陇西及周边地区地震活动与震磁效应的高阶统计量研究
|
摘要
高阶统计量是研究非高斯过程,非最小相位信号和非线性系统的有力工具,其应用领域已涉及通信、地球物理、生物医学、故障诊断等。本文对嘉峪关和乾陵2个地磁台站1998～2002年期间的地磁垂直分量(Z)资料运用高阶统计量方法进行处理,探讨陇西及其周边地区地震活动与地磁场变化之间相互关系,并将其结果与相关分析和线性拟合方法的结果进行对比。结果发现,高阶统计量异常一般早于两台Z分量相关系数和嘉峪关台Z分量线性拟合差异常1～2个月出现,且三阶矩(三阶累积量)的异常变化幅度在5～10之间,四阶矩的异常幅度在50以上,甚至达到150,四阶累积量的异常幅度在10～60之间。这些表明在震磁效应的统计分析研究中引入高阶统计量方法的必要性及其在地震预报中的潜力和良好前景。
Higher-Order Statistics (HOS) is a useful tool for studying non-Gaussian process, non-minimum phase system and nonlinear systems. The HOS method has been applied widely in different fields such as communication, geophysics, biomedical, failure diagnosis etc. the data of geomagnetic vertical component in Jiayuguan and Qianling Seismostation from 1998 to 2002 is processed with HOS method, and the relationship between the seismicity and the variations of geomagnetic field in west Gansu and its adjacent area is discussed. Through comparing the results with those from correlative analysis and linear fitting we found that the anomalies of HOS appeared 1~2 months earlier than those of correlation coefficient and linear fitting before an earthquake. Usually, the variation of third order moment anomaly (third order cumulant) is from 5 to 10, and of the fourth order is from 10 to 60. The magnitude of fourth order moment anomalies is over 50, sometimes even reach 150. The results show it is significant to introduce HOS method into the statistics of seismo-magnetic effect.
Higher-Order Statistics (HOS) is a useful tool for studying non-Gaussian process, non-minimum phase system and nonlinear systems. The HOS method has been applied widely in different fields such as communication, geophysics, biomedical, failure diagnosis etc. the data of geomagnetic vertical component in Jiayuguan and Qianling Seismostation from 1998 to 2002 is processed with HOS method, and the relationship between the seismicity and the variations of geomagnetic field in west Gansu and its adjacent area is discussed. Through comparing the results with those from correlative analysis and linear fitting we found that the anomalies of HOS appeared 1~2 months earlier than those of correlation coefficient and linear fitting before an earthquake. Usually, the variation of third order moment anomaly (third order cumulant) is from 5 to 10, and of the fourth order is from 10 to 60. The magnitude of fourth order moment anomalies is over 50, sometimes even reach 150. The results show it is significant to introduce HOS method into the statistics of seismo-magnetic effect.
引文
[1] 熊仲华,刘运生.地磁观测技术[M].北京:地震出版社,1997,167 182.
[2] 卡·纳·阿勃杜拉别可夫.地壳中的电磁现象[M].蒋宏耀译.北京:学术书刊出版社,1990.
[3] E.M.扬诺夫斯基.地磁学[M].刘洪学,周姚秀译.北京:地质出版社,1981.
[4] Duma,G.SeismicitycyclesintheMt.Vesuviusareaandtheirrelationtosolarfluxandthevariationsoftheearth'smagneticfield[J].PhysicsandChemistryoftheEarth,1998,23(9 10):927 931.
[5] Sasai,Y.Piezomagneticfieldsproducedbydislocationsources[J].SurveysinGeophysics,1994,15(4):363 382.
[6] Schmucker,U.Anintroductiontoinductionanomalies[J].J.Geomag.Geoeletr.1970,22,9 33.
[7] 丁鉴海,卢振业,黄雪香.地震地磁学[M].北京:地震出版社,1994,220 284.
[8] 张贤达.时间序列分析-高阶统计量方法[M].北京:清华大学出版社,1996.
[9] 李宏伟,程乾生.高阶统计量与随机信号分析[M].武汉:中国地质大学出版社,2002,1 58.
[10] Giannakis,G.B.Cumulants:APowerfulToolinSignalProcessing[J].Proc,IEEE,1987,75,1333 1334.
