Journal of Petrology Advance Access originally published online on November 3, 2007
Journal of Petrology 2007 48(12):2407-2430; doi:10.1093/petrology/egm066
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Fate of High-Angle Dipping Slabs in the Subduction Factory: an Integrated Trace Element and Radiogenic Isotope (U, Th, Sr, Nd, Pb) Study of Stromboli Volcano, Aeolian Arc, Italy
1Dipartimento Di Scienze Della Terra, Università Degli Studi Di Firenze, VIA LA PIRA 4, I-50121 Firenze, Italy
2Geowissenschaftliches Zentrum, Universität Göttingen, Abteilung: Geochemie, Goldschmidtstrasse 1, D-37077 Göttingen, Germany
RECEIVED NOVEMBER 11, 2006; ACCEPTED OCTOBER 9, 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEODYNAMIC SETTINGANALYTICAL TECHNIQUESSAMPLE SELECTION AND GENERAL...THE POTENTIAL SOURCES OF...THE MANTLE WEDGE ENRICHMENTCONCLUDING REMARKSREFERENCES |
|---|
The subaerial part of the Stromboli stratovolcano was built up in the last 100 kyr through six periods of activity; the erupted magmas record the largest compositional variation of all the Aeolian arc volcanoes (calc-alkaline, shoshonitic, and potassic alkaline magma series). The trace element characteristics of the less evolved magmas of each period of activity are coherently correlated with their radiogenic isotope (Sr, Nd, Pb) composition, and are typical of volcanic arc rocks. In terms of U-series isotopes, samples from the different magma series have both 238U and 230Th excesses, and this distinctive feature provides additional constraints on source enrichment processes within the mantle wedge and on the mechanism of partial melting. Overall the complete set of data demonstrates that the genesis of the different magma series at Stromboli can be accommodated in a mantle source that experienced two distinct enrichment processes by different parts of the subducting oceanic crust of the Ionian slab. The first was caused by supercritical liquids originating from the basaltic and sedimentary parts of the subducting slab at >5 GPa and
900°C. The second was induced by aqueous fluids, again originating from the basaltic and sedimentary parts of the slab, released from a shallower part of the subducted Ionian slab (< 5 GPa and 800°C). U–Th disequilibria constrain the timing of the first metasomatic event (Stage I: supercritical liquids) at >435 ka, whereas the second event (Stage II: aqueous fluids) occurred at 100 ka. The high-angle dip of the Ionian slab (70°) caused the superimposition of the metasomatizing agents of the two enrichment processes in the same volume of the mantle wedge, explaining the occurrence of such different magma series in a single volcanic edifice. The U–Th disequilibria provide evidence for dynamic melting of the metasomatized mantle wedge combined with an ageing effect resulting from the restoration of secular equilibrium after the perturbation caused by the U-rich aqueous fluids of Stage II. The trace element and radiogenic isotope (U, Th, Sr, Nd, Pb) signature of the mantle source of the magmas at Stromboli is thus dependent upon the amount of supercritical liquids and aqueous fluids released by the two components of the subducted slab, whereas the distinctive 238U and 230Th excesses of the magmas result from a combination of mantle ageing and time-dependent dynamic melting. The geochemical and radiogenic isotope signature of the mantle source beneath Stromboli places important constraints on the isotopic polarity from Southern Latium to the Aeolian arc attributed to the effect of a HIMU mantle component following either lateral inflow of foreland mantle material or upwelling of a mantle plume in the centre of the Tyrrhenian basin. Our geochemical model demonstrates that the high 206Pb/204Pb of the putative ‘HIMU’ mantle component could be equally formed during metasomatism of the pre-existing mantle wedge by either the supercritical liquid (Stage I) or aqueous fluid (Stage II) released by the subducted altered basalt of the Ionian plate. KEY WORDS: radiogenic isotopes; U–Th disequilibria; mantle metasomatism; supercritical liquid; aqueous fluid; Stromboli
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEODYNAMIC SETTINGANALYTICAL TECHNIQUESSAMPLE SELECTION AND GENERAL...THE POTENTIAL SOURCES OF...THE MANTLE WEDGE ENRICHMENTCONCLUDING REMARKSREFERENCES |
|---|
Stromboli is part of the Aeolian Islands volcanic arc, whose formation has been attributed to the subduction of the Ionian plate beneath Calabria (Fig. 1) (e.g. Wortel & Spakman 2000
; Faccenna et al., 2003). Element recycling within the ‘subduction factory’ is a fundamental component in the long-term global geochemical dynamics of the mantle–crust system. A number of studies have demonstrated that subduction zone magmas carry contributions from several components including the mantle wedge, the recycled oceanic crust (altered oceanic basalt and sediment), and the continental or oceanic crust underlying the arc (e.g. Hawkesworth et al., 1991, 1997; Elliott et al., 1997; Turner et al., 1997; Plank & Langmuir, 1998; Sigmarsson et al., 1998). In the case of Stromboli, the geochemical characteristics of the erupted magmas reveal a distinct contribution of both altered oceanic basalt and sediment of the subducted slab (e.g. Class et al., 2000; Hochstaedter et al., 2001). The key points that make Stromboli a unique volcano are the extreme geochemical diversity of its magmas (e.g. Francalanci et al., 1988, 1989, 2007; Luais, 1988), and the high-angle dip (70°) of the subducting Ionian plate, along with its almost complete consumption (e.g. Gvirtzman & Nur, 1999, 2001; Meletti et al., 2000; Pontevivo & Panza, 2006; Faccenna et al., 2007).
|
In this study, we consider the petrogenesis of the different magma series at Stromboli in terms of mantle components, time-scales of mantle enrichment, and partial melting processes. These goals are achieved using an extensive, new, dataset of key chemical tracers including trace elements, radiogenic isotopes (Sr, Nd, Pb) and U-series isotopes, in combination with appropriate geophysical data to constrain the geodynamic setting of the magmatism.
The trace element and radiogenic isotope signature of the Stromboli basalts places important constraints on the isotopic polarity from Southern Latium to the Aeolian arc, ascribed to an increasing within-plate component southward (e.g. Hawkesworth & Vollmer, 1979; Ellam et al., 1989; Conticelli et al., 2002). The within-plate component has been referred either to lateral inflow of foreland mantle material into the wedge as a result of subduction roll-back (e.g. Peccerillo, 2001; Trua et al., 2002), or to a hot mantle plume piercing the subducted slab in the centre of the Tyrrhenian basin (e.g. Gasperini et al., 2002; Bell et al., 2004), neglecting the potential contribution of the subducted altered basalt of the Ionian plate.
|
GEODYNAMIC SETTING |
|---|
|
TOPABSTRACTINTRODUCTIONGEODYNAMIC SETTINGANALYTICAL TECHNIQUESSAMPLE SELECTION AND GENERAL...THE POTENTIAL SOURCES OF...THE MANTLE WEDGE ENRICHMENTCONCLUDING REMARKSREFERENCES |
|---|
A schematic outline of the geodynamic evolution of the Mediterranean region is first summarized to highlight a number of key points that are critical to assess the mantle source characteristics of the magmas at Stromboli.
The Mediterranean domain provides a present-day geodynamic analogue for the early stages of continent–continent collision. The present-day structure of the Italian region is a consequence of the convergence of the African and Eurasian plates (for recent reviews, see, e.g. Gueguen et al., 1998; Faccenna et al., 2003, 2007; Mattei et al., 2004). During the Tertiary, Africa has been slowly converging with Eurasia at a rate of 1–2 cm/year; and the total convergence in the central Mediterranean region is estimated to have reached 400–500 km during the past 90 Myr (e.g. Dewey et al., 1989). The oceanic plates have been subducted or obducted almost completely, except for the last remnants occurring in the Ionian basin and southeastern Mediterranean (Cavazza & Wezel, 2003; Finetti, 2005; Stampfli, 2005). This process, combined with the Neogene retreat of the trench (e.g. Gueguen et al., 1998; Gvirtzman & Nur, 1999, 2001; Wortel & Spakman, 2000; Faccenna et al., 2003), has caused the progressive closure of the intervening Mesozoic oceanic basins of the Western Tethyan domain, with the formation of a complex arcuate orogenic belt (Apennine–Maghrebide belt) and extensional back-arc basins (Ligurian–Provençal and Tyrrhenian basins). The Ionian–Adriatic lithosphere has been continuously subducted toward the NW underneath the Eurasian plate. The present-day remnants of the Ionian plate consist of Triassic–Jurassic oceanic crust underneath a thick pile of Mesozoic and Cenozoic sediments (Catalano et al., 2001; Finetti, 2005; Stampfli, 2005).
