地表动态大地测量资料反映的孕震断层变形机制研究

发布时间：2018-09-18 08:36

【摘要】：构造地震孕育、发生、震后调整过程中应力应变积累、释放必然伴随有相应的地壳形变发生。利用大地测量动态资料研究与地震过程密切关联的地壳形变时空动态特征是地震预测研究的重要途径之一。由于构造地震发生在活动断层上(通常在10公里以下的深部),因此如何建立逼近真实的地表形变动态变化与孕震断层深浅部应力应变状态的关系以揭示孕震断层变形机理,是地震机理与预测研究中至关重要的科学问题。开展这方面的研究具有重要的科学意义。本论文首先研究了利用GPS观测资料获取与地震孕育关联的地表形变动态变化特征方法,进一步研究了深浅部地壳运动状态与断层应变积累之间的关系。主要研究内容、结果和取得认识如下：1.地壳变形动态特征分析方法研究及其在南北地震带的应用。研究解决了速度场稳定基准选取和变形频域一致等建立多期可比较速度场的关键技术问题；研究讨论了微元转动参数的基本特性及其在构造形变分析中的意义。a.利用“中国地壳运动观测网络”和“中国大陆构造环境监测网络"GPS区域站1999年～2013年7期复测资料,建立了南北地震带地区相对于稳定华南地块统一参考基准的多期速度场,基于该具有可比性速度场,对2008年汶川8.0级地震前后的地壳形变动态过程进行了分析。(1)为构建各期速度场的统一参考基准,发展了拟准检定法(QUAD法)剔除观测点群中的不稳定点,改进了初选指标稳定点选取准则,有效抑制了初算结果可能出现的基准偏移对稳定点的筛选影响,提高了结果的可靠性。(2)选择统一且适当的协方差衰减参数,利用最小二乘配置进行GPS速度场拟合推估,保证了各期速度场具有相同的变形频域,并解决了不同期资料测点分布存在差异的问题。(3)对任意两期网格化速度场求差分获取了这段时间的地表位移动态变化。汶川地震前后的两期速度场差值结果显示,在2007-2009年间汶川地震的影响范围较大,包括祁连地块、柴达木地块东部都有明显的南东东向运动响应,但距破裂带更近的鄂尔多斯西南缘响应很小,可能属于背景应力应变积累水平较高的地带,但巴颜喀拉地块北边界带东部的东昆仑断裂带左旋剪切的响应明显。汶川地震震后效应对龙门山断裂带南段表现为显著的应变加载过程,而鲜水河断裂带则表现为与应变积累背景相反的右旋扭动响应。(4)汶川地震发生后,川滇地块中北部及其边界有近东西向挤压增强,但整个川滇地块南东向挤出滑移的运动背景并没有增强,直到2011-2013时段南东向运动才有所增强。b、通过应变张量计算得到的转动参数反映计算单元的有限转动但不包含纯应变信息,与应变参数配合使用可以更客观的反映地表的变形状态。(1)该特征与剪应变参数不同,转动参数与坐标方向的选取无关,值的大小表示微小单元的转动。(2)转动参数与参考基准的选取有关。当研究多期资料时,需考虑参考基准的统一,可采用研究区域无整体旋转基准速度场进行计算。(3)转动参数反映了微元在变形中主应变轴的偏转。基于区域无整体旋转基准速度场计算的转动参数可表示最大剪应变在两个方向的不对称,从而确定区域构造运动决定的实际剪切变形最大方向。(4)对于纯走滑断层,震间位移曲线的斜率反映了转动参数的大小。当断层的变形宽度较窄时,转动参数非零值的区域较窄；当断层的变形宽度较宽时,转动参数非零值的区域较宽。当区域剪应变较大时,如果转动参数量值较大,说明断层的应变积累程度可能较高；当区域剪应变较大,而转动参数量值较小,说明断层的应变积累程度可能较低,甚至断层处于蠕滑状态。2.一般倾角断层下地表震间／同震位移场的变形特征研究。通过对特定断层的剖面分析研究,从力学特性分析了断层倾角对地表位移的影响。在原有公式基础上推导给出了带断层倾角的震间、同震的走滑、倾滑地表位移公式,该公式能够较好的拟合与断层闭锁关联的地表位移,与位错模型中的复杂公式相比是一种简化公式,但已能达到相当高的拟合精度,便于对用实测GPS速度场资料做拟合的实际应用。(1)对于非直立型走滑断层而言,震间变形中心一般不位于断层出露地表处,而是位于断层的滑动段上边沿,即断层闭锁段与滑动段的分界线在地表投影处。(2)无论走滑断层还是倾滑断层的震间形变,断层闭锁段与滑动段的分界线在地表投影与断层地表出露处之间的距离doftset、断层闭锁深度d和断层倾角δ之间存在关系tanδ=d/doffs,此关系揭示了断层的闭锁深度与断层倾角之间的内在关系。(3)由于断层倾角的影响,震间形变曲线以断层闭锁段与滑动段的分界线在地表投影为中心,同震时上下盘错动沿断层发生错动,导致上下盘同震释放的位移不对称。而当发震断层为盲断层时,地震错动不达到地表时的特例情况与震间的情况类似。(4)基于实际观测数据利用带倾角的公式拟合安宁河断裂带的倾角,得到的倾角结果小于地质考察结果,原因可能由于实际断层是曲面而非平面,在地表处断层陡峭,随着断层深度增加倾角变小。由于断层闭锁段以下滑动,对地表位移的影响,相当于断层出露地表处与断层闭锁与滑动分界之间存在一条“平面断层”对地表的作用。利用反正切函数拟合的倾角结果为这条“平面断层”的倾角。3.应用三维数值流形方法研究孕震断层深浅部力学特性。基于三维数值流形方法的连续与非连续耦合计算的优势,通过断层切割算法构建了三维模型,对孕震断层的力学特性进行研究。设置断层上部闭锁,下部滑动,对研究块体侧向和底部加载“动力源”进行走滑剪切数值试验,给出了不同加载方式下地表位移场分布差异特性。基于该结果,对汶川地震后龙门山断裂带和川滇块体东边界断裂带闭锁深度反演的动态变化给出了机理上的解释。(1)利用三维数值流形方法的断层切割算法建立了三维块体模型,块体中的断层设置为上部闭锁,下部滑动,对研究区域块体加载“推挤”和“拖曳”两种力源来模拟块体所受其他块体对其的推挤作用和脆性层底部软流物质对其拖曳作用。模拟结果显示,两种动力模式下地表位移场均呈现较精确的反正切函数特性,与位错解析结果有很好的一致性。(2)针对两种不同“力源”输入,模拟得到的位移分布曲线表现出显著差异,在一定程度上反映力源的力学特征。“推挤”力源的力学特性是沿水平方向传递,得到的反正切曲线在断层远端相对平直,曲线远端基本反映“推挤力源”的加载量,变形宽带较小,利用反正切函数拟合得到的闭锁深度小于模型设定值。“拖曳”力源的力学特性是垂直向上传递,得到的反正切曲线在断层远端略有上翘,由于力在垂直方向传递时衰减导致曲线远端不能反映“推挤力源”的加载量,曲线变形宽带较大,利用反正切函数拟合得到的闭锁深度大于模型设定值。(3)实际块体应既受到其他块体对其的推挤作用,又受到下部软流物质对其的拖曳作用。将两种力源按不同比例同时加载,得到的结果表明：哪种力源增强,地表形变动态变化就体现出此力源加载下的曲线特征。基于此认识可以对汶川地震之后川滇地区地壳深浅部运动特征和DEFNODE负位错反演动态结果给出可能的解释。汶川地震之后,龙门山断裂带南段西侧的巴颜喀拉块体南东向运动增强,是由于龙门山断裂带脆性层同震释放所导致,可理解为块体的“推挤”作用,所以反演得到的龙门山断裂带南段闭锁深度变浅。上部脆性层的突然加速对下部软流物质提供加载动力。获得加载动力的巴颜喀拉块体下部软流物质由通道进入川滇块体下部,对川滇块体上部脆性层提供了“拖曳”力源,由此计算的小江断裂闭锁深度增加。(4)“推挤”与“拖曳”力源计算得到断层不同深度的位移结果均表明断层的闭锁段以下位移量随深度逐渐加大,并非位错理论中闭锁与滑动的突变分界而是渐变的过程,这与Tse and Rice (1986)、Scholz (1998)的模型的示意图具有一致性。