DEVELOPMENT OF FLOOD FORECASTING SYSTEM APPLIED DATA ASSIMILATION TECHNIQUE TO THE OBSERVATION WATER LEVEL IN THE REACH INCLUDING THE JUNCTION OF RIVER YODO

Abstract

This study is to build the flood forecasting system comprised of a distributed runoff model and a one dimensional hydrodynamic model for the whole Yodogawa water system using the data assimilation technique for the real-time observation water level by the particle filter. We built data assimilation technique to let observation water level perform a fitting of the forecast of one dimensional hydrodynamic model and inspected the compatibility with the results flood in the model. Results of having judged prediction data with Nash values and RMSE show extremely precise.