[11] Mendel,M.TutorialonHigher-OrderStatistics(spectra)inSignalProcessingandSystemTheoryResultsandSomeApplications[J].Proc.,IEEE.,1991,79(3):278 305.
[12] Nikias,L.C.,RaghuveerM.R.Bispectrumestimation:AdigitalsignalProcessingframework[J].Proc,IEEE,1987,75,869 891.
[13] 尹成,伍志明,邓怀群.高阶统计量方法在地震勘探中的应用[J].地球物理学进展,2003,18(3):546 550.
[14] 唐兵,尹成.基于高阶统计的非最小相位地震子波恢复[J].地球物理学报,2001,44(3):404 410.
[15] 张浩,赵正予,谢树果,等.地磁脉动信号的双谱分析[J].武汉大学学报(理学版),2001,47(3):351 354.
[16] 沈民奋,孙丽萍,沈风麟.心音信号的非高斯AR模型双谱分析[J].中国生物医学工程学报,2000,19(1):72 77.
[17] 马石庄.地磁场长期变化的自洽正态统计模型[J].地球物理学报,1997,40(5):616 626.
[18] 赵明,林云芳,曾小苹,等.1996年12月16日高丽营地震的磁效应[J].地震,1998,18(增刊):98 102.
[19] 林美,沈斌等.地磁场垂直分量相关分析与地震的对应关系[J].地震研究,1982,5(2):212 219.
[2] 卡·纳·阿勃杜拉别可夫.地壳中的电磁现象[M].蒋宏耀译.北京:学术书刊出版社,1990.
[3] E.M.扬诺夫斯基.地磁学[M].刘洪学,周姚秀译.北京:地质出版社,1981.
[4] Duma,G.SeismicitycyclesintheMt.Vesuviusareaandtheirrelationtosolarfluxandthevariationsoftheearth'smagneticfield[J].PhysicsandChemistryoftheEarth,1998,23(9 10):927 931.
[5] Sasai,Y.Piezomagneticfieldsproducedbydislocationsources[J].SurveysinGeophysics,1994,15(4):363 382.
[6] Schmucker,U.Anintroductiontoinductionanomalies[J].J.Geomag.Geoeletr.1970,22,9 33.
[7] 丁鉴海,卢振业,黄雪香.地震地磁学[M].北京:地震出版社,1994,220 284.
[8] 张贤达.时间序列分析-高阶统计量方法[M].北京:清华大学出版社,1996.
[9] 李宏伟,程乾生.高阶统计量与随机信号分析[M].武汉:中国地质大学出版社,2002,1 58.
[10] Giannakis,G.B.Cumulants:APowerfulToolinSignalProcessing[J].Proc,IEEE,1987,75,1333 1334.
[11] Mendel,M.TutorialonHigher-OrderStatistics(spectra)inSignalProcessingandSystemTheoryResultsandSomeApplications[J].Proc.,IEEE.,1991,79(3):278 305.
[12] Nikias,L.C.,RaghuveerM.R.Bispectrumestimation:AdigitalsignalProcessingframework[J].Proc,IEEE,1987,75,869 891.
[13] 尹成,伍志明,邓怀群.高阶统计量方法在地震勘探中的应用[J].地球物理学进展,2003,18(3):546 550.
[14] 唐兵,尹成.基于高阶统计的非最小相位地震子波恢复[J].地球物理学报,2001,44(3):404 410.
[15] 张浩,赵正予,谢树果,等.地磁脉动信号的双谱分析[J].武汉大学学报(理学版),2001,47(3):351 354.
[16] 沈民奋,孙丽萍,沈风麟.心音信号的非高斯AR模型双谱分析[J].中国生物医学工程学报,2000,19(1):72 77.
[17] 马石庄.地磁场长期变化的自洽正态统计模型[J].地球物理学报,1997,40(5):616 626.
[18] 赵明,林云芳,曾小苹,等.1996年12月16日高丽营地震的磁效应[J].地震,1998,18(增刊):98 102.
[19] 林美,沈斌等.地磁场垂直分量相关分析与地震的对应关系[J].地震研究,1982,5(2):212 219.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心