东江流域降水时间序列的混沌特征分析  

Chaotic Characteristics of Precipitation Time Series in the Dongjiang River Valley

作　　者：石教智[1] 陈晓宏[1] 林汝颜[2] 

机构地区：[1]中山大学水资源与环境系,广东广州510275 [2]广东省水利电力勘测设计研究院,广东广州510170

出　　处：《中山大学学报（自然科学版）》2006年第4期111-115,共5页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：国家自然科学基金资助项目(50579078);广东省自然科学基金资助项目(04009805)

摘　　要：以广东省东江流域月降雨序列为例,在介绍相空间重构原理的基础上,探讨了混沌分析的主要定量指标:饱和关联维数D2和最大Lyapunov指数λ。得到该时间序列的饱和关联维数D2=3.93,最小嵌入维数m=8,最佳嵌入滞时τ=3个月,最大Lyapunov指数λ=0.253。并且采用主分量方法进一步验证了该序列具有混沌特性,指出该序列的预测时限不应超过4个月,对此结论则用ARMA(p,q)模型作了验证,为东江流域月降雨预测提供了较为科学的依据。Based on introducing the phase space reconstruction theory of chaotic time series, the main quantitative indexes of saturation correlation dimension D2 and maximal Lyapunov exponent A for chaotic analysis are discussed with the monthly rainfall time series of the Dongjiang River Valley in Guangdong province. The saturation correlation dimension, minimum embedding dimension, optimal built-in delay time and maximal Lyapunov exponent are calculated and given, that is D2 = 3.93, m = 8, τ = 3 and λ = 0. 253. On the basis of this, primary component analysis method is applied to validate its chaotic character. It is suggested that the forecasting length for this rainfall time serial should not exceed 4 months and this is verified with the model of ARMA. This time series chaotic analy- sis will provide a scientific gist for monthly rainfall forecasting in the Dongjiang River Valley.

关 键 词：东江流域 相空间重构 关联维数 主分量分析 LYAPUNOV指数 

分 类 号：P333.6[天文地球—水文科学]