Modelling the flood peril - coupling multi-scale fluid dynamics with weather and rainfall statistics

A proper modelling of flood risk involves the understanding and description of complex physical processes that need to be resolved at the finest scale. The dynamics of rivers has to be investigated for complex geometries involving both river bed variations, natural and urban landscapes. Moreover, these dynamics have to be coupled with input and output conditions corresponding to rainfall and storm surge events in particular. In extremely violent cases, the river beds and surrounding landscape can face rapid and important variations due to erosion processes and transport of different materials. The erosion dynamics of flood protection infrastructures (dams, dykes etc...) must also be taken into account. Specific properties of the soil may complicate the modelling even further: in particular, the drainage capacity of a soil will depend strongly on its location (urban or agricultural) but also on its wetting state (already saturated with water or dry soil for instance), which will influence the flooding process. Finally, modelling the risk of flooding requires a statistical and probabilistic analysis of the range of weather events and their consequences.
The proposed research project aims at developing a finer flood modelling methodology that will be used to generate risk statistics and geographical risk ratings based on captured exposure data. The modelling approach proposed is new as it seeks to develop a dynamical flood model coupling a wide range of physical processes. More precisely, the model will couple statistical models of spatial and temporal rainfall distribution, soil-dependent processes such as run-off, percolation and evaporation, river flows and erosion of flood defenses and tides/storm surges. Instead of creating static flood footprints (most common existing approaches), the goal is to dynamically simulate water precipitation and run-off. A key element of this approach will be the complementary expertise of the participants in developing physical and mathematical fluid dynamics models of shallow-water flows, coupled with external forcing. The academic team will bring its expertise in physical modelling of the different mechanisms and in developing adapted numerical method and codes, the AXA team will support and have depth of experience of dealing with large datasets and undertaking reviews statistical reviews, both for the analysis of weather events and for the probabilistic treatment of the modelling output.

To add or modify information on this page, please contact us at the following address: community.research@axa.com