Geochemistry of marine sediments of the world's major subduction zones

This paper presents new major and trace-element data and Lu-Hf and Sm-Nd isotopic compositions for representative suites of marine sediment samples from 14 drill sites outboard of the world's major subduction zones. These suites and samples were chosen to represent the global range in lithology, Lu/Hf ratios, and sediment flux in subducting sediments worldwide. The data reported here represent the most comprehensive data set on subducting sediments and define the Hf-Nd isotopic variations that occur in oceanic sediments and constrain the processes that caused them. Using new marine sediment data presented here, in conjunction with published data, we derive a new Terrestrial Array given by the equation, epsilon-Hf = 1.55 * epsiolon-Nd + 1.21. This array was calculated using >3400 present-day Hf and Nd isotope values. The steeper slope and smaller y-intercept of this array, compared to the original expression (epsilon-Hf = 1.36 * epsilonNd + 2.89; Vervoort et al., 1999, doi:10.1016/S0012-821X(99)00047-3) reflects the use of present day values and the unradiogenic Hf of old continental samples included in the array. In order to examine the Hf-Nd isotopic variations in marine sediments, we have classified our samples into 5 groups based on lithology and major and trace-element geochemical compositions: turbidites, terrigenous clays, and volcaniclastic, hydrothermal and hydrogenetic sediments. Compositions along the Terrestrial Array are largely controlled by terrigenous material derived from the continents and delivered to the ocean basins via turbidites, volcaniclastic sediments, and volcanic inputs from magmatic arcs. Compositions below the Terrestrial Array derive from unradiogenic Hf in zircon-rich turbidites. The anomalous compositions above the Terrestrial Array largely reflect the decoupled behavior of Hf and Nd during continental weathering and delivery to the ocean. Both terrigenous and hydrogenetic clays possess anomalously radiogenic Hf, reflecting terrestrial sedimentary and weathering processes on the one hand and marine inheritance on the other. This probably occurs during complementary processes involving preferential retention of unradiogenic Hf on the continents in the form of zircon and release of radiogenic Hf from the breakdown of easily weathered, high Lu-Hf phases such as apatite.

Element analyses and isotopic composition of basalts from the western Pacific

New Sr- Nd- and Pb-isotopic and trace element data are presented on basalts from the Sulu and Celebes Basins, and the submerged Cagayan Ridge Arc (Western Pacific), recently sampled during Ocean Drilling Program Leg 124. Drilling has shown that the Sulu Basin developed about 18 Ma ago as a backarc basin, associated with the now submerged Cagayan Ridge Arc, whereas the Celebes Basin was generated about 43 Ma ago, contemporaneous with a general plate reorganisation in the Western Pacifc, subsequently developing as an open ocean receiving pelagic sediments until the middle Miocene. In both basins, a late middle Miocene collision phase and the onset of volcanic activity on adjacent arcs in the late Miocene are recorded. Covariations between 87Sr/86Sr and 143Nd/144Nd show that the seafoor basalts from both the Sulu and Celebes Basins are isotopically similar to depleted Indian mid-ocean ridge basalts (MORB), and distinct from East Pacifc Rise MORB, defining a single negative correlation. The Cagayan Arc volcanics are different, in that they have distinctly lower epsilon-Ne(T) for a given epsilon-Sr(T), compared to Sulu and Celebes basalts. In the 207Pb/204Pb and 208Pb/204Pb versus 206Pb/204Pb diagrams, the Celebes, Sulu and Cagayan rocks all plot distinctly above the Northern Hemisphere Reference Line, with high Delta 7/4 Pb (5.3-9.3) and Delta 8/4 Pb (46.3-68.1) values. They define a single trend of radiogenic lead enrichment from Celebes through Sulu to Cagayan Ridge, within the Indian Ocean MORB data field. The data suggest that the overall chemical and isotopic features of the Sulu, Cagayan and Celebes rocks may be explained by partial melting of a depleted asthenospheric N-MORB-type ("normal") mantle source with isotopic characteristics similar to those of the Indian Ocean MORB source. This asthenospheric source was slightly heterogeneous, giving rise to the Sr-Nd isotopic differences between the Celebes and Sulu basalts, and the Cagayan Ridge volcanics. In addition, a probably slab-derived component enriched in LILE and LREE is required to generate the elemental characteristics and low Ne(T) of the Cagayan Ridge island arc tholeiitic and calcalkaline lavas, and to contribute to a small extent in the backarc basalts of the Sulu Sea. The results of this study confirm and extend the widespread Indian Ocean MORB signature in the Western Pacifc region. This signature could have been inherited by the Indian Ocean mantle itself during the rupture of Gondwanaland, when fragments of this mantle could have migrated towards the present position of the Celebes, Sulu and Cagayan sources.

