Influence of testbed characteristics on community resilience using agent-based modeling

Abstract

There has been a large increase in the number of days per year with numerous EF1-EF5 tornadoes. Given the significant damage incurred by tornadoes upon communities, community resilience analyses for tornado-stricken communities have been gaining momentum. As the community resilience analysis aims to guide how to lay out effective hazard mitigation strategies to decrease damage and improve recovery, a comprehensive and accurate approach is necessary. Agent-based modeling, an analysis approach in which different types of agents are created with their properties and behavior clearly defined to simulate the processes of those agents in an external environment, is the most comprehensive and accurate approach so far to conducting community resilience simulations and investigating the decision-making for mitigation and recovery under natural hazards. In this paper, agent-based models (ABMs) are created to simulate the recovery process of a virtual testbed based on the real-world community in Joplin City, MO. The tornado path associated with the real-world tornado event that occurred in May 2011 is adopted in the tornado hazard modeling for the Joplin testbed. In addition, agent-based models are created for another virtual community in the Midwest United States named Centerville using an assumed tornado scenario of the same EF-scale as that in Joplin. The effects of hazard mitigation strategies on the two communities are also explored. A comparison between the analysis results of these two testbeds can indicate the influence of the characteristics of a tornado-prone community on the resilience of the community as well as on the effects of hazard mitigation strategies. It is observed that a community's level of development significantly impacts the tornado resilience. In addition, the effects of a specific type of hazard mitigation strategy on the recovery process are contingent upon testbed characteristics.