Low Shear-wave Velocity Provinces Research Articles

Heavy Mo isotope enrichment in the Pitcairn plume: Implications for the subduction cycle of anoxic sediments

Subduction redistributes elements between Earth's principal geochemical reservoirs, modifying the chemical composition of Earth's mantle, crust, atmosphere, and hydrosphere, and consequently having an impact on the evolution of life itself. Subduction of surface material that has been geochemically modified by low-temperature processes leads to mineralogical and chemical heterogeneities in mantle reservoirs over time and is recorded in modern ocean island basalts. One of the principal geochemical end members of the heterogeneous deep mantle, the enriched mantle 1 (EM-1) source of Pitcairn Island, has been attributed to the contribution of crustal material with vastly different chemical compositions and ages. The Mo isotope composition of lavas from Pitcairn Island constrains the nature of this recycled crustal component. Pitcairn lavas have elevated δ98/95Mo relative to the depleted mantle. The high δ98/95Mo is associated with high time-integrated 232Th/238U and 87Rb/86Sr, and low time-integrated 147Sm/144Nd and 238U/204Pb. These characteristics can be attributed to the recycling of nearly pristine pelagic sediments that were deposited in a Proterozoic anoxic deep-ocean into the sources of the Pitcairn Island lavas. The isotope composition of these lavas is similar to that of EM-1 hotspots from the South Atlantic, indicating the addition of reduced sediments in both of Earth's large low shear wave velocity provinces (LLSVPs). Consistent data from both locations imply that the subduction cycling of sedimentary redox-sensitive elements such as Mo, S, Se, and U into arc magmas was in these cases inefficient in the Precambrian and the chemical and isotopic signature of reduced sediments is preserved in the source of ocean island basalts bearing the EM-1 characteristics.

New evidence for the Ontong Java Nui hypothesis

The formation of the Ontong Java Nui super oceanic plateau (OJN), which is based on the model that the submarine Ontong Java Plateau (OJP), Manihiki Plateau (MP), and Hikurangi Plateau (HP) were once its contiguous fragments, could have been the largest globally consequential volcanic event in Earth’s history. This OJN hypothesis has been debated given the paucity of evidence, for example, the differences in crustal thickness, the compositional gap between MP and OJP basalts and the apparent older age of both plateaus relative to HP remain unresolved. Here we investigate the geochemical and 40Ar-39Ar ages of dredged rocks recovered from the OJP’s eastern margin. Volcanic rocks having compositions that match the low-Ti MP basalts are reported for the first time on the OJP and new ~ 96–116 Ma and 67–68 Ma 40Ar-39Ar age data bridge the temporal gap between OJP and HP. These results provide new evidence for the Ontong Java Nui hypothesis and a framework for an integrated tectonomagmatic evolution of the OJP, MP, and HP. The isotopic data imply four mantle components in the source of OJN that are also expressed in present-day Pacific hotspots sources, indicating origin from (and longevity of) the Pacific Large Low Shear-wave Velocity Province.

Contrasts in 2-D and 3-D system behaviour in the modelling of compositionally originating LLSVPs and a mantle featuring dynamically obtained plates

