Search

International audience; Oceanic detachment faulting is a major mode of seafloor accretion at slow and ultraslow spreading mid-ocean ridges, and is associated with dramatic changes in seafloor morphology. Detachments form expansive dome structures with corrugated surfaces known as oceanic core complexes (OCCs), and often transition to multiple regularly-spaced normal faults that form abyssal hills parallel to the spreading axis. Previous studies have attributed these changes to along-axis gradients in lithospheric strength or magma supply. However, despite the recognition that magma supply can influence fault style and seafloor morphology, the mechanics controlling the transition from oceanic detachment faults to abyssal hill faults and the relationship to along-axis variations in magma supply remain poorly understood. This study investigates this issue using two complementary modeling approaches. The first consists of semi-analytical, two-dimensional (2-D) cross-axis models designed to address the fundamental mechanical controls on the longevity of normal faults. These 2-D model sections are juxtaposed in the along-axis direction to examine the response of the plan-view pattern of faults to along-axis variations in magmatic accretion in the absence of along-axis mechanical coupling. The second approach uses three-dimensional (3-D), time-dependent numerical models that simulate faulting and magma intrusion in a visco-elasto-plastic continuum. The primary variable studied through both approaches is the along-axis gradient in the fraction M of seafloor spreading that is accommodated by magmatism. The 2-D and 3-D results predict different abyssal hill spacing and orientation, however the plan-view geometry of self-emerging detachment faults predicted by the 3-D numerical models are well explained by the juxtaposed 2-D models. This indicates a first-order control by cross-axis effects of changing values of M. These models are also shown to explain the along-axis extent and plan-view curvature of the well-developed 13 • 20 N and Mt. Dent OCCs (Mid-Atlantic Ridge and Cayman Rise) in terms of quantifiable along-axis gradients in magma emplacement rates.

There is debate to what degree mid-ocean ridge basalts (MORB) can be used to infer the composition of the sub-ridge mantle and whether the upper mantle contains components that are difficult to recognize by observations on MORB alone. Here we report trace element and isotope data on a fossil mantle section exposed in the ophiolitic Lanzo South massif, Italy. Our findings show the existence of clinopyroxenes in geochemically depleted melt-infiltrated harzburgites with an extreme radiogenic Hf-isotopic composition but MORB-like Nd isotopic compositions. The simplest way of explaining the depleted trace elements in combination with the decoupled isotopic compositions is that the harzburgites were infiltrated by a melt with isotopic compositions inherited from an anciently (>1 Ga) depleted source. The existence of melts from ultra-depleted sectors of the asthenosphere allows constraining of the configuration of the heterogeneities in the sub-ridge mantle. The parallel arrays in Hf–Nd isotope composition space of different Mid-Ocean ridge segments imply the ubiquitous presence of anciently depleted residual mantle as the matrix in which more enriched components occur in pockets.

Ocean ridges are one of the primary connections between Earth's mantle and surface. Therefore, understanding the evolution of ocean ridge processes through Earth history is important for understanding how Earth's surface and mantle have evolved. We combine an analytic model of mantle convection with a petrologic model of mantle melting in the presence of CO2 to compute the mantle temperature, plate speed, melt production, and CO2 outgassing flux at ocean ridges for the past 4 billion years. We explore a large suite of realistic mantle and lithospheric parameters to map out a full range of possible thermal histories. Our results show that in order to satisfy thermal constraints, the Earth must have started in a sluggish-lid mode of plate tectonics and transitioned to an active-lid mode. Furthermore, we show that plate speed and CO2 outgassing flux do not necessarily scale with mantle temperature, and that it is possible to reach present-day mantle temperatures and plate speeds with a simple force balance that does not invoke any feedbacks (e.g. grain size evolution, dehydration stiffening) or a fully stagnant-lid mode of convection in the Precambrian. The solutions show a range of evolutionary behaviors depending on the parameters chosen. The transition from sluggish- to active-lid can be smooth, abrupt, or somewhere in between. A smooth transition is difficult to distinguish from a purely active-lid evolutionary path based on temperature, plate speed, melt production, and CO2 outgassing flux. In contrast, an abrupt transition leads to a large increase in the average plate speed with a corresponding increase in melt production and CO2 outgassing flux. As an abrupt transition would have a much larger effect on CO2 outgassing and melt production than a change in ridge length, we investigate the impact of rapidly changing plate speeds and do not consider changes in ridge length. Finally, we show that carbon recycling is required for a large part of Earth history in order to explain present-day CO2 outgassing. Our model highlights the importance of understanding the style of mantle convection when calculating melt production and volatile fluxes through Earth's history.

Iceland is a natural laboratory where plate separation and mid-ocean ridge formation can be studied directly in the field. In this setting, it is not fully understood how magma interacts with faulting and fissuring, and how rift zones propagate. In order to improve the understanding of how these processes operate, we conducted a detailed field study about the kinematics and propagation of 33 Holocene faults along the N–S Theistareykir rift (North Volcanic Zone) and mapped all tension fractures. This rift hosts the Holocene Theistareykir central volcano, for which ground deformation data (InSAR and GPS) indicate an inflating shallow magma chamber throughout 2006–2008. Analyses of the cumulative fault slip distribution at 696 sites show two opposite directions of fault/rift propagation: The first, more developed, is directed northwards and affects the area located north of Theistareykir volcano, whereas south of the volcano, the rift propagates southwards. Tension fractures increase in frequency, with respect to faults, outwards from the volcano, along a N–S to NNE–SSW direction. The presence of aligned vents, dykes and topographic bulging along some of the N–S structures, and the active magma chamber below Theistareykir volcano, suggest that faults and tension fractures propagate following repeated dyke intrusions from the magma chamber outward along the plate margin. The fact that the rift is more developed north of the volcano than south of it, is interpreted as the effect of buttressing from another, coterminous magma chamber sitting below a neighbouring central volcano. Other causes for the general observed pattern such as tectonic stresses, mechanical interaction of faults, and changes in the rheological characteristics of rocks, are discussed as well.

Combined Ce and Nd isotope ratios provide a time-integrated record of the light rare earth element (LREE) abundances of their source materials. Here, we present new high precision Ce isotope data for chondrites and basalts from ocean islands (OIB) and mid ocean ridges (MORB). The new Ce isotope ratios in chondritic meteorites define a precise new CHUR reference value. In Ce–Nd isotopic space, the MORB and OIB form a well-defined array that intersects with the Ce–Nd chondritic reference value. The simplest first-order explanation is that the bulk silicate Earth (BSE) has chondritic LREE and Ce–Nd isotope ratios. We show, however, that the intercept and slope of the Ce–Nd isotope mantle array depend on several factors. Perhaps most important are whether the BSE is chondritic and how closely the mantle average reflected in the MORB and OIB data corresponds to the Ce–Nd isotope ratio of the BSE. A significant difference between the accessible mantle's average Ce–Nd isotope ratio and that of BSE could result from the permanent storage of a considerable proportion of Earth's total LREE budget in the continental crust or potential isolated reservoirs. We show that the formation of isolated reservoirs either has a minor effect on the average Ce–Nd composition of the average mantle, or is geochemically and geodynamically implausible. If, due to formation of the continental crust, a significant shift in the average mantle's Ce–Nd isotope composition relative to BSE occurs, this shift is parallel to the Ce–Nd mantle array, and does not affect its chondritic intercept. The chondritic intercept of the Ce–Nd isotope mantle array therefore is strong evidence that BSE's relative LREE and Ce–Nd isotope composition is chondritic. However, an apparent difference between the accessible mantle's average Ce–Nd isotope ratio and that of BSE could also result if MORB and OIB do not sample the accessible mantle in a representative manner. Although we cannot entirely exclude the latter, it would require a fortuitous combination of factors to cause the observed chondritic intercept of the Ce–Nd isotope mantle array. We therefore conclude that the bulk silicate Earth has chondritic LREE and Ce–Nd isotope ratios.

We report 3He/4He, He and CO2 concentrations for the dissolved (glass) and vapor (vesicle) phase of basalts, for 23 lavas from the 2005–2006 eruption and 4 lavas from 1991–92 at 9°50′N on the East Pacific Rise. We also determined δ 13 C and δ 18 O of CO2 in vesicles in 18 of these lavas. These sample suites provide a rare opportunity to study volatile systematics related to magma recharge at a mid-ocean ridge volcano and to quantify degassing prior to and during the eruptions. Our study covers the spatial and temporal extent of the 2005–06 eruption and is complementary to previous studies of variations in basaltic volatiles with distance from the ridge axis during part of this multi-stage eruption ( Soule et al., 2012 , Gardner et al., 2016 ). 3He/4He shows minor variation with a mean value of 8.51 ± 0.03 RA. All but 1 sample lies in a narrow range of 280–420 ppm for total (vesicles + glass) CO2 indicating volatile saturation at 1.2–2.3 km depth in the crust, similar to depths of the seismically imaged magma lenses in the region. Vesicle He and CO2 concentrations vary throughout the lava flow field by factors of 17 and 250, respectively. The vapor phase CO2/He ratio co-varies positively with the fraction of CO2 contained in vesicles due to kinetic fractionation between He and CO2 during vesiculation. Vesicle δ 13 CPDB co-varies with the fraction of CO2 in vesicles ( r 2 = 0.83 ) , ranging from −2.6 to −5.0‰, and straddles the isotope composition of EPR vent fluids. The vesicle 13C-enrichment ( δ vapor − δ melt ) is +3.3‰ and comparable to the experimentally determined fractionation factor for basaltic systems. Collectively the observations indicate closed-system degassing during rapid ascent of melt from the axial magma lens through the conduit to its eruption and emplacement on the seafloor, with minimal bubble loss. In contrast, CO2/Ba systematics and a comparison of δ 13 C for lavas and vent fluids indicate an earlier period of open-system degassing during which 50% or more of the initial (parental magma) CO2 inventory was lost prior to magma storage in the shallow crust.

