Seismic Structure of the Upper Mantle Beneath Eastern Asia From Full Waveform Seismic Tomography

Abstract

To better understand the subsurface behavior of subducting slabs and their relation to the tectonic evolution of the overriding plate, we conduct a full waveform inversion on a large data set to determine a high‐resolution seismic model, FWEA18 (Full Waveform inversion of East Asia in 2018), of the upper mantle beneath eastern Asia. FWEA18 reveals sharper, more intense high‐velocity slabs in the upper mantle under the southern Kuril, Japan, and Ryukyu arcs, than previous studies have found. The subducting Pacific plate is imaged as a roughly 100 km thick high‐velocity slab to near 550 km depth indicating relatively little deformation. Stagnation near 600 km depth is observed over horizontal distances of 600 km or less. The Pacific plate we image accounts for roughly 25 Myr of subduction with older slab likely located in the lower mantle. The Philippine plate, subducting beneath the Ryukyu Islands, has a clear termination at about 450 km depth. This may indicate a tearing event in the past or that less Philippine Sea plate has subducted than previously thought. We found a double‐layer high‐velocity anomaly above and below 660 km under the Yellow Sea and eastern coast of North China. This may correspond to parts of the Philippine Sea plate that detached in the past and Pacific plate that have intersected at depth or a complicated behavior of the Pacific plate at that depth. Slow cylindrical anomalies cross the entire upper mantle are imaged beneath major Holocene volcanoes, which are likely upwellings associated with the edges of deep slabs that are entering the lower mantle.

Seismic Structure Beneath the Tibetan Plateau From Iterative Finite‐Frequency Tomography Based on ChinArray: New Insights Into the Indo‐Asian Collision

“…Besides, the total number of measurement windows, which were automatically selected by FLEXWIN, also increase considerably by 25%. The seismic structure of TP2019 reveals more details than M00 (Figure ) and depicts the eastern cratons as clearly as previous high‐resolution tomographic models (Shen et al, ; Tao et al, ).…”

“…Seismic structure of M00 and TP2019 and several tomographic models. SL2013sv (Schaeffer & Lebedev, ), FWEA18 (Tao et al, ), and Shen2016 (Shen et al, ) are high‐resolution seismic models of the globe, East Asia, and mainland China, respectively.…”

“…The continuous deployment of seismic stations opened the avenue for an improvement of seismic tomography resolving power (Liu & Gu, ). Due to the high performance computer and technical progress in numerical methods (e.g., Komatitsch & Tromp, , ), full‐waveform tomography based on 3‐D numerical simulations at regional and global scales has become feasible (e.g., Bozdağ et al, ; Chen, Niu, et al, ; Konishi et al, ; Fichtner et al, ; Tao et al, ; Tape et al, ; Wang et al, ; Zhu & Tromp, ; Zhu et al, ). In this paper, we present a new seismological model with enhanced clarity on the deep structure of Tibet and its surrounding areas, taking advantage of both the regional dense‐array and iterative finite‐frequency tomography (Tape et al, ).…”

Three‐Dimensional S Velocity Structure of the Mantle Transition Zone Beneath Central America and the Gulf of Mexico Inferred Using Waveform Inversion

“…Hence, the total computation time is 2,106 CPU hours, or ~105 CPU hours per event. In contrast, a recent study inferring the 3‐D structure of the MTZ beneath Japan using adjoint tomography reported using ~896 CPU hours per event per iteration (using 336 cores), and a total of 18 iterations (Tao et al, ), although CPU performance used in each study may be different from each other.…”

“…However, they are usually not used in 3‐D tomography using large data sets. Exceptions are (1) the recent models of the Berkley group (e.g., French & Romanowicz, ), who used relatively long‐period waveforms (periods of 30 s and longer), which can thus resolve relatively broad features, and also used fewer stations in the United States than this study; (2) the study of Van der Lee and Nolet (), who inverted radial component S wave waveforms to image the 3‐D S velocity structure beneath the western United States and Central America; (3) the global S wave travel time tomography model of Grand (); (4) the P wave travel time tomography study of Takeuchi et al (); and (5) the adjoint tomography study of Tao et al (), who used a cross‐correlation misfit function to image the upper mantle and transition zone beneath Japan with higher resolution than in previous models. Other recent studies (e.g., Zhu et al, ) used waveform inversion to image the upper mantle and transition zone beneath the contiguous United States using a phase misfit function.…”