| description abstract | Dike risk management requires breaching parameters that can be estimated rapidly to support the prediction of inundation zones and decision making. The objectives of this paper are to compile a database of dike breaching cases, study common dike failure mechanisms, and develop a set of empirical equations for estimating breaching length, depth, and peak discharge. A database of over 1,000 dike failure cases was collected, with information on pre-breach dike geometry, materials, type of dike, failure mechanisms, breaching length, depth, and peak discharge. A set of regression models are formulated using five control variables: dike height, width, material, type of dike, and failure mechanism. The standard error is set as a selection criterion and the Akaike information criterion is used to optimize the proposed empirical models. The new models are validated using independent cases and compared with available empirical equations for dikes and man-made dams. | |
