Small-scale Convection Research Articles

3-D shear wave radially and azimuthally anisotropic velocity model of the North American upper mantle

Using a combination of long period seismic waveforms and SKS splitting measurements, we have developed a 3-D upper-mantle model (SAWum_NA2) of North America that includes isotropic shear velocity, with a lateral resolution of ∼250 km, as well as radial and azimuthal anisotropy, with a lateral resolution of ∼500 km. Combining these results, we infer several key features of lithosphere and asthenosphere structure. A rapid change from thin (∼70–80 km) lithosphere in the western United States (WUS) to thick lithosphere (∼200 km) in the central, cratonic part of the continent closely follows the Rocky Mountain Front (RMF). Changes with depth of the fast axis direction of azimuthal anisotropy reveal the presence of two layers in the cratonic lithosphere, corresponding to the fast-to-slow discontinuity found in receiver functions. Below the lithosphere, azimuthal anisotropy manifests a maximum, stronger in the WUS than under the craton, and the fast axis of anisotropy aligns with the absolute plate motion, as described in the hotspot reference frame (HS3-NUVEL 1A). In the WUS, this zone is confined between 70 and 150 km, decreasing in strength with depth from the top, from the RMF to the San Andreas Fault system and the Juan de Fuca/Gorda ridges. This result suggests that shear associated with lithosphere–asthenosphere coupling dominates mantle deformation down to this depth in the western part of the continent. The depth extent of the zone of increased azimuthal anisotropy below the cratonic lithosphere is not well resolved in our study, although it is peaked around 270 km, a robust result. Radial anisotropy is such that, predominantly, ξ > 1, where ξ= (Vsh/Vsv)2, under the continent and its borders down to ∼200 km, with stronger ξ in the bordering oceanic regions. Across the continent and below 200 km, alternating zones of weaker and stronger radial anisotropy, with predominantly ξ < 1, correlate with zones of small lateral changes in the fast axis direction of anisotropy, and faster than average Vs below the LAB, suggesting the presence of small scale convection with a wavelength of ∼2000 km. Finally, in the western United States, complex 3-D patterns of isotropic velocity and anisotropy reflect mantle dynamics associated with the rich tectonic history of the region.

Geochemistry of 24 Ma basalts from NE Egypt: source components and fractionation history

Abstract Subalkaline basalts from NE Egypt represent an episode of magmatism at c . 24 Ma, coincident with widespread eruptive activity in northern Africa. New geochemical data provide insight into the mineralogical and isotopic characteristics of the underlying mantle. The basalts show little geochemical variation, with incompatible trace element abundances similar to those of ocean island basalts. They display fairly smooth primitive mantle-normalized incompatible trace element patterns. Trace element abundances and Sr–Nd–Pb–Hf isotopic signatures are consistent with contributions from two distinct source regions, one similar to the Afar plume and the other located within the metasomatized spinel-facies subcontinental lithosphere. Mixing of melts from these two domains was followed by minor crustal contamination during prolonged ascent or emplacement. Integrating the geochemical data with available tomographic information allows us to develop a framework for understanding mid-Tertiary magmatic activity throughout northern Africa. A model for this widespread volcanism involves ascent of upwelling mantle derived from the margins of the South African Superplume rooted at the core–mantle boundary and/or through small-scale convection at the 660 km discontinuity. Ascent of magmas to the surface was facilitated by pre-existing structures within the lithosphere, including those associated with incipient rifting of the Red Sea. Supplementary material: Mineral chemistry data are available at http://www.geolsoc.org.uk/SUP18483 .

Charles Henry Brian Priestley. 8 July 1915 — 18 May 1998

Charles Henry Brian Priestley was born and educated in England. After completing the Mathematical Tripos at the University of Cambridge, he joined the Meteorological Office in 1939. For the next seven years he was engaged mostly in wartime work, including a two-year spell in Canada (1941–43) and three years with the Meteorological Office upper-air unit at Dunstable, UK (1943–46). In 1946, aged 31 years, he took up an Australian appointment with the Council for Scientific and Industrial Research (later to become the Commonwealth Scientific and Industrial Research Organization (CSIRO)) to establish and develop a group to undertake research in meteorological physics. Thereafter he was based in Melbourne, Australia, with his career in the CSIRO extending to 1977. Priestley’s own early research focused on large-scale atmospheric systems, including substantial work on global-scale transport, and later on small-scale atmospheric convection and heat transfer, in which he established some significant results. He had a leading role in the development of the atmospheric sciences in Australia, and was strongly involved in international meteorology.