SUMMARY More than two decades of systematic investigation has made steady progress towards generating plate-like surface behaviour in models of vigorous mantle convection. Accordingly, properties required to obtain dynamic plates from mantle convection have become widely recognized and used in both 2-D and 3-D geometries. Improving our understanding of the properties required to obtain durable (or replenishable) deep mantle features with LLSVP-like characteristics has received interest for a period with similar longevity. Investigation ultimately focuses on discovering the properties able to produce the presence of a detached pair of 3-D features, distinct from the ambient mantle. Here, we assume the large low shear-wave velocity provinces (LLSVPs) have a chemical origin by incorporating a compositionally anomalous and intrinsically dense (CAID) mantle component comprising 2–3.5 per cent of the total mantle volume. The feedback between plate formation and the presence of a CAID mantle component is investigated in both 2-D and 3-D spherical geometries. We explore the impact of both an intrinsic contrast in density and viscosity for the CAID component, with the objective of finding system parameter values that encourage the formation of a pair of LLSVP-like assemblages and a surface that exhibits the principle features of terrestrial plate tectonics; including recognizable and narrowly focused divergent, convergent and (in 3-D) transform plate boundaries that separate 8–16 distinct plate interiors. We present the results of nine 2-D and 11 3-D calculations and show that for some of the cases examined, a pair of CAID material provinces can be freely obtained in 2-D cases while maintaining a surface characterized by plate-like behaviour. However, specifying the same system parameters in the 3-D model does not readily yield a pair of enduring provinces for any values of the parameters investigated. Moreover, the inclusion of the CAID component in the mantle can affect the global geotherm so that in comparison to the surface behaviour obtained for the initial condition isochemical model, the surface behaviour of the cases incorporating the dense component are less exemplary of plate tectonics. In general, CAID material components that are 3.75–5 per cent denser than the surrounding mantle (at surface temperatures), and up to a factor of 100 times greater in intrinsic viscosity, form layers populated by voids, or nodes connected by tendril-like ridges that reach across the core–mantle boundary (CMB), rather than distinct piles resembling LLSVPs. Due to its inherently heavy and stiff character, in equilibrated systems, we find the CAID material becomes especially hot so that the temperature-dependence of its density and viscosity results in reduced distinction between the intrinsically dense assemblages and the ambient mantle. Accordingly, the CAID material forms masses on the CMB that are relatively less dense (0.625–1.5 per cent) and viscous than the adjacent mantle material, in comparison to the percentage differences obtained at common temperatures. We find that by adjusting our yield stress model to account for the influence of the CAID material on the geotherm, a highly satisfactory plate-like surface can be re-attained, however, the formation of a pair of LLSVP-shaped masses remains elusive.

D″ anisotropy inverted from shear wave splitting intensity

The D″ layer, located at the bottom of the mantle, is an active thermochemical boundary layer. The upwelling of mantle plumes, as well as possible plate subduction in the D″ layer, could lead to large-scale material transformation and mineral deformation, which could result in significant seismic anisotropy. However, owing to limited observations and immense computational cost, the anisotropic structures and geodynamic mechanisms in the D″ layer remain poorly understood. In this study, we proposed a new inversion method for the seismic anisotropy in the D″ layer quantitatively with shear wave splitting intensities. We first proved the linearity of the splitting intensities under the ray-theory assumption. The synthetic tests showed that, with horizontal axes of symmetry and ray incidences lower than 30º in the D″ layer (typical SKS phase), the anisotropy is well resolved. We applied the method to the measured dataset in Africa and Western Europe, and obtained strong D″ anisotropy in the margins of the large low shear-wave velocity provinces and subducting slabs. The new method makes it possible to obtain D″ anisotropy, which provides essential constraints on the geodynamical processes at the base of the mantle.

Spatial Characteristics of Recycled and Primordial Reservoirs in the Deep Mantle

AbstractThe spatial distribution of the geochemical domains hosting recycled crust and primordial (high‐3He/4He) reservoirs, and how they are linked to mantle convection, are poorly understood. Two continent‐sized seismic anomalies located near the core‐mantle boundary—called the Large Low Shear Wave Velocity Provinces (LLSVPs)—are potential geochemical reservoir hosts. It has been suggested that high‐3He/4He hotspots are spatially confined to the LLSVPs, hotspots sampling recycled continental crust are associated with only one of the LLSVPs, and recycled continental crust shows no relationship with latitude. We reevaluate the links between LLSVPs and isotopic signatures of hotspot lavas using improved mantle flow models including plume conduit advection. While most hotspots with the highest‐3He/4He can indeed be traced to the LLSVP interiors, at least one high‐3He/4He hotspot, Yellowstone, is located outside of the LLSVPs. This suggests high‐3He/4He is not geographically confined to the LLSVPs. Instead, a positive correlation between hotspot buoyancy flux and maximum hotspot 3He/4He suggests that it is plume dynamics (i.e., buoyancy), not geography, which determines whether a dense, deep, and possibly widespread high‐3He/4He reservoir is entrained. We also show that plume‐fed EM hotspots (enriched mantle, with low‐143Nd/144Nd), signaling recycled continental crust, are spatially linked to both LLSVPs, and located primarily in the southern hemisphere. Lastly, we confirm that hotspots sampling HIMU (“high‐μ,” or high 238U/204Pb) domains are not spatially limited to the LLSVPs. These findings clarify and advance our understanding of deep mantle reservoir distributions, and we discuss how continental and oceanic crust subduction is consistent with the spatial decoupling of EM and HIMU.

