Estimating the Risk of Flood Severity over the United States

The characterization of flood severity is of overriding importance to floods monitoring, to support the flood hazard mitigation process and to evaluate operational hydrologic models. Currently, flood severity levels (action, minor, moderate, major) are defined by local stakeholders at approximately 3490 gauged locations (comprising 34% of the total) and are unknown over most of the conterminous United States (CONUS). This contribution presents a method for estimating flood stages and their probability distributions at ungauged locations across the CONUS using data the United States Geologic Survey’s stream gauge network of the United States’ Geological Survey (USGS). The innovative approach quantifies the relation between flood stage and geophysical factors including soil characteristics, land cover/land use, geomorphology and precipitation climatology. The probabilistic stages acknowledge the uncertainty arising from the impact-based flood stage estimates and evolving channel characteristics. They are estimated at ungauged locations over the CONUS at 1-km scale through regionalization of model parameters. Probabilistic flood stages provide the basis for flood probability maps and flooding risk assessment. They are implemented in the FLASH distributed hydrologic modeling approach to assess the impending severity of flash flooding. It sets the stage for the characterization of uncertainty in a hydrologic modeling system and probabilistic forecasting framework at ungauged locations. Case studies will be presented.