Subduction zones are arguably the most important geological features of our planet, where plates plunge into the deep, metamorphic reactions take place, large earthquakes happen and melting induces volcanism and creation of continental crust. None of these processes would be possible without the cycling of volatiles, and this session aims to explore their role in convergent margins. Questions to address include the following. Do Atlantic and Pacific subduction zones cycle volatiles in different ways? What dynamic or chemical roles are played by subducted fracture zones and plate bending faults? How do fluids and melts interact with the mantle wedge and overlying lithosphere? Why do some of the Earth’s largest mineral resources form in subduction settings? We aim to bring together geodynamicists, geochemists, petrologists, seismologists, mineral and rock physicists, and structural geologists to understand how plate hydration/slab dynamics/dehydration, and subsequent mantle wedge melting/fluid percolation, and ultimately melt segregation/accumulation lead to the diverse range of phenomena observed at convergence zones around the globe.

Includes Augustus Love Medal by Harro Schmeling
Invited Speaker: Nestor Cerpa (University of Montpellier, France)

Co-organized by GMPV2/SM6/TS7
Convener: Jeroen van Hunen | Co-conveners: Jenny Collier, Colin Macpherson, Andreas Rietbrock, Jamie Wilkinson
| Attendance Wed, 06 May, 14:00–15:45 (CEST)

Files for download

Download all presentations (24MB)

Chat time: Wednesday, 6 May 2020, 14:00–15:45

D1387 |
| Highlight
| Augustus Love Medal Lecture
Harro Schmeling

Melting at convergent plate boundaries

At divergent plate boundaries hot mantle upwelling is associated with abundant melt generation and volcanism. At convergent plate boundaries such as subduction zones and continental collision zones thick and cold plates feed mantle downwellings. Yet these "cold" regions also show abundant volcanic activity with mean volcanic output rates of almost similar order of magnitudes (White et al., 2006, G-cubed). Responsible melt generation mechanisms are addressed including a) volatile driven decrease of the solidus temperature, b) decompressional melting in the mantle wedge or in shallow asthenosphere associated with delamination, or c) increased radiogenic heating within thickened continental crust.     

Melt transport mechanisms

The above processes form partially molten regions. By which mechanism(s) does the melt segregate out of the melt source region and rise through the mantle or crust. The basic mechanism is two-phase flow, i.e. a liquid phase percolates through a solid, viscously deforming matrix. The corresponding equations and related issues such as compaction or effective matrix rheology are addressed. Beside simple Darcy flow, special solutions of the equations are addressed such as solitary porosity waves. Depending on the bulk to shear viscosity ratio of the matrix and the non-dimensional size of these waves, they show a variety of features: they may transport melt over large distances, or they show transitions from rising porosity waves to diapiric rise or to fingering. Other solutions of the equations lead to channeling, either mechanically or chemically driven. One open question is how do such channels transform into dykes which have the potential of rising through sub-solidus overburden. A recent hypothesis addresses the possibility that rapid melt percolation may reach the thermal non-equilibrium regime, i.e. the local temperature of matrix and melt may evolve differently.  Once dykes have been formed they may propagate upwards driven by melt buoyancy and controlled by the ambient stress field. As another magma ascent mechanism diapirism is addressed.  

Modelling magmatic systems in thickened continental crust

Once basaltic melts rise from subducting slabs, they may underplate continental crust and generate silicic melts. Early dynamic models (Bittner and Schmeling, 1995, Geophys. J. Int.) showed that such silicic magma bodies may rise to mid-crustal depth by diapirism. More recent approaches (e.g. Blundy and Annan, 2016, Elements) emplace sill intrusions into the crust at various levels and calculate the thermal and melting effects responsible for the formation of mush zones. Recently Schmeling et al. (2019, Geophys. J. Int.) self-consistently modelled the formation of crustal magmatic systems, mush zones and magma bodies by including two-phase flow, melting/solidification and effective power-law rheology. In these models melt is found to rise to mid-crustal depths by a combination of compaction/decompaction assisted two-phase flow, sometimes including solitary porosity waves, and diapirism. An open question in these models is whether or how dykes may self-consistently form to transport the melts to shallower depth. First models which combine the two-phase flow crustal models with elastic dyke-propagations models (Maccaferri et al., 2019, G-cubed) are promising.      


