Application of Stochastic System Theory for Rainfall Runoff Analysis

通过流域蓄、流量非线性关系的时变参数变化，研究了山区小流域洪水径流过程。本文应用随机系统理论，提出了一种参数辨识方法，主要研究结果如下：1。利用随机系统理论，在状态空间中进行参数论证.随时间变化的蓄水参数的变化与以沟道径流为代表的坡面快速流的产生密切相关，通过识别蓄水参数的变化可以反推坡面平均径流.本文介绍了一个以蓄量Sc来联结以贯流为主和以急流（以沟流为代表）为主时流域蓄流量关系的模型。结果表明，仅用降雨和流量资料，可通过非线性滤波估计Sc。给出了利用Sc进行参数辨识的方法.明确径流模型参数的空间变化规律是集总问题研究的重要内容。

Studies are made on flood runoff processes in small mountainous basin through variation in timespace-varying parameters of the nonlinear relationship between basin storage and dischage. An identification method of the parameters is used by the stochastic system theory.The results are summarized as follows:1. Parameter argumentation is utilized in state space formulation by using stochashic system theory.2. The variations of time-varying storage parameters are closely related to the generation of fast flowe on a hillslope as represented by gully runoffs,and it is possible to inversely estimate the average flow-ststes by identification of the variations.3. A model,in which storage Sc is used to join the ralationship between basin storage and discharge when throughflows predominate and that when fast flows (represented by gully flow) predominate,is introduced. It is shown that Sc can be estimated by nonlinear filtering with only rainfall and discharge data. The method of identifying the parameters using Sc are also shown.4. It is important for lumping problem to clarify space variations of runoff model parameters.