In subduction zones, fluids rise from the slab to the mantle, causing metasomatism and flux melting of the mantle to produce arc magmas. The transfer of material from slab to mantle and, in turn, to arc crust is an important control on the long-term chemical evolution of the mantle and continental crust. In this study, we investigate the transport of K in subduction zones by exploring the systematics of K stable-isotope variations in lavas of the Izu arc. We find that the Izu lavas have isotopically heavy K relative to estimates for midocean ridge basalt (MORB)-source upper mantle. Moreover, the δ 41 K values of the lavas are clearly heavier than those of subducting sediments and are probably heavier than subducting altered ocean crust. An across-arc decrease in δ 41 K values is apparent. Arc-front lavas are heavier than the mantle by about 0.22‰ (median), whereas rear-arc lavas are heavier by only about 0.08‰ (median). The heavy K-isotope compositions of the arc lavas may arise from isotopic fractionation during slab dehydration, where light K is preferentially retained in phases such as phengite in the slab. The across-arc decrease in δ 41 K values may be due to progressive breakdown of these phases, and to associated depletion of the slab in heavy K. Variations in the relative contributions of different source materials—igneous ocean crust, sediment, and mantle peridotite—may also play a role. In particular, we explore a possibility, motivated by radiogenic-isotope studies, that the slab signal in K isotopes may be attenuated in the rear arc as a result of extensive fluid-mantle interaction. If K isotopes do track slab dehydration, then K isotopes provide insight into the transfer of K and similarly mobile elements out of the slab and into the upper mantle and arc crust. Lastly, we observe extreme isotopic variations in some of the lavas, which we interpret to result from crustal-level or Earth-surface processes that affect only a subset of the lavas.

Many oceanic plateaus have been emplaced at or adjacent to mid-ocean ridges. To explain plateau volume and thickened crust compared to normal oceanic crust, hotspot–ridge interaction is commonly assumed, but the manner of interaction remains unclear. The Shatsky Rise oceanic plateau is a large volcanic mountain that formed at a triple junction during Late Jurassic and Early Cretaceous time. Recent drilling and seismic investigations suggest that the intermediate edifice in the rise, Ori Massif, is a central volcano. Paradoxically, magnetic lineations were traced across parts of Ori Massif, implying formation at a spreading ridge. In this study, we examined magnetic anomalies over and around Ori Massif to obtain insights about the formation of this volcanic edifice. Magnetic data from 21 cruises were corrected, combined, and gridded to construct a magnetic anomaly map. Forward and inverse magnetic modeling was done to investigate the magnetic structure of Ori Massif. The results imply that this large volcanic edifice is predominantly characterized by linear magnetic anomalies resulting from alternating normal and reversed polarity magnetization blocks, analogous to magnetic anomalies recorded by spreading-ridges. This magnetic structure is not expected for a central volcano that was built by long runout lava flows, implying that Ori Massif eruptions must have been constrained near the ridge axis. Magnetic bights on the north and south boundaries of Ori Massif imply that it was bracketed by triple junctions, indicating complex ridge tectonics during the formation of Shatsky Rise. The surprising finding that Ori Massif is traversed by coherent linear magnetic anomalies indicates that oceanic plateaus can record seafloor spreading magnetic anomalies despite large crustal thickness. Other oceanic plateaus also record linear magnetic anomalies, implying a link between divergent plate boundaries and oceanic plateau volcanism.

The fluxes of volatile elements from the mantle have long been used to understand mantle structure and evolution, and are critical controls on Earth's climate stability. Because of the ubiquity of magmatic degassing, inferring pre-degassing volatile concentrations from measured basalts requires the application of a degassing model. Such models, including the commonly-applied equilibrium Rayleigh distillation, typically assume equilibrium or solubility-based partitioning between melt and vapor. Here, we demonstrate that ratios of radiogenic isotopes of He, Ne, Ar and, especially, Xe measured in global mid-ocean ridge basalts (MORBs) are inconsistent with equilibrium degassing models, even when not considering He. We conclude that kinetic disequilibrium is a crucial process affecting volatile abundances during degassing. We present a simple disequilibrium Rayleigh distillation model to reconstruct pre-degassing MORB noble gas and carbon concentrations, which predicts that He and Ne achieve nearly equilibrium partitioning between melt and vesicles, but the slower-diffusing heavier noble gases are strongly affected by disequilibrium, resulting in non-equilibrium fractionation of noble gas elemental ratios, and other volatile element ratios like CO2/He. We apply our model to a large set of MORB data, and find average pre-degassing 3He, 22Ne, and 36Ar concentrations of 4.4 ± 0.9 × 10−10, 6.6 ± 1.4 × 10−11, and 6.8 ± 4.5 × 10−10 ccSTP/g (2σ), with variations of approximately 2 orders of magnitude, similar to other highly incompatible elements. Pre-degassing noble gas concentrations imply a mid-ocean ridge 3He flux of 800 ± 170 mol/yr and upper mantle 3He/22Ne and 3He/36Ar ratios of 6.6 ± 2.0 and 0.64 ± 0.44, but substantial variability in these ratios between samples. Applying our model to CO2, we calculate an average mantle CO2/3He molar ratio of 1.67 ± 0.21 × 109, which, when combined with our estimate of 3He flux, implies an upper mantle CO2 flux of 5.9 ± 1.0 × 1013 g/yr and a CO2 concentration of approximately 110 ppm. Our estimate of the mantle 3He flux is the first determined independently of oceanographic 3He measurements, and consequently represents a time-integrated flux substantially longer than ∼1000 years. And although at the high end of the range of previous estimates, our estimate does not resolve the long-standing heat-helium paradox. Additionally, our requirement for heterogeneous pre-degassing 3He/22Ne and 3He/36Ar ratios between samples is contrary to conclusions of previous applications of disequilibrium degassing models, which advocated uniform ratios. Furthermore, we find that CO2/Ba ratios are highly variable in MORB samples, but still consistent with an average mantle mass ratio of ∼100. However, estimating pre-degassing CO2 concentrations and the mantle CO2 flux depend strongly on the poorly-constrained carbon diffusivity. Consequently, our demonstration of the prevalence of disequilibrium during mid-ocean ridge degassing, and the potential for disequilibrium in other volcanic settings, highlights the need for better characterization of the physical parameters associated with volcanic degassing.

The oxygen fugacity (fO2) of the oceanic upper mantle has fundamental implications for the production of magmas and evolution of the Earth's interior and exterior. Mid-ocean ridge basalts and peridotites sample the oceanic upper mantle, and retain a record of oxygen fugacity. While fO2 has been calculated for mid-ocean ridge basalts worldwide (>200 locations), ridge peridotites have been comparatively less well studied (33 samples from 11 locations), and never in the same geographic location as basalts. In order to determine whether peridotites and basalts from mid-ocean ridges record congruent information about the fO2 of the Earth's interior, we analyzed 31 basalts and 41 peridotites from the Oblique Segment of the Southwest Indian Ridge. By measuring basalts and peridotites from the same ridge segment, we can compare samples with maximally similar petrogenetic histories. We project the composition and oxygen fugacity of each lithology back to source conditions, and evaluate the effects of factors such as subsolidus diffusion in peridotites and fractional crystallization in basalts. We find that, on average, basalts and peridotites from the Oblique Segment both reflect a source mantle very near the quartz–fayalite–magnetite (QFM) buffer. However, peridotites record a significantly wider range of values (nearly 3 orders of magnitude in fO2), with a single dredge recording a range in fO2 greater than that previously reported for mid-ocean ridge peridotites worldwide. This suggests that mantle fO2 may be heterogeneous on relatively short length scales, and that this heterogeneity may be obscured within aggregated basalt melts. We further suggest that the global peridotite fO2 dataset may not provide a representative sample of average basalt-source mantle. Our study motivates further investigation of the fO2 recorded by ridge peridotites, as peridotites record information about the fO2 of the Earth's interior that cannot be gleaned from analysis of basalts alone.

Vanadium isotope compositions of igneous rocks have the potential to constrain variations of physico-chemical conditions such as oxidation states during magmatism. Here, we present V isotope data for 27 fresh lavas (ranging from basaltic to dacitic compositions) from mid-ocean ridges, 31 altered basalts and gabbros from IODP site 1256 near the East Pacific Rise (EPR), and 2 back arc basin basalts (BABB). Our analyses of fresh mid-ocean ridge basalt (MORB) provide new constraints on the V isotope composition of MORBs, i.e. δ 51 V = − 0.84 ± 0.02 ‰ (2SE, n = 22 ). In addition, the mean δ 51 V of MORBs from individual segments is correlated with the mean ridge depth and Na8.0 of the segment, which might reflect the effect of melting extent on V isotope fractionation during mantle melting. The mafic profile of intact altered oceanic crust (AOC) from the IODP site 1256 has δ 51 V ranging from −1.01 to − 0.77 ‰ , similar to that of fresh MORBs, suggesting that V isotope fractionation is limited during alteration of oceanic crust. These results also indicate the V isotopic homogeneity of the bulk oceanic crust with average δ 51 V of − 0.85 ± 0.02 ‰ (2SE, n = 53 ), which is unaffected by ocean water and hydrothermal fluid alteration. Our results provide a guideline for application of V isotopes into studies of low and high temperature geochemical processes. The evolved lavas (basaltic andesites, andesites, and dacites) from the East Pacific Rise (EPR) show apparent shifts towards heavy δ 51 V values with increasing degree of differentiation, which can be explained by the crystal–liquid fractionation during crystallization with an inferred isotope fractionation factor of Δ 51 V mineral - melt = − 0.15 × 10 6 / T 2 . The enrichment of 51V with increasing differentiation degree for the 9°N Overlapping Spreading Center (OSC) lavas is consistent with direction of the isotope shift observed in lavas from Anatahan Island (Northern Mariana Arc) and Hekla Volcano (Iceland), but the magnitude (0.3‰) is much smaller than that (2‰) reported in Prytulak et al. (2017) . Modeling of V isotope fractionation between mineral and melt shows that variations in redox condition are important for controlling V isotope fractionation, but insufficient to explain the dramatically different Δ 51 Vmineral-melt between 9°N OSC lavas and Anatahan/Hekla suites. More studies are necessary for better understanding of mechanisms of V isotope fractionation during magmatism.

