Jean-Philippe Eissen

Can slab melting be caused by flat subduction

Slab melting has been suggested as a likely source of adakitic arc magmas (i.e., andesitic and dacitic magmas strongly depleted in Y and heavy rare earth elements). Existing numerical and petrologic models, however, restrict partial melting to very young (≤ 5 Ma) oceanic crust (typically at 60–80 km depth). Paradoxically, most of the known Pliocene-Quaternary adakite occurrences are related to subduction of 10–45 Ma lithosphere, which should not be able to melt under normal subduction-zone thermal gradients. We propose an unusual mode of subduction known as flat subduction, occurring in ∼10% of the worlds convergent margins, that can produce the temperature and pressure conditions necessary for fusion of moderately old oceanic crust. Of the 10 known flat subduction regions worldwide, eight are linked to present or recent (<6 Ma) occurrences of adakitic magmas. Observations from Chile, Ecuador, and Costa Rica suggest a three-stage evolution: (1) steep subduction produces a narrow calc-alkaline arc, typically ∼300 km from the trench, above the asthenospheric wedge; (2) once flat subduction begins, the lower plate travels several hundred kilometers at nearly the same depth, thus remaining in a pressure-temperature window allowing slab melting over this broad distance; and (3) once flat subduction continues for several million years, the asthenospheric wedge disappears, and a volcanic gap results, as in modern-day central Chile or Peru. The proposed hypothesis, which reconciles thermal models with geochemical observations, has broad implications for the study of arc magmatism and for the thermal evolution of convergent margins.

Mio–Pliocene adakite generation related to flat subduction in southern Ecuador: the Quimsacocha volcanic center

New geochemical and geochronological data on the Miocene^Pliocene Quimsacocha volcanic center (QVC) have led to the recognition of adakitic lavas generated by slab melting related to the flat slab subduction in southern Ecuador and northern Peru. The QVC, located in the presently inactive southern part of the Ecuadorian arc, was built up during three distinctive volcanic phases. The first phase generated a basal edifice with mainly andesitic lava flows, while the second phase is characterized by the emplacement of cryptodomes, domes and related outflow breccias comprised of andesites and some dacites. The last phase released rhyolitic ignimbrites associated with the formation of a large caldera, which was later partly filled by dacitic^rhyolitic domes. Geochemical data for the QVC indicate higher Al2O3, TiO2 ,N a 2O, Zr and Sr contents and lower Fe2O3*, MgO, Y, MREE and HREE abundances, compared to other eruptive rocks of the Plio^Quaternary volcanic front of Ecuador. Such geochemical features, as well as the frequent presence of an associated epithermal gold deposit, are characteristic of the involvement of slab melts, also known as adakites [1,2], in the generation of these magmas. After a calc-alkaline arc magmatism phase, slab horizontalization ^ in response to the subduction of a buoyant oceanic plateau ^ results in increased involvement of a slab melting component in the magmas produced. However, pristine adakites were generated and emplaced during a relatively short period, as indicated by zircon fission-track ages. Then volcanic activity stopped and a volcanic gap formed. The identification of these adakites, their location and age support a model of slab melting associated with flat slab subduction [M.A. Gutscher et al., Geology 28 (2000) 535^538]. fl 2001 Elsevier Science B.V. All rights reserved.

Magmatic response to early aseismic ridge subduction: the Ecuadorian margin case (South America)

