断层破裂传播速度与破坏性地震的烈度分布——浅源地震的断层不均匀性与短周期波动的多普勒效应

地震的等震线不仅反映了地震的大小,而且也反映了地震断层过程的类型和破裂速度.单侧破裂的断层过程呈现出蛋型的等震线,而双侧破裂的断层过程呈现出椭圆的等震线,等震线的长轴与短轴之比对于破裂传播速度十分敏感.本研究利用理论等震线与观测结果相比较的方法,确定了1964年日本新地震、1983年日本海中部地震、1975年中国海城地震和1976年中国唐山地震的破裂传播速度、断层类型以及断层走向.得到的断层破裂速度是剪切波的0.7—0.9倍.这些值比用长周期地震波所确定的要稍大.产生这种差异的原因是:对烈度起主要影响的短周期地震波强烈地依赖于小尺度的断层不均匀破裂过程,以及局部的破裂传播的开始和终止;而由长周期波所得到的破裂速度却反映了在整个断层上破裂传播的平均过程.根据等震线图所得出的断层类型及断层走向与其它独立方法的结果相一致.这意味着本方法可以应用于推断某些历史地震的断层类型、破裂传播速度及破裂传播方向.     

基于动摩擦应力下调滑移弱化机制，利用地表破裂同震位移约束2001年昆仑山口西MS8.1地震破裂传播速度

利用2001年昆仑山口西MS8.1地震现场观测所提供的地表破裂同震位移数据,使用简单滑移弱化破裂模型,估算了发震主断层上的破裂传播速度. 该模型中考虑了断层破裂时动摩擦过程中应力上调和下调机制对地震波辐射能量分配的影响. 对比Bouchon和Valleacute;e有关昆仑山口西地震主断层破裂传播速度超过剪切波速度,甚至达到P波速度的结果, 采用动摩擦应力下调时的滑移弱化模型 (分数应力降模型),结果表明,伴随较高的地震波辐射效率,主断层的平均破裂传播速度等于或小于瑞利波速度,这与许力生和陈运泰的体波反演结果,以及陈学忠震源应力场估算的结果是一致的. 最后,联系到由地表破裂现象所反映出的断层力学特征,如与视应力相关的分数应力降 (动摩擦应力下调), 基于滑移弱化模型, 讨论了可能的震源破裂机制.      

断层自发破裂动力学过程的有限单元法模拟及其在地震研究中的应用

地震是断层的自发破裂动力学过程。数值模拟断层的自发破裂动力学过程对于认识地震的力学本质、减轻地震灾害等有着重要的科学意义及应用价值。本文首先对经典的滑移弱化摩擦关系进行了改进,然后对断层的破裂过程进行动态数值模拟。模拟结果表明,利用改进后的摩擦关系能够产生脉冲型(pulse-like)破裂模式。断层自发破裂过程受初始应力场及摩擦关系影响,若初始应力场中的剪应力水平较低或滑移弱化摩擦本构关系中的动摩擦系数较大,则容易产生脉冲型破裂;反之,则容易产生裂纹型(crack-like)破裂。另外,为了研究双材料(bimaterial)断层破裂对强地面运动的影响,我们采用正则化的速率-状态相关摩擦本构关系计算了破裂沿着双材料断层传播的二维有限元模型。模拟结果表明,双材料机制对地震破裂过程以及断层周边区域的强地面运动有显著影响。由断层破裂辐射出的地震波导致的强地面运动在整个空间上的分布是不对称的,其不对称性会随着断层两侧材料差异程度的增加而增加。断层破裂能否跨越断层阶区(stepover)继续传播,从而引发更大震级的地震,地震时断层是否发生超剪切破裂导致地震灾害加剧,都是震源动力学研究的重要内容。本文利用有限单元方法模拟断层阶区对地震破裂传播的控制作用以及对产生超剪切地震破裂的促进作用。研究结果表明:断层面上的摩擦系数减小、断层周边区域内初始剪应力增大以及较小的阶区间距等,都将增加断层破裂跳跃阶区传播的可能性;此外,这些物理因素都会对破裂的传播速度产生影响。在一定条件下,破裂传播速度会由在初始断层上的亚剪切波速度转为在次级断层上的超剪切波速度。结合以上在概念模型中对断层自发破裂过程的模拟研究结果,我们根据汶川地震和玉树地震发震断层的实际几何情况分别构建有限单元数值模型,研究了汶川地震单侧破裂过程的动力学机制以及玉树地震产生超剪切破裂过程的动力学机制。2008年汶川大地震的破裂过程极其复杂,向东北方向的破裂距离长达300 km,而向西南方向的破裂长度很小,呈现出单侧破裂的主要特征。文中模拟并分析了汶川地震的破裂过程,结果表明:龙门山断裂带两侧的物性差异是造成汶川大地震单侧传播的决定性因素。由于2010年玉树地震(Ms=7.1)产生了超剪切地震破裂,所以地震灾害特别严重。文中在模拟并分析玉树地震的破裂过程后认为:玉树地震发震断层走向与初始主应力方向之间的关系断层破裂是亚剪切转化为超剪切破裂的可能原因。     

