Abstract
With the increase of extreme climate events due to the global climate change, urban areas have been becoming more prone to flooding. This trend highlights the importance of effective schemes to mitigate the damage and loss due to inundation. The advancement of numerical prediction method for heavy rainfall is a key technology for early warning. The objectives of this study are twofold. One is to clarify the applicability of the combination of WRF (Weather and Research Forecasting) model and a data assimilation method LETKF (Local Ensemble Transform Kalman Filter). The other is to discuss the evolution of rainfall system that caused heavy rainfall based on the numerical results. We carried out numerical simulation of heavy rainfall in western Shimane prefecture in August, 2013. Comparison between the calculated results and those measured by a radar-AMeDAS shows that the numerical results fail to reproduce the intensity and the location of the heavy rainfall with small success during the latter half of the rainfall event. It is found that the ensemble mean rainfall provides a slightly better results about the peak value and peak time of rainfall compared to those by the deterministic approach without assimilation, showing the effectiveness of LETKF. The heavy rainfall event over Shimane prefecture is considered to have occurred by the southwest moisture flux converge near the front with the orographic effect of the Chugoku distract.