The tungsten-182 record of kimberlites above the African superplume: Exploring links to the core-mantle boundary

Many volcanic hotspots are connected via ‘plume’ conduits to thermochemical structures with anomalously low seismic velocities at the core-mantle boundary. Basaltic lavas from some of these hotspots show anomalous daughter isotope abundances for the short-lived 129I-129Xe, 146Sm-142Nd, and 182Hf-182W radioactive decay systems, suggesting that their lower mantle sources contain material that dates back to Earth-forming events during the first 100 million years in solar system history. Survival of such ‘primordial’ remnants in Earth's mantle places important constraints on the evolution and inner workings of terrestrial planets.Here we report high-precision 182W/184W measurements for a large suite of kimberlite volcanic rocks from across the African tectonic plate, which for the past 250 million years has drifted over the most prominent thermochemical seismic anomaly at the core-mantle boundary. This so-called African LLSVP, or ‘large low shear-wave velocity province’, is widely suspected to store early Earth remnants and is implicated as the ultimate source of global Phanerozoic kimberlite magmatism. Our results show, however, that kimberlites from above the African LLSVP, including localities with lower mantle diamonds such as Letseng and Karowe Orapa A/K6, lack anomalous 182W signatures, with an average μ182W value of 0.0 ± 4.1 (2SD) for the 18 occurrences studied.If kimberlites are indeed sourced from the African LLSVP or superplume, then the extensive 182W evidence suggests that primordial or core-equilibrated mantle materials, which may contribute resolvable μ182W excesses or deficits, are only minor or locally concentrated components in the lowermost mantle, for example in the much smaller ‘ultra-low velocity zones’ or ULVZs. However, the lack of anomalous 182W may simply suggest that low-volume kimberlite magmas are not derived from hot lower mantle plumes. In this alternative scenario, kimberlite magmas originate from volatile-fluxed ambient convecting upper mantle domains beneath relatively thick and cold lithosphere from where previously ‘stranded’ lower mantle and transition zone diamonds can be plucked.

Distinct formation history for deep-mantle domains reflected in geochemical differences

The Earth’s mantle is currently divided into the African and Pacific domains, separated by the circum-Pacific subduction girdle, and each domain features a large low shear-wave velocity province (LLSVP) in the lower mantle. However, it remains controversial as to whether the LLSVPs have been stationary through time or dynamic, changing in response to changes in global subduction geometry. Here we compile radiogenic isotope data on plume-induced basalts from ocean islands and oceanic plateaus above the two LLSVPs that show distinct lead, neodymium and strontium isotopic compositions for the two mantle domains. The African domain shows enrichment by subducted continental material during the assembly and breakup of the supercontinent Pangaea, whereas no such feature is found in the Pacific domain. This deep-mantle geochemical dichotomy reflects the different evolutionary histories of the two domains during the Rodinia and Pangaea supercontinent cycles and thus supports a dynamic relationship between plate tectonics and deep-mantle structures. Earth’s deep-mantle domains are geochemically distinct. The African domain is enriched in subducted material, which suggests a different history from the Pacific domain and a dynamic relationship between plate tectonics and deep-mantle structures.