How to cite: Schmeling, H.: Melting processes at convergent plate boundaries: from melt segregation, extraction to the formation of crustal magmatic systems , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-6912, https://doi.org/10.5194/egusphere-egu2020-6912, 2020.

D1388 |
Timm John, Esther Schwarzenbach, Jay Ague, and Jilei Li

One of the most pressing riddles of the subduction cycle to be solved is linked to the fO2 of the slab-released fluids. It is now well accepted that the fluids liberated during slab dehydration play a crucial role in subduction zone seismicity, element cycling, and arc magmatism. However, whether these slab fluids are oxidizing or reducing transport agents is poorly understood and thus, there is still a lot we need to understand. This is of particular importance for gaining a mechanistic view on the formation processes of economically important arc related ore deposits, which certainly require understanding of the behavior of redox sensitive mobilities of the relevant elements. In brief, while some field-based studies from the slab perspective are suggesting rather reduced conditions (e.g., based on sulfides and sulfur isotope work, ref. 1) others, mainly related to higher temperature systems (e.g., based on bulk-rock – rutile systems and molybdenum isotope work, ref. 2), are indicative of more oxidizing slab fluids. Especially for mélange-like structures developed at the plate interface, studies on sulfur-bearing minerals result in contrasting fO2 of the related slab fluids (ref. 3 vs ref. 4). It appears that at least during retrogression along the plate interface the reactively flowing fluids tend to have a more oxidizing potential (ref. 5). Interestingly, the prime fluid source of subducting slabs, i.e. dehydrating slab mantle serpentinites, is thought to release reduced fluids (ref. 6) but melt inclusions in arc volcanic rocks are often oxidized. Recent studies suggest that this is likely linked to fluid-rock interaction at local scales (ref. 7) and/or possibly within the magma reservoirs that comprise rather low-melt-fraction mush (ref. 8). This in turn would suggest that the slab fluids might change their fO2 during reactive intra-slab fluid flow, or would not need to be oxidized prior to melt inclusion entrapment and that the oxidizing potential of the fluids may be the result of magmatic processes during melt ascent in the arc. In this contribution we review the current state of knowledge, provide new ideas and models regarding channelized though reactive intra-slab fluid flow, and illustrate the next steps to unravel this exiting and thus far poorly understood topic of subduction zone element cycling.


1]        Li, J.-L., et al. (2020). Nature Communications. https://doi.org/10.1038/s41467-019-14110-4

2]        Chen, S., et al. (2019). Nature Communications. http://doi.org/10.1038/s41467-019-12696-3

3]        Schwarzenbach, E.M., et al. (2018). Scientific Reports 8, 15517.

4]        Walters, J. B., et al. (2019). Geochemistry Geophysics Geosystems, 286, 185–28. http://doi.org/10.1029/2019GC008374

5]        Li, J.-L., et al. (2016). Contributions to Mineralogy and Petrology, 171:72. http://doi.org/10.1007/s00410-016-1284-2

6]        Piccoli, F., et al. (2019). Scientific Reports, 1–7. http://doi.org/10.1038/s41598-019-55944-8

7]        Tollan, P. & Hermann, J. (2019). Nature Geoscience 12, 667–671.

8]        Jackson, M. D., et al. (2018). Nature, 564, 405–409. http://doi.org/10.1038/s41586-018-0746-2


How to cite: John, T., Schwarzenbach, E., Ague, J., and Li, J.: On the fO2 of slab fluids in subduction zone systems, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-13454, https://doi.org/10.5194/egusphere-egu2020-13454, 2020.