Previous models do not simultaneously reproduce the observed range of axial relief and fault patterns at plate spreading centers. We suggest that this failure is due to the approximation that magmatic dikes open continuously rather than in discrete events. During short – lived events, dikes open not only in the strong axial lithosphere but also some distance into the underlying weaker asthenosphere. Axial valley relief affects the partitioning of magma between the lithosphere and asthenosphere during diking events. The deeper the valley, the more magma goes into lithospheric dikes in each event and so the greater the average opening rate of those dikes. The long-term rate of lithospheric dike opening controls faulting rate and axial depth. The feedback between axial valley depth D and lithospheric dike opening rate allows us to analytically relate steady-state values of D to lithospheric thickness H L and crustal thickness H C . A two-dimensional model numerical model with a fixed axial lithospheric structure illustrates the analytic model implications for axial faulting. The predictions of this new model are broadly consistent with global and segment-scale trends of axial depth and fault patterns with H L and H C .

Stresses from mantle convection deflect Earth's surface vertically, producing dynamic topography that is important for continental dynamics and sea-level change but difficult to observe due to overprinting by isostatic topography. For long wavelengths (∼104 km), the amplitude of dynamic topography is particularly uncertain, with mantle flow models typically suggesting larger amplitudes (>1000 m) than direct observations. Here we develop a new constraint on the amplitude of long-wavelength dynamic topography by examining asymmetries in seafloor bathymetry across mid-ocean ridges. We compare bathymetric profiles across the Mid-Atlantic Ridge (MAR) and the East Pacific Rise (EPR) and we find that the South American flank of both ridges subsides faster than its opposing flank. This pattern is consistent with dynamic subsidence across South America, supported by downwelling in the lower mantle. To constrain the amplitude of dynamic topography, we compare bathymetric profiles across both ridges after correcting bathymetry for several different models of dynamic topography with varying amplitudes and spatial patterns. We find that long-wavelength dynamic topography with an amplitude of only ∼500 m explains the observed asymmetry of the MAR. A similar model can explain EPR asymmetry but is complicated by additional asymmetrical topography associated with tectonic, crustal thickness, and/or asthenospheric temperature asymmetries across the EPR. After removing 500 m of dynamic topography, both the MAR and EPR exhibit a slower seafloor subsidence rate (∼280–290 m/Myr1/2) than previously reported. Our finding of only ∼500 m of long-wavelength dynamic topography may indicate the importance of thermochemical convection and/or large viscosity variations for lower mantle dynamics.

Two distinct igneous differentiation trends – the tholeiitic and calc-alkaline – give rise to Earth's oceanic and continental crust, respectively. Mantle melting at mid-ocean ridges produces dry magmas that differentiate at low-pressure conditions, resulting in early plagioclase saturation, late oxide precipitation, and Fe-enrichment in mid-ocean ridge basalts (MORBs). In contrast, magmas formed above subduction zones are Fe-depleted, have elevated water contents and are more oxidized relative to MORBs. It is widely thought that subduction of hydrothermally altered, oxidized oceanic crust at convergent margins oxidizes the mantle source of arc magmas, resulting in erupted lavas that inherit this oxidized signature. Yet, because our understanding of the calc-alkaline and tholeiitic trends largely comes from studies of erupted melts, the signals from shallow crustal contamination by potentially oxidized, Si-rich, Fe-poor materials, which may also generate calc-alkaline rocks, are obscured. Here, we use deep crustal cumulates to “see through” the effects of shallow crustal processes. We find that the tholeiitic and calc-alkaline trends are indeed reflected in Fe-poor mid-ocean ridge cumulates and Fe-rich arc cumulates, respectively. A key finding is that with increasing crustal thickness, arc cumulates become more Fe-enriched. We propose that the thickness of the overlying crustal column modulates the melting degree of the mantle wedge (lower F beneath thick arcs and vice versa) and thus water and Fe3+ contents in primary melts, which subsequently controls the onset and extent of oxide fractionation. Deep crustal cumulates beneath thick, mature continental arcs are the most Fe-enriched, and therefore may be the “missing” Fe-rich reservoir required to balance the Fe-depleted upper continental crust.

Extension at slow- and intermediate-spreading mid-ocean ridges is commonly accommodated through slip on long-lived faults called oceanic detachments. These curved, convex-upward faults consist of a steeply-dipping section thought to be rooted in the lower crust or upper mantle which rotates to progressively shallower dip-angles at shallower depths. The commonly-observed result is a domed, sub-horizontal oceanic core complex at the seabed. Although it is accepted that detachment faults can accumulate kilometre-scale offsets over millions of years, the mechanism of slip, and their capacity to sustain the shear stresses necessary to produce large earthquakes, remains subject to debate. Here we present a comprehensive seismological study of an active oceanic detachment fault system on the Mid-Atlantic Ridge near 13°20′N, combining the results from a local ocean-bottom seismograph deployment with waveform inversion of a series of larger teleseismically-observed earthquakes. The unique coincidence of these two datasets provides a comprehensive definition of rupture on the fault, from the uppermost mantle to the seabed. Our results demonstrate that although slip on the deep, steeply-dipping portion of detachment faults is accommodated by failure in numerous microearthquakes, the shallow, gently-dipping section of the fault within the upper few kilometres is relatively strong, and is capable of producing large-magnitude earthquakes. This result brings into question the current paradigm that the shallow sections of oceanic detachment faults are dominated by low-friction mineralogies and therefore slip aseismically, but is consistent with observations from continental detachment faults. Slip on the shallow portion of active detachment faults at relatively low angles may therefore account for many more large-magnitude earthquakes at mid-ocean ridges than previously thought, and suggests that the lithospheric strength at slow-spreading mid-ocean ridges may be concentrated at shallow depths.

Mid-ocean ridge peridotites record significantly greater variability in major and trace elements, isotopic compositions, and thermodynamic properties such as oxygen fugacity (fO2) than do their basaltic counterparts. This variability may derive from modern ridge processes related to melting and melt-rock interaction or from long-lived source heterogeneity related to recycled material or ancient melting events. In this study, we investigate variations in spinel geochemistry as well as silicate major and trace element chemistry and oxygen fugacity of a suite of peridotites from a single segment of the Southwest Indian Ridge (SWIR). We present new petrographic analysis and trace element data for samples with previously-published fO2 results and combine this with new data for a suite of SWIR gabbro-veined peridotites. We find that SWIR residual lherzolites record low spinel Cr# (Cr# = 100*Cr/(Cr+Al) 30) record both melt extraction as well as melt-rock interaction. In these samples, spinel Cr# has been substantially elevated by reaction of spinel to form plagioclase during melt addition, complicating the use of spinel Cr# in mid-ocean ridge peridotites as a proxy for degree of melt extraction alone. While spinel Cr# remains a robust proxy for melt extraction within residual, non-melt-influenced samples, mid-ocean ridge peridotites must first be evaluated to ensure that modification by melt-rock reaction has not occurred. Although addition of MORB melt to a peridotite residue modifies spinel Cr#, this melt addition does not result in significant changes to the fO2 recorded by the peridotite. Residual SWIR lherzolites record fO2 of 0.66±0.39 relative to the quartz-fayalite-magnetite buffer (QFM), statistically indistinguishable from melt-influenced and veined SWIR samples (QFM+1.13±0.61). In contrast to other tectonic settings, such as subduction zones, ocean islands, and continental cratons—locations where peridotite is oxidized by petrogenetically unrelated, presumably high-fO2 melts/fluids—ridge peridotites interact with MORB, which has little to no oxidizing power over its own mantle residues. Thus, modern processes beneath the ridge modify peridotite major and trace elements, but do not generate variability in oxygen fugacity.

Serpentinisation of mantle rocks requires the generation and maintenance of transport pathways for water. The solid volume increase during serpentinisation can lead to stress build-up and trigger cracking, which ease fluid penetration into the rock. The quantitative effect of this reaction-induced cracking mechanism on reactive surface generation is poorly constrained, thus hampering our ability to predict serpentinisation rate in geological environments. Here we use a combined approach with numerical modelling and observations in natural samples to provide estimates of serpentinisation rate at mid-ocean ridges. We develop a micromechanical model to quantify the propagation of serpentinisation-induced cracks in olivine. The maximum crystallisation pressure deduced from thermodynamic calculations reaches several hundreds of megapascals but does not necessary lead to crack propagation if the olivine grain is subjected to high compressive stresses. The micromechanical model is then coupled to a simple geometrical model to predict reactive surface area formation during grain splitting, and thus bulk reaction rate. Our model reproduces quantitatively experimental kinetic data and the typical mesh texture formed during serpentinisation. We also compare the model results with olivine grain size distribution data obtained on natural serpentinised peridotites from the Marum ophiolite and the Papuan ultramafic belt (Papua New Guinea). The natural serpentinised peridotites show an increase of the number of olivine grains for a decrease of the mean grain size by one order of magnitude as reaction progresses from 5 to 40%. These results are in agreement with our model predictions, suggesting that reaction-induced cracking controls the serpentinisation rate. We use our model to estimate that, at mid-ocean ridges, serpentinisation occurs up to 12 km depth and reaction-induced cracking reduces the characteristic time of serpentinisation by one order of magnitude, down to values comprised between 10 and 1000 yr. The increase of effective pressure with depth also prevents cracking, which positions the peak in serpentinisation rate at shallower depths, 4 km above previous predictions.

