From People - School of Science, The University of Tokyo
Structure of the Earth's interior, Seismic wave propagation, Numerical simulation
< Research Objectives > Our group's main research interest is determining the structure of the Earth's interior by inversion of seismic waveform data. Better knowledge of the elastic and anelastic properties of the Earth's interior will assist in obtaining a better understanding of geodynamic processes. In contrast to methods that analyze intermediate parameters such as travel times of body waves or phase velocities of surface waves, waveform inversion allows all of the information contained in the observed seismograms to be utilized. In order to make it practical to perform waveform inversion we have developed accurate and efficient methods for computing synthetic seismograms using optimally accurate numerical operators.
< Research to Date > Many workers have inverted surface wave waveform data for the 3-D structure of the upper mantle, but the problems associated with inversion of body wave waveform data for 3-D Earth structure are far more computationally intensive. For this reason, although our methods are formulated to allow inversion of body-wave waveform data for 3-D structure, we have focused to date on inversion of body-wave waveform data for localized 1-D S-velocity structure.
Up until about 2006 our work was primarily theoretical and computational, focusing on developing methods, algorithms, and software for waveform inversion, but since then we have focused mainly on analyses of observed data to determine Earth structure, with particular emphasis on the D" layer (the lowermost several hundred km at the base of the mantle) and the transition zone between the upper mantle and lower mantle. Waveform inversion allows us to observe the variation of S-velocity with depth within the D" layer, as opposed to just the average velocity in D" yielded by travel-time tomography.
In our recent work we found that in regions where the average velocity in D" is faster than the global average model PREM, such as beneath central America, the north pole, and central Asia, the velocity in the upper half of D" is considerably faster than PREM, whereas in the lower half of D" the velocity drops down to that of PREM. In contrast, in regions beneath the Pacific where the average velocity in D" is slower than PREM, we found an "S-shaped" model in which the velocity decreases at about 300 km above the core, and then increases at greater depths. The above results, taken together, provide important information for constraining the thermal state and material properties of the lowermost mantle.
< Future Research > We are also continuing to refine our methods and software for the forward and inverse problems. We have also carried out inversion for the local 1-D models of the variation of the anelasticity parameter (Q). In the near future we are planning to carry out inversion of body-wave waveform data for localized 3-D seismic velocity and anelastic structure.