2:15 PM - 2:30 PM
[SSS08-12] Systematic changes in seismic radiated energy in the foreshock activities of 2016 Kumamoto earthquake
Keywords:source spectrum, sacled energy, seismci radiated energy, foreshock activity
It has been pointed out that intensive seismic activities prior to a large earthquake may be caused by aseismic slip related to or unrelated to the mainshock nucleation processes. A related hypothesis to the latter is that fluid movement and the induced aseismic slip trigger the foreshock activity and mainshocks. However, to date, the mechanism of foreshock generation has not been well clarified based on observational data. A possible observational approach is to systematically investigate the radiation characteristics of earthquakes before and after the mainshock.
Approximately 28 hours before the 2016 M7.3 Kumamoto earthquake, an M6.5 earthquake occurred, and many small earthquakes occurred between the two events, making it suitable for studying the characteristics of foreshock activities.
We estimated source spectra and scaled energies for 15756 small earthquakes (1.5<Mjma<2.5 occurring after 2008) in the focal region of the 2016 Kumamoto earthquake, and investigate their characteristics.
We determined source spectra by removing site and path effects from observed spectra. The site response and seismic attenuation Q-1 were estimated based on the coda normalization method (Aki & Chouet, 1975; Aki, 1980). Based on the results, we derived the source spectrum, radiated energy, seismic moment, and scaled energy for each event. Source spectra and scaled energies (eR ) were estimated for 9381 earthquakes. The obtained eR values are between 10-5 and 10-4 (the mean and standard deviation of log10eR are -4.93 and 0.28, respectively), similar to the results obtained in previous studies (e.g., Abercrombie et al., 1995; Kanamori & Brodsky, 2004; Yoshida & Kanamori, 2023).
We investigated the characteristics of spatiotemporal changes in eR . We found that eR was systematically low in the period between the M6.5 earthquake and the M7.3 mainshock. It is difficult to interpret this trend uniquely, but since scaled energy is correlated with stress drop and rupture velocity (e.g., Kanamori & Brodsky, 2004), the systematic decrease in eR during the foreshock period may reflect the reductions in stress drop and/or rupture velocity for earthquakes in the foreshock period. A possible cause is the reduction of effective normal stress due to the intrusion of crustal fluids in the period between the M6.5 earthquake and M7.3 mainshock (e.g., Goertz-Allmann et al., 2011; Kwiatek et al., 2014; Yoshida et al., 2017).
We estimated b -values and found characteristic spatiotemporal changes. We compared the b and eR values to investigate their relationship. We estimated b -values by fixing the cut-off magnitude (Mc) at 2.0 and applying the maximum likelihood method (Aki, 1965) to the earthquakes within 5 km from the hypocenters of each earthquake. We found that the eR values are inversely correlated with b values. Some previous studies (e.g., Scholz, 1968) suggested that b values may be inversely correlated with differential stress. The relation in the eR and b values may suggest that the derived eR values reflect differential stress and effective normal stress.
Additionally, we investigated the relationships in Mjma , Mw and eR . We found that Mw is systematically larger than Mjma for these small earthquakes, as reported by previous studies (e.g., Edwards et al., 2010). We found that the discrepancy tends to be greater for earthquakes with lower eR. The obtained trend in Mjma , Mw and eR may suggest that Mjma may deviate from Mw for earthquakes whose eR are smaller than the typical value (~10-5 ). For accurate estimation of source parameters for earthquakes with smaller than Mjma 3~4, Mw should be estimated directly rather than converted from Mjma based on empirical relations.
Approximately 28 hours before the 2016 M7.3 Kumamoto earthquake, an M6.5 earthquake occurred, and many small earthquakes occurred between the two events, making it suitable for studying the characteristics of foreshock activities.
We estimated source spectra and scaled energies for 15756 small earthquakes (1.5<Mjma<2.5 occurring after 2008) in the focal region of the 2016 Kumamoto earthquake, and investigate their characteristics.
We determined source spectra by removing site and path effects from observed spectra. The site response and seismic attenuation Q-1 were estimated based on the coda normalization method (Aki & Chouet, 1975; Aki, 1980). Based on the results, we derived the source spectrum, radiated energy, seismic moment, and scaled energy for each event. Source spectra and scaled energies (eR ) were estimated for 9381 earthquakes. The obtained eR values are between 10-5 and 10-4 (the mean and standard deviation of log10eR are -4.93 and 0.28, respectively), similar to the results obtained in previous studies (e.g., Abercrombie et al., 1995; Kanamori & Brodsky, 2004; Yoshida & Kanamori, 2023).
We investigated the characteristics of spatiotemporal changes in eR . We found that eR was systematically low in the period between the M6.5 earthquake and the M7.3 mainshock. It is difficult to interpret this trend uniquely, but since scaled energy is correlated with stress drop and rupture velocity (e.g., Kanamori & Brodsky, 2004), the systematic decrease in eR during the foreshock period may reflect the reductions in stress drop and/or rupture velocity for earthquakes in the foreshock period. A possible cause is the reduction of effective normal stress due to the intrusion of crustal fluids in the period between the M6.5 earthquake and M7.3 mainshock (e.g., Goertz-Allmann et al., 2011; Kwiatek et al., 2014; Yoshida et al., 2017).
We estimated b -values and found characteristic spatiotemporal changes. We compared the b and eR values to investigate their relationship. We estimated b -values by fixing the cut-off magnitude (Mc) at 2.0 and applying the maximum likelihood method (Aki, 1965) to the earthquakes within 5 km from the hypocenters of each earthquake. We found that the eR values are inversely correlated with b values. Some previous studies (e.g., Scholz, 1968) suggested that b values may be inversely correlated with differential stress. The relation in the eR and b values may suggest that the derived eR values reflect differential stress and effective normal stress.
Additionally, we investigated the relationships in Mjma , Mw and eR . We found that Mw is systematically larger than Mjma for these small earthquakes, as reported by previous studies (e.g., Edwards et al., 2010). We found that the discrepancy tends to be greater for earthquakes with lower eR. The obtained trend in Mjma , Mw and eR may suggest that Mjma may deviate from Mw for earthquakes whose eR are smaller than the typical value (~10-5 ). For accurate estimation of source parameters for earthquakes with smaller than Mjma 3~4, Mw should be estimated directly rather than converted from Mjma based on empirical relations.