We describe progress towards a coupled human-natural system model that integrates hydrologic modeling of flooding hazard with an agent-based model (ABM) that accounts for economic, land use, and critical infrastructure expansion decisions. The development of such a model attempts to integrate the key physical, demographic, economic, and institutional drivers of coastal development for real-world coastal systems. Utilizing an initial prototype of the model, we conduct stylized experiments interrogating the relative influence of structural versus parametric uncertainty in the representation of residential location choice on urban development trajectories in flood-prone housing markets. The stylized model is applied to the Baltimore landscape, initialized using baseline population and demographic data from the U.S. Census and home price data from Zillow’s ZTRAX database. Future efforts will involve enhancement of the model to conduct a retrospective analysis of urban development patterns in Baltimore from the 1950s through present and to simulate future trajectories of 21^st century urban development patterns under scenarios of climate-driven flood hazard.