Recent tropical storms that have resulted in flood events with large economic impacts in the Northeastern US have catalyzed efforts to understand the complex interactions between human and natural systems. Specifically, the resilience of our transportation infrastructure to climate and the impact of our transportation systems on the aquatic environment are of significant interest on both the local and regional scales to state and federal agencies. It is important that new, innovative approaches be developed that consider both the robustness of our infrastructure today and its ability to cope with forecasted extremes due to climate change. Because few streams are gauged to record their flows, road-stream crossings are almost always designed without adequate knowledge of what floods flows that will occur in the future. Hydrologic predictions at these ungauged locations are difficult and few techniques exist that can accurately estimate extreme flows without long-term precipitation and streamflows records. Robust analysis using either statistical or physically based models within a multi-model framework is a useful approach in quantifying the degree of uncertainty and variability across models at these ungauged, road-stream crossing locations. This paper reviews background information associated with the development of a hydrologic vulnerability protocol system for road-stream crossings currently being piloted in the Deerfield River basin, focused primarily on predicting flows in ungauged basins and estimating flood flows for current and future climate scenarios. Finally, a framework for a decision support tool is discussed within the context of providing hydrologic flood predictions at road-stream crossings within an online interactive map interface.