Physically-Based Probabilistic Fluvial Flood Hazard Screening Framework for Civil Infrastructure

Statistical flood frequency analysis yields probabilistic discharge estimates, yet current flood hazard modelling practice rarely translates these distributions into corresponding probabilistic hydraulic parameters. This disconnect limits the consideration of probabilistic information in infrastructure design, where water depth and velocity remain primary decision variables typically considered with several assumptions in practice and in state-of-the-art literature. This study introduces the first physically-based methodology that propagates hydrologic uncertainty into hydraulic flood hazard modelling using site-specific hydraulic geometry response functions. Results show that once uncertainty is propagated, depth and velocity at longer return periods span ranges that cross critical flood risk classification thresholds, and that return period migration, due to uncertainty and non-stationarity, materially increases encounter probabilities relevant to design service life. Application to representative European and North American freeboard assignments shows that reductions in effective freeboard due to uncertainty can substantially increase predicted damages when coupled with standard global depth–damage functions, with the North American cases exhibiting larger sensitivity due to lower design elevations and steeper damage curves. The framework remains compatible with standard flood hazard modelling workflows while offering a computationally efficient pathway for probabilistic flood hazard characterisation. By providing continuous hazard curves, the method supports risk-based screening for infrastructure design, spatial planning, and code calibration across diverse regional contexts, as well as in progressing flood resilience towards advanced performance-based engineering applications matured in other natural hazards.