A geochemical and isotopic study of lavas from Pichincha, Antisana and Sumaco volcanoes in the Northern Volcanic Zone (NVZ) in Ecuador shows their magma genesis to be strongly influenced by slab melts. Pichincha lavas (in fore arc position) display all the characteristics of adakites (or slab melts) and were found in association with magnesian andesites. In the main arc, adakite-like lavas from Antisana volcano could be produced by the destabilization of pargasite in a garnet-rich mantle. In the back arc, high-niobium basalts found at Sumaco volcano could be produced in a phlogopite-rich mantle. The strikingly homogeneous isotopic signatures of all the lavas suggest that continental crust assimilation is limited and confirm that magmas from the three volcanic centers are closely related. The following magma genesis model is proposed in the NVZ in Ecuador: in fore arc position beneath Pichincha volcano, oceanic crust is able to melt and produces adakites. En route to the surface, part of these magmas metasomatize the mantle wedge inducing the crystallization of pargasite, phlogopite and garnet. In counterpart, they are enriched in magnesium and are placed at the surface as magnesian andesites. Dragged down by convection, the modified mantle undergoes a first partial melting event by the destabilization of pargasite and produces the adakite-like lavas from Antisana volcano. Lastly, dragged down deeper beneath the Sumaco volcano, the mantle melts a second time by the destabilization of phlogopite and produces high-niobium basalts. The obvious variation in spatial distribution (and geochemical characteristics) of the volcanism in the NVZ between Colombia and Ecuador clearly indicates that the subduction of the Carnegie Ridge beneath the Ecuadorian margin strongly influences the subduction-related volcanism. It is proposed that the flattening of the subducted slab induced by the recent subduction (<5 Ma?) of the Carnegie Ridge has permitted the progressive warming of the oceanic crust and its partial melting since ca. 1.5 Ma. Since then, the production of adakites in fore arc position has deeply transformed the magma genesis in the overall arc changing from ‘typical’ calc-alkaline magmatism induced by hydrous fluid metasomatism, to the space- and time-associated lithology adakite/high-Mg andesite/adakite-like andesite/high-Nb basalts characteristic of slab melt metasomatism.

Giant tuff cone and 12-km-wide associated caldera at Ambrym Volcano (Vanuatu, New Hebrides Arc)

Abstract Ambrym, in the New Hebrides arc, has been considered as an effusive, basaltic volcano. The present paper discusses the general structure of this edifice, which consists of a basal shield volcano topped by an exceptionally large tuff cone surrounding a 12-km-wide summit caldera. Dacitic pyroclastic flow deposits are exposed in the lower part of the tuff series; they grade upward into composite sequences of bedded surtseyan-type hyaloclastites, ash flow deposits, and fallout tephra which are essentially basaltic in composition, in such a way that the tuff cone may be considered as mainly basaltic. The relationship between the eruption of pyroclastics and the collapse event precludes a classical model of caldera formation at a basaltic volcano in which “quiet” subsidence (i.e. Kilauea type) is the dominant mechanism. Interpretation of the tuff series implies intervention of external water and suggests both explosive and collapse mechanisms. A model of caldera formation which assumes an enlargement of the ring fracture during a first plinian and dacitic, then essentially hydrobasaltic eruption is proposed.

Kuwae (≈ 1425 A.D.): the forgotten caldera

In Vanuatu, Tongoa and Epi islands once formed part of a larger landmass, Kuwae, which was partly destroyed during a cataclysmic seismo-volcanic event that is recorded in local folklore. It led to the formation of a 12-kmlong and 6-km-wide oval-shaped submarine caldera with two distinct basins and a total area of - 60 km2 at the level of the rim. The age ofthis eniption, 1420-1430 A.D., and the structure of the related collapse are discussed and a composite log ( 143 m) of the pyroclastics surrounding the caldera is presented. They comprise thick hydromagmatic deposits belonging to a terminal hydromagmatic phase of the pre-caldera edifice, which grade upwards into two major sequences of pyroclastic flow deposits, clearly related to the caldera event. Collapse near the caldera edge was at least in the range 650 to 950 m, and may have been as much as 800 to 1100 m. The volume of rocks engulfed during the caldera formation is - 32-39 km3, suggesting the same volume of magma was erupted. Even if two coalescent collapse structures were formed, it is worth noting that the Kuwae caldera is not a reactivated structure, but the result of a single event of short duration which occurred in the first half of the Fifteenth century. This event is one of the seven biggest caldera-forming events during the last 10,000 years, and is comparable with the Santorini Minoan eruption and the Crater Lake eruption.

North Fiji Basin basalts and their magma sources:Part I.Incompatible element constraints