To understand how the mantle cools beneath mid-ocean ridge spreading centers, we applied a REE-in-two-pyroxene thermometer and major element thermometers to peridotites from the Wadi Tayin massif in the southern part of the Samail ophiolite in the Sultanate of Oman, which represent more than 10 km of structural depth beneath the paleo-Moho. Closure temperatures for REEs in pyroxenes deduced from the REE-in-two-pyroxene thermometer ( T REE ) decrease smoothly and systematically with depth in the section, from >1300 °C near the crust to ∼ 10 − 3 °C/y at a depth of six km below the MTZ. Cooling rates derived from Ca-in-olivine thermometry also decrease moving deeper into the section. These variations in cooling rate are most consistent with conductive cooling of the mantle beneath a cold overlying crust. In turn, this suggests that hydrothermal circulation extended to the MTZ near the axis of the fast-spreading ridge where the igneous crust of the Samail ophiolite formed. These observations are consistent with the Sheeted Sills model for accretion of lower oceanic crust, and with previous work demonstrating very rapid cooling rates in the crust of the Wadi Tayin massif. Our observations, combined with previous results, suggest that efficient hydrothermal circulation beneath fast spreading centers cools the uppermost mantle from magmatic temperatures to

Magmatism at mid ocean ridges is one of the main pathways of S outflux from deep Earth to the surface reservoirs and is a critical step in the global sulfur cycle, yet our understanding of the behavior of sulfide during decompression melting of the upper mantle is incomplete. In order to constrain the sulfur budget of the mantle and reconcile the sulfur and chalcophile element budget of mantle partial melts parental to primitive mid-ocean ridge basalts (MORBs), here we developed a model to describe the behavior of sulfide and Cu during decompression melting by combining the pMELTS thermodynamic model and empirical sulfur contents at sulfide concentration (SCSS) models, taking into account the effect of the presence of Ni and Cu in sulfides on SCSS of mantle-derived melts. Calculation of SCSS along melting adiabat at mantle potential temperature of 1380 °C with variable initial S content in the mantle indicates that the complete consumption or partial survival of sulfide in the melting residue depends on initial S content and degree of melting. Primitive MORBs (Mg# > 60) with S and Cu mostly concentrated in 800–1000 ppm and 80–120 ppm are likely mixture of sulfide undersaturated high degree melts and sulfide saturated low degree melts derived from depleted peridotite containing 100–200 ppm S. Model calculations to capture the effects of variable mantle potential temperatures (1280–1420 °C) indicate that for a given abundance of sulfide in the mantle, hotter mantle consumes sulfide more efficiently than colder mantle owing to the effect of temperature in enhancing sulfide solubility in silicate melt, and higher mantle temperature stabilizing partial melt with higher FeO ⁎ and lower SiO 2 and Al 2 O 3 , all of which generally enhance sulfide solubility. However, sulfide can still be exhausted by ∼ 10 – 15 % melting with bulk S of 100–150 ppm in the mantle when T P is as low as 1300 °C. We also show that although variation of D Cu peridotite / melt and initial Cu in the mantle can all affect the Cu concentration of primitive MORBs, 100–200 ppm S in the MORB source mantle can satisfy both S and Cu geochemistry of partial melts parental to ocean floor basalts.

How the thickened crust of a large igneous province on an incoming oceanic plate is accommodated at a subduction zone remains an open question. New Zealand is one of the few places to study this, as at ca. 105 Ma the ca. 35 km-thick Hikurangi Plateau impacted the Gondwana subduction zone in what is now the South Island. Here we report on results from a forty-station portable seismograph array in the southern South Island, designed to delineate the leading edge of the subducted plateau. Three-dimensional images of Vp and Vp/Vs reveal the southwestern part of the plateau was a relatively narrow salient, and the first part to be subducted. The plateau then rotated clockwise about this salient until the southern edge of the plateau was parallel to subduction strike and subduction ceased at ca. 100 Ma. Our results suggest that the global-scale plate reorganization event at 105–100 Ma was due to a cessation of subduction caused by the Hikurangi Plateau choking the Gondwana subduction zone, rather than the subduction of mid ocean ridges as previously proposed. The choking of Gondwana subduction by the plateau also led to a concentration of slab pull in the adjacent subducted oceanic crust, explaining the episode of basin opening and intraplate magmatism there that occurred at the same time. Our study underlines the havoc caused by impact of a large igneous province with a subduction zone.

Magma supply for dike injection can be highly variable within a segment of a slow-spreading mid-ocean ridge but the tectonic impact of this variability is not fully understood. Here, we use three-dimensional numerical models to quantify the effects of variable diking rates on the faulting mode at a 20 km-long slow spreading ridge segment. In addition to end-member faulting modes in which long-lived detachment faults or short-lived normal faults form along the whole segment, we newly identify a transitional mode in which a detachment and a short-lived normal fault form simultaneously but in respective domains separated by a transfer fault. Different faulting modes can be better correlated with the average dike intrusion rate, rather than the highest or lowest rate along the segment. Along-axis stress coupling tends to homogenize fault offset along the segment, inhibiting the domination of a particular faulting mode associated with an extreme local diking rate. This homogenizing effect explains why detachment faults can sometimes form even in the regions previously considered as unfavorable. Our results further suggest that a long (>15 km) and continuous detachment, partially overlain by younger faults, can create an oceanic core complex when faults weaken fast and diking rate is low. When faults weaken slow and diking rate is moderate, however, faulting occurs in the transitional mode, producing a detachment over only a part of the segment length.

Crustal features of the Reykjanes Ridge have been attributed to mantle plume flow radiating outward from the Iceland hotspot. This model requires very rapid mantle upwelling and a “rheological boundary” at the solidus to deflect plume material laterally and prevent extreme melting above the plume stem. Here we propose an alternative explanation in which shallow buoyant mantle upwelling instabilities propagate along axis to form the crustal features of the ridge and flanks. As only the locus of buoyant upwelling propagates this mechanism removes the need for rapid mantle plume flow. Based on new geophysical mapping we show that a persistent sub-axial low viscosity channel supporting buoyant mantle upwelling can explain the current oblique geometry of the ridge as a reestablishment of its original configuration following an abrupt change in opening direction. This mechanism further explains the replacement of ridge-orthogonal crustal segmentation with V-shaped crustal ridges and troughs. Our findings indicate that crustal features of the Reykjanes Ridge and flanks are formed by shallow buoyant mantle instabilities, fundamentally like at other slow spreading ridges, and need not reflect deep mantle plume flow.

Basalts from the East Pacific Rise (EPR), Siqueiros transform zone and Mid-Atlantic Ridge (MAR) area have been analyzed for the platinum-group elements (PGE) and a wide range of incompatible elements. The low PGE content of the most primitive mid-ocean ridge basalts (MORB) suggests that they leave the mantle sulfide-saturated but become sulfide under-saturated, as a consequence of decompression, during their ascent from the mantle to the MOR magma chamber. Because the pressure drop is small, the ascending magma enters the magma chamber slightly sulfide under-saturated and requires only a small amount of fractional crystallization to return it to sulfide-saturation. Sulfide-saturation is marked by the Pd content of the melt falling by over an order of magnitude at ca. 9.5 wt.% MgO. However, once the magma has become sulfide-saturated, it shows no evidence of a further decline in Pd with decreasing MgO as the system evolves. This must result from regular replenishment of the underlying axial magma chamber by fresh batches of primitive magma. Most replenishments occur when the MgO content of the resident melt in the magma chamber is between ∼9 to 6.5 wt.% MgO. We combine the tight constraints imposed by highly compatible Pd abundances, with those imposed by strongly incompatible elements, to produce the first model that successfully accounts for the variations of both classes of elements in open system mid-ocean ridges magma chambers for normal-type MORB (N-MORB). We show that the lack of decline in the Pd content of sulfide-saturated mid-ocean ridge basalts, with decreasing MgO, requires frequent small replenishments (between 1% and 4%) of similar magnitude, rather than large initial inputs that systematically decline from near 100% to near zero during the life of an individual ridge system. Assuming the spreading rate of ∼110 mm/yr for the EPR at 9°N, and 14 yr between the 1991–1992 and 2005–2006 eruptions, the calculated volume of erupted and intruded magma is 2.3 × 103 m3/m of ridge length. If the expelled melt represents 0.5% of the accessible magma in the chamber, as suggested by the Pd modeling, the volume of the magma accessible during tapping is 4.6 × 105 m3/m of ridge length.

