5:15 PM - 6:30 PM
[SSS06-P06] Whether the change in shear strain energy causes change in b value? Case study for the 2016 Kumamoto Earthquakes and the interplate coupling along the Nankai Trough
Keywords:seismicity, shear strain energy, b value
The b value in the Gutenberg–Richter law, one of the typical parameters representing the characteristics of seismic activity, is often considered to depend on the differential stress from observations in laboratory experiments and regional differences in seismic activity (Scholz, 2015). Regarding the physical background of the stress dependence, it is reasonable to think that the larger the differential stress, the higher the probability of crack growth and the number of relatively large-scale earthquakes increase, as proposed by Scholz (1968). However, estimating the stress field including its absolute value involves great uncertainty, and it is difficult to estimate the differential stress on a scale corresponding to actual seismic activity. Therefore, when considering the use for real-time monitoring of earthquakes or earthquake forecasts, it would be unrealistic to discuss the change in b value focusing only on the relationship with the differential stress.
On the other hand, in recent years, changes in shear strain energy (ΔSSE, hereafter) have been proposed to be used as an index that quantitatively evaluates the degree of shear stress accumulation from relative changes in the stress field without assuming a fault plane (Saito et al., 2018; Matsu'ura et al., 2019; Terakawa et al., 2020; Noda et al., 2020). An increase in shear strain energy means an average increase in shear stress when fault planes are randomly oriented in the crust, and a relationship with an increase in the number of earthquakes has already been reported. If the increase in shear strain energy contributes the growth of rupture in earthquake faults, it is expected that the b value will decrease correspondingly.
Based on this view, we investigated the difference in the size distribution of seismic activity corresponding to ΔSSE due to the 2016 Kumamoto earthquakes (Noda et al., 2020). Estimates of shear strain energy have uncertainty mainly due to the uncertainty of the assumed effective friction coefficient. However, unless the effective friction coefficient is 0, the spatial distribution of increase and decrease of shear strain energy is almost constant except in the very vicinity of the fault. Therefore, we first, identified the region where the shear strain energy increased/decreased (positive/negative ΔSSE) in any case with the effective friction coefficient of 0.1 to 0.4. Then, for each region, the earthquakes that occurred in the region were extracted from the Japan Meteorological Agency unified seismic catalog. Using the extracted hypocenter data, b values before the Kumamoto earthquakes (January 2000-April 13, 2016), and those after the decrease of detectability due to the largest earthquake has almost recovered (April 16, 2016, 8:00 to August 2020) were estimated with some different lower limit of M(Mth, hereafter) in the range from 2.0 to 3.0.
As a result, in the positive ΔSSE region, the b value decreased after the Kumamoto earthquakes compared to before the earthquakes. The amount of decrease exceeded the standard error for almost all Mth. After the Kumamoto earthquake, the b value in the positive ΔSSE region was lower than the b value in the negative ΔSSE region for all Mth. This difference was slightly smaller than the standard error for most Mth. In the negative ΔSSE region, no clear change in b value was observed before and after the Kumamoto earthquake. There is no significant difference in the distribution of focal depths in each of these comparison targets.
In addition to the above results, we investigate the relationship between ΔSSE due to interplate coupling along Nankai Trough (Saito et al., 2018) and the b value. Comparing the size distributions of seismic activity corresponding to the positive and negative of the shear strain energy change in the entire analysis region, the estimated b values tend to be slightly higher in the region where the shear strain energy change is negative. Though there are some parts that cannot be simply compared since the comparison is made over a wide area where the regional differences in b value can be affected not only by ΔSSE considered here but also by various causes such as focal depth and effective normal stress, on average, the results are consistent with the expectations. We would like to discuss these results together in the presentation.
On the other hand, in recent years, changes in shear strain energy (ΔSSE, hereafter) have been proposed to be used as an index that quantitatively evaluates the degree of shear stress accumulation from relative changes in the stress field without assuming a fault plane (Saito et al., 2018; Matsu'ura et al., 2019; Terakawa et al., 2020; Noda et al., 2020). An increase in shear strain energy means an average increase in shear stress when fault planes are randomly oriented in the crust, and a relationship with an increase in the number of earthquakes has already been reported. If the increase in shear strain energy contributes the growth of rupture in earthquake faults, it is expected that the b value will decrease correspondingly.
Based on this view, we investigated the difference in the size distribution of seismic activity corresponding to ΔSSE due to the 2016 Kumamoto earthquakes (Noda et al., 2020). Estimates of shear strain energy have uncertainty mainly due to the uncertainty of the assumed effective friction coefficient. However, unless the effective friction coefficient is 0, the spatial distribution of increase and decrease of shear strain energy is almost constant except in the very vicinity of the fault. Therefore, we first, identified the region where the shear strain energy increased/decreased (positive/negative ΔSSE) in any case with the effective friction coefficient of 0.1 to 0.4. Then, for each region, the earthquakes that occurred in the region were extracted from the Japan Meteorological Agency unified seismic catalog. Using the extracted hypocenter data, b values before the Kumamoto earthquakes (January 2000-April 13, 2016), and those after the decrease of detectability due to the largest earthquake has almost recovered (April 16, 2016, 8:00 to August 2020) were estimated with some different lower limit of M(Mth, hereafter) in the range from 2.0 to 3.0.
As a result, in the positive ΔSSE region, the b value decreased after the Kumamoto earthquakes compared to before the earthquakes. The amount of decrease exceeded the standard error for almost all Mth. After the Kumamoto earthquake, the b value in the positive ΔSSE region was lower than the b value in the negative ΔSSE region for all Mth. This difference was slightly smaller than the standard error for most Mth. In the negative ΔSSE region, no clear change in b value was observed before and after the Kumamoto earthquake. There is no significant difference in the distribution of focal depths in each of these comparison targets.
In addition to the above results, we investigate the relationship between ΔSSE due to interplate coupling along Nankai Trough (Saito et al., 2018) and the b value. Comparing the size distributions of seismic activity corresponding to the positive and negative of the shear strain energy change in the entire analysis region, the estimated b values tend to be slightly higher in the region where the shear strain energy change is negative. Though there are some parts that cannot be simply compared since the comparison is made over a wide area where the regional differences in b value can be affected not only by ΔSSE considered here but also by various causes such as focal depth and effective normal stress, on average, the results are consistent with the expectations. We would like to discuss these results together in the presentation.