5:15 PM - 7:15 PM
[SCG45-P46] Foreshocks, mainshocks, and aftershocks in multi-scale circular patch model of quasi-dynamic numerical simulation of earthquake generation cycle
We conducted quasi-dynamic numerical simulations of earthquake generation cycles based on the rate- and state-dependent friction law by adopting a multiscale circular patch model (Ide & Aochi, 2013) for off Sanriku region, northern segment of the Japan Trench. To determine the spatial heterogeneity of seismic events, historical earthquakes with M ≧ 5.6 recorded since 1896 in the study area were referenced. We divided them into four groups according to magnitude. Source area was modeled as a circular patch of the same size centered on the epicenter. As the magnitude of the group decreased, the radius halved. The characteristic slip distance was set to be constant within each group and proportional to the patch radius. We assumed uniform A-B and velocity weakening regardless of the magnitude of the earthquake.
As a result, 43 M ≧ 8 earthquakes with a recurrence interval of approximately 61 years. For each M ≧ 8 mainshock, there were at most 11 events occurred in one year before the mainshock. However, not all earthquakes that occur far from the rupture initiation point can be considered as foreshocks because they are not related to the rupture process of the mainshock. We then visually defined foreshocks as those close in space to the rupture initiation of each M ≧ 8 earthquake. We found in more than 80% of the 43 M >= 8 earthquakes that at least one foreshock-like earthquake occurred within approximately one year prior to the mainshock. In the remaining cases, it appears to have been triggered by slow slip.
On the other hand, we can treat all the earthquakes that occurred in the model domain after each mainshock as aftershocks. These aftershocks occurred around the source of each mainshock, especially in the isolated small patches, and were generally small in magnitude. For each M ≧ 8 mainshock, the number of all earthquakes occurred in the model domain during 10 years after each mainshock roughly followed the modified Omori formula.
By adopting a multiscale structure, we were able to show different triggering scenarios for M ≧ 8 earthquakes. Our results suggest that a hierarchical structure is fundamental for simulating not only the dynamic rupture process, but also the preparation process, propagation process, and subsequent relaxation process of long-term earthquake cycles.
As a result, 43 M ≧ 8 earthquakes with a recurrence interval of approximately 61 years. For each M ≧ 8 mainshock, there were at most 11 events occurred in one year before the mainshock. However, not all earthquakes that occur far from the rupture initiation point can be considered as foreshocks because they are not related to the rupture process of the mainshock. We then visually defined foreshocks as those close in space to the rupture initiation of each M ≧ 8 earthquake. We found in more than 80% of the 43 M >= 8 earthquakes that at least one foreshock-like earthquake occurred within approximately one year prior to the mainshock. In the remaining cases, it appears to have been triggered by slow slip.
On the other hand, we can treat all the earthquakes that occurred in the model domain after each mainshock as aftershocks. These aftershocks occurred around the source of each mainshock, especially in the isolated small patches, and were generally small in magnitude. For each M ≧ 8 mainshock, the number of all earthquakes occurred in the model domain during 10 years after each mainshock roughly followed the modified Omori formula.
By adopting a multiscale structure, we were able to show different triggering scenarios for M ≧ 8 earthquakes. Our results suggest that a hierarchical structure is fundamental for simulating not only the dynamic rupture process, but also the preparation process, propagation process, and subsequent relaxation process of long-term earthquake cycles.