ALBERT

Description: The Ontong Java and Manihiki oceanic plateaus are believed to have formed through high-degree melting of a mantle plume head. Boninite-like, low-Ti basement rocks at Manihiki, however, imply a more complex magma genesis compared with Ontong Java basement lavas that can be generated by ∼30% melting of a primitive mantle source. Here we show that the trace element and isotope compositions of low-Ti Manihiki rocks can best be explained by re-melting of an ultra-depleted source (possibly a common mantle component in the Ontong Java and Manihiki plume sources) re-enriched by ≤1% of an ocean-island-basalt-like melt component. Unlike boninites formed via hydrous flux melting of refractory mantle at subduction zones, these boninite-like intraplate rocks formed through adiabatic decompression melting of refractory plume material that has been metasomatized by ocean-island-basalt-like melts. Our results suggest that caution is required before assuming all Archaean boninites were formed in association with subduction processes.

Description: Highlights • Four of the seven seamounts northeast of the Galápagos Platform are drowned islands • The ages of the seamounts range from 5.2 Ma to 0.5 Ma • Seamount morphology changes from conical to elongate at ~1.5 Ma • The locus of volcanism appears to migrate eastward at the rate of Nazca plate motion Abstract We present new geochemical and 40Ar/39Ar analyses from seven seamounts located off the northeastern margin of the shallow Galápagos Platform. Initial volcanism at 5.2 Ma created a small island (Pico) over the current location of the hotspot with geochemically enriched lavas. There is no further record of magmatism in the study area until 3.8 to 2.5 Ma, during which four roughly conical volcanoes (Sunray, Grande, Fitzroy, and Beagle) formed through eruption of lavas derived from a depleted mantle source. Sunray, Fitzroy, and Grande were islands that existed for ~3 m.y. ending with the submergence of Fitzroy at ~0.5 Ma. The youngest seamounts, Largo and Iguana, do not appear to have been subaerial and were active at 1.3 Ma and 0.5 Ma, respectively, with the style of edifice changing from the previous large cones to E-W elongate, composite structures. The progression of magmatism suggests that Pico erupted near 91.5°W near the location of the Galápagos plume while the others formed well east of the plume center. If the locations of initial volcanism are calculated using the eastward velocity of the Nazca plate, there appears to be a progression of younger volcanism toward the east, opposite what would be expected from a fixed mantle plume source. The rate that initial volcanism moves eastward is close to the plate velocity. A combination of higher temperature and geochemical enrichment of the thickened lithosphere of the Galápagos platform could have provided a viscosity gradient at the boundary between the thick lithosphere and the thinner oceanic lithosphere to the northeast. As this boundary moved eastward with the Nazca plate, it progressively triggered shear-driven mantle upwelling and volcanism.

Description: The Nifonea submarine volcano rises 1000 m above the seafloor of the Vate Trough back-arc basin behind the New Hebrides island arc. This large volcanic edifice has a caldera of ∼8 km diameter and is connected to two ∼20 km long volcanic rift zones in the back-arc basin. We present new chemical and isotope data for volcanic glasses and whole-rocks from both the volcano and its rift zones. Lavas from Nifonea volcano show an evolution towards more incompatible element enrichment, with the most enriched lavas being the youngest eruption products on the caldera floor. These are products of significant fractional crystallization, show minor contamination by hydrothermal fluids (〈0·3%) and reflect mixing of melts derived from depleted upper mantle and melts from an enriched source similar to those occurring in the North Fiji Basin. The enrichment in Nb of these lavas is comparable with that of some lavas from the New Hebrides island arc (e.g. Mota Lava island), where these coexist with typical island arc basalts. The lavas erupted along the rift zones in the Vate Trough back-arc basin are relatively depleted in incompatible elements, indicating melting of depleted upper mantle with a minor addition of a sediment-derived fluid. Our observations suggest that the mantle beneath Vate Trough is heterogeneous on a small scale (〈20 km) and that the occurrence of these enriched and fertile mantle portions has a stronger control on melting processes than the influx from the subducting slab, as all samples were recovered at a similar distance from the trench.