远场地震波辐射能计算以及对近断层地面运动的约束

从圆盘断层模型出发,根据地震波能量表象定理推导出了滑移弱化过程中远场S-波辐射能量表达式,并同已有的动力学模型作了比较.结果表明,得到的模型能量值或视应力的取值强烈地依赖于断层上的动态、静态应力降和破裂传播速度,而破裂速度则对应了断层带模型中断层破坏过程所耗散的能量.动摩擦应力上调和应力下调力学机制在能量求解中得到了充...     

安徽两次中等地震发震断层参数测定

根据地震波传播过程中的多普勒效应。利用拐角频率的方向变化,用最小二乘法反演了2005年寿县ML4．1级、2006定远ML4.7级地震的发震断层参数。结果表明,反演得到的中等地震的发震断层参数与震中附近实际断层的参数基本一致,破裂速度均小于S波的速度,且破裂长度为0．2-0．3km。     

断层阶区对产生超剪切地震破裂的促进作用

地震时若断层发生超剪切破裂,地震灾害会显著加剧.因此研究超剪切破裂的形成机理有着非常重要的科学意义.本文利用动力有限单元方法,模拟断层破裂从初始断层跳跃传播到另一条平行的次级断层(断层阶区)时破裂速度的变化情况,并分析断层阶区几何特征等物理参数对产生超剪切地震破裂的促进作用.计算结果表明,断层阶区的诸多物理因素(如:重叠长度、相隔距离以及摩擦系数等)都会对破裂的传播速度产生影响.在一定条件下,破裂传播速度会由在初始断层上的亚剪切波速度,转换为在次级断层上的超剪切波速度.在破裂速度转换过程中,断层间隔起着重要作用,当断层阶区中两断层垂直间隔距离小到一定程度时,破裂跳跃阶区后,破裂速度不会发生变化.所以对于分段断层(可视为一种特殊的断层阶区),由于其断层垂直间隔为0,也就不会出现破裂速度变化的现象,模拟结果对此也进行了证实.然而,若断层间隔太大,当其距离超过一定的限度后,破裂通常无法跨越断层阶区继续传播,而是终止在初始断层上.模拟结果还表明,初始断层与次级断层之间的重叠距离也十分重要,只有当断层阶区中两平行断层之间的重叠部分达到一定长度后,断层的破裂速度才有可能发生转换.此外,计算结果显示,破裂过程中断层面上的应力变化可能是破裂传播速度发生转换的直接原因.最后,模拟还发现,当破裂跨越断层阶区发生速度转换时,破裂需要停顿一定的时间,以便积聚足够的能量来实现破裂速度的增快.     

最小能量准则用于1995年河北沙城ML 4.1地震序列破裂性质的探讨

运用变分原理,我们得到了最小地震波辐射能量约束准则并用于研究震源的物理过程.通过研究1995年M_L4.1河北沙城地震序列主震和余震的动力学过程,可知主震和余震震源的动态破裂过程明显不同;M_L4.1主震的破裂速度与瑞利波速相近,约为剪切波速度的0.89倍;而28个余震的破裂速度远远小于剪切波速度,大约是剪切波速度的0.05到0.55倍.根据裂纹扩展模型,计算得到其余震的地震波辐射效率多在10％以下,这也说明了余震的地震效率较低.我们认为余震震源的动态破裂过程应与断层内部新生裂纹的扩展有关,而非简单的岩体间的相对滑动.余震震源的动态破裂传播与破裂能占主导地位的小地震有关.这些小震所带来的破裂能也导致了断层的进一步扩展.在对该地震序列的研究中,我们发现主震与余震的震源破裂过程在能量分配上有着本质的区别.因此当地震断层尺度相当小时,破裂能的贡献不能忽略,它的大小将显著地影响地震波辐射能的大小.     

