Automatic determination of focal mechanism solutions using initial motion polarities of P and S waves

This study proposes an automatic method of determining a focal mechanism solution. This method includes picking initial motion polarities of P and S waves and judging reliability of the solution. The method of picking the polarities is as follows. Displacement waveform data are calculated only for S waves. A Bessel band-pass filter is then applied. Finally, arrival times of the P and S wave initial motions are detected by using an autoregressive (AR) model and an AR filter is applied. A grid search method is used to determine the focal mechanism solution. The reliability of the solution is evaluated synthetically from the stability of the solution and by comparing it with focal mechanism solutions of earthquakes that occurred near the event. About 71% of the polarities of P waves picked manually by the Japan Meteorological Agency (JMA) can be picked automatically by applying this method. Moreover, the ratio of inconsistent polarities between those picked in this study and by JMA is less than 4.0%. This proves that automatic picking of the polarities of P waves is sufficiently accurate. About 85% of the polarities of S waves satisfy focal mechanism solutions that are determined with acceptable reliability. This also demonstrates that automatic picking of the polarities of S waves is sufficiently accurate. Furthermore, 2.8 times the solutions obtained by JMA are determined with sufficient reliability by applying the method. We can occasionally determine the stable solutions of inland shallow earthquakes that are less than 2.0 in magnitude, or of deep earthquakes whose magnitudes cannot be calculated because of their small sizes. Applying this method, we can effectively determine focal mechanism solutions with sufficient reliability.