Slab-wedge Interface Research Articles

Variant across-forearc compositions of slab-fluids recorded by serpentinites: Implications on the mobilization of FMEs from an active subduction zone (Mariana forearc)

Serpentinite mud volcanism in the Mariana forearc provides a window into the shallow portions of an active subduction zone. Fluid–rock interactions and related mass transfers into the mantle wedge can be assessed by studying the trace element compositions of slab-derived fluids and serpentinized mantle wedge materials brought to the seafloor by the serpentinite mud volcanoes. We investigated variably serpentinized ultramafic clasts from the Yinazao, Fantangisña, and Asùt Tesoru mud volcanoes recovered on International Ocean Discovery Program Expedition 366 to examine the transfer of fluid-mobile elements (FMEs) from the slab to the wedge. These mud volcanoes sample the slab–wedge interface at depths of ~13–18 km and estimated temperatures of 80–250 °C. Our samples represent the serpentinized forearc and exhibit a multi-phase serpentinization history, as apparent from microfabrics, mineralogy, and in situ major and trace elemental analyses of distinct generations of serpentine. Initial hydration of the forearc mantle occurred under reducing conditions by Si-rich fluids. Early serpentine is characterized by generally high concentrations of Li, Sr, Rb, Cs, and Ba. Subsequent fluid–rock interactions were driven by Si-rich and FME-poor fluids and at later stages by Si- and FME-poor fluids in the mud volcano conduits, the latter of which resulted in the abundant formation of Fe-rich brucite. Iowaite and hematite indicate that less reducing conditions prevailed during the alteration of clasts after their emplacement at the seafloor. Concentrations of B are generally high but our dataset does not allow distinguishing slab- from seawater-derived B.Serpentine from the shallow-sourced Yinazao exhibits high Rb/Cs ratios of ≤37, highest concentrations of Li, but lowest Rb, Sr, Ba, and Cs contents. The serpentinizing fluids were derived from expulsion of sedimentary pore waters and by the breakdown of opal in the subducted sediments. Serpentine at the intermediate-sourced Fantangisña has Rb/Cs ratios of <10, similar Li, Sr, and Ba concentrations as Yinazao, but higher Rb and Cs contents. These patterns likely reflect dewatering and FME-release from clays in the subducted sediments. Fluids at the deeply sourced Asùt Tesoru as well originate from clay-breakdown, but increased concentrations of Rb, Sr, Cs, and Ba are further indicative of beginning dehydration of altered oceanic crust.Including data from the South Chamorro serpentinite mud volcano (18 km slab depth; Kahl et al., 2015, Lithos), we provide a detailed record of slab dehydration reactions at shallow forearc depths and the related mobilization of FMEs as well as their transport into the mantle wedge. Our study demonstrates that slab-derived fluids undergo extensive alteration during the interaction with mantle wedge peridotite. Pore waters from the serpentinite mud volcanoes hence provide incomplete insight into the processes at depth; fluid signatures at the slab–wedge interface as well as their across-forearc changes are best recorded in early hydration products such as serpentine.

H2O-rich mantle melting near the slab–wedge interface

To investigate the first melts of the mantle wedge in subduction zones and their relationship to primitive magmas erupted at arcs, the compositions of low degree melts of hydrous garnet lherzolite have been experimentally determined at 3.2 GPa over the temperature range of 925–1150 °C. Two starting compositions with variable H2O contents were studied; a subduction-enriched peridotite containing 0.61% Na2O, 0.16 K2O% (wt%) with 4.2 wt% H2O added (Mitchell and Grove in Contrib Mineral Petrol 170:13, 2015) and an undepleted mantle peridotite (Hart and Zindler in Chem Geol 57:247–267, 1986) with 14.5% H2O added (Till et al. in Contrib Mineral Petrol 163:669–688, 2012). Saturating phases include olivine, orthopyroxene, clinopyroxene, garnet and rutile. Melting extent is tracked from near solidus (~ 5 wt%) to 25 wt%, which is close to or beyond the point where clinopyroxene and garnet are exhausted. The beginning of melting is a peritectic reaction where 0.54 orthopyroxene + 0.17 clinopyroxene + 0.13 garnet react to produce 1.0 liquid + 0.88 olivine. The melt production rate near the solidus is 0.1 wt% °C−1 and increases to 0.3 wt% °C−1 over the experimentally studied interval. These values are significantly lower than that observed for anhydrous lherzolite (~ 1 wt% °C−1). When melting through this reaction is calculated for a metasomatized lherzolite source, the rare earth element characteristics of the melt are similar to melts of an eclogite, as well as those observed in many subduction zone magmas. Moreover, since rutile is stable up to ~ 8 wt% melting, the first melts of a hydrous lherzolite source could also show strong high field strength element depletions as is observed in many subduction zone lavas. The silicate melts measured at the lowest temperatures and melting extents (< 10 wt%) are high silica andesites (56–60 wt% SiO2) and contain very low Ca/Al and high alkalis. These deep low degree andesitic melts, if added to experimentally produced hydrous liquids from melting (20–25 wt%) of harzburgite residues at shallow pressures (1.0–1.2 GPa, Mitchell and Grove in Contrib Mineral Petrol 170:13, 2015), can match the compositional characteristics of primitive natural basaltic andesite and magnesian andesite lavas found globally in arcs. In addition to a silicate melt phase, there is a small amount of silicate dissolved in the H2O supercritical fluid that coexists with the silicate liquid and solids in our experiments. The composition of this fluid is in equilibrium with the Mg-rich minerals and it is granitic. The results presented here are used to develop a model for producing hydrous arc magmas. We hypothesize that mantle wedge melting produced by the flux of hydrous fluid from the slab occurs over a range of depths that begins at the base of the mantle wedge and ends at shallow mantle depths. These melts ascend and remain isolated until they mix in the shallow, hottest part of the mantle wedge. In this melting scenario, the metasomatic “slab melt” contributions to arc magmas is small (~ 5 wt%), but its effect on the alkali, REE and incompatible trace element budget of the derivative magmas is large and able to reproduce the trace elemental characteristics of the primitive andesites. Higher proportions of slab or sediment melt do not resemble primitive high magnesian arc andesites and basaltic andesites.