最小能量准则用于1995年河北沙城ML 4.1地震序列破裂性质的探讨

运用变分原理,我们得到了最小地震波辐射能量约束准则并用于研究震源的物理过程.通过研究1995年ML4.1河北沙城地震序列主震和余震的动力学过程,可知主震和余震震源的动态破裂过程明显不同;ML4.1主震的破裂速度与瑞利波速相近,约为剪切波速度的0.89倍;而28个余震的破裂速度远远小于剪切波速度,大约是剪切波速度的0.05到0.55倍.根据裂纹扩展模型,计算得到其余震的地震波辐射效率多在10%以下,这也说明了余震的地震效率较低.我们认为余震震源的动态破裂过程应与断层内部新生裂纹的扩展有关,而非简单的岩体间的相对滑动.余震震源的动态破裂传播与破裂能占主导地位的小地震有关.这些小震所带来的破裂能也导致了断层的进一步扩展.在对该地震序列的研究中,我们发现主震与余震的震源破裂过程在能量分配上有着本质的区别.因此当地震断层尺度相当小时,破裂能的贡献不能忽略,它的大小将显著地影响地震波辐射能的大小.     

断层破裂速度对地震动及其地震灾害影响的有限单元法模拟

断层破裂传播速度通常会影响强地面运动加速度的空间分布及地震的灾害程度,但究竟是如何影响的,目前未见全面的定量分析.为此,本研究设断层破裂速度从小到大发生改变(从亚剪切波速度到超剪切破裂速度),同时利用有限元方法计算在不同破裂传播速度的情况下,破裂产生的地震动及强地面运动峰值加速度(PGA)的空间分布,计算时保持在所有情...     

2010年青海玉树Ms7.1地震近断层地面运动估计

2010年4月14日在我国青海省玉树藏族自治州发生了M_s7.1级地震.地震给当地群众的生命财产带来巨大损失,其中最严重的破坏发生在震中东南的玉树县城.我们利用反演远场地震波形得到的震源破裂模型和Crust2.0提供的当地的地壳模型,计算了近断层地面运动（速度场和位移场）,再现了地震发生时当地的地面运动图像.根据这个图像,震源的单侧破裂过程导致了明显的多普勒效应.地震波自震中向四周扩散开去,但沿着断层向东南方向传播的地震波的幅度更大、波峰更密.玉树县城恰好位于地面静态位移较大的区域,也位于较强烈的位移波和速度波穿越的地带.与实际的地面运动相比,虽然我们的计算还不够精确,但无容置疑,破裂过程导致的多普勒效应是玉树县城遭受严重破坏的主要原因.     

Seismic intensity distribution of shallow earthquakes due to rupture velocities and faulting modes

Isoseismals of seismic intensity distributions are represented by earthquake source size, faulting mode, and rupture velocity of fault propagation. Unilateral faulting forms egg—shaped isoseimals, while bilateral faulting forms elliptical ones. It is found that the ratio of major to minor axes of isoseismals is sensitive to rupture velocity. Rupture velocity, faulting mode, and fault trend have been determined from the seismic intensity maps of the 1964 Niigata and the 1983 Japan Sea earthquakes in China by matching theoretical isoseismals. Rupture velocities thus estimated are mostly 70% to 90% of shear wave velocity. The difference would be considered as follows: short—waves which determine the seismic intensity are strongly dependent on the rupture of small—scale fault heterogeneities and on the jerky—onsets and abrupt terminations of local rupture propagations. On the other hand, rupture velocity from long—waves represents an average rupture propagation along the whole fault length. Faulting mode and fault trend estimated from seismic intensity maps match with each earthquake faulting independently determined. This suggests that the present method would be applicable to some historical earthquakes with known seismic intensity distribution to obtain detailed information on the faulting process.     

震源机制与烈度分布的关系（二）

除了通常熟知的震源深度和介质吸收系数对烈度分布有较大影响外,震源的破裂方式（单侧或双侧）,破裂速度的大小和方向,断层长度,断层面倾角和错动方式都对烈度分布有较大影响。尤其是破裂速度的大小和方向对烈度分布的影响更大。破裂速度大,而且方向向上,可使震中烈度增高一度多,极震区面积扩大1-2倍;反之,可使震中裂度降低一度、极震区面积减小许多。 根据震源参数对烈度分布影响的分析,可由历史地震的等震线反推其震源参数。 文中试行反推了三个历史地震的震源参数。     

Fracture characteristics of the 1997 Jiashi, Xinjiang, China, earthquake swarm inferred from source spectra