D1389 |
Nicolas Brantut and Emmanuel David

High Vp/Vs ratio is a commonly used diagnostic for elevated fluid pressure when interpreting seismological data. The physical basis for this interpretation comes from rock physical data and models of isotropic, cracked rocks. Here, we establish precise conditions under which this interpretation is correct, by using an effective medium approach for fluid-saturated rocks. While the usual result of an increasing Vp/Vs with increasing fluid-saturated porosity holds for crack-like pores, we find that Vp/Vs ratio is not always monotonically increasing with increasing fluid content if the porosity shape deviates from thin cracks, and if the initial Vp/Vs of the rock (without porosity) is already quite high. This is specifically the case of dehydrating rocks, where initial Vp/Vs may already be high (>1.9 for lizardite, for instance), and where the porosity created by the dehydration reaction may be in the form of elongated needles. The model predictions are supported by existing experimental data obtained during dehydration experiments in gypsum and lizardite, which both show a significant decrease in Vp/Vs as dehydration proceeds. Although no experimental data is yet availbale on antigorite, we make a prediction that antigorite dehydration may not lead to any strong increase in Vp/Vs ratio under typical subduction zone conditions. We present our theoretical results in the form of simple closed-form solution (valid asymptotically for a range of limiting cases), which should help guide the interpretation of Vp/Vs ratio from seismological data.

How to cite: Brantut, N. and David, E.: Vp/Vs ratio and dehydration reactions in subduction zones, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-18243, https://doi.org/10.5194/egusphere-egu2020-18243, 2020.

D1390 |
| Highlight
Nestor G. Cerpa, José Alberto Padrón-Navarta, and Diane Arcay

The subduction of water via lithospheric-mantle hydrous phases have major implications for the generation of arc and back-arc volcanism, as well as for the global water cycle. Most of the current numerical models use Perple_X [Connolly et al., 2009] to quantify water release from the slab and subsequent fluid migration in the mantle wedge. At UHP conditions, the phase diagrams generated with this thermodynamic code suggest that the breakdown of serpentine and chlorite leads to the near complete dehydration of the lithospheric mantle before reaching a 200-km depth. Laboratory experiments, however, have observed the stability of the 10-Å phase and the phase E in natural bulk compositions, which may hold moderate amounts of water, beyond the stability field of serpentine and chlorite [Fumagalli and Poli, 2005; Maurice et al., 2018]. Here, using 2D thermo-mechanical models, we explore to what extent the presence of these hydrous phases may favor a deeper subduction of water than those predicted by Perple_X.

We perform end-member models in terms of slab temperature and thickness of hydrated lithospheric mantle entering at trench. The computed geotherms within the uppermost subducted mantle show that the stability field of mantle hydrous phases around 600-800°C and 6-8 GPa is crucial for predictions of water fluxes. We point out that the lack of systematic experiments at these P-T conditions, as well as the absence of 10-Å and E phases in current thermodynamic databases, prevent accurate estimates of deep water transfers. We nonetheless build a phase diagram based on current experimental constraints that includes approximations of their stability field and qualitatively discuss the potential implications for fluid migration in the back-arc mantle wedge and water fluxes.

How to cite: Cerpa, N. G., Padrón-Navarta, J. A., and Arcay, D.: Uncertainties in the stability field of UHP hydrous phases (10-A phase and phase E) and deep-slab dehydration: potential implications for fluid migration and water fluxes at subduction zones, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-4783, https://doi.org/10.5194/egusphere-egu2020-4783, 2020.

D1391 |
Antonio Caracausi, Attilio Sulli, Maurizio Gasparo Morticelli, Marco Pantina, Paolo Censi, Vincenzo Stagno, Andrea Billi, Martina Coppola, and Claudia Romano

The central Mediterranean is a very complex area constituted by a puzzle of different lithosphere segments, whose geological evolution is controlled by the interaction between the European and African plates. Within this geological domain, the northern Sicily continental margin and adjacent coastal belt represent a link between the Sicilian chain and the Tyrrhenian extensional (back-arc) area in the north-south direction, whereas in the east-west direction a transition from a subduction type B (Ionian-Tyrrhenian) to a continental collisional system, subduction type A, (Sicilian-Maghrebian Chain) is recognized.

The structure of the lithosphere in this area is matter of a strong debate. Most uncertainties on the geologic evolution of the boundary between the European and African plate at depth rise from the lack, up to now, of constraints and clear evidence of geometry of the lithosphere down to the crust-mantle interface.

