Subduction zones are convergent zones in plate tectonics and are one of the most important parts of the geodynamic system of the Earth’s interior. For oceanic subduction zones, the presence of even small amounts of serpentinite in the subducted lithosphere is important for the water cycle into the deep earth. The breakdown of serpentinite to olivine, orthopyroxene, and water results in a very large density increase of the subducted material, but water may also be transported in serpentines and dragged down with the slab, which means the interior structure of the subducted slabs are likely showing more complicated internal structure than a single high velocity zone. The resolution of the fine scale velocity structure of the slab can allow insight into many processes related to material cycling into the deep earth. At the same time, Seismograms carry a lot of detailed information about the heterogeneous seismic velocity structure along the path that the seismic wave front passed and distorted (Figure 1), which could be recorded by a dense network.
Our project incorporates two and three dimensional high-accuracy seismic modelling and inversion, including the finite-difference and spectral-element discretisation (Gokhberg and Fichtner, 2016) of the seismic wave equation combined with adjoint techniques which could support 3D heterogeneous visco-elastic rheologies with radial anisotropy. Taking advantage of the multi-phases of teleseismic down-dip and regional up-dip seismic wavefields, we target to extract seismic velocity structure of the slabs including its shape and sharpness from different ray paths and even the interior anisotropic structure of the slabs.