Observations of seismic anisotropy are a powerful tool to explore deformation and flow in the deep mantle. Recent work has explored how flow in the deepest mantle interacts with major structures such as subducting slab remnants, large low velocity provinces (LLVPs), and mantle plumes. However, a comprehensive framework describing the patterns and drivers of flow in the mantle’s bottom boundary layer is only starting to emerge. Here we target the lowermost mantle beneath Australia and the surrounding region, which encompasses slab remnants, the edge of the Pacific LLVP, and a previously identified possible mantle plume that has not yet reached the surface. We apply a recently developed approach that relies on array processing of SmKS phases, which increases signal-to-noise ratios and enables analysis of low-amplitude phases such as S3KS. We supplement our differential SmKS splitting measurements with analyses of ScS phases that sample our study area. We infer strong seismic anisotropy in localized regions, including along the southwestern edge of the Pacific LLVP and in a region south of Australia that is dominated by high seismic velocities. To provide an interpretive framework for our observations, we compare them with the results of instantaneous mantle flow models and with whole-mantle S wave tomography models. Our results support an emerging view of lowermost mantle dynamics that involves slab-driven flow, interactions between mantle flow and structures such as LLVPs, and strong deformation at the root of mantle plumes, including a plume that has not yet reached the surface.

The bottom of Earth’s mantle hosts strong seismic wave speed heterogeneities, such as ultra-low velocity zones (ULVZs). ULVZ heterogeneities are commonly detected via forward modeling of seismic waveforms, which can include time-consuming waveform synthesis and visual inspection. Furthermore, ULVZ imaging has been most commonly carried out with waves that have limited global coverage. In this work, we investigate the efficacy of the diffracted S (Sdiff) wavefield, which has global coverage to map CMB heterogeneity. We implement a Kirchhoff migration algorithm to objectively investigate the presence or absence of postcursors to Sdiff due to ULVZ heterogeneity. The Kirchhoff approach is efficient, taking less than one CPU-minute per earthquake (for ~1000 receivers) for our implementation. Our approach makes use of the expected moveout of ULVZ-born Sdiff post cursors as a function of distance from great-circle path at the base of the mantle. We investigate epicentral distances 95°, where Sdiff includes asymptotic S and ScS up to diffraction. We test the algorithm using synthetic waveforms calculated for models that include lowermost mantle wave speed heterogeneity. These results demonstrate that the migration approach can well resolve the location of heterogeneity structures in the azimuthal direction, but is less accurate at constraining the along-great circle path location in the absence of crossing ray paths. Lastly, our real-data examples detect CMB heterogeneity that agrees with past ULVZ work. Our algorithm provides a quantitative assessment of the magnitude of the waveform anomalies and, therefore, how anomalous the structures are that are producing them.

A document by Jonathan Wolf. Click on the document to view its contents.

The dependence of seismic wavespeeds on propagation or polarization direction, called seismic anisotropy, is a relatively direct indicator of mantle deformation and flow. Mantle seismic anisotropy is often inferred from measurements of shear-wave splitting. A number of standard techniques to measure shear-wave splitting have been applied globally; for example, *KS splitting is often used to measure upper mantle anisotropy. In order to obtain robust constraints on anisotropic geometry, it is necessary to sample seismic anisotropy from different directions, ideally using different seismic phases with different incidence angles. However, many standard analysis techniques can only be applied for certain epicentral distances and source-receiver ge-ometries. In this work, we apply a "wavefield differencing" approach to (systematically) understand what parts of the seismic wavefield are most affected by seismic anisotropy in the mantle. We systematically analyze differences between synthetic global wavefields calculated for isotropic and anisotropic input models, incorporating seismic anisotropy at different depths. Our results confirm that the seismic phases that are commonly used in splitting techniques are indeed strongly influenced by mantle anisotropy. However, we also identify less commonly used phases whose waveforms reflect the effects of anisotropy. For example, PS is strongly affected by upper mantle seismic anisotropy. We show that PS can be used to fill in gaps in global coverage in shear wave splitting datasets (for example, beneath ocean basins). We find that PcS is also a promising phase, and present a proof-of-concept example of PcS splitting analysis across the contiguous United States using an array processing approach. Because PcS is recorded at at much shorter distances than *KS phases, PcS splitting can therefore fill in gaps in backazimuthal coverage. The insights provided by a wavefield differencing approach provide promising new strategies for improving our ability to detect and characterize seismic anisotropy in the mantle.

