11:15 〜 11:30
[SIT16-03] レシーバ関数とマルチモード表面波の同時インバージョンによる豪州大陸の上部マントル不連続面マッピング
キーワード：上部マントル、リソスフェア、アセノスフェア、レーマン面、レシーバ関数、表面波
Seismic discontinuities in the upper mantle (such as LAB: Lithosphere-Asthenosphere Boundary) are essential to investigate the evolutionary history of tectonic plates and mantle dynamics. Recent seismological studies have revealed the spatial distributions of upper mantle discontinuities beneath Australia on a large scale using surface-wave tomography (Yoshizawa & Kennett, 2015, GRL) and receiver functions (RFs) (Birkey et al., 2021, JGR). Surface-wave dispersion (SWD) enables us to reconstruct a 3-D S-wave speed model, while RFs can effectively detect the interfaces which generate converted signals from P to S (and S to P). Thus, joint inversions of SWD and RF allow us to retrieve S-wave speed models with an improved depth resolution of discontinuities (e.g., Calo et al., 2016, EPSL; Taira & Yoshizawa, 2020, GJI). In this study, we employ the joint inversion of SWD and RF in a Bayesian framework to estimate the anisotropic 1-D S-wave velocity model under permanent seismic stations in the Australian continent.
Here we apply the hierarchical trans-dimensional Bayesian inversion method to multimode SWD and RFs for incoming P-waves (P-RFs). This method can treat the data noises and the number of layers as unknown parameters so that it does not require a priori information. In this study, localized multimode SWD is derived from phase velocity maps by Yoshizawa (2014, PEPI). For the RF analysis, we subdivide the event groups by the epicentral distance range of 10º and the back-azimuth range of 5º. Such distance and back-azimuth restrictions allow us to pinpoint the location of the P-to-S conversion under the station. We have applied the method to GSN (Global Seismographic Network) stations in Australia to recover the 1-D radially anisotropic S-wave speed models with multiple discontinuities.
Figure 1 shows two selected stations (COEN and CTAO) and seismic events used in this study. Figures 2 (a) and (c) show examples of Bayesian inversion results for COEN and CTAO, where several candidates of LAB and Lehmann discontinuity (L-D) are indicated. Figures 2 (b) and (d) display potential conversion points corresponding to LAB and their depths estimated from retrieved 1-D models using various event groups. We can see some negative velocity jumps at 50-90km accompanying the radial anisotropy with faster SH than SV (Figure 2-a, c). Such shallower interfaces detected around the station (Figure 2-b, d) represent the candidates of LAB, reflecting a thinner lithosphere in eastern Australia comprised of the Phanerozoic basement. In the CTAO station, we could also detect a clear negative SV-velocity jump at 125-135km for an incoming P-wave from the west of the station (Figure 2-c). Conversion points of such deeper interfaces are about 100km away from the CTAO and located near the Tasman line. This indicates that the LAB depths may rapidly change beneath this area (Figure 2-d), supporting the stepwise LAB hypothesis in eastern Australia (Fishwick et al., 2008, Tectonics). In addition, we can observe the positive SV-velocity jumps below 200km depth, representing the candidates of L-D. They appear to accompany the rapid change of radial anisotropy. Although the Bayesian inversion requires a huge computation time, the extensive application of our inversion method to a large number of permanent and temporary stations across the continent can clarify the spatial distribution of the upper mantle discontinuity and their relationship with radial anisotropy.
Here we apply the hierarchical trans-dimensional Bayesian inversion method to multimode SWD and RFs for incoming P-waves (P-RFs). This method can treat the data noises and the number of layers as unknown parameters so that it does not require a priori information. In this study, localized multimode SWD is derived from phase velocity maps by Yoshizawa (2014, PEPI). For the RF analysis, we subdivide the event groups by the epicentral distance range of 10º and the back-azimuth range of 5º. Such distance and back-azimuth restrictions allow us to pinpoint the location of the P-to-S conversion under the station. We have applied the method to GSN (Global Seismographic Network) stations in Australia to recover the 1-D radially anisotropic S-wave speed models with multiple discontinuities.
Figure 1 shows two selected stations (COEN and CTAO) and seismic events used in this study. Figures 2 (a) and (c) show examples of Bayesian inversion results for COEN and CTAO, where several candidates of LAB and Lehmann discontinuity (L-D) are indicated. Figures 2 (b) and (d) display potential conversion points corresponding to LAB and their depths estimated from retrieved 1-D models using various event groups. We can see some negative velocity jumps at 50-90km accompanying the radial anisotropy with faster SH than SV (Figure 2-a, c). Such shallower interfaces detected around the station (Figure 2-b, d) represent the candidates of LAB, reflecting a thinner lithosphere in eastern Australia comprised of the Phanerozoic basement. In the CTAO station, we could also detect a clear negative SV-velocity jump at 125-135km for an incoming P-wave from the west of the station (Figure 2-c). Conversion points of such deeper interfaces are about 100km away from the CTAO and located near the Tasman line. This indicates that the LAB depths may rapidly change beneath this area (Figure 2-d), supporting the stepwise LAB hypothesis in eastern Australia (Fishwick et al., 2008, Tectonics). In addition, we can observe the positive SV-velocity jumps below 200km depth, representing the candidates of L-D. They appear to accompany the rapid change of radial anisotropy. Although the Bayesian inversion requires a huge computation time, the extensive application of our inversion method to a large number of permanent and temporary stations across the continent can clarify the spatial distribution of the upper mantle discontinuity and their relationship with radial anisotropy.