Improving the resolution of earth's mantle structure through seismic imaging is important for our understanding of the deep dynamic processes that are reflected in surface tectonics and ultimately give rise to earthquakes and volcanic eruptions. Seismic tomography utilizes seismic waves originating from natural earthquakes to illuminate the earth's interior using similar principles as used in medical imagery. While the long wavelength, smooth structure is now well documented, we strive to sharpen the picture and resolve smaller scale features that will help understand in detail how the internal circulation drives the motions of tectonic plates, and vice-versa, and obtain a better view of how heat is channeled from the deep interior to the surface, resulting not only in the system of volcanoes mid-ocean ridges but also in mid-plate volcanoes called "hotspots", such as the Hawaiian chain. Because hotter than average regions correspond to low seismic velocities, which are particularly difficult to image, their accurate imaging has not been possible until recently, with the advent of numerical methods to accurately compute the seismic wavefield propagation across the earth's mantle. This project builds upon our recent efforts to develop global seismic images of the entire mantle using such computations, and improve the sharpness of these images.
The main goal of this project is to finalize the construction of higher resolution (up to 200-400 km laterally) global 3D radially anisotropic shear velocity model of the whole mantle based on a time domain waveform inversion approach that has so far been applied only to the upper mantle, with promising results. This waveform inversion methodology builds upon 20 years of experience of the PI in waveform tomography using mode-based wavefield computations, as well as two generations of global upper mantle shear velocity models developed using long period fundamental mode and overtone waveforms (periods >60s) and numerical wavefield computations (Spectral Element Method). In order to resolve lower mantle structure, isolating wavepackets that contain body wave energy is necessary, therefore we propose to extend the period range of our dataset to 30s - this will also increase spatial resolution further in the upper mantle, while keeping computations manageable. Further goals of the proposal are to (1) obtain higher resolution in the lateral variations of radial anisotropy throughout the mantle, (2) test the sensitivity of the 30 s dataset to inversion for a global averaged S/P velocity conversion factor and (3) collect additional, shorter window (~30mn after origin time) waveforms down to 20s to include more P wave energy and perform additional iterations to obtain a global P velocity model of the lower mantle.
