水災害・防災 (1)

[II-166] Real Time River Stage Forecasting Case Study using Dynamical Systems Theory

*熊井 暖陽¹、奥野 峻也²、戸井 隆¹ (1. 株式会社　構造計画研究所、2. 東京大学生産技術研究所)

Keywords:River Stage Forecasting, Flood Forecasting, medium and small size rivers, Nonlinear time series analysis, Dynamical system, Machine learning, Disaster Prevention

This study was inspired by a concern from the local government on advancing the flood control system. In order to support this concern, we propose a real-time river stage forecast using dynamical systems theory. By using this method, we can easily predict river stage from only the observed river stage and rainfall and forecast rainfall data. The purpose of this study is to verify the accuracy of the prediction method that will be applied on the case study area. The 6 hours ahead prediction results suggested that it is possible to utilize this method for decision making. However, it was revealed that the forecast rainfall error greatly affects the accuracy of river stage prediction.