Plume interaction and mantle heterogeneity: A geochemical perspective

Mantle plumes originating from the Core-Mantle Boundary (CMB) or the Mantle Transition Zone (MTZ) play an important role in material transfer through Earth’s interior. The hotspot-related plumes originate through different mechanisms and have diverse processes of material transfer. Both the Morganian plumes and large low shear wave velocity provinces (LLSVPs) are derived from the D″ layer in the CMB, whereas the Andersonian plumes originate from the upper mantle. All plumes have a plume head at the Moho, although the LLSVPs have an additional plume head at the MTZ. We compare the geochemical characteristics of various plumes in an attempt to evaluate the material exchange between the plumes and mantle layers. The D″ layer, the LLSVPs and the Morganian plumes are consisted of subducted slab and post-perovskite from the lower mantle. Bridgmanite would crystallize during the upwelling process of the LLSVPs and the Morganian plumes in the lower mantle, and the residual is a basalt-trachyte suite. Unlike the Morganian plumes, the crystallization in the LLSVPs is insufficient that material accumulates beneath the MTZ to form a plume head. Typically, the secondary plumes above the plume head occur at the edge of the LLSVPs because it is easier for bridgmanite crystal separating from the plume head at the edge, and the residual material with low density upwells to form the secondary plumes. Meanwhile, Na and K are enriched during the long-term crystallization process, and then the basalt-phonolite suite appears in the LLSVPs. The geochemical characteristics of Andersonian plumes suggest that the basalt-rhyolite suite is the major component in the upper mantle. Meanwhile the basalt-rhyolite suite also appears in the LLSVPs and the Morganian plumes because of the assimilation and contamination in the plume head beneath the Moho.

The dynamics and impact of compositionally originating provinces in a mantle convection model featuring rheologically obtained plates

SUMMARY Previous geodynamic studies have indicated that the presence of a compositionally anomalous and intrinsically dense (CAID) mantle component can impact both core heat flux and surface features, such as plate velocity, number and size. Implementing spherical annulus geometry mantle convection models, we investigate the influence of intrinsically dense material in the lower mantle on core heat flux and the surface velocity field. The dense component is introduced into a system that features an established plate-like surface velocity field, and subsequently we analyse the evolution of the surface velocity as well as the interior thermal structure of the mantle. The distribution and mobility of the CAID material is investigated by varying its buoyancy ratio relative to the ambient mantle (ranging from 0.7 to 1.5), its total volume (3.5–10 per cent of the mantle volume) and its intrinsic viscosity (0.01–100 times the ambient mantle viscosity). We find at least three distinct distributions of the dense material can occur adjacent to the core–mantle boundary (CMB), including multiple piles of varying topography, a core enveloping layer and two diametrically opposed provinces (which can on occasion break into three distinct piles). The latter distribution mimics the morphology of the seismically observed large low shear wave velocity provinces (LLSVPs) and can occur over the entire range of CAID material viscosities. However, diametrically opposed provinces occur primarily in cases with CAID material buoyancy numbers of 0.7–0.85 (corresponding to contrasts in density between ambient and CAID material of 130 and 160 kg m−3, respectively) in our model (with an effective Rayleigh number of order 106). Steep and high topography piles are also obtained for cases featuring buoyancy ratios of 0.85 and viscosities 10–100 times that of the ambient mantle. An increase in relative density, as well as larger volumes of CAID material, lead to the development of a core enveloping layer. Our findings show that when two provinces are present core heat flux can be reduced by up to 50 per cent relative to cases in which CAID material is absent. Surface deformation quantified by Plateness is minimally influenced by variation of the properties of the dense material. Surface velocity is found to be reduced in general but mostly substantially in cases featuring high CAID material viscosities and large volumes (i.e. 10 per cent) or buoyancy ratios.

Controls on Global Mantle Convective Structures and Their Comparison With Seismic Models

AbstractSeismic observations suggest (1) significant accumulation of subducted slabs above the 670‐km discontinuity in many subduction zones, (2) possible structure change at ~1,000‐km depth, and (3) the large low shear wave velocity provinces above the core‐mantle boundary in the African and Pacific lower mantle be associated with chemical heterogeneity. Global mantle convection models with realistic plate motion history reproduce most of these structures. However, it remains unclear how the convection models compare with seismic models at different spatial wavelengths and depths. By conducting quantitative analysis between mantle convection and seismic models, we found that mantle convective structures show significant correlations with seismic structures in the upper mantle and mantle transition zone for wavelengths up to spherical harmonic degree 20. However, the global correlation is weak at intermediate to short wavelengths (for degrees 4 and higher) in the lower mantle below ~1,000‐km depth. A weak layer beneath the spinel‐to‐postspinel phase change help consistently reproduce stagnant slabs in the western Pacific, while having insignificant effects elsewhere, that is, the large low shear wave velocity province structures. The cold slab structures and their correlations with the seismically fast anomalies are nearly identical for our convection models with and without the plumes, indicating that seismically fast anomalies in the mantle mainly result from the subducted slabs. Models with viscosity increase at 1,000‐km depth and the 670‐km depth phase change may reproduce seismic slab structures including the stagnant slabs in the mantle transition zone equally well as models with a thin weak layer below the 670‐km phase boundary.

