蒸发蒸腾量时间序列的混沌特征分析  

The Chaotic Features of Evapotranspiration Time Series

作　　者：战国隆[1] 马孝义[1] 刘继龙[1] 朱亚磊[1] 吕静渭[1] 

机构地区：[1]西北农林科技大学旱区农业水土工程教育部重点实验室,陕西杨凌712100

出　　处：《中国农村水利水电》2010年第3期22-24,28,共4页China Rural Water and Hydropower

基　　金：国家自然科学基金项目(50879072);国家科技支撑计划(2006BAD11B04);国家"863"计划课题(2006AA100209)

摘　　要：利用成都1951-2000年的气象资料和地理参数,用Penman-Monteith公式计算蒸发蒸腾量(ET0)。根据混沌理论,采用Welch法和小数据量法识别ET0的混沌特征,根据G-P关联维和C-C方法得出主要的混沌特征指标,并利用RBF神经网络和Volterra级数进行一步预测。研究表明:ET0序列存在一定的混沌特征,最小嵌入维数m=6时对应的吸引子维数D=2.025,最大李雅诺夫指数大于0;利用RBF神经网络和Volterra级数自适应方法对ET0进行预测一步预时,Volterra级数自适应一步预测结果误差较小,预测精度较高。因此,在实际中可采用Volterra级数自适应模型进行ET0一步预测。According to the meteorological data and geographical parameters between 1951 and 2000, the evapotranspiratJon （ETo） for Penman-Monteith model is calculated. This paper analyzes the chaos characteristic of ETo and displays the main chaos character- istic target according to the chaotic theory. And RBF neural network and Volterra series forecast for one step is used. This result indicates that ETo has chaos characteristic. When the minimum embedding dimension is 6, the corresponding attracts dimension is 2. 025, and the maximum Lyapunov index is bigger than zero. Using the RBF neural network and the Volterra series forecast ETo for one step, the Voherra series forecast ETo for one step forecasting result error is small, the precision of forecast is high. Therefore, ETo can be predicted with the Volterra series forecast in actual application.

关 键 词：作物蒸发蒸腾量 混沌理论 李雅诺夫指数 

分 类 号：S161.4[农业科学—农业气象学]