16:30 〜 16:45
[SSS10-27] Analysis of acoustic emissions and stress evolution in large-scale laboratory experiments to reveal the foreshock occurrence mechanism
Foreshocks have been actively studied since they are considered to be closely related to the mainshock occurrence process and may also provide short-term forecasting before large earthquakes. For example, Yamashita et al. (2021) found that the foreshock activity and the process of mainshock occurrence depend on the heterogeneity in the distribution of the gouge layer on the fault plane by using STA/LTA to detect acoustic emissions (AEs) in rock experiments. This study aims to bring new insights into the topics described above using shear stress data and a more detailed AE catalog derived from the rock experiments used by Yamashita et al. (2021). The enhanced AE catalog was obtained using a modified matched filter technique that we have developed. The methodology was already described and presented by Ishiyama et al. (2024, JPGU). We describe below the main results.
Our AEs detections revealed that very different foreshock activity occurred immediately before some successive mainshocks: many foreshocks and their migration toward the mainshock were observed for one mainshock, while almost no foreshocks were observed for the subsequent mainshock. The observed shear stress changes suggest that slow slip may have occurred along with foreshock migration. The slow slip was localized and decelerated. Therefore, it may be interpreted differently from the aseismic nucleation process of the mainshock predicted by the 'preslip' model. Since we observe localized slow-slip that occurs closely to the foreshock locations and likely initiates after foreshock occurrence, our interpretation is that the observed slow-slip is a by-product of foreshock activity. On the other hand, in cases where foreshocks are scarce, the mainshock may have been triggered by the foreshocks as described by the 'cascade' model.
The difference in the number of foreshocks detected for each of the 31 mainshocks was found to be positively correlated with the heterogeneity of the distribution of the initial shear stress on the fault plane immediately after the previous mainshock. In other words, fewer foreshocks occur when the initial shear stress is homogeneous, and more foreshocks occur when the shear stress is heterogeneous. This may be because foreshocks are smoothing the distribution of the shear stress along the laboratory fault plane, so that the mainshock failure is facilitated. The relationship between the number of foreshocks and the heterogeneity of the initial shear stress indicates that foreshock activity and the timing of mainshock occurrence can be predicted to some extent based on information from previous mainshocks. Because the distribution of the normal stress on the fault plane did not change significantly during the experiment, the differences in foreshock activity observed in this study are likely independent of the setting of the fault and may reflect processes that take place in nature, on tectonic faults of various geological settings.
Our AEs detections revealed that very different foreshock activity occurred immediately before some successive mainshocks: many foreshocks and their migration toward the mainshock were observed for one mainshock, while almost no foreshocks were observed for the subsequent mainshock. The observed shear stress changes suggest that slow slip may have occurred along with foreshock migration. The slow slip was localized and decelerated. Therefore, it may be interpreted differently from the aseismic nucleation process of the mainshock predicted by the 'preslip' model. Since we observe localized slow-slip that occurs closely to the foreshock locations and likely initiates after foreshock occurrence, our interpretation is that the observed slow-slip is a by-product of foreshock activity. On the other hand, in cases where foreshocks are scarce, the mainshock may have been triggered by the foreshocks as described by the 'cascade' model.
The difference in the number of foreshocks detected for each of the 31 mainshocks was found to be positively correlated with the heterogeneity of the distribution of the initial shear stress on the fault plane immediately after the previous mainshock. In other words, fewer foreshocks occur when the initial shear stress is homogeneous, and more foreshocks occur when the shear stress is heterogeneous. This may be because foreshocks are smoothing the distribution of the shear stress along the laboratory fault plane, so that the mainshock failure is facilitated. The relationship between the number of foreshocks and the heterogeneity of the initial shear stress indicates that foreshock activity and the timing of mainshock occurrence can be predicted to some extent based on information from previous mainshocks. Because the distribution of the normal stress on the fault plane did not change significantly during the experiment, the differences in foreshock activity observed in this study are likely independent of the setting of the fault and may reflect processes that take place in nature, on tectonic faults of various geological settings.