[A-AS05] スーパーコンピュータを用いた気象・気候・環境科学

[AAS05-05] Flood forecasting using 1000-member ensemble prediction for a severe flood event

*大泉 伝^1,2、川畑 拓矢²、Le Duc^3,2、小林 健一郎^4,2、斉藤 和雄^5,2,1、太田 琢磨⁶ (1.一般社団法人気象業務支援センター、2.気象研究所、3.東京大学、4.神戸大学、5.東京大学大気海洋研究所、6.気象庁)

キーワード：洪水予測、アンサンブル気象予報、洪水災害

This study is the first study realizing “Impact-based forecasts and warnings” for severe flooding. “Impact-based forecasts and warnings” is a new concept proposed by WMO that predicts not only extreme weather conditions but also damage levels of disasters. We developed an ensemble flood forecast system that predicts the probability of flooding nationwide in three damage levels. The feature of this system is that it uses an operational flood forecast model "Rainfall Index Model" in Japan in combination with a large ensemble weather forecast. This study investigated the impacts of a large ensemble size (1000) on the probability of occurrence of flooding for the different sizes of river basins. The system predicted the maximum probability of exceeding the historical magnitude flood of 60% 11 hours before the flood event.