王乐洋,
谷旺旺,
孙龙翔
东华理工大学测绘工程学院, 南昌 330013

基金项目: 国家自然科学基金（41874001）资助


详细信息

作者简介: 王乐洋, 男, 1983年出生, 博士, 教授, 主要研究方向为大地测量反演及大地测量数据处理.E-mail: wleyang@163.com

中图分类号: P227

收稿日期:2021-01-04
修回日期:2021-05-23
上线日期:2021-08-10

A cooperation inverse approach to invert the slip distribution of earthquake sequence

WANG LeYang,
GU WangWang,
SUN LongXiang
Faculty of Geomatics, East China University of Technology, Nanchang 330013, China



MSC: P227

--> Received Date: 04 January 2021
Revised Date: 23 May 2021
Available Online: 10 August 2021


摘要
摘要:在利用大地测量数据对地震序列进行滑动分布反演中，由于地震序列中多次地震的发震间隔短、InSAR数据时间跨度大，包含了多次地震产生的形变，难以利用InSAR数据反演单次地震的滑动分布，故本文利用协同反演方法结合GPS数据分离地震序列中单次地震产生的InSAR形变，从而反演单次地震的滑动分布.为验证本文方法的有效性，分别采用联合反演方法和本文方法进行模拟实验，其中联合反演采用方差分量估计法确定相对权比，实验表明本文方法能够有效地反演出单次地震滑动分布，且本文方法无需确定两类数据的相对权比.采用本文方法反演Ridgecrest地震M_W6.4前震和M_W7.1主震的滑动分布，结果表明M_W7.1主震为西北向右旋走滑断层，M_W6.4前震包括西北向右旋走滑和东北向左旋走滑断层.两次地震产生的最大滑动量分别为4.26 m和1.0 m，累积释放地震矩4.49×10¹⁹N·m和5.28×10¹⁸N·m，对应矩震级为M_W7.07和M_W6.45，反演结果与已有文献结果较为一致.模拟实验与加州Ridgecrest地震序列反演结果表明，协同反演方法能够利用GPS数据和包含多次地震形变的InSAR数据有效地反演出单次地震的滑动分布，且协同反演方法较联合反演方法而言，无需确定两类数据的相对权比.但由于协同反演方法分别采用不同的数据作为先验信息进行反演，故利用精度较低的数据作为约束时，会造成反演精度较差，而联合反演通过权比减小了这种影响.
关键词: 联合反演/
协同反演/
Ridgecrest地震/
共轭断层

Abstract:In the using the geodetic data to invert the slip distribution of seismic sequences, due to the short recurrence interval of multiple earthquakes in the earthquake sequence, and multiple deformations from the earthquake included in the InSAR data with large time span. Therefore, it is difficult to use the InSAR data to invert slip distribution of an earthquake. To deal with the issue, a cooperative inversion method to invert the slip distribution of an earthquake in a seismic sequence is proposed. Furthermore, the suitability of the joint inversion method and the proposed method are respectively validated with the simulated tests, and the variance component estimation method is used to determine relative weight ratio. The simulated experiment results show that the proposed method can effectively invert the slip distribution of an earthquake sequence, and the proposed method need not determine relative weight ratio. Naturally, we also invert the slip distribution of M_W6.4 foreshock and M_W7.1 main shock of the Ridgecrest earthquake using the proposed method. The real experiments results demonstrate that the M_W7.1 main shock is a right-lateral strike-slip fault in the northwest direction, and the M_W6.4 foreshock consists of right-lateral strike-slip fault in northwest direction and left-lateral strike-slip fault in northeast direction. The maximum slips for both earthquakes are 4.26m and 1.0m, respectively, and the cumulative releasing seismic moments are 4.49×10¹⁹N·m and 5.28×10¹⁸N·m with the corresponding moment magnitudes of M_W7.07 and M_W6.45, respectively. Interestingly, the experimental results are identical to that of the existing investigations. The results of the simulation experiment and the Ridgecrest seismic sequence inversion in California show that the proposed method can successfully use the GPS data and InSAR data with multiple seismic deformations to invest the slip distribution of an earthquake, and the proposed method need not determine relative weight ratio. Due to different kinds of data are used in cooperation as prior information, so the results inverted by using low accuracy data will be worse. However, joint inversion can reduce the influence through posed different relative weight ratio factor in data with low accuracy.
Key words:Joint inversion/
Cooperation inversion/
Ridgecrest earthquake/
Conjugate fault



PDF全文下载地址:

http://www.geophy.cn/data/article/export-pdf?id=ed63ee7d-7ba0-40a6-922b-fb244032a4c1