Where land meets the ocean: compound flood risk and transition zones along U.S. coastal rivers

Transition zones are stretches of rivers where extreme water levels can arise through the synchronous or successive interaction of multiple drivers such as elevated river discharge, storm surge, and high tide. These flooding drivers are often correlated as they can be caused by the same synoptic weather situation. Hence, combining independently derived fluvial and coastal flood maps, as is currently being implemented for example by the U.S. Federal Emergency Management Agency (FEMA) as part of the National Flood Insurance Program, could mischaracterize the true flood risk. An efficient way to overcome this challenge is to use hybrid modelling approaches where hydraulic models are forced by boundary conditions derived from statistical models capable of capturing dependence between the relevant flooding drivers. The hybrid modelling framework can also be complemented by surrogate models which are trained with the input and output of the hydraulic model to limit the number of required simulations. Here we introduce a scalable hybrid framework for compound flood risk assessment and apply it to multiple coastal rivers, including the Suwannee River (FL), Savannah River (GA), Potomac River (MD), and Nueces River (TX). The location and extent of the transition zone is quantified for extreme events with different return periods and compared to relevant catchment characteristics.