(5)考虑下部软流物质应为渐变的可能性,对“拖曳”力源进行改进,使之逐渐加载,加载量与距断层距离呈现单调递增函数关系。模拟得到的地表位移结果表明,渐变“拖曳”力源的加载使得地表位移曲线相比于无渐变“拖曳”力源得到的变形宽度更大,因此实际断层由于可能存在软流物质渐变拖曳而使得地表变形宽度更宽。(6)南北地震带中段三维数值流形模拟结果表明,在弹性本构模型下水平主应力增量和水平主应变增量的方向具有一致性,但主张应力增量与主压应力增量的比值和主张应变增量与主压应变增量的比值存在差异。汶川地震前龙门山断裂带处于挤压面应力积累状态,汶川震源区处于该挤压应力速率高值区的边缘,以汶川震源区为界龙门山断裂带南段的挤压应力速率明显快于龙门山断裂带北东段。汶川地震后的2009-2013年龙门山断裂带中北段处于震后调整过程,而龙门山断裂带南段则处于应力积累状态。鲜水河断裂带中北段以张性应力增量为主,而其中南段以张剪应力增量为主。芦山地震震源区处于龙门山断裂带中北段和鲜水河断裂带中南段最大剪应力快速积累区的弱化地带。安宁河断裂带的最大剪应力积累速率小于其北侧的鲜水河断裂带南段以及其南侧的则木河断裂带。总体而言,本文通过定量描述地表形变动态变化特征；从理论上研究断层倾角对震间和同震的地表位移影响；利用三维数值流形方法研究断层深浅部力学特性对地表位移的影响,分析了地壳动态变形特征及其反映的孕震断层力学特性,对孕震断层应变积累状态与地表位移的动态变化之间关系有了初步认识。为今后进一步结合不同期GPS资料,利用三维数值流形方法研究实际断层不同时间段地表形变动态变化的深部动力加载影响、科学揭示孕震断层闭锁与深浅部应力应变状态奠定了理论基础。
[Abstract]:The accumulation and release of stress and strain in the process of tectonic earthquake preparation, occurrence and post-earthquake adjustment must be accompanied by corresponding crustal deformation. So how to establish the relationship between the dynamic change of surface deformation and the stress-strain state of the shallow and deep part of the fault to reveal the deformation mechanism of the fault is a very important scientific problem in the study of earthquake mechanism and prediction. Firstly, the method of acquiring dynamic characteristics of surface deformation associated with earthquake preparation by GPS observation data is studied, and the relationship between crustal movement and fault strain accumulation is further studied. The key technical problems of establishing multi-period comparable velocity field, such as selection of stable datum of velocity field and uniformity of deformation frequency domain, are studied and solved. The basic characteristics of micro-element rotation parameters and their significance in structural deformation analysis are discussed. Based on the comparable velocity field, the dynamic process of crustal deformation before and after the 2008 Wenchuan M8.0 earthquake is analyzed. (1) In order to construct the unified velocity field of each period, the multi-period velocity field in the North-South seismic zone is established relative to the unified reference datum of the stable South China block. Reference datum, Quasi-Accurate calibration (QUAD) method is developed to eliminate the unstable points in the observation point group, and the selection criterion of the stable points of the primary selection index is improved. The influence of the datum offset that may appear in the initial calculation results on the selection of the stable points is effectively suppressed, and the reliability of the results is improved. (2) Unified and appropriate covariance attenuation parameters are selected, and the minimum is used. Second-order collocation is used to fit and estimate the GPS velocity field, which ensures that the velocity field of each period has the same deformation frequency domain, and solves the problem that the distribution of the observation points in different periods is different. (3) The dynamic variation of the surface displacement in this period is obtained by calculating the difference of the grid velocity field of any two