多源布格重力异常数据贝叶斯融合算法及其在川滇地区的应用  

A Bayesian data fusion algorithm for multi-source Bouguer gravity anomalies and its application in the Sichuan-Yunnan region

作　　者：李红蕾 陈石[1,2] 庄建仓 张贝 卢红艳 LI HongLei;CHEN Shi;ZHUANG JianCang;ZHANG Bei;LU HongYan(Institute of Geophysics,China Earthquake Administration,Beijing 100081,China;Beijing Baijiatuan Earth Sciences National Observation and Research Station,Beijing 100095,China;The Institute of Statistical Mathematics,Research Organization of Information and Systems,Tokyo 190-8562,Japan)

机构地区：[1]中国地震局地球物理研究所,北京100081 [2]北京白家疃国家地球科学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

出　　处：《地球物理学报》2021年第9期3232-3245,共14页Chinese Journal of Geophysics

基　　金：国家重点研发专项项目(2018YFC1504506,2017YFC1500503,2018YFC1503201);国家自然科学基金项目(42004074,41774090);地震联合基金项目(U1839210,U1939205);中国地震科学实验场项目(2019CSES0105);中国地震局地球物理研究所基本科研业务(DQJB20Z04,DQJB19B37)联合资助.

摘　　要：布格重力异常数据是研究地壳结构和变形的重要依据,由于观测手段、处理方法等存在差异,不同手段获得的布格重力异常可能存在数据噪声水平及起算基准不一致,从而引起多源布格重力异常数据难以有效融合的难题.为此,本文在传统等效源重力数据融合方法的基础上,通过引入贝叶斯参数优化准则,提出了一种新的多源布格重力异常融合算法.随后对川滇地区两条实测重力剖面布格异常和WGM2012模型计算布格异常开展了融合实验,结果表明:新算法不仅能显著降低不同异常中的非相干噪声,而且能将多源异常的系统偏差有效纠正.与已有布格重力异常相比,融合所得布格重力异常特征与川滇地区主要线性构造分布形态具有更好的一致性.本文给出的川滇地区融合布格重力异常同时兼顾了地表实测重力的高精度和卫星重力场全球模型的高分辨率,可为研究川滇地区地壳结构和构造区划提供更加有价值的地球物理基础资料.Bouguer gravity anomaly data are important sources for the study of crustal structure and deformation.Due to different observation means,processing methods and so on,there may be obvious inconsistent in noise level and datum among Bouguer anomalies which obtained by different approaches.This leads to the difficulty of effective integration of multi-source Bouguer gravity anomaly data.Therefore,based on the traditional gravity data fusion method that is based on the equivalent-source principle,this paper proposes a new data fusion algorithm for multi-source Bouguer gravity anomaly by introducing the Bayesian optimization criteria.At the same time,a numerical experiment was carried out on two actual gravity profile Bouguer anomalies and the Bouguer anomaly calculated by the WGM2012 model in the Sichuan-Yunnan area.The results show that the new algorithm can not only significantly reduce the incoherent noise in different anomalies,but also effectively correct the system deviation of multi-source anomalies.Compared with the existing results of Bouguer gravity anomaly,the characteristics of the Bouguer gravity anomaly obtained by this data fusion algorithm are more consistent with the distribution of main linear structures in the Sichuan-Yunnan area.The Bouguer gravity anomaly model presented in this paper incorporates to both the high accuracy of ground measured gravity and the high resolution of global satellite gravity model,and provides valuable basic geophysical data for the study of crustal structure and tectonic zoning in the Sichuan-Yunnan region.

关 键 词：等效源反演 贝叶斯优化 数据融合 布格重力异常 川滇地区 

分 类 号：P312[天文地球—固体地球物理学]