Charles Henry Brian Priestley 1915 - 1998

Charles Henry Brian Priestley (known as Bill) was born and educated in England. After completing the Mathematical Tripos at the University of Cambridge, he joined the Meteorological Office in 1939. In 1946, aged 31 years, he took up an Australian appointment with the Council for Scientific and Industrial Research (CSIR, later to become CSIRO) to establish and develop a group to undertake research in meteorological physics. Thereafter he was based in Melbourne, Australia. The group earned world recognition, particularly for its investigations of turbulent transfer in the lower atmosphere, and evolved to become the CSIRO Division of Atmospheric Research. Priestley's own early research focused on large-scale atmospheric systems, including substantial work on global-scale transport, and later on small-scale atmospheric convection and heat transfer, in which he established some significant results. He had a leading role in the development of the atmospheric sciences in Australia, and was strongly involved in international meteorology. His career with CSIRO extended to 1977, and he finally retired from all professional commitments in the mid-1980s. After several years of declining health, he died on 18 May 1998, seven weeks before he turned 83.

Towards a better understanding of African topography: a review of passive-source seismic studies of the African crust and upper mantle

Abstract Explaining the cause and support of Africa's varied topography remains a fundamental question for our understanding of the long-term evolution of the continent. As geodynamical modelling becomes more frequently used to investigate this problem, it is important to understand the seismological results that can be incorporated into these models. Crustal thickness estimates are crucial for calculating components of topography that are isostatically compensated. Variations in seismic velocity help constrain variations in subsurface temperature and density and thus buoyancy; measurements of anisotropy can also be used to determine the contribution of the mantle flow field to dynamic topography. In this light, we review the results of passive seismic studies across Africa. At the continental scale there are significant differences in crustal models, meaning large uncertainties in corrections for isostatic topography. In east Africa, multiple seismic experiments have provided firm constraints on crustal and mantle structure. Tomographic images illuminate a broad ( c. 500 km wide) low-velocity region in the upper mantle, with possible connection to the African Superplume in the lower mantle. These observations, alongside the variations in radial anisotropy, strongly suggest that the mantle flow field contributes significantly to the uplift of the region. Beneath southern Africa, low velocities are observed near the base of the continental lithosphere; the depth to transition zone discontinuities however suggests that they are not linked to the superplume beneath. It is thus less clear what role the sublithospheric mantle plays in supporting the region's high topography. Many of Africa's secondary topographic features (e.g. Atlas, Hoggar, Bie Dome) are underlain by slow velocities at depths of 100–150 km and are adjacent to rapid changes in lithospheric thickness. Whether these variations in lithospheric structure promote small-scale convection or simply guide the larger-scale mantle flow field remains ambiguous.

Young volcanism in the Borborema Province, NE Brazil, shows no evidence for a trace of the Fernando de Noronha plume on the continent

