【作者】 李皎皎；

【导师】 黄金莉；

【作者基本信息】 中国地震局地震预测研究所 ， 固体地球物理学， 2012， 硕士

【摘要】 中国东北地区地处西伯利亚块体与中朝块体之间，东临西太平洋边缘海，区域内构造格局错综复杂，主要地质构造单元如松辽盆地、大兴安岭和张广才岭等均呈近北东向展布并被交错复杂的断裂系统所分割，著名的郯庐断裂带北段贯穿全区。本区经历多期构造作用，岩石圈减薄已得到地球化学和地震面波层析成像结果的证实。鉴于东北地区地质构造的复杂性和构造运动的特殊性，研究其岩石圈地震波速度结构，对于揭示该区区域构造特征及深部动力学机制等都具有重要意义。面波层析成像是通过反演面波频散曲线得到群速度分布，不同周期的群速度反映不同深度范围内的速度结构，周期越大，反映的深度越深；而且，与体波成像相比，面波成像具有更高的垂向分辨率。因此，该方法被广泛应用于地壳上地幔速度结构的研究。由于面波的固有衰减和散射，用地震面波成像的方法难以得到浅层的速度结构，而近年来发展的背景噪声成像方法摆脱了依赖有源地震的限制，能够提取短周期面波频散曲线，得到高分辨率的浅层结构。而从背景噪声中又难以提取可靠的长周期面波频散资料，因此仅用背景噪声成像很难得到深层的结构信息。本研究应用东北地区的黑龙江、吉林、辽宁和内蒙古四省区域地震台网119个宽频带台站记录的波形资料，结合背景噪声成像和传统的单台地震面波成像方法获得了研究区从浅至深直至约100km的群速度分布，充分发挥了两种方法的优势。由于两种方法在提取瑞利面波群速度频散曲线时所用资料和数据处理方法不尽相同，本研究先分别利用背景噪声和地震面波提取频散曲线，再由这些频散曲线反演面波群速度分布。其中，利用背景噪声测量频散曲线时，选取的数据为区域台网记录的2009年1月至2010年12月连续波形资料的垂向分量；数据处理分三步进行：（1）对单台数据进行预处理得到背景噪声；（2）计算并叠加台站对间背景噪声的互相关函数；（3）采用多重滤波方法测量频散曲线并对其进行人工筛选和质量控制。地震面波成像应用区域地震台网记录的2007年8月至2011年7月的波形资料，从302个震级大于5.5级、震中距60°以内、震源深度小于50km的地震事件中精心筛选出瑞利面波发育良好、频散特征明显的116个地震事件用于本次研究；对面波资料进行数据预处理后，采用多重滤波法测量出瑞利面波群速度频散曲线。经过上述数据处理，本研究分别应用背景噪声及地震面波提取的5－30s和25－70s瑞利面波群速度频散曲线，采用层析成像方法反演获取东北地区8－70s瑞利面波群速度分布，进而得到研究区从浅到深直至约100km的岩石圈速度结构。对于反演结果的可靠性，采用检测板分辨率分析方法进行检测。结果显示，背景噪声部分分辨率较高，能达到1°×1°；地震面波部分由于射线分布相对稀疏，大部分地区的反演结果能在2°×2°网格划分下得到较好恢复。对比两种方法在周期为25s和30s时的群速度图像说明其结果较为一致。反演结果表明东北地区岩石圈速度结构存在明显的横向不均匀性。周期为8s至15s的短周期群速度分布与地表构造特征有较好的对应关系；盆岭边界的大型断裂如依兰－伊通断裂对上地壳速度结构的影响明显，松辽盆地表现出较厚的低速沉积盆地特征。周期为20s至30s的群速度高低速分布与短周期相比出现了反转变化，表明以大兴安岭重力梯度带为界，西部地区莫霍面深度大于东部地区。周期为50s至70s的长周期群速度图表现为随着周期的增加东部低速区西移，而西部表现为高速稳定的特征，这可能反映了研究区受太平洋板块俯冲影响，大兴安岭以东地区软流圈热物质上涌的特征。更多还原