Description: We present a comprehensive major and trace element and Sr–Nd–Pb isotope dataset from the major volcanic units exposed on La Palma and show how geochemical and volcanic evolution can be linked to asthenospheric and lithospheric processes. Lavas from the northern shield (from Basal Complex to Taburiente to Bejenado volcanism, 3–4 Ma to 400 ka) become more alkalic and SiO2-undersaturated with decreasing age, but show little change in MgO-normalized trace element compositions. Their high (Nb,Ta)/U and Ba/Th but low La/Nb ratios suggest assimilation of amphibole, probably in the lithospheric mantle that was metasomatized by earlier melts. Lavas from the Cumbre Vieja unit (〈125 ka) in the southern half of La Palma are more incompatible-element enriched and probably formed through lower degrees of melting than those from the northern shield, which are nearly identical isotopically. Their Nb/U ratios are mostly within the range 47 ± 10, significantly below those of the earlier lavas. In 206Pb/204Pb versus 143Nd/144Nd, 208Pb/204Pb and 208Pb/206Pb isotope diagrams, the Basal Complex rocks and lavas from the adjacent El Hierro island form a separate trend compared with the younger subaerial La Palma lavas. Both groups share a common depleted end-member but require separate, enriched HIMU-like end-members, believed to be located within the asthenosphere. The temporal and spatial variations in the composition of La Palma and El Hierro lavas could be explained within the context of NE-directed plate motion over a zoned Canary plume. After La Palma moved away from the asthenospheric source domain of the Basal Complex, El Hierro formed above the same domain, whereas the younger units on La Palma tapped a distinct asthenospheric domain located further north. The short-lived Bejenado volcano that formed directly after the giant Cumbre Nueva sector collapse at c. 560 ka produced the isotopically most depleted lavas reported from La Palma thus far. Their compositions suggest incorporation of a depleted pyroxenitic component. The Bejenado lavas also extend to the highest Nb/U and Ba/Th and lowest La/Nb ratios of all La Palma lavas, consistent with increased melting of amphibole within the lithospheric mantle or lower crust. We propose that the collapse is related to the migration of magmatism to the south of La Palma, and led to short-term enhanced decompression melting of amphibole and pyroxenite within the lithosphere.

Description: Silicate-sulfide liquid immiscibility in mantle-derived magmas has important control on the budget of siderophile and chalcophile metals, and is considered to be instrumental in the origin orthomagmatic sulfide deposits. Data on primitive sulfide melts in natural samples, even those representing most voluminous magmatism in oceanic rifts, are very scarce due to the small size and poor preservation of incipient sulfide melt globules. Here we present the first detailed report of the crystallized sulfides melts in the oceanic picrites of the (presumably) Cretaceous age Kamchatsky Mys ophiolite complex in Eastern Kamchatka (Far East Russia). Sulfide melts are present in three forms; (1) as inclusions in olivine (87.1–89.6 mol% Fo), (2) interstitial to the groundmass minerals (clinopyroxene, plagioclase, and Ti-magnetite) of studied picrites, and (3) as daughter phases in silicate melt inclusions hosted by olivine and Cr-spinel phenocrysts. The sulfide melt inclusions in olivine and the groundmass of studied rocks are composed of several sulfide phases that correspond to the monosulfide (Fe–Ni; Mss) and intermediate (Fe–Cu–Ni; Iss) solid solutions. Several 〈0.5 μm Pd–Sn, Pt–Ag, and Au–Ag phases are recorded within the matrix sulfides, commonly along phase boundaries and fractures. Major elements (S, Fe, Cu, Ni, Co), platinum group elements (PGE), and gold analyzed in the homogenized olivine-hosted sulfide melt inclusions, and phases identified in the matrix sulfides record the range of magmatic sulfide compositions. The most primitive sulfide liquids are notably enriched in Ni and Cu [(Ni+Cu)/Fe, at% 〉 0.5], continuously evolve with crystallization of (e.g., increasing Cu/Ni and Au/PGE) and demonstrate metal fractionation between Mss and Iss. Although the compositional systematics found in this study are consistent with those previously recorded, the compositions of individual sulfide phases are strongly affected by the noble metal (PGE, Au) “nuggets” that exsolve at subsolidus temperatures and form during serpentinization of the rocks. We conclude that the budget of noble metals in the studied picrites is controlled by sulfides, but the abundances of Pt and Au are influenced by mobility in post-magmatic alteration. Our data can be also used for modeling sulfide saturation at crustal pressures and understanding behavior of the noble metals in primitive oceanic magmas.

