11:00 AM - 1:00 PM
[SSS10-P02] Heterogeneity of rupture velocity on the source fault estimated from the kinematic source model
Keywords:Rupture velocity, Kinematic source model, Source time function
The rupture velocity on the fault plane is one of the essential parameters to understand the rupture process of earthquakes. In the multi-time window waveform inversion analysis (e.g., Hartzell and Heaton, 1983), homogeneous rupture propagation spreading in concentric circles from the rupture initiation point is assumed to estimate the kinematic source model, and the rupture propagation velocity is often set to explain the observed waveform. This propagation velocity is known as the first-time-window rupture propagation velocity, which triggers the rupture of the point source set at the center of each sub-fault. However, the estimated source time function of each sub-fault does not necessarily indicate that the slip starts at the first-time window (Miyakoshi et al., 2004). Therefore, if we can estimate the rupture time that defines the start of slip, we can understand the heterogeneity of the rupture propagation velocity that better represents the actual rupture process. Furthermore, the rupture velocity 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 rupture propagation velocity of the characterized source model is set to be 0.72 β (S-wave velocity in the source region) by Geller (1976), unless there is detailed information about the source region. In this study, we discuss the heterogeneity of the rupture velocity estimated from the kinematic source models of inland crustal earthquakes by specifying the slip onset based on the source time function for each sub-fault.
We generally follow the procedure of Miyakoshi and Petukhin (2004), to estimate the rupture velocity for each sub-fault considering with the slip initiation specified from the kinematic source model. The onset of slip is defined as the time when the amount of moment on each sub-fault reaches an amount of moment equal to 0.3 times the average slip of the entire fault, and the rupture propagation time of each sub-fault is calculated using this definition of the onset of slip. The rupture propagation time of each sub-fault is calculated using the rupture propagation time between the sub-faults and the distance between the point sources, and the rupture propagation velocity of each sub-fault is calculated from the vector of the slowness calculated from the central difference between each strike and dip directions. The four earthquakes targeted in this study are the 2013 northern Tochigi, the 2013 Awaji Island, the 2016 northern Ibaraki, and the 2019 Ridgecrest (Mw 6.5) earthquakes, each of which was estimated by multi-time-window waveform inversion from strong motion records. We calculated the arithmetic mean and standard deviation of the rupture velocities of the sub-faults in the asperity area and the off-asperity area (background area) extracted by the Somerville et al. (1999) criterion for each earthquake, and found that the rupture velocity in the asperity area is 0.66β ± 0.06, and the background area is 0.58β ± 0.07. Miyakoshi and Petukhin (2004), which conducted a similar study for 12 inland crustal earthquakes, found a rupture velocity of 0.73β ± 0.14 in the asperity area and 0.69β ± 0.19 in the background area. Although the rupture velocity estimated in this study is slow, it is similar to that the rupture velocity is slightly faster in the asperity area than that in the background area. We will compare the results with the heterogeneity of the rupture velocity obtained by the dynamic simulations. In addition, the method of defining the slip initiation and the handling of complex source time functions should be discussed in the future.
Acknowledgement: This study was based on the 2021 research project “Examination for uncertainty of strong ground motion prediction for inland crustal earthquakes” by The Secretariat of the Nuclear Regulation Authority (NRA), Japan.
We generally follow the procedure of Miyakoshi and Petukhin (2004), to estimate the rupture velocity for each sub-fault considering with the slip initiation specified from the kinematic source model. The onset of slip is defined as the time when the amount of moment on each sub-fault reaches an amount of moment equal to 0.3 times the average slip of the entire fault, and the rupture propagation time of each sub-fault is calculated using this definition of the onset of slip. The rupture propagation time of each sub-fault is calculated using the rupture propagation time between the sub-faults and the distance between the point sources, and the rupture propagation velocity of each sub-fault is calculated from the vector of the slowness calculated from the central difference between each strike and dip directions. The four earthquakes targeted in this study are the 2013 northern Tochigi, the 2013 Awaji Island, the 2016 northern Ibaraki, and the 2019 Ridgecrest (Mw 6.5) earthquakes, each of which was estimated by multi-time-window waveform inversion from strong motion records. We calculated the arithmetic mean and standard deviation of the rupture velocities of the sub-faults in the asperity area and the off-asperity area (background area) extracted by the Somerville et al. (1999) criterion for each earthquake, and found that the rupture velocity in the asperity area is 0.66β ± 0.06, and the background area is 0.58β ± 0.07. Miyakoshi and Petukhin (2004), which conducted a similar study for 12 inland crustal earthquakes, found a rupture velocity of 0.73β ± 0.14 in the asperity area and 0.69β ± 0.19 in the background area. Although the rupture velocity estimated in this study is slow, it is similar to that the rupture velocity is slightly faster in the asperity area than that in the background area. We will compare the results with the heterogeneity of the rupture velocity obtained by the dynamic simulations. In addition, the method of defining the slip initiation and the handling of complex source time functions should be discussed in the future.
Acknowledgement: This study was based on the 2021 research project “Examination for uncertainty of strong ground motion prediction for inland crustal earthquakes” by The Secretariat of the Nuclear Regulation Authority (NRA), Japan.