Abstract: With the extensive application of focal mechanism solutions of micro-earthquakes,more and more attentions have been paid to the reliability of focal mechanism solution itself.The application of focal mechanism solution has been a significant progress in seismology,and it has played a non-fungible role in the recognition of seismic structure and seismogenesis.However,from the very beginning of the method development,the error has existed in the solution.In recent years,many methods and techniques for deriving focal mechanism solutions have been developed.Therefore,the analyses on the accuracy and reliability of these methods should also be made. The Capital-circle Digital Seismic Network was put into operation in 2002.In the network,107 seismographs may simultaneously record and transmit seismic data,and among them some are broadband seismograph with large dynamic range and some are short-period seismographs with greatly improved dynamic range.About half of the stations in the network had recorded clearly the two M=4 earthquakes occurred successively in Ninghe,Tangshan seismic area in April,2003,providing an opportunity to make the reliability analysis. For the M 4.1 earthquake occurred on April 23,2003,48 distinct and identifiable initial motion signals were selected from the P-wave initial motions recorded by stations with epicentral distance of less than 320km among the 107 stations.For the data recorded at 12 stations with epicentral distance of less than 43km,the emulation was made to obtain the record of displacement,and then the maximum amplitudes of the vertical-P and -S were read.The focal mechanism solution was then calculated using the grid-point trial method and the amplitude ratio method,respectively.The results indicate that this earthquake can be assigned to the strike-slip type with an upright nodal plane.The distribution of P-wave initial motion signals on the spherical surface used by the grid-point trial method is rather homogeneous and the nodal line can be better controlled,but the initial motion signals are a little bit deviated from the center of the net.The inconsistent ratio of the average focal mechanism solutions is 0.208,the minimum inconsistent ratio is 0.188 and the cut-off solutions are 60.The discrete area of the focal mechanism solutions is relatively small,so the results are reliable.The focal mechanism solution obtained by the amplitude ratio method is quite similar to the average focal mechanism solution obtained by the grid-point trial method,and the 3 stress axes are exactly located within the discrete area of the corresponding solutions.It can be seen that the solutions obtained by the two methods with different data coincide very well with each other,indicating that both methods are reliable.For further verification,the CMT solution of Harvard University for the largest earthquake occurred in this region in 1976 is used for comparison.The result shows that the difference of the solution with the solution of this moderate earthquake is very small.Therefore,it not only indicates the inheritance of the seismic activity,but also proves that the focal mechanism solutions are not randomly distributed. For the M 4.3 earthquake occurred on April 24,2003,55distinct and identifiable initial motion signals recorded at stations with epicentral distance of less than 320km were selected.And the maximum displacement amplitudes for the vertical-P and -S were obtained from the emulation of the data recorded at11stations with epicentral distance of less than 52km.The focal mechanism solution was calculated respectively by the above-mentioned two methods.The result shows that the focal mechanism solution of this event is of the dipslip type.The inconsistent ratio of focal mechanism solution obtained by the grid-point trial method is 0.182,the minimum inconsistent ratio is 0.127 and the cut-off solutions are 15.The discrete area of the focal mechanism solutions is relatively small,so the results are reliable.The focal mechanism solution obtained by the amplitude ratio method is quite similar to the average focal mechanism solutions ohtained by the grid-point trial method. For further verification, the CMT solution of Harvard University for the largest earthquake occurred in this region in 1977 was used for comparison. The result shows that it is basically consistent with the solution of this moderate earthquake.In the above-mentioned analyses, two different methods were used and two earthquakes with different rupture features were comparatively analyzed, and the results have proved the reliability of the focal mechanism solution determined by the amplitude ratio of -P and -S recorded by the regional seismic network. However,there are various factors that might cause the errors in the focal mechanism solutions,and the occurrence of moderate and small earthquakes is usually random. It is suggested, therefore, that the focal mechanism solutions of the moderate and small earthquakes inferred from the amplitude ratio should not be analyzed as a single item.The analysis should always be made by using the statistic method based on the focal mechanism solutions derived from a large amount of moderate and small earthquakes to abstract the average information and the variation of statistic characteristics. Only in this way can the obtained conclusigns be objective and reasonable.

Key words: focal mechanism solutions, central moment tensor, amplitude ratio method, reliability analysis

摘要： 震源机制解的可靠性越来越受到人们的重视,已有的检验工作证实中国早期或者手工作图测定的震源机制解可靠性极差。经比较,由远震波形反演的矩张量解,同一地震不同作者的结果之间差异甚小,可靠性较高;而基于P波初动方向测定震源机制解的格点尝试法所给出的解空间比其他方法要好。对2003年4月发生在首都圈数字化地震台网范围内的2次4级地震进行了分析。取用不同的资料,由格点尝试法和-P,-S垂直向振幅比方法分别测定了震源机制解,二者基本一致,并且和过去当地发生的 5级以上地震的哈佛大学CMT解接近。这既说明了地震活动的继承性,也证实了震源机制解并非随机分布。由于中小地震发生的随机性,用统计的方法对用振幅比资料测定的震源机制解进行分析,结果是客观、恰当的。

关键词: 震源机制解, 矩张量, 振幅比测定震源机制方法, 可靠性分析