【Abstract】 Northeast China is located between Siberian block and Sino-Korean block，aswell as to the west of the West Pacific marginal seas. The main geological tectonicunits such as the Songliao Basin，Daxinganling and Zhangguangcai Ridge show anearly north-east trending and are separated by the complex fault system，includingthe northern section of famous Tanlu fault zone running through the whole area. Thisregion has been underwent multiphase tectonism，the lithospheric thinning wasconfirmed by the results from geochemical studies and surface wave tomography.Therefore，it is significant for us to study the lithospheric structure under theNortheast China to reveal its regional tectonic features and deep dynamicsmechanism.Surface wave tomography provides the distribution of group velocity byinversing dispersion measurements. Surface waves at different periods are sensitive toEarth structure at different depths，with the longer period waves exhibiting sensitivityto greater depths. Compared to body wave tomography，surface wave tomography hasa higher vertical resolution，which is widely used in imaging the structure of crust andupper mantle. However，it is difficult to obtain the velocity structure at shallow depthsfrom teleseismic events due to intrinsic attenuation and scattering along ray paths.The development of ambient noise tomography now permits high resolution shallowdepths imaging by enabling the measurement of short-period surface wave dispersionmeasurements between pairs of seismic stations without earthquake，but it is difficultto extract long-period dispersion. In this study，we applied ambient noise tomography（ANT）together with classical single-station surface wave tomography（SS），takingadvantage of the two methods to image the lithospheric group velocity down to about100km depth under the Northeast China，using the waveform data recorded by119broadband stations from Heilongjiang，Jilin，Liaoning and Inner Mongolia regionalseismic network.The data and methods are different for ANT and SS to measure Rayleigh wave group velocity dispersion curves. In this study，we first measured dispersion curvesfrom ambient noise correlation and surface wave，respectively，and then inverted themto obtain the distribution of group velocity. The data we used in ANT was the verticalcomponent of continuous seismic observations from Jan.2009to Dec.2010whoseprocessing procedure was divided into three phases：（1）single station data preprocessto obtain the ambient noise；（2）cross-correlation and temporal stack to extract theinter-station correlation function；（3）measurement of group velocity dispersion bymultiple-filter method and quality control. To SS data，we selected116seismic eventswith clear Rayleigh wave signals and obvious dispersion characteristics out of302events which occurred from Aug.2007to Jul.2008. The magnitudes of these eventsare restricted Ms5.5and greater，their epicentral distances are less than60degreesand focal depths are less than100km. After preprocessing was executed，multiple-filter method was being applied to measure Rayleigh wave group velocitydispersion curves.After the data processing，group velocity dispersion measurements at periods of5－30s and25－70s had been obtained from ambient noise correlation and surfacewave，respectively，which were used to invert for group velocity structure of thelithosphere under the Northeast China to about100km depth. We used checkerboardtest to estimate resolution. It shows the resolution of inversion results from ANTreaches to1°×1°and that from SS is2°×2°due to the sparse ray paths distribution.The group velocities at the relatively periods like25s and30s are similar for bothmethods.Our results show obvious lateral heterogeneity of the lithosphere under NortheastChina. At the short periods of8－15s，group velocities are notably correlated with thesurface geology and controlled apparently by the large faults such as Yilan-Yitongfault. The Songliao Basin exhibits the significant low velocity which suggestsrelatively thick sediment. Compared with the group velocities of8－25s periods，thedistribution of the low-velocity（low-V）and high-velocity（high-V）at periods of20－30s show reverse changes，this result reveals that the Moho depths in the westernregion of North-South gravity lineament (NSGL) are deeper than that in the eastern region. At the long periods of50－70s，with the increase of the periods，the low-V inthe eastern regions move to the west；meanwhile，the western regions show stablehigh-V features. The results may reflect asthenospheric upwelling under the easternregions of NSGL due to the subduction of Pacific slab.更多还原