We present 40Ar/39Ar ages for four volcanic bodies from a group of volumetrically minor alkaline plugs, necks, and dikes in northeastern Brazil, previously ascribed to passage over the purported Fernando de Noronha plume. The rocks are relatively primitive (9.5–14.7wt.% MgO), typically nepheline-normative basanites with ocean island basalt (OIB)-like trace-element compositions similar to alkalic basalts from the Fernando de Noronha archipelago. Highly fractionated REE coupled with relative depletions of K and Rb indicates that the silica undersaturated magmas were generated by small degrees of melting in the presence of residual garnet and a hydrous metasomatic phase. Three of the four units (Caracarazinho, Cabugizinho da Arara and Serra Preta de Bodó) were heretofore undated. The fourth body (Cabelo de Negro) was included to facilitate comparison with published K–Ar dates. 40Ar/39Ar age determinations by the laser incremental-heating method on duplicate grains of groundmass reveal the youngest continental volcanism in Brazil, with emplacement ages between 8.9±0.5 and 7.1±0.3Ma. Our age for Cabelo de Negro (7.9±0.3Ma) is roughly 20Ma younger than the published K–Ar date for this plug. The reproducibility of our duplicate analyses and the consistency of the plateau, ideogram and isochron ages for this sample attest to the reliability of the new 40Ar/39Ar results. Our geochronological results show that volcanic activity on the continent did not shut down prior to the onset of volcanism on the island of Fernando de Noronha. Both areas were active contemporaneously for at least 5Ma. We argue that the extended duration, small volume and lack of a clear age progression suggest that this example of alkaline intraplate volcanism is more likely the surface manifestation of the upwelling flow seen in an edge-driven convection mode, rather than tracking passage over a deep-seated mantle plume. This hypothesis is supported by xenolith thermobarometery, heat-flow data and seismic tomography, which collectively provide evidence for a mild thermal anomaly (upwelling) in the upper mantle under the eastern-most equatorial Brazil that extends into the Atlantic and appears to travel with the plate. Small-scale convection is also consistent with the seismic evidence for a downwelling to about 600km depth in the mantle beneath the cratonic margin of eastern Brazil. Given the inherently sluggish upwelling associated with such small-scale convection, it appears that the entrainment of metasomatically enriched lithospheric mantle is likely vital to melt generation. If so, the spatial distribution of edge-driven magmatism may provide a tool for mapping fertility of the subcontinental lithospheric mantle.

Small-scale convection signatures associated with a strong plage solar magnetic field

In this work, we study and quantify properties of strong-field small-scale convection and compare observed properties with those predicted by numerical simulations. We analyze spectropolarimetric 630.25 nm data from a unipolar ephemeral region near sun center. We use line-of-sight velocities and magnetic field measurements obtained with Milne-Eddington inversion techniques along with measured continuum intensities and Stokes V amplitude asymmetry at a spatial resolution of 0.15 arcseconds to establish statistical relations between the measured quantities. We also study these properties for different types of distinct magnetic features, such as micropores, bright points, ribbons, flowers and strings. We present the first direct observations of a small-scale granular magneto-convection pattern within extended regions of strong (more than 600 G average) magnetic field. Along the boundaries of the flux concentrations we see mostly downflows and asymmetric Stokes V profiles, consistent with synthetic line profiles calculated from MHD simulations. We note the frequent occurrence of bright downflows along these boundaries. In the interior of the flux concentrations, we observe an up/down flow pattern that we identify as small-scale magnetoconvection, appearing similar to that of field-free granulation but with scales 4 times smaller. Measured RMS velocities are 70% of those of nearby field-free granulation, even though the average radiative flux is not reduced. The interiors of these flux concentrations are dominated by upflows.

Subduction-triggered magmatic pulses: A new class of plumes?

A variety of atypical plume-like structures and focused upwellings that are not rooted in the lower mantle have recently been discussed, and seismological imaging has shown ubiquitous small-scale convection in the uppermost mantle in regions such as the Mediterranean region, the western US, and around the western Pacific. We argue that the three-dimensional return flow and slab fragmentation associated with complex oceanic subduction trajectories within the upper mantle can generate focused upwellings and that these may play a significant role in regional tectonics. The testable surface expressions of this process are the outside-arc alkaline volcanism, topographic swell, and low-velocity seismic anomalies associated with partial melt. Using three-dimensional, simplified numerical subduction models, we show that focused upwellings can be generated both ahead of the slab in the back-arc region (though ~ five times further inward from the trench than arc-volcanism) and around the lateral edges of the slab (in the order of 100 km away from slab edges). Vertical mass transport, and by inference the associated decompression melting, in these regions appears strongly correlated with the interplay between relative trench motion and subduction velocities. The upward flux of material from the depths is expected to be most pronounced during the first phase of slab descent into the upper mantle or during slab fragmentation. We discuss representative case histories from the Pacific and the Mediterranean where we find possible evidence for such slab-related volcanism.