Hf isotope evidence for variable slab input and crustal addition in basalts and andesites of the Taupo Volcanic Zone, New Zealand

Crustal contamination complicates the identification of primary mantle-derived magma compositions in continental arcs. However, when crustal processes and components are well characterised, it is possible to extrapolate through continental arc magma compositional arrays towards the Hf and Nd isotope compositions of uncontaminated primary magmas. This is because of the similar behaviour of Hf and Nd during fractional crystallisation and mantle melting, and the subsequent limited variation in Hf/Nd in mantle-derived magmas and in many crustal lithologies, resulting in linear contamination trends for Hf-Nd isotopes. Here we present new Hf isotope data for a selection of volcanic rocks and crustal lithologies from the Taupo Volcanic Zone (TVZ), New Zealand and propose that the scatter in Hf-Nd isotopes indicates heterogeneity in the parental magmas prior to interactions with crustal lithologies. The observed variations likely represent variability in primary magma compositions as a result of different degrees of sediment addition at the slab-wedge interface. Coupled variations in isotopic composition, LILE/HFSE ratios (e.g. Rb/Zr and Ba/La) and SiO2 also clearly indicate that shallower level crustal interactions have occurred. Andesites from Ruapehu Volcano have more consistent parental magma compositions, and require greater amounts of a source sediment contribution. Notably, the compositions of older Ruapehu lavas can be modelled by interactions between mantle-derived magmas and lower crustal granulites, whereas younger lavas have probably interacted more with mid- to shallow crustal meta-sedimentary greywacke-argillite lithologies of the Permian to Cretaceous composite Torlesse Terrane. Hf-Nd isotopic compositions of meta-igneous granulite xenoliths from Mt. Ruapehu are consistent with previous interpretations that they are derived from oceanic crust that underlies the Torlesse meta-sediments. The results indicate that interactions with sediments at both the slab-wedge interface and in the lithosphere must be considered when evaluating trace element and isotopic variations in continental arcs.

Controls on the stability and composition of amphibole in the Earth’s mantle

Presented here is a suite of new experiments aimed at quantifying the effects of pressure, temperature, bulk composition, and H2O content on the stability and composition of amphibole in the Earth’s mantle. Experiments have been performed from 2 to 4 GPa and 950 to 1100 °C on fertile and depleted mantle compositions. H2O contents of most experiments are 0.65 wt%. In the fertile mantle composition, pargasitic amphibole is stable up to ~3.8 GPa at 1000 °C, approximately 0.5 GPa higher than any previous study. The upper stability limit of amphibole in depleted mantle is 0.7 GPa and 40 °C lower than in fertile mantle. The addition of 3 wt% H2O to fertile mantle destabilizes amphibole by 0.5 GPa and 40 °C relative to the 0.65 wt% H2O experiments. Compared to existing experiments on amphibole stability, these experiments indicate that pargasitic amphibole may be stable in mantle lithosphere to almost 4 GPa (0.5 GPa higher (15 km deeper) than previously thought). The extremely strong destabilizing effect of H2O suggests that deeper portions of the strongly fluid-fluxed mantle wedge may be amphibole-free even at low temperatures near the slab–wedge interface. The molar alkali content of amphibole is shown to be a linear function (R 2 = 0.98) of pressure and temperature and is relatively insensitive to bulk compositional differences between fertile and depleted mantle. This relationship is used to produce an empirical thermobarometer for pargasite-bearing spinel and garnet lherzolites. Comparison to existing experimental data shows that this thermobarometer has predictive ability over the pressure range of 1–4 GPa. Comparisons with pressure–temperature estimates of garnet + amphibole peridotites further corroborate the applicability of this thermobarometer for natural samples. Pressure estimates are presented for four examples of metasomatized spinel peridotites otherwise lacking pressure information, and future avenues for refinement of the thermobarometer are discussed.

