Sound flood risk management decision making is underpinned by flood risk analysis. Current methods applied at regional and local scales are often limited in their consideration of the potential for defences to fail. Ultimately this can lead to underestimates of the true risk and subsequent difficulties in justifying mitigation measures such as maintenance and replacement of defences. A methodology has been developed for assessing flood risk arising from fluvial and coastal sources that explicitly considers defence failures represented through fragility curves. This method requires consideration of flooding scenarios involving multiple defence section failures and flood events ranging in severity. It has therefore been necessary to develop a purpose-specific flood spreading method that is capable of simulating many flood events in practical timescales. The method has been applied to the Thames Estuary, where outputs including spatial maps of flood risk and defences attributed with residual risk have been used to support decisions relating to strategic flood risk management over the coming century.

Flood risk analysis increasingly involves the integration of a full range of loading conditions as well as multiple defence system states, overlaid by uncertainty analysis. This type of analysis involves the simulation of many thousands of flood events. To keep model runtimes to practical levels an efficient yet robust flood inundation model is required. To accommodate this need a rapid flood spreading model (RFSM) has been developed that utilises the availability of good quality topography data and advanced GIS techniques. This paper describes recent improvements to the RFSM that have focused on incorporating additional physical processes within the spreading algorithm (multiple spilling and friction). This improved model is applied to a number of different sites with comparisons made to a more complex hydrodynamic model. The findings of this comparison demonstrate a good degree of similarity between the RFSM and more complex models, with a significantly reduced runtime overhead.

Detailed testing of a new two‐dimensional hydraulic modelling system is presented. The methodology consists of applying the hydraulic model to a set of theoretical tests, for which analytical solutions are known, and then comparing a model simulation with a real flood event. The water‐at‐rest test, three dam‐break tests and the Seiche test constitute the set of tests with analytic solutions. The flood event is the Boscastle 2004 flood, for which observed water levels are available. The model yields results that compare closely with the analytical solutions in all of the tests, except for the Seiche test. When applied to the real flood event, the simulated maximum levels are close to the observed levels. Comparisons with two similar commercial software packages show similar results between the models.

A significant amount of effort is being directed towards finding ways to improve the management of urban drainage to reduce flooding. It is recognized that an integrated modelling approach is required, linking rivers and their floodplains with surface water and foul drainage systems. As a result of this, flood inundation modelling methods are starting to converge. The existing 1D InfoWorks software has recently been enhanced to include a 2D flood modelling capability. This 2D hydrodynamic modelling software incorporates links with the already existing 1D software for rivers (InfoWorks RS) and network systems (InfoWorks CS). This paper describes the results of a series of analytical tests used to validate the robustness of the new 2D modelling engine, results of its application to a real case study and, in some cases, comparisons of the results with other 2D flood models.

No abstract