中国农业科学 ›› 2011, Vol. 44 ›› Issue (10): 2060-2069.doi: 10.3864/j.issn.0578-1752.2011.10.011

• 土壤肥料·节水灌溉·农业生态环境 • 上一篇    下一篇

密云水库上游白河磷浓度时间序列的傅立叶与小波分析

 张微微, 李红, 孙丹峰, 周连第   

  1. 1.北京市农林科学院农业综合发展研究所,北京 100097
    2.中国农业大学资源与环境学院,北京 100193
  • 收稿日期:2010-06-23 出版日期:2011-05-15 发布日期:2010-10-11
  • 通讯作者: 通信作者李红,Tel:010-51503312;E-mail:lihsdf@cau.edu.cn
  • 作者简介:张微微,Tel:13146394890;E-mail:zhangwei492@163.com
  • 基金资助:

    国家科技支撑项目(2006BAB15B05)、北京市财政局项目

Fourier and Wavelet Analysis of Phosphorus Concentration Time Series in Baihe River in Miyun Reservoir Upstream Watershed

 ZHANG  Wei-Wei, LI  Hong, SUN  Dan-Feng, ZHOU  Lian-Di   

  1. 1.北京市农林科学院农业综合发展研究所,北京 100097
    2.中国农业大学资源与环境学院,北京 100193
  • Received:2010-06-23 Online:2011-05-15 Published:2010-10-11

摘要: 【目的】分析密云水库上游流域水质长期监测数据,获取流域水质时间格局特征。【方法】以密云水库上游白河上的S1和S2监测点总磷(TP)浓度1986—2003年时间序列为例,采用时域分析、傅立叶和小波分析等方法对比来综合揭示磷污染物的周期模式以及时间格局特征。【结果】S1和S2监测点时间序列不存在自相关性和异方差性,不适合用时域方法进行水质的时间演化规律的研究。傅立叶分析得到两个监测点的隐含周期,S1和S2都有6年的周期模式,S2还包含一个2年左右的周期行为。小波分析得到了不同尺度的时间格局特征,S1和S2点都有中尺度的周期格局特征,S2点还有小尺度的周期格局特征。【结论】傅立叶分析和小波分析这两种方法都可以应用在水质的时间演化规律的研究中,傅立叶分析方法比较粗糙,而小波分析方法以其多尺度、时频相结合的特点为研究水质时间特征的演化提供了更精细的方法。

关键词: 傅立叶, 小波分析, 总磷浓度, 水质, 密云水库

Abstract: 【Objective】Analyzing the long-term water quality monitoring data to find temporal pattern of water quality signals over different temporal scales can assist us in understanding various processes affecting water quality temporal evolution at different scales and further ensuring watershed water quality security. 【Method】 Taking the phosphorus monitoring data of two sites S1 and S2 in the period 1986-2003 in Baihe River lying Miyun reservoir upstream watershed as a case, time domain analysis method, Fourier and Wavelet analysis were adopted to explore and compare the periodic patterns and temporal pattern characteristics of the two sites. 【Result】 The results showed that the Time Domain analysis method was not adopted due to no serial correlation and heteroscedasticity in each site series. The periodic patterns of two sites were discovered using Fourier analysis. The site S1 had a period of six years, while the site S2 had two periodic patterns of two years around and six years. The temporal pattern characteristics at different scales were obtained through wavelet analysis, which were at moderate scale for the site S1, while at moderate and small scales for the site S2. 【Conclusion】 The Fourier and wavelet analysis method can both be used in the study of surface water quality temporal change pattern, the first is a coarse method and the latter is a more detailed method for analyzing surface water quality temporal pattern characteristics.

Key words: Fourier, Wavelet analysis, phosphorus, water quality, Miyun reservoir