Broadband P and S waves source spectra of 12 MS5.0 earthquakes of the 1997 Jiashi, Xinjiang, China, earthquake swarm recorded at 13 GDSN stations have been analyzed. Rupture size and static stress drop of these earthquakes have been estimated through measuring the corner frequency of the source spectra. Direction of rupture propagation of the earthquake faulting has also been inferred from the azimuthal variation of the corner frequency. The main results are as follows: ①The rupture size of MS6.0 strong earthquakes is in the range of 10～20 km, while that of MS=5.0～5.5 earthquakes is 6～10 km.② The static stress drop of the swarm earthquakes is rather low, being of the order of 0.1 MPa. This implies that the deformation release rate in the source region may be low. ③ Stress drop of the earthquakes appears to be proportional to their seismic moment, and also to be dependent on their focal mechanism. The stress drop of normal faulting earthquakes is usually lower than that of strike-slip type earthquakes. ④ For each MS6.0 earthquake there exists an apparent azimuthal variation of the corner frequencies. Azimuthally variation pattern of corner frequencies of different earthquakes shows that the source rupture pattern of the Jiashi earthquake swarm is complex and no uniform rupture expanding direction exists.     

Synthetic isoseismals of three earthquakes in California and Nevada

Recent tests on a series of earthquakes in California and Nevada suggest that in some regions the approximate shapes of the territories with equal earthquake-induced damage (expressed in terms of macroseismic intensity) could be synthetically traced out with a simple formula. This formula takes into account some gross features of the source: depth and length, unilateral or bilateral rupture, radiation patterns, rupture velocity, and directivity. Having been formulated on an empirical basis, the formula is however compatible with the so-called asymptotic approach, in which the far-field component of the Green's function is used. This paper presents the synthetic isoseismals of the earthquakes at Cedar Mountain, Nevada, 1932; Fairview Peak-Dixie Valley, Nevada, 1954; and Coalinga, California, 1983. An overall consistency, from acceptable to remarkable, between the observed intensity patterns and the synthetically back-predicted intensity has been obtained for them. Where the detailed modelling techniques available today are inapplicable, due to insufficient information on the features of the seismic sources, or to save time and money, the new formula may be utilizable for improving seismic hazard calculations.The formula was also used inversely for back-predicting geometric-kinematic parameters of the Coalinga 1983 earthquake from macroseismic maps. This gave characteristics for its source which are in good agreement with the majority of data inferred from modelling and from analyzing modern instrumental recordings. This striking result opens new perspectives in retrieving information on the source of ancient earthquakes for which only macroseismic information is available.     

基于实时观测数据快速判定地震破裂方向的方法研究

针对地震烈度速报工作中快速评估大震断层破裂方向的需求,使用我国西部多个地震的观测资料,根据有限移动源理论,利用趋势面分析方法探讨了震中一定范围内台站初始P波参数的方向差异特征,提出一种利用初始P波快速判定地震破裂断层方向的方法。并通过断层距模型和最小二乘法,统计回归最优拟合参数以确定地震断层破裂方式。通过实际震例对比分析了根据所提方法获得的判定结果与实际地震断层破裂分布的吻合度,证实了方法的可用性。本研究为利用地震实时监测资料快速进行地震断层破裂方向判定提供了新思路,可为地震烈度速报中破裂方向的快速判定和动态修正提供参考。     

Apparent stress scaling for tectonic and induced seismicity: Model and observations

A macroscopic model of seismic sources provides a scaling relationship for the apparent stress, treated as a function of three independent parameters: seismic moment, rupture area size, and average slip acceleration. These parameters represent three different factors: kinematic, geometric and material. This relationship allows us to distinguish and explain the following statistical characteristics of the log apparent stress versus log seismic moment plot. The regional trends, represented by a series of 1/2 slope lines, are related to the averaged shape of slip velocity pulses, so they reflect kinematic characteristics of the rupture process. The global trend, represented by the 1/6 slope line, is expected to characterize sets of events of wide range of rupture area sizes and assumes dependence of rupture area size on total slip, so it is related to the rupture initiation, propagation and arrest conditions; therefore, it reflects earthquake rupture dynamics. Additional shiftings among the trend lines obtained for the smallest induced tremors, larger tectonic earthquakes, and slow tsunami earthquakes, reflect differences between the intact rock failure and the frictional slip failure, that is, between fracture energies of these different earthquake classes.     