The Apennine orogeny and associated back-arc extension were accompanied by long-lived volcanic activity that continues today. The magmas produced represent one of the most impressive features of the geodynamic evolution of Italy, and are characterized by a considerable petrological, geochemical and isotopic diversity (e.g. Conticelli & Peccerillo, 1992; Conticelli et al., 2002; Peccerillo, 2003; Avanzinelli et al., 2007).
The main present-day geodynamic feature of the Italian region is a well-defined Benioff Zone (roughly 200 km wide, dipping NW at 70°) down to about 500 km, which reveals the still active process, although now almost exhausted, of subduction of the Ionian oceanic plate below Calabria and the Tyrrhenian Sea (e.g. Meletti et al., 2000; Wortel & Spakman, 2000; Faccenna et al., 2001, 2007; Panza et al., 2003). Thermal and rheological modelling of the subducting Ionian slab indicates a very cold subduction zone (Carminati et al., 2005; Pasquale et al., 2005), perhaps one of the coldest on Earth, implying the occurrence of dehydration and melting reactions at greater depths than in many volcanic arcs.
The Aeolian Volcanic Arc
The Aeolian volcanic arc (Fig. 1) is related to the subduction beneath Calabria of the Ionian plate, whose depth below the arc is 200 km (Di Stefano et al., 1999; Gvirtzman & Nur, 1999, 2001; Meletti et al., 2000; Faccenna et al., 2001, 2007; Pontevivo & Panza, 2006; Montuori et al., 2007). The ‘arc’ is formed of seven volcanic islands (Alicudi, Filicudi, Salina, Lipari, Vulcano, Panarea, Stromboli) and seven seamounts (Lametini, Alcione, Palinuro, Marsili, Sisifo, Enarete, Eolo), which form an approximately crescent-shaped structure (for a recent review, see Francalanci et al., 2007). The arc is located on the southeastern continental slope of the Tyrrhenian abyssal plain, overlying 15–20 km thick continental crust (Morelli et al., 1975). The age of the subaerial and submerged Aeolian arc volcanoes ranges from at least 1·3 Ma to the present (Beccaluva et al., 1985; Gillot & Keller, 1993; Santo et al., 1995; De Rosa et al., 2003). The volcanic edifices of Vulcano and Stromboli are still active, and historic eruptions have been recorded at Lipari (AD 580).
The Aeolian volcanic rocks belong to the typical magmatic series occurring in subduction-related settings, varying from island arc tholeiite (limited to seamounts, Beccaluva et al., 1985) to calc-alkaline, high-K calc-alkaline and shoshonitic magma series closely associated in space and time (Fig. 1). The magmas with the most depleted isotopic signature crop out in the volcanic island of Alicudi (Peccerillo et al., 1993; Francalanci et al., 1993, 2007). Shoshonitic magmas at Stromboli and Vulcano have a large range of K2O contents at a given silica content, hence a potassic series (KS) is also defined to distinguish the different potassium enrichment (K2O >3·5 wt % at SiO2 <56 wt %). There is no unique time-dependent variation of magmatic series in the different sectors of the arc, although a general increase of K2O with time is observed at Vulcano and Stromboli. The western, central and eastern branches of the arc are distinct on the basis of available petrological, geochemical and geophysical data (Ellam et al., 1989; Francalanci et al., 1993, 2007; Falsaperla et al., 1999; Peccerillo & Panza, 1999; De Astis et al., 2000; Tonarini et al., 2001).
|
ANALYTICAL TECHNIQUES |
|---|
|
TOPABSTRACTINTRODUCTIONGEODYNAMIC SETTINGANALYTICAL TECHNIQUESSAMPLE SELECTION AND GENERAL...THE POTENTIAL SOURCES OF...THE MANTLE WEDGE ENRICHMENTCONCLUDING REMARKSREFERENCES |
|---|
New analytical data have been obtained for chemically well-defined samples for which major and selected trace element data had been acquired in previous studies (Francalanci et al., 1989
, 2004). Briefly, about 100–200 mg of rocks powder were dissolved in screwtop Teflon beakers using a mixture of HF–HNO3–HClO4, then dried down, and repeatedly converted into nitrate and chloride form. Subsequently, sample solutions were split into several aliquots for inductively coupled plasma mass spectrometry (ICP-MS) analysis and sequential U, Th, and Pb, Sr, Nd separation by standard chromatographic ion-exchange techniques [details have been given by Heumann et al. (2002)]. A range of trace element concentrations were determined by ICP-MS using a VG PQ2 system at the Geowissenschaftliches Zentrum, University of Göttingen. Analytical accuracies are synonymous with the significant digits of concentrations reported in Table 1.
|
The concentrations of U and Th were determined by isotope dilution using 229Th and 236U tracer solutions at the University of Göttingen. Isotopic ratios were measured by thermal ionization mass spectrometry (TIMS), by peak jumping or multi-statically switching the ion beam between the ion-counting and Faraday detectors on a MAT262 RPQ+ TIMS system. For uranium, results were processed off-line following an exponential law of mass fractionation after correction for the added tracer solution. The external precision of Th isotope ratio measurements was monitored with an in-house standard which had a 230Th/232Th value within the range of the analysed samples and a reproducibility (2
) of ±0·7% (n = 8). Calculated activity ratios, indicated in parentheses, are based on decay constants from the compilation of Bourdon et al. (2003).
Sr and Nd isotope compositions were measured in dynamic mode on a Thermo Finnigan Triton-Ti TIMS system equipped with nine moveable collectors at the Department of Earth Sciences, University of Firenze (Avanzinelli et al., 2005). The data are corrected for mass fractionation using an exponential law based on 86Sr/88Sr = 0·1194 and 146Nd/144Nd = 0·7219. The external precision of NIST SRM987 was 87Sr/86Sr = 0·710251 ± 9 (2, n = 20), and that of the La Jolla standard was 143Nd/144Nd = 0·511845 ± 4 (2, n = 11). The isotopic composition of Pb was determined on a Thermo Finnigan Neptune high-resolution multi-collector (MC) ICP-MS system at the Department of Earth Sciences, University of Bristol. Samples were spiked with a Tl solution and measured in a single analytical session. A sample–standard bracketing technique was used with NBS981 as standard, for which we obtained 206Pb/204Pb = 16·928 ± 12, 207Pb/204Pb = 15·494 ± 10 and 208Pb/204Pb = 36·709 ± 23 (n = 7). The isotopic composition of the samples is reported in Table 2.
|
|
SAMPLE SELECTION AND GENERAL GEOCHEMICAL CHARACTERISTICS |
|---|
|
TOPABSTRACTINTRODUCTIONGEODYNAMIC SETTINGANALYTICAL TECHNIQUESSAMPLE SELECTION AND GENERAL...THE POTENTIAL SOURCES OF...THE MANTLE WEDGE ENRICHMENTCONCLUDING REMARKSREFERENCES |
|---|
Stromboli is a stratovolcano forming the northernmost island of the Aeolian volcanic arc (Fig. 1). It has an elevation of
2900 m above the Tyrrhenian sea floor (924 m above sea level) with a volume of 300 km3, and lies upon 20 km thick continental crust. The subaerial part of the main cone was built up during the last 100 kyr through six periods of activity (e.g. Hornig-Kjarsgaard et al., 1993) characterized by prevalent lava flows and minor explosive eruptions (Palaeostromboli I, II, III, Vancori, Neostromboli and Recent). The Strombolicchio neck, located 1·7 km NNE offshore from Stromboli and belonging to the same submarine cone, represents the oldest subaerial portion of the volcano, with an age of 200 ka. The volcanic rocks of Stromboli record the greatest range of compositional variation of all the Aeolian arc volcanoes (e.g. Francalanci et al., 1988, 1989, 2007; Luais, 1988), ranging from calc-alkaline (CA: Strombolicchio, Palaeostromboli II) to high-K calc-alkaline (HKCA: Palaeostromboli I, III, Recent Period), shoshonitic (SHO: Vancori, Recent Period), and potassic–alkaline (KS: Neostromboli) magma series (Fig. 2).
