SMG水文系统模型的实时校正

本文在分析讨论了具有物理成因概念的系统模型(SMG模型)误差和BP网络应用问题的基础上,对其加以改进建立了实时预报校正模型。将该模型应用于洮河流域红旗站日径流实时预报,结果显著地提高了预报方案的确定性系数,对系统模型的洪峰预报精度也有一定的提高。

Real-time Calibration of System Model of SMG

On the basis of discussion on Hydrological System Model of generic error and problems in application of the BP neural network,a real-time calibration model for hydrologic forecast was built with modified BP neural network in this paper.The function of the model is to predict a calibrated value used to update the forecast out of rainfall-runoff model with observed date of the previous time section.As a case study,the model is employed in daily runoff forecasting at Hongqi gauging station on the Taohe river in Qinghai province.The results shows that this model can improve the coefficient of the model efficiency significantly and daily peak discharge to some extend.