Date of Award

May 2022

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Atmospheric Science

First Advisor

Paul J Roebber

Committee Members

Rebecca E Morss, Clark Evans, Jon D Kahl

Keywords

agent-based, decision-making, evacuation, hurricane, model, traffic

Abstract

In the mainland US, the hurricane-forecast-evacuation system is uncertain, dynamic, and complex. As a result, it is difficult to know whether to issue warnings, implement evacuation management strategies, or how to make forecasts more useful for evacuations. This dissertation helps address these needs, by holistically exploring the system’s complex dynamics from a new perspective. Specifically, by developing – and using – an empirically informed, agent-based modeling framework called FLEE (Forecasting Laboratory for Exploring the Evacuation-system). The framework represents the key, interwoven elements to hurricane evacuations: the natural hazard (hurricane), the human system (information flow, evacuation decisions), the built environment (road infrastructure), and connections between systems (forecasts and warning information, traffic). The dissertation’s first article describes FLEE’s conceptualization, implementation, and validation, and presents proof-of-concept experiments illustrating its behaviors when key parameters are modified. In the second article, sensitivity analyses are conducted on FLEE to assess how evacuations change with evacuation management strategies and policies (public transportation, contraflow, evacuation order timing), evolving population characteristics (population growth, urbanization), and real and synthetic forecast scenarios impacting the Florida peninsula (Irma, Dorian, rapid-onset version of Irma). The third article begins to explore how forecast elements (e.g., track and intensity) contribute to evacuation success, and whether improved forecast accuracy over time translates to improved evacuations outcomes. In doing so, we demonstrate how coupled natural-human models – including agent-based models –can be a societally-relevant alternative to traditional metrics of forecast accuracy. Lastly, the fourth article contains a brief literature review of inequities in transportation access and their implication on evacuation modeling. Together, the articles demonstrate how modeling frameworks like FLEE are powerful tools capable of studying the hurricane-forecast-evacuation system across many real and hypothetical forecast-population-infrastructure scenarios. The research compliments, and builds-upon empirical work, and supports researchers, practitioners, and policy-makers in hazard risk management, meteorology, and related disciplines, thereby offering the promise of direct applications to mitigate hurricane losses.

Download

Share

COinS