Formation of the orogen-foredeep system: A geodynamic model and comparison with the data of the northern Forecaucasus

The disturbance of mechanical and thermal equilibria in the upper shell of the Earth as a result of mantle or local within-plate processes related to periodic tectonic activity gives rise to the formation of convective flows in the low-viscosity asthenosphere. These flows affect the lithosphere and create domains of subsidence and uplift, which can continue to develop long after the cessation of active periods. If the density of the lithosphere does not decrease with depth, then small-scale flows increase uplift in zones of compression of the continental lithosphere and create domains of extension at their margins. In our opinion, small-scale convection is the main geodynamic factor that forms foredeeps. The results of detailed numerical modeling of foredeep formation at the margins of adjoining orogens are presented in the current paper. In order to set the initial conditions for the stage of continental collision, the precollision stages of the foldbelt evolution are considered, including the stage of trough formation on the thinned continental crust or on the oceanic lithosphere and the stage of sedimentary basin formation; depending on the degree of extension, this can be an inner sea or a passive continental margin. Such initial conditions were used in modeling of the compression stage (continental collision), when the orogen-foredeep system is formed. The parameters of the model and the tectonic processes are chosen so as to bring the results of numerical computation in line with the data on the Greater Caucasus and northern Forecaucasus, including the thickness of the crustal layers and sedimentary cover, structure of the foredeeps, rate of tectonic subsidence, heat flow, etc. Comparison of the numerical modeling results with the formation history of the Caucasus foredeeps confirms that the first stage of regional compression of the Greater Caucasus took place before the deposition of the Maikop sediments. At least three compression stages followed: 16.6–15.8 Ma (Tarchanian), 14.3–12.3 Ma (Konkian-early Sarmatian), and 7.0–5.2 Ma (Pontian). The next stage of regional compression is apparently occurring at present.

Small-Scale Mantle Convection Produces Stratigraphic Sequences in Sedimentary Basins

Cyclic sedimentary deposits link stratigraphic sequences that are now geographically distant but were once part of the same depositional environment. Some of these sequences occur at periods of 2 to 20 million years, and eustatic sea-level variations or regional tectonic events are likely causes of their formation. Using numerical modeling, we demonstrate that small-scale mantle convection can also cause the development of stratigraphic sequences through recurrent local and regional vertical surface movements. Small-scale convection-driven stratigraphic sequences occur at periods of 2 to 20 million years and correlate only at distances up to a few hundred kilometers. These results suggest that previous sequence stratigraphic analyses may contain erroneous conclusions regarding eustatic sea-level variations.

Complex subduction and small-scale convection revealed by body-wave tomography of the western United States upper mantle

New high-resolution P- and S-wave tomography of the United States upper mantle from the Pacific Coast to the Great Plains reveals strong multi-scale heterogeneity closely correlated with tectonic and magmatic activity. We invert teleseismic travel-time residuals from the EarthScope Transportable Array and more than 1700 additional temporary and permanent stations for 3-D velocity perturbations to a depth of 1000 km. The inversion uses recent advances in western U.S. crust models to better isolate the mantle component of travel-time residuals, and frequency-dependent 3-D sensitivity kernels to map travel-time residuals, measured in multiple frequency bands, into velocity structure. In addition to separate V P and V S models, we jointly invert the two datasets for V P/V S perturbations by imposing a smoothness constraint on the δ lnV S/δ lnV P field. The joint inversion helps us identify regions where partial melt is probable. The amplitude of V P, V S, and V P/V S variations is greatest in the upper 200 km of the mantle and the form of velocity anomalies suggests a provincially heterogeneous lithosphere and the occurrence of widespread small-scale convection. Partially molten mantle is inferred beneath Yellowstone and the eastern Snake River Plain (SRP), the Salton Trough, and the Clear Lake volcanic field. The inferred depth extent of partial melt is consistent with a generally hydrated upper mantle and elevated temperatures beneath the eastern SRP and Yellowstone. Despite continuous subduction since the Cretaceous, the distribution of sub-lithospheric high-velocity anomalies is dissected (similar to other recent studies). Based on our new tomography models, western U.S. geologic history, and plate–tectonic reconstructions, we infer patchy and incomplete removal of the flat-subducting Laramide slab and slab tearing associated with Eocene accretion in the northwestern U.S.