Pacific‐Panthalassic Reconstructions: Overview, Errata and the Way Forward

AbstractWe have devised a new absolute Late Jurassic‐Cretaceous Pacific plate model using a fixed hot spot approach coupled with paleomagnetic data from Pacific large igneous provinces (LIPs) while simultaneously minimizing plate velocity and net lithosphere rotation (NR). This study was motivated because published Pacific plate models for the 83.5‐ to 150‐Ma time interval are variably flawed, and their use affects modeling of the entire Pacific‐Panthalassic Ocean and interpretation of its margin evolution. These flaws could be corrected, but the revised models would imply unrealistically high plate velocities and NR. We have developed three new Pacific realm models with varying degrees of complexity, but we focus on the one that we consider most realistic. This model reproduces many of the Pacific volcanic paths, modeled paleomagnetic latitudes fit well with direct observations, plate velocities and NR resulting from the model are low, and all reconstructed Pacific LIPs align along the surface‐projected margin of the Pacific large low shear wave velocity province. The emplacement of the Shatsky Rise LIP at ~144 Ma probably caused a major plate boundary reorganization as indicated by a major jump of the Pacific‐Izanagi‐Farallon triple junction and a noteworthy change of the Pacific‐Izanagi seafloor spreading direction at around chron M20 time.

Radial thermo-chemical structure beneath Western and Northern Pacific from seismic waveform inversion

The Earth's deep mantle seismic structure is dominated by two large low shear-wave velocity provinces (LLSVPs) located beneath Africa and the Pacific. While the existence of these structures has been attested by many studies and data sets, their detailed nature, purely thermal or thermo-chemical, is still a matter of debate. Discriminating between these hypotheses requires constraints independent from seismic velocity structure. Seismic shear-wave attenuation, measured by the quality factor QS, strongly depends on temperature but not (or weakly) on composition. It may bring key information on temperature, resolving in turn the trade-off between temperature and composition. Here, we invert seismic waveform data for radial models of shear-wave velocity anomalies (dlnVS), and QS at two different locations in the Pacific, and from a depth of 2000 km down to the core-mantle boundary (CMB). We show that Western Pacific (WP) models, sampling the western tip of the Pacific LLSVP and the Caroline plume, cannot be explained by thermal anomalies alone, but require excess in iron of ∼4.0% from the CMB up to 2600 km. By contrast, Northern Pacific models (NP), if unaffected by seismic anisotropy, may have a purely thermal origin. Based on these observations, we build radial thermo-chemical models at WP and NP.

Implementation of a Volume-of-Fluid method in a finite element code with applications to thermochemical convection in a density stratified fluid in the Earth’s mantle

We describe the implementation of a second-order accurate Volume-of-Fluid interface tracking algorithm in the open source finite element code ASPECT that is designed to model convection and other processes in the Earth’s mantle. This involves the solution of the incompressible Stokes equations coupled to an advection diffusion equation for the temperature, a Boussinesq approximation that governs the dependence of the density on the temperature, and an advection equation for a marker indicating two initial (constant) density states, that are passively advected in the underlying flow field. The Volume-of-Fluid method in ASPECT is fully parallelized and fully integrated with ASPECT’s adaptive mesh refinement algorithm. We present the results of several interface tracking benchmarks in order to demonstrate the accuracy of the method, as well as the results of several benchmarks commonly used in the computational mantle convection community. Finally, we present the results of computations with and without adaptive mesh refinement of a model problem involving thermochemical convection in a computationally stratified fluid designed to provide insight into how thermal plumes, that eventually reach the Earth’s surface as ocean island basalts, originate at structures near the core-mantle boundary known as Large Low Shear wave Velocity Provinces or “LLSVPs”. LLSVPs are structures in parts of the lowermost portion of the Earth’s mantle, characterized by slow shear wave velocities and higher density than the surrounding mantle, which were discovered by seismic tomography of the deep Earth.

