Dynamic routing model with real-time roughness updating for flood forecasting

A flood routing model with time-varying roughness updating was developed to simulate flows through natural channels based on the dynamic wave theory. Taking observed stages as the targets, a roughness updating technique was developed using the Gauss-Newton method to update the Manning n in each time step of the routing processes. The technique provides a reasonable roughness coefficient estimate and reliable initialization of stage profile for the forecast. The examinations including the initialization of stage profile, conservation of mass, iteration convergence, effectiveness evaluation, and convergence with different initial values were conducted to verify the predictive capability of the roughness updating technique. The forecasting results show that the stage recalculated by updating the Manning n in current time has a good agreement with the observed stage. The presented model can improve the forecast for a lead time up to 6 h in the Tanshui River of northern Taiwan.