HR: 17:45h
AN: S34C-08    [Abstracts]
TI: Shear Velocity Structure of the Lowermost Mantle as Revealed by Multiscale Finite-Frequency Differential Traveltime Tomography
AU: * Hung, S
EM: shung@ntu.edu.tw
AF: Department of Geosciences, National Taiwan University, Taipei, 106 Taiwan
AU: Zhao, L
EM: zhaol@usc.edu
AF: Institute of Earth Sciences, Academia Sinica, Nankang, Taipei, 115 Taiwan
AU: Chiao, L
EM: chiao@ntu.edu.tw
AF: Institute of Oceanography, National Taiwan University, Taipei, 106 Taiwan
AU: Kuo, B
EM: byk@ntu.edu.tw
AF: Institute of Earth Sciences, Academia Sinica, Nankang, Taipei, 115 Taiwan
AB: Recent progress in finite-frequency tomography has led to compelling evidence for deep origins of some hotspots, perhaps down to the core-mantle boundary (CMB). Finite-frequency Fréchet kernels which translate observed body wave delay times into 3D velocity perturbations help better resolve the strength and scale length of mantle heterogeneity beyond the resolution limits of ray-based models. Though many seismological studies indicate that the degree of elastic property heterogeneity is enhanced and present on all scales near the base of the mantle, only long wavelength structures are well correlated among various models. Not only will different interpretation theories result in different tomographic images, but so will different choices of parameterization and regularization schemes because of uneven wave path coverage. Parameterization in terms of spherical harmonics or regular blocks tends to suppress either spatial or spectral resolutions in good-sampling areas. Data-adaptive multiscale parameterizations have been recently developed to bring these two extreme models closer together by recovering the variable-scale robust features locally with the preservation of their spectral contents. In this study we explore the deep mantle by mapping measured S(S_d)-SKS and ScS-S differential times into 3D shear velocity variation. Despite that the ray-theoretical traveltime kernels applied in the current teleseismic tomography are very computationally efficient allowing the use of massive data sets, they are not accurate enough for the core grazing and diffracting S waves providing unique sampling of the lowermost mantle. With the advanced power of parallel computing, we are able to construct S(S_d)-SKS differential kernels by means of the general normal mode coupling summation. Moreover, the resolvable shear velocity heterogeneity is automatically retrieved through multiresolution representation of the pursued model based on 3D spherical wavelets.
DE: 3260 Inverse theory
DE: 3285 Wave propagation (0689, 2487, 4275, 4455, 6934)
DE: 7208 Mantle (1212, 1213, 8124)
DE: 7260 Theory
DE: 7270 Tomography (6982, 8180)
SC: Seismology [S]
MN: Fall Meeting 2005