A PREDICTION METHOD FOR BANK EROSION WIDTH CONSIDERING MULTIPLE CHANNEL CONDITIONS

Abstract

 This paper presents a screening method for predicting flood-induced bank erosion around bridges. In August 2016, four typhoons hit Hokkaido and a number of bridges were damaged. For efficiently extracting vulnerable bridges, multiple regression analysis is applied to predict the erosion width of the channel using stream parameters such as basin area, bed-slope, channel curvature, precipitation and soil properties. The results show that four geological conditions vary considerably in erosion width, indicating that the parameters are unsuitable for screening. However, without soil properties of targeted sections, predicted normalized erosion width shows good agreement with the measured one when applied to a wide range of basin area. This method can be an effective measure to efficiently extract the potentially dangerous bridges.