-based Evacuation (ISE) tool to support hurricane evacuation decision making. Specifically, it is designed to help emergency managers decide when and where to issue official evacuation orders, taking into account the uncertainty in how the hurricane will evolve and how the situation changes over time. The decreased uncertainty in the forecast as the hurricane approaches landfall is implicitly captured by the process. The ISE framework supports adaptive decision-making, providing evacuation order recommendations that offer an advance in the state-of-the-art because they: (1) are based on an integrated hazard assessment including inland flooding, (2) explicitly balance the sometimes competing objectives of minimizing risk and minimizing travel time, (3) offer a solution that is robust under the range of ways the hurricane might evolve, and (4) leverage the substantial value of increasing information (or decreasing degree of uncertainty) over the course of a hurricane event. Input from stakeholders (FEMA, North Carolina Division of Emergency Management, and the American Red Cross) helped guide development of the ISE framework. We have demonstrated the tool for full-scale case studies in North Carolina for Hurricane Isabel (2003), Hurricane Matthew (2016), and Hurricane Florence (2018) and are pursuing funding with a private partner to transition research findings into operations. The project’s goal is to reduce deaths, injuries, and unnecessary expenses associated with ill-planned or ill-executed emergency responses to a hurricane event. A cadre of students--undergraduate, graduate, and post-doctoral – have benefited by being active participants in the research and development of this cutting-edge tool.

There are two key components to the ISE framework. First, a comprehensive hazard assessment is computed using advanced computer models that represent the primary hazard components. These are: (1) the meteorological conditions that represent the hurricane itself, including the surface winds, atmospheric pressure, and precipitation; (2) overland and riverine flooding due to the hurricane precipitation; and (3) storm surge and wind waves driven by the hurricane wind field. Because the models use information from each other, the hazard assessment is integrated, comprehensive and coordinated across the models. That is, storm surge caused by the hurricane winds occurs with the proper timing relative to inland and riverine flooding. This approach of coordinating and integrating the hazard information is different from the typical way of using relatively independent sources of information for each hazard component, which may generally not represent the proper timing in the arrival of strong hurricane winds, peak storm surge, and inland flooding. Uncertainty in the hurricane event is captured by simulating an ensemble of hurricane events that represent uncertainty in how the hurricane is modeled. 

The second key component is the evacuation model, which takes as input the ensemble of hazard scenarios and develops a set of contingency evacuation plan recommendations. Each plan recommends when and where to issue evacuation orders if the hurricane evolves in a particular way. Together they offer plans for all the possible ways the hurricane may evolve. These recommendations are made in a way that maximizes safety for the impacted population and minimizes the total travel times. It incorporates information about the evacuation road network and its carrying capacity, the location of shelters, and the expected behavior of the population. This is different from the traditional approach, wherein only a single plan is recommended and only the time it takes to clear the impacted population from the hazard area is considered (so-called “clearance time”); there are not multiple contingency plans and other critical risk factors are not included.

Last Modified: 09/22/2019
Modified by: Rachel A Davidson