|
On the basis of this variety of magma series in <100 kyr of activity, we have focused the study on the less evolved samples of each period to assess the geochemical characteristics of the mantle source at Stromboli (Fig. 2). We have also selected a number of evolved samples to constrain the effects of magmatic differentiation (i.e. fractional crystallization ± assimilation) on the trace element and radiogenic isotope signature. The increase in the potassium content of the studied samples is positively correlated with their incompatible trace element content and radiogenic isotope signature (Sr and Nd isotopes). For example, K2O/Na2O and 87Sr/86Sr increase systematically from CA to SHO and KS rocks (Fig. 3), providing evidence of a continuous enrichment of lithophile elements in the mantle source. In Fig. 3, we have also reported data for representative samples of Alicudi, Panarea and Tyrrhenian Sea basalts. The rationale for this choice is based on the following: (1) Alicudi, although located some 60 km to the west of Stromboli, represents the volcano of the Aeolian Islands with the most depleted isotopic signature (e.g. Peccerillo et al., 1993
; Francalanci et al., 2007) and can help to constrain the enrichment process(es) of the mantle wedge underneath the Aeolian arc volcanoes; (2) Panarea, also with a relatively depleted geochemical signature (e.g. Francalanci et al., 1993), is the closest island to Stromboli and could, therefore, share a common pre-enrichment mantle source; (3) the Tyrrhenian Sea basalts can help to set constraints on the geochemical signature of the upwelling asthenospheric mantle along with potential mixing processes with the overlying lithosphere. The Alicudi and Panarea basalts (Fig. 3) define the unradiogenic end of the Stromboli trend consistent with derivation from a depleted mantle source (e.g. Peccerillo et al., 1993; Calanchi et al., 2002; Francalanci et al., 2007), whereas the Tyrrhenian Sea basalts have distinctly lower K2O/Na2O defining a trend subparallel to that of the Alicudi–Panarea–Stromboli basalts.
|
Primitive mantle normalized trace element patterns are similar for all of the selected Stromboli samples (Fig. 4) and are typical of volcanic arc rocks, with pronounced negative anomalies in Nb–Ta and Ti, and positive anomalies in K and Pb. Rb and Ba are slightly less enriched than U and Th. Despite the similar patterns, the samples have different enrichment factors: the lowest are for the CA rocks of Palaeostromboli II, and the highest for the KS rocks of Neostromboli. Rare earth element (REE) patterns (not shown) are also similar for all of the selected samples and have fractionated light REE [LREE; (La/Sm)n = 4·2–2·8] and heavy REE [HREE; (Gd/Yb)n = 3·5–1·7], and a small negative Eu anomaly (Eu/Eu* = 0·82–0·74). Similar to the other incompatible trace elements, the REE enrichment and the fractionation of HREE increase systematically from the CA rocks of Palaeostromboli II to the KS rocks of Neostromboli (Table 1).
|
The radiogenic isotope signature of the rocks of Stromboli (Fig. 5) trends towards the EM II mantle component (e.g. Zindler & Hart, 1986
). Taken together with representative volcanic rocks from Alicudi and Panarea, the rocks of Stromboli define a trend between an EM II-like and a high-µ (high 238U/204Pb) mantle component (Fig. 5). In all radiogenic isotope diagrams, with no exception, the rocks of Alicudi are the closest to the high-µ mantle end-member, whereas the rocks of Stromboli trend toward the EM II end-member. The Sr (87Sr/86Sr = 0·7037 ± 2, 1), Nd (143Nd/144Nd = 0·51285 ± 3) and Pb (208Pb/206Pb = 2·019 ± 11) isotope signature, and key trace element ratios [e.g. (Th/Nb)n = 4·6 ± 1·0], of the Alicudi basalts (Francalanci et al., 1993, 2007; Peccerillo et al., 1993, 2004) suggest that the high-µ mantle component differs from the classical HIMU or FOZO mantle components defined on the basis of ocean island basalt (OIB) geochemistry (Zindler & Hart, 1986; Stracke et al., 2005). On the other hand, the Tyrrhenian Sea basalts define a trend between the EM II component and a mid-ocean ridge basalt (MORB)-like component (Fig. 5), consistent with new asthenospheric mantle impinging at the base of the lithosphere and interacting with it.
|
The Stromboli lavas exhibit a systematic increase in 87Sr/86Sr from the CA (Palaeostromboli II) to the KS (Neostromboli) rocks. In plots of large ion lithophile elements (LILE)/Nb vs 87Sr/86Sr (e.g. K/Nb vs 87Sr/86Sr, Fig. 6), the data delineate a striking concave-upward pattern: the samples of the CA and SHO series display a negative correlation, whereas the samples of the KS series exhibit a positive correlation. When combined with the samples of Alicudi and Panarea, the whole dataset defines a saw-tooth trend (Fig. 6), with the Panarea basalts plotting on the extension of the CA and SHO series trend toward unradiogenic Sr isotope compositions, whereas the Alicudi basalts define a wide range of K/Nb at relatively constant 87Sr/86Sr.
|
U and Th concentrations in all the magma series are remarkably high, ranging from 2 to 6·5 ppm and from 7 to 25 ppm, respectively (Table 2). The CA samples of Palaeostromboli II have the lowest contents (2 ppm U, 7 ppm Th), whereas the KS samples of Neostromboli have the highest U (up to 6·5 ppm) and Th (up to 25 ppm) contents. In terms of U-series isotopes (Fig. 7), the selected samples have (230Th/232Th) typical of volcanic arc rocks whose geochemical signature is dominated by crustal material (e.g. Hawkeworth et al., 1997
). Most sediment-dominated arcs plot, however, on the equiline (e.g. Condomines & Sigmarsson, 1993), whereas the rocks of Stromboli have both 238U and 230Th excesses. In particular, the Palaeostromboli rocks trend from the equiline towards the right (up to 18% 238U excess), whereas the Neostromboli samples display both 238U (up to 8%) and 230Th (up to 18%) excesses. The rocks of the Recent period have 230Th excesses up to 14%, and a range in (230Th/232Th) that covers 50% of the entire sample variation. We also note a general increase, albeit not systematic, of (230Th/232Th) from the samples of the oldest periods (Palaeostromboli I and II) to those of the youngest periods (Neostromboli and Recent). Another key feature is the roughly positive correlation between the Sr and Th isotope compositions of all the samples, except for those of the Recent period that have constant 87Sr/86Sr but variable (230Th/232Th) (Fig. 8). This characteristic is contrary to what is expected during mantle metasomatism dominated by subducted sediments that typically have high 87Sr/86Sr and low (230Th/232Th) and should, therefore, produce a negative correlation (e.g. Hawkesworth et al., 1997).
|
|
The 238U excess, which is normally interpreted to reflect the addition of fluid-mobile elements (e.g. Ba, U, Pb) released during dehydration of the downgoing slab (e.g. Brenan et al., 1995
; Keppler, 1996; Elliott et al., 1997), provides some contradictory evidence (Fig. 9): (1) in the samples of the CA and SHO series, the U excess of the Palaeostromboli II samples (up to 18%, Fig. 7) is only partially coupled with the U/Th and Ba/Th enrichment; (2) in the KS series of Neostromboli, however, the sample having the most marked Th excess (18% STR 50, Fig. 7) has the highest Ba/Th. In other words, Ba was decoupled from U during the process that created the distinctive U–Th signature of the Stromboli magmas.
|
|
THE POTENTIAL SOURCES OF THE MAGMAS AT STROMBOLI |
|---|
|
TOPABSTRACTINTRODUCTIONGEODYNAMIC SETTINGANALYTICAL TECHNIQUESSAMPLE SELECTION AND GENERAL...THE POTENTIAL SOURCES OF...THE MANTLE WEDGE ENRICHMENTCONCLUDING REMARKSREFERENCES |
|---|
Stromboli, along with other volcanoes of the Aeolian Islands, represents the product of subduction of the Ionian plate beneath Calabria (e.g. Faccenna et al., 2003
; Mattei et al., 2004). On the basis of geophysical studies (Panza, 1985; Bassi et al., 1997; Di Stefano et al., 1999; Gvirtzman & Nur, 1999, 2001; Meletti et al., 2000; Wortel & Spakman, 2000; Faccenna et al., 2001, 2007; Panza et al., 2003; Scrocca et al., 2003; Pontevivo & Panza, 2006; Montuori et al., 2007), the deep structure underneath Stromboli consists of Calabrian continental crust to a depth of 20 km and a thin lithospheric mantle lid to a depth of 40 km, separated by a convective asthenospheric mantle wedge to a depth of 200 km, and then the subducting Ionian slab. A major obstacle in modelling the geochemical characteristics of the mantle wedge, the source of magmas at Stromboli, is the lack of direct data on the various components that could potentially contribute to the trace element budget of the magmas. In the following sections we outline the assumptions we had to make to provide constraints on the geochemical and isotopic characteristics of the different crustal and mantle components beneath Stromboli.