periods. The results show that the Wenchuan Earthquake in 2007-2009 has a large influence range, including the Qilian Block and the eastern Qaidam Block, which have obvious southeast-east movement response, but the southwestern margin of Ordos, which is nearer to the rupture zone, has a small response and may belong to the zone with higher background stress-strain accumulation, but the eastern part of the northern boundary zone of the Bayanhar Block. The response of the East Kunlun fault zone to sinistral shear is obvious. The post-earthquake effect of Wenchuan earthquake on the southern section of Longmenshan fault zone shows a significant strain loading process, while the Xianshuihe fault zone shows a dextral torsion response contrary to the background of strain accumulation. (4) After the Wenchuan earthquake, the north-central Sichuan-Yunnan block and its boundary have nearly E-W compression and increase. However, the movement background of the whole Chuan-Yun block is not enhanced until 2011-2013. B. The rotation parameters calculated by the strain tensor reflect the finite rotation of the calculation element but do not contain the pure strain information. The combination of the rotation parameters and the strain parameters can reflect the surface more objectively. (2) The rotational parameters are related to the selection of reference datum. When studying multi-period data, the unification of reference datum should be considered, and the velocity field without reference datum in the study area can be calculated. (3) The rotational parameters reflect the deflection of the principal strain axis of the element in deformation. The rotational parameters calculated based on the region without global rotational reference velocity field can indicate the asymmetry of the maximum shear strain in two directions, thus determining the maximum direction of actual shear deformation determined by regional tectonic movement. (4) For pure strike-slip faults, the obliquity of the interseismic displacement curve When the deformation width of the fault is narrow, the region with non-zero rotation parameters is narrower; when the deformation width of the fault is wide, the region with non-zero rotation parameters is wider. The larger strain and the smaller rotational parameter indicate that the strain accumulation of faults may be lower, even the faults are in creep state. 2. Study on deformation characteristics of surface inter-seismic/co-seismic displacement field under general dip-angle faults. On the basis of the original formula, the formulas of the earth surface displacement with fault dip angle, strike slip and dip slip are deduced and given. The formulas can well fit the ground surface displacement associated with fault locking. Compared with the complex formulas in the dislocation model, the formulas are simplified, but they have reached fairly high fitting accuracy and are convenient for the GPS velocity measurement. (1) For non-vertical strike-slip faults, the center of interseismic deformation is not located on the surface of the fault, but on the upper edge of the fault slip section, i.e. the boundary between