11:00 AM - 1:00 PM
[SSS10-P03] Heterogeneity of rupture velocity on the source fault estimated from the dynamic cycle source model validated by the observed data
Keywords:rupture velocity, dynamic source model, asperity
The rupture velocity (Vr) on the fault plane is one of essential parameters to understand rupture process of earthquakes. From multi time-window inversions we can estimate the rupture time and Vr respectively and understand details of the rupture propagation. Furthermore, Vr is an essential parameter for the evaluation and prediction of ground motions generated from the source model. In the recipe for predicting strong ground motions, the average Vr is set to be 0.72 β (S-wave velocity) by Geller (1976), unless there is a detailed information. Furthermore, Miyakoshi and Petukhin (2005) and Somei et al. (2022, this meeting) found that Vr in asperities should be larger than Vr in background area. In this study, we discuss the heterogeneity of Vr estimated from the validated dynamic rupture models.
Due to a small number of time windows, Vr estimated from source inversions may have poor resolution. Realistic dynamic rupture modelling validated by observed earthquakes, is an alternative approach. In this study, we use set of spontaneous physically self-consistent rupture models, whose rupture process is consistent with the spatio-temporal heterogeneity of previous earthquakes on the same fault, as the result of cycle simulations under the rate-and-state friction law on strike-slip fault (Galvez et al., 2021). Models are well validated by comparison with: (1) source scaling relations from the seismic inversions, (2) GMPE, and (3) observed fault displacements.
Earthquake Research Committee (2008) showed the result of the long-period ground motions calculated based on the recipe vary greatly depending on the setting of the Vr. From another side, Petukhin et al. (2021) demonstrated that Vr correlates with other kinematic parameters of rupture and cannot be prescribed independently. Nevertheless, analysis of residuals of average PGV values and Vr values over a rupture demonstrate positive correlation between them (Fig.1). In this way, by physics-based modelling we additionally confirm that Vr is a key parameter for generation of ground motions and require detail study.
In contrast to seismic inversion models, dynamic ruptures here have homogeneous distribution of the Slip, but heterogeneous peak slip rate distribution (PSR), see example in Fig.2. Petukhin et al. (2021) found that PSR correlates with reverse value of characteristic distance 1/Dc, key parameter of rate-and-sate friction model. Asperities are defined as areas of small Dc, and we can extrapolate definition of asperity as an area of large PSR or high-rate area (HRA). This definition disagrees to definition of asperity as the large-slip area (LSA), used to analyze source inversion results. We calculated average and standard deviation of Vr in the HRA asperities and the off-asperity area extracted by the Somerville et al. (1999) criterion and found that Vr in asperity area is 0.63β ± 0.12, and in background area is 0.39β ± 0.11. In comparison to the source inversion results, dynamic Vr have much larger ratio of the asperity values to the background values, although average values are smaller while standard deviations are larger. This result is consistent with result of Petukhin et al. (2021), they found that Vr correlates well with PSR values. However, to apply this result to the characterized source modelling for ground motion prediction, it may be necessary to tune Vr values in asperities and background area so that to reproduce ground motions.
Acknowledgement: This study was based on the research project “Examination for uncertainty of strong ground motion prediction for inland crustal earthquakes” for 2020 and 2021 yrs., by The Secretariat of the Nuclear Regulation Authority (NRA), Japan.
References.
Galvez et al. (2021), Bull.Seismol.Soc.Am., doi: 10.1785/0120210104.
Miyakoshi and Petukhin (2005), JpGU 2005, Chiba, Japan, S046p-002.
Petukhin et al. (2021), JpGU 2021, Chiba, Japan, SSS11-12.
Somei et al. (2022), JpGU 2022, Chiba, Japan (submitted).
Due to a small number of time windows, Vr estimated from source inversions may have poor resolution. Realistic dynamic rupture modelling validated by observed earthquakes, is an alternative approach. In this study, we use set of spontaneous physically self-consistent rupture models, whose rupture process is consistent with the spatio-temporal heterogeneity of previous earthquakes on the same fault, as the result of cycle simulations under the rate-and-state friction law on strike-slip fault (Galvez et al., 2021). Models are well validated by comparison with: (1) source scaling relations from the seismic inversions, (2) GMPE, and (3) observed fault displacements.
Earthquake Research Committee (2008) showed the result of the long-period ground motions calculated based on the recipe vary greatly depending on the setting of the Vr. From another side, Petukhin et al. (2021) demonstrated that Vr correlates with other kinematic parameters of rupture and cannot be prescribed independently. Nevertheless, analysis of residuals of average PGV values and Vr values over a rupture demonstrate positive correlation between them (Fig.1). In this way, by physics-based modelling we additionally confirm that Vr is a key parameter for generation of ground motions and require detail study.
In contrast to seismic inversion models, dynamic ruptures here have homogeneous distribution of the Slip, but heterogeneous peak slip rate distribution (PSR), see example in Fig.2. Petukhin et al. (2021) found that PSR correlates with reverse value of characteristic distance 1/Dc, key parameter of rate-and-sate friction model. Asperities are defined as areas of small Dc, and we can extrapolate definition of asperity as an area of large PSR or high-rate area (HRA). This definition disagrees to definition of asperity as the large-slip area (LSA), used to analyze source inversion results. We calculated average and standard deviation of Vr in the HRA asperities and the off-asperity area extracted by the Somerville et al. (1999) criterion and found that Vr in asperity area is 0.63β ± 0.12, and in background area is 0.39β ± 0.11. In comparison to the source inversion results, dynamic Vr have much larger ratio of the asperity values to the background values, although average values are smaller while standard deviations are larger. This result is consistent with result of Petukhin et al. (2021), they found that Vr correlates well with PSR values. However, to apply this result to the characterized source modelling for ground motion prediction, it may be necessary to tune Vr values in asperities and background area so that to reproduce ground motions.
Acknowledgement: This study was based on the research project “Examination for uncertainty of strong ground motion prediction for inland crustal earthquakes” for 2020 and 2021 yrs., by The Secretariat of the Nuclear Regulation Authority (NRA), Japan.
References.
Galvez et al. (2021), Bull.Seismol.Soc.Am., doi: 10.1785/0120210104.
Miyakoshi and Petukhin (2005), JpGU 2005, Chiba, Japan, S046p-002.
Petukhin et al. (2021), JpGU 2021, Chiba, Japan, SSS11-12.
Somei et al. (2022), JpGU 2022, Chiba, Japan (submitted).