Lowermost mantle thermal conductivity constrained from experimental data and tomographic models

SUMMARY Heat transfer through Earth's mantle is sensitive to mantle thermal conductivity and its variations. Thermal conductivities of lower mantle minerals, bridgmanite (Bm) and ferropericlase (Fp), depend on pressure, temperature, and composition. Because temperature and composition are expected to strongly vary in the deep mantle, thermal conductivity may also vary laterally. Here, we compile self-consistent data on lattice thermal conductivities of Bm and Fp at high pressure to model lower mantle thermal conductivity and map its possible lateral variations. Importantly, our data set allows us to quantify the influence of iron content on mantle conductivity. At the bottom of the mantle, the thermal conductivity for a pyrolitic mantle calculated along an adiabat with potential temperature 2000 K is equal 8.6 W m–1 K–1. Using 3-D thermochemical models from probabilistic tomography, which include variations in temperature, iron content, and bridgmanite fraction, we then calculate possible maps of conductivity anomalies at the bottom of the mantle. In regions known as low shear-wave velocity provinces, thermal conductivity is reduced by up to 26 per cent compared to average mantle, which may impact mantle dynamics in these regions. A simple analysis of threshold and saturation effects related to the iron content shows that our estimates of thermal conductivity may be considered as upper bounds. Quantifying these effects more precisely however requires additional mineral physics measurements. Finally, we estimate variations in core–mantle boundary heat flux, and find that that these variations are dominated by lateral temperature anomalies and are only partly affected by changes in thermal conductivity.

Geodynamic origin of carbonatites from the absolute paleotectonic reconstructions

Geodynamic origin of carbonatites is debated for several decades. One of hypotheses links their origin to large-volume mantle plumes rising from the core-mantle boundary (CMB). Some evidence exists for temporal and spatial relationships between the occurrences of carbonatites and large igneous provinces (LIPs), and both carbonatites and LIPs can be related to mantle plumes. A good example is the carbonatites of the Maymecha-Kotuy Province in the Polar Siberia, which were formed at the same time as the Siberian superplume event at ca. 250 Ma. In this study we use a recently published absolute plate kinematic modelling to reconstruct the position of 155 Phanerozoic carbonatites at the time of their emplacement. We demonstrate that 69% of carbonatites may be projected onto the central or peripheral parts of the large low shear-wave velocity provinces (LLSVPs) in the lowermost mantle. This correlation provides a strong evidence for the link between the carbonatite genesis and the locations of deep-mantle plumes. A large group of carbonatites (31%) has no obvious links to LLSVPs and, on the contrary, they plot above the “faster-than-average S-wave” zones in the deep mantle, currently located beneath North and Central America and China. We propose that their origin may be associated with remnants of subducted slabs in the mantle.

Tungsten isotopes in mantle plumes: Heads it's positive, tails it's negative

The lowermost mantle is driven to Earth's surface by mantle plumes, providing a volcanic record of its structure and composition. Plumes comprise a head and tail, which melt to form large igneous provinces (LIPs) and ocean island basalts (OIBs), respectively. Recent analyses have shown that LIPs and OIBs exhibit tungsten (W) isotope heterogeneity that was created in the first ∼60 million years of our solar system's evolution. Moreover, the isotopic signature found in LIPs differs to that found in OIBs, revealing that the melt products of plume heads must be dominated by a different ancient mantle reservoir to that of plume tails. However, existing geodynamical studies suggest that plume heads and tails sample the same deep-mantle source region and, therefore, cannot account for any systematic differences in composition. Here, we present a suite of numerical simulations of thermo-chemical plumes and an isotopic model for W sources in the mantle. Our results demonstrate that the W isotope systematics of LIPs and OIBs can, under certain conditions, arise as a dynamical consequence of plumes forming in a heterogeneous, thermo-chemical boundary layer. We find that ultra low-velocity zones (ULVZs), which sit on the core–mantle boundary (CMB), likely contribute to the chemical diversity observed in OIBs but not LIPs, while any dense components residing inside large low shear-wave velocity provinces (LLSVPs) may contribute to both. This study places geochemical observations from Earth's surface in a geodynamically consistent framework and illuminates their relationship with seismically imaged features of the deep mantle.

