Resolving the Degree of Chemical Heterogeneities in the Mantle by Dual Bootstrap Stack

点击翻译按钮获取中文摘要

On Earth, surface plates are continuously recycled into the deep mantle by subduction, but it is still very poorly understood how quickly these subducted plates are mixed and homogenized by mantle convection. Observational constraints are limited in spite of innumerable speculations on the chemical state of the mantle. Previous attempts to deterministically detect small-scale heterogeneities in the mantle have relied on scattered waves with relatively high amplitudes, which can be identified in original seismic data before stacking. This limitation on signal strength leads to a bias in our detection capability, restricting us to spatially extensive features such as lateral discontinuities and dipping reflectors. Recently, a powerful signal detection technique called dual bootstrap stack (DBS) has been developed. Built with rigorous statistical tests for signal significance and coherency, DBS can achieve a clean separation of signal and noise. Unlike conventional nonlinear stacks, it can detect signals even when the signal-to-noise ratio (S/N) is lower than unity, and compared to simple linear stack, the number of traces required for unambiguous signal detection is reduced by up to two orders of magnitude. The proposed project aims to conduct a systematic test of this new signal detection technique by revisiting published data and to better understand its potential for producing convincing observational constraints on the chemical state of the mantle. More specifically, we will revisit the study of Kaneshima and Helffrich [2010], who reported an extensive study on small-scale heterogeneities in the mid-lower mantle beneath the circum-Pacific area. The existing DBS-based migration code will be modified to handle S-to-P scattering for this purpose. The number of events they used is: 23 for the Izu-Bonin region, 11 for the Kuril region, 28 for the Fiji region, and 12 for the South America. In addition to S-to-P scattering, which was exploited by this previous study, we will also seek evidence for P-to-P scattering, and if time permits, we will investigate other events to see how reproducible detected scatterers are, in light of DBS-based teleseismic migration.