Abstract
The verification of a dam-break flood-routing model was conducted by using the 1960-1961 laboratory data from the U S Army Engineer Waterways Experiment Station. A series of measured stage, velocity, and discharge hydrographs at selected stations upstream and downstream of a breached dam were compared to computed results under various test conditions, such as instantaneous full or partial breaches, with or without an initial flow, and on hydraulically smooth or rough surfaces. Extended application of the Rankine-Hugoniot shock equations to the case of a nearly dry bed has made the model generally applicable to both dry and wet channels. The large reduction in the computational inaccuracies and oscillations at a severe contraction and expansion was achieved by first introducing the actual storage width in the equation of continuity and the imaginary conveyance width in the equation of motion and then solving them with the help of internal boundary conditions.