The mantle wedge
The trace element composition of the convective mantle wedge below Stromboli prior to the recent subduction enrichment process can be estimated using the composition of the magmas generated as a consequence of asthenospheric upwelling in the central Tyrrhenian Sea. As the available data for Tyrrhenian Sea basalts are from dredged samples (Beccaluva et al., 1990) (i.e. likely to have suffered some hydrothermal and sea-floor alteration) we have made our selection with a bias towards the least enriched compositions and we have also corrected their 87Sr/86Sr on the basis of their 143Nd/144Nd using the mantle array (DePaolo & Wasserburg, 1979) as a reference. In our inverse modelling we have assumed that the composition of the transitional MORB of the Tyrrhenian Sea (Beccaluva et al., 1990; Gasperini et al., 2002) has been generated by 10% batch melting of the asthenospheric mantle, using the bulk partition coefficients of Workman & Hart (2005). The estimated geochemical composition of the mantle wedge (Table 3) mimics that of enriched MORB-source mantle (E-DMM, Workman & Hart, 2005), and although speculative, it is not critical to the model proposed below, as its low trace element budget is overwhelmed by the trace element and isotopic signal of the subducting components during the recent enrichment processes.
|
The subducted slab
The geodynamic setting of the Aeolian Islands differs from that of most subduction-related volcanoes in the high-angle (
70°) geometry of the Benioff Zone and a very low geothermal gradient (Carminati et al., 2005; Pasquale et al., 2005). This results in the occurrence of dehydration reactions at greater depths than in other arcs (Davies & Stevenson, 1992; Peacock et al., 1994). Consequently, this characteristic must be taken into account in considering the fate of the oceanic crust in the subduction factory. The subducted Ionian oceanic slab consists of sediment, basalt and oceanic lithospheric mantle. In terms of the trace element budget delivered to the overlying mantle wedge, the contribution of the oceanic lithospheric mantle, commonly regarded as being a refractory harzburgite, is considered negligible, excluding any H2O released during the dehydration of serpentine, chlorite and other high-pressure H2O-bearing phases (e.g. Stalder et al., 2001; Iwamori, 2004; Rüpke et al., 2004). These minerals, upon reaching their stability limits, dehydrate and release supercritical fluids, which flush the oceanic crust and transfer fluid-mobile elements to the overlying mantle wedge, lowering the solidus of the peridotite and triggering melting.
On the other hand, the two major components of the oceanic crust (altered basalt, including the intrusive counterpart, and sediment cover) play a fundamental role in recycling lithophile elements back into the mantle wedge through the prograde dehydration reactions occurring during slab subduction (e.g. Hawkesworth et al., 1991, 1997; Peacock et al., 1994; Elliott et al., 1997; Turner et al., 1997; Plank & Langmuir, 1998). The geochemical characteristics of altered oceanic basalt (AOB) mostly reflect hydrothermal alteration at mid-ocean ridges, which continues more slowly as the sea-floor ages (e.g. Staudigel et al., 1995; Becker et al., 2000). The most significant modifications are the enrichments in K, Ba, Rb, Cs, U, and radiogenic Sr (Staudigel et al., 1995). In the case of Stromboli, however, we cannot use the AOB composition as it stands because during slab subduction, the high P–low T metamorphic regime must have induced mineralogical and geochemical transformations of the variably hydrated basalt and gabbro to eclogite with the release of H2O and fluid-mobile elements (e.g. Peacock et al., 1994; Schmidt et al., 2004).
For this reason, we chose to use as a proxy of the basaltic component (Table 3) an intermediate composition between two eclogites (Becker et al., 1999) occurring in the Adula nappe (central Alps, Switzerland) and in the Münchberg Gneiss Massif (Variscan fold belt, Germany). The two eclogites are considered to represent the AOB after the loss of a fluid phase at high-P (2–3 GPa) and relatively low-T (600–650°C) conditions, corresponding to a geothermal gradient <10°C/km, which is likely to be similar to the P–T regime below Stromboli (Carminati et al., 2005; Pasquale et al., 2005). In terms of Pb isotopes, U enrichment during mid-ocean ridge hydrothermal alteration of the basaltic component results in an increase in 238U/204Pb (µ), and consequently in a time-integrated increase of 206Pb/204Pb. Assuming an original 206Pb/204Pb of 18·65, intermediate between DMM and E-DMM (Workman & Hart, 2005), and a µ of 35 (e.g. Staudigel et al., 1995), developed during sea-floor alteration, the basaltic component could reach a 206Pb/204Pb of 19·47–19·75 in 150–200 Myr, a time span consistent with the age of the Western Tethyan ocean basins (Catalano et al., 2001; Bortolotti & Principi, 2005; Finetti, 2005; Stampfli, 2005).
The subducted pelagic sediments of the Ionian plate are the other major component capable of recycling lithophile elements back into the mantle. The relevant data are the sediment samples collected from Ocean Drilling Program (ODP) leg 160, Site 964; unfortunately, these are limited to Sr and Nd isotope data only (Weldeab et al., 2002), and hence we have to make some assumptions about the trace element budget of the sediment pile. A potential proxy for the subducted sediments could be the sediments cropping out in the Apennine chain. They constitute the terrigenous load accumulating in the Tethyan ocean basins, of which the Ionian plate represents one of the last remnants. Unfortunately, there is no comprehensive geochemical database on the Apennine sediments and this limits their use as a proxy of the subducted sediment.
A global survey of sediments recycled into the mantle has been undertaken by Plank & Langmuir (1998) encompassing the main arc–trench systems around the world; this resulted in an average composition of globally subducted sediments (GLOSS). It is tempting to use GLOSS as a proxy of the subducting sediments of the Ionian plate. The available data on sediments cropping out in the Apennine chain reveal, however, systematic low Ba/Th <60 (e.g. Di Battistini et al., 2001; Di Leo et al., 2002; Melluso et al., 2003; Avanzinelli et al., 2007), whereas GLOSS has a Ba/Th = 110. In contrast, the average composition of Post Archean Shales (Taylor & McLennan, 1985) has Ba/Th <60, whereas the other trace elements are not greatly different from GLOSS. We therefore used the Post Archean Shale composition as a proxy of the subducting sediment pile (Table 3).
The Calabrian lithosphere
The trace element budget of the continental crust exceeds greatly that of the continental lithospheric mantle (e.g. Taylor & McLennan, 1985; McDonough, 1990). Consequently, the 20 km thick crust forming the basement below Stromboli has a greater potential than the 20 km thick lithospheric mantle lid to modify the composition of the mantle-derived magmas en route to the surface, as a consequence of low-pressure assimilation–fractional crystallization processes. On the other hand, the frozen outer portion of the magma chamber and feeding dykes can serve as chemical insulators, preventing significant assimilation of the wall-rocks. Previous studies on the different magma series of Stromboli (Fig. 2) have already established that the CA, HKCA, and SHO rocks experienced negligible wall-rock assimilation, whereas the evolution of the KS series was accompanied by complex open-system processes (Francalanci et al., 1988, 1989, 2007).
Details of the evolution of single magma series are beyond the scope of this study; we simply wish to emphasize the general geochemical characteristics of the selected samples. The occurrence of open-system processes can be established using a number of key trace element ratios and radiogenic isotopes because systematic variations are to be expected. For example, the increase of Ba/K with 87Sr/86Sr is a robust tool to discriminate between crustal assimilation and source enrichment (Turner et al., 1996). Available data for the Calabrian basement confirm that the geochemical characteristics of the Stromboli, Alicudi and Panarea magmas are dominantly controlled by source enrichment processes (Fig. 10). The whole set of samples from Stromboli, both evolved and of more primitive composition, exhibit negligible crust involvement. Also, considering the Rb–Sr systematics of the Stromboli magmas alone (Tables 1 and 2), there is no net increase of 87Sr/86Sr with differentiation in each period of activity, reflecting a general closed-system differentiation environment. It should be noted, however, that studies of the Stromboli KS series evolution have proposed the involvement of open-system processes (Francalanci et al., 1988, 1989, 2007). This does not necessarily conflict with our interpretation, which is based on the geochemical signature of the entire set of samples and not on a single magma series. For example, the slight shift to the right of some of the KS series samples in Fig. 10 (indicated by the hatched field), could reflect some crustal assimilation (Francalanci et al., 1989), although it is clear that the displacement towards more radiogenic Sr isotope compositions is minimal and cannot challenge the general interpretation that the main process controlling the geochemical signature of the samples is due to mantle source enrichment and not crustal contamination.
|
|
THE MANTLE WEDGE ENRICHMENT |
|---|
|
TOPABSTRACTINTRODUCTIONGEODYNAMIC SETTINGANALYTICAL TECHNIQUESSAMPLE SELECTION AND GENERAL...THE POTENTIAL SOURCES OF...THE MANTLE WEDGE ENRICHMENTCONCLUDING REMARKSREFERENCES |
|---|
In this section we assess the enrichment processes responsible for the geochemical signature of the mantle source of the magmas feeding Stromboli and the role played by subducted components. The Stromboli lavas have among the highest Th contents of volcanic arc magmas (up to almost 25 ppm excluding the evolved samples, Table 2), including those from the Philippines and Indonesia (e.g. McDermott et al., 1993
; Vroon et al., 1993, 1996; Turner & Foden, 2001). This suggests, along with the low U/Th of the samples (U/Th <0·3, Table 1) and other trace element (Fig. 4) and radiogenic isotope characteristics (Fig. 5), that their geochemistry is dominated by a sedimentary component in the mantle source rather than slab fluids (e.g. Hawkesworth et al., 1997), leaving the open question of the process responsible for the development of both Th and U excesses (Fig. 7), and the positive correlation between Sr and Th isotopes (Fig. 8), which contradicts a model of mantle metasomatism by subducted sediment.