Introduction to the special issue on SFB 574 “Volatiles and fluids in subduction zones: climate feedback and trigger mechanisms for natural disasters”

combine field work on land with marine cruises probing the seafloor. During the course of the SFB 574, many different aspects of subduction zone processes were investigated. Geophysical investigations identified and quantified the input of water through hydration of the bend-faulted subducting plate as far as mantle depths. This process and water release from the subducting slab deeper in the subduction zone was investigated by numerical modeling. These studies highlighted the important role of hydration and dehydration of mantle rocks for the global water cycle. Subducted oceanic fracture zones are another conduit for water transported deep into the subduction zone. Compactional and thermal dehydration of the subducted sediment layer affects the strength of interplate coupling and the depth and lateral extension of the seismogenic zone where disastrous earthquakes are generated. The fluids, largely generated by clay mineral transformation, are expelled through the forearc by splay faults. At the seafloor, the sites of cold seeps, often associated with mud volcanoes, are populated by biota that control the carbon transfer to the ocean. The most prominent manifestations of cold seeps are authigenic carbonates that form from anaerobic oxidation of methane and which serve as archives of forearc processes. The largest cold seep emissions occur at faults generated by the subduction of volcanic seamounts. Petrological analyses of ancient, exhumed subduction zone metamorphic rocks revealed the important role of pervasive, typically channelized fluids in high-pressure metamorphic reactions of dewatering subducted igneous crust and mantle. Slab-derived fluids hydrate the mantle wedge along the slab–wedge interface and form a subduction channel in which mixing of different fluids and fluid–rock interaction causes metasomatic overprinting. The element redistributions associated with the liberation of fluids from Sonderforschungsbereiche (SFBs) are a successful funding model in use by the German Science Foundation (DFG) for over 30 years to strengthen basic research first locally at universities and later also supra-regionally by including academic institutions at different cities and states. Literally translated, SFB means “special research area” that comprises research that complements but does not duplicate research at participating institutions and departments. The English terminology used by the DFG is “Collaborative Research Centre,” which better describes the expected approach by emphasizing collaboration and interdisciplinary efforts in such a way that the overall result is better than the sum of individual results. The SFB 574 had united more than 70 scientists with expertise in structural geology, geophysics, sedimentology, geochemistry, empirical and experimental petrology, volcanology, and biology for 11 years (2001–2012). The overarching theme addressed the role of volatiles in subduction zone tectonic, hydrological, metamorphic and magmatic processes, and resulting hazards. The main areas of research were the subduction zones of Central America and southern Chile. Both extend across shorelines from deepsea trenches to arc-volcano summits and thus required to

Rheological contrast between serpentine species and implications for slab–mantle wedge decoupling

Serpentine is considered the main hydrous mineral to cause significant weakening at the slab–mantle interface; however, despite the possible presence of lizardite/chrysotile (liz/ctl) and antigorite (i.e., low- and high-temperature serpentine species, respectively) at the base of the mantle wedge, the differences in rheological properties between these two serpentine types are poorly understood. To investigate the effect of serpentine speciation on slab–mantle decoupling, we performed a series of two-layer shear deformation experiments on liz/ctl, antigorite, and olivine at P = 1 GPa and T = 250–300 °C, using a Griggs-type solid-medium apparatus. With increasing shear strain (γ up to ~ 5), the liz/ctl and antigorite samples evolved to deform predominantly by slip or glide on the (001) basal plane, whereas the olivine sample deformed by a combination of brittle and crystal-plastic mechanisms. The contrasting amount of rotation of strain markers between layers demonstrates that the shear strain of liz/ctl is significantly lower than that of antigorite, possibly arising from differences in interlayer bond strength. The strain contrasts between serpentines and olivine were γliz/ctl/γol = 9–12 and γatg/γol = 1–2. Our results indicate that assuming a thin, serpentinized layer upon a subducting slab, significant decoupling can occur only where liz/ctl is stable in the layer. Thus, the distribution of serpentine species in the hydrated mantle wedge is a critical factor that controls the degree of decoupling along the slab–wedge interface, at depths below the seismogenic zone.