1988年云南澜沧-耿马地震的烈度分布及地表破裂

1988年11月6日,在云南省西南部的澜沧-耿马断裂带上发生了两次大于7级的地震.地震造成的严重破坏和人员伤亡主要是由于极震区内抗震性能极差的毛石房、砖柱土坯房的大量倒塌所致.澜沧地震的震中基本烈度可达Ⅸ度,耿马地震极震区烈度达Ⅹ度.澜沧地震构造活动的地表证据主要是出现在极震区内的张性地裂缝带和小断层陡坎.地裂缝带和小断坎主要出现在四条相对连续的北北西走向的狭窄地带内,其长度从几百 m 到6km 不等.澜沧地震地表破裂带长约35km,宽约3km,最大垂直位移量和最大右旋水平位移量分别为1.5m 和1.4m.耿马地震地表断裂活动的明显证据包括一系列北北西走向的地裂缝带和一段长约5km的地震断层陡坎.耿马地震地表破裂带长约24km,其最大垂直位移3.5m,最大右旋水平位移为3m.两次地震形成的地表破裂带均具有右旋-正断层性质.本文讨论了这两次地震的 度分布及地表破裂特征.      

Paleoseismic study of the Kamishiro Fault on the northern segment of the Itoigawa–Shizuoka Tectonic Line,Japan

The M_w 6.2 (Mj 6.8) Nagano (Japan) earthquake of 22 November 2014 produced a 9.3-km long surface rupture zone with a thrust-dominated displacement of up to 1.5 m, which duplicated the pre-existing Kamishiro Fault along the Itoigawa–Shizuoka Tectonic Line (ISTL), the plate-boundary between the Eurasian and North American plates, northern Nagano Prefecture, central Japan. To characterize the activity of the seismogenic fault zone, we conducted a paleoseismic study of the Kamishiro Fault. Field investigations and trench excavations revealed that seven morphogenic paleohistorical earthquakes (E2–E8) prior to the 2014 M_w 6.2 Nagano earthquake (E1) have occurred on the Kamishiro Fault during the last ca. 6000 years. Three of these events (E2–E4) are well constrained and correspond to historical earthquakes occurring in the last ca. 1200 years. This suggests an average recurrence interval of ca. 300–400 years on the seismogenic fault of the 2014 Kamishiro earthquake in the past 1200 years. The most recent event prior to the 2014 earthquakes (E1) is E2 and the penultimate and antepenultimate faulting events are E3 and E4, respectively. The penultimate faulting event (E3) occurred during the period of AD 1800–1400 and is associated with the 1791 M_w 6.8 earthquake. The antepenultimate faulting event (E4) is inferred to have occurred during the period of ca. AD 1000–700, likely corresponding to the AD 841 M_w 6.5 earthquake. The oldest faulting event (E8) in the study area is thought to have occurred during the period of ca. 5600–6000 years. The throw rate during the early Holocene is estimated to be 1.2–3.3 mm/a (average, 2.2 mm/a) with an average amount of characteristic offset of 0.7–1.1 m produced by individual event. When compared with active intraplate faults on Honshu Island, Japan, these slip rates and recurrence interval estimated for morphogenic earthquakes on the Kamishiro Fault along the ISTL appear high and short, respectively. This indicates that present activity on this fault is closely related to seismic faulting along the plate boundary between the Eurasian and North American plates.     

Dynamic faulting under rate-dependent friction

We discuss the effects of rate-dependent friction on the propagation of seismic rupture on active faults. Several physicists using Burridge and Knopoff's box and spring model of faulting have proposed that fault complexity may arise from the spontaneous development of a self-similar stress distribution on the fault plane. If this model proves to be correct, it has important consequences for the origin of the complexity of seismic sources. In order to test these ideas on a more realistic earthquake model, we developed a new boundary integral equation method for studying rupture propagation along an antiplane fault in the presence of nonlinear rate-dependent friction. We study rupture dynamics of models with single and twin asperities. In our models, asperities are places on the fault with a higher value of prestress. Othewise all fault parameters are homogeneous. We show that for models with such asperities, a slip velocity weakening friction leads to the propagation of supersonic healing phases and to the spontaneous arrest of fracture if the prestress outside the asperities is low enough. For models with asperities, we can also observe narrow slip velocity pulses, qualitatively similar to the so-called Heaton pulses observed in some earthquake accelerograms. We also observe a complex distribution of stress after the rupture that depends on details of the initial distribution of asperities and on the details of the friction law.     

地震预警断层参数实时识别方法(FinDer)详解及其性能初步评价

详细阐述了美国加州ShakeAlert地震预警系统中使用的断层破裂参数实时测定方法(FinDer FinDer:Finite Fault Rupture Detector)的技术细节.利用1999年集集Mw7.6级地震、2008年汶川Ms8.0级地震及2013年芦山Ms7.0级地震,以及日本K-NET台网记录的80次中...