Small-scale convection at the edge of the Colorado Plateau: Implications for topography, magmatism, and evolution of Proterozoic lithosphere

The Colorado Plateau of the southwestern United States is characterized by a bowl-shaped high elevation, late Neogene–Quaternary magmatism at its edge, large gradients in seismic wave velocity across its margins, and relatively low lithospheric seismic wave velocities. We explain these observations by edge-driven convection following rehydration of Colorado Plateau lithosphere. A rapidly emplaced Cenozoic step in lithosphere thickness between the Colorado Plateau and adjacent extended Rio Grande rift and Basin and Range province causes small-scale convection in the asthenosphere. A lithospheric drip below the plateau is removing lithosphere material from the edge that is heated and metasomatized, resulting in magmatism. Edgedriven convection also drives margin uplift, giving the plateau its characteristic bowl shape. The edge-driven convection model shows good consistency with features resolved by seismic tomography.

Excitation of small-scale convection in the continental lithosphere by disturbances in the topography of the Earth’s surface

After the buildup of a steady convective motion in a layer, the topography of its surface is determined only by the displacements of the material particles lying on its surface.

Shaping mobile belts by small-scale convection

Mobile belts are long-lived deformation zones composed of an ensemble of crustal fragments, distributed over hundreds of kilometres inside continental convergent margins. The Mediterranean represents a remarkable example of this tectonic setting: the region hosts a diffuse boundary between the Nubia and Eurasia plates comprised of a mosaic of microplates that move and deform independently from the overall plate convergence. Surface expressions of Mediterranean tectonics include deep, subsiding backarc basins, intraplate plateaux and uplifting orogenic belts. Although the kinematics of the area are now fairly well defined, the dynamical origins of many of these active features are controversial and usually attributed to crustal and lithospheric interactions. However, the effects of mantle convection, well established for continental interiors, should be particularly relevant in a mobile belt, and modelling may constrain important parameters such as slab coherence and lithospheric strength. Here we compute global mantle flow on the basis of recent, high-resolution seismic tomography to investigate the role of buoyancy-driven and plate-motion-induced mantle circulation for the Mediterranean. We show that mantle flow provides an explanation for much of the observed dynamic topography and microplate motion in the region. More generally, vigorous small-scale convection in the uppermost mantle may also underpin other complex mobile belts such as the North American Cordillera or the Himalayan-Tibetan collision zone.

Driving Earth's surface motions

Density variations within Earth's mantle may be a significant driver of both horizontal and vertical surface movements. The fingerprints of such mantle processes have been found in the Mediterranean region. Claudio Faccenna and Thorsten Becker compute global mantle flow based on a recent, high-resolution seismic tomography model, to investigate the role of buoyancy-driven and plate-motion-induced mantle circulation for the Mediterranean. They show that mantle flow provides an explanation for much of the observed dynamic topography and microplate motion in the region and conclude that vigorous small-scale convection in the uppermost mantle may also hold the key for understanding other complex mobile belts, such as those in the North American Cordillera and the Himalayan–Tibetan collision zone.

The consequences of Canadian Cordillera thermal regime in recent tectonics and elevation: a reviewThis article is one of a series of papers published in this Special Issue on the theme Lithoprobe — parameters, processes, and the evolution of a continent.Geological Survey of Canada Contribution 20090195.

The crust and upper mantle thermal regime of the Canadian Cordillera and its tectonic consequences were an important part of the Cordillera Lithoprobe program and related studies. This article provides a review, first of the thermal constraints, and then of consequences in high surface elevation and current tectonics. Cordillera and adjacent craton temperatures are well constrained by geothermal heat flow, mantle tomography velocities, upper mantle xenoliths, and the effective elastic thickness, Te. Cordillera temperatures are very high and laterally uniform, explained by small scale convection beneath a thin lithosphere, 800–900 °C at the Moho, contrasted to 400–500 °C for the craton. The high temperatures provide an explanation for why the Cordillera has high elevation in spite of a generally thin crust, ∼33 km, in contrast to low elevation and thicker crust, 40–45 km, for the craton. The Cordillera is supported ∼1600 m by lithosphere thermal expansion. In the Cordillera only the upper crust has significant strength; Te ∼ 15 km, in contrast to over 60 km for the craton. The Cordillera is tectonically active because the lithosphere is sufficiently weak to be deformed by plate boundary and gravitational forces; the craton is too strong. The Canadian Cordillera results have led to new understandings of processes in backarcs globally. High backarc temperatures and weak lithospheres explain the tectonic activity over long geological times of mobile belts that make up about 20% of continents. They also have led to a new understanding of collision orogenic heat in terms of incorporation of already hot backarcs.

