首都圈地区Pb震相典型特征与康拉德界面研究

Pb震相是近震震相中的一个重要震相.关于Pb震相典型特征的总结和分析,对指导区域台网工作人员如何识别该震相,丰富台网观测报告震相产出,提高地震定位精度和确定康拉德界面等相关研究具有重要意义.但由于识别困难,国内很少有系统进行有关Pb震相的识别和研究工作.本研究采用首都圈地区高密度台网2009-2015年记录到的369个ML≥2.5地震事件的波形资料,重新分析震相并识别出1153条Pb震相.基于震相资料,利用时频分析、多项式拟合、射线分析、最小二乘法、联合反演、理论走时计算等方法进行研究,并在结合前人研究结果的基础上,我们得出:在首都圈及邻区,在肉眼可识别时频域内的特征来看,通常容易识别的Pb的振幅或频率高频部分相对比初至Pn和Pg大或高,也有振幅变小或者频率变化不明显的情况,这可能与震源机制、台站方位、场地响应、仪器类型等方面有关.时频分析、功率谱密度和肉眼识别分析的结果表明,P波的主要能量集中在相对低频部分,Pn,Pb,Pg,PmP四种震相(本研究以后提到的震相顺序只考虑这四种震相)有很强的共性,区别在于传播路径上的不同,频率或观测记录周期上的小幅度差别.在Pg作为初至波时,Pb震相的低频主频部分与Pg震相的低频主频部分带宽差不多(受到包含Pg震相的影响),但是高频主频部分频率更高,Pb到时在Pg之后,PmP之前.Pn作为初至波时,Pn震相低频主频部分带宽比Pb宽,但是Pb高频主频部分频率相对更高,Pb在Pn之后,在Pg之前.鉴于震源深度对Pb到时顺序的影响,及其在定位结果中精确度最差的情况,在震中距约在80~140km范围内时,得考虑区域地壳厚度横向不均匀、震中距、震源深度等情况并结合波形特征,来判定Pb是否为初至震相.Pb震相在康拉德界面的平均传播速度约为7.0km·s^-1,康拉德界面平均深度约为23km.Pb射线的分布情况直接证明了康拉德界面在首都圈地区的分布是连续的.基于本研究利用Pg,Pb,Pn,PmP震相走时联合反演所得模型计算的理论走时结果和实际观测结果一致进一步证明了我们结果的可靠性.

Characteristics of seismic phase Pb and the Conrad interface beneath the Capital Circle region around Beijing

Seismic phase Pb is an important one in near events. Identifying this phase and summarizing its characteristics are important for enriching phase catalogs of regional networks, improving earthquake location accuracy and studying the Conrad interface. Due to many difficulties in identifying the phase, few studies have focused on this issue. This work employed waveform data from 369 M_L ≥ 2.5 local earthquakes between 2009 and 2015 recorded by the high density seismic network in the Capital Circle region around Beijing to identify 1153 Pb phases. We used a series of techniques to analyze these data, including time-frequency analysis, polynomial fitting, ray analysis, least square fitting, joint inversion, and theoretical travel time calculation and study the time-frequency features of the Pb phase and its propagating speed, and to infer the distribution of the Conrad interface. Combining with previous studies, the results show that in the study area, Pb arrivals that can be easily recognized generally have larger amplitudes and more high-frequency content than the first arrival Pn or Pg phases. There are also some cases in which the Pb phase has relatively weak amplitudes and no visible frequency changes, which can be attributed to differences in the focal mechanism, station azimuth, site response, and instrument type. Through time-frequency analysis, power spectral density and visual identification, we find that in the study area, the main energy of P-waves is concentrated in relatively low frequencies, including four main phases Pn, Pb, Pg and PmP (which are the only phases considered) share this common feature, with small variations due to different propagation paths. When Pg is the first arrival, the dominant low frequency content in the Pb phase is similar to that of the Pg phase (likely due to contamination of Pg phase), Pb, however, has relatively higher frequencies, and comes after Pg but before PmP. When Pn is the first arrival, its dominant low frequency content has broader bandwidth than the Pb, while Pb has higher frequencies and comes after Pn but before Pg. Given that the first appearance of Pb is highly influenced by source depth, and which has the least accuracy in source solutions, we suggest that for epicenter distances between 80 and 140 km, regional variations in the crust thickness, epicentral distance, source depth, and waveform characteristics must be combined to determine whether the Pb phase is the first arrival or not. The propagation speed of the Pb phase at the Conrad interface is about 7.0 km·s^-1, and the average depth of the Conrad interface is about 23 km. The distribution of the Pb ray paths suggests a continuous distribution of the Conrad interface in the study area. The theoretical travel times, calculated using the velocity model determined in a joint travel time inversion of Pg, Pb, Pn, and PmP phases, agree well with the actual observations, suggesting that our results are robust.