Deformation and mantle flow beneath the Sangihe subduction zone from seismic anisotropy

Subduction of oceanic lithosphere is the most direct feedback between the Earth’s surface and deep interior. However, the detail of its interaction with the broader convecting mantle is still unclear. Mantle flow around subduction zones can be constrained using seismic anisotropy, but despite many such studies, a simple global picture is lacking. The Sangihe subduction zone (where the Molucca Sea microplate is subducting westward beneath the Eurasian plate) is part of the tectonically complex Sulawesi–Philippine region, and an ideal natural laboratory to study complex subduction processes. We investigate the anisotropic structure of the Sangihe subduction zone with shear wave splitting measurements of local S and SKS phases at two stations (MNI in Sulawesi, DAV in the Philippines), as well as downgoing S phases at five stations at teleseismic distances. Combining different phases allows a better vertical resolution of anisotropic fabrics than is possible with a single phase. The broad depth distribution of local events (∼60–630km) allows us to observe a change in splitting behaviour at ∼380km depth: above, fast directions (ϕ) are trench-parallel and delay times (δt) are ∼0.34–0.53s with no increase with depth. We suggest this anisotropy is caused by aligned cracks, possibly melt-filled beneath the volcanic arc, and fossil anisotropy in the overriding plate. Below ∼380km, ϕ is predominantly trench-normal and δt are slightly higher (∼0.53–0.65s). As no correlation is observed with inferred distance travelled inside the slab, we attribute this anisotropy to shear layers atop the slab, which are coherent from ∼200 to 400km depth and perhaps extend into the transition zone. SKS and source-side measurements show larger δt (∼1.53 and 1.33s, respectively) and trench-parallel ϕ. Since these phases predominantly sample sub-slab mantle, we consider along-strike lateral flow associated with the double-sided subduction of the Molucca Sea microplate to be the most likely explanation. We thus infer three dominant regions of anisotropy at the Sangihe subduction zone: one within the overriding lithosphere, one along the slab–wedge interface, and one below the subducting Molucca Sea slab. The mantle wedge above 200km depth and the slab itself do not seem to contribute notably to the measured anisotropy. This study demonstrates the insight seismic anisotropy can provide into mantle dynamics even in tectonically complex subduction systems.

Changes in dip of subducted slabs at depth: Petrological and geochronological evidence from HP–UHP rocks (Tianshan, NW-China)

High-resolution seismic imaging has previously revealed sudden changes in the dip of subducted oceanic plates. This ‘kinking’ feature is crucial because in many subduction zones it is thought to coincide with the disappearance of the low-velocity layers associated with subducted oceanic crust. In this study, we present petrological evidence for this phenomenon derived from oceanic blueschist- and eclogite-facies rocks from the Chinese Tianshan. The investigated samples span a large range of peak metamorphic conditions, with temperatures between 330 and 580 °C at 1.5 to 2.3 GPa. Such variable peak metamorphic conditions, together with the intimate interlayering of high- and ultrahigh-pressure rocks, have also been reported from other Tianshan localities. These observations suggest that the rocks were derived from varying depths within the subduction zone and then juxtaposed during exhumation in the subduction channel. Multi-point Lu–Hf isochrons from four high-pressure rocks yield consistent garnet-growth ages of 313 ± 12, 315.8 ± 2.9, 313.9 ± 4.8, and 315.2 ± 1.6 Ma. These, in conjunction with the ~ 311 Ma cluster of published 40Ar– 39Ar and Rb–Sr white mica ages from the same vicinities imply fast exhumation rates. The Lu–Hf ages confirm that the eclogite-facies metamorphism of the Tianshan high-pressure rocks occurred during a single subduction event in the Late Carboniferous. However, the previously reported peak metamorphic P-T estimates from UHP metasediments and eclogites all lie on a lower geothermal gradient – and thus on a colder P-T path at the slab-wedge interface – than the HP eclogites and meta-volcaniclastic rocks from this study. This suggests that the slab-subduction angle steepened sharply at a depth greater than approximately 90 km, i.e., between the depths at which the HP and UHP rocks equilibrated. In addition to the negative buoyancy of mafic UHP rocks in subduction zones, the kinking may act as a geometric hindrance to the exhumation of such rocks, thereby explaining their rarity.

