中文摘要：

在对混沌理论及其在径流系统应用的适应性分析的基础上,提出河川径流混沌分析方法;以黄河为研究对象对月径流序列进行相空间重构,并进行混沌特征识别和分析,得到如下结论：天然月径流比实测月径流的饱和关联维数要大,要恰当描述实测月径流序列的变化特征,进行动力系统建模,最少需要4个独立变量,最多需要8个独立变量,要描述天然径流序列则最少需要5～6个,最多需要12个独立变量;同一水文站、同一时期的实测径流序列和天然序列的混沌特征不同;下游的混沌特征要强于上游;从上世纪五十年代到本世纪初黄河干流月径流的混沌特性比上世纪二十年代到上世纪七十年代月径流的混沌特性要稍强;所采用径流时间序列的长、短对混沌特征的识别有影响,序列越长,所表现的混沌特征就越强。黄河径流具有混沌特征,为径流系统的建模和混沌预测提供了依据。

英文摘要：

This paper proposes an analysis method of river runoff chaotic character with a discussion on chaos theory and applicability of this method to river runoff system,applies the method to the Yellow River and reconstructs a phase space of its monthly runoff.The conclusions are as follow.Satiable correlation dimension of natural runoff is greater than the measured runoff.For a dynamical model,at least 4 variables and at most 8 variables are needed to depict the variation character of measured runoff,while at least 5-6 and at most 12 for natural runoff.Measured runoff has different chaotic character from that of natural runoff at the same hydrological station in the same time period,and the downstream reach has stronger chaotic character than the upstream.The chaotic character in the period from 1950s to the beginning of this century is somewhat stronger than that of 1920s to 1970s.The chaotic character depends on the length of runoff time series,and longer series has a stronger character.The chaotic character revealed in the present study would be a basis for development of dynamical models and runoff forecast for the Yellow River.