High‐Resolution Mantle Transition Zone Imaging Using Multi‐Dimensional Reconstruction of SS Precursors

Abstract

AbstractThe SS precursors have been extensively utilized in mapping the mantle transition zone (MTZ). However, their applications are often challenged by weak phases that arise from small impedance contrasts of the mantle discontinuities, noise contamination, and localized thermal/compositional heterogeneities. We develop a new data processing workflow for more reliable MTZ imaging by adopting the recently proposed robust damped rank‐reduction (RDRR) method from exploration seismology. This method exploits the signal coherency in the four‐dimensional data and allows simultaneously attenuating noise and interpolating missing traces. We utilize synthetic data sets generated with realistic earth structures and MTZ topography to evaluate the capability of the proposed workflow. Our test results show that the RDRR method can well capture the topographic variation of mantle discontinuities, improving the SNR of the SS precursor data by an order of magnitude. The application to SS precursors from the western Pacific successfully removes contaminating noises, mitigates imaging artifacts of small‐scale anomalies, and improves the lateral coherency of the MTZ structure, revealing a clear first‐order structural transition. Compared with earlier global and regional models, our model reveals more structural details including a localized thin MTZ near the Changbai volcano in East Asia. This observation, in conjunction with the reported low‐velocity structure in the MTZ in earlier tomographic studies, may support the presence of deep mantle up‐welling through a slab gap. In summary, our work enables resolving MTZ structures with high fidelity and highlights the importance of advanced array methods in improving SS precursor imaging.