Author(s): T. Fewtrell; P. Bates; M. Horritt; M. Trigg
Linked Author(s): Paul Bates
Keywords: Hydraulic modelling; Scale; Urban flooding; Time-step
Abstract: Computational cellular based approaches to flood inundatio n modelling have not yet been widely applied to urban environments, where the need for flood manag ement is greatest. The availability of high resolution elevation data obtained from LiDAR, combined with MasterMap R (cid: 13) topological data, has enabled the representation of the small-scale structural elements characteristic of urban areas. Practical application of two dimensional (2D) flood inundat ion models is currently restricted by the computational cost incurred on high resolution grids. In this paper, LISFLOOD-FP, a 2D storage cell model, is applied to a hypothetical flooding scenario in the G reenfields region of Glasgow using relaxed time-stepping and coarse grid resolutions. A large time step, based on analysis of the water height distribution at maximum flood extent, delivers some d egradation in model results compared to the full adaptive time step solution. However, the difference is of the order of the error of LiDAR elevation data. A new grid resampling method where the bare earth and building locations are considered as separate units is presented. Comparison of the new method with standard techniques shows improvement in model results, particularly at coarse resolutions. Combining the relaxed time step and new grid resampling method shows clear scope for relaxing the strict constraints placed on models of urban flooding.
Year: 2007