Feedbacks between river morphodynamics and overbank flooding

Abstract

River morphodynamics (the way rivers adjust their shape and form) can affect overbank flooding if changes to cross-sectional area or roughness reduce the flow conveyance capacity of the channel. Extreme floods can also cause drastic adjustments to river morphology on relatively short timescales. The potential simultaneity of these processes raises the question: how do flood dynamics and river morphology co-evolve? Additionally, are river morphodynamics a substantial contributor to flood hazards? Existing research on channel adjustment and recovery during a flood hydrograph has been limited by a paucity of field data recording river bathymetry during peakflows. Additionally, the computational expense of morphodynamic modeling has challenged the application of models to investigate feedbacks between morphodynamic river adjustments and flood hazard. This thesis seeks to fill the gaps in scientific knowledge about feedbacks between morphodynamics in flooding by (I) examining the importance of river morphodynamics in modifying flood hazards; (II) explaining causes of spatial variation in morphodynamic response to floods; and (III) investigating interactions between flood hydraulics and river morphodynamics during a flood event in a location with persistent overtopping during peak flows. Part I investigates long-term changes to river channel conveyance in northwest Washington State, U.S., and compares the relative importance of these to shifts in moderate flood streamflow using 50 long term river gaging datasets. River conveyance is unsteady in most rivers in western Washington, but its importance for modifying flood risk depends on the style of conveyance response. Conveyance responses can be linear, oscillating, punctuated by sudden sediment-supply events or influenced by flow regulation. The relative behavior and importance of conveyance and streamflow variability depends on flow regulation; moderate flood streamflows have increased in unregulated rivers, but this trend is suppressed and/or reversed in regulated rivers. Flow regulation does not necessarily reduce total flood hazard, because downstream channel conveyance losses can exceed reductions in flood flows. Part I concludes that channel conveyance unsteadiness is an important modifier of total flood hazard variability and incorporating this factor may improve flood risk predictions. Part II explains the spatial variation of river bed response during floods and recovery during low flows in relation to floodplain geometry and topography using 20 years of bi-weekly bathymetry data from the Waal River, an engineered channel in the Netherlands. Bed response and recovery to peak flows are a result of floodplain geometry and topography that forces flood flows to enter or exit the main-channel including: 1) streamwise changes in floodplain width, and 2) topographic routing of floodplain flow that causes flow to plunge into and out of the main-channel during floods. The latter factor is primarily important for spatial variation in main-channel bed response and recovery to floods. Where floodplain flows enter the main-channel, erosion (deposition) occurs during floods (low flows) and where floodplain flows exit the main-channel, deposition (erosion) occurs during floods (low flows). Part II concludes that bed elevation within and between flood events may be understood by floodplain topography and geometry that results in flow exchange between the main channel and floodplain. Part III applies a hydro-morphodynamic model to analyze hydraulic and morphodynamic feedbacks during the November 2021 flood in the Nooksack River, Washington State–a system with persistent overbank flooding during peak flows at an historic avulsion node. Cross-sectionally averaged patterns of modeled bed elevation change suggest that meter-scale adjustments in bed elevation occurred during the flood. Bed elevation changes in the main-channel are commonly located where flow spatially accelerates or decelerates during the flood; in particular, flow deceleration is observed at locations of overbank flooding. At the historic overflow location, bed deposition co-occurs with abrupt flow deceleration related to overbank flooding. Part III demonstrates that the hydrodynamics of overbank flooding can cause channel adjustments that affect channel conveyance capacity for future floods. These results augment our understanding of the importance of river morphodynamics in modifying overbank flooding and the feedbacks between overbank flooding and river bed response. The thesis concludes with a suite of possible questions and future research directions relating to: 1) the preferential style of morphodynamic adjustment to peak flows, 2) river bed response to climate variability, 3) feedbacks between floods, sediment supply, and channel response, and 4) compound channel structure as it relates to morphodynamic response to floods. Overall, this work furthers our understanding of fluvial response to floods with implications for improving flood safety of river-bounding communities.