Upper mantle anisotropy has been mapped beneath continents at high spatial resolution. Beneath the oceans, however, shear wave splitting constraints on upper mantle anisotropy are sparse, due to the paucity of seismic receivers. The technique that does not require the availability of seismic stations close to the region under study is differential PS-SKS splitting. Here, we use global wavefield simulations to investigate circumstances under which PS-SKS splitting can be applied, and then use this technique to measure upper mantle anisotropy beneath the Pacific Ocean basin. Our results demonstrate that upper mantle anisotropy in our study region mostly reflects shearing due to the Pacific plate. North of Fiji, we observe a rotation of fast polarization directions, away from the direction of absolute plate motion of the Pacific plate. We attribute this to far-field mantle flow effects associated with the subduction of the Australian plate beneath the Pacific.

We apply beamforming to enhance SKS, S2KS, S3KS, and PKS signals recorded across the United States, Mexico, and Central America. These phases are then used to investigate polarization- and propagation direction-dependent shear wave speeds (seismic anisotropy) in the deepest mantle beneath large swaths of North America and the northeastern Pacific Ocean. A differential splitting approach enables us to robustly identify contributions from D" anisotropy. Our results show strong seismic anisotropy in approximately half of our study region, indicating that D" anisotropy may be more prevalent than commonly thought. In some regions, the anisotropy may be induced by flow driven by sinking cold slabs, and in other, more compact regions, by upwelling flow. Measured splitting due to lowermost mantle anisotropy is sufficiently strong to be non-negligible in interpretations of SKS splitting due to upper mantle anisotropy in certain regions, which may prompt future re-evaluations of upper mantle anisotropy beneath North and Central America.

Many regions of the Earth’s mantle are seismically anisotropic, including portions of the lowermost mantle, which may indicate deformation due to convective flow. The splitting of ScS phases, which reflect once off the core-mantle boundary (CMB), is commonly measured to identify lowermost mantle anisotropy, although some challenges exist. Here, we use global wavefield simulations to evaluate commonly used approaches to inferring a lowermost mantle contribution to ScS splitting. We show that due to effects of the CMB reflection, only the epicentral distance range between 60◦ and 70◦ is appropriate for ScS splitting measurements. For this distance range, splitting is diagnostic of deep mantle anisotropy if no upper mantle anisotropy is present; however, if ScS is also split due to upper mantle anisotropy, the reliable diagnosis of deep mantle anisotropy is challenging. Moreover, even in the case of a homogeneously anisotropic deep mantle region sampled from a single azimuth by multiple ScS waves with different source polarizations (in absence of upper mantle anisotropy), different apparent fast directions are produced. We suggest that ScS splitting should only be measured at “null” stations and conduct such an analysis worldwide. Our results indicate that seismic anisotropy is globally widespread in the deep mantle.  

The dynamics of Earth’s D” layer at the base of the mantle plays an essential role in Earth’s thermal and chemical evolution. Mantle convection in D” is thought to result in seismic anisotropy; therefore, observations of anisotropy may be used to infer lowermost mantle flow. However, the connections between mantle flow and seismic anisotropy in D” remain ambiguous. Here we calculate the present-day mantle flow field in D” using 3D global geodynamic models. We then compute strain, a measure of deformation, outside the two large-low velocity provinces (LLVPs) and compare the distribution of strain with previous observations of anisotropy. We find that, on a global scale, D” material is advected towards the LLVPs. Strain is highest at the core-mantle boundary (CMB) and decreases with height above the CMB. Material outside the LLVPs mostly undergoes lateral stretching, with the stretching direction often, but not always, aligning with mantle flow direction. Strain generally increases towards the LLVPs and reaches a maximum at their edges, although models that consider recrystallization suggest that anisotropy may actually be weaker near LLVP edges. The depth-averaged strain in D” is >1.5 in almost all regions, consistent with widespread observations of seismic anisotropy. The mantle flow field and strain in D” outside of LLVPs are not very sensitive to LLVP density but are strongly controlled by local density and viscosity variations outside the LLVPs. Flow directions inferred from anisotropy observations often (but not always) align with predictions from geodynamic modeling calculations.