U-series disequilibria in Kick’em Jenny submarine volcano lavas: A new view of time-scales of magmatism in convergent margins

We present data for U and its decay series nuclides 230Th, 226Ra, 231Pa, and 210Po for 14 lavas from Kick’em Jenny (KEJ) submarine volcano to constrain the time-scales and processes of magmatism in the Southern Lesser Antilles, the arc having the globally lowest plate convergence rate. Although these samples are thought to have been erupted in the last century, most have ( 226Ra)/( 210Po) within ±15% of unity. Ten out of 14 samples have significant 226Ra excesses over 230Th, with ( 226Ra)/( 230Th) up to 2.97, while four samples are in 226Ra– 230Th equilibrium within error. All KEJ samples have high ( 231Pa)/( 235U), ranging from 1.56 to 2.64 and high 238U excesses (up to 43%), providing a global end-member of high 238U and high 231Pa excesses. Negative correlations between Sr, sensitive to plagioclase fractionation, and Ho/Sm, sensitive to amphibole fractionation, or K/Rb, sensitive to open system behavior, indicate that differentiation at KEJ lavas was dominated by amphibole fractionation and open-system assimilation. While ( 231Pa)/( 235U) does not correlate with differentiation indices such as Ho/Sm, ( 230Th)/( 238U) shows a slight negative correlation, likely due to assimilation of materials with slightly higher ( 230Th)/( 238U). Samples with 226Ra excess have higher Sr/Th and Ba/Th than those in 226Ra– 230Th equilibrium, forming rough positive correlations of ( 226Ra)/( 230Th) with Sr/Th and Ba/Th similar to those observed in many arc settings. We interpret these correlations to reflect a time-dependent magma differentiation process at shallow crustal levels and not the process of recent fluid addition at the slab–wedge interface. The high 231Pa excesses require an in-growth melting process operating at low melting rates and small residual porosity; such a model will also produce significant 238U– 230Th and 226Ra– 230Th disequilibrium in erupted lavas, meaning that signatures of recent fluid addition from the slab are unlikely to be preserved in KEJ lavas. We instead propose that most of the 238U– 230Th, 226Ra– 230Th, and 235U– 231Pa disequilibria in erupted KEJ lavas reflect the in-growth melting process in the mantle wedge (reflecting variations in U/Th, daughter–parent ratios, fO 2, and thermal structure), followed by modification by magma differentiation at crustal depths. Such a conclusion reconciles the different temporal implications from different U-series parent–daughter pairs and relaxes the time constraint on mass transfer from slab to eruption occurring in less than a few thousand years imposed by models whereby 226Ra excess is derived from the slab.

Imaging a steeply dipping subducting slab in Southern Central America

Subduction of oceanic lithosphere leads to large changes in seismic velocities that are observable through converted and scattered teleseismic waves. In this study, we use teleseismic data from a dense broadband seismic experiment in Central America to image a slab that dips > 60°. Although strongly bent, up going P-to- S mode conversions show a strong signal from 60–200 km depth; reverberations do not reach the surface because the slab dips steeply here. The subducted oceanic plate is successfully imaged to 200 km depth in Nicaragua where station geometry is adequate, but only at periods near 6 s at depths > 100 km, compared with 1–2 s periods for shallower structure. The long-period nature of the deeper conversion indicates a gradational boundary rather than a discontinuity, probably marking the thermal boundary layer at the slab-wedge interface. A sharper discontinuity is imaged near the top of the slab at depths < 80 km, probably denoting the Cocos plate Moho. Seismicity lies above the boundary at depths < 80 km, and below the inferred thermal boundary at greater depth. The change of character of the discontinuity with depth may reflect partial eclogitization of the subducting crust and associated reduction of the velocity contrast at the Moho, with seismicity lying within subducting crust but the dominant velocity change migrating out of it with increasing depth. In the upper plate, images map the Moho in both Costa Rica and Nicaragua, with Moho conversions weakest directly beneath the arc. Complex dipping structure at 25–50 km depth beneath the Nicaragua forearc perhaps reveals the deep roots of an accreted arc terrane.

Kinematic variables and water transport control the formation and location of arc volcanoes

The processes that give rise to arc magmas at convergent plate margins have long been a subject of scientific research and debate. A consensus has developed that the mantle wedge overlying the subducting slab and fluids and/or melts from the subducting slab itself are involved in the melting process. However, the role of kinematic variables such as slab dip and convergence rate in the formation of arc magmas is still unclear. The depth to the top of the subducting slab beneath volcanic arcs, usually approximately 110 +/- 20 km, was previously thought to be constant among arcs. Recent studies revealed that the depth of intermediate-depth earthquakes underneath volcanic arcs, presumably marking the slab-wedge interface, varies systematically between approximately 60 and 173 km and correlates with slab dip and convergence rate. Water-rich magmas (over 4-6 wt% H(2)O) are found in subduction zones with very different subduction parameters, including those with a shallow-dipping slab (north Japan), or steeply dipping slab (Marianas). Here we propose a simple model to address how kinematic parameters of plate subduction relate to the location of mantle melting at subduction zones. We demonstrate that the location of arc volcanoes is controlled by a combination of conditions: melting in the wedge is induced at the overlap of regions in the wedge that are hotter than the melting curve (solidus) of vapour-saturated peridotite and regions where hydrous minerals both in the wedge and in the subducting slab break down. These two limits for melt generation, when combined with the kinematic parameters of slab dip and convergence rate, provide independent constraints on the thermal structure of the wedge and accurately predict the location of mantle wedge melting and the position of arc volcanoes.