In order to investigate the regional crust-mantle tectonics, here we discuss recent deep seismic reflection data, gravimetric modelling, the regional fluid geochemistry coupled to the seismicity that clearly indicate presence, along this sector of the Central Mediterranean, of a hot mantle-wedging at about 28 km of depth. This wedge lies just below a thick-skinned deformed belt cut by a dense system of faults down to the Mohorovicic discontinuity.

We also discuss new geochemical data in mineralization (fluorite) of hydrothermal deposits along the main regional faults above the mantle wedge. The mineralization is strongly enriched in saline fluid inclusions that allowed high precision analyses of the trapped volatiles (H2O, CO2 and noble gases).

 Notwithstanding the region is far from any evidence of volcanism (Etna volcano and Aeolian Islands are in about 80km), the new geochemical data highlight the presence of mantle-derived volatiles that degas through the crust (e.g., He isotopes, up to 1.4Ra, Ra is the He isotopic ratio in atmosphere). An excess of heat sourced from the mantle characterizes the region. This is a rare case of occurrence of mantle volatiles together with heat in a collisional system.

The active regional seismicity indicates that the mantle fluids move from the mantle wedge to the surface, hence across the ductile crust that could be thought as a barrier to the advective transfer of fluids because of its low permeability on long time scales. Here we reconstruct the deep faults by the deep seismic reflection data that works as a network of pathways that actively sustains the advective transfer of the mantle fluids through the entire continental crust.     

Finally, the new geochemical data strongly supports that 1) the mantle wedge and possible associated magmatic intrusions as the source of the mantle volatiles outgassing in the region. A comparison of the noble gases isotopic signature of fluids coming from the mantle wedge and those emitted from the Mt Etna volcano furnish new constrain on the mantle composition below the central Mediterranean getting new constrains to the processes that controlled the geodynamic evolution of the central Mediterranean (i.e., delamination processes).

How to cite: Caracausi, A., Sulli, A., Gasparo Morticelli, M., Pantina, M., Censi, P., Stagno, V., Billi, A., Coppola, M., and Romano, C.: Mantle degassing in a collisional zone: Subduction types A & B in the Central Mediterranean , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-15868, https://doi.org/10.5194/egusphere-egu2020-15868, 2020.

D1392 |
Margarita Merkulova, Antoine Triantafyllou, and Bernard Charlier

Enrichment of subduction derived magmas in chalcophile and siderophile elements plays a key role in the formation of economically important ore deposits. Melting of mantle wedge peridotite induced by fluxing slab-derived fluids is the first step in the chemical transfer from the slab to melts (Spandler and Pirard, 2013). Changes in redox conditions, sulfur content, amount of fluid, and P-T conditions affect the elements partitioning between melts and residual phases (Dale et al., 2009; Li and Audetat, 2012; Perchuk et al., 2018; Rielli et al., 2018). However, to date there is a few systematic data on mineral-melt partitioning for economically important elements during partial melting of peridotite in mantle wedge conditions.

We determined experimentally mineral-melt partition coefficients for a range of chalcophile and siderophile elements (V, Co, Cu, Zn, As, Se, Mo, Ru, Rh, Pd, Ag, Cd, Sb, Re, Os, Ir, Pt, Au, Tl, Bi) at different P-T conditions. Slightly serpentinized peridotite (3 wt.% H2O; Debret et al., 2013) was used as a starting material for all experiments. A set of experiments with 1 wt.% of FeS added to the peridotite was also performed in order to study the effect of S on element partitioning. Starting materials were doped with 100-200 ppm of targeted elements. The experiments were performed at pressures of 1-2 GPa and at temperatures between 1100 and 1300°C in end-loaded 0.5” piston-cylinder apparatus in the newly established high-pressure laboratory at the University of Liege (Belgium). In order to avoid Fe-loss, chemical reduction and volatile loss of experimental charges, double Au80Pd20 capsules pre-saturated with Fe were used. Major and trace element composition in synthesized experimental products were measured by electron microprobe and LA-ICP-MS respectively.

In this study, we report a wide range of partition coefficients determined between coexisting silicate, oxide, sulfide minerals and melt as a function of P-T and S content. The results provide further insights into mobility of economically important elements during genesis of ore-forming magmas in subduction settings.


