Laboratory models of the thermal evolution of the mantle during rollback subduction

The subduction of oceanic lithosphere plays a key role in plate tectonics, the thermal evolution of the mantle and recycling processes between Earth's interior and surface. Information on mantle flow, thermal conditions and chemical transport in subduction zones come from the geochemistry of arc volcanoes, seismic images and geodynamic models. The majority of this work considers subduction as a two-dimensional process, assuming limited variability in the direction parallel to the trench. In contrast, observationally based models increasingly appeal to three-dimensional flow associated with trench migration and the sinking of oceanic plates with a translational component of motion (rollback). Here we report results from laboratory experiments that reveal fundamental differences in three-dimensional mantle circulation and temperature structure in response to subduction with and without a rollback component. Without rollback motion, flow in the mantle wedge is sluggish, there is no mass flux around the plate and plate edges heat up faster than plate centres. In contrast, during rollback subduction flow is driven around and beneath the sinking plate, velocities increase within the mantle wedge and are focused towards the centre of the plate, and the surface of the plate heats more along the centreline.     

Trench-parallel flow and seismic anisotropy in the Mariana and Andean subduction systems

Shear-wave splitting measurements above the mantle wedge of the Mariana and southern Andean subduction zones show trench-parallel seismically fast directions close to the trench and abrupt rotations to trench-perpendicular anisotropy in the back arc. These patterns of seismic anisotropy may be caused by three-dimensional flow associated with along-strike variations in slab geometry. The Mariana and Andean subduction systems are associated with the largest along-strike variations of slab geometry observed on Earth and are ideal for testing the link between slab geometry and solid-state creep processes in the mantle. Here we show, with fully three-dimensional non-newtonian subduction zone models, that the strong curvature of the Mariana slab and the transition to shallow slab dip in the Southern Andes give rise to strong trench-parallel stretching in the warm-arc and warm-back-arc mantle and to abrupt rotations in stretching directions that are accompanied by strong trench-parallel stretching. These models show that the patterns of shear-wave splitting observed in the Mariana and southern Andean systems may be caused by significant three-dimensional flow induced by along-strike variations in slab geometry.     

Evidence of lower-mantle slab penetration phases in plate motions

It is well accepted that subduction of the cold lithosphere is a crucial component of the Earth's plate tectonic style of mantle convection. But whether and how subducting plates penetrate into the lower mantle is the subject of continuing debate, which has substantial implications for the chemical and thermal evolution of the mantle. Here we identify lower-mantle slab penetration events by comparing Cenozoic plate motions at the Earth's main subduction zones with motions predicted by fully dynamic models of the upper-mantle phase of subduction, driven solely by downgoing plate density. Whereas subduction of older, intrinsically denser, lithosphere occurs at rates consistent with the model, younger lithosphere (of ages less than about 60 Myr) often subducts up to two times faster, while trench motions are very low. We conclude that the most likely explanation is that older lithosphere, subducting under significant trench retreat, tends to lie down flat above the transition to the high-viscosity lower mantle, whereas younger lithosphere, which is less able to drive trench retreat and deforms more readily, buckles and thickens. Slab thickening enhances buoyancy (volume times density) and thereby Stokes sinking velocity, thus facilitating fast lower-mantle penetration. Such an interpretation is consistent with seismic images of the distribution of subducted material in upper and lower mantle. Thus we identify a direct expression of time-dependent flow between the upper and lower mantle.     

Seismic detection of folded, subducted lithosphere at the core-mantle boundary

Seismic tomography has been used to infer that some descending slabs of oceanic lithosphere plunge deep into the Earth's lower mantle. The fate of these slabs has remained unresolved, but it has been postulated that their ultimate destination is the lowermost few hundred kilometres of the mantle, known as the D' region. Relatively cold slab material may account for high seismic velocities imaged in D' beneath areas of long-lived plate subduction, and for reflections from a seismic velocity discontinuity just above the anomalously high wave speed regions. The D' discontinuity itself is probably the result of a phase change in relatively low-temperature magnesium silicate perovskite. Here, we present images of the D' region beneath the Cocos plate using Kirchhoff migration of horizontally polarized shear waves, and find a 100-km vertical step occurring over less than 100 km laterally in an otherwise flat D' shear velocity discontinuity. Folding and piling of a cold slab that has reached the core-mantle boundary, as observed in numerical and experimental models, can account for the step by a 100-km elevation of the post-perovskite phase boundary due to a 700 degrees C lateral temperature reduction in the folded slab. We detect localized low velocities at the edge of the slab material, which may result from upwellings caused by the slab laterally displacing a thin hot thermal boundary layer.     