The structure and evolution of the lithosphere–asthenosphere boundary beneath the Atlantic–Mediterranean Transition Region

The present-day thermal and compositional 3D structure of the lithosphere beneath the Atlantic–Mediterranean Transition Region and the lithosphere–asthenosphere interaction from Jurassic times to present has been studied. The Atlantic–Mediterranean Transition Region comprises the western segment of the Africa–Eurasia plate boundary, encompassing two main large-scale tectonic domains: the Gibraltar Arc System and the Atlas Mountains. An integrated and self-consistent geophysical–petrological methodology ( LitMod3D) has been applied that combines elevation, gravity, geoid, surface heat flow, and seismic data and allows modelling of compositional heterogeneities within the lithospheric mantle. Our results reveal large variations in the depth of the Moho and the lithosphere–asthenosphere boundary (LAB) as well as a lack of spatial correlation between the thicknesses of these two boundaries. The Moho essentially mimics the topography with depths ranging from ∼ 10 km beneath the oceanic domains of the Atlantic abyssal plains and the Algerian Basin to > 34 km in the Eastern Betics and Rif, the High Atlas mountains, and the Sahara Platform. In contrast, the LAB is shallower beneath the central and eastern Alboran Basin (∼ 70 km) and all along the High, Middle and Anti Atlas (< 100 km) coinciding with the loci of Cenozoic volcanism. Deeper LAB depths are found along the central and western Betics and the Moroccan Atlantic margin (> 140 km) with values exceeding 230 km beneath the Rif and the Sahara Platform. The average bulk composition of the lithospheric mantle corresponds to that of a typical Tecton (i.e. Phanerozoic) domain, with the exceptions of the Sahara Platform, the Alboran Basin, and Atlas Mountains. Distinct mantle compositions are required in these areas to make model predictions and geophysical observables compatible. It is proposed that the highly irregular LAB topography is the result of the superposition of three different geodynamic processes: i) shortening and thickening related to NW–SE Iberia–Africa convergence lasting from Late Cretaceous to Recent, ii) impingement of a baby-like mantle plume or small-scale convection beneath the High-Middle Atlas and Anti Atlas commencing in the mid Eocene, and iii) slab roll-back or mantle delamination in the Betic–Rif–Alboran realm acting from early to late Miocene.

A simple three-dimensional model of thermo–chemical convection in the mantle wedge

In order to understand the possible existence of small-scale convection in the mantle wedge, we have constructed a simple three-dimensional model of convection driven by both thermal and chemical buoyancies above the subducting slab. In this model, a chemical agent, which affects both the density and the viscosity of mantle, is introduced from the top of the subducting slab and the associated density and viscosity decreases are treated as parameters. The model does not include the along-arc variation of the source of the chemical agent. We found that the major effects of low density chemical anomaly are to suppress the three-dimensional instability and make the flow two-dimensional, i.e., the flow velocity is normal to the plate boundary. The chemically polluted region tends to stay in the corner of the mantle wedge because of its low density and this results in the low temperature zone there. This suggests the importance of chemical buoyancy on the origin of cold mantle part or “nose” in the corner of the mantle wedge. We also studied the hybrid case: The region closer to the trench is in the low density and viscosity state and the region in the back arc is in the low viscosity state only. This case shows the existence of the low temperature nose and the small-scale thermally driven convection in the back arc. We also investigated the nature of the flip-flop phenomenon of the thermally driven convection and found that the thickness of the thermal boundary layer under the back arc controls it. This flow pattern in the back arc may have a close connection with the temporal and spatial variation of volcano distribution.