How to cite: Merkulova, M., Triantafyllou, A., and Charlier, B.: Partitioning of chalcophile and siderophile elements during partial melting of serpentinized peridotite in subduction settings, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-6946, https://doi.org/10.5194/egusphere-egu2020-6946, 2020.

D1393 |
Anshuman Giri and Rajagopal Anand

The archaean greenstone belts, dominated by mafic to felsic volcanic rocks followed by younger granitic intrusions occurs associated with volcano-sedimentary sequences. The Dharwar Super group (2600 to 2900 Ma) of rocks in western Dharwar craton, underlie the older TTG gneisses. The Shimoga greenstone belt (SGB) of WDC constitute the basal polymictic conglomerate along with quartzite, pyroclastic rocks, carbonaceous rocks, greywacke-argillite sequences with a thick pile of mafic and felsic metavolcanic rocks (BADR). These rocks are suffered from greenschist to lower amphibolite grade of metamorphism. The Medur metavolcanic volcanic rocks give an age of 2638 ± 66 Ma (1), whereas the Daginakatte felsic volcanic rocks give an age of 2601 ± 6 Ma (2). The present studied age of 2638 ± 66 Ma, tells about the cessation of mafic magmatism in WDC. The metavolcanic rocks of the Medur formation are tholeiitic to calc-alkaline in nature. These rocks show flat to LREE enriched REE pattern with negative europium anomaly. And also show enrichment in LILE and depletion in HFSE elements with significant Nb-Ta anomaly. The geochemical and the isotope data suggest the involvement of partial melting of the depleted mantle by the slab components and assimilation fractional crystallization (AFC) processes for the magma generation. The SGB metavolcanic rocks have 143Nd/144Nd ratios (0.511150 to .513076) and εNd values of -3.1 to -5.5 and the negative εNd values  for the rocks is due to the crustal contamination of the magma in a shallow marine subduction setting. The parental magmas were derived from melting in the mantle wedge fluxed by slab derived fluids and slab components followed by assimilation fractional crystallization (AFC) processes involving continental crust in an active continental margin.

  • (1) Giri et al., 2019. Lithos, 330-331, 177-193
  • (2) Trendall et al., 1997a. J. Geol. Soc. India, 50, 25-50.

How to cite: Giri, A. and Anand, R.: Geochemistry and isotope studies of the metavolcanic rocks of Shimoga greenstone belt, Western Dharwar craton - an effort to deduce the Petrogenesis., EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-21529, https://doi.org/10.5194/egusphere-egu2020-21529, 2020.

D1394 |
Dimitri Sverjensky and Simon Matthews

It is well documented that subducting slabs influence arc volcanics. Slab components are transferred to the mantle wedge by fluids and/or melts. Aqueous fluids released from the slab are thought to trigger partial melting in the mantle wedge and potentially influence the chemistry of the lavas that erupt in island arcs. Both fluids and melts from the slab have been proposed to transfer chemical elements to the mantle wedge. However, exactly how this occurs chemically and physically remains unclear. Recent progress in developing a Deep Earth Water model calibrated with experimental mineral and rock solubility data under sub-arc conditions now enables the chemical mass transfer from slab to mantle wedge to be predicted for comparison with natural samples.

            We report a new aqueous speciation model for Ti-species calibrated with experimental data Kessel and co-workers and Antignano and Manning that includes a neutral Ti-OH species, a Na-Ti-silicate anion, and a Ti-silicate-bicarbonate anion. The Ti-OH species is only important in almost pure water. However, the Na-Ti-silicate anion is important in high-silica fluids (e.g. in equilibrium with quartz or coesite-bearing mafic eclogites) but is overtaken in importance by the Ti-silicate-bicarbonate complex in CO2-bearing fluids.

