A method of flood severity assessment for predicting local flood hazards in small ungauged catchments

A flash flood caused by heavy storm in a short period of time is now one of the common natural disasters worldwide. The aim of this paper is to develop a new flood hazard index for use in local flood severity predictions in small ungauged catchments through the regression analysis between the flood hazard index and rainfall features. The flood events with the peak flow above a threshold discharge are targeted to estimate the flood hazard index among the annual maximum flood runoff hydrographs generated from a rainfall-runoff model for the long-term-observed rainfall data. To quantify characteristics of flood runoff hydrographs simulated for ungauged catchments, the new flood hazard index is measured by the average of the three normalized relative severity factors, such as the flood magnitude ratio, the rising curve gradient, and the rising average runoff to the peak time. The proposed flood hazard index shows a high correlation with logarithm of the short-duration rainfall for the two selected small ungauged catchments in the Korean Peninsula. The best-fit regression equation between the flood hazard index and rainfall patterns can be used in predicting the relative flash flood severity for small ungauged catchments.