Three-Dimensional Structure of the Mantle Transition Zone Revealed by High-Resolution Wavefield Imaging with USArray Dat

The USArray is the largest seismic array ever deployed to study the deep interior of the Earth. It enables us to study the nature of the mantle transition zone in unprecedented detail that was impossible to image before. In this study, I developed a novel method of processing teleseismic P-waves called the generalized iterative deconvolution method to improve the resolution of receiver functions for imaging. I then applied the recently developed three-dimensional, plane wave migration method to image the mantle down to 1000 km under the contiguous US. The result is the highest resolution image of the three-dimensional structure of the transition zone ever produced. I use this image to study two features of the transition zone: (1) topographic variation of the 410-km and 660-km discontinuities and (2) seismic heterogeneity within the transition zone. I found that the 410 and 660 discontinuities are characterized by topographic variation at all resolvable scales, which is limited by diffraction, not data density. The large variation of amplitude that correlates with topography on the discontinuities suggests that both the 410 and 660 discontinuities are rough at scales smaller than inferred from any previous study. The correlation of the roughness with the inferred location of the subducted Farallon slab implies that the discontinuities are dynamic. The final section of this study brings a new insight on the transition zone made possible by the unprecedented resolution. I found that seismic waves are scattered strongly within the transition zone with a strong preference towards negative scattering objects. I argue that two different models of compositional heterogeneity can explain this observation: a set of randomly distributed blobs of relatively low-density material, or diapir structures from small-scale instabilities. The origin of heterogeneity could be either compositional relics of subducting slabs or alteration of mineralogy by water. The latter is preferred in light of related work in geodynamic modeling and mineral physics. If true, this result will be helpful in estimating the spatial distribution of water within the transition zone with broader implications for the global-scale volume of water in the planet.