Stage I: the role of supercritical liquids
A key factor that helps to understand the unique characteristics of the magmas of Stromboli is the geometry of the Benioff Zone, which dips at a high angle (70°), associated with a very low geothermal gradient. A number of studies have suggested that the transfer of sedimentary components from the subducted slab into the mantle wedge occurs as a partial melt rather than by bulk addition (e.g. Hawkesworth et al., 1997; Turner et al., 1997; Sigmarsson et al., 1998; Johnson & Plank, 1999), despite the fact that most thermal models of subduction zones predict temperatures at the slab–wedge interface that are too low for sediment melting (e.g. Peacock et al., 1994; Rüpke et al., 2004). In the case of Stromboli, this dichotomy can possibly be resolved because of the high-angle dip of the Ionian plate, which causes dehydration and ‘melting’ reactions to occur at greater depths (equivalent to >4 GPa) than in many volcanic arcs (e.g. Davies & Stevenson, 1992). The increase of the dehydration and ‘melting’ pressure for a given lithology results in the convergence of the composition of the low-density aqueous fluid (low-temperature side) and the dense hydrous silicate melt (high-temperature side) along a miscibility gap, which eventually disappears; the intersection of the fluid-saturated solidus with the closure of the fluid–melt solvus at high pressure is termed the second critical end-point of the solidus (Boettcher & Wyllie, 1969). Beyond this pressure, the discontinuous solidus reaction involving the aqueous fluid (low T) and the hydrous silicate melt (high T) is replaced by a continuous reaction involving a supercritical liquid (Kessel et al., 2005), which spans a chemical continuum between the composition of the aqueous fluids and hydrous silicate melts at lower pressure. In particular, the low-pressure difference between the partition coefficients of fluid-mobile (e.g. Rb, Ba, Pb, U) and fluid-immobile (e.g. Th, REE) trace elements ceases to exist in supercritical liquids, and all the lithophile trace elements have melt-like solubilities (Kessel et al., 2005).
In basaltic and pelitic systems, the second critical end-point of the solidus is at 5–6 GPa (Schmidt et al., 2004; Kessel et al., 2005). Melting of the slab at P > 5 GPa generates supercritical liquids whose K content should be positively correlated with pressure as demonstrated by the experiments of Johnson & Plank (1999), albeit limited to 2 and 4 GPa. In the case of Stromboli, the positive correlation between K2O/Na2O and 87Sr/86Sr (Fig. 3) provides evidence of mantle metasomatism by supercritical liquids enriched in K and radiogenic Sr.
As the subducted slab includes both basalt and sediment, their distinct trace element leverage on the composition of the mantle wedge can possibly generate a different geochemical signal in the erupted magmas (e.g. Class et al., 2000; Hochstaedter et al., 2001). In the case of Stromboli, the distinctive concave-upward pattern in the K/Nb vs 87Sr/86Sr diagram (Fig. 6) requires two geochemically distinct enriching agents both having high K/Nb (and also other LILE/Nb): one with a relatively unradiogenic Sr isotope composition (87Sr/86Sr <0·705), possibly originating from the basaltic component, and the other with a more radiogenic Sr isotope composition (87Sr/86Sr >0·705), possibly related to the subducted sediment component.
Accordingly, we have attempted to model the enrichment of the mantle wedge overlying the subducted slab using the relevant bulk partition coefficients (D) for a number of key trace elements that have been recently determined for supercritical liquids in equilibrium with basalts at 6 GPa (Kessel et al., 2005); that is, beyond the second critical end-point of the solidus (Boettcher & Wyllie, 1969).
As starting end-members we used (1) our estimated composition of the mantle wedge, (2) the composition of the basaltic crust after the escape of a fluid phase at lower pressure, and (3) the sediment component (Table 3). The bulk partition coefficients have been measured only for supercritical liquids in equilibrium with basalt, but at pressures >4 GPa sediments have the same eclogite-facies mineralogy (although different mineral proportions) as basalt, except for the occurrence of phengite because of the higher K2O and Al2O3 contents of the sediments (Schmidt et al., 2004). The high solubility of K2O in supercritical liquids, however, results in the rapid dissociation of phengite and its removal from the residual sediment (Schmidt et al., 2004). Thus, the bulk partition coefficients measured for supercritical liquids in equilibrium with basalt can be equally applied to sediments (Kessel et al., 2005), although the high amount of coesite and kyanite in the residual sediment (50%, Schmidt et al., 2004) requires that the bulk partition coefficients of the sedimentary component be roughly reduced by a factor of 0·5 compared with those of the basaltic component.
We have determined the best fit to the Stromboli data using a trial and error approach on the parameters controlling the trace element composition of the two supercritical liquids: the partial melting degree and the bulk partition coefficients. The compositional variability of the subducted oceanic crust (basalt and sediment) has not been considered in our modelling but it is another parameter that can potentially affect the trace element characteristics of the two enriching agents. The best solution has been found using the bulk partition coefficients measured by Kessel et al. (2005) at 6 GPa and 900°C, and a partial melting degree of 4% and 20% for the basaltic and sedimentary component, respectively (Table 3). The P–T conditions represent only nominal values of the available bulk partition coefficients that have been adopted (Kessel et al., 2005), and are not strictly related to the actual P–T conditions of slab melting below Stromboli; they simply indicate the high-pressure melting regime (>5 GPa) of the subducted slab with the formation of supercritical liquids, consistent with the low geothermal gradient of the Aeolian arc subduction zone. In our modelling, the bulk partition coefficients for the sedimentary (reduced by a factor of 0·5; see above) and basaltic components have been allowed to vary within 1 of the measured values (Kessel et al., 2005), and the partial melting degrees have been estimated according to the results of Schmidt et al. (2004) and Kessel et al. (2005). Moreover, assuming all Ti formed rutile, consistent with the experimental results of Schmidt et al. (2004) and Kessel et al. (2005), Nb has been considered to have a less incompatible behaviour in the basaltic than in the sedimentary component of the subducted slab (DNb = 1 and 0·12, respectively, Table 3).
The results of the modelling are presented in Figs 11 and 12 for a number of key trace element ratios (Th/Nd, Th/Nb, U/Nb, Ba/Nb) and radiogenic isotopes (Sr, Nd, Pb). The mantle source of the Stromboli magmas can be reproduced by metasomatism of the mantle wedge with different mixtures of the two supercritical liquids originating from the subducted slab. As an example, we have reported in Figs 11 and 12 three mixing lines starting from the pre-enrichment mantle wedge and representing three different mixtures of the two supercritical liquids corresponding to 15% (SL1), 30% (SL2) and 60% (SL3) relative of the sediment-derived supercritical liquid (Table 3). The two supercritical liquids, unlike aqueous fluids originating at shallower depths (<120 km), have melt-like solubilities for LILE and high field strength elements (HFSE) (Kessel et al., 2005). This means that high LILE/HFSE values can be obtained not only for ratios involving fluid-mobile elements (e.g. K/Nb, U/Nb) but also for ratios of fluid-immobile elements such as Th/Nb, depending on the amount of rutile remaining in the residue. The basaltic component of the slab will deliver a supercritical liquid with relatively unradiogenic 87Sr/86Sr and high Th/Nb (DTh = 0·04, DNb = 1, Table 3), whereas the sedimentary component will yield a supercritical liquid with more radiogenic 87Sr/86Sr and relatively lower Th/Nb (DTh = 0·03, DNb = 0·12, Table 3; Fig. 12a). This determines, along with the mixing proportions of the two supercritical liquids with the mantle wedge, the development of the characteristic concave-upward pattern in Figs 6 and 12a.
