A framework for risk assessment of radiological dispersion device (RDD) events: Integrating physical dispersion and behavioral response models

A radiological dispersion device (RDD) or "dirty" bomb is a conventional explosive wrapped in radiological material. Terrorists may use an RDD to disperse radioactive material across a populated area, causing casualties and/or economic damage. Current RDD risk assessment models do not account for behavioral factors affecting risk. In this dissertation, we develop a behaviorally realistic risk assessment for RDD events. The methodology was developed using a comprehensive influence diagram, combining (1) physical dispersion variables, (2) behavioral response variables, (3) official detection, communication, and response variables, and (4) public health variables. We develop a plausible RDD scenario using the influence diagram. A physical dispersion and behavioral response model is applied to generate risk assessments for the RDD scenario. The physical dispersion and behavioral response model predicts trauma fatalities from the conventional explosive and long-term cancer fatalities from radiation exposure during sheltering and evacuating. Public behavior is treated parametrically in the model. The RDD scenario was also integrated into an elicitation procedure to understand how the public might behave in response to official orders to evacuate or shelter-in-place. Open-ended interviews were conducted with academic experts. Results from the interviews were used to develop a confirmatory questionnaire conducted with first responders in Pittsburgh, PA. Data collected from the confirmatory questionnaires were used to develop realistic distributions of public behavior during an RDD event. Elicited behavioral distributions alone suggest that resources should be invested to improve public compliance with blanket recommendations. We evaluate the costs of several policies for improving public compliance. By integrating these behavioral distributions with RDD risk assessments, we find that partial compliance reduces societal risk significantly more than perfect compliance. We suggest that resources should focus on studying communications and compliance for partial evacuations. A theoretical decision analysis concept is introduced known as the value of communication. Using the value of communication, the impact of various risk communications for emergency management of RDD events are evaluated and compared. The approach described in this thesis is a novel and effective approach for incorporating behavioral assumptions into risk assessments. It is also an effective structured approach for building an understanding of how the public will behave during an emergency event.