Geochemical tracing of Pacific-to-Atlantic upper-mantle flow through the Drake passage

The Earth's convecting upper mantle can be viewed as comprising three main reservoirs, beneath the Pacific, Atlantic and Indian oceans. Because of the uneven global distribution and migration of ridges and subduction zones, the surface area of the Pacific reservoir is at present contracting at about 0.6 km2 x y(r-1), while the Atlantic and Indian reservoirs are growing at about 0.45 km2 x yr(-1) and 0.15 km2 x yr(-1), respectively. Garfunkel and others have argued that there must accordingly be net mantle flow from the Pacific to the Atlantic and Indian reservoirs (in order to maintain mass balance), and Alvarez further predicted that this flow should be restricted to the few parts of the Pacific rim (here termed 'gateways') where there are no continental roots or subduction zones that might act as barriers to shallow mantle flow. The main Pacific gateways are, according to Alvarez, the southeast Indian Ocean, the Caribbean Sea and the Drake passage. Here we report geochemical data which confirm that there has been some outflow of Pacific mantle into the Drake passage--but probably in response to regional tectonic constraints, rather than global mass-balance requirements. We also show that a mantle domain boundary, equivalent to the Australian-Antarctic discordance, must lie between the Drake passage and the east Scotia Sea.     

Arc-parallel flow in the mantle wedge beneath Costa Rica and Nicaragua

Resolving flow geometry in the mantle wedge is central to understanding the thermal and chemical structure of subduction zones, subducting plate dehydration, and melting that leads to arc volcanism, which can threaten large populations and alter climate through gas and particle emission. Here we show that isotope geochemistry and seismic velocity anisotropy provide strong evidence for trench-parallel flow in the mantle wedge beneath Costa Rica and Nicaragua. This finding contradicts classical models, which predict trench-normal flow owing to the overlying wedge mantle being dragged downwards by the subducting plate. The isotopic signature of central Costa Rican volcanic rocks is not consistent with its derivation from the mantle wedge or eroded fore-arc complexes but instead from seamounts of the Galapagos hotspot track on the subducting Cocos plate. This isotopic signature decreases continuously from central Costa Rica to northwestern Nicaragua. As the age of the isotopic signature beneath Costa Rica can be constrained and its transport distance is known, minimum northwestward flow rates can be estimated (63-190 mm yr(-1)) and are comparable to the magnitude of subducting Cocos plate motion (approximately 85 mm yr(-1)). Trench-parallel flow needs to be taken into account in models evaluating thermal and chemical structure and melt generation in subduction zones.     

Seismic evidence of negligible water carried below 400-km depth in subducting lithosphere

Strong evidence exists that water is carried from the surface into the upper mantle by hydrous minerals in the uppermost 10-12?km of subducting lithosphere, and more water may be added as the lithosphere bends and goes downwards. Significant amounts of that water are released as the lithosphere heats up, triggering earthquakes and fluxing arc volcanism. In addition, there is experimental evidence for high solubility of water in olivine, the most abundant mineral in the upper mantle, for even higher solubility in olivine's high-pressure polymorphs, wadsleyite and ringwoodite, and for the existence of dense hydrous magnesium silicates that potentially could carry water well into the lower mantle (deeper than 1,000?km). Here we compare experimental and seismic evidence to test whether patterns of seismicity and the stabilities of these potentially relevant hydrous phases are consistent with a wet lithosphere. We show that there is nearly a one-to-one correlation between dehydration of minerals and seismicity at depths less than about 250?km, and conclude that the dehydration of minerals is the trigger of instability that leads to seismicity. At greater depths, however, we find no correlation between occurrences of earthquakes and depths where breakdown of hydrous phases is expected. Lastly, we note that there is compelling evidence for the existence of metastable olivine (which, if present, can explain the distribution of deep-focus earthquakes) west of and within the subducting Tonga slab and also in three other subduction zones, despite metastable olivine being incompatible with even extremely small amounts of water (of the order of 100?p.p.m. by weight). We conclude that subducting slabs are essentially dry at depths below 400?km and thus do not provide a pathway for significant amounts of water to enter the mantle transition zone or the lower mantle.     

