Risk-based system for coastal flooding early warning – RISCOPE

Most of the existing forecast and early warning systems rely on the prediction of hydrodynamic conditions at the coast, or exceptionally on flood computations inland, but under deep simplifications. Recent scientific progresses now allow properly modeling coastal flooding events. Such models are nevertheless very expensive in terms of computation time (>hours) which prevents any use for forecast and warning or even for estimating the coastal flood hazard return period together with uncertainties. These challenges raise several questions: “How to improve coastal flooding early warning systems?”, “How to improve coastal flooding assessment and ensure taking into account all possible events?”, “How to deal with uncertainties and raise awareness on climate change effects?”.

RISCOPE aims at addressing part of these challenges by developing a risk-based method contributing to forecast, early warning and prevention of coastal flooding risks. In this bottom-up approach, the starting point is the key information useful for decision making (e.g. water level on a strategic asset). This allows to produce targeted warnings but also to identify all the scenarios that may lead the flood to exceed given thresholds. The method should be robust, fast and integrate the complexity of coastal flood processes (e.g. overtopping, interaction with structures) and cascade effect (e.g. coastal defense failure). RISCOPE should thus constitute a breakthrough alternative to existing systems and methods.

One of the challenges is to develop fast and robust coastal flooding models accounting for processes complexity and cascade effect. The solution explored in RISCOPE relies on meta-models, i.e. mathematical functions which estimate with good precision and at a negligible computational cost (<minutes) the results which could be obtained with the numerical model. However, the use of meta-models for coastal flooding applications raises a number of challenges: manipulation of complex spatio-temporal data, non-linear and shocks processes, etc. They will be addressed during the project. An early-warning oriented demonstrator will be set up using the developed meta-models and local data (offshore conditions, GIS data, etc.). The user interface will be designed with decision makers. It will be used in a forecast/warning mode (few days), but also in a scenario mode. This scenario mode will be used to raise risk-awareness (simulations performed during workshops) and to estimate coastal flooding hazard return periods and associated uncertainties, under present and future climate. To develop the method and the demonstrator, the site of Gâvres (Morbihan) has been selected as it is representative of a significant part of French mainland coasts in terms of variety and complexity of flooding processes, as well as available offshore data.<br />
RISCOPE will call on several scientific domains and will require interactions with decision makers. It is thus based on (1) a multidisciplinary team (coastal flooding, statisticians, numerical development, GIS and Decision Support System experts, climate change impact, etc.); (2) a user group gathering the local decision makers (e.g. Gâvres mayor and Lorient Agglomeration, who agreed to take part to this group); (3) an advisory committee composed of researchers, national operators and risk prevention actors (e.g. DGPR, Meteo-France). In addition to scientific papers and reports, the main expected outcomes are: (1) a better knowledge of local risks, (2) a new method for prevention, warning and estimation of coastal flooding; (3) a demonstrator at the site of development; (4) guidelines for the implementation of the method to other sites and (5) open-source mathematical advances in meta-modeling which can be useful in other domains (natural risks, industry, civil engineering, structural engineering, etc.).