Seismic anisotropy has been detected at many depths of the Earth, including its upper layers, the lowermost mantle, and the inner core. While upper mantle seismic anisotropy is relatively straightforward to resolve, lowermost mantle anisotropy has proven to be more complicated to measure. Due to their long, horizontal raypaths along the core-mantle boundary, S waves diffracted along the core-mantle boundary (Sdiff) are potentially strongly influenced by lowermost mantle anisotropy. Sdiff waves can be recorded over a large epicentral distance range and thus sample the lowermost mantle everywhere around the globe. Sdiff therefore represents a promising phase for studying lowermost mantle anisotropy; however, previous studies have pointed out some difficulties with the interpretation of differential SHdiff-SVdiff travel times in terms of seismic anisotropy. Here, we provide a new, comprehensive assessment of the usability of Sdiff waves to infer lowermost mantle anisotropy. Using both axisymmetric and fully 3D global wavefield simulations, we show that there are cases in which Sdiff can reliably detect and characterize deep mantle anisotropy when measuring traditional splitting parameters (as opposed to differential travel times). First, we analyze isotropic effects on Sdiff polarizations, including the influence of realistic velocity structure (such as 3D velocity heterogeneity and ultra-low velocity zones), the character of the lowermost mantle velocity gradient, mantle attenuation structure, and Earth’s Coriolis force. Second, we evaluate effects of seismic anisotropy in both the upper and the lowermost mantle on SHdiff waves. In particular, we investigate how SHdiff waves are split by seismic anisotropy in the upper mantle near the source and how this anisotropic signature propagates to the receiver for a variety of lowermost mantle models. We demonstrate that, in particular and predictable cases, anisotropy leads to Sdiff splitting that can be clearly distinguished from other waveform effects. These results enable us to lay out a strategy for the analysis of Sdiff splitting due to anisotropy at the base of the mantle, which includes steps to help avoid potential pitfalls, with attention paid to the initial polarization of Sdiff and the influence of source-side anisotropy. We demonstrate our Sdiff splitting method using three earthquakes that occurred beneath the Celebes Sea, measured at many Transportable Array (TA) stations at a suitable epicentral distance. We resolve consistent and well-constrained Sdiff splitting parameters due to lowermost mantle anisotropy beneath the northeastern Pacific Ocean.

Shear-wave splitting measurements are commonly used to resolve seismic anisotropy in both the upper and lowermost mantle. Typically, such techniques are applied to SmKS phases that have reflected (m-1) times off the underside of the core-mantle boundary before being recorded. Practical constraints for shear-wave splitting studies include the limited number of suitable phases as well as the large fraction of available data discarded because of poor signal-to-noise ratios (SNRs) or large measurement uncertainties. Array techniques such as beamforming are commonly used in observational seismology to enhance SNRs, but have not been applied before to improve SmKS signal strength and coherency for shear wave splitting studies. Here, we investigate how a beamforming methodology, based on slowness and backazimuth vespagrams to determine the most coherent incoming wave direction, can improve shear-wave splitting measurement confidence intervals. Through the analysis of real and synthetic seismograms, we show that (1) the splitting measurements obtained from the beamformed seismograms (beams) reflect an average of the single-station splitting parameters that contribute to the beam; (2) the beams have (on average) more than twice as large SNRs than the single-station seismograms that contribute to the beam; (3) the increased SNRs allow the reliable measurement of shear wave splitting parameters from beams down to average single-station SNRs of 1.3. Beamforming may thus be helpful to more reliably measure splitting due to upper mantle anisotropy. Moreover, we show that beamforming holds potential to greatly improve detection of lowermost mantle anisotropy by demonstrating differential SKS-SKKS splitting analysis using beamformed USArray data.