Obtaining Reasonable Breach Hydrographs Based on Breach Parameter Estimates from Regression Methods

摘要

Dam breach simulation models, such as HEC-RAS, are widely used to calculate dam breach hydrographs using breach parameter estimates from regression methods. The resulting information is used in dam failure emergency action plans and for estimating potential life loss or economic consequences of dam failure for use in risk assessments. In this paper we address several important and often overlooked practical issues in obtaining reasonable breach hydrographs using breach parameter estimates based on regression methods. These include addressing inconsistences between peak breach flow, reservoir volume and breach formation time estimates and the tendency for peak flow estimates from HEC-RAS to be significantly overestimated. We also discuss a fairly recently developed regression methodology by Xu and Zhang (2009), which has the significant advantage of being the first approach to incorporate erodibility as a predictor variable. We address several questions that have been raised about this approach and provide a revised version of approach that addresses these questions. The paper is illustrated with data from case histories and dam breach simulations and a comparison of the original and revised Xu and Zhang equations for the 1976 Teton Dam failure.