S31B-01
Localized seismic scatterers near the core-mantle boundary beneath the Caribbean Sea: evidence from PKP precursors
The nature of the core-mantle boundary (CMB) region, the D", has been a focus of research for decades because it is a crucial part in understanding the evolution of geodynamics and whole earth structure. The D" region is very heterogeneous and has complicated seismic structures that require involvements of both chemical and thermal processes. One representative and well studied area is Central America and Caribbean, where the D" region is featured by a well defined seismic discontinuity underlain by anisotropic and higher velocity materials. Here we used seismic phase arrivals from earthquakes in the Western Pacific recorded by broadband seismic stations in northern South America and the Caribbean to investigate the core-mantle boundary and the D" region beneath the central Caribbean plate. We identified precursors to the PKP phases in 14 events, which can be explained by a region of heterogeneous scattering located above the CMB beneath the central Caribbean plate. The seismic scatterers are localized to the vicinity between approximately 5°N to 20°N and -80°W to - 65°W. Other similar distance events that display no precursors, interpreted to have traversed non- scattering regions on the core-mantle boundary, lie to the south and southwest of the identified cluster of seismic scatterers. We found that the small-scale seismic scatterers are not uniformly distributed, are concentrated, and are surrounded by relatively fast shear velocity perturbations in the lower mantle imaged with global seismic tomography. The fast velocity perturbations may be residual slab from the subducted Farallon plate and the scatterers to the east-northeast of the high velocity material could be remnants of heterogeneous material, perhaps former basaltic crust, that has been transported ahead or on top of the subducted slab material as it impacts and bends at the core-mantle boundary.
S31B-02
Lower mantle phase-boundary variability
A lower mantle S-wave triplication with a Scd branch occurring between S and ScS appears to be explained by a recently discovered Perovskite (PV) to Post-Perovskite (PPV) phase-change. It is predicted to have a positive Clapeyron slope (γ) between 5 to 13 MPa/K with a small S-velocity jump (1.5 to 4%) and an even smaller 1 to 2% jump in P-velocity. Seismic observations indicate that Scd arrives earlier and stronger beneath fast regions (circum-Pacific) than slow regions (super plumes) indicating a positive γ. However, it proves difficult to separate effects produced by downwelling (slab debris) from upwelling (plumes) in refining the actual physical properties. Here we model dense record sections collected from USArray and existing PASSCAL data to isolate effects produced by lower mantle structure as evidenced by P and S tomographic models, to better define the seismic phase-change properties beneath Central America. We find that the PV-PPV velocity jump is twice as strong beneath slow regions as fast regions requiring distinct reference heights indicative of changing chemistry. Moreover, the edges of the supposed buckled slabs deliminated by both P and S-waves display very rapid changes in phase-boundary heights producing Scd multipathing. These features can explain the unstable nature of this phase with easy detection to no detection commonly observed. The fine structure at the base of the mantle beneath these edges contains particularly strong reflections indicative of local ultralow velocity zones, which is predicted by some dynamic models.
S31B-03
Empirical transfer functions: Application to the determination of outermost core velocity structure using teleseismic SmKS phases
Teleseismic SmKS waves propagate as S-waves in the mantle and compressional (K) waves in the core, with m-1 underside bounce points at the core-mantle boundary. For long-period or broadband recordings at epicentral distances of 115-135°, higher-order SmKS waves (3 ≤ m < ∞) are not often discernible as distinct pulses. Instead, they are typically manifested as a weakly dispersive waveform that lags SKKS by ~ 12-32s. In a ray-theoretical representation of this process, there is a strong geometrical similarity between the coalescence of SmKS turning waves to form a composite arrival and the interference of mantle S waves to form teleseismic Love waves. SmKS waves can thus be viewed as a type of pseudo-interface wave, the dispersive properties of which depend strongly on the fine-scale velocity structure of the outermost core. In order to analyze SmKS arrivals, we have developed an empirical transfer-function (ETF) technique that uses SKKS as a reference phase. An ETF is a wave-shaping filter that transforms the observed SKKS pulse into the observed SmKS pulse. We obtain this filter by windowing the respective pulses and applying frequency- domain Wiener deconvolution. Each ETF contains SmKS-SKKS differential arrival-time, phase-shift and relative-amplitude information; it also implicitly removes the source-time function and instrument response, thus facilitating the merging of results from different stations and events. Here, we apply this approach to global observations of SmKS phases and invert the results to yield a new velocity model for the outermost core region.
S31B-04
Post-perovskite Lens and Heat Flux at the CMB Under the Cocos Plate
Investigations of structure in the lowermost mantle have revealed evidence for one or more velocity discontinuities in the lowermost mantle. The existence of a lower mantle velocity discontinuity can be interpreted as evidence for a phase change, with the most commonly discussed possibility being the transformation from perovskite to post- perovskite in relatively low temperature regions of the boundary layer. Subduction of Farallon Plate may account for low mantle temperatures in the deep mantle under the Caribbean and, if material has displaced westward above the core-mantle boundary, under the Cocos plate. But, even in a cooled region, the temperature at the core- mantle boundary will be the same as elsewhere, so a very steep thermal gradient is expected in the very regions where post-perovskite may exist, creating favorable conditions for a second intersection of the phase boundary and conversion back to perovskite. This would form a lens of post-perovskite bounded by velocity discontinuities above and below, of opposite sign and equal magnitude. Such paired discontinuities are being sought, as they can intrinsically provide two P-T points in a given region that allow the thermal gradient to be determined, and with assumption of a thermal conductivity, the heat flux can be estimated. The evidence for such structures is considered and the implications for heat flux in the putatively low temperature region near the core-mantle boundary below the Cocos plate is discussed.