Chemistry and boron isotope ratios of Izu arc front volcanic rocks

The process of fluid release from the subducting slab beneath the Izu arc volcanic front (Izu VF) was examined by measuring B concentrations and B isotope ratios in the Neogene fallout tephra (ODP Site 782A). Both were measured by secondary ion mass spectrometry, in a subset of matrix glasses and glassy plagioclase-hosted melt inclusions selected from material previously analyzed for major and trace elements (glasses) and radiogenic isotopes (Sr, Nd, Pb; bulk tephra). These tephra glasses have high B abundances (~10-60 ppm) and heavy delta11B values (+4.5? to +12.0?), extending the previously reported range for Izu VF rocks (delta11B, +7.0? to +7.3?). The glasses show striking negative correlations of delta11B with large ion lithophile element (LILE)/Nb ratios. These correlations cannot be explained by mixing two separate slab fluids, originating from the subducting sediment and the subducting basaltic crust, respectively (model A). Two alternative models (models B and C) are proposed. Model B proposes that the inverse correlations are inherited from altered oceanic crust (AOC), which shows a systematic decrease of B and LILE with increasing depth (from basaltic layer 2A to layer 3), paralleled by an increase in delta11B (from ~ +1? to +10? to +24?). In this model, the contribution of sedimentary B is insignificant (<4% of B in the Izu VF rocks). Model C explains the correlation as a mixture of a low-delta11B (~ +1?) 'composite' slab fluid (a mixture of metasediment- and metabasalt-derived fluids) with a metasomatized mantle wedge containing elevated B (~1-2 ppm) and heavy delta11B (~ +14?). The mantle wedge was likely metasomatized by 11B-rich fluids beneath the outer forearc, and subsequently down dragged to arc front depths by the descending slab. Pb-B isotope systematics indicate that, at arc front depths, ~ 53% of the B in the Izu VF is derived from the wedge. This implies that the heavy delta11B values of Izu VF rocks are largely a result of fluid fractionation, and do not reflect variations in slab source provenance (i.e. subducting sediment vs. basaltic crust). Since the B content of the peridotite at the outer forearc (7-58 ppm B, mean 24 +/- 16 ppm) is much higher than beneath the arc front (~1-2 ppm B), the hydrated mantle wedge must have released a B-rich fluid on its downward path. This 'wedge flux' can explain (1) the across-arc decrease in B and delta11B (e.g. Izu, Kuriles), without requiring a progressive decrease in fluid flux from the subducting slab, and (2) the thermal structure of volcanic arcs, as reflected in the B and delta11B variations of volcanic arc rocks.

Rock magnetic properties of ODP Leg 134 sites

During Leg 134, the influence of ridge collision and subduction on the structural evolution of island arcs was investigated by drilling at a series of sites in the collision zone between the d'Entrecasteaux Zone (DEZ) and the central New Hebrides Island Arc. The DEZ is an arcuate Eocene-Oligocene submarine volcanic chain that extends from the northern New Caledonia Ridge to the New Hebrides Trench. High magnetic susceptibilities and intensities of magnetic remanence were measured in volcanic silts, sands, siltstones, and sandstones from collision zone sites. This chapter presents the preliminary results of studies of magnetic mineralogy, magnetic properties, and magnetic fabric of sediments and rocks from Sites 827 through 830 in the collision zone. The dominant carrier of remanence in the highly magnetic sediments and sedimentary rocks in the DEZ is low-titanium titanomagnetite of variable particle size. Changes in rock magnetic properties reflect variations in the abundance and size of titanomagnetite particles, which result from differences in volcanogenic contribution and the presence or absence of graded beds. Although the anisotropy of magnetic susceptibility results are difficult to interpret in terms of regional stresses because the cores were azimuthally unoriented, the shapes of the susceptibility ellipsoids provide information about deformation style. The magnetic fabric of most samples is oblate, dominated by foliation, as is the structural fabric. The variability of degree of anisotropy (P) and a factor that measures the shape of the ellipsoid (q) reflect the patchy nature of deformation, at a micrometer scale, that is elucidated by scanning electron microscope analysis. The nature of this patchiness implies that deformation in the shear zones is accomplished primarily by motion along bedding planes, whereas the material within the beds themselves remains relatively undeformed.