Development of anisotropic structure in the Earth's lower mantle by solid-state convection

Seismological observations reveal highly anisotropic patches at the bottom of the Earth's lower mantle, whereas the bulk of the mantle has been observed to be largely isotropic. These patches have been interpreted to correspond to areas where subduction has taken place in the past or to areas where mantle plumes are upwelling, but the underlying cause for the anisotropy is unknown-both shape-preferred orientation of elastically heterogeneous materials and lattice-preferred orientation of a homogeneous material have been proposed. Both of these mechanisms imply that large-strain deformation occurs within the anisotropic regions, but the geodynamic implications of the mechanisms differ. Shape-preferred orientation would imply the presence of large elastic (and hence chemical) heterogeneity whereas lattice-preferred orientation requires deformation at high stresses. Here we show, on the basis of numerical modelling incorporating mineral physics of elasticity and development of lattice-preferred orientation, that slab deformation in the deep lower mantle can account for the presence of strong anisotropy in the circum-Pacific region. In this model-where development of the mineral fabric (the alignment of mineral grains) is caused solely by solid-state deformation of chemically homogeneous mantle material-anisotropy is caused by large-strain deformation at high stresses, due to the collision of subducted slabs with the core-mantle boundary.     

Contrasting crustal production and rapid mantle transitions beneath back-arc ridges

The opening of back-arc basins behind subduction zones progresses from initial rifting near the volcanic arc to seafloor spreading. During this process, the spreading ridge and the volcanic arc separate and lavas erupted at the ridge are predicted to evolve away from being heavily subduction influenced (with high volatile contents derived from the subducting plate). Current models predict gradational, rather than abrupt, changes in the crust formed along the ridge as the inferred broad melting region beneath it migrates away from heavily subduction-influenced mantle. In contrast, here we show that across-strike and along-strike changes in crustal properties at the Eastern Lau spreading centre are large and abrupt, implying correspondingly large discontinuities in the nature of the mantle supplying melt to the ridge axes. With incremental separation of the ridge axis from the volcanic front of as little as 5?km, seafloor morphology changes from shallower complex volcanic landforms to deeper flat sea floor dominated by linear abyssal hills, upper crustal seismic velocities abruptly increase by over 20%, and gravity anomalies and isostasy indicate crustal thinning of more than 1.9?km. We infer that the abrupt changes in crustal properties reflect rapid evolution of the mantle entrained by the ridge, such that stable, broad triangular upwelling regions, as inferred for mid-ocean ridges, cannot form near the mantle wedge corner. Instead, the observations imply a dynamic process in which the ridge upwelling zone preferentially captures water-rich low-viscosity mantle when it is near the arc. As the ridge moves away from the arc, a tipping point is reached at which that material is rapidly released from the upwelling zone, resulting in rapid changes in the character of the crust formed at the ridge.     

The Earth's mantle

Seismological images of the Earth's mantle reveal three distinct changes in velocity structure, at depths of 410, 660 and 2,700 km. The first two are best explained by mineral phase transformations, whereas the third-the D" layer-probably reflects a change in chemical composition and thermal structure. Tomographic images of cold slabs in the lower mantle, the displacements of the 410-km and 660-km discontinuities around subduction zones, and the occurrence of small-scale heterogeneities in the lower mantle all indicate that subducted material penetrates the deep mantle, implying whole-mantle convection. In contrast, geochemical analyses of the basaltic products of mantle melting are frequently used to infer that mantle convection is layered, with the deeper mantle largely isolated from the upper mantle. We show that geochemical, seismological and heat-flow data are all consistent with whole-mantle convection provided that the observed heterogeneities are remnants of recycled oceanic and continental crust that make up about 16 and 0.3 per cent, respectively, of mantle volume.     

Possible presence of high-pressure ice in cold subducting slabs

During the subduction of oceanic lithosphere, water is liberated from minerals by progressive dehydration reactions and is thought to be critical to several geologically important processes such as island-arc volcanism, intermediate-depth seismicity and chemical exchange between the subducting lithosphere and mantle. Although dehydration reactions would yield supercritical fluid water in most slabs, we report here that the stable phase of H2O should be solid ice VII in portions of the coldest slabs. The formation of ice VII as a dehydration product would affect the generation, storage, transport and release of water in cold subduction zones and equilibrium conditions of dehydration would shift, potentially affecting the depths of seismogenesis and magmagenesis. Large amounts of pure ice VII might accumulate during subduction and, as a sinking slab warms, eventual melting of the ice would release large amounts of water in a small region over a short period of time, with a significant positive volume change. Moreover, the decreasing availability of fluid water, owing to the accumulation of ice VII and its subsequent reaction products in a cooling planetary interior (for example, in Mars or the future Earth), might eventually lead to a decline in tectonic activity or its complete cessation.     