Geochemistry and Distribution of Recycled Domains in the Mantle Inferred From Nd and Pb Isotopes in Oceanic Hot Spots: Implications for Storage in the Large Low Shear Wave Velocity Provinces

AbstractSubduction of continental and oceanic crust is thought to give rise to geochemically distinct reservoirs in the mantle called EM (enriched mantle) and HIMU (high μ = 238U/204Pb), respectively. However, the locations of EM and HIMU domains in the Earth's interior are poorly constrained. We explore the geographic distribution of extreme EM (143Nd/144Nd ≤ 0.512630) and HIMU (206Pb/204Pb ≥ 20) geochemical signatures in ocean island basalts erupted at hot spots, highlighting three observations. First, hot spots geographically associated with the two large low shear wave velocity provinces (LLSVPs) have a similar range of EM compositions. If these hot spots are sourced by the LLSVPs via upwelling plumes, this observation is consistent with the hypothesis that the LLSVPs formed by similar processes and have similar geodynamic histories. Second, the EM and HIMU domains exhibit different latitudinal zonation: oceanic hot spots with the most extreme EM compositions (143Nd/144Nd < 0.5125) are concentrated in the Southern Hemisphere (14°S to 52°S latitude), while oceanic hot spots hosting HIMU compositions are found primarily near the tropical latitudes (38°N to 29°S). Third, all 13 oceanic hot spots with EM compositions (143Nd/144Nd ≤ 0.512630) are geographically associated with the LLSVPs; oceanic hot spots located far from the LLSVPs exhibit only non‐EM (143Nd/144Nd > 0.512630) compositions. In contrast, the HIMU domains do not show a clear geographic association with the LLSVPs. Therefore, EM and HIMU domains in the Earth's mantle exhibit different spatial distributions. This may reflect differences in subduction inputs of these two components, or differences in how they segregate or accumulate in the deep Earth.

Deformation, crystal preferred orientations, and seismic anisotropy in the Earth's D″ layer

We use a forward multiscale model that couples atomistic modeling of intracrystalline plasticity mechanisms (dislocation glide ± twinning) in MgSiO3 post-perovskite (PPv) and periclase (MgO) at lower mantle pressures and temperatures to polycrystal plasticity simulations to predict crystal preferred orientations (CPO) development and seismic anisotropy in D″. We model the CPO evolution in aggregates of 70% PPv and 30% MgO submitted to simple shear, axial shortening, and along corner-flow streamlines, which simulate changes in flow orientation similar to those expected at the transition between a downwelling and flow parallel to the core–mantle boundary (CMB) within D″ or between CMB-parallel flow and upwelling at the borders of the large low shear wave velocity provinces (LLSVP) in the lowermost mantle. Axial shortening results in alignment of PPv [010] axes with the shortening direction. Simple shear produces PPv CPO with a monoclinic symmetry that rapidly rotates towards parallelism between the dominant [100](010) slip system and the macroscopic shear. These predictions differ from MgSiO3 post-perovskite textures formed in diamond-anvil cell experiments, but agree with those obtained in simple shear and compression experiments using CaIrO3 post-perovskite. Development of CPO in PPv and MgO results in seismic anisotropy in D″. For shear parallel to the CMB, at low strain, the inclination of ScS, Sdiff, and SKKS fast polarizations and delay times vary depending on the propagation direction. At moderate and high shear strains, all S-waves are polarized nearly horizontally. Downwelling flow produces Sdiff, ScS, and SKKS fast polarization directions and birefringence that vary gradually as a function of the back-azimuth from nearly parallel to inclined by up to 70° to CMB and from null to ∼5%. Change in the flow to shear parallel to the CMB results in dispersion of the CPO, weakening of the anisotropy, and strong azimuthal variation of the S-wave splitting up to 250 km from the corner. Transition from horizontal shear to upwelling also produces weakening of the CPO and complex seismic anisotropy patterns, with dominantly inclined fast ScS and SKKS polarizations, over most of the upwelling path. Models that take into account twinning in PPv explain most observations of seismic anisotropy in D″, but heterogeneity of the flow at scales <1000 km is needed to comply with the seismological evidence for low apparent birefringence in D″.

