Sensitivity of hydrodynamic and sediment transport models to uncertainties in bottom sediments characterization.
Résumé
An original method coupling a spherical factor analysis and a series of kriging is applied in the English Channel to the spatial interpolation of a sampled grain size distribution dataset. The number of factorial axes produces various spatial repartitions of bottom sediments numerically degraded with respect to the original data. These surficial sediments characterizations varying by their correlation degrees towards the sampled data are integrated to a hydrodynamic and sediment transport model. The modeling procedure is based on the coupling of the three-dimensional tidal circulation model COHERENS with multicomponent sediment transport modules. The analysis identifies the variability of “key” parameters of (i) hydrodynamic like the bottom roughness, the current amplitude and the bottom shear stress and (ii) sediment transport such as bedload and suspended load transport rates. A low sensitivity of current predictions is obtained. This result contrasts with the noticeable variability of sediment load transport rates exacerbated for the suspension load.