Natural disasters could result in unnecessary loss of life and disproportionate suffering to families and communities if evacuation plans are not in place or understood by the public. In the Pacific Northwest, a magnitude 9.0 earthquake and tsunami from the Cascadia Subduction Zone (CSZ) represents one of the most pressing natural disasters with an astonishing high 7%-12% chance of occurrence by 2060. The destructive nature of earthquakes and the subsequent near-field tsunami, and also the retrofitting challenges of infrastructure network motivates us to accurately model the tsunami evacuation to reduce the number of fatalities. This thesis presents an agent-based multi-modal near-field tsunami evacuation modeling framework in Netlogo.
The goals of this study are two folded. The first objective is to investigate how (1) decision time, (2) choice of modes of transportation (i.e., walking and automobile), and in general (3) different variables involved in the evacuation scenario (e.g., walking speed and driving speed) impact the estimation of casualties. Using the city of Seaside, Oregon as a study site, which is one of the most vulnerable cities on the Oregon coast, different evacuation scenarios are included in the model to assess the impact of parameters involved on the mortality rate of the tsunami event. The results show that (1) evacuation mode choice strongly influences the expected number of casualties; (2) the mortality rate is strongly correlated with decision-making time (τ ); and (3) the mortality rate is sensitive to the variations in walking speed of the evacuee population.
Secondly, this study extends the agent-based modeling framework to assess the transportation network vulnerability in Seaside, OR, under unplanned disruptions (i.e., bridge failures) due to the Cascadia Subduction Zone earthquake initiating a near-field tsunami on the coast of State of Oregon. The criticality of each link in the entire network is evaluated iteratively by connecting the impacts of link failures on the resultant mortality rate. After assessing all the links, an innovative method is developed to identify the most critical links within the network. Further assessment is conducted on the identified critical links to formulate an optimal network retrofitting plan to minimize the mortality rate considering the limited amount of resources. The framework has been tested on the transportation network of city of Seaside, OR, and the results show that the critical links are not necessarily the bridges in the network. Therefore, the identification of the critical links requires a systematic assessment of the entire transportation network, and furthermore, minimizing the mortality rate necessitates the logical use of available resources.