15:30 〜 17:00
[SCG56-P04] 空間分解能向上のための応力空間パターン推定手法の改良
キーワード：応力場、空間パターン、P波初動、Delaunay三角形分割、高解像度
A Bayesian approach to estimating a spatial stress pattern from P-wave first motions was developed [Iwata, 2018, JGR]. In this approach, the spatial pattern is represented by the cubic B-spline function. In a usual case, the knot intervals of the spline are constant over a study area; if we make the intervals small to enhance the spatial resolution, the number of the coefficients of the spline or estimated parameters dramatically increases. Additionally, if the spatial density of data points (epicenter in this case) is not uniform, the knot intervals should be changed with the density of the data points for a more reasonable estimation.
For this problem, in a seismicity analysis, Ogata [2004, JGR] introduced the Delaunay triangulation for the estimation of the spatial (and temporal) pattern of the ETAS parameters. The triangulation is applied to the epicenters, and it was assumed that, within each of the triangles generated by the triangulation, the values of the parameters were changed linearly in space. The values of the parameters were estimated at each epicenter (i.e., vertices of the triangles) with smoothness constraints. Consequently, as the density of epicenters is thickened, the spatial resolution of the estimation is higher.
Following Ogata [2004], in this study, the Delaunay triangulation is incorporated to represent the spatial stress pattern. As a demonstration, the modified approach was applied to the P-wave first motion dataset taken from the aftershocks of the 2000 Western Tottori earthquake, which is the same as the one analyzed in Iwata [2018]. The general feature of the spatial pattern does not change, but the spatial variation on a small scale around the main fault is clarified in this approach.
For this problem, in a seismicity analysis, Ogata [2004, JGR] introduced the Delaunay triangulation for the estimation of the spatial (and temporal) pattern of the ETAS parameters. The triangulation is applied to the epicenters, and it was assumed that, within each of the triangles generated by the triangulation, the values of the parameters were changed linearly in space. The values of the parameters were estimated at each epicenter (i.e., vertices of the triangles) with smoothness constraints. Consequently, as the density of epicenters is thickened, the spatial resolution of the estimation is higher.
Following Ogata [2004], in this study, the Delaunay triangulation is incorporated to represent the spatial stress pattern. As a demonstration, the modified approach was applied to the P-wave first motion dataset taken from the aftershocks of the 2000 Western Tottori earthquake, which is the same as the one analyzed in Iwata [2018]. The general feature of the spatial pattern does not change, but the spatial variation on a small scale around the main fault is clarified in this approach.