Poster #223, Seismology

Quantification of the Complexities in Rupture Processes of Small Earthquakes by Multiple Spectral Ratio Analyses

Takahiko Uchide
Poster Image: 

Poster Presentation

2021 SCEC Annual Meeting, Poster #223, SCEC Contribution #11476 VIEW PDF
The earthquake source processes have been characterized by seismological data analysis such as finite-fault slip inversion analyses and back-projection analyses, although the resolution is limited. Some earthquakes are unilateral, while some others are bilateral. Some have single rupture patches, while some others have multiple ones. Such diversity may be due to the stress and frictional properties of the fault being ruptured, but it is not easy to measure them directly. Nevertheless, the complexity of the rupture process may provide a clue to the physical properties of such faults.

Seismic spectral analyses quantify characteristics of the rupture process of many earthquakes. ...
By removing propagation and site effects from the observed spectra and fitting them to a spectral model (e.g., the omega-square model), the characteristics of the earthquake source process are reduced to parameters such as seismic moment and corner frequency, which can be converted into stress drops by assuming a circular crack model (e.g., Brune, 1970).

The spectral ratio method removes propagation and site effects by taking the ratio of the spectrum of the target event to that of a nearby smaller earthquake (empirical Green's function (EGF)). The multiple spectral ratio method proposed by Uchide and Imanishi (2016) further reduces the errors caused by choice of EGF events. We fit the EGF and target events' spectral ratios by the omega-square model and then obtain the residuals. By median-stacking the residual spectra, we can obtain the residual spectra independent of the choice of EGF. We consider that this residual spectrum is due to the source spectrum of the target event. Therefore, the residual spectrum superimposed on the omega-square model is considered the source spectrum of the earthquake to be analyzed.

We applied the multiple spectral ratio method to inland earthquakes at depths greater than 20 km in Japan. In some cases, the residual spectra are almost flat, and the omega-square model holds well. In contrast, in other cases, the residual spectra are undulating, and the source spectra deviate from the omega-square model so much that they appear to have two corner frequencies. A similar trend is seen in the aftershocks of the 2019 Ridgecrest earthquake. The regional variety will be a clue to the earthquake source complexity.