Speaker: Laurent Marescot
Recent advances in modelling technologies are enabling the industry to manage risk and exposure better. The availability of more data to calibrate models, increasing capabilities to run models at higher resolution and faster, and at the same time at much larger geographic scales, helps the industry to identify and grow new coverage opportunities, address the protection gap and better understand the challenges around changing climate.
Recent advances in fully probabilistic flood modelling in Europe, using the most advanced methodologies and data, have offered the (re)insurance industry its first continental-scale, high-resolution view of flood risk that correctly captures peril correlation, antecedent condition, flood defences, and hours clauses. In addition to the use of growing profitable flood insurance business, this new modelling solutions can be used to assess implications of extreme flood events by combining different realistic situations that were observed in the past.
As an illustration, Germany experienced severe flooding in 2002 characterized by "dry" antecedent conditions and extreme precipitation, and flooding in 2013 characterized by very "wet" antecedent conditions but only severe precipitation (i.e. not extreme, not record breaking). Advanced modelling solutions allow to evaluate the impact of combining record-breaking wetness (like in 2013) with the extreme precipitation similar to 2002, in a fully accurate and realistic way.
This plausible worst-case scenario, that could happen in the future, shows that an increase in flow of ~40% can lead to ~250% increase in flood extent and ~400% increase in overall losses. This shows an important non-linearity in the loss generation process and the need for realistic models to assess the future flood loss impact, including issues with a changing climate.