Basalts erupted within intra-transform spreading centers can be valuable probes of geochemical components in Earth's upper mantle, and provide constraints on the proportions of mantle carbon that are juvenile (primordial) vs. tectonically recycled. Here we present new results for submarine basalt glasses, erupted within the Garrett Transform Fault (GTF) and to its north and south along the East Pacific Rise (EPR). Analyses of 3He/4He, and He and CO2 concentrations in vesicles and glass were performed, through a series of crushing and melting experiments plus FTIR spectroscopy. Trace elements were analyzed by laser ablation ICP-MS. The GTF basalts provide further tests for the origin of volatile-undersaturated basalts using CO2/Ba and CO2/Nb systematics. CO2 is highly correlated with Ba and Nb in basalts erupted in the transform domain (n = 13, including 5 undersaturated basalts) and does not show the variability expected from mixing between undegassed and variably degassed melts. Rather, the melts appear to originate from a heterogeneous mantle source that was variably depleted through partial melting, and limited mixing of melts is involved in their generation. The CO2/Ba and CO2/Nb weight ratios of 106 ± 8 and 308 ± 27, respectively, are similar to values determined previously for a global suite of undersaturated mid-ocean ridge basalts (MORBs). The ridge and transform domains have distinct 3He/4He ratios. Along the nearby EPR, 3He/4He = 8.5 - 9.1 RA, while within the Garrett Transform Fault 3He/4He = 9.2 - 10.1 RA. These two basalt populations are also distinct in their Pb-Sr-Nd isotope compositions based on earlier regional studies. The distinct populations result from partial melting of two different mantle source compositions. Melting of depleted mantle containing a small amount (∼1 to 5%) of enriched, ancient heterogeneities occurs beneath the EPR. Melting of ultradepleted mantle (in which the heterogeneities have been removed by earlier melting beneath the EPR) occurs beneath the GTF. This explains the distinction between intra-transform and spreading ridge domains for 3He/4He, if the heterogeneities were enriched in U, Th and He and had low 3He/4He as would be found in tectonically recycled material. The enriched mantle component sampled by the EPR basalts has molar CO2/3He = 2 × 10 9 , and it dominates the CO2/3He ratio generally ascribed to the upper mantle source for mid-ocean ridge basalts. In contrast, the ultradepleted MORB mantle component sampled by the GTF basalts has CO2/3He = 3 × 10 8 or less. This indicates that most of the carbon in Earth's upper mantle originates from tectonic recycling.

We report total helium concentrations (vesicles + glass) for a suite of thirteen ultradepleted mid-ocean ridge basalts (UD-MORBs) that were previously studied for volatile contents (CO2, H2O) plus major and trace elements. The selected basalts are undersaturated in CO2 + H2O at their depths of eruption and represent rare cases of undegassed MORBs. Sample localities from the Atlantic (2), Indian (1) and Pacific (7) Oceans collectively show excellent linear correlations (r2=0.75–0.92r2=0.75–0.92) between the concentrations of helium and the highly incompatible elements C, K, Rb, Ba, Nb, Th and U. Three basalts from Gakkel Ridge in the Arctic were also studied but show anomalous behavior marked by excess lithophile trace element abundances. In the Atlantic–Pacific–Indian suite, incompatible element concentrations vary by factors of 3–4.3, while helium concentration varies by a factor of 13. The strong correlations between the concentrations of helium and incompatible elements are explained by helium behavior as the most incompatible element during mantle melting. Partial melting of an ultradepleted mantle source, formed as a residue of earlier melt extraction, accounts for the observed concentrations. The earlier melting event involved removal of a small degree melt (∼1%) at low but non-zero porosity (0.01–0.5%), leading to a small amount of melt retention that strongly leveraged the incompatible element budget of the ultradepleted mantle source. Equilibrium melting models that produce the range of trace element and helium concentrations from this source require a bulk solid/melt distribution coefficient for helium that is lower than that for other incompatible elements by about a factor of ten. Alternatively, the bulk solid/melt distribution coefficient for helium could be similar to or even larger than that for other incompatible elements, but the much larger diffusivity of helium in peridotite leads to its more effective incompatibility and efficient extraction from a larger volume of mantle during melting. In either case, partial melting leaves a mantle residue with elevated (U + Th)/3He. Consequently, peridotite residues of mantle melting cannot be the source of high 3He/4He observed at ocean island hotspots such as Hawaii and Iceland. The extreme effective incompatibility of helium entails that high 3He/4He mantle sources, isolated before 4.45 Ga based on Xe and W isotopes, have not experienced any melt extraction since they were isolated from convecting portions of the mantle.

Molybdenum (Mo) stable isotopes hold great potential to investigate the processes involved in planetary formation and differentiation. However their use is currently hampered by the lack of understanding of the dominant controls driving mass-dependent fractionations at high temperature. Here we investigate the role of magmatic processes and mantle source heterogeneities on the Mo isotope composition of Mid-Ocean Ridges Basalts (MORBs) using samples from two contrasting ridge segments: (1) the extremely fast spreading Pacific–Antarctic (66–41°S) section devoid of plume influence and; (2) the slow spreading Mohns–Knipovich segment (77–71°N) intercepted by the Jan Mayen Plume (71°N). We show that significant variations in Mo stable isotope composition exist in MORBs with δ98/95Mo ranging from − 0.24 ‰ to + 0.15 ‰ (relative to NIST SRM3134). The absence of correlation between δ98/95Mo and indices of magma differentiation or partial melting suggests a negligible impact of these processes on the isotopic variations observed. On the other hand, the δ98/95Mo variations seem to be associated with changes in radiogenic isotope signatures and rare earth element ratios (e.g., (La/Sm)N), suggesting mantle source heterogeneities as a dominant factor for the δ98/95Mo variations amongst MORBs. The heaviest Mo isotope compositions correspond to the most enriched signatures, suggesting that recycled crustal components are isotopically heavy compared to the uncontaminated depleted mantle. The uncontaminated depleted mantle shows slightly sub-chondritic δ98/95Mo, which cannot be produced by core formation and, therefore, more likely result from extensive anterior partial melting of the mantle. Consequently, the primitive δ98/95Mo composition of the depleted mantle appears overprinted by the effects of both partial melting and crustal recycling.

Eruption rates at the mid-ocean ridges (MORs) are believed to strongly control the morphology and length of lava flows emplaced along the ridge axis, and thus the structure and porosity of the upper oceanic crust. Eruption rate also represents one of the few tools to gain insight into the driving pressures within sub-ridge magmatic systems. As eruption rate is inferred to vary systematically along the global mid-ocean ridge system, understanding of how to assess eruption rate in submarine systems and how it maps to observable features of the ridge axis would provide a powerful tool to understand Earth's largest magmatic system. Eruption rates at MORs are poorly constrained, however, because of a lack of direct observations, preventing the duration of an eruption to be quantified. This study uses decompression experiments of MORB samples and numerical modeling of CO 2 degassing to reconstruct the timescales for magma ascent and lava emplacement during the 2005–06 eruption of the East Pacific Rise at ca. 9°51′N. Samples collected from the lava flow are all supersaturated in dissolved CO 2 contents, but CO 2 decreases with distance from the vent, presumably as a consequence of progressive CO 2 diffusion into growing bubbles. Samples collected at the vent contain ∼10 5 vesicles per cm 3 . Pieces of these samples were experimentally heated to 1225 °C at high pressure and then decompressed at controlled rates. Results, plus those from numerical modeling of diffusive bubble growth, indicate that magma rose from the axial magma chamber to the seafloor in ≤1 h and at a rate of ≥2–3 km h −1 . Our modeling, as validated by experimental decompression of MORB samples with ∼10 6 vesicles cm −3 , also suggests that CO 2 degassed from the melt within ∼10–100 min as the vesicular lava traveled ∼1.7 km along the seafloor, implying a volumetric flow rate on order of 10 3–4 m 3 s −1 . Given an ascent rate of ≥0.2 m s −1 , the width of a rectangular dike feeding the lava would have been ∼1–2 m wide. MORB samples from the Pacific ridge are generally more supersaturated in dissolved CO 2 than those from slower spreading Atlantic and Indian ridges. Our results suggest that Pacific MORBs ascend to the seafloor faster than Atlantic or Indian MORBs.

Variations in the volume and major element composition of basalt erupted along the global mid-ocean ridge system have been attributed to differences in mantle potential temperature, mantle composition, or plate spreading rate and lithosphere thickness. Abyssal peridotites, the residues of mantle melting beneath mid-ocean ridges, provide additional information on the melting process, which could be used to test these hypotheses. We compiled a global database of abyssal peridotite compositions averaged over the same ridge segments defined by Gale et al. (2013). In addition, we calculated the distance of each ridge segment to the nearest hotspots. We show that Cr# in spinel in abyssal peridotites is negatively correlated with Na90 in basalts from the same ridge segments on a global scale. Ridge segments that erupt basalts apparently produced by larger degrees of mantle melting are thus underlain by peridotites from which large amounts of melt have been extracted. We find that near-ridge hotspots have a more widespread influence on mid-ocean ridge basalt (MORB) composition and ridge depth than previously thought. However, when these hotspot-influenced ridge segments are excluded, the remaining segments show clear relationships between MORB composition, peridotite composition, and ridge depth with spreading rate. Very slow-spreading ridges (

Decades of exploration for venting sites along spreading ridge crests have produced global datasets that yield estimated mean site spacings of ∼ 12 – 220 km . This conclusion demands that sites where hydrothermal fluid leaks from the seafloor are improbably rare along the 66 000 km global ridge system, despite the high bulk permeability of ridge crest axes. However, to date, exploration methods have neither reliably detected plumes from isolated low-temperature, particle-poor, diffuse sources, nor differentiated individual, closely spaced (clustered within a few kilometers) sites of any kind. Here we describe a much lower mean discharge spacing of 3–20 km, revealed by towing real-time oxidation–reduction–potential and optical sensors continuously along four fast- and intermediate-rate (>55 mm/yr) spreading ridge sections totaling 1470 km length. This closer spacing reflects both discovery of isolated sites discharging particle-poor plumes (25% of all sites) and improved discrimination (at a spatial resolution of ∼1 km) among clustered discrete and diffuse sources. Consequently, the number of active vent sites on fast- and intermediate-rate spreading ridges may be at least a factor of 3–6 higher than now presumed. This increase provides new quantitative constraints for models of seafloor processes such as dispersal of fauna among seafloor and crustal chemosynthetic habitats, biogeochemical impacts of diffuse venting, and spatial patterns of hydrothermal discharge.