|
|
The mantle source of the CA series of Palaeostromboli II requires a metasomatizing agent consisting of only 15–20% relative of the sediment-derived supercritical liquid (SL1), whereas the mantle source of the KS series of Neostromboli requires a supercritical liquid formed by some 50–60% relative of the sedimentary component (SL3). In terms of absolute amounts, the mantle source of the CA series of Palaeostromboli II requires
7% of the supercritical liquid mixture SL1 (Figs 11 and 12), corresponding to 1% and 6% absolute of the sediment-derived and basalt-derived supercritical liquid, respectively. On the other hand, the mantle source of the KS series of Neostromboli requires 3% of the supercritical liquid mixture SL3 (Figs 11 and 12), corresponding to 2% and 1% absolute of the sediment-derived and basalt-derived supercritical liquid, respectively. These supercritical liquid proportions, both absolute and relative, are consistent with the difference in radiogenic isotope signature and potassium enrichment between the two magma series. The other magma series (HKCA and SHO) fall between these two extreme compositions and require intermediate mixtures of the two supercritical liquids. Another important geochemical characteristic is that the more the relative proportion of the supercritical liquid from the basaltic component, the higher the 206Pb/204Pb (SL3, SL2, SL1, Fig. 11b) owing to the high µ developed during mid-ocean ridge hydrothermal alteration. Consequently, the involvement of two distinct supercritical liquids enriching the mantle wedge can account for the trend between the EM II and high-µ mantle components delineated by the rocks of Stromboli, Alicudi and Panarea (Fig. 5). We can thus envisage a scenario where these high-P supercritical liquids migrate through the overlying mantle wedge at a very fast rate owing to their low viscosity, and form a metasomatized mantle that subsequently, upon melting, originates the different types of magma series building up the volcanic edifice of Stromboli.
Despite the general agreement between the model calculations and the Stromboli lavas, there are a number of discrepancies that require further explanation. For example, the high Th/Nd values outside the range of the model metasomatized mantle are restricted to evolved samples (Fig. 11c), and could be due to low-pressure differentiation processes. On the other hand, the fluid-mobile trace element signature (U, Ba, and partly 206Pb/204Pb, Figs 11b and 12b, c) of the model metasomatized mantle wedge does not match that of the Stromboli magmas and is unlikely to be the result of low-pressure differentiation as suggested above for Th/Nd. This observation, along with the U–Th disequilibria in the samples (Fig. 7), is discussed below.
Stage II: the fluid signal
In terms of U–Th systematics, the high-P supercritical liquids invariably have 230Th excesses (Table 3 and Kessel et al., 2005); the Stromboli samples thus require additional processes to explain their distribution on the equiline diagram (Fig. 7). Also, the metasomatizing process outlined above is unable to account for both the positive correlation between Sr and Th isotopes (Fig. 8), and the fluid-mobile trace element signature (Figs 11 and 12). A plausible explanation is that the Stage I enrichment occurred >435 kyr ago, allowing the 230Th excess created in the mantle wedge by the supercritical liquids to decay. This delay between mantle enrichment (Stage I) and the eruption of magmas of the Palaeostromboli I (85 ka) period, although it helps in modelling the composition of the Stromboli samples, is not sufficient to explain the 238U excess of some of the Stromboli magmas (Fig. 7).
On the basis of the experimental data of Kessel et al. (2005), the only way to produce a metasomatizing agent with 238U excess in this high-P regime involves an aqueous fluid at a pressure below the second critical end-point of the solidus in basaltic and pelitic systems (< 5 GPa), as established in a number of papers dealing with U–Th systematics of subduction zone magmas (e.g. Brenan et al., 1995; Keppler, 1996; Elliott et al., 1997). Consequently, it is tempting to suggest that the high-P supercritical liquids represent a first and old stage of mantle wedge enrichment, and a second, more recent stage occurred at lower P and involved a shallower part of the subducting Ionian slab. In our modelling we have assumed, although not strictly necessary, that the mantle wedge was back to secular equilibrium before the onset of the second metasomatizing stage. Also, the aqueous fluids liberated from the subducting oceanic crust (e.g. Tatsumi, 1989; Poli & Schmidt, 1995) have perhaps been helped by H2O released during dehydration of serpentine and chlorite in the underlying harzburgite, clearly occurring at greater depths (e.g. Stalder et al., 2001; Iwamori, 2004; Rüpke et al., 2004). These fluids migrated rapidly into the upper portion of the overlying mantle wedge—already metasomatized during Stage I—and triggered melting of the mantle source to form the different magma series building up the volcanic edifice of Stromboli.
This second stage must have taken place not long before the eruption of the Palaeostromboli I magmas (85 ka) to maintain the 238U excess signature of the Palaeostromboli II magmas at 60 ka (Table 2). The 18% 238U excess in the sample STR143 (Palaeostromboli II, Table 2) confines the time-window for aqueous fluid release between 120 and 85 ka, and in our model we have assumed 100 ka. The exact time is not critical and can only slightly modify the melting rate parameter during dynamic melting (see below).
On the basis of these arguments, we have tried to model the second stage of mantle enrichment by aqueous fluids released from another part of the subducting Ionian slab located at shallower depth (equivalent to <5 GPa), using the same trial and error approach as for Stage I. The best-fit solution has been found using the bulk partition coefficients measured for aqueous fluids in equilibrium with basalt (and sediment) at 4 GPa and 800°C (Kessel et al., 2005), and an amount of H2O release of 2% and 3% from the basaltic and sedimentary component, respectively (Table 3), consistent with the amount of fluid added to the mantle wedge beneath volcanic arcs (see Turner et al., 2003, for a review). As for Stage I, the P–T conditions are only nominal values and indicate a shallower depth of slab dehydration with the formation of aqueous fluids instead of supercritical liquids. The most prominent effects of this second stage are limited to U, Ba, 206Pb/204Pb and partly 87Sr/86Sr (i.e. fluid-mobile elements), although the mobility of Ba (DBa = 0·13, Table 3) from the sedimentary component has been assumed to be reduced with respect to the experimental data of Kessel et al. (2005), owing to the stability of phengite at 800°C and 4 GPa (Schmidt et al., 2004).
Also during this stage, the two distinct aqueous fluids can mix in different proportions to account for the specific fluid-mobile characteristics of the Stromboli lavas. For the sake of clarity, in Figs 11 and 12 we have reported the effects of the Stage II metasomatism only for fluid-mobile elements and considering the two compositionally extreme magma series of the Palaeostromboli II (CA series) and Neostromboli (KS series). The result of the Stage II model indicates that the Neostromboli (SL3) and Palaeostromboli (SL1) mantle sources after the Stage I enrichment are flushed and hybridized by two mixtures corresponding to 10% (AF1) and 60% (AF2) relative of the sediment-derived aqueous fluid. The absolute amount of the aqueous fluid mixture is 7% for AF1 and 3% for AF2 (white arrows, Figs 11b and 12b, c). This process causes a shift towards higher Ba/Nb and U/Nb of the two mantle sources (Fig. 12b and c), along with an opposite change of their Pb isotope compositions (Fig. 11b). The fluid-mobile trace element characteristics of the other magma series (HKCA and SHO), falling between these two extreme compositions (Figs 11b and 12b, c), require intermediate mixtures of the basalt- and sediment-derived aqueous fluids. Notably, the samples with high U/Nb (STR 143 Palaeostromboli II, and STR 2 Neostromboli, Fig. 12c) have also the most extreme 238U excess (Fig. 7).
Another striking feature is that Th is slightly mobilized in the aqueous fluids (Kessel et al., 2005). This means that the aqueous fluids (e.g. AF1, AF2), although having high (238U/232Th), also have (230Th/232Th) higher than the metasomatized mantle wedge back to secular equilibrium. Consequently, this second step of metasomatism shifts the mantle source towards the 238U excess part of the equiline and produces a slightly inclined array (Fig. 13a).
|
The only sample that does not fit the model is the sample STR50 (KS series) because, as with Ba/Th (Fig. 9), its high Ba/Nb (Fig. 12b) would require a significant amount of aqueous fluid which disagrees with its low U/Th (Fig. 9), U/Nb (Fig. 12c) and 230Th excess (Fig. 7). Also, its Ba content is not greater than that of other KS series samples (Table 1). We have no argument to explain this anomalous behaviour except that, perhaps, the geochemical signature of this sample has been determined by either low-pressure differentiation (Francalanci et al., 1989
, 2007) or a slightly different metasomatized mantle source with respect to other magmas of the KS series. Given this scenario, we are left to explain how it is possible to produce a series of magmas with 230Th excess, mostly during the Recent and Neostromboli periods, and with (230Th/232Th) greater than that of their metasomatized mantle source (Fig. 13a). Any ageing effect will ultimately raise the (230Th/232Th) of the metasomatized mantle source (Fig. 13a) but cannot form magmas with Th excess starting from the U excess part of the equiline diagram. A potential explanation of the shift toward the Th excess part of the equiline diagram could be related to the presence of garnet in the residue after partial melting of the mantle wedge, allowing a net U–Th fractionation. Undoubtedly, the HREE fractionation of the samples (Table 1) requires an increasing amount of residual garnet from CA to SHO and KS series. However, after testing this possibility, we found that residual garnet alone is unable to greatly modify the (238U/232Th) of the mantle source. What is more important, residual garnet cannot account for the roughly constant (Gd/Yb)n and variable (238U/232Th) of the Neostromboli samples (Table 1 and Fig. 13a).
