BMA方法在黄河龙门站含沙量过程 预报中的应用研究

以黄河流域吴堡站到龙门站为研究区域，建立龙门站含沙量过程的线性动态和BP神经网络两种统计模型，并结合BMA方法给出两种模型的综合预报结果。结果表明，在线性动态模型和BP神经网络模型的基础上，采用BMA方法进行含沙量预报模型的合成预报不仅可以提高预报精度，而且可以提供概率预报值。

Study on application of BMA method to forecasting of sediment concentration hydrograph at Longmen Hydrological Station on Yellow River

Taking the area from Wubao Hydrological Station to Longmen Hydrological Station within the Yellow River Basin as the study area，two statistical models for the sediment concentration hydrograph at Longmen Hydrological station，i.e. a linear dynamic model and a BP neural network model are established，and then the comprehensive forecasting results from both the models are given in combination with BMA method. The results show that on the basis of the linear dynamic model and the BP neural network model，the application of BMA method to the combined forecasting with the sediment concentration models can not only enhance the accuracy of the forecasting，but can also provide the probabilistic forecasting value.