            In the present study, we modeled the metasomatic reactions when a fluid in equilibrium with a mafic eclogite leaves a subducting slab and encounters lherzolite in the overlying mantle wedge. Initially, the mafic eclogitic fluid was in equilibrium with clinopyroxene, garnet, coesite, diamond, magnesite solid-solution, and rutile at 700°C and 4.0 GPa. Whilst the presence of CO2 enables the modelled fluid to carry 600 mg/kg H2O of nominally immobile Ti from the slab into the wedge, the fluid transports a factor of 30 more K. The fluid was then heated to 950°C and simultaneously reacted irreversibly with lherzolite containing 0.86 wt% K2O and 0.084 wt% TiO2. The resultant metasomatized peridotite consisted of olivine, orthopyroxene, clinopyroxene, and garnet to which phlogopite-rich biotite had been added, and from which the TiO2 component was subtracted. Overall, the metasomatism resulted in K-enrichment and Ti-depletion in the metasomatized part of the mantle wedge. The final fluid was enriched in Ti (2,830 mg/kg H2O) with lowered K (11,600 mg/kg H2O). Both the remaining fluid and metasomatized mantle may serve as sources of the elevated K/Ti ratios in arc volcanics relative to MORB.

How to cite: Sverjensky, D. and Matthews, S.: K and Ti metasomatism of the mantle wedge by fluids under sub-arc conditions, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-17987, https://doi.org/10.5194/egusphere-egu2020-17987, 2020.

D1395 |
Robert Allen, Benedikt Braszus, Saskia Goes, Andreas Rietbrock, and Jenny Collier and the The VoiLA Team

The Caribbean plate has a complex tectonic history, which makes it  particularly challenging to establish the evolution of the subduction zones at its margins. Here we present a new teleseismic P-wave tomographic model under the Antillean arc that benefits from ocean-bottom seismometer data collected in our recent VoiLA (Volatile Recycling in the Lesser Antilles) project. We combine this imagery with a new plate reconstruction that we use to predict possible slab positions in the mantle today. We find that upper mantle anomalies below the eastern Caribbean correspond to a stack of material that was subducted at different trenches at different times, but ended up in a similar part of the mantle due to the large northwestward motion of the Americas. This stack comprises: in the mantle transition zone, slab fragments that were subducted between 70 and 55 Ma below the Cuban and Aves segments of the Greater Arc of the Caribbean; at 450-250 km depth, material subducted between 55 and 35 Ma below the older Lesser Antilles (including the Limestone Caribees and Virgin Islands);  and above 250 km, slab from subduction between 30 and 0 Ma below the present Lesser Antilles to Hispaniola Arc. Subdued high velocity anomalies in the slab above 200 km depth coincide with where the boundary between the equatorial Atlantic and proto-Caribbean subducted, rather than as previously proposed, with the North-South American plate boundary. The different phases of subduction can be linked to changes in the age, and hence buoyancy structure, of the subducting plate.

How to cite: Allen, R., Braszus, B., Goes, S., Rietbrock, A., and Collier, J. and the The VoiLA Team: Evolution of Caribbean subduction from P-wave tomography and plate reconstruction, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-9018, https://doi.org/10.5194/egusphere-egu2020-9018, 2020.

D1396 |
Jianfeng Yang and Manuele Faccenda

Most magmatism occurring on Earth is conventionally attributed to passive mantle upwelling at mid-ocean ridges, slab devolatilization at subduction zones, and mantle plumes. However, the widespread Cenozoic intraplate volcanism in northeast China and the peculiar petit-spot volcanoes offshore the Japan trench cannot be readily associated with any of these mechanisms. Furthermore, the seismic tomography images show remarkable low velocity zones (LVZs) sit above and below the mantle transition zone which are coincidently corresponding to the volcanism. Here we show that most if not all the intraplate/petit-spot volcanism and LVZs present around the Japanese subduction zone can be explained by the Cenozoic interaction of the subducting Pacific slab with a hydrous transition zone. Numerical modelling results indicate that 0.2-0.3 wt.% H2O dissolved in mantle minerals which are driven out from the transition zone in response to subduction and retreat of a stagnant plate is sufficient to reproduce the observations. This suggests that critical amounts of volatiles accumulated in the mantle transition zone due to past subduction episodes and/or delamination of volatile-rich lithosphere could generate abundant dynamics triggered by recent subduction event. This model is probably also applicable to the circum-Mediterranean and Turkish-Iranian Plateau regions characterized by intraplate/petit-spot volcanism and LVZs in the underlying mantle.

