Practical appraisal of extra-probabilistic approach to support decision-making under deep uncertainty for future coastal flooding
Résumé
Decision-making in the area of coastal adaptation is facing a major challenge due to the ambiguity in the selection of a unique probability model (deep uncertainty) to represent the lack of knowledge in future sea level rise (SLR). A handful of methods exist to address this problem, but many lack practical recommendations to bring them to an operative state. Our work aims to fulfil this requirement by providing practical recommendations to support decision making under deep uncertainty using the possibility theory, i.e. an extra-probabilistic approach that avoids selecting one unique probability model (CDF) by bounding all the plausible ones consistent with the available data. The framework is applied at a local low-lying coastal French urban area on the Mediterranean coast to assess the uncertainties on the flooding probability by 2100, i.e. SLR uncertainties, the choice in Representative Concentration Pathway scenario, the ranking of high-end scenarios, the regional bias, the vertical ground motion, the contributions of extremes and waves. Our results highlight where the current knowledge prevents the assignment of a unique CDF to future flooding risks. We provide different and complementary information from the probabilistic and possibilistic viewpoints by informing the decision-maker with a single “effective” CDF, which both reflects the impact of ambiguity and his/her attitudes toward risk. If a large ambiguity exists, we ultimately propose to support the setting of learning scenarios by applying the scenario discovery process to search across the set of all plausible probabilistic models with respect to an acceptable flooding probability and of ambiguity.
Origine : Fichiers produits par l'(les) auteur(s)