Improving flood hazard prediction models

Errors associated with inaccurate components of hydrodynamic flood hazard prediction models can misrepresent reality, particularly when a model is run for conditions extrapolated outside of a calibrated range. Hydrograph exceedance probabilities, LiDAR errors, roughness mapping, bathymetry and flow resistance are discussed. There is particular uncertainty surrounding the treatment of flow resistance. Typically, 2D numerical flood models have low resolution mapping of roughness compared to mapping of topography. Formulation of friction within the hydrodynamic model code and the derivation of depth-averaged velocity equations are investigated. Significant areas with high roughness and low flow depth can occur in flood plain models yet conventional flow resistance equations break down under these conditions. The paper gives recommendations for improving mapping of roughness, new equations for better representation of flow resistance in the modelling code and a chart for converting the more common "n" values to "Z(o)" roughness values.