Séminaire Megumi Watanabe (université de Tokyo, Japon)

Résumé

Ensemble flood forecasting provides probabilistic information with a long lead time. They are expected to be used for early disaster warning. However, the know-how to effectively utilize ensemble flood forecast information and link it to disaster prevention actions has yet to be established. This is because ensemble flood forecasting has different characteristics from conventional deterministic flood forecasting. Local government officials and others responsible for disaster prevention actions during a flood may need help interpreting ensemble flood forecast results. In this study, we examine what actions Japanese municipal officials can take to counter floods based on ensemble flood forecasting. We also examine effective methods of visualizing the forecast results for municipal officials. 26 municipal officials from 13 municipalities in the Iwaki River basin that experienced flooding in 2022 were interviewed. More than one in four municipal officials can act based on medium- to high-reliability (33-100%) forecasts. Early warnings with medium confidence (33% or higher) at least 12 hours before flooding can increase the types of measures (supplies, transportation, etc.) that can be taken to reduce flood damage. Probabilistic forecasts may be acceptable if limited to an appropriate amount of information. Further research is needed to understand the influence of municipal officials' flood experience on their forecast preferences (Watanabe et al., in prep).

Pour en savoir plus sur les travaux de Megumi Watanabe : Google Scholar