陈石1,2,,,
庄建仓3,
张贝1,2,
石磊1,2
1. 中国地震局地球物理研究所, 北京 100081
2. 北京白家疃地球科学国家野外科学观测研究站, 北京 100095
3. 日本数理统计研究所, 东京 190-8562
基金项目: 科技部重点研发专项项目（2018YFC0603502，2018YFC1503201，2017YFC1500503），中国地震科学实验场项目（2019CSES0105），国家自然科学基面上项目（41774090），地震联合基金项目（U1939205，U1839210）和中国地震局地球物理研究所基本科研业务（DQJB204，DQJB19B37）联合资助
详细信息
作者简介: 李红蕾, 博士, 助理研究员, 主要研究方向为重力反演方法与解释.E-mail: lihonglei@cea-igp.ac.cn
通讯作者: 陈石, 男, 研究员, 研究方向为重力学与地球动力学.E-mail: chenshi@cea-igp.ac.cn
中图分类号: P312收稿日期:2020-04-04
修回日期:2020-12-19
上线日期:2021-04-10
Gravity inversion method base on Bayesian-assimilation and its application in constructing crust density model of the Longmenshan region
LI HongLei1,2,,CHEN Shi1,2,,,
ZHUANG JianCang3,
ZHANG Bei1,2,
SHI Lei1,2
1. Institute of Geophysics, China Earthquake Administration, Beijing 100081, China
2. Beijing Baijiatuan Earth Sciences National Observation and Research Station, Beijing 100095, China
3. The Institute of Statistical Mathematics, Research Organization of Information and Systems, Tokyo 190-8562, Japan
More Information
Corresponding author: CHEN Shi,E-mail:chenshi@cea-igp.ac.cn
MSC: P312--> Received Date: 04 April 2020
Revised Date: 19 December 2020
Available Online: 10 April 2021
摘要
摘要:重力异常对地壳横向密度变化敏感，而无约束重力反演得到的密度模型其垂向分辨能力往往不理想.为了改善反演结果的垂向分辨率，本文参考已有先验分层模型，基于贝叶斯原理，提出了一种重震联合反演的新策略，可实现多种参考模型和复杂加权参数条件下的最大后验概率估计.理论模型测试结果表明，对于深度加权、多参考模型约束等多种问题，本文提出的新方法都可以稳健地获得最优化的模型参数.本文同时以中国地震科学台阵在龙门山地区及周边的一维接收函数分层模型和地震层析成像结果为参考，通过此方法对该区的重力异常进行反演，获得了该区的高精度三维密度结构，其水平分辨率优于10 km，垂直分辨率优于5 km.结合四条通过汶川和芦山地震震中的剖面进行分析后发现，反演得到的密度结构模型在过强震震源区位置横向变形显著，其揭示的分层地壳结构和变形模式与地表已知断裂构造具有相关性.本文提出的重震联合反演新策略，可为研究潜在强震风险源区的地壳结构和物性特征提供有效的科技方法支撑.
关键词: 重力反演/
贝叶斯推断/
三维密度结构/
龙门山构造/
数据同化
Abstract:Gravity anomalies are sensitive to the lateral change of crust density. While the vertical resolution of the density model obtained by unconstrained gravity inversion is not satisfactory. To solve this problem, a new strategy for joint inversion of gravity and seismic data is proposed by referring to the existing prior hierarchical model based on Bayesian assimilation. This new strategy can obtain the optimal maximum posterior probability estimation derived from the multiple reference models and complex weighted parameters. Tests on synthetic models show that the new method proposed in this paper can obtain the optimal model parameters stably as for depth weighting, multi-reference model constraint and other problems. Meanwhile, the three-dimensional high-precision density structure in and around the Longmenshan area is obtained by using the one-dimensional receiver function stratification model and seismic tomographic results of the China Seismic Array in the Longmenshan area as a high precision three-dimensional density structure with horizontal resolution better than 10 km and vertical resolution better than 5 km is obtained by the gravity inversion in terms of this method. Combined with the four profiles through the epicenter of the Wenchuan and Lushan earthquakes, the density structure model obtained by the inversion has significant lateral deformation at the location of the over-intense earthquake source area, and the layered crustal structure and deformation pattern revealed by the inversion are correlated with the known fault structures on the surface. The new gravity and seismic inversion strategy proposed in this paper can provide an effective methodology support for the study of crustal structure and physical characteristics in the source areas of potential major earthquakes.
Key words:Gravity inversion/
Bayesian inference/
3D density structure/
Longmenshan tectonic/
Data assimilation
PDF全文下载地址:
http://www.geophy.cn/data/article/export-pdf?id=6076842bc23e6710c26d5b32
