11:00 AM - 1:00 PM
[SCG44-P06] Roughness analysis of Nankai trough subduction plate interface
Slow earthquake, which is a slip faster than the plate convergence motion and significantly slower than the normal earthquake, has been discovered in subduction plate interface in recent years, and its relationship with regular earthquakes is one of issues to understand mechanisms of the slow and fast slips (Obara and Kato, 2016). The observations indicate that plate interface slips with various slip velocities such as regular earthquakes and slow earthquakes. Previous studies have proposed a model that the scale-hierarchical distribution of physical properties can control the various slips (Ide et al., 2014). There are various debates about the factors of the distribution of physical properties itself. The variation in physical properties could be related to heterogenetic distribution of lithology, fluid pressure and geometry of plate interface, but these causes are only conceptual without any evidence-based interpretation from natural example.
Therefore, in this study, we targeted the actual subduction plate interface. The three-dimensional seismic reflection profiles in the Nankai Trough off Kii Peninsula were employed to examine the geometrical distribution of roughness. The study area is covering about 11 km east-west and about 26 km north-south. After clarifying hierarchical distribution of rough area by roughness analysis, the conceptual model is verified. Specifically, the amplitude distribution of the topography obtained by the roughness analysis of the natural decollement topography was illustrated, and the hierarchical distribution of high-amplitude regions on different scales with the difference in segment length was examined.
In this study, roughness analysis was performed using a double-logarithmic graph between the wavelength and PSD obtained by the Fourier transform of the topographical waveform with some variations of the segment length. It was confirmed that there are linear relationships, indicating a power relationship is established between the wavelength and the amplitude. With a small segment length, patches with relatively smaller area were observed in places. Conversely, for large segment lengths, high amplitude regions are distributed over a relatively wide area. The size of high-amplitude (rough) area increases with segment length, showing hierarchal distribution of rough areas with the variation in segment length, which was evidenced form the natural plate interface.
In Candela et al. (2011), the roughness analysis of the natural fault plane gave a linear relationship in the space between wavelength and PSD. In other words, it was confirmed that the Hurst index showing this slope is constant. Using this relationship, it can be estimated that the amplitude increases as the wavelength increases.
Further, by deriving the relationship between stress drop and wavelength, the deviation of the stress drop is depending on the size of the rupture region. Specifically, assuming that the area of the high amplitude area at a certain amplitude threshold value is the rupture region, the larger the rupture region corresponds the area with lower stress drop in the distribution of initial shear stress according to the topography on the fault plane.
As mentioned above, a hierarchical distribution of high-amplitude area was observed as a topographical effect in response to variation in segment length in this study. Based on Candela et al. (2011), the results of this study may be said to be a hierarchical model in terms of stress drop. However, if the results of Candela et al. (2011) are correct, the relationship between the size of rupture area and the stress drop is that the smaller the area of the rupture area, the greater the stress drop. This is different from the seismological image.
As a future prospect, the roughness of various factors such as slip tendency and dilation tendency would be examined and compare them with the roughness itself to understand the topographic control on the stress distributions.
Therefore, in this study, we targeted the actual subduction plate interface. The three-dimensional seismic reflection profiles in the Nankai Trough off Kii Peninsula were employed to examine the geometrical distribution of roughness. The study area is covering about 11 km east-west and about 26 km north-south. After clarifying hierarchical distribution of rough area by roughness analysis, the conceptual model is verified. Specifically, the amplitude distribution of the topography obtained by the roughness analysis of the natural decollement topography was illustrated, and the hierarchical distribution of high-amplitude regions on different scales with the difference in segment length was examined.
In this study, roughness analysis was performed using a double-logarithmic graph between the wavelength and PSD obtained by the Fourier transform of the topographical waveform with some variations of the segment length. It was confirmed that there are linear relationships, indicating a power relationship is established between the wavelength and the amplitude. With a small segment length, patches with relatively smaller area were observed in places. Conversely, for large segment lengths, high amplitude regions are distributed over a relatively wide area. The size of high-amplitude (rough) area increases with segment length, showing hierarchal distribution of rough areas with the variation in segment length, which was evidenced form the natural plate interface.
In Candela et al. (2011), the roughness analysis of the natural fault plane gave a linear relationship in the space between wavelength and PSD. In other words, it was confirmed that the Hurst index showing this slope is constant. Using this relationship, it can be estimated that the amplitude increases as the wavelength increases.
Further, by deriving the relationship between stress drop and wavelength, the deviation of the stress drop is depending on the size of the rupture region. Specifically, assuming that the area of the high amplitude area at a certain amplitude threshold value is the rupture region, the larger the rupture region corresponds the area with lower stress drop in the distribution of initial shear stress according to the topography on the fault plane.
As mentioned above, a hierarchical distribution of high-amplitude area was observed as a topographical effect in response to variation in segment length in this study. Based on Candela et al. (2011), the results of this study may be said to be a hierarchical model in terms of stress drop. However, if the results of Candela et al. (2011) are correct, the relationship between the size of rupture area and the stress drop is that the smaller the area of the rupture area, the greater the stress drop. This is different from the seismological image.
As a future prospect, the roughness of various factors such as slip tendency and dilation tendency would be examined and compare them with the roughness itself to understand the topographic control on the stress distributions.