A systematic compilation of the geochemical data collected during the starmer programme is presented, in addition to all published data gathered along the North Fiji Basin (NFB) spreading system. From the 194 samples selected, the geochemical variations observed along the different spreading segments can be related to uni- or multi-modal origin of distinctive magma sources. These variations are interpreted as directly linked to the geodynamical features which surround the NFB. In the southern NFB, on the N174°E segment, N-MORB are present, but the low rate (or dead) subduction along the Hunter ridge still marks its influence as several basalts show a significant subduction-related contamination with negative Nb anomaly and high LOI. Several others are marked by a weak E-MORB source contribution, that may be related to subducted-OIB seamounts from the South Fiji Basin. In the central NFB, along the N-S segment which represents the most active and morphologically regular spreading ridge of the NFB, the magma source produces only N-MORB, with depleted LILE, HFSE, and LREE patterns, except in its northernmost part where magmatic signatures similar to that of the N15° segment appear. Along the N15° segment, three distinctive sources coexist and produce a wide range of geochemical signatures on the basalts collected; (1) a N-MORB source signature; (2) a transitional towards E-MORB source signature with a negative Nb anomaly indicating that some subduction related contamination still exists beneath the NFB (the basalts derived from this source might also have been called BABB previously); and (3) a source signature transitional towards E-MORB or OIB marking the start of an influence that increases towards the northern NFB as the Rotuma-Samoan hot spot lineament is approched. Around 17°S, on the Kaiyo station 4 site explored and sampled by the Nautile deep-sea submersible, as well as in the triple junction area, the same variability derived from three distinctive mantle sources is observed. Along the N160° segment, the three sources still coexist, but the influence of the E-MORB or OIB source (hot spot-related) increases, whereas the influence of the subduction-related source decreases. Along the Pandora-Rotuma ridge, the OIB-derived lava type is the only one present, the eventual contribution from other sources being completely diluted. Thus, the geochemistry of the NFB basalts is directly influenced by (1) the regional geodynamic environment, such as subduction zones and/or hot spot trails, (2) the geodynamical regime and stability of this type of back-arc system. The influence of the New Hebrides subduction, located some 500 km west of the active spreading ridge, is still perceptible, although weak, in the whole northern half of the NFB. However, this infuence is not directly linked to the presently active subduction, but originates rather from the partial melting of an upper mantle source that suffered subduction contamination during the clockwise rotation of the New Hebrides arc leading to the opening of the NFB during the past 12 Myr. In the northern NFB, many basalts result from the mixing of an N-MORB and a OIB source, similar to transitional alkalic lavas from oceanic intra-plate magmatism. This sources mixing increases northwards from 18°20′S to 12°S.

The North Fiji Basin basalts and their magma sources: Part II. Sr-Nd isotopic and trace element constraints

Abstract Sr-Nd isotope and trace element data are reported for basalts from the North Fiji Basin (NFB). NFB basalts are characterized by extreme variations in isotopic ratios and trace element abundances which are related to mantle heterogeneities. Values of 87 Sr 86 Sr for basalts from the northern segments, N160° and triple junction range from 0.7029 to 0.7041, whereas 143 Nd 144 Nd values vary from 0.51281 to 0.51313. Most of the basalts from these segments are characterized by strong relative enrichments in Rb, Ba, Sr, K, Nb, Ta La, Ce and Ti that are comparable to OIB components. The central segments, N15° and N-S have 143 Nd 144 Nd ratios between 0.51298 and 0.51363 and 87 Sr 86 Sr ratios between 0.7029 and 0.7033, and are depleted in large ion lithophile, light rare-earth and high field strength elements similar to N-MORB. Covariation between trace element and isotopic ratios among NFB basalts supports a model in which melts from the NFB rift system are derived by mixing of OIB-type and depleted N-MORB mantle components. The Sr-Nd isotopic and trace element variability indicates that the NFB basalt source is heterogeneous on the scale of individual melt batches.

Formation of the mid-fifteenth century Kuwae caldera (Vanuatu) by an initial hydroclastic and subsequent ignimbritic eruption