Description: Oceanic basalts reflect the heterogeneities in the earth's mantle, which can be explained by five mantle end members. The HIMU end member, characterized by high time-integrated μ (238U/204Pb), is defined by the composition of lavas from the ocean islands of St. Helena, South Atlantic Ocean and Mangaia and Tubuai (Cook-Austral Islands), South Pacific Ocean. It is widely considered to be derived from a mantle reservoir that is rarely sampled and not generally involved in mixing with the other mantle components. On the other hand, the FOZO end member, located at the FOcal ZOne of oceanic volcanic rock arrays on isotope diagrams, is considered to be a widespread common component with slightly less radiogenic 206Pb/204Pb and intermediate Sr-Nd-Hf isotopic compositions. Here we present new major and trace element, Sr-Nd-Pb-Hf isotope and geochronological data from the Walvis Ridge and Richardson Seamount in the South Atlantic Ocean and the Manihiki Plateau and Eastern Chatham Rise in the southwest Pacific Ocean. Our new data, combined with literature data, document a more widespread (nearly global) distribution of the HIMU end member than previously postulated. Our survey shows that HIMU is generally associated with low-volume alkaline, carbonatitic and/or kimberlitic intraplate volcanism, consistent with derivation from low degrees of melting of CO2-rich sources. The majority of end member HIMU locations can be directly related to hotspot settings. The restricted trace element and isotopic composition (St. Helena type HIMU), but near-global distribution, point to a deep-seated, widespread reservoir, which most likely formed in the Archean. In this context we re-evaluate the origin of a widespread HIMU reservoir in an Archean geodynamic setting. We point out that the classic ocean crust recycling model cannot be applied in a plume-lid dominated tectonic setting, and instead propose that delamination of carbonatite- metasomatized subcontinental lithospheric mantle could be a suitable HIMU source.

Description: Highlights • First comprehensive data set of the seamounts from the Walvis Ridge. • The seamounts are 20–40 Myr younger than the age progressive Walvis Ridge basement. • The composition of the seamounts extends from the St. Helena HIMU to EMORB. • The seamounts are derived from a distinct source compared to the Walvis Ridge. • The temporal change from EM I to HIMU could reflect the compositional heterogeneities of the LLSVP. Abstract Volcanic activity at many oceanic volcanoes, ridges and plateaus often reawakens after hiatuses of up to several million years. Compared to the earlier magmatic phases, this late-stage (rejuvenated/post-erosional) volcanism is commonly characterized by a distinct geochemical composition. Late-stage volcanism raises two hitherto unanswered questions: Why does volcanism restart after an extended hiatus and what is the origin of this volcanism? Here we present the first 40Ar/39Ar age and comprehensive trace element and Sr–Nd–Pb–Hf isotopic data from seamounts located on and adjacent to the Walvis Ridge in the South Atlantic ocean basin. The Walvis Ridge is the oldest submarine part of the Tristan-Gough hotspot track and is famous as the original type locality for the enriched mantle one (EM I) end member. Consistent with the bathymetric data, the age data indicates that most of these seamounts are 20–40 Myr younger than the underlying or nearby Walvis Ridge basement. The trace element and isotope data reveal a distinct compositional range from the EM I-type basement. The composition of the seamounts extend from the St. Helena HIMU (high time-integrated 238U/204Pb mantle with radiogenic Pb isotope ratios) end member to an enriched (E) Mid-Ocean-Ridge Basalt (MORB) type composition, reflecting a two-component mixing trend on all isotope diagrams. The EMORB end member could have been generated through mixing of Walvis Ridge EM I with normal (N) MORB source mantle, reflecting interaction of Tristan-Gough (EM I-type) plume melts with the upper mantle. The long volcanic quiescence and the HIMU-like geochemical signature of the seamounts are unusual for classical hotspot related late-stage volcanism, indicating that these seamounts are not related to the Tristan-Gough hotspot volcanism. Two volcanic arrays in southwestern Africa (Gibeon-Dicker Willem and Western Cape province) display similar ages to the late-stage Walvis seamounts and also have HIMU-like compositions, suggesting a larger-scale event at ∼77–49 Ma. We propose that the EM I-like mantle plumes rise from the edges of the African Large Low Shear Velocity Province (LLSVP; Tristan-Gough, Discovery and Shona hotspot), whereas the HIMU-dominated intraplate lavas (St. Helena, Gibeon-Dicker Willem and Western Cape province) and the late-stage Walvis seamounts tap material from internal portions of the African LLSVP, suggesting possible lateral and/or vertical chemical zonation of the African LLSVP.