Major and trace element concentration of melt inclusions from Zhupanovsky volcano, Kamchatka

The paper presents data on naturally quenched melt inclusions in olivine (Fo 69-84) from Late Pleistocene pyroclastic rocks of Zhupanovsky volcano in the frontal zone of the Eastern Volcanic Belt of Kamchatka. The composition of the melt inclusions provides insight into the latest crystallization stages (~70% crystallization) of the parental melt (~46.4 wt % SiO2, ~2.5 wt % H2O, ~0.3 wt % S), which proceeded at decompression and started at a depth of approximately 10 km from the surface. The crystallization temperature was estimated at 1100 ± 20°C at an oxygen fugacity of deltaFMQ = 0.9-1.7. The melts evolved due to the simultaneous crystallization of olivine, plagioclase, pyroxene, chromite, and magnetite (Ol: Pl: Cpx : (Crt-Mt) ~ 13 : 54 : 24 : 4) along the tholeiite evolutionary trend and became progressively enriched in FeO, SiO2, Na2O, and K2O and depleted in MgO, CaO, and Al2O3. Melt crystallization was associated with the segregation of fluid rich in S-bearing compounds and, to a lesser extent, in H2O and Cl. The primary melt of Zhupanovsky volcano (whose composition was estimated from data on the most primitive melt inclusions) had a composition of low-Si (~45 wt % SiO2) picrobasalt (~14 wt % MgO), as is typical of parental melts in Kamchatka and other island arcs, and was different from MORB. This primary melt could be derived by ~8% melting of mantle peridotite of composition close to the MORB source, under pressures of 1.5 ± 0.2 GPa and temperatures 20-30°C lower than the solidus temperature of 'dry' peridotite (1230-1240°C). Melting was induced by the interaction of the hot peridotite with a hydrous component that was brought to the mantle from the subducted slab and was also responsible for the enrichment of the Zhupanovsky magmas in LREE, LILE, B, Cl, Th, U, and Pb. The hydrous component in the magma source of Zhupanovsky volcano was produced by the partial slab melting under water-saturated conditions at temperatures of 760-810°C and pressures of ~3.5 GPa. As the depth of the subducted slab beneath Kamchatkan volcanoes varies from 100 to 125 km, the composition of the hydrous component drastically changes from relatively low-temperature H2O-rich fluid to higher temperature H2O-bearing melt. The geothermal gradient at the surface of the slab within the depth range of 100-125 km beneath Kamchatka was estimated at 4°C/km.

Chemical composition of Late Cenozoic volcanic and volcaniclastic deposits from the western Woodlark Basin area, SW Pacific