Melting above the anhydrous solidus controls the location of volcanic arcs

Segregation of magma from the mantle in subduction zones is one of the principal mechanisms for chemical differentiation of the Earth. Fundamental aspects of this system, in particular the processes by which melt forms and travels to the Earth's surface, remain obscure. Systematics in the location of volcanic arcs, the surface expression of this melting, are widely considered to be a clue to processes taking place at depth, but many mutually incompatible interpretations of this clue exist (for example, see refs 1-6). We discriminate between those interpretations by the use of a simple scaling argument derived from a realistic mathematical model of heat transfer in subduction zones. The locations of the arcs cannot be explained by the release of fluids in reactions taking place near the top of the slab. Instead, the sharpness of the volcanic fronts, together with the systematics of their locations, requires that arcs must be located above the place where the boundary defined by the anhydrous solidus makes its closest approach to the trench. We show that heat carried by magma rising from this region is sufficient to modify the thermal structure of the wedge and determine the pathway through which both wet and dry melts reach the surface.     

Origin of the Changbai intraplate volcanism in Northeast China： Evidence from seismic tomography

Seismic images of the mantle beneath the active Changbai intraplate volcano in Northeast China determined by teleseismic travel time tomography are presented. The data are measured at a new seismic network consisting of 19 portable stations and 3 permanent stations. The results show a columnar low-velocity (-3%) anomaly extending to 400 km depth under the Changbai volcano. High velocity anomalies are visible in the mantle transition zone, and deep earthquakes occur at depths of 500--600 km under the region,suggesting that the subducting Pacific slab is stagnant in the transition zone, as imaged clearly also by global tomography.These results suggest that the Changbai intraplate volcano is not a hotspot like Hawaii but a kind of back-arc volcano related to the upwelling of hot asthenospheric materials associated with the deep subduction and stagnancy of the Pacific slab under northeast Asia.     

Subduction dynamics and the origin of Andean orogeny and the Bolivian orocline

The building of the Andes results from the subduction of the oceanic Nazca plate underneath the South American continent. However, how and why the Andes and their curvature, the Bolivian orocline, formed in the Cenozoic era (65.5?million years (Myr) ago to present), despite subduction continuing since the Mesozoic era (251.0-65.5?Myr ago), is still unknown. Three-dimensional numerical subduction models demonstrate that variations in slab thickness, arising from the Nazca plate's age at the trench, produce a cordilleran morphology consistent with that observed. The age-dependent sinking of the slab in the mantle drives traction towards the trench at the base of the upper plate, causing it to thicken. Thus, subducting older Nazca plate below the Central Andes can explain the locally thickened crust and higher elevations. Here we demonstrate that resultant thickening of the South American plate modifies both shear force gradients and migration rates along the trench to produce a concave margin that matches the Bolivian orocline. Additionally, the varying forcing along the margin allows stress belts to form in the upper-plate interior, explaining the widening of the Central Andes and the different tectonic styles found on their margins, the Eastern and Western Cordilleras. The rise of the Central Andes and orocline formation are directly related to the local increase of Nazca plate age and an age distribution along the margin similar to that found today; the onset of these conditions only occurred in the Eocene epoch. This may explain the enigmatic delay of the Andean orogeny, that is, the formation of the modern Andes.     

Faulting induced by precipitation of water at grain boundaries in hot subducting oceanic crust

Dehydration embrittlement has been proposed to explain both intermediate- and deep-focus earthquakes in subduction zones. Because such earthquakes primarily occur at shallow depths or within the core of the subducting plate, dehydration at relatively low temperatures has been emphasized. However, recent careful relocation of subduction-zone earthquakes shows that at depths of 100-250 km, earthquakes continue in the uppermost part of the slab (probably the former oceanic crust that has been converted to eclogite) where temperatures are higher. Here we show that at such pressures and temperatures, eclogite lacking hydrous phases but with significant hydroxyl incorporated as defects in pyroxene and garnet develops a faulting instability associated with precipitation of water at grain boundaries and the production of very small amounts of melt. This new faulting mechanism satisfactorily explains high-temperature earthquakes in subducting oceanic crust and could potentially be involved in much deeper earthquakes in connection with similar precipitation of water in the mantle transition zone (400-700 km depth). Of potential importance for all proposed high-pressure earthquake mechanisms is the very small amount of fluid required to trigger this instability.     

