Mass Evacuation of Halifax, Canada: A Dynamic Traffic Microsimulation Modeling Approach

This study develops a dynamic traffic microsimulation model for testing a mass evacuation of the Halifax Peninsula, Canada. The model is developed utilizing multiple data sources and calibrated using a Latin Hypercube (LH) sampling technique. The proposed microsimulation modeling framework adopts a Dynamic Traffic Assignment process to capture temporal variations in travel time and traffic congestion propagation during evacuation in the Halifax transport network. The study demonstrates the efficacy of the model in predicting traffic flows for evacuation due to floods of water level 2.9m, 3.9m, and 7.9m. Network level impacts as well as individuals’ evacuation experiences are examined utilizing the developed model. The results suggest that it takes 22 hours to evacuate the Peninsula in the case of a flood of water level 2.9m. In case of all scenarios, it takes 6-8 hours to complete 50% evacuation, whereas the rest is evacuated in 14-16 hours depending on the extent of the road damages caused by the floods. Individuals departing early experience higher network and local level traffic congestion during evacuation. The model results will be useful for policy makers to develop evacuation plans and countermeasures.