In the mid-fifteenth century, one of the largest eruptions of the last 10 000 years occurred in the Central New Hebrides arc, forming the Kuwae caldera (12x6 km). This eruption followed a late maar phase in the pre-caldera edifice, responsible for a series of alternating hydromagmatic deposits and airfall lapilli layers. Tuffs related to caldera formation (≈ 120 m of deposits on a composite section from the caldera wall) were emitted during two main ignimbritic phases associated with two additional hydromagmatic episodes. The lower hydromagmatic tuffs from the precaldera maar phase are mainly basaltic andesite in composition, but clasts show compositions ranging from 48 to 60% SiO2. The unwelded and welded ashflow deposits from the ignimbritic phases and the associated intermediate and upper hydromagmatic deposits also show a wide compositional range (60–73% SiO2), but are dominantly dacitic. This broad compositional range is thought to be due to crystal fractionation. The striking evolution from one eruptive style (hydromagmatic) to the other (magmatic with emission of a large volume of ignimbrites) which occurred either over the tuff series as a whole, or at the beginning of each ignimbritic phase, is the most impressive characteristic of the caldera-forming event. This strongly suggests triggering of the main eruptive phases by magma-water interaction. A three-step model of caldera formation is presented: (1) moderate hydromagmatic (sequences HD 1–4) and magmatic (fallout deposits) activity from a central vent, probably over a period of months or years, affected an area slightly wider than the present caldera. At the end of this stage, intense seismic activity and extrusion of differentiated magma outside the caldera area occurred; (2) unhomogenized dacite was released during a hydromagmatic episode (HD 5). This was immediately followed by two major pyroclastic flows (PFD 1 and 2). The vents spread and intense magma-water interaction at the beginning of this stage decreased rapidly as magma discharge increased. Subsequent collapse of the caldera probably commenced in the southeastern sector of the caldera; (3) dacitic welded tuffs were emplaced during a second main phase (WFD 1–5). At the beginning of this phase, magma-water interaction continued, producing typical hydromagmatic deposits (HD 6). Caldera collapse extended to the northern part of the caldera. Previous C14 dates and records of explosive volcanism in ice from the south Pole show that the climactic phase of this event occurred in 1452 A.D.

Ignimbrites of basaltic andesite and andesite compositions from Tanna, New Hebrides Arc

Tanna island is part of a large volcanic complex mainly subsided below sea-level. On-land, two series of hydroclastic deposits and ignimbrites overlie the subaerial remains of a basal, mainly effusive volcano. The ‘Older’ Tanna Ignimbrite series (OTI), Late Pliocene or Pleistocene in age, consists of ash flows and ash- and scoria-flow deposits associated with fallout tephra layers, overlain by indurated pumice-flow deposits. Phreatomagmatic features are a constant characteristic of these tuffs. The ‘younger’ Late Pleistocene pyroclastics, the Siwi sequence, show basal phreatomagmatic deposits overlain by two successive flow units, each comprising a densely welded layer and a nonwelded ash-flow deposit. Whole-rock analyses of 17 juvenile clasts from the two sequences (vitric blocks from the phreatomagmatic deposits, welded blocks, scoriaceous bombs and pumices from the ignimbrites) show basaltic andesite and andesite compositions (SiO2=53–60%). In addition, 296 microprobe analyses of glasses in these clasts show a wide compositional range from 51 to 69% SiO2. Dominant compositions at ∼54, 56, 58.5 and 61–62% SiO2 characterize the glass from the OTI. Glass compositions in the lower — phreatomagmatic — deposits from the Siwi sequence also show multimodal distribution, with peaks at SiO2=55, 57.5, 61–62 and 64% whereas the upper ignimbrite has a predominant composition at 61–62% SiO2. In both cases, mineralogical data and crystal fractionation models suggest that these compositions represent the magmatic signature of a voluminous layered chamber, the compositional gradient of which is the result of fractional crystallization. During two major eruptive stages, probably related to two caldera collapses, the OTI and Siwi ignimbrites represent large outpourings from these magmatic reservoirs. The successive eruptive dynamics, from phreatomagmatic to Plinian, emphasize the role of water in initiating the eruptions, without which the mafic and intermediate magmas probably would not have erupted.

Active Spreading and Hydrothermalism in North Fiji Basin (SW Pacific). Results of Japanese French Cruise Kaiyo 87

The aim of the Japanese-French Kaiyo 87 cruise was the study of the spreading axis in the North Fiji Basin (SW Pacific). A Seabeam and geophysical survey allowed us to define the detailed structure of the active NS spreading axis between 16° and 22° S and its relationships with the left lateral motion of the North Fiji Fracture Zone. Between 21° S and 18°10′ S, the spreading axis trends NS. From 18°10 S to 16°40 S the orientation of the spreading axis changes from NS to 015°. North of 16°40′ S the spreading axis trends 160°. These two 015° and 160° branches converge with the left lateral North Fiji fracture zone around 16°40′ S to define an RRFZ triple junction. Water sampling, dredging and photo TV deep towing give new information concerning the hydrothermal activity along the spreading axis. The discovery of hydrothermal deposits associated with living communities confirms this activity.