Full waveform tomography for radially anisotropic structure: New insights into present and past states of the Australasian upper mantle

We combine spectral-element simulations and adjoint techniques in a non-linear full seismic waveform tomography for the radially anisotropic structure of the Australasian upper mantle. Our method allows us to explain 30 s waveforms in detail, and it yields tomographic images with locally 2 ∘ lateral resolution. In the course of 19 conjugate-gradient iterations the total number of exploitable waveforms increased from 2200 to nearly 3000. The final model, AMSAN.19, thus explains data that were not initially included in the inversion. This is strong evidence for the effectiveness of the inversion scheme and the physical consistency of the tomographic model. AMSAN.19 confirms long-wavelength heterogeneities found in previous studies, and it allows us to draw the following inferences concerning the past and present states of the Australian upper mantle and the formation of seismic anisotropy: (1) Small-scale neutral to low-velocity patches beneath central Australia are likely to be related to localised Paleozoic intraplate deformation. (2) Increasing seismic velocities between the Moho and 150 km depth are found beneath parts of Proterozoic Australia, suggesting thermochemical variations related to the formation and fragmentation of a Centralian Superbasin. (3) Radial anisotropy above 150 km depth reveals a clear ocean–continent dichotomy: We find strong v sh > v sv beneath the Coral and Tasman Seas. The anisotropy is strongest at the top of the inferred asthenospheric flow channel, where strain is expected to be largest. Radial anisotropy beneath the continent is weaker but more variable. Localised patches with v sh < v sv appear, in accord with small-scale intraplate deformation. (4) The ocean–continent dichotomy disappears gradually between 150 and 250 km depth, where the continental lithospheric mantle and the oceanic asthenosphere pass into the underlying convecting mantle. (5) Significant anisotropy exists below 250 km depth. Its character can be explained by sublithospheric small-scale convection and a change in the dominant glide system of olivine.

Small-scale sublithospheric convection reconciles geochemistry and geochronology of ‘Superplume’ volcanism in the western and south Pacific

Cretaceous volcanism in the West Pacific Seamount Province (WPSP), and Tertiary volcanism along the Cook-Australs in the South Pacific are associated with the same broad thermochemical anomaly in the asthenosphere perhaps related to the Pacific ‘Superplume.’ Abundant volcanism has usually been attributed to secondary plumelets rising from the roof of the Superplume. The Cook-Australs display distinct geochemical trends that appear to geographically project, backward in time, to corresponding trends in the WPSP. However, the implied close proximity of source regions (i.e., ∼ 1000 km) with very different geochemical fingerprints and their longevity over geological time (> 100 Myrs) appear to be at odds with the secondary plumelet hypothesis, a mechanism with a typical timescale of ∼ 30 Myrs. Moreover, ages sampled along the individual volcano chains of the Cook-Australs, and of the WPSP violate the predictions of the plumelet hypothesis in terms of linear age–distance relationships. Our numerical models indicate that small-scale sublithospheric convection (SSC) as likely triggered by the thermochemical anomaly of the ‘Superplume’ instead reconciles complex age–distance relationships, because related volcanism occurs above elongated melting anomalies parallel to plate motion (‘hot lines’). Furthermore, SSC-melting of a mantle source that consists of pyroxenite veins and enriched peridotite blobs in a matrix of depleted peridotite creates systematic geochemical trends over seafloor age during volcanism. These trends arise from variations in the amount of pyroxenite-derived lavas relative to peridotite-derived lavas along a ‘hot line,’ therefore stretching between the geochemical end-members HIMU and EMI. These predicted trends are consistent with observed trends in radiogenic isotopic composition from the Wakes, Marshalls, Gilberts (i.e., the individual volcano groups of the WPSP) and the Cook-Australs. For increasing mantle temperatures, volcanism is further predicted to occur at greater seafloor ages and with a more EMI-like signature, a result that can explain many of the observed systematics. Thus, SSC explains many of the geochemical observations with long-term temporal variations in mantle temperature, instead of persistent intermediate-scale (∼ 1000 km) compositional heterogeneity.