The previous orthogonal ridge/transform staircase geometry south of Iceland is being progressively changed to the present continuous oblique Reykjanes Ridge spreading geometry as North America–Eurasia transform faults are successively eliminated from north to south. This reorganization is commonly interpreted as a thermal phenomenon, caused by warmer Iceland plume mantle progressively interacting with the ridge, although other diachronous seafloor spreading reorganizations are thought to result from tectonic rift propagation. New marine geophysical data covering our reinterpretation of the reorganization tip near 57°N show successive transform eliminations at a propagation velocity of ∼110 km/Myr, ten times the spreading half rate, followed by abrupt reorganization slowing at the Modred transform as it was converted to a migrating non-transform offset. Neither the simple thermal model nor the simple propagating rift model appears adequate to explain the complicated plate boundary reorganization process.

Highlights • Magmatism, detachment faulting and changing symmetry of crustal accretion at a ridge segment end. • PmPs from an OCC argue for constant magma production during detachment faulting. • Refraction seismic modelling and PmP events reveal very thin (4 km) oceanic crust at a segment end. A wide-angle seismic section across the Mid-Atlantic Ridge just south of the Ascension transform system reveals laterally varying crustal thickness, and to the east a strongly distorted Moho that appears to result from slip along a large-offset normal fault, termed an oceanic detachment fault. Gravity modelling supports the inferred crustal structure. We investigate the interplay between magmatism, detachment faulting and the changing asymmetry of crustal accretion, and consider several possible scenarios. The one that appears most likely is remarkably simple: an episode of detachment faulting which accommodates all plate divergence and results in the westward migration of the ridge axis, is interspersed with dominantly magmatic and moderately asymmetric (most on the western side) spreading which moves the spreading axis back towards the east. Following the runaway weakening of a normal fault and its development into an oceanic detachment fault, magma both intrudes the footwall to the fault, producing a layer of gabbro (subsequently partially exhumed).

The degassing of primordial gases from Earth's interior is evidenced by the high He 3 / 4 He ratios in submarine hydrothermal plumes, vent fluids, and rock samples with mantle origin from active hydrothermal systems in mid-ocean ridges (MOR) and subduction zone volcanism. As the largest aquifer on Earth, the uppermost 40-500 m of permeable submarine ridge flank basement (1-65 million-years-old, Myr) holds ∼2% of the ocean volume and accounts for 70% of the seafloor hydrothermal heat flux. However, the degassing of primordial gases through the oceanic ridge flank crust has not yet been directly quantified. Here, we show that high integrity hydrothermal (65 °C) fluids from the sediment-buried 3.5 Myr basaltic crust from the eastern flank of the Juan de Fuca Ridge (JdFR) contain elevated 3He. The He 3 / 4 He for the basaltic fluid is 4.5 ± 0.1 R a (relative to the air ratio), which is greatly elevated when compared to deep seawater (1.05 R a ), but is half of that observed for high-temperature vent fluids (∼8 R a ) emitting from MOR. Only a small fraction of the 3He in ridge flank fluids is derived from the entrainment of high-temperature ridge-axis fluids and is better explained by degassing of the mantle through the mantle-crust boundary. The lower than MOR He 3 / 4 He ratios indicate that radiogenic 4He originates from aged uranium and thorium decay within the mantle as well as from the ridge-flank basalts. The 3He outgassing through warm ridge flanks (4.9 to 36 mol/yr) accounts for 0.7-6% of the global 3He outgassing, exceeded only by degassing through mid-ocean ridges and subduction volcanism. The presence of mantle 3He suggests that the abiogenic methane present in the ridge flank fluids might be mantle-derived. Based on the 3He outgassing flux, a possibly mantle-derived abiotic methane production rate at the ridge flank is estimated to be 0.3 – 35 × 10 8 mol / yr .

Partial melting of asthenospheric mantle generates magma that supplies volcanic systems. The timescale of melt extraction from the mantle has been hotly debated. Microstructural measurements of permeability typically suggest relatively slow melt extraction (1 m/yr) whereas geochemical (Uranium-decay series) and geophysical observations suggest much faster melt extraction (100 m/yr). The deglaciation of Iceland triggered additional mantle melting and magma flux at the surface. The rapid response has been used to argue for relatively rapid melt extraction. However, this episode must, at least to some extent, be unrepresentative, because the rates of magma eruption at the surface increased about thirty-fold relative to the steady state. Our goal is to quantify this unrepresentativeness. We develop a one-dimensional, time-dependent and nonlinear (far from steady-state), model forced by the most recent, and best mapped, Icelandic deglaciation. We find that 30 m/yr is the best estimate of the steady-state maximum melt velocity. This is a factor of about 3 smaller than previously claimed, but still relatively fast. We translate these estimates to other mid-ocean ridges accounting for differences in passive and active upwelling and degree of melting. We find that fast melt extraction greater than about 10 m/yr prevails globally., Accepted for publication in Earth and Planetary Science Letters (2020). Total 17 pages and 5 figures (including Supplementary Material)

Recent studies suggest that eustatic sea level fluctuations induced by glacial cycles in the Pleistocene influence mantle-melting and volcanic eruptions at mid-ocean ridges, with models predicting variations in oceanic crustal thickness and seafloor bathymetry linked to sea level change. Analyses of seafloor bathymetry have found evidence of significant spectral energy at frequencies consistent with Milankovic cycles of 1/23, 1/41, and 1/80-1/120 ka−1. However, other studies emphasize the need for crustal thickness observations to test the “sea level hypothesis”. Here we investigate the hypothesis of climate driven periodicity in mid-ocean ridge magmatism through analysis of a unique bathymetry and crustal thickness dataset derived from a 3D multi-channel seismic investigation of the East Pacific Rise from 9°42' to 57′N. Crustal thickness data spans the last ∼235 ka in age and reveals three axis-parallel zones of 200-800 m thicker crust. The amplitude and spacing of these thick crust ridges, which are most prominent on the east flank of the ridge, are consistent with predictions of sea level modulated mantle melting. Similarly spaced ridges are apparent in the longer duration (470 ka) seafloor bathymetry data. Spectral analysis of these datasets shows peaks centered near 1/80 ka−1 and locally near 1/41 ka−1 on the east flank in both bathymetry and crustal thickness. West flank spectral results show intermittent peaks near 1/100 ka−1 and 1/41 ka−1 in crustal thickness and no coherent peak frequencies in bathymetry data. We attribute differences between the east and west flank to the impacts of spatially variable asymmetric spreading and small changes in the locus of accretion. Observed half spreading rates are dominantly faster to the east with small ridge jumps transferring crust from the west flank. Lagged cross-correlations between sea level and crustal thickness indicate a maximum when the latter is lagged by ∼45 ka, which align thick crust zones with the ∼100 ka periods of lower sea level. Crustal thickness is also directly compared with seafloor bathymetry, indicating a component of compensated topography with RMS relief at the seafloor of 10 to 29% of crustal thickness variations. However, complexity inherited from variable asymmetric spreading and seafloor faulting is also apparent, and results provide new insights into how the crustal accretion filter modulates the recording of magma supply variations in the crust and in seafloor relief. While the significance of the statistical analysis of these ridge records is limited by the short duration of the available crustal thickness dataset and effects of asymmetric spreading, the novel observations of crustal thickness varying at timescales of ∼80-100 ka require a mechanism, and the sea level hypothesis provides a plausible explanation.

Slow-spreading mid-ocean ridges develop large offset normal faults, or detachments, that exhume mantle-derived peridotites to the seafloor. Successive detachments that flip polarity (flip-flop detachments) have been documented in corridors of nearly amagmatic spreading at the eastern Southwest Indian Ridge (SWIR). The depth of earthquake hypocenters supports the existence of a very thick axial lithosphere with a brittle lid thicker than 20 km. This is also consistent with our petrological observations on dredged peridotites from the area (plagioclase-free, spinel-bearing dynamically recrystallized assemblages in shear zones) that reveal high stress (80–270 MPa) and high temperature (>800 °C) deformation combining brittle and ductile mechanisms. Thermo-mechanical models using strain dependent cohesion weakening to localize deformation do not produce detachments for a brittle lithosphere thicker than 20 km. Here, we explore two other weakening mechanisms observed in situ, and for which we have temperature constraints from rock sampling: serpentinization (at T 350 °C) and grain size reduction by dynamic recrystallization (at T∼800–1000 °C). Serpentinization and more generally hydrous alteration minerals have been shown to localize strain at detachment-dominated slow-spreading ridge locations. Here we show that grain size reduction due to dynamic recrystallization also occurs in peridotites from the eastern SWIR. Our models explore the effect of implementing these weakening mechanisms separately or together in an amagmatic spreading context, of varying the threshold conditions for their activation, and the thermal conditions on-axis. We show that serpentinization alone does not produce flip-flop detachments, while grain size reduction alone does, albeit with unrealistically high associated topographic relief. Combining the two, we observe the activation of several modes of axial deformation: horst mode, spider mode, and long or short-lived flip-flop detachments. Flip-flop detachment faulting, once initiated, can be activated over long periods of time, with fault topography and offsets that are consistent with geological observations in the eastern SWIR nearly amagmatic study area. The transition between these modes appears to depend on the amount and location of serpentine that has been produced in previous stages of faulting.

Volcanic seamounts are spectacular bathymetric features representing ‘undersea mountains’ comparable in size to volcanic islands. They are often sites of active volcanism, represent major marine habitats, and contain potentially globally significant commodities of trace metals. However, we know comparatively little about how they form, the extent of their volcanic lifecycle, and what risks they pose from periodic flank collapse. Here, we study a uniquely extensive 43 Myr record of volcaniclastic and calciclastic turbidites accumulated on the Madeira Abyssal Plain in the NE Atlantic, which originate from large submarine landslides from the adjacent Great Meteor-Cruiser seamount complex on the flanks of the Mid-Atlantic Ridge. These turbidites present the first long-term temporal record of the recurrence, magnitude, and emplacement mechanism of seamount flank collapses from a single seamount complex throughout its lifecycle. Landslides from the Great Meteor-Cruiser seamount complex are comparable size to those from the Canary Islands, with average volumes of 30-40 km3 but can be as large as 130 km3, comparable to many volcanic island landslides. These seamount landslides were predominantly multi-stage failures similar to those form neighbouring volcanic islands, but in contrast occurred preferentially during sea-level lowstands. We show that volcanic seamounts above mantle plumes have long protracted lifecycles and that substantial flank failures are capable of occurring at all stages of development. Temporal change in turbidite composition has, for the first time, enabled reconstruction of the inception (around 43-45 Myrs ago), ascension (taking around 20-25 Myrs), emergence (after 30 Myrs) and later recession (last 9 Myrs), cessation in volcanism (last 4 Myrs) and subsidence of the seamounts (from 2 Myrs onwards). These seamounts above a mantle plume have taken around 26 Myrs to emergence above sea level from inception, while their volcanic island contemporaries only took 2-to-5 Myrs. Here, we use the history of turbidites derived from landslides from the seamounts to show how the seamounts developed and suffered mass wasting during their life cycles.

A plethora of observations have been made at mid-ocean ridges and mantle sections of ophiolites: presence of tabular replacive dunites, highly depleted LREE in residual abyssal peridotites, U-series disequilibria in fresh basalts, crustal thickness, and highly attenuated seismic and magnetotelluric structures beneath spreading centers. These independent observations can become more powerful if they are jointly analyzed in a self-consistent model. The difficulty for such a model is to resolve and evaluate the effect of fine scale petrologic feature such as dunite channels in a tectonic scale geodynamic model. Here we present a two-dimensional double-porosity ridge model that consists of low-porosity residual lherzolite and harzburgite matrix and high-porosity interconnected channel network. The geodynamic simulation features a spatially distributed channel network and anisotropic permeability that gradually develops as the upwelling mantle is deformed. We compute the porosity, melt and solid flow fields for several choices of channel distribution and permeability model. We use the calculated porosity distribution and velocity fields to model the variations of REE in residual mantle and U-series disequilibria in melts. The present study underscores the importance of deep channel networks and permeability model to the interpretation of first order geophysical and geochemical observations at mid-ocean ridge spreading centers. The anisotropic permeability of channels can enhance melt focusing by 60%, resulting in thicker crust. The attenuated seismic and magnetotelluric structures require a channel network starting from 60 km depth beneath the ridge axis. The depleted REE patterns in clinopyroxenes in residual abyssal peridotites are also consistent with melt extraction into channels starting from 60 km depth. Although deep channels are conducive to producing U-series disequilibria in eruptible melts, the present model still cannot explain the full ranges of the observed U series disequilibria data in MORB samples. Additional factors, processes, and models are discussed. And finally, we found that, within the uncertainty of the permeability, the porosity varies by three folds and the excess of 230Th varies by up to two folds.

Magmatism and volcanism transfer carbon from the solid Earth into the climate system. This transfer may be modulated by the glacial/interglacial cycling of water between oceans and continental ice sheets, which alters the surface loading of the solid Earth. The consequent volcanic-carbon fluctuations have been proposed as a pacing mechanism for Pleistocene glacial cycles. This mechanism is dependant on the amplitude and lag of the mid-ocean ridge response to sea-level changes. Here we develop and analyse a new model for that response, eliminating some questionable assumptions made in previous work. Our model calculates the carbon flux, accounting for the thermodynamic effect of mantle carbon: reduction of the solidus temperature and a deeper onset of melting. We analyse models forced by idealised, periodic sea level and conclude that fluctuations in melting rate are the prime control on magma and carbon flux. We also discuss a model forced by a reconstruction of eustatic sea level over the past 800 kyr. It indicates that peak-to-trough variations of magma and carbon flux are up to about 20% and 10% of the mean flux, respectively. Peaks in mid-ocean ridge emissions lag peaks in sea-level forcing by less than about 20 kyr and the lag could well be shorter. The amplitude and lag are sensitive to the rate of melt segregation. The lag is much shorter than the time it takes for melt to travel vertically across the melting region., 28 pages including supplementary material. 9 figures plus 6 supplementary figures. Accepted for publication in Earth and Planetary Science Letters

Oceanic crust formed along spreading centers in the Lau back-arc basin exhibits a dramatic change in structure and composition with proximity to the nearby Tofua Arc. Results from recent seismic studies in the basin indicate that crust formed near the Tofua Arc is abnormally thick (8–9 km) and compositionally stratified, with a thick low-velocity (3.4–4.5 km/s) upper crust and an abnormally high-velocity (7.2–7.4+ km/s) lower crust ( Arai and Dunn, 2014 ). Lava samples from this area show arc-like compositional enrichments and tend to be more vesicular and differentiated than typical mid-ocean ridge basalts, with an average MgO of ∼3.8 wt.%. We propose that slab-derived water entrained in the near-arc ridge system not only enhances mantle melting, as commonly proposed to explain high crustal production in back-arc environments, but also affects magmatic differentiation and crustal accretion processes. We present a petrologic model of Lau back-arc crustal formation that successfully predicts the unusual crustal stratification imaged in the near-arc regions of the Lau basin, as well as the highly fractionated basaltic andesites and andesites that erupt there. Results from phase equilibria modeling using MELTS indicate that the high water contents found in near-arc parental melts can lead to crystallization of an unusually mafic, high velocity cumulate layer. Best-fit model runs contain initial water contents of ∼0.5–1.0 wt.% H2O in the parental melts, and successfully reproduce geochemical trends of the erupted lavas while crystallizing a cumulate assemblage with calculated seismic velocities consistent with those observed in the near-arc lower crust. Modeled residual melts are also lower density than their dry equivalents, which aids in melt segregation from the cumulate layer. Low-density, water-rich residual melts can lead to the eruption of vesicular lavas that are unusually evolved for an oceanic spreading center.

We describe a variety of one- and two-dimensional mathematical modeling approaches to characterizing diffuse flow circulation at mid-ocean ridge hydrothermal systems. The goal is to estimate the potential extent of the sub-seafloor microbial biosphere based on subsurface contours of the 120 °C isotherm as determined from the various models. The models suggest that the sub-seafloor depth for microbial life may range from less than 1 m in some places to the thickness of crustal layer 2A of ∼ 500 m in others. This depth depends primarily on how diffuse flow is driven. The 120 °C isotherm tends to be much deeper if diffuse flow is induced as boundary layer flow near high-temperature plumes, than if it results from conductive cooling or mixing near the seafloor. Because the heat flow alone may not allow identification of the flow regime in the subsurface, we highlight the use of chemical tracers as an additional constraint that sheds light into the flow and reaction patterns associated with vents. We use thermodynamic modeling, which connects the temperature of the diffuse fluid to its chemical composition. As the temperature-composition relationships differ for mixing versus conductive heating and cooling, the fluid geochemistry can shed light on subsurface transport. Using methane as an example, the geochemical models indicate subsurface microbial methane production and consumption in different regions of the vent field near EPR 9 ° 50 ′ N .

A new geospeedometer, based on diffusion modeling of Mg in plagioclase, is used to determine cooling rates of the upper section of the lower oceanic crust formed at fast-spreading mid-ocean ridges. The investigated natural sample suites include gabbroic rocks formed at three different locations along the fast-spreading East Pacific Rise. These samples cover a depth interval of 0–840 m below the sheeted dike/gabbro boundary and therefore allow the variation of cooling rate as a function of depth within the upper plutonic sequence to be determined. We demonstrate that the cooling rates we obtained are robust (reproducible and consistent across different vertical sections at fast spreading ridges) and decrease significantly with increasing sample depth (covering almost 4 orders of magnitude, ranging from ∼1 °C y −1 for the shallowest samples to 0.0003 °C y −1 for the deepest samples). Both the absolute cooling rates, and the rate of decrease of cooling rate with depth, are consistent with conductive thermal models. In contrast, the absolute cooling rates determined from the deeper samples (>300 m below DGB), and the large decrease in cooling rate with depth are inconsistent with thermal models that include substantial cooling by off-axis hydrothermal circulation within the upper plutonic section of the crust.

Chemical gradients in the subsurface of mid-ocean ridge hydrothermal systems create an environment where minerals precipitate and dissolve and where chemosynthetic organisms thrive. However, owing to the lack of easy access to the subsurface, robust knowledge of the nature and extent of chemical transformations remains elusive. Here, we combine measurements of vent fluid chemistry with geochemical and transport modeling to give new insights into the under-sampled subsurface. Temperature–composition relationships from a geochemical mixing model are superimposed on the subsurface temperature distribution determined using a heat flow model to estimate the spatial distribution of fluid composition. We then estimate the distribution of Gibb's free energies of reaction beneath mid oceanic ridges and by combining flow simulations with speciation calculations estimate anhydrite deposition rates. Applied to vent endmembers observed at the fast spreading ridge at the East Pacific Rise, our results suggest that sealing times due to anhydrite formation are longer than the typical time between tectonic and magmatic events. The chemical composition of the neighboring low temperature flow indicates relatively uniform energetically favorable conditions for commonly inferred microbial processes such as methanogenesis, sulfate reduction and numerous oxidation reactions, suggesting that factors other than energy availability may control subsurface microbial biomass distribution. Thus, these model simulations complement fluid-sample datasets from surface venting and help infer the chemical distribution and transformations in subsurface flow.

