Characterizing multi-subevents earthquakes using Brune source model

Brune source model has been widely used in ground motion prediction and source parameter estimation in a wide range of magnitudes. The source model describes the source process as a single pulse with a characteristic time scale corresponding to a corner frequency in the frequency domain. However, one source pulse could be oversimplified for large earthquakes when 80% of Mw>5.5 earthquakes in SCARDEC Source Time Functions Database are found to have multiple pulses that represent subevents in the rupture process (Danr et al., 2019). Here we investigate the performance of Brune source model in characterizing complex earthquake sources. We first build synthetic source time functions with two subevents and vary corner frequency, seismic moment, and the occurrence time of the subevents. We then construct synthetic waveforms of the two-subevents master event and its empirical Green's function using the Frequency-Wavenumber (FK) method and model their spectral ratios using the Brune source model. OUr results show that the master event corner frequency estimated from the Brune source model is more related to the subevent with a larger seismic moment. We also find that the master event corner frequency strongly depends on the source duration when the time difference between two subevents is relatively small (<2.5 s in our two subevent models).

To verify the results using real data, we decompose the SCARDEC source time functions of deep-focus earthquakes into a sum of subevent source time functions represented by Brune sources and measure the time difference between two subevents with the largest seismic moments. We then estimate the Brune corner frequencies of these deep-focus earthquakes using the spectral ratio method (Liu et al., 2020) and unveil a similar dependence of the master event corner frequency on the subevent time difference. Our results show that the time difference between two subevents and the subevent with the largest moments play important roles in characterizing multi-subevents earthquakes when the Brune source model is used.