The only possibility left to cross the equiline has to do with a time-dependent melting model as opposed to instantaneous melting models, such as batch or fractional melting, which implicitly assume instantaneous melt generation and extraction. These time-dependent melting models (for a review, see, e.g. Elliott, 1997), although still requiring a net U–Th fractionation at the onset of mantle melting (i.e. residual garnet), consider the time elements spend in the melting column until a critical porosity is reached and the melt is extracted. For stable elements there is no difference between instantaneous and time-dependent melting models, but for short-lived nuclides such as 230Th the result of melting during a finite time period causes 230Th ingrowth in the residual mantle and produces 230Th excess in the extracted magmas. The different time-dependent mantle melting processes at subduction zones have been recently reviewed by Bourdon et al. (2003), who suggested, on the basis of U-series systematics, that the most appropriate mechanism is a dynamic melting model following fluid addition, where melt is produced within a porous matrix and then rapidly channelized to the surface with little or no re-equilibration.
We have, therefore, combined an ageing effect starting from 100 ka with a dynamic melting model using the formalism of Zou & Zindler (2000). The results (Fig. 13a) are presented for the magmas of the CA series (60 ka, Palaeostromboli II), the KS series (10 ka, Neostromboli), and the SHO series (0 ka, Recent period). For the sake of clarity, only a few arrows are reported (vertical dashed arrows, Fig. 13a), which show the vertical displacement of the mantle source upon melting, applying (1) a melting rate of some 10–3 to 10–4 kg/m3 per year, (2) a mantle porosity of 0·2%, and (3) bulk distribution coefficients for U and Th of 0·003 and 0·0014, respectively. These parameters are in agreement with those adopted by Bourdon et al. (2003) in their review of U-series disequilibria in arc-related magmatism. Notably, different melting rates during dynamic melting can account for the (230Th/232Th) variability of the Recent period magmas, which covers 50% of the entire sample range.
The ageing of the metasomatized mantle wedge after aqueous fluid addition coupled with dynamic melting is also able to explain the apparent controversy related to the positive correlation between Sr and Th isotopes, including the variable (230Th/232Th) at constant 87Sr/86Sr of the Recent period samples (Fig. 13b). The first stage of metasomatism by supercritical liquids (Stage I) resulted in a progressive increase of the radiogenic Sr isotope signature of the mantle source (Figs 11a and 13b) from the CA series (Palaeostromboli II) to the SHO series (Recent) and the KS series (Neostromboli), while keeping an almost constant (230Th/232Th) of 0·62 once back to secular equilibrium. The second stage of metasomatism by aqueous fluids (Stage II) shifted the mantle source towards the U excess part of the equiline. This caused, during restoration of secular equilibrium, an increase of (230Th/232Th) of the mantle source (Fig. 13a) dependent upon the time elapsed since aqueous fluid addition and the amount of aqueous fluid added; that is, the greater the fluid added and the time elapsed, the greater the (230Th/232Th) increase. At the time corresponding to the eruption of each magma series, the (230Th/232Th) of the mantle wedge was further increased by a time-dependent melting process reflecting the rate of melting; that is, the lower the melting rate the higher the (230Th/232Th) increase.
Thus, the CA series of Palaeostromboli II have the lowest (230Th/232Th) and 87Sr/86Sr because they were erupted early (60 ka) and they originated from a mantle source metasomatized by a supercritical liquid released mainly from the basaltic component of the subducted slab (SL1, Figs 11 and 12). In constrast, the KS series of Neostromboli have the highest 87Sr/86Sr and among the highest (230Th/232Th) because they were erupted much more recently (10 ka) and they originated from a mantle source metasomatized by a supercritical liquid mixture consisting of 60% relative of the sedimentary component of the subducted slab (SL3, Figs 11 and 12). The variable and highest (230Th/232Th) at constant 87Sr/86Sr of the Recent period samples is probably a consequence of variable melting rates within a mantle wedge metasomatized by a relatively constant amount of the supercritical liquid mixture SL2 (3% absolute, Figs 11 and 12).
Geodynamic model of the Stromboli subduction factory
A realistic geodynamic model of the Stromboli subduction factory must take into account both the key geochemical characteristics of the erupted magmas and the structure of the Aeolian arc subduction zone. The predominant geochemical characteristics of the mantle source of the Stromboli magmas have been imparted by the different mixtures of the two supercritical liquids of Stage I (SL1, SL2, SL3, Figs 11 and 12). This process cannot account, however, for fluid-mobile trace element characteristics (U, Ba, K, and partly Pb) and U–Th disequilibria of the samples. It is then proposed that another mantle enrichment stage occurred at shallower depths induced by aqueous fluids (Stage II), along with a delay of 
350 kyr between the two stages to allow the decay of the 230Th excess of the enriched mantle source after Stage I. Fitting these geochemical constraints into a geodynamic model is not straightforward, although some constraints can be made based on the model proposed for the Mariana arc lavas by Elliott et al. (1997), adapting it to the structure of the Aeolian arc subduction zone.
The simplest model (Fig. 14), perhaps not the only one, must be based upon the low geothermal gradient of the Aeolian arc subduction zone, which dips at a high angle (70°), leading to melting of the subducting slab at depths >150 km with the formation of supercritical liquids (Stage I). The supercritical liquids from the basalt and sediment components of the slab rise into the overlying mantle wedge and react to hybridize it. We cannot say whether or not this stage caused melting of the mantle source. The only observation we can make is that any magmas that might have been generated during this first stage did not erupt in the last 85 kyr (Palaeostromboli I). It is possible, however, that magmas produced during Stage I could have been erupted prior to the Palaeostromboli I period (>85 kyr ago), as our samples are limited to the subaerial part of the Stromboli volcano, and below sea level there are some 300 km3 of volcanic rocks building up the volcanic edifice from the base of the Tyrrhenian sea floor (2000 m below sea level). During the delay between the two stages, the enriched mantle wedge is dragged downward with the subducting slab, and the 230Th excess created by the addition of the supercritical liquids decays.
|
In the second stage (Stage II), dehydration reactions in the subducted oceanic crust and the underlying harzburgite occurs at depths shallower than 150 km, releasing aqueous fluids into the overlying enriched mantle wedge and creating the 238U excess. The aqueous fluid enrichment is critical to lower the solidus of the mantle wedge and trigger melting. Thus the mantle wedge, on returning to secular equilibrium (i.e. increasing 230Th/232Th, Fig. 13a), forms, upon melting, the different Stromboli magmas whose fluid-immobile trace element characteristics are dominated by the Stage I mantle enrichment process.
The critical parameter in this geodynamic model is the delay between the two stages. If we take at face values the nominal pressures (4 and 6 GPa) and temperatures (800 and 900°C) adopted in our model—recalling that they represent only the experimental P–T conditions used to measure the trace element partitioning between aqueous fluids and supercritical liquids in equilibrium with basalt (Kessel et al., 2005)—the geothermal gradient should vary from 5°C/km (Stage I) to 6·6°C/km (Stage II). It is beyond the scope of our study to discuss thoroughly the causes of these P–T variations of the descending slab in 350 kyr, but we believe that this small difference could be caused by variations of the shear heating parameter (e.g. Davies & Stevenson, 1992; Roselle et al., 2002) in response to slab roll-back (e.g. Gvirtzman & Nur, 1999, 2001; Faccenna et al., 2001) and the likely slowing down of the convergence rate of the almost exhausted Ionian slab.
Whatever the reason and the actual P–T–t path of the Ionian slab, the important point outlined by the geochemical characteristics of the Stromboli magmas is the occurrence of two mantle enrichment stages (supercritical liquids and aqueous fluids) that are separated by a time interval of 350 kyr and retain their discrete chemical characteristics until they are brought together in the mantle source, shortly before melting.
|
CONCLUDING REMARKS |
|---|
|
TOPABSTRACTINTRODUCTIONGEODYNAMIC SETTINGANALYTICAL TECHNIQUESSAMPLE SELECTION AND GENERAL...THE POTENTIAL SOURCES OF...THE MANTLE WEDGE ENRICHMENTCONCLUDING REMARKSREFERENCES |
|---|
We have presented a comprehensive trace element and radiogenic isotope study of the eruptive products of Stromboli, one of the most intriguing volcanoes of the world. The variety of magma compositions erupted in the last 100 kyr is unique among the Aeolian Island volcanoes. This variety requires a complex model of mantle source enrichment and time-dependent mantle melting mechanisms.