the fault block and the slip section is projected on the surface. (2) Whether the strike-slip fault or the dip-slip fault, the fault is closed. Doftset is the distance between surface projection and outcrop of faults. There is a relationship between fault locking depth D and fault dip angle delta. This relationship reveals the intrinsic relationship between fault locking depth and fault dip angle. (3) Interseismic deformation curve is closed by fault due to the influence of fault dip angle. The boundary between lock section and sliding section is centered on the surface projection, and the dislocation of upper and lower wall occurs along the fault during the same earthquake, resulting in the dissymmetry of the displacement released by the same earthquake. The dip angle of the Anning River fault zone is smaller than that of the geological investigation. The reason may be that the actual fault is a curved surface rather than a plane, the fault is steep on the surface, and the dip angle decreases with the increase of the depth of the fault. There is a "plane fault" acting on the earth's surface between the lock and the slip boundary. The dip angle of the "plane fault" is obtained by fitting the tangent function. 3. The mechanical properties of the depth and shallowness of the seismogenic fault are studied by using the three-dimensional numerical manifold method. A three-dimensional model is constructed by using fault cutting algorithm to study the mechanical properties of seismogenic faults.The upper part of the fault is locked and the lower part slides.The numerical tests of strike-slip shear are carried out on the "power source" loaded laterally and at the bottom of the block.The distribution characteristics of surface displacement field under different loading modes are given.Based on the results,the Wenchuan area is studied. The dynamic change of the inversion of the locking depth of the Longmenshan fault zone and the eastern boundary fault zone of the Sichuan-Yunnan block after the earthquake is explained in mechanism. (1) A three-dimensional block model is established by using the fault cutting algorithm of the three-dimensional numerical manifold method. The fault in the block is set up as upper locking and lower sliding, and the block in the study area is loaded with "push" and "push". The simulation results show that the surface displacement field under the two dynamic modes exhibits more precise arc-tangent function characteristics, which is in good agreement with the results of dislocation analysis. (2) Two different "force sources" are used to simulate the pushing action of the block on the block and the dragging action of the soft flow material at the bottom of the brittle layer. The mechanical characteristics of the push force source are transmitted along the horizontal direction, and the arc-tangent curve is relatively straight at the far end of the fault. The far end of the curve basically reflects the load of the push force source, and the deformation broadband is small. The locking depth obtained by fitting the arc-tangent function is less than the set value of the model. The latch-up