Incorporating river bed level changes into flood risk modelling

Typically, flood risk assessments (FRAs) are conducted for single N year extreme flow events using a single survey ofthe river channel and floodplains. This approach is fundamentally flawed as it does not account for any changes in theconveyance capacity of the channel that will occur due to sediment transport and morphological adjustment. Therefore,to provide a more robust estimate of future flood risk, the uncertainties related to these changes should be incorporatedinto inundation modelling. This paper proposes a modelling methodology that combines: a stochastic model, forestimating streamflow; a 1D sediment transport model (HEC-RAS), to estimate morphological change; and, a 1D/2Dlinked model (TUFLOW) to estimate inundation. This can be considered as the first quantitative modelling methodologyto account for sediment-related uncertainty in FRA and provides valuable insights into the sensitivity of future flood riskto this variable. The methodology is demonstrated through a conceptual implementation that evaluates the change ininundation for eight flood events with Return Periods (RPs) of 1, 2, 5, 10, 25, 50, 100 and 200 years, along an alluvialriver reach (River Caldew, UK) subjected to 50 years of sediment transport. Results show that, whilst all events exhibitan increase in flooded area and volume, these changes are more pronounced at the lower, more frequent, RPs (160%compared to 9% increase in flood extent for 1 and 200 year RPs respectively).