Doctor of Philosophy (PhD)


Geography and Anthropology

Document Type



The empirical study of business responses to disasters is relatively scarce, despite that they are a fundamental part of the cities, providing services, jobs and, taxes that are essential for urban sustainability. This dissertation develops an agent-based simulation model to represent and understand the businesses reopening process in a dynamic environment in New Orleans after Hurricane Katrina. The objectives are two-fold: 1) To identify the main reopening predictors involved and estimate their relative importance through time, using an empirical data set collected from another study; 2) To represent the business reopening process through a computer simulation model, using the parameters derived from the first objective. The results show that businesses located in flooded areas had lower reopening probabilities, however the effect was significant only in the first nine months after the disaster. Larger businesses had better reopening probabilities than smaller ones, although this variable stopped being significant after six months. Variables associated with higher social vulnerability, such as percent non-white population and percent population under 18, had a negative effect on the business reopening probabilities at different points of time. The influence of neighboring firms using 1-km buffer was found significantly positive only immediately after the disaster; it became significantly negative one year after the disaster. The simulation model developed proved to mimic the reopening process at a suitable level. The model was used to simulate two scenarios: 1) First, the flood depth was reduced by 1 meter as a way to represent the implementation of measures designed to increase the buildings and infrastructure resistance to floods. The simulation results indicate that there are specific areas that would obtain greater benefit from these measures, however ten months after the disaster the effect of the measures tends to diminish. 2) Second, the spatial effects of aids were simulated by making a limited number of businesses in specific locations totally resilient to the disaster. The results indicate that the beneficial effect is influenced by variables such as business density and socio-economic conditions of the area. The positive effect is perceivable until four months after the disaster, after this point it diminishes.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Lam, Nina S.