Description: The 541 ± 4 Ma-old magnesian, weakly peraluminous, calc-alkalic Donkerhoek Onanis granite is part of the ca. 6000 km2 large Donkerhoek batholith in the Southern Zone of the Damara orogen of Namibia. Linear major and trace element variations and decreasing MgO, FeO, Al2O3, CaO, K2O, Na2O, Ba and Sr concentrations with increasing SiO2 indicate that this part of the batholith represent a coherent mass and underwent fractional crystallization processes. The Donkerhoek Onanis granites are isotopically evolved (initial εNd: −4.7 to −12.3, initial 87Sr/86Sr: 0.7099–0.7157) with moderately radiogenic Pb isotope ratios (206Pb/204Pb: 17.26–18.22; 207Pb/204Pb: 15.59–15.67; 208Pb/204Pb: 37.60–38.06). Beside heterogeneities imparted by the sources, an evaluation of LREE fractionation and Nd isotope data suggests that AFC processes also modified some samples. Based on the chemical and isotope data, the Donkerhoek Onanis granites cannot be derived by partial melting of Al- and Fe-rich metasedimentary rocks of the Kuiseb formation in which they intruded. Instead, melting of meta-igneous crustal sources with Proterozoic crustal residence ages is more likely. Three igneous to meta-igneous rock suites from the area (Matchless amphibolites, Proterozoic mafic to felsic gneisses from the southern Kalahari craton basement, syn-tectonic Salem granodiorites to granites) are potential sources. An evaluation of chemical and isotope data suggests that remelting of early syn-orogenic Salem-type granites is the most likely process which would also explain the existence of ca. 563 ± 4 Ma-old zircon in the Donkerhoek Onanis granites. Comparison of the Donkerhoek Onanis granites with experimentally derived melt compositions from an intermediate igneous parent indicates temperatures between 800 and 850 °C. It is suggested that the Pan-African igneous activity in this part of the Damara Belt was a moderate-temperature intra-crustal event. Although there are some compositional similarities with juvenile granites generated in subduction zones, unradiogenic Pb isotope ratios and moderately radiogenic Sr and unradiogenic Nd isotopes suggest that reprocessed crustal rocks are more likely sources. Previously obtained high δ18O values of the Donkerhoek Onanis granites ranging from 11.8 to 13.6‰, covering the range of δ18O values obtained on Salem-type granites from the area (12.5–13.3‰) confirm this view. In contrast to igneous processes along active continental margins that produce juvenile batholiths with calc-alkaline affinities, this igneous event was not a major crust-forming episode and the Donkerhoek Onanis granites represent reprocessed crustal material.