中文摘要：

本文基于2004苏门答腊地震后全球地球动力学计划观测网中9台超导重力仪13200h的连续重力观测数据,联合利用整体经验模态分解（EEMD）、最优序列估计（OSE）、积谱密度分析（PSA）和自回归估计（AR）方法,探测地球固态内核平动振荡模态（Slichter模）的三重分裂信号.利用EEMD提取常规预处理后的重力残差中包含目标频段的本征模态函数（IMFs）作为最终的重力残差之后,将每个台站的重力残差平均分成无重叠的3个子数据块,再应用OSE和PSA方法获取Slichter模三重分裂谱线的积谱密度,并采用AR方法估计各积谱中弱共振信号的中心频率及其误差.结果表明,结合OSE和PSA方法探测到三个清晰的谱峰,周期分别为5.8307±7.1×10^-4h、5.2161±8.1×10^-4 h和4.7536±5.2×10^-4 h,分别对应Slichter模三重分裂谱线m=-1、m=0和m=＋1,且三个谱峰的周期与Crossley（1992）、Rochester和Peng（1993）、Peng（1997）和Rogister（2003）基于PREM地球模型给出的Slichter模理论周期非常接近,也与Ding和Shen（2013）建议的＂可能的观测结果＂较为接近.参考现有的理论和观测结果,本文认为这三个信号可能是Slichter模三重分裂谱线.

英文摘要：

Based on 13200-hour records by nine superconducting gravimeters（SGs）under a worldwide network of the Global Geodynamics Project after the 2004 Sumatra earthquake,a series methods including the ensemble empirical mode decomposition（EEMD）,the optimal sequence estimation（OSE）,the product spectrum analysis（PSA）and the autoregressive（AR）are applied jointly to detect the Earth′s inner core translational mode triplet（Slichter mode triplet）.We first obtain the intrinsic mode functions（IMFs）of the target frequency band extracted from the residual sequence after the routine pretreatment as the final gravity residual,and each of the nine residual gravity series is divided into three segments of the same length without overlap,and the product spectra of Slichter triplet are obtained by using OSE and PSA.Then,the AR method is used to estimate the central frequencies of the week resonance signals and the corresponding errors.The results show that three clear spectral peaks are detected usingOSE and PSA,which correspond respectively to the singlets m=-1,m=0and m=＋1.The periods corresponding to the three spectral peaks are close to the prediction of the PREM model given by Crossley（1992）,Rochester and Peng（1993）,Peng（1997）and Rogister（2003）,as well as the＂possible observation results＂suggested by Ding and Shen（2013）.Referring to the existing theory and observed results of the Slichter triplet,we conclude that the three observed signals in this study might correspond to the Slichter triplet.