Effects of iron on the lattice thermal conductivity of Earth’s deep mantle and implications for mantle dynamics

Iron may critically influence the physical properties and thermochemical structures of Earth's lower mantle. Its effects on thermal conductivity, with possible consequences on heat transfer and mantle dynamics, however, remain largely unknown. We measured the lattice thermal conductivity of lower-mantle ferropericlase to 120 GPa using the ultrafast optical pump-probe technique in a diamond anvil cell. The thermal conductivity of ferropericlase with 56% iron significantly drops by a factor of 1.8 across the spin transition around 53 GPa, while that with 8-10% iron increases monotonically with pressure, causing an enhanced iron substitution effect in the low-spin state. Combined with bridgmanite data, modeling of our results provides a self-consistent radial profile of lower-mantle thermal conductivity, which is dominated by pressure, temperature, and iron effects, and shows a twofold increase from top to bottom of the lower mantle. Such increase in thermal conductivity may delay the cooling of the core, while its decrease with iron content may enhance the dynamics of large low shear-wave velocity provinces. Our findings further show that, if hot and strongly enriched in iron, the seismic ultralow velocity zones have exceptionally low conductivity, thus delaying their cooling.

Modelling Earth’s surface topography: Decomposition of the static and dynamic components

Contrasting results on the magnitude of the dynamic component of topography motivate us to analyse the sources of uncertainties affecting long wavelength topography modelling. We obtain a range of mantle density structures from thermo-chemical interpretation of available seismic tomography models. We account for pressure, temperature and compositional effects as inferred by mineral physics to relate seismic velocity with density. Mantle density models are coupled to crustal density distributions obtained with a similar methodology. We compute isostatic topography and associated residual topography maps and perform instantaneous mantle flow modelling to calculate the dynamic topography. We explore the effects of proposed mantle 1-D viscosities and also test a 3D pressure- and temperature-dependent viscosity model. We find that the patterns of residual and dynamic topography are robust, with an average correlation coefficient (r) of respectively ∼0.74 and ∼0.71, upper-lower quartile ranges of 0.86–0.65 for residual topography and 0.83–0.62 for dynamic topography maps. The amplitudes are, on the contrary, poorly constrained. For the static component, the inferred density models of lithospheric mantle give an interquartile range of isostatic topography that is always higher than 100m, reaching 1.7km in some locations, and averaging ∼720m. Crustal density models satisfying the same compressional velocity structure lead to variations in isostatic topography averaging 350m, with peaks of 1km in thick crustal regions, and always higher than 100m. The uncertainties on isostatic topography are strong enough to mask, if present, the contribution of mantle convection to surface topography. For the dynamic component, we obtain a peak-to-peak dynamic topography amplitude exceeding 3km for all our mantle density and viscosity models. These extremely high values would be associated with a magnitude of geoid undulations that is not in agreement with observations. Considering chemical heterogeneities in correspondence with the lower mantle Large Low Shear wave Velocity Provinces (LLSVPs) helps to decrease the peak-to-peak amplitudes of dynamic topography and geoid, but significantly reduces the correlation between synthetic and observed geoid. The correlation coefficients between all our residual and dynamic topography maps (a total of 220 and 198, respectively) is <0.55 (average=∼0.19). The correlation slightly improves when considering only the very long-wavelength components of the maps (average=∼0.23). We therefore conclude that a robust determination of dynamic topography is not feasible since current uncertainties affecting crustal density, mantle density and mantle viscosity are still too large. A truly interdisciplinary approach, combining constraints from the geological record with a multi-methodological interpretation of geophysical observations, is required to tackle the challenging task of linking the surface topography to deep processes.