Metamorphic evolution of garnet‐bearing ultramafic rocks from the Gongen area, Sanbagawa belt, Japan

AbstractGarnet‐bearing ultramafic rocks including clinopyroxenite, wehrlite and websterite locally crop out in the Higashi‐akaishi peridotite of the Besshi region in the Cretaceous Sanbagawa metamorphic belt. These rock types occur within dunite as lenses, boudins or layers with a thickness ranging from a few centimetres to 1 metre. The wide and systematic variation of bulk‐rock composition and the overall layered structure imply that the ultramafic complex originated as a cumulate sequence. Garnet and other major silicates contain rare inclusions of edenitic amphibole, chlorite and magnetite, implying equilibrium at relatively low P–T conditions during prograde metamorphism. Orthopyroxene coexisting with garnet shows bell‐shaped Al zoning with a continuous decrease of Al from the core towards the rim, consistent with rims recording peak metamorphic conditions. Estimated P–T conditions using core and rim compositions of orthopyroxene are 1.5–2.4 GPa/700–800 °C and 2.9–3.8 GPa/700–810 °C, respectively, implying a high P/T gradient (&gt; 3.1 GPa/100 °C) during prograde metamorphism. The presence of relatively low P–T conditions at an early stage of metamorphism and the steep P/T gradient together trace a concave upwards P–T path that shows increasing P/T with higher T, similar to P–T paths reported from other UHP metamorphic terranes. These results suggest either (1) down dragging of hydrated mantle cumulate parallel to the slab–wedge interface in the subduction zone by mechanical coupling with the subducting slab or (2) ocean floor metamorphism and/or serpentinization at early stage of subduction of oceanic lithosphere and ensuing HP–UHP prograde metamorphism.

Sulfur isotope systematics of basaltic lavas from Indonesia: implications for the sulfur cycle in subduction zones

We report sulfur isotope compositions of basaltic and basaltic andesite lavas from selected volcanoes in the Indonesian arc system covering the spectrum from low-K tholeiitic to high-K calc–alkaline compositions. The results of 25 samples from seven volcanoes, which are associated with different subduction regimes, show a range in δ34S values of +2.0–+7.8‰ (VCDT) with an average of +4.7±1.4‰ (1σ). Averages and within-suite variations of two larger sets of samples from Batur and Soputan volcanoes (+4.2±1.3‰ with n=9 and +5.7±1.4‰ with n=7, respectively) are comparable to those of the entire sample set. Sulfur concentrations are low (mostly between 2 and 74 ppm, average=19 ppm) and do not show correlations with sulfur isotope composition and whole-rock chemistry, or systematic changes with time in any of the lava suites. From model calculations we infer that basaltic magmas will undergo sulfur isotope fractionation during degassing, most commonly towards lower δ34S values, but that the extent is limited at P–T conditions and oxidation states of interest. Hence, δ34S signatures of basaltic lavas will generally be within a few permil from primary magmatic values, even in cases of extensive sulfur loss. Consequently, magmas in the Indonesian arc system originate from mantle sources that are enriched in 34S relative to MORB and OIB sources and are likely to have δ34S values of about +5–+7‰. The enrichment in 34S is considered to reflect addition of slab-derived material, presumably from sediments rather than altered oceanic crust, with fluids being the most likely transport medium. Absence of correlation between δ34S values of Indonesian basalts and chemical proxies for source components or processes at the slab–wedge interface suggests that sulfur isotopes are relatively insensitive to variations in subduction setting and dynamics. This is supported by the modest range in δ34S of the Indonesian volcanoes studied despite significant variations in the nature and amount of subducted material, and by the similarity with average 34S enrichments in other oceanic arc systems.