Tephra fallout layers and volcaniclastic deposits, derived from volcanic sources around and on the Papuan Peninsula, form a substantial part of the Woodlark Basin marine sedimentary succession. Sampling by the Ocean Drilling Program Leg 180 in the western Woodlark Basin provides the opportunity to document the distribution of the volcanically-derived components as well as to evaluate their chronology, chemistry, and isotope compositions in order to gain information on the volcanic sources and original magmatic systems. Glass shards selected from 57 volcanogenic layers within the sampled Pliocene-Pleistocene sedimentary sequence show predominantly rhyolitic compositions, with subordinate basaltic andesites, basaltic trachy-andesites, andesites, trachy-andesites, dacites, and phonolites. It was possible to correlate only a few of the volcanogenic layers between sites using geochemical and age information apparently because of the formation of strongly compartmentalised sedimentary realms on this actively rifting margin. In many cases it was possible to correlate Leg 180 volcanic components with their eruption source areas based on chemical and isotope compositions. Likely sources for a considerable number of the volcanogenic deposits are Moresby and Dawson Strait volcanoes (D'Entrecasteaux Islands region) for high-K calc-alkaline glasses. The Dawson Strait volcanoes appear to represent the source for five peralkaline tephra layers. One basaltic andesitic volcaniclastic layer shows affinities to basaltic andesites from the Woodlark spreading tip and Cheshire Seamount. For other layers, a clear identification of the sources proved impossible, although their isotope and chemical signatures suggest similarities to south-west Pacific subduction volcanism, e.g. New Britain and Tonga- Kermadec island arcs. Volcanic islands in the Trobriand Arc (for example, Woodlark Island Amphlett Islands and/or Egum Atoll) are probable sources for several volcaniclastic layers with ages between 1.5 to 3 Ma. The Lusancay Islands can be excluded as a source for the volcanogenic layers found during Leg 180. Generally, the volcanogenic layers indicate much calc-alkaline rhyolitic volcanism in eastern Papua since 3.8 Ma. Starting at 135 ka, however, peralkaline tephra layers appear. This geochemical change in source characteristics might reflect the onset of a change in geotectonic regime, from crustal subduction to spreading, affecting the D'Entrecasteaux Islands region. Initial 143Nd/144Nd ratios as low as 0.5121 and 0.5127 for two of the tephra layers are interpreted as indicating that D'Entrecasteaux Islands volcanism younger than 2.9 Ma occasionally interacted with the Late Archean basement, possibly reflecting the mobilisation of the deep continental crust during active rift propagation.

Mineralogy of sea-floor surface sediments, Sonne cruise SO202/1 to the North Pacific and Bering Sea

During expedition 202 of research vessel SONNE in 2009, 39 sea-floor surface sediments were sampled over a wide area across the North Pacific and the Bering Sea, which are well suited as reference archives of modern environmental processes. In this study, we used the samples to infer the documentation of land-ocean linkages of terrigenous sediment supply. We followed an integrated approach of grain-size analysis, bulk mineralogy, and clay mineralogy in combination with statistical data evaluation (end-member modelling of grain-size data, fuzzy-cluster analysis of mineralogical data), in order to identify the significant sources and modes of sediment transport in an overregional context. We also compiled literature data on clay mineralogy and updated those with the new data. Today, two processes of terrigenous sediment supply prevail in the study area: far-distant aeolian sediment supply to the pelagic North Pacific as well as hemipelagic sediment dispersal from nearby land sources by ocean currents along the continental margins and island arcs of the study area. The aeolian particles show the finest grain sizes (clay and fine silt), while the hemipelagic sediments have high abundances of sortable silt, particles >10 microns.

He isotope, U/Th isotope, Calcium carbonate, biogenic opal, rare earth element concentrations, and grain size distribution measured on core-top sediments during SONNE cruise SO202 (INOPEX)

Eolian dust is a significant source of iron and other nutrients that are essential for the health of marine ecosystems and potentially a controlling factor of the high nutrient-low chlorophyll status of the Subarctic North Pacific. We map the spatial distribution of dust input using three different geochemical tracers of eolian dust, 4He, 232Th and rare earth elements, in combination with grain size distribution data, from a set of core-top sediments covering the entire Subarctic North Pacific. Using the suite of geochemical proxies to fingerprint different lithogenic components, we deconvolve eolian dust input from other lithogenic inputs such as volcanic ash, ice-rafted debris, riverine and hemipelagic input. While the open ocean sites far away from the volcanic arcs are dominantly composed of pure eolian dust, lithogenic components other than eolian dust play a more crucial role along the arcs. In sites dominated by dust, eolian dust input appears to be characterized by a nearly uniform grain size mode at ~4 µm. Applying the 230Th-normalization technique, our proxies yield a consistent pattern of uniform dust fluxes of 1-2 g/m**2/yr across the Subarctic North Pacific. Elevated eolian dust fluxes of 2-4 g/m**2/yr characterize the westernmost region off Japan and the southern Kurile Islands south of 45° N and west of 165° E along the main pathway of the westerly winds. The core-top based dust flux reconstruction is consistent with recent estimates based on dissolved thorium isotope concentrations in seawater from the Subarctic North Pacific. The dust flux pattern compares well with state-of-the-art dust model predictions in the western and central Subarctic North Pacific, but we find that dust fluxes are higher than modeled fluxes by 0.5-1 g/m**2/yr in the northwest, northeast and eastern Subarctic North Pacific. Our results provide an important benchmark for biogeochemical models and a robust approach for downcore studies testing dust-induced iron fertilization of past changes in biological productivity in the Subarctic North Pacific.

