Hydrologic scaling for hydrogeomorphic floodplain mapping: Insights into human-induced floodplain disconnectivity

Hydrogeomorphic approaches for floodplain modelling are valuable tools for water resource and flood hazard management and mapping, especially as the global availability and accuracy of terrain data increases. Digital terrain models implicitly contain information about floodplain landscape morphology that was produced by hydrologic processes over long time periods, as well as recent anthropogenic modifications to floodplain features and processes. The increased availability of terrain data and distributed hydrologic datasets provide an opportunity to develop hydrogeomorphic floodplain delineation models that can quickly be applied at large spatial scales. This research investigates the performance of a hydrogeomorphic floodplain model in two large urbanized and gauged river basins in the United States, the Susquehanna and the Wabash basins. The models were calibrated by a hydrologic data scaling technique, implemented through regression analyses of USGS peak flow data to estimate floodplain flow levels across multiple spatial scales. Floodplain model performance was assessed through comparison with 100-year Federal Emergency Management Agency flood hazard maps. Results show that the hydrogeomorphic floodplain maps are generally consistent with standard flood maps, even when significantly and systematically varying scaling parameters within physically feasible ranges, with major differences that are likely due to infrastructure (levees, bridges, etc.) in highly urbanized areas and other locations where the geomorphic signature of fluvial processes has been altered. This study demonstrates the value of geomorphic information for large-scale floodplain mapping and the potential use of hydrogeomorphic models for evaluating human-made impacts to floodplain ecosystems and patterns of disconnectivity in urbanized catchments.