2018年6月18日大阪府北部の地震 (<i>M</i>_J 6.1) の震源過程解析および疑似点震源モデルによる強震動シミュレーション

<p>The M_J 6.1 Northern Osaka earthquake of 18 June, 2018, is a crustal earthquake that occurred near the Osaka metropolitan area. While moment tensor solutions and several previous studies indicate that this earthquake involved the rupture of two fault planes, one having a predominantly dip-slip mechanism and the other having a predominantly strike-slip mechanism, contribution of each rupture to observed strong motions has not been sufficiently studied. In this study, we aimed at investigating the generation process of the strong ground motions during the earthquake and modeling the process with the pseudo point-source model. In the pseudo point-source model, the Fourier amplitude of strong ground motions is expressed as the multiplication of source, path, and site effects, where source spectra following the omega square model and empirically evaluated site amplification factors are used. In addition, Fourier phase characteristics of weak motion records are used. The model has shown great applicability to megathrust and intra-slab earthquakes despite its simplified source description, however, its applicability to crustal earthquakes has not been fully studied. Therefore, investigating its applicability to a crustal earthquake was among our objectives. First, the rupture process of the earthquake was estimated with the inversion of strong motion records using empirical Green’s functions. Reverse and strike-slip faults were used following a moment tensor solution. We found that slip was concentrated around the hypocenter on the reverse fault and slightly above the hypocenter on the strike-slip fault. Seismic moment was distributed almost evenly to both faults. These results indicated that up-dip rupture propagation caused the forward directivity effect around the epicenter and that the rupture on both faults contributed to strong motions. Then two types of pseudo point-source models were constructed. One involved only one subevent and the other involved two subevents. Either the reverse or the right-lateral fault was used for the one-subevent models and both the faults were used in the two-subevent model. The one-subevent model well captured the characteristics of observed strong motions in terms of acceleration Fourier spectra and velocity waveforms, indicating excellent applicability of the pseudo point-source model to this crustal earthquake. The two-subevent model did not lead to evident improvement of the results because the result for the one-subevent model was already satisfactory. This result does not necessarily mean that only one fault plane contributed to the generation of strong ground motions during the earthquake. Instead, this indicates that the contributions from two faults were not separated from each other on the time axis and the ground motion could be well explained with one subevent by the pseudo point-source model. The effect of forward directivity was generally small for this earthquake, however, at OSK002 just above the fault, the Fourier spectrum was underestimated at high frequencies between 2-6 Hz, which can be attributed to the up-dip rupture propagation of the strike-slip fault. Our future works include introducing a variable corner frequency model accounting for directivity effect to the pseudo point-source model and to see the model can improve the results.</p>