滇西实验场微震衰减与场地响应

MICROSEISMIC ATTENUATION AND SITE EFFECT IN WESTERN YUNNAN EARTHQUAKE PREDICTION TEST AREA

本文使用滇西实验场区PDR—2数字化台网的77个微震(M_L<3)资料,分析了场区内的下关、洱源、邓川、双廊、乔后和漾濞6个台站的微震振幅谱。采用振幅谱频率比方法,研究区域衰减效应和局部场地响应。在一固定接收器场地条件下对不同距离不同方向的几十个地震,测量5Hz和20Hz谱振幅比,估计场区内地壳上部P波和S波的Q值。6个台站分别属于不同的构造背景,各台Qp和Qs值差异与局部区域构造差异强烈相关,在所研究的小区域范围内衰减机制和场地响应差异可分为三个条带状区域。6个台站的平均Qp为210,Qs为460。Qs≈2.2Qp,与Frankel估计的Qp≈Qs对比,表明地壳上部的衰减机制存在明显的地区性差异。场地响应的另一主要特征是谱的拐角频率,对于浅源微震拐角频率不是震源距和地震位置的函数,统计结果表明,各场地具有明显的优势拐角频率,它与局部场地构造差异相关。

In this paper, according to the data of 77 earthquakes (ML<3 ) recorded by the PDR-2 dugital recorder network in the Western Yunnan Earthquake Predietion Test Area, the author has studied the micrososmic amplitude spectra at Xiaguan, Er-yuan, Dengchuan, Shuanglang, Qiaohou and Yangbi seismo-logical stations located in the area. Furthermore, with the method of amplitude ratio with different frequency selected, the areal attenuation effect and local site response have been studied. According to a few dozens of microearthquakes recorded by a PDR-2 recorder fixed at a station (Each event's epicentral distance and direction are different), the spectral amplitude ratio (A20Hz/A5Hz) has been calculated respectively, and the Q values of P and S wave in the upper crust in the area estimated by using the initial parts of P and S wave. Each of the six stations is located on a unique fault zone. The Qp and Qs values of each station are closely related to the areal structure on which the station is located. The studied area is divided into three zones, each zon'e's attenuation mechanism and site effect being different. The average Qp, value of the six stations is 210, and the Qs value 460 (Qs≈2.2 Qp). As compared with Frankel's result(Qp≈Qs), it is shown that there exists the obvious areal difference of attenuation mechanism in the upper crust. Another main feature of site response is shown by the spectral corner frequency. With regard to a shallow focus microearthquake, its corner frequency is not the function of focal distance and position. The statistical results show that each event has its obvious pre-dominant corner frequency which has nothing to do with its local site structure.