How to cite: Yang, J. and Faccenda, M.: Intraplate and petit-spot volcanism originating from hydrous mantle transition zone, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5804, https://doi.org/10.5194/egusphere-egu2020-5804, 2020.

D1397 |
Ben Mather, Dietmar Müller, and Tobias Keller

Chemical heterogeneities in the mantle are typically introduced by recycling oceanic lithosphere through subduction, which transports volatiles into the mantle. The provenance of volatiles, such as carbon, with the down-going plate is varied; here we show how the spatial distribution of carbon evolves through time with the motion of the tectonic plates. Carbon is sequestered at mid-ocean ridges, as new oceanic crust forms, and is transported similar to a conveyor belt until it is recycled at subduction zones. We budget the amount of carbon that has been recycled at subduction zones over the past 230 million years using a global plate reconstruction. The present-day distribution of in-plate carbon, taking into consideration the last 230 million years of plate influx, is predominantly distributed in the Atlantic. In contrast, most of the carbon that was sequestered in Pacific seafloor from 230 Ma has since been subducted. Therefore, it is likely that the carbon stored in Pacific seafloor has played an important role in stimulating volcanic activity along the Ring of Fire.

How to cite: Mather, B., Müller, D., and Keller, T.: Tracing the global consumption of carbon at subduction zones over the last 230 million years, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-14838, https://doi.org/10.5194/egusphere-egu2020-14838, 2020.

D1398 |
Carolina Rosca, Stephan König, Marie-Laure Pons, and Ronny Schoenberg

Our understanding regarding the behaviour of the fluid mobile element Zn at destructive plate margins is limited. In particular the fractionation mechanisms and input-output mass-balance remains to be investigated due to implications for the spatio-temporal cycling of this vital and socio-economically relevant element. In this study, we investigate the Zn isotope systematics of subduction input provided by IODP samples from the SW Pacific in comparison to lavas from the central Tonga arc, addressed as a worldwide endmember in terms of pre-subduction mantle wedge depletion. With an improved analytical precision, we report subtle, yet resolvable Zn isotope variations between the central Tongan islands, with an overall statistically relevant variation of 0.05‰ (at ±0.014‰ 2SD). The signatures are all > 0.1‰ lighter than the subduction input at this site, suggesting a fractionation process during subduction. After careful extraction of the isotopic effect caused by mantle melting processes (using DMM δ66/64Zn JMC-Lyon provided by Sossi et al. (2018) and Wang et al. (2017) and melt extraction indices such as Sm/La, Zr/Nb, and Zn/La), a pronounced negative correlation is observed between the Zn isotopic composition of the lavas and key fluid indicators such as Ba/Th and Ce/Pb. Together with predictions from ab initio calculations and mixing models performed between Indian DMM and Rayleigh dehydration of the subducting slab, we attribute the remaining, subtle Zn isotope variations to additions by Cl-rich fluids to the individual mantle wedges. A maximum of 5% chlorine-fluid contribution is suggested for the magmatic source of Tofua, whereas smaller proportions are estimated for Kao, Late and Ata. Overall, this study sheds new light on Zn fractionation mechanisms in sediment-poor subduction zones. Implications for the long-term Zn recycling will be addressed in this presentation.


Sossi, P.A., Nebel, O., O’Neill, H.S.C., Moynier, F., 2018. Zinc isotope composition of the Earth and its behaviour during planetary accretion. Chemical Geology 447, 73-84.

Wang, Z.-Z., Liu, S.-A., Liu, J., Huang, J., Xiao, Y., Chu, Z.-Y., Zhao, X.-M., Tang, L., 2017. Zinc isotope fractionation during mantle melting and constraints on the composition of Earth’s upper mantle. Geochimica et Cosmochimica Acta 198, 151- 167.

How to cite: Rosca, C., König, S., Pons, M.-L., and Schoenberg, R.: Zinc isotope fractionation at destructive plate margins and potential implications for the global recycling signature, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-20001, https://doi.org/10.5194/egusphere-egu2020-20001, 2020.