双差地震定位法误差定量分析*

(1.中国地震局第二监测中心,陕西 西安 710054; 2.长安大学 地质工程与测绘学院,陕西 西安 710054)

双差重定位; 误差分析; 初始震源; 台站分布; 震源对

Error Quantities Analysis on Double-difference Earthquake Location Algorithm
ZHAO Yun-Feng1, WANG Wei-Dong2, ZHU Yi-Qing1, LIU Fang1, BIAN Long2

(1.Second Crust Monitoring and Application Center, CEA, Xi'an 710054, Shannxi, China)(2.School of Geology Engineering and Geomatics, Chang'an University, Xi'an 710054, Shannxi, China)

double-difference location; error analysis; initial hypocenter; station distribution; focal pcrirs

备注

利用双层地壳模型,在下地壳中构造若干地震事件构成事件簇,分别在真实震源位置中添加不同程度的噪声后作为初始震源位置进行重定位,重定位中分别使用均匀分布和仅分布于事件簇一侧的台站,分析了初始震源位置和台站分布情况对双差重定位结果的影响。研究结果表明,在台站均匀分布情况下,当初始震中位置偏离真实震中小于22.3 km时,双差重定位后事件的震中位置平均偏离真实震中小于3.1 km,事件间的相对震中距离为2.9 km与真实值3.35 km接近,重定位后事件深度分布比真实事件集中; 重定位后事件间的相对震中位置与添加的噪声水平关系不大,表明双差重定位法得到的事件间的相对震中位置更加稳定。当台站仅分布于震群一侧时,重定位结果与台站均匀分布时基本相同,表明台站分布对双差重定位结果影响不大。最后对双差重定位中最大震源对距离参数MAXSEP的分析表明,当事件簇中震源对平均距离比MAXSEP稍大时,重定位结果中部分事件的震源位置可能是不可靠的。

Based on the double-layer crust model, we constructed some earthquakes in lower crust to constitute the event cluster, and added different decibels noise to the real hypocenter location as initial hypocenter location in double-difference relocation when the stations are in uniform distribution and on one side of the clusters respectively, and then analyzed the influence of initial hypocenter location and stations distribution on the double-difference relocation results. The result indicates that in the situation of the stations are in uniform distribution, when initial epicenter location deviates less than 22.3 km from real epicenter, the double-difference relocated epicenter of earthquakes averagely deviates less than 3.1 km from real epicenter, and relative epicenter distance between the earthquakes is 2.9 km which approaches the real value of 3.35 km. The relocated depth distribution is more concentrated than that of real events. The relocated relative epicenter location between the earthquakes which is not sensitive to the different decibels noise is more stable obtained by double-difference location. The relocated results with stations are in uniform distribution or all on one side of the cluster are almost same, which shows that the stations distribution has little influence on the double-difference location. The analysis on the search radius MAXSEP shows that relocated focal location of some events may not be reliable when average distance between the focal pairs in the events cluster is slightly larger than MAXSEP.