利用基于集合经验模态分解(EEMD)方法而改进的Hilbert-Huang变换(HHT)方法,分析了吕宋岛西北海域从1997年9月至2009年7月近12年的月平均遥感叶绿素浓度观测序列以及相关环境物理要素时间序列,分离出各要素的特征振荡模态(IMFs);在此基础上以叶绿素与相关要素间具有相同或相近频率的IMF对偶之间相位差余弦值的方差为指标,探讨了该海域叶绿素浓度与环境场之间的联系.结果表明:1)海区各研究变量都具有明显的季节和年际振荡特征,叶绿素准年周期模态方差贡献达81％,年际变化中准两年振荡是海区诸要素共同的波动类型,此外叶绿素浓度还具有4年左右周期的振荡.2)除埃克曼抽吸速度在准年周期振荡上与叶绿素浓度显著正相关、Ni(n)o3区海表温度在准两年周期上有弱的正相关关系外,其余要素均与叶绿素浓度在不同时间尺度上呈负相关关系.这些结果说明HHT是气候序列多时间尺度分析中的一种有力工具.%An improved Hilbert-Huang transform (HHT) based on ensemble empirical mode decomposition (EEMD) is applied to analyze monthly averaged time series of remotely-sensed surface chlorophyll concentration from September 1997 to July 2009 and the relevant hydrological and meteorological factors northwest of the Luzon Island in the South China Sea. On the basis of modes separated by EEMD, we develop a correlation index using the cosine values of the phase differences between a pair of intrinsic mode functions (IMFs), of which one is from chlorophyll and the other is from another variable. Our explora-tion of the relation between chlorophyll and physical environment shows that a seasonal pattern and an interannual oscillation are common characteristics in the time series. Quasi-annual periodicity accounts for 81% of the total variance of chlorophyll, and there are two other modes (the quasi-biennial oscillation and 4-year period, respectively). Further more, it is revealed that a significant positive correlation between seasonal modes of Ekman pumping velocity and chlorophyll, and a weak positive relationship between quasi-biennial modes of Nino3 SST and chlorophyll. Besides, all the other variables show out-of-phase correlation with chlorophyll at different time-scales. These findings demonstrate the usefulness of HHT in climate time seriesanalysis.
展开▼
