洪水预报实时校正技术研究进展

Research advances on real-time correction methods for flood forecasting

回顾国内外洪水预报实时校正的产生背景,评述其理论与方法的发展历程。在此基础上,将实时校正方法归纳为终端误差校正和过程误差校正两类,并梳理出各自的典型校正方法以及联合校正方法,概述不同方法的研究成果及进展。重点介绍其中的反馈模拟技术、误差自回归算法(AR)、递推最小二乘算法(RLS)、卡尔曼滤波技术(KF)和动态系统响应曲线算法(DSRC)等5种代表性的实时校正技术,阐述其计算过程,并分析其特点与适用性。对洪水预报实时校正的未来发展方向及研究热点进行了展望。

The background of real-time correction for flood forecasting is reviewed,and the research advances in real-time correction are summarized.On this basis,real-time correction was divided into two categories such as the terminal bias correction and process bias correction.The correction methods in these two approaches were sorted out,and the research results and progress were presented.Five representative real-time correction algorithms,i.e.,feedback simulation,auto-regressive(AR),recursive least squares(RLS),Kalman filtering(KF),and dynamic system response curve(DSRC),were introduced,and their characteristics and applicability were analyzed.The future development direction and research hotspots of real-time correction for flood forecasting were predicted.