The main results obtained from our study demonstrate that the magmas of Stromboli originated from a mantle source that experienced two distinct enrichment processes by different parts of the subducting oceanic crust of the Ionian slab. The first was caused by supercritical liquids originating from the basaltic and sedimentary component of the subducting slab at >5 GPa and 900°C. The second was caused by aqueous fluids, originating again from the basaltic and sedimentary components of the slab, released from another part of the subducted Ionian slab located at shallower depth (equivalent to <5 GPa) and 800°C. The high-angle dip of the Ionian slab determined the superimposition of the metasomatizing agents of the two enrichment processes (Stage I: supercritical liquids; Stage II: aqueous fluids) in the same volume of mantle wedge, explaining the occurrence of such different magma series in a single volcanic edifice (Fig. 14).
U–Th disequilibria constrain the timing of the first metasomatism (Stage I) at 435 ka, whereas the second event (Stage II) occurred at 100 ka. Moreover, the (230Th/232Th) and 238U and 230Th excesses in the Stromboli samples provide evidence for the occurrence of dynamic melting of the metasomatized mantle wedge, combined with an ageing effect resulting from the restoration of secular equilibrium after the perturbation caused by the U-rich aqueous fluids of the second stage.
As a corollary, the geochemical and isotopic signature of the mantle source of the Stromboli magmas places important constraints on the isotopic polarity from Southern Latium to the Aeolian arc (e.g. Hawkesworth & Vollmer, 1979; Ellam et al., 1989; Conticelli et al., 2002), attributed to an increasing within-plate (HIMU) component. For the sake of clarity, we did not include in the model of the Stromboli magmas (Figs 11 and 12) representative samples from Alicudi and Panarea, but we have demonstrated (Fig. 11b) that the trend towards high 206Pb/204Pb of the putative ‘HIMU’ mantle component delineated by the Alicudi and Panarea magmas (Fig. 5) can be equally formed during metasomatism of the pre-existing mantle wedge by the supercritical liquid (Stage I) and subsequently aqueous fluid (Stage II) released by the subducted altered basalt of the Ionian plate. Similarly, the trend towards high K/Nb (and also high Th/Nb, Francalanci et al., 1993, 2007; Peccerillo et al., 1993, 2004) and relatively low 87Sr/86Sr of the Alicudi and Panarea basalts (Fig. 6) can be explained by the predominance of the enriching agents (supercritical liquid and aqueous fluid) from subducted altered basalt in the mantle source of the Alicudi and Panarea magmas.
The involvement of the subducted basalt of the Ionian slab in forming the high 206Pb/204Pb mantle component is in keeping with the orogenic trace element characteristics of the Alicudi and Panarea magmas (Francalanci et al., 1993, 2007; Peccerillo et al., 1993, 2004), and requires neither lateral inflow of foreland mantle material (e.g. Peccerillo, 2001; Trua et al., 2002) nor a hot mantle plume in the centre of the Tyrrhenian basin (e.g. Gasperini et al., 2002; Bell et al., 2004). The former is a very sluggish process, and the latter is contrary to the radiogenic isotope signature of the Tyrrhenian Sea basalts.
|
ACKNOWLEDGEMENTS |
|---|
This paper benefited from discussions with Jon Davidson and Gerhard Wörner during the ERUPT workshop held on Stromboli in 2005. The criticism of Sandro Conticelli on an early draft of the manuscript and the discussion with Giuliano Panza on the vertical structure beneath the Aeolian Islands is also acknowledged. The careful review of Marjorie Wilson and that of Ronit Kessel greatly helped to improve the manuscript. We would like to thank Maurizio Ulivi for helping with Sr and Nd isotope analysis in Firenze, and Gabriele Mengel and Klaus Simon for their expertise in ion-chromatography and IPC-MS analysis in Göttingen. Tim Elliott provided access to the Thermo Finnigan Neptune facility in Bristol for Pb isotope analysis. This research was supported by an ERUPT grant (EVG1-CT-2002-00058) issued to L.F.
*Corresponding author. Telephone: +39 055 2756347. Fax: +39 055 2756242. E-mail: toms{at}unifi.it
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGEODYNAMIC SETTINGANALYTICAL TECHNIQUESSAMPLE SELECTION AND GENERAL...THE POTENTIAL SOURCES OF...THE MANTLE WEDGE ENRICHMENTCONCLUDING REMARKSREFERENCES |
|---|
Arrighi S, Rosi M, Tanguy JC, Courtillot V. Recent eruptive history of Stromboli (Aeolian islands, Italy) determined from high-accuracy archeomagnetic dating. Geophysical Research Letters (2004) 31:L19603. doi:10.1029/2004GL020627.[CrossRef]
Avanzinelli R, Boari E, Conticelli S, Francalanci L, Guarnieri L, Perini G, Petrone CM, Tommasini S, Ulivi M. High precision Sr, Nd, and Pb isotopic analyses using new generation thermal ionisation mass spectrometer: aims and perspective for isotope geology applications. Periodico di Mineralogia (2005) 74:147–166.
Avanzinelli R, Elliott T, Tommasini S, Conticelli S. Constraints on the genesis of the potassium-rich Italian volcanics from U/Th disequilibrium. Journal of Petrology (2007) 48 (in press).
Bassi G, Sabadini R, Rebai S. Modern tectonic regime in the Tyrrhenian area: observations and models. Geophysical Journal International (1997) 129:330–346.[CrossRef][Web of Science]
Beccaluva L, Gabbianelli G, Lucchini F, Rossi PL, Savelli C. Petrology and K/Ar ages of volcanics dredged from the Aeolian seamounts: implication for the geodynamic evolution of the Southern Tyrrenian basin. Earth and Planetary Science Letters (1985) 74:187–208.[CrossRef][Web of Science]
Beccaluva L, Bonatti E, Dupuy C, Ferrara G, Innocenti F, Lucchini F, Macera P, Petrini R, Rossi PL, Serri G, Seyler M. Geochemistry and mineralogy of volcanic rocks from ODP sites 650, 651, 655 and 654 in the Tyrrhenian Sea. In: Proceedings of the Ocean Drilling Program, Scientific Results, 107—Stewart NJ, ed. (1990) College Station, TX: Ocean Drilling Program. 49–74.
Becker H, Jochum KP, Carlson RW. Constraints from high-pressure veins in eclogites on the composition of hydrous fluids in subduction zones. Chemical Geology (1999) 160:291–308.[CrossRef][Web of Science]
Becker H, Jochum KP, Carlson RW. Trace element fractionation during dehydration of eclogites from high-pressure terranes and the implications for element fluxes in subduction zones. Chemical Geology (2000) 163:65–99.[CrossRef][Web of Science]
Bell K, Castorina F, Lavecchia G, Rosatelli G, Stoppa F. Is there a mantle plume below Italy? EOS Transactions, American Geophysical Union (2004) 85:541–547.
Boettcher AL, Wyllie PJ. The system CaO–SiO2–CO2–H2O—III. Second critical end-point on the melting curve. Geochimica et Cosmochimica Acta (1969) 33:611–632.[CrossRef][Web of Science]
Bortolotti V, Principi G. Tethyan ophiolites and Pangea break-up. Island Arc (2005) 14:442–470.[CrossRef]
Bourdon B, Turner S, Dosseto A. Dehydration and partial melting in subduction zones: Constraints from U-series disequilibria. Journal of Geophysical Research (2003) 108(B6):2291. doi:10.1029/2002JB001839.[CrossRef]
Boynton WV. Cosmochemistry of the rare earth elements: meteorite studies. In: Rare Earth Element Geochemistry—Henderson P, ed. (1984) Amsterdam: Elsevier. 63–114.
Brenan JM, Shaw HF, Phinney DL, Ryerson FJ. Mineral–aqueous fluid partitioning of trace elements at 900°C and 2·0 GPa: constraints on the trace element chemistry of mantle and deep crustal fluids. Geochimica et Cosmochimica Acta (1995) 59:3331–3350.[CrossRef][Web of Science]
Calanchi N, Peccerillo A, Tranne CA, Lucchini F, Rossi PL, Kempton P, Barbieri M, Wu TW. Petrology and gochemistry of volcanic rocks from the island of Panarea: implications for mantle evolution beneath the Aeolian island Arc (southern Tyrrhenian Sea). Journal of Volcanological and Geothermal Research (2002) 115:367–395.[CrossRef]
Carminati E, Negredo AