Search

Modern oceans contain large bathymetric highs (spreading oceanic ridges, aseismic ridges or oceanic plateaus and inactive arc ridges) that, in total, constitute more than 20–30% of the total area of the world’s ocean floor. These bathymetric highs may be subducted, and such processes are commonly referred to as ridge subduction. Such ridge subduction events are not only very common and important geodynamic processes in modern oceanic plate tectonics, they also play an important role in the generation of arc magmatism, material recycling, the growth and evolution of continental crust, the deformation and modification of the overlying plates, and metallogenesis at convergent plate boundaries. Therefore, these events have attracted widespread attention. The perpendicular or high-angle subduction of mid-ocean spreading ridges is commonly characterized by the occurrence of a slab window, and the formation of a distinctive adakite–high-Mg andesite–Nb-enriched basalt-oceanic island basalt (OIB) or a mid-oceanic ridge basalt (MORB)-type rock suite, and is closely associated with Au mineralization. Aseismic ridges or oceanic plateaus are traditionally considered to be difficult to subduct, to typically collide with arcs or continents or to induce flat subduction (low angle of less than 10°) due to the thickness of their underlying normal oceanic crust (> 6–7 km) and high topography. However, the subduction of aseismic ridges and oceanic plateaus occurred on both the western and eastern sides of the Pacific Ocean during the Cenozoic. On the eastern side of the Pacific Ocean, aseismic ridges or oceanic plateaus are being subducted flatly or at low angles beneath South and Central American continents, which may cause a magmatic gap. But slab melting can occur and adakites, or an adakite–high-Mg andesite–adakitic andesite–Nb-enriched basalt suite may be formed during the slab rollback or tearing. Cu-Au mineralization is commonly associated with such flat subduction events. On the western side of the Pacific Ocean, however, aseismic ridges and oceanic plateaus are subducted at relatively high angles (>30°). These subduction processes can generate large scale eruptions of basalts, basaltic andesites and andesites, which may be derived from fractional crystallization of magmas originating from the subduction zone fluid-metasomatized mantle wedge. In addition, some inactive arc ridges are subducted beneath Southwest Japan, and these subduction processes are commonly associated with the production of basalts, high-Mg andesites and adakites and Au mineralization. Besides magmatism and Cu-Au mineralization, ridge subduction may also trigger subduction erosion in subduction zones. Future frontiers of research will include characterizing the spatial and temporal patterns of ridge subduction events, clarifying the associated geodynamic mechanisms, quantifying subduction zone material recycling, establishing the associated deep crustal and mantle events that generate or influence magmatism and Cu-Au mineralization, establishing criteria to recognize pre-Cenozoic ridge subduction, the onset of modern-style plate tectonics and the growth mechanisms for Archean continental crust.

Flat subduction refers to low-angle (

The Brazilian continental margin includes several volcanic islands, submerged volcanic seamounts, and a unique non-volcanic archipelago located in a transform segment of the Equatorial South Atlantic. The mechanism of formation of these islands is related to post-breakup magmatic episodes dated as Late Cretaceous to Pleistocene. Diverse Late Cretaceous to Paleogene alkaline magmatic episodes are registered in southeast Brazil, resulting in igneous plugs onshore and volcanic structures offshore. The Abrolhos Volcanic Complex in eastern Brazil is characterized by several volcanic features on the continental shelf, including small islands that expose Paleogene sedimentary layers interbedded with volcanic sequences. The adjacent Vitoria-Trindade Chain extends to oceanic crust forming basaltic to alkaline seamounts that outcrop at the Trindade Archipelago, the easternmost islands in Brazil with the youngest volcanic eruptions. The Fernando de Noronha lineament in northeast Brazil is characterized by Neogene alkaline igneous plugs. The small islets in the Sao Pedro—Sao Paulo archipelago, located near the mid-Atlantic ridge, are formed by exhumed mantle rocks related to compressional episodes a transform fault zone. The Rio Grande Rise in southern Brazil is characterized by shallow Paleogene seamounts and a large oceanic plateau probably related to subaerial spreading centers formed in the Late Cretaceous. Multiple mechanisms are responsible for the origin and evolution of the volcanic islands offshore Brazil in continental, transitional, and oceanic crust settings, including volcanic build-ups, leaking fracture zones, and hotspots. Some of the islands might be related to mantle plume activity, as indicated by comparisons with modern mantle plume analogues in the South Atlantic.

Oceanic plateaus are a significant type of large igneous provinces in the oceans, providing insights to regional tectonic events and mantle behavior. The three world’s largest oceanic plateaus, the Ontong Java Plateau, the Kerguelen Plateau and the Shatsky Rise, are representatives in displaying extraordinary fluxes of magma from mantle to lithosphere. Detailed description incorporating transdisciplinary observations on marine geology, geophysics and geochemistry allow us to test the two lively-debated oceanic plateau formation hypotheses (mantle plume and plate boundary models). Predictions from either hypothesis merely obtain partial support. It is therefore unclear to differentiate one model from another one regarding the oceanic plateau formation. Careful comparisons of the three oceanic plateaus show many commonalities and even more differences in their formation and evolution. This diversity signifies one may not be typical of all. Notably, several key common features, i.e., massive and rapid eruption and near-ridge formation setting, imply that the lithospheric volcanic emplacement of oceanic plateaus was controlled by seafloor spreading despite a mantle plume exists peripherally. If a coincidence of mantle plume and spreading ridge occurs, it may indicate a plume-ridge interaction. One possible mechanism is that spreading ridge is dragged by a plume and migrates to the location of the plume. Another possibility is that the asthenosphere is fed by a plume nearby and generates melting anomalies along the spreading ridge.

Oceanic plateaus are large igneous provinces in the oceans, created by massive underwater eruptions, but their late-stage volcanism is poorly understood. With the addition of recent high-quality bathymetric data to existing data, 286 secondary cones were discovered over Shatsky Rise oceanic plateau. These cones with steeper flank slopes (mean 6.1° ± 4.4°) and smaller sizes (102–1923 m in height) are morphologically distinct from the plateau, and they are thought to have formed after the main volcanic episodes. Cone height and characteristic height (420 m) are close to seamounts in the Pacific Ocean, whereas greater than those in the Atlantic Ocean. Mean flatness of Shatsky Rise’s cones (0.25 ± 0.20) are similar to that of seamounts in both Pacific and Atlantic Oceans, but notably density in cone distribution (0.56 km−3) and their mean slope are significantly lower than those of seamounts in the two oceans. Lower slopes of secondary cones within Shatsky Rise may be explained by higher effusion rates of remaining magma. Although cone formation was expected to have a link to rifting by seafloor spreading, weak relationship between cone orientation and magnetic anomaly pattern implies that the expectation is negative. Moreover, weak correlation between the cone height and depth indicates it is not true that volcanic cones grow taller when they occur closer to the massif summits with thicker oceanic crust, which was suggested as the increase in hydraulic pressure. Cone height and flatness are also not strongly related, implying that remaining magma supply was too limited to foster the cones to critical height.

Tamu Massif is an immense Mesozoic submarine volcano, the main edifice of the Shatsky Rise oceanic plateau. It is located at a spreading ridge triple junction, but considered to be a shield volcano formed by effusive volcanism from an emerging mantle plume. However, it is unclear how Tamu Massif eruptions interacted with the spreading ridges, which are enormous linear volcanoes themselves. Here we create a magnetic anomaly map for Tamu Massif, which can provide clues about crustal formation. For Tamu Massif, we find dominantly linear magnetic field anomalies caused by crustal blocks of opposite magnetic polarity. This pattern suggests that Tamu Massif is not a shield volcano, but was emplaced by voluminous, focused ridge volcanism. If the magma source at the Shatsky Rise was a plume, it was closely connected to and controlled by seafloor spreading. By implication, even the largest oceanic plateau edifices can be formed by seafloor spreading. We suggest that the widely accepted analogy between continental flood basalts and oceanic plateaus requires reconsideration. The Shatsky Rise oceanic plateau formed by spreading ridge volcanism, according to analyses of linear magnetic anomalies over the Tamu Massif submarine volcano.

Newly collected, high resolution multi-beam sonar data are combined with previous bathymetry data to produce an improved bathymetric map of Shatsky Rise oceanic plateau. Bathymetry data show that two massifs within Shatsky Rise are immense central volcanoes with gentle flank slopes declining from a central summit. Tamu Massif is a slightly elongated, dome-like volcanic edifice; Ori Massif is square shaped and smaller in area. Several down-to-basin normal faults are observed on the western flank of the massifs but they do not parallel the magnetic lineations, indicating that these faults are probably not related to spreading ridge faulting. Moreover, the faults are observed only on one side of the massifs, which is contrary to expectations from a mechanism of differential subsidence around the massif center. Multi-beam data show many small secondary cones with different shapes and sizes that are widely-distributed on Shatsky Rise massifs, which imply small late-stage magma sources scattered across the surface of the volcanoes in the form of lava flows or explosive volcanism. Erosional channels occur on the flanks of Shatsky Rise volcanoes due to mass wasting and display evidence of down-slope sediment movement. These channels are likely formed by sediments spalling off the edges of summit sediment cap.

The geological evolution of the Earth during the mid-Cretaceous were shown to be anomalous, e.g., the pause of the geomagnetic field, the global sea level rise, and increased intra-plate volcanic activities, which could be attributed to deep mantle processes. As the anomalous volcanic activities occurred mainly in the Cretaceous Pacific, here we use basalt chemical compositions from the oceanic drilling (DSDP/ODP/IODP) sites to investigate their mantle sources and melting conditions. Based on locations relative to the Pacific plateaus, we classified these sites as oceanic plateau basalts, normal mid-ocean ridge basalts, and near-plateau seafloor basalts. This study shows that those normal mid-ocean ridge basalts formed during mid-Cretaceous are broadly similar in average Na-8, La/Sm and Sm/Yb ratios and Sr-Nd isotopic compositions to modern Pacific spreading ridge (the East Pacific Rise). The Ontong Java plateau (125-90 Ma) basalts have distinctly lower Na-8 and Nd-143/Nd-144, and higher La/Sm and Sr-87/Sr-86 than normal seafloor basalts, whereas those for the near-plateau seafloor basalts are similar to the plateau basalts, indicating influences from the Ontong Java mantle source. The super mantle plume activity that might have formed the Ontong Java plateau influenced the mantle source of the simultaneously formed large areas of seafloor basalts. Based on the chemical data from normal seafloor basalts, I propose that the mantle compositions and melting conditions of the normal mid-ocean ridges during the Cretaceous are similar to the fast spreading East Pacific Rise. Slight variations of mid-Cretaceous normal seafloor basalts in melting conditions could be related to the local mantle source and spreading rate.

The mafic–ultramafic Fariman complex in northeastern Iran has been interpreted as a Paleo-Tethyan ophiolitic fragment with subduction- and plume-related characteristics as well as a basin deposit on an active continental margin. Contributing to this issue, we present geochemical, geochronological, and mineralogical data for transitional and tholeiitic basalts. Thermodynamic modeling suggests picritic parental magmas with 16–21 wt% MgO formed at plume-like mantle potential temperatures of ca. 1460–1600 °C. Rare pyroxene spinifex textures and skeletal to feather-like clinopyroxene attest to crystallization from undercooled magma and high cooling rates. Chromium numbers and TiO2 concentrations in spinel are similar to those in intraplate basalts. 40Ar–39Ar dating of magmatic hornblende yielded a plateau age of 276 ± 4 Ma (2σ). Transitional basalt with OIB-like trace element characteristics is the predominant rock-type; less frequent are tholeiitic basalts with mildly LREE depleted patterns and picrites with intermediate trace element characteristics. All samples show MORB-OIB like Pb/Ce, Th/La, and Th/Nb ratios which preclude subduction-modified mantle sources and felsic crustal material. Tholeiitic basalts and related olivine cumulate rocks show MORB-like initial eNd values of + 9.4 to + 6.2 which define a mixing line with the data for the transitional basalts (eNd ca. + 2.6). Initial 187Os/188Os ratios of 0.124–0.293 support mixed sources with a high proportion of recycled mafic crust in the transitional basalts. High concentrations of highly siderophile elements are in agreement with the high mantle potential temperatures and inferred high-melting degrees. It is argued that the Fariman complex originated by melting of a mantle plume component as represented by the OIB-like transitional basalt and entrained asthenosphere predominant in the MORB-like tholeiites. Two lines of evidence such as association of the Fariman complex with pelagic to neritic sedimentary rocks and the tectonic position at the boundary of two continental blocks defined by ophiolites and accretionary complexes of different ages suggest formation in an oceanic domain. Thus, we interpret it as a fragment of an oceanic plateau, which escaped subduction and was accreted as exotic block in the Paleo-Tethyan suture zone.

The late Archaean Shimoga schist belt in the Western Dharwar Craton, with its huge dimensions and varied lithological associations of different age groups, is an ideal terrane to study Archean crustal evolution. The rock types in this belt are divided into Bababudhan Group and Chitradurga Group. The Bababudhan Group is dominated by mafic volcanic rocks followed by shallow marine sedimentary rocks while the Chitradurga Group is dominated by greywackes, pillowed basalts, and deep marine sedimentary rocks with occasional felsic volcanics. The Nb/Th and Nb/La ratios of the studied metabasalts of the Bababudhan Group indicate crustal contamination. They were extruded onto the vast Peninsular Gneisses through the rifting of the basement gneiss. The Nb/Yb ratios of high-magnesium basalts and tholeiitic basalts of Chitradurga Group suggest the enrichment of their source magma. Based on the flat primitive mantle-normalized multi-element plot with negative Nb anomalies and Th/Ta-La/Yb ratios, the high-magnesium basalts and tholeiitic basalts are considered to have erupted in an oceanic plateau setting with minor crustal contamination. The high-magnesium basalts and tholeiitic basalts formed two different pulses of same magma type, in which the first pulse of magma gave rise to high-magnesium basalts which were derived from deep mantle sources and underwent minor crustal contamination en route to the surface, while the second pulse of magma gave rise to tholeiitic basalts formed at similar depths to that of high-magnesium basalts and escaped crustal contamination. The associated lithological units found with the studied metavolcanic rock types of Bababudan and Chitradurga Groups of Dharwar Supergroup of rocks in Shimoga schist belt of Western Dharwar Craton confirm the mixed-mode basin development with a transition from shallow marine to deep marine settings.

We disagree the transitional supra-subduction zone model of Feng et al. (Int J Earth Sci (Geol Rundsch) 105:491–505, 2016) for the tectonic setting of Jifeng ophiolite suite in NE China. Existence of the komatiite in the Jifeng ophiolite indicates an oceanic plateau environment for this ophiolite suite within the so-called Xinlin–Xiguitu ocean.

This study presents laser step-heating 40Ar/39Ar age determinations of basaltic lava samples from Tamu Massif, the oldest and largest edifice of the submarine Shatsky Rise in the northwest Pacific and Earth’s proposed largest volcano. The rocks were recovered during Integrated Ocean Drilling Program Expedition 324, which cored 160 m into the igneous basement near the summit of Tamu Massif. The analyzed lavas cover all three major stratigraphic groups penetrated at this site and confirm a Late Jurassic/Early Cretaceous age for the onset of Shatsky Rise volcanism. Lavas analyzed from the lower and middle section of the hole yield plateau ages between 144.4 ± 1.0 and 143.1 ± 3.3 Ma with overlapping analytical errors (2σ), whereas a sample from the uppermost lava group produced a significantly younger age of 133.9 ± 2.3 Ma suggesting a late or rejuvenated phase of volcanism. The new geochronological data infer minimum (average) melt production rates of 0.63–0.84 km3/a over a time interval of 3–4 million years consistent with the presence of a mantle plume.

Most oceanic plateaux are massive basaltic volcanoes. However, the structure of these volcanoes, and how they erupt and evolve, is unclear, because they are remote and submerged beneath the oceans. Here we use multichannel seismic profiles and rock samples taken from Integrated Ocean Drilling Program core sites to analyse the structure of the Tamu Massif, the oldest and largest edifice of the Shatsky Rise oceanic plateau in the north-western Pacific Ocean. We show that the Tamu Massif is a single, immense volcano, constructed from massive lava flows that emanated from the volcano centre to form a broad, shield-like shape. The volcano has anomalously low slopes, probably due to the high effusion rates of the erupting lavas. We suggest that the Tamu Massif could be the largest single volcano on Earth and that it is comparable in size to the largest volcano in the Solar System, Olympus Mons on Mars. Our data document a class of oceanic volcanoes that is distinguished by its size and morphology from the thousands of seamounts found throughout the oceans.

Subduction of both the Iapetus and Rheic oceans began relatively soon after their opening. Vestiges of both the Iapetan and Rheic oceanic lithospheres are preserved as supra-subduction ophiolites and related mafic complexes in the Appalachian–Caledonian and Variscan orogens. However, available Sm–Nd isotopic data indicate that the mantle source of these complexes was highly depleted as a result of an earlier history of magmatism that occurred prior to initiation of the Iapetus and Rheic oceans. We propose two alternative models for this feature: either the highly depleted mantle was preserved in a long-lived oceanic plateau within the Paleopacific realm or the source for the basalt crust was been recycled from a previously depleted mantle and was brought to an ocean spreading centre during return flow, without significant re-enrichment en-route. Data from present-day oceans suggest that such return flow was more likely to have occurred in the Paleopacific than in new mid-ocean ridges produced in the opening of the Iapetus and Rheic oceans. Variation in crustal density produced by Fe partitioning rendered the lithosphere derived from previously depleted mantle more buoyant than the surrounding asthenosphere, facilitating its preservation. The buoyant oceanic lithosphere was captured from the adjacent Paleopacific, in a manner analogous to the Mesozoic–Cenozoic “capture” in the Atlantic realm of the Caribbean plate. This mechanism of “plate capture” may explain the premature closing of the oceans, and the distribution of collisional events and peri-Gondwanan terranes in the Appalachian–Caledonian and Variscan orogens.

Northwestern Costa Rica is built upon an oceanic plateau that has developed chemical and geophysical characteristics of the upper continental crust. A major factor in converting the oceanic plateau to continental crust is the production, evolution, and emplacement of silicic magmas. In Costa Rica, the Caribbean Large Igneous Province (CLIP) forms the overriding plate in the subduction of the Cocos Plate—a process that has occurred for at least the last 25 my. Igneous rocks in Costa Rica older than about 8 Ma have chemical compositions typical of ocean island basalts and intra-oceanic arcs. In contrast, younger igneous deposits contain abundant silicic rocks, which are significantly enriched in SiO2, alkalis, and light rare-earth elements and are geochemically similar to the average upper continental crust. Geophysical evidence (high Vp seismic velocities) also indicates a relatively thick (~40 km), addition of evolved igneous rocks to the CLIP. The silicic deposits of NW Costa Rica occur in two major compositional groups: a high-Ti and a low-Ti group with no overlap between the two. The major and trace element characteristics of these groups are consistent with these magmas being derived from liquids that were extracted from crystal mushes—either produced by crystallization or by partial melting of plutons near their solidi. In relative terms, the high-Ti silicic liquids were extracted from a hot, dry crystal mush with low oxygen fugacity, where plagioclase and pyroxene were the dominant phases crystallizing, along with lesser amounts of hornblende. In contrast, the low-Ti silicic liquids were extracted from a cool, wet crystal mush with high oxygen fugacity, where plagioclase and amphibole were the dominant phases crystallizing. The hot-dry-reducing magmas dominate the older sequence, but the youngest sequence contains only magmas from the cold-wet-oxidized group. Silicic volcanic deposits from other oceanic arcs (e.g., Izu-Bonin, Marianas) have chemical characteristics distinctly different from continental crust, whereas the NW Costa Rican silicic deposits have chemical characteristics nearly identical to the upper continental crust. The transition in NW Costa Rica from mafic oceanic arc and intra-oceanic magma to felsic, upper continental crust-type magma is governed by a combination of several important factors that may be absent in other arc settings: (1) thermal maturation of the thick Caribbean plateau, (2) regional or local crustal extension, and (3) establishment of an upper crustal reservoir.

A mixed conodont fauna-bearing limestone breccia, informally designated as Necoslie breccia, crops out as an internal sediment hosted by the Lower Moscovian limestone unit of the Pope limestone succession (Bashkirian to Asselian) of Cache Creek Complex in central British Columbia, reconstructed as a mid-oceanic buildup upon a seamount, or an oceanic plateau. Necoslie breccia consists predominantly of limestone fragments derived from the host rocks, with subordinate isolated skeletal debris, all embedded in a matrix of fine-grained clastic carbonates. The mixed conodont fauna yields the youngest element indicative of Carnian. Our microscopic examination and conodont paleontological data show that Necoslie breccia is better defined as a neptunian dyke formed by infilling sedimentation of carbonates in a fracturing-generated submarine cavity, rather than a karstic cave-fill deposit. We postulate that infilling sedimentation took place immediately after the formation of the cavity, most likely in a transgression period during Carnian time.

The volcanic stage of the 2.7-Ga Abitibi greenstone belt, Canada, is dominated by bimodal arc magma series and komatiite–basalt sequences. The latter represents an aerially extensive oceanic plateau erupted from an anomalously hot super-plume. Komatiites define a linear array of Nb/Th vs. Nb/U, extending from Nb/Th=8–20, and Nb/U=26–58, whereas basalts plot on a separate, but overlapping, field extending to higher Th/U but lower Nb/Th values. Inter-element ratios of Th, U, Nb, and LREE of komatiites and basalts plot with Phanerozoic and modern ocean plateau basalts. Th, U, Nb, and LREE are fractionated in subduction zones into low Nb/Th, Nb/U, and Nb/LREE arc crust, and complementary high Nb/Th, Nb/U, and Nb/LREE residual slab. Accordingly, the Archean komatiite–basalt association may be explained by a plume that likely originated from the core–mantle boundary with komatiites erupted from a hot axis containing recycled oceanic crust, and basalts erupted from the plume annulus that entrained upper mantle containing recycled oceanic and continental crust. High Nb/Th and Nb/U of plume-related volcanic sequences documented in Abitibi, Yilgarn, and Baltic Archean greenstone belts suggest that the extraction and recycling of continental crust may have occurred early in the Archean.

The Mikabu and Sorachi–Yezo belts comprise Jurassic ophiolitic complexes in Japan, where abundant basaltic to picritic rocks occur as lavas and hyaloclastite blocks. In the studied northern Hamamatsu and Dodaira areas of the Mikabu belt, these rocks are divided into two geochemical types, namely depleted (D-) and enriched (E-) types. In addition, highly enriched (HE-) type has been reported from other areas in literature. The D-type picrites contain highly magnesian relic olivine phenocrysts up to Fo93.5, and their Fo–NiO trend indicates fractional crystallization from a high-MgO primary magma. The MgO content is calculated as high as 25 wt%, indicating mantle melting at unusually high potential temperature (T p) up to 1,650 °C. The E-type rocks represent the enrichment in Fe and LREE and the depletion in Mg, Al and HREE relative to the D-type rocks. These chemical characteristics are in good accordance with those of melts from garnet pyroxenite melting. Volcanics in the Sorachi–Yezo belts can be divided into the same types as the Mikabu belt, and the D-type picrites with magnesian olivines also show lines of evidence for production from high T p mantle. Evidence for the high T p mantle and geochemical similarities with high-Mg picrites and komatiites from oceanic and continental large igneous provinces (LIPs) indicate that the Mikabu and Sorachi–Yezo belts are accreted oceanic LIPs that were formed from hot large mantle plumes in the Late Jurassic Pacific Ocean. The E- and D-type rocks were formed as magmas generated by garnet pyroxenite melting at an early stage of LIP magmatism and by depleted peridotite melting at the later stage, respectively. The Mikabu belt characteristically bears abundant ultramafic cumulates, which could have been formed by crystal accumulation from a primary magma generated from Fe-rich peridotite mantle source, and the HE-type magma were produced by low degrees partial melting of garnet pyroxenite source. They should have been formed later and in lower temperatures than the E- and D-type rocks. The Mikabu and Sorachi Plateaus were formed in a low-latitude region of the Late Jurassic Pacific Ocean possibly near a subduction zone, partially experienced high P/T metamorphism during subduction, and then uplifted in association with (or without, in case of Mikabu) the supra-subduction zone ophiolite. The Mikabu and Sorachi Plateaus may be the Late Jurassic oceanic LIPs that could have been formed in brotherhood with the Shatsky Rise.

The northern part of the western Kunlun (southern margin of the Tarim basin) represents a Sinian rifted margin. To the south of this margin, the Sinian to Paleozoic Proto-Tethys Ocean formed. South-directed subduction of this ocean, beneath the continental southern Kunlun block during the Paleozoic, resulted in the collision between the northern and southern Kunlun blocks during the Devonian. The northern part of the Paleo-Tethys Ocean, located to the south of the southern Kunlun, was subducted to the north beneath the southern Kunlun during the Late Paleozoic to Early Mesozoic. This caused the formation of a subduction-accretion complex, including a sizeable accretionary wedge to the south of the southern Kunlun. A microcontinent (or oceanic plateau?), which we refer to as “Uygur terrane,” collided with the subduction complex during the Late Triassic. Both elements together represent the Kara-Kunlun. Final closure of the Paleo-Tethys Ocean took place during the Early Jurassic when the next southerly located continental block collided with the Kara-Kunlun area. From at least the Late Paleozoic to the Early Jurassic, the Tarim basin must be considered a back-arc region. The Kengxiwar lineament, which “connects” the Karakorum fault in the west and the Ruogiang-Xingxingxia/Altyn-Tagh fault zone in the east, shows signs of a polyphase strike-slip fault along which dextral and sinistral shearing occurred.

The lithosphere of the Arabian–Nubian shield was mainly formed during an interval of about 150 million years near the end of the Proterozoic aeon. The events recorded in the rocks of the shield indicate that an oceanic plateau, formed by the head of an upwelling mantle plume, was later overprinted with continent-like characteristics by plate convergence and its associated magmatism. Similar sequences of events are seen in the geological record from Archaean to recent times, suggesting that the transformation from plume head to continental lithosphere has been an important component of continent generation throughout Earth history.

The island of Curacao in the southern Caribbean Sea is composed mainly of a thick sequence (>5 km) of pillow lavas, grading upwards from picrites at the base of the exposed section, to basalts nearer the top. Modelling suggests that picrites are related to the basalts by fractional crystallisation. Initial radiogenic isotope ratios of the picrites have a restricted compositional range: ɛNd=+6.1 to +6.6, 87Sr/86Sr=0.70296–0.70319; whereas the basalts display a wider range of compositions: ɛNd=+6.6 to +7.6, 87Sr/86Sr=0.70321–0.70671. This variation in isotope ratios between basalts and picrites may be due to the assimilation of altered oceanic crust (or possibly partial melts of such crust) by a picritic magma along with fractional crystallisation. The relatively narrow range of Nd and Pb isotopic compositions in the Curacao lavas suggests either that the source region was homogeneous, or that melts from a heterogeneous mantle source were well mixed before eruption. Chondritic to slightly light rare earth element enriched patterns, combined with long-term light rare earth element depletion (positive ɛNd), suggest that the lavas were formed by polybaric melting of spinel lherzolite, with small a contribution from garnet lherzolite melts. High-MgO lavas, the absence of a subduction related chemistry, and the chemical similarity to other oceanic plateaux, suggest a mantle plume origin for the Curacao lava succession. The Curacao volcanic sequence is part of an oceanic plateau formed at about 88–90 Ma, fragments of which are dispersed around the Caribbean as well as being obducted onto the western margin of Colombia and Ecuador. The occurrence of high-Mg lavas throughout this Cretaceous Caribbean–Colombian igneous province requires anomalously hot mantle (>200° C hotter than ambient upper mantle) over a large part of a putative plume head, which is inconsistent with some mantle plume models.

SHATSKY Rise, in the northwest Pacific Ocean, is probably the oldest extant oceanic plateau, and as with most such features, its origin is uncertain. Both oceanic plateaus and continental flood basalts are thought to be formed by rapid, voluminous eruptions that occur when the 'head' of a newly born mantle plume ascends to the base of the lithosphere1–3. High eruption rates have been estimated for flood basalts (for example, 1.5km3 yr−1 for the Deccan Traps2) from dating of lava flows, but the inaccessibility of oceanic plateaus makes it necessary to extrapolate dating information from a small number of samples and sites4. Here we estimate the eruption rate of Shatsky Rise by a method that is indirect, but has the virtue of 'sampling' the entire volume of the plateau above the surrounding sea floor. The main, southern part of the plateau has a positive magnetic anomaly, corresponding to a reversed geomagnetic polarity at the time of eruption. Using age constraints to identify the longest period of reversed polarity during which the plateau could have formed, we estimate that 2 × 106 km3 of material erupted at a minimum rate of 1.7km3 yr−1. This issomewhat less than the rate of 8–22 km3 yr−1 estimated for the Ontong–Java Plateau4, but still represents a massive eruption, consistent with the plume-head hypothesis.

Within-plate ocean island basalts probably originate by the partial melting of upwelling plumes coming from deep within the Earth's mantle1. Part of the observed heterogeneity in Sr, Nd and Pb isotopic ratios in oceanic basalts may result from the interaction of plumes with the lithosphere 2–4. and asthenosphere2, 5. Kerguelen and Heard Islands, in the southern Indian Ocean, are the latest products of a large and long-lived plume system6, comparable to the Hawaiian–Emperor chain in the Pacific. Kerguelen Island began forming ∼40 Myr ago on the Antarctic plate, near to the embryonic Southeast Indian Ridge (SEIR), and represents the continuation of the hotspot activity preserved on the Indian plate as the Ninety east Ridge7 (Fig. 1). With continuing sea-floor spreading, younger volcanism on Kerguelen has occurred further away from the influence of the SEIR. Here we show that the isotopic compositions of Kerguelen basalts vary in sympathy with this changing tectonic environment, and suggest not only that depleted asthenosphere, but also that Cretaceous oceanic plateau litho-sphere, may have contributed significantly to the earlier magmatic history of the island.

Numerous models have been presented to explain the late Cretaceous/early Cenozoic Laramide orogeny, which affected the foreland region of the western cordillera within the US1. The most attractive models invoke low-angle subduction2–4, which can develop for various reasons5,6. Evidence from South America indicates that one such reason may be the subduction of buoyant ocean floor. We here extend the low-angle subduction models by attributing the shallowing of the subduction angle during the Laramide orogeny to subduction of a large oceanic plateau of anomalously thick and buoyant oceanic crust. Pacific plate history indicates that a twin to the Hess Rise may represent the oceanic plateau that triggered Laramide events.