Abstract
The hydroelectric power dams, which don't have the function of the flood control, become more important for the flood control in Japan, because many abnormal floods occurred recently. In order to reduce the flood discharge of the hydroelectric power dams at Kumano river basin, GSM(Japan Meteorological Agency's Global Spectral Model) is applied to the dam operation, but the precipitation of GSM at the basin is underestimated frequently. So, in order to improve the underestimation of GSM at the basin, a new method of correcting the underestimation by using two indexes is proposed and evaluated for the dam operation in this study. The one index is the spatial variation of GSM, another is the variation of the ensemble rain forecast of JMA.