【作者】 姜慧；

【导师】 胡聿贤； 赵凤新；

【作者基本信息】 中国地震局地球物理研究所 ， 固体地球物理学， 2005， 博士

【摘要】 随机模拟方法是一种建立在地震观测资料的基础上的半理论、半经验的方法。在频域内,场地的地震动傅立叶幅值谱,可以表示为震源谱、传播路径影响因子和场地效应的乘积。其优点是能反映本地区的震源、传播路径和场地特性。 大家知道,在同一地区,用中小地震确定的传播路径影响和软基岩场地效应,同样适用于强震。在强震资料有限的情况下,如何用中小地震的观测记录研究本地区的传播路径和场地影响,服务于强地面运动随机模拟,是本论文的出发点。 中小地震的震源谱可以用“布龙”理论模型表示,传播路径中的几何衰减因子可以用三段几何衰减模型表示。但传播路径的非弹性衰减和软基岩场地的放大和衰减联合效应影响难以分开,如何分离非弹性衰减和场地放大和衰减联合效应（简称“场地联合效应”）是本论文研究的关键问题。 经过对随机模拟预测方程中各因子的物理意义研究,根据“非弹性衰减因子是震源距的函数,而场地效应基本不随震源距变化”（Boore,1983,2003）这一特性,通过对大量的、不同距离的观测记录分析和反演计算,可以把软基岩的场地联合效应和全路径非弹性衰减因子在1-10Hz频率范围内分离、并分别确定出来。 本论文选择了两个研究区,分别研究了软基岩场地的联合效应和厚覆盖土层场地的联合效应。 在数字地震观测资料较为丰富的新疆北天山研究区,可以用随机模拟预测方程反演法,分离软基岩场地联合效应和非弹性衰减的影响,并把二者分别确定出来。不考虑地形效应的影响,研究区的平均软基岩场地效应,在2-4Hz频率范围为1.5倍左右,可以直接用于研究区软基岩场地的地面运动随机模拟。研究区的S波品子因子为Q_S=278f^0.346,客观地反映了研究区的非弹性衰减。在此基础上,本论文根据单拐角频率震源谱模型,对研究区中强地震的加速度反应谱进行了随机模拟。 另外,本论文用H/V谱比方法研究了软基岩台站的放大特性,分析结果表明:研究区观测记录的H/V谱比与震源距无关,能反映场地的放大特性。软基岩场地的H/V谱与参考硬基岩场地的H/V谱的比值,可近似表示软基岩场地的相对放大效应。但它们不仅反映了研究区软基岩台站场地自身放大效应,还包含了个别台站的地形效应,远大于1。 在具有丰富的中小地震强震记录和钻孔剪切波速资料的伽师-巴楚地区,用两种方法研究了厚覆盖土层场地的联合效应。首先,根据近场中小地震的强震记录及其传播特性,用地震动随机模拟预测方程估计出厚覆盖土层场地的联合效应。然后,采用联合方法-1/4波长近似法确定的场地放大效应和强震记录确定的场地Kappa衰减效应的乘积,估计出厚覆盖土层场地的联合效应。结果显示,两种方法确定的厚覆盖土层场地的联合效应很接近,在2-6Hz频率范围,厚覆盖土层场地联合效应可达3-5倍。最后,通过对研究区6-7级地震的地面运动随机模拟,修正研究区中小地震作用下的覆盖土层场地的联合效应的谱,标定出6-7级地震作用下的覆盖土层场地的联合效应。标定后的覆盖土层场地的联合效应的,可以用于研究区6-7级地震的地面运动随机模拟。更多还原

【Abstract】 Stochastic simulation method is a kind of semi-theoritical and semi-experiential method based on seismic observation data. The Fourier amplitude spectrum used in a seismological model can be expressed as a product of source spectrum, propagation path factor and site effect in frequency domain. One advantage of this approach is that it can reflect the characteristics from source, propagation path and site effects in selected area.It is usually accepted that the path effect and the soft rock site effect determined from small-moderate earthquakes are fit for strong earthquakes in the same region.This paper aims at the study of the effect of propagation path and site in selected area based on observation data from small-moderate earthquakes in case of lack of strong motion records. The results of research will serve the stochastic simulation of strong ground motions.The source spectrum of small-moderate earthquake can be represented by theoretical "Brune" model, the geometric attenuation factor of whole path can be represented by three-segment geometric model. But it’s very difficult to divide the effects of anelastic attenuation of the whole path and the effect of the soft rock site, the key problem of the paper is to separate the anelastic attenuation and the combined effect of site(i.e. the combied effect of site amplification and attenuation ).Through studying the physical meanings of every factors in prediction equation of stochastic simulation, according to the physical nature that anelastic attenuation factor is a function of source distance and site effects is almost independent of source distance(Boore,1983, 2003), the anelastic attenuation and site effects can be separated and be respectively confirmed within 1-10Hz by analyzing large of seismic data from different distance.Two research regions are chosen, i.e. to study the combined effects of soft rock site and the combined effects of thick soil cover site repectively.In the first research region lied in Northern Tian-shan of Xinjiang, there are large of digital seismic data. The combined effects of soft rock sites have been estimated using inversion method of stochastic simulation prediction equation. The average site effects of soft rock reach 1.5 or so in 2-4Hz, and can be directly applied to stochastic modelingground motions in the region. The S wave quality factor is obtained as QS = 278f 0.346,which reflects objectively the anelastic attenuation in the region. On the basis of researches mentioned above, the acceleration spectrum of moderate-strong earthquakes on soft rock sites has been simulated according to source spectrum of singnal corner frequency using stochastic method.On the other hand, the amplification characteristic is also studied using H/V spectral ratio method, the results show: (1) The H/V spectral ratios of observation data are independent of source distance, and reflect the amplification characteristic of soft rock sites. (2)The ratios of soft rock sites are greater than 1, because the ratios not only reflects the site effect themselves, but also the topographic effect of some stations.In another region in Jiashi-Bachu area, there are a lot of digital seismic data and更多还原