Preparing a Small Town for a Hazardous Materials Incident: An Examination of Evacuation Routing Algorithms and Plume Models

Evacuation and shelter in place are two common protective action measures during hazardous events that involve the release of hazardous materials. These responses are complex and require advanced planning to determine their appropriateness to reduce human exposure to hazardous materials and minimize related health risks. Evacuation and shelter in place responses were assessed for people in the town of Erwin, Tennessee, USA, a small, rural town in the mountains of Northeast Tennessee, using a release of uranium hexafluoride (UF6). The population at risk was identified using historical meteorological data and the Radiological Assessment System for Consequence Analysis tool to create plume models for a hypothetical release of UF6 from a nuclear fuel facility that downblends highly enriched uranium. Two hypothetical evacuation scenarios were modeled. One uses the total road network in Erwin and the other involves a train impeding access to an arterial evacuation route. Two routing algorithms available within the custom network analyst routing tool (ArcCASPER) were used for each scenario: 1) a basic shortest path algorithm and 2) a capacity-aware shortest path evacuation routing algorithm. Post-hoc analyses of each scenario and algorithm indicated that the capacity-aware algorithm predicted the quickest evacuation times for both scenarios. Roads with the longest evacuation times and all critical facilities that would benefit from sheltering in place were identified. The study concluded that the capacity-aware algorithm available within ArcCASPER is the most realistic for the town of Erwin.