中文摘要：

传统的水沙数学模型基于水沙运动守恒规律，在给定模型参数和边界条件下封闭未知变量，产生可求解的静定方程组。本文在传统水沙数学模型的基础上，引入水沙实时观测值，利用集合卡尔曼滤波等方法，构造水位、流量和含沙量等未知变量的状态空间方程，通过滤波求解获得他们的优化值，并实时更新模型方程的初始场，将传统的水沙数学模型发展为水沙实时预测数学模型。该实时校正模式应用于2011年黄河下游花园口至利津河段调水调沙实验，取得了满意的效果。

英文摘要：

Traditional numerical models for water and sediment transport are solved on the basis of the principle of conservation law and the identification of parameters. In this paper, available observations of water level and sediment concentration are combined into the traditional numerical model to improve the ac-curacy of prediction. The traditional model is transformed into state-space equations, in which the variables including water level,discharge and sediment concentration are changed into input-output system. The opti-mal values of the variables are calculated using the Ensemble Kalman filter and then used to update the initial conditions in the next time step. In this way, the traditional numerical model develops into the dy-namic model. The water and sediment regulation in 2011 was taken as a case to test the performance of the dynamic model. It finds out that the dynamic model can largely improve the accuracy of the predicted water level,discharge and sediment concentration.