Petrology and P–T evolution of garnet peridotites from central Sulawesi, Indonesia

Alpine‐type orogenic garnet‐bearing peridotites, associated with quartzo‐feldspathic gneisses of a 140–115 Ma high‐pressure/ultra‐high‐pressure metamorphic (HP‐UHPM) terrane, occur in two regions of the Indonesian island of Sulawesi. Both exposures are located within NW–SE‐trending strike–slip fault zones. Garnet lherzolite occurs as &lt;10 m wide fault slices juxtaposed against Miocene granite in the left‐lateral Palu‐Koro (P‐K) fault valley, and as 10–30 m wide, fault‐bounded outcrops juxtaposed against gabbros and peridotites of the East Sulawesi ophiolite within the right‐lateral Ampana fault in the Bongka river (BR) valley. Six evolutionary stages of recrystallization can be recognized in the peridotites from both localities. Stage I, the precursor spinel lherzolite assemblage, is characterized by Ol+Cpx+Opx±Prg‐Amp ± Spl±Rt±Phl, as inclusions within garnet cores. Stage II, the main garnet lherzolite assemblage, consists of coarse‐grained Ol+Opx+Cpx+Grt; whereas finer‐grained, neoblastic Ol+Opx+Grt+Cpx±Spl±Prg‐Amp±Phl constitutes stage III. Stages IV and V are manifest as kelyphites of fibrous Opx+Cpx+Spl in inner coronas, and Opx+Spl+Prg‐Amp±Ep in outer coronas around garnet, respectively. The final (greenschist facies) retrogressive stage VI is accompanied by recrystallization of Serp+Chl±Mag±Tr±Ni sulphides±Tlc±Cal. P–T conditions of the hydrated precursor spinel lherzolite stage I were probably about 750 °C at 15–20 kbar. P–T determinations of the peak stage IIc (from core compositions) display considerable variation for samples derived from different outcrops, with clustering at 26–38 kbar, 1025–1210 °C (P‐K &amp; BR); 19–21 kbar, 1070–1090 °C (P‐K), and 40–48 kbar, 1205–1290 °C (BR). Stage IIr (derived from rim compositions) generally records decompression of around 4–12 kbar accompanied by cooling of 50–240 °C from the IIc peak stage. Stage III, which post‐dates a phase of ductile deformation, yielded 22±2 kbar at 750±25 °C (P‐K) and 16±2 kbar at 730±40 °C (BR). The granulite–amphibolite–greenschist decompression sequence reflects uplift to upper crustal levels from conditions of 647–862 °C at P=15 kbar (stage IV), through 580–635 °C at P=10–12 kbar (stage V) to 350–400 °C at P=4–7 kbar (stage VI), respectively, and is identical to the sequence recorded in associated granulite, gneiss and eclogite. Sulawesi garnet peridotites are interpreted to represent minor components of the extensive HP‐UHP (peak P &gt;28 kbar, peak T of c. 760 °C) metamorphic basement terrane, which was recrystallized and uplifted in a N‐dipping continental collision zone at the southern Sundaland margin in the mid‐Cretaceous. The low‐T , low‐P and metasomatized spinel lherzolite precursor to the garnet lherzolite probably represents mantle wedge rocks that were dragged down parallel to the slab–wedge interface in a subduction/collision zone by induced corner flow. Ductile tectonic incorporation into the underthrust continental crust from various depths along the interface probably occurred during the exhumation stage, and the garnet peridotites were subsequently uplifted within the HP‐UHPM nappe, suffering a similar decompression history to that experienced by the regional schists and gneisses. Final exhumation from upper crustal levels was clearly facilitated by entrainment in Neogene granitic plutons, and/or Oligocene trans‐tension in deep‐seated strike–slip fault zones.

HFSE Mobility at the Slab-Wedge Interface: Hafnium-Isotope Evidence from the East Sunda-Banda Arc

Models describing trace-element behaviour in the petrogenesis of arc volcanics generally assume that high-field-strength elements (HFSE) are relatively immobile during element transfer between the subducted slab and the overlying mantle wedge. However, recent studies in the East Sunda-Banda Arcs (Vroon et al., t993; Hoogewerff et al., 1997) modify this view, and suggest that conditions in subduction settings can be favourable for the mobility of a wide range of slab-derived elements, including the HFSE-group. Sr-Nd-Pb isotopic signals in the ESA-BA are strongly dominated by subducted sediments of continental origin (SCM). Correlations between radiogenic isotopes and HFSE ratios provide indications for (1) an SCM-derived contribution to the HFSE budgets, and (2) increased mobility of HFSE down the dip of the subduction zone as a result of changes of the transfer medium from hydrous fluid to siliceous melt. Here, we present Hf isotope data for volcanic centres distributed along and across the ESA-BA, and use the results to trace a slab-derived origin of HFSE in the magmas.

