Using ensemble data assimilation to forecast hydrological flumes

Abstract. Data assimilation, commonly used in weather forecasting, means combining a mathematical forecast of a target dynamical system with simultaneous measurements from that system in an optimal fashion. We demonstrate the benefits obtainable from data assimilation with a dam break flume simulation in which a shallow-water equation model is complemented with wave meter measurements. Data assimilation is conducted with a Variational Ensemble Kalman Filter (VEnKF) algorithm. The resulting dynamical analysis of the flume displays turbulent behavior, features prominent hydraulic jumps and avoids many numerical artifacts present in a pure simulation.