Identifiability of distributed floodplain roughness values in flood extent estimation

In this paper we explore the potential for identifying roughness values for distributed land use types using a comprehensive calibration data set of the 1995 flood event in the River Meuse, including gauged levels, flood extent maps and distributed flood plain level observations. The reach studied is modelled using an integrated 1D\2D hydrodynamic model, with floodplain flow modelled in the 2D domain. Detailed information on floodplain land use types is aggregated to form one, two or five classes of floodplain roughness. Sensitivity analysis of model performance against the calibration data shows that as the number of floodplain classes increases, sensitivity to these roughness values decreases, given allocation of prior roughness values on the basis of constituent land use types and associated roughness values found from literature. Evaluating the identifiability of the roughness in these classes using the Generalised Likelihood Uncertainty Estimation (GLUE) method confirms this insensitivity. As a consequence, application of complex formulae to establish roughness values for changed floodplain land use would seem inappropriate, and evaluation of such changes within a probabilistic framework is suggested.