Seismic imaging of the global boundaries of the mantle transition zone provides important constraints regarding mantle composition and convection. Additionally, modern hypotheses about mantle convection and increasing seismic data resources motivate attempts to image more complex regional transition zone structures such as dipping and discontinuous interfaces. Here we extend a 3-D prestack migration method to make it more applicable for transition zone imaging with Ps receiver functions. After validation with 1-D synthetic data, two types of hypothetical structures are adopted to demonstrate the strengths and weaknesses of different practical imaging parameters with 2-D synthetic tests. The results show that the method can resolve dipping anomalies and laterally discontinuous low velocity layers. However, receiver spacing of 0.5°is inadequate to accurately migrate scattering from all possible angles in the transition zone at~0.3 Hz. Application of a slowness cutoff window about the Ps raypath can mitigate artifacts due to low receiver density, but the tradeoff is a decrease in the maximum resolvable dip angle. Further synthetic tests evaluate source wavelet effects, addition of observed seismic noise to synthetic data, and imaging grid dimensions. Cumulatively, the tests indicate that generic application of migration algorithms to the limited source-receiver distributions relevant for the transition zone should be treated cautiously. Imaging parameters, such as the slowness cutoff and wavelet, should be chosen based on synthetic tests for hypothetical targets and realistic source-receiver geometries. Finally, observational Ps receiver functions from the USArray are migrated with the new method.Plain Language Summary This study extends a seismic imaging method to make it more applicable to the mantle transition zone depths (from 300 to 800 km) because seismic imaging of sharply defined structures in this layer of Earth provide important insights into mantle composition and convection. The method was tested with realistic combinations of synthetic earthquake sources and seismic network densities to evaluate how it would perform for hypothetical transition zone structures. The results are encouraging in that they show the possibility of detecting increasingly complex structures in this layer of Earth's interior, but they also show that realistic combinations of earthquakes and seismographs present challenges for generic application of imaging methods to the transition zone. Examples from a suite of synthetic tests outline how the new imaging method may be optimized to resolve some complex structures with realistic seismic data resources and avoid introducing artificial features that could confound structural interpretations.

Seismic observations of the mantle discontinuities at depths of 410 and 660 km provide crucial information to constrain the thermal and compositional states and further shape our understanding of mantle dynamics. The topography and impedance contrast of the mantle discontinuities, at both global and regional scales, have been imaged through differential travel times and amplitude ratios between SS and its precursors. However, a relatively simple reference phase (SS) is often assumed, which may overlook potential bias caused by near‐surface complexities. To assess these effects, we developed an algorithm to simulate the waveforms of SS and its precursors efficiently based on the propagation matrix method. We found that the crustal thickening increases the differential travel times but does not change the amplitude ratios much, while the sedimentary layer significantly affects both the differential travel times and amplitude ratios. The global anomalies of differential travel times and amplitude ratios are mapped incorporating earth models from CRUST1.0 and PREM. The time shift can reach ∼9 s, which will bring a depth bias of mantle discontinuity of more than 20 km, while the amplitude ratio anomaly is up to ∼40%, which will make the impedance contrast of the mantle discontinuity overestimated. The near‐surface complexities cannot be ignored in the investigation of mantle discontinuities with SS precursors, especially when the bounce points located in the regions with sediment.

The 410 km discontinuity (410) represents the solid phase transition between olivine and wadsleyite, and the 660 km discontinuity (660) is primarily attributed to the decomposition of ringwoodite into bridgmanite and ferropericlase (e.g., Frost, 2008). However, the majorite garnet to bridgmanite transition may also contribute to changing seismic properties near the 660 and debate continues about whether the 660 may be partly attributed to compositional changes across a leaky boundary in the mantle convection system (

No abstract