Urban traffic is made of a variety of mobility modes that have to be taken into account to explore the impact of catastrophic event. From individual mobility behaviors to macroscopic traffic dynamics, agent-based modeling provides an interesting conceptual framework to study this question. Unfortunately, most proposals in the domain do not provide any simple way to model these multi-modal trajectories, and thus fell short at simulating in a credible way the outcomes of a catastrophic event, like natural or industrial hazards. This paper presents an agent-based framework implemented with the GAMA modeling platform that aims at overcoming this lack. An application of this model for the study of flood crisis in a district of Hanoi (Vietnam) is presented.