Bending-related faulting and mantle serpentinization at the Middle America trench

The dehydration of subducting oceanic crust and upper mantle has been inferred both to promote the partial melting leading to arc magmatism and to induce intraslab intermediate-depth earthquakes, at depths of 50-300 km. Yet there is still no consensus about how slab hydration occurs or where and how much chemically bound water is stored within the crust and mantle of the incoming plate. Here we document that bending-related faulting of the incoming plate at the Middle America trench creates a pervasive tectonic fabric that cuts across the crust, penetrating deep into the mantle. Faulting is active across the entire ocean trench slope, promoting hydration of the cold crust and upper mantle surrounding these deep active faults. The along-strike length and depth of penetration of these faults are also similar to the dimensions of the rupture area of intermediate-depth earthquakes.     

Trace element signature of subduction-zone fluids, melts and supercritical liquids at 120-180 km depth

Fluids and melts liberated from subducting oceanic crust recycle lithophile elements back into the mantle wedge, facilitate melting and ultimately lead to prolific subduction-zone arc volcanism. The nature and composition of the mobile phases generated in the subducting slab at high pressures have, however, remained largely unknown. Here we report direct LA-ICPMS measurements of the composition of fluids and melts equilibrated with a basaltic eclogite at pressures equivalent to depths in the Earth of 120-180 km and temperatures of 700-1,200 degrees C. The resultant liquid/mineral partition coefficients constrain the recycling rates of key elements. The dichotomy of dehydration versus melting at 120 km depth is expressed through contrasting behaviour of many trace elements (U/Th, Sr, Ba, Be and the light rare-earth elements). At pressures equivalent to 180 km depth, however, a supercritical liquid with melt-like solubilities for the investigated trace elements is observed, even at low temperatures. This mobilizes most of the key trace elements (except the heavy rare-earth elements, Y and Sc) and thus limits fluid-phase transfer of geochemical signatures in subduction zones to pressures less than 6 GPa.     

Mineralogical evidence for continental deep subduction

Diamond is an index mineral to prove ultrahigh pressure (UHP) metamorphic conditions because it is only stable at the pressures above 3.3 GPa. Its occurrence in eclogite-facies metamorphic rocks suggests plate subduction to depths over 120 km assuming the normal gradient of lithostatic pressure. Because UHP eclogites are the metamorphic products of basaltic rocks, the occurrence of diamond in the eclogites demonstrates a complete geodynamic cycle in that mafic crustal rocks were subducted t…     

The subducted component in island arc lavas: constraints from Be isotopes and B-Be systematics

Although high concentrations of beryllium-10 and boron are taken as unequivocal indicators of the contribution of subducting plates, controversy persists about the processes by which material is transferred from slabs to the sources of arc magmas. Data on (10)Be/Be and B/Be ratios from four arcs suggest that the contribution from the slab is compositionally homogeneous in each arc and that subducted boron is not stored in the sub-arc mantle. The link between subduction and magma-tism at convergent margins seems to be well regulated.     

Sound velocities of majorite garnet and the composition of the mantle transition region

The composition of the mantle transition region, characterized by anomalous seismic-wave velocity and density changes at depths of approximately 400 to 700 km, has remained controversial. Some have proposed that the mantle transition region has an olivine-rich 'pyrolite' composition, whereas others have inferred that it is characterized by pyroxene- and garnet-rich compositions ('piclogite'), because the sound velocities in pyrolite estimated from laboratory data are substantially higher than those seismologically observed. Although the velocities of the olivine polymorphs at these pressures (wadsleyite and ringwoodite) have been well documented, those of majorite (another significant high-pressure phase in the mantle transition region) with realistic mantle compositions have never been measured. Here we use combined in situ X-ray and ultrasonic measurements under the pressure and temperature conditions of the mantle transition region to show that majorite in a pyrolite composition has sound velocities substantially lower than those of earlier estimates, owing to strong nonlinear decreases at high temperature, particularly for shear-wave velocity. We found that pyrolite yields seismic velocities more consistent with typical seismological models than those of piclogite in the upper to middle parts of the region, except for the potentially larger velocity jumps in pyrolite relative to those observed at a depth of 410 km. In contrast, both of these compositions lead to significantly low shear-wave velocities in the lower part of the region, suggesting possible subadiabatic temperatures or the existence of a layer of harzburgite-rich material supplied by the subducted slabs stagnant at these depths.