U-series, SrNdPb isotope and trace-element systematics across an active island arc-continent collision zone: Implications for element transfer at the slab-wedge interface

We present U-series, SrNdz.sbnd;Pb isotope and trace-element results of a regional study of geochemical systematics across an island arc-continent collision zone in the East Sunda Arc of Indonesia. Samples from four active volcanoes exhibit a striking compositional range from low-K tholeiitic to ultrapotassic, but all are characterised by high 87 Sr 86 Sr (0.7053–0.7067), radiogenic lead isotope ratios ( 206 Pb 204 Pb = 18.99–19.15), low ( 230 Th) ( 232 Thz) (0.66–0.85), and low 143 Nd 144 Nd (0.51255–0.51272), except for high 143 Nd 144 Nd (>0.51286) at the volcanic front. Low ( 230 Th) ( 232 Th) ratios are also found in terrigenous sediments in front of the arc, which, in combination with Srz.sbnd;Ndz.sbnd;Pb isotopic constraints, indicates that subducted continental material contributes to magma sources in this arc sector. The volcanoes close to the trench show a large excess of 238U over . 230Th (up to 80%) and of 226Ra over 230Th (up to 800%). In addition, they are enriched in elements thought to be mobile in hydrous fluids during slab-wedge transfer, such as Ba, Pb, and Sr. In contrast, Uz.sbnd;Thz.sbnd;Ra systematics are close to equilibrium in the volcanoes behind the front. Abundance patterns of incompatible trace elements in these rocks are similar to those of the terrigenous sediments, so that, in comparison with the arc-front lavas, they possess low Ba/La, Ba/Th, La/Th, Pb/Ce, and Zr/Nb. Higher concentration levels and less interelement fractionation form conspicuous differences with the front volcanics. Our combined isotopic and trace element data are consistent with three-component mixing whereby a slab-derived hydrous fluid and a siliceous melt are both added to the sub-arc mantle source. The hydrous fluid largely controls the input in the shallow part of the subduction zone, whereas the siliceous melt dominates the flux at deeper levels. Sedimentary material is considered to be the primary source of both. The large U-Th-Ra disequilibria at the front suggests that element transfer is a currently active process associated with present-day subduction of continental material.

Correction to “Physical model of source region of subduction zone volcanics” by J. Huw Davies and D. J. Stevenson

where I•=•/Vol2/2•:. Repeating he xample in appendix; i.e., where/•=5 km,/2=5 km V0=7 cm/yr, T•=I.0, To=0.0, and •:= 10 -6 m 2 s 'z, we find that Tin=0.73. This shows that such a high velocity impinging on a stationary region can lead to high temperatures at interface. In contrast, mantle flow in a subduction zone impinges on a moving slab; it is downward advection of cold lithosphere by subducting slab that is critical in reducing interface temperatue in this region. The above error does not affect any of results or conclusions in body of paper, since body of paper is independent of Appendix A. In section A.2 of Appendix A, axis convention should read the x axis is perpendicular to slab (negative out into wedge). Figure 7c has a younger oceanic plate than Figure 7d, so it is not strictly an enlargement of Figure 7d, as stated in figure caption. The temperatures in regions of interest, slabwedge interface and mantle wedge, are virtually unaffected by age of subducting plate as stated in paper. The temperature of subducting plate though does depend upon its age, but since temperatures in slab were subsolidus, their exact values had no bearing on discussion of magmatism and conclusions. We would like to acknowledge K. Bose, who pointed out convention error in Appendix A.2 and error in figure caption of Figure 7.

Thermal effects of ridge subduction

The temperature structure of a subduction zone during subduction of an oceanic spreading ridge has been calculated by a finite-difference procedure. The plates preceding and following the ridge crest into the trench had convergence rates of 90 and 20 mm/yr, respectively, corresponding to a spreading half-rate for the ridge of 35 mm/yr and to the velocity balance for a simple three-plate model of ridge subduction. Abrupt temperature increases in the range 50–500°C occur at shallow locations(≤35km) in the outer 40 km of the arc-trench gap within a few million years of the time that the ridge crest enters the trench. Geological manifestations of these changes could include anomalous near-trench magmatism (gabbroic or granitic), termination of high-pressure metamorphism, enhancement of low-pressure metamorphism, and generation of hydrocarbons. Along the slab-wedge interface the depth of dehydration of hydrous phases in the subducted oceanic crust rapidly migrates upward as the ridge approaches. Consequently, substantial volumes of water may be released at depths of 50–75 km just before the ridge crest is subducted, and relatively shallow dehydration(<100km) may continue for 20–30 m.y., in contrast to the depth range of 95–135 km during the preceding interval of steady-state subduction. If water release at these greater depths is a requisite for normal arc magmatism, the shallowing of the dehydration front could produce a decrease or hiatus in this magmatism.