基于CORS观测网络的地壳时变位移场研究与地震三要素反演

发布时间：2018-07-01 12:31

本文选题：GPS地震学 + CORS　；参考：《西南交通大学》2014年博士论文

【摘要】：GPS (Global Positioning System)地震学作为20世纪末的一门新兴学科,在浅源强震地壳形变研究中占有明显的优势。首先,时间尺度上它可实现数秒到数年的地壳形变监测。其次,该观测手段可直接提供时变位移场观测信息,弥补了传统观测手段在强震观测时出现的不足,如饱和、漂移等现象。多个固定GPS测站构成的CORS (Continuously Operating Reference System)观测网络近几年来在地震地壳形变监测方面发展迅猛。然而,深入、系统地将CORS运用于地震地壳形变,进一步发掘其应用潜能还亟待发展。本文在此背景下,以四川2008年Ms8.0汶川地震和2013年Ms7.0芦山地震为主要研究对象,通过解算分析区域CORS观测资料,主要研究了三部分内容,分别为：汶川地震前后,板块运动特性的探索；高频GPS地震信号的提取和分析；准实时地震三要素(发震时刻、震源和震级)的反演。本文的研究工作和贡献主要表现在如下几个方面：1.分析了四川盆地CORS站在汶川地震前后的运动特征。利用GAMIT/GLOBK解算出的以日为单位的坐标时序,通过时序建模对比分析了区域CORS站在汶川地震前后的水平运动速度场和垂直方向的振幅,初步确认汶川地震改变了四川盆地原有的运动特性。具体表现在：震后共模误差的第一贡献率明显降低了20%-40%；地震后PIXI MYAN、CHDU等站水平运动速度场改变明显,但YAAN、QLAI等测站处于闭锁状态；四川盆地CORS站在U分量的平均周年振幅在震前一年明显偏大,此后逐年减小。2.以2013年4月20日的Ms7.0级芦山地震为例,高频GPS信号的时频分析表明,地震对高频CORS站时序的频谱影响范围集中在0.02Hz-0.2Hz,与无震时相比,其功率谱密度明显偏大。10Hz的GPS位移数据和200Hz的强震仪加速度数据通过卡尔曼滤波融合后,可获取更加完整和客观的200Hz位移波形,表明两种技术手段之间可相互取长补短。3.利用S变换和本文提出的三维搜索法分析近震高频CORS站坐标时序,在震后60s内可得出与实际情况较相符的发震地点和时刻。利用CORS站观测到的峰值位移和震中距,在300s内拟合出芦山地震的面波震级为6.6级。矩震级在震后数小时内,通过多个CORS站的静态位移,利用fastCMT算法也可得出与美国地质调查局相一致的结果。4.采用差分定位策略时,共模误差对区域CORS网络坐标时序的影响不容忽视。主成分滤波能有效减弱共模误差的影响,无论是针对汶川地震前CORS站单日解的连续观测时序中,还是短时高频CORS站的坐标时序,均能使数据的均方根(RMS, Root Mean Square)值得到约30%-60%的改善。5.通过有色噪声模型建模,利用极大似然估计法,对近七年(2006-2012年)单日解CORS站坐标时序的噪声特性分析表明,四川区域CORS站的最优噪声模型为白噪声+闪烁噪声和白噪声+幂律噪声模型,两者共占到了78%；高频CORS站坐标时序的谱指数经功率谱估计在1.6—2之间,可认为此时序噪声符合白噪声+随机游走模型。
[Abstract]:GPS (Global Positioning system) seismology, as a new subject at the end of the 20th century, has an obvious advantage in the study of crustal deformation of shallow strong earthquakes. First, on a time scale, it can be used to monitor crustal deformation for several seconds to several years. Secondly, the observation method can directly provide time-varying displacement field observation information, which makes up for the shortcomings of traditional observation methods in strong earthquake observation, such as saturation, drift and so on. The continuous operating reference system (cors), which is composed of several fixed GPS stations, has developed rapidly in recent years in the field of crustal deformation monitoring. However, the application of cors to seismic crustal deformation needs to be developed. In this background, the earthquake of Ms8.0 Wenchuan in 2008 and the Lushan earthquake of Ms7.0 in 2013 are taken as the main research objects in this paper. Through the calculation and analysis of the regional cors observation data, three parts are studied: before and after the Wenchuan earthquake. Exploration of plate motion characteristics, extraction and analysis of high frequency GPS seismic signals, inversion of three elements of quasi real time earthquakes (time of occurrence, source and magnitude). The research work and contribution of this paper are as follows: 1. The movement characteristics of cors station in Sichuan basin before and after Wenchuan earthquake are analyzed. By using the coordinate time series calculated by GAMIT / GLOBK, the horizontal velocity field and the amplitude of vertical direction of the cors station before and after the Wenchuan earthquake are compared and analyzed by time series modeling. It is preliminarily confirmed that Wenchuan earthquake changed the original movement characteristics of Sichuan basin. The main results are as follows: the first contribution rate of the common mode error after the earthquake obviously decreased by 20-40, the horizontal velocity field of PIXI MYANN CHDU and other stations changed obviously after the earthquake, but the YAAN-QLAI and other stations were in a closed state. The average annual amplitude of the U component of the cors station in Sichuan Basin is obviously larger one year before the earthquake, and then decreases by 0.2. Taking the Lushan earthquake of Ms7.0 on April 20, 2013 as an example, the time-frequency analysis of high frequency GPS signal shows that the frequency spectrum influence range of the earthquake on the time series of high frequency cors station is 0.02Hz-0.2Hz. When the GPS displacement data with the power spectral density of .10Hz and the acceleration data of the 200Hz strong seismograph are fused by Kalman filter, a more complete and objective 200Hz displacement waveform can be obtained, which shows that the two techniques can complement each other. The coordinate sequence of the high frequency cors station of near earthquakes is analyzed by using S transform and the 3D search method proposed in this paper. Within 60 seconds after the earthquake, the location and time of occurrence can be obtained in accordance with the actual situation. Based on the peak displacement and epicentral distance observed by the cors station, the surface wave magnitude of the Lushan earthquake is estimated to be 6. 6 in 300 s. Within a few hours after the earthquake, the moment magnitude can also be obtained by using FastCMT algorithm through the static displacement of several cors stations, which is consistent with that of the US Geological Survey (USGS). When using differential positioning strategy, the influence of common-mode error on coordinate timing of regional cors network can not be ignored. Principal component filtering can effectively reduce the influence of common-mode error, not only in the continuous observation time series of the single day solution of the cors station before the Wenchuan earthquake, but also in the coordinate time series of the short-time high-frequency cors station. Both make RMS (Root mean Square) worth about 30% to 60% better. By modeling the colored noise model and using the maximum likelihood estimation method, the noise characteristics of the single day solution of the cors station coordinates in the last seven years (2006-2012) are analyzed. The optimal noise models of cors station in Sichuan region are white noise flicker noise and white noise power law noise model, and the spectral exponents of high frequency cors station coordinate sequence are estimated to be between 1.6-2 by power spectrum. It can be considered that this time series noise accords with the white noise random walk model.
【学位授予单位】：西南交通大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：P228.4;P315.7

【参考文献】