depth obtained by fitting the arc-tangent function is greater than the set value of the model. (3) The actual block should be pushed by other blocks and dragged by the lower asthenospheric material. Based on this understanding, it is possible to explain the characteristics of crustal movement in the deep and shallow part of Sichuan-Yunnan region and the results of DEFNODE negative dislocation inversion after Wenchuan earthquake. Because of the co-seismic release in the upper brittle layer, the locking depth in the southern section of the Longmenshan fault zone becomes shallower. The sudden acceleration of the upper brittle layer provides the loading power for the lower asthenospheric material. The upper brittle layer of the Yunnan block provides a "drag" force source, and the locking depth of the Xiaojiang fault calculated by this method increases. (4) The displacement results of different depths of the fault calculated by "push" and "drag" force sources show that the displacement below the locking section of the fault increases with the depth, not the sudden boundary between locking and sliding in the dislocation theory. It is a gradual process, which is consistent with the schematic diagram of Tse and Rice (1986) and Scholz (1998). (5) Considering the possibility that the underlying asthenosphere material should be gradual change, the drag force source is improved to gradually load, and the load and the distance from the fault show a monotonically increasing function relationship. The loading of the gradient dragging force source makes the deformation width of the surface displacement curve larger than that of the non-gradient dragging force source, so the deformation width of the actual fault is wider because of the possible existence of the gradient dragging of the asthenic material. (6) Three-dimensional numerical manifold simulation results of the mid-section of the North-South seismic belt show that the elastic constitutive model is applicable to the elastic constitutive model. The direction of the increment of the horizontal principal stress and the increment of the horizontal principal strain are consistent, but the ratio of the assertive stress increment to the increment of the principal compressive stress and the ratio of the assertive strain increment to the increment of the principal compressive strain are different. At the margin of the high value area, the compressive stress rate in the southern section of the Longmenshan fault zone bounded by the Wenchuan earthquake source region is obviously faster than that in the northeastern section of the Longmenshan fault zone. The focal area of the Lushan earthquake is in the weakening zone of the fast accumulation area of the maximum shear stress in the middle-north section of the Longmenshan fault zone and the middle-south section of the Xianshuihe fault zone. The Zemuhe fault zone on the side. Generally speaking, this paper quantitatively describes the dynamic change characteristics of surface deformation, and theoretically studies the dip angle of the fault on the interseismic sum.
【学位授予单位】：中国地震局地质研究所
【学位级别】：博士
【学位授予年份】：2015
【分类号】：P315.2

【相似文献】