Geochemistry and radiogenic isotope ratios of the volcanic sequences from Izu-Bonin Arc, Japan

New radiogenic isotope and trace element data are presented for the volcanic sequences along 600 km of the active Izu-Bonin arc, the Oligocene Izu arc, and their associated rift basins. As with many intra-oceanic island arcs, the Pliocene-Recent Izu-Bonin frontal-arc lavas are highly depleted in Zr, Nb and the rare-earth elements relative to typical mid-ocean ridge basalt (MORB), indicating that the mantle wedge source has undergone a previous episode of melting. Ratios between these elements (such as Nb/Zr and La/Sm), as well as 143Nd/144Nd, do not vary significantly along the length of the frontal-arc. These parameters suggest that each of the arc volcanoes is derived from similar melt fractions of the mantle wedge. However, Ba/Zr, Ba/Rb and 87Sr/86Sr increase along the frontal-arc to the north. This leads us to propose that a variable enrichment in Ba and radiogenic Sr is superimposed on the mantle wedge. Sr-Nd and Pb-Nd isotope variation indicate that both Sr and Pb become more radiogenic after fluid addition. However, Pb isotope ratios do not correlate with increases in Pb concentration or ratios such as Ba/Zr and Nb/Pb. In other words, the Pb isotopic composition of the arc lavas appears to be independent of the amount of Pb introduced by subduction fluids into the mantle source. This buffering of Pb isotopes along the frontal-arc means that the isotopic composition of the lavas is indistinguishable from that of the fluid. Isotopic mixing models presented for the arc are only illustrative of the many plausible combinations of components and quantities. Despite this, we are able to determine that the mantle wedge has isotopic characteristics similar to Indian Ocean MORB, and that the subduction-fluid solute is primarily derived from subducted oceanic basalt with a <2% contribution from subducted sediment. Lavas in the Oligocene Izu arc and fore-arc basin were derived from a mantle wedge of similar composition to the active arc. Despite levels of Pb enrichment comparable to those of the modern arc, the Pb isotopes of the Oligocene volcanics indicate a lower sediment input into the melting region.

Biogenic opal, biogenic barium, calcium carbonate and U/Th isotope data from core-top sediments of SONNE cruise SO202 (INOPEX)

Sedimentary proxies used to reconstruct marine productivity suffer from variable preservation and are sensitive to factors other than productivity. Therefore, proxy calibration is warranted. Here we map the spatial patterns of two paleoproductivity proxies, biogenic opal and barium fluxes, from a set of core-top sediments recovered in the Subarctic North Pacific. Comparisons of the proxy data with independent estimates of primary and export production, surface water macronutrient concentrations and biological pCO2 drawdown indicate that neither proxy shows a significant correlation with primary or export productivity for the entire region. Biogenic opal fluxes, when corrected for preservation using 230Th-normalized accumulation rates, show a good correlation with primary productivity along the volcanic arcs (tau = 0.71, p = 0.0024) and with export productivity throughout the western Subarctic North Pacific (tau = 0.71, p = 0.0107). Moderate and good correlations of biogenic barium flux with export production (tau = 0.57, p = 0.0022) and with surface water silicate concentrations (tau = 0.70, p = 0.0002) are observed for the central and eastern Subarctic North Pacific. For reasons unknown, however, no correlation is found in the western Subarctic North Pacific between biogenic barium flux and the reference data. Nonetheless, we show that barite saturation, uncertainty in the lithogenic barium corrections and problems with the reference datasets are not responsible for the lack of a significant correlation between biogenic barium flux and the reference data. Further studies evaluating the factors controlling the variability of the biogenic constituents in the sediments are desirable in this region.