Multi-channel seismic (MCS) reflection profiles across the Pacific Plate south of the Alaska Peninsula reveal the internal structure of mature oceanic crust (48–56 Ma) formed at fast to intermediate spreading rates during and after a major plate re-organization. Oceanic crust formed at fast spreading rates (half spreading rate ∼ 74 mm / yr ) has smoother basement topography, thinner sediment cover with less faulting, and an igneous section that is at least 1 km thicker than crust formed at intermediate spreading rates (half spreading rate ∼ 28 – 34 mm / yr ). MCS data across fast-spreading oceanic crust formed during plate re-organization contain abundant bright reflections, mostly confined to the lower crust above a highly reflective Moho transition zone, which has a reflection coefficient (RC) of ∼0.1. The lower crustal events dip predominantly toward the paleo-ridge axis at ∼10–30°. Reflections are also imaged in the uppermost mantle, which primarily dip away from the ridge at ∼10–25°, the opposite direction to those observed in the lower crust. Dipping events in both the lower crust and upper mantle are absent on profiles acquired across the oceanic crust formed at intermediate spreading rates emplaced after plate re-organization, where a Moho reflection is weak or absent. Our preferred interpretation is that the imaged lower crustal dipping reflections within the fast spread crust arise from shear zones that form near the spreading center in the region characterized by interstitial melt. The abundance and reflection amplitude strength of these events ( RC ∼ 0.15 ) can be explained by a combination of solidified melt that was segregated within the shear structures, mylonitization of the shear zones, and crystal alignment, all of which can result in anisotropy and constructive signal interference. Formation of shear zones with this geometry requires differential motion between the crust and upper mantle, where the upper mantle moves away from the ridge faster than the crust. Active asthenospheric upwelling is one possible explanation for these conditions. The other possible interpretation is that lower crustal reflections are caused by magmatic (mafic/ultramafic) layering associated with accretion from a central mid-crustal magma chamber. Considering that the lower crustal dipping events have only been imaged in regions that have experienced plate re-organizations associated with ridge jumps or rift propagation, we speculate that locally enhanced mantle flow associated with these settings may lead to differential motion between the crust and the uppermost mantle, and therefore to shearing in the ductile lower crust or, alternatively, that plate reorganization could produce magmatic pulses which may lead to mafic/ultramafic banding.

ODP Hole 1256D in the eastern equatorial Pacific is the first penetration of a complete section of fast spread ocean crust down to the dike–gabbro transition, and only the second borehole to sample in situ sheeted dikes after DSDP Hole 504B. Here a high spatial resolution record of whole rock and mineral strontium isotopic compositions from Site 1256 is combined with core observations and downhole wireline geophysical measurements to determine the extent of basalt–hydrothermal fluid reaction and to identify fluid pathways at different levels in the upper ocean crust. The volcanic sequence at Site 1256 is dominated by sheet and massive lava flows but the Sr isotope profile shows only limited exchange with seawater. However, the upper margins of two anomalously thick (>25 m) massive flow sequences are strongly hydrothermally altered with elevated Sr isotope ratios and appear to be conduits of lateral low-temperature off-axis fluid flow. Elsewhere in the lavas, high 87 Sr/ 86 Sr are restricted to breccia horizons. Mineralised hyaloclastic breccias in the Lava–Dike Transition are strongly altered to Mg-saponite, silica and pyrite, indicating alteration by mixed seawater and cooled hydrothermal fluids. In the Sheeted Dike Complex 87 Sr/ 86 Sr ratios are pervasively shifted towards hydrothermal fluid values (∼0.705). Dike chilled margins display secondary mineral assemblages formed during both axial recharge and discharge and have higher 87 Sr/ 86 Sr than dike cores, indicating preferential fluid flow along dike margins. Localised increases in 87 Sr/ 86 Sr in the Dike–Gabbro Transition indicates the channelling of fluids along the sub-horizontal intrusive boundaries of the 25 to 50 m-thick gabbroic intrusions, with only minor increases in 87 Sr/ 86 Sr within the cores of the gabbro bodies. When compared to the pillow lava-dominated section from Hole 504B, the Sr isotope measurements from Site 1256 suggest that the extent of hydrothermal circulation in the upper ocean crust may be strongly dependent on the eruption style. Sheet and massive flow dominated lava sequences typical of fast spreading ridges may experience relatively restricted circulation, but there may be much more widespread circulation through pillow lava-dominated sections. In addition, the Hole 1256D sheeted dikes display a much greater extent of Sr-isotopic exchange compared to dikes from Hole 504B. Because seawater-derived hydrothermal fluids must transit the dikes during their evolution to black smoker-type fluids, the different Sr-isotope profiles for Holes 504B and 1256D suggest there are significant variations in mid-ocean ridge hydrothermal systems at fast and intermediate spreading ridges, which may impact geochemical cycles of elements mobilised by fluid–rock exchange at different temperatures.

The West Iberia continental margin is a type locale for magma-poor rifting, and studies there have been instrumental in changing the classical view of the ocean–continent transition (OCT) from a discrete boundary juxtaposing continental and oceanic crust, into a more complicated zone of varying width that can include exhumed mantle. This study examines two new seismic lines in the Galicia Bank area extending west of the Peridotite Ridge, showing high resolution images of five new ridges. These ridges could be hyperextended continental crust, exhumed continental mantle, or rough ultra-slow spreading oceanic crust. There are no tilted fault blocks with pre-syn rift stratigraphy that would indicate continental crust. There are also no faults indicating mid-ocean spreading with seismic layer stratigraphy indicating normal oceanic crust. The ridges have no coherent internal seismic structure, and some resemble the topographic profile of the Peridotite Ridge. Therefore, it is likely the western ridges are also mainly composed of serpentinized mantle. These western ridges are also similar to small oceanic core complexes observed along the active part of the Mid-Atlantic Ridge, which also contain exhumed serpentinized mantle. This implies that there is a gradual transition within our study area from continental extension to seafloor spreading. Exhumation of continental mantle results in the formation of peridotite ridges, then transitions to episodic volcanism, which produces local thin basaltic crust, and exhumation of oceanic core complexes. Asymmetric processes during initial rifting and spreading result in contrasting structures on the two resulting margins.

Shallow, seismically imaged melt lenses are a ubiquitous feature of mid-ocean ridges with high magma supply; melt lenses deepen and become less continuous along axis as the rate of magma supply decreases. Despite compelling petrologic evidence for evolution of magma within the crust prior to eruption at lower magma supply, melt lenses are rarely detected along ridge segments with rates of magma supply less than 0.3 × 10 6 m 3 / yr / km , and the depths of sub-axial magma reservoirs are therefore poorly known. We use ion microprobe measurements of H2O and CO2 concentrations of olivine-hosted melt inclusions to calculate vapor saturation pressures that constrain crystallization depths at two locations along the Galapagos Spreading Center (94.2°W and 95°W). These sites were chosen to examine crystallization pressures in the presence (94.2°W) and absence (95°W) of a seismically imaged melt lens. At 95°W, where magma supply is too low to sustain a seismically resolvable melt lens, samples were selected from each of the three most recent eruptive units, allowing us to document temporal variations in vapor saturation pressures and the depth of magma residence at this location. Clusters in melt inclusion entrapment depths for these eruptions range from 3.0 to 3.4 km below the seafloor, indicating that magmas at 95°W resided at a narrow range of mid-crustal depths prior to eruption, generally consistent with the global trend of increasing melt lens depth with decreasing rate of magma supply. A discrepancy between seismic data and the peak in melt inclusion entrapment depths at 94.2°W may reflect temporal variability of magmatic systems at this location. This study demonstrates the potential for using measurements of the concentrations of H2O and CO2 in olivine-hosted melt inclusions to determine the depths of crustal magmatic systems that feed mid-ocean ridge eruptions, even in locations where seismic studies have not detected melt lenses.

The Tristan-Gough hotspot trail on the African plate consists of the Walvis Ridge and a younger province of seamounts and islands. In order to determine the relative motion between the African plate and the Tristan-Gough hotspot it is essential to resolve changes in the age and morphology of the Walvis Ridge. A significant problem is, however, to establish how the vigor and flow of hotspot material to the mid-ocean ridge constructed the Walvis Ridge. We have addressed this issue by measuring an 40Ar/39Ar stratigraphy at three sites across the central Walvis Ridge sampled by Ocean Drilling (DSDP Leg 74). The age-distance relation of volcanism, together with geophysical, geochemical and paleodepth information, suggests collectively that hotspot volcanism was occurring locally c. 72 Ma on an elevated segment of the mid-ocean ridge located close to the Tristan-Gough hotspot. As the mid-ocean ridge migrated away from the hotspot (c. 36 km/Ma) between c. 72 Ma and 68 Ma, hotspot material continued flowing to the mid-ocean ridge and the Walvis Ridge shoaled rapidly (c. 500 m/Ma) to the west, on seafloor that might have been subsiding at a rate consistent with normal crustal cooling. This apparent correlation points to the possibility of an inverse relation between the volume flux of hotspot volcanism and the distance between the mid-ocean ridge and the Tristan-Gough hotspot. We infer that since c. 93 Ma the geometry and motion of the mid-ocean ridge determined where the hotspot material that built the Walvis Ridge was channeled to the plate surface. Furthermore, interplay between hotspot flow, and the changing geometry of the mid-ocean ridge as it migrated relative to the Tristan-Gough hotspot, might explain the age and morphology of the Walvis Ridge. Our finding provides further evidence that the distribution of hotspot volcanism in the southeast Atlantic expresses interaction between deep mantle (plume) and shallow plate tectonic and asthenosphere processes.