Solving For Focal Mechanisms Of Small And Moderate Earthquakes Using Initial P Waveforms And Its Preliminary Application For Source Reversion In Beijing Metropolitan Area

As the density of seismic stations in regional seismic networks gradually increasing, focal mechanism determination can be possible for small and moderate regional earthquakes that occur frequently. This will provide more important constraints on regional tectonic activities and stress field status, and also has led to an expansion of seismic network functions. Developing automatic or semi-automatic focal mechanism inversion methods to provide more reliable results is a very important way to improve seismic network data utility ratio and extend information service.The aims of this paper is to introduce an automatic focal mechanism inversion method that uses small and moderate earthquakes'initial P waveforms, with both theoretical basis and application examples presented. As the digital seismic networks being established more and more widely, solving for focal mechanisms by digital seismic waveform inversion is becoming more convenient and tends to be an international trend. The grid search method was used in this study. Based on the idea of CAP(Cut and Paste) method, this method searches and finds focal mechanisms with a two-step strategy by fully utilizing the initial P-waveform information: in the first step, we search for focal mechanism with large grid size in 3-dimensional parameter space by cross-correlating initial P waveforms between synthetic seismograms and recorded waveforms, and limit the parameter range for the search in the next step by comparing multi-stations'results. In the second step, the final earthquake focal mechanism will be found by repeating the search procedure with smaller grid size in the limited parameter range.This paper will first introduced a method of calculating synthetic seismogram—the frequency-wavenumber integral (F-K) method, and Green's function database method. The Green's function database adopted here is for the efficiency of the this method and future data processing which will determine a large number of earthquake focal mechanisms of small and moderate earthquake in Beijing metropolitan area. To build the database, one-dimensional P-wave velocity model of Beijing metropolitan area was established, and Green's functions in Beijing metropolitan region were calculated using the F-K program. A search and match script was also made at this time. To test and verify this inversion method, the second part of this paper did some numerical tests. Four scenario earthquakes were used to simulate focal mechanism inversion, and the results showed that this method is capable of retrieving focal mechanisms of small and moderate earthquakes using initial P waveforms, which in turn validates the feasibility of this method from the numerical simulation perspective. Furthermore, the capability to invert focal mechanism with only a fewer stations was tested. The test showed that results similar to those using many stations can be achieved by this method if more than 4 stations with uniform azimuth distribution were available and their azimuthal coverage was around 180°.The last part of this paper is focused on inversion method's application on Wen'an earthquake which occurred on July 4th, 2006 in Beijing metropolitan area. Using initial P waveforms of 33 stations in the Beijing metropolitan area, based on two-step search procedure, one section of the Wen'an earthquake focal mechanism was inverted as: strike= 20°, dip =82°, slip= 158°. The initial P waveform comparisons between synthetic seismograms calculated with inverted focal mechanism and the recorded waveforms of 33 stations showed good consistency. As small earthquake radiates energy so weak that the number of recorded stations would be limited and the azimuthal coverage of recorded stations would be poor, which can restrict focal mechanism inversion. So earthquake focal mechanism inversions using fewer stations were tried. These testes showed that focal mechanism can be recovered as good as 33 stations'result by this method on the Wen'an earthquake data when over 5 stations were available with about 180°station azimuth coverage and an even station azimuth distribution. These results also confirmed the previous numerical cases.Through the numerical test and Wen'an Earthquake's real data inversion, the method of this study were proved to be feasible in utilizing initial P waveforms to invert for focal mechanisms of small and moderate earthquakes, it also showed that this method has potential advantages for small earthquake focal mechanism inversion when only fewer stations are available. The research in this paper has provided theoretical basis and application examples for automatic focal mechanism inversion of small and moderate earthquakes using initial P waveforms.