New strategies for upscaling high-resolution flow and overbank sedimentation models to quantify floodplain sediment storage at the catchment scale

Abstract

Quantitative models of catchment scale floodplain sediment storage must balance competing demands. For example, such approaches must incorporate a physically-based representation of controls on overbank sedimentation rates at the reach scale, but should also be computationally efficient so that catchment scale analysis remains a realistic goal. This paper reports the development of a novel nested modelling strategy that combines a high-resolution hydraulic model based on the shallow water form of the Navier–Stokes equations, with a reduced complexity overbank sedimentation model and coarse-resolution catchment sediment budget model. The approach is implemented within a Monte-Carlo framework to allow an assessment of uncertainty in the parameterisation of overbank sedimentation processes, and to derive uncertainty-bounded estimates of floodplain sedimentation rates over a range of spatial scales. This strategy is applied to a 26 km reach of the River Culm, Devon, UK. The relative performance of a wide range of model structures is evaluated by comparing model predictions with estimates of actual mean annual sedimentation rates derived by analysis of the caesium-137 content of floodplain sediment cores at 20 locations within each of eight study sites distributed throughout the catchment. The results of the current model application demonstrate the potential of the nested modelling strategy as a means of upscaling physically-based flow and sediment transport codes. Furthermore, the novel reduced-complexity overbank sedimentation model presented here is shown to provide a means of simulating complex patterns of suspended sediment transport and deposition, while reducing computing costs by 2–3 orders of magnitude compared with conventional high-resolution advection–diffusion codes. Uncertainty-bounded estimates of floodplain sediment storage for the River Culm confirm that floodplain sedimentation represents a primary component of the fine sediment budget of lowland catchments. Consequently, the modelling strategies developed here may also be of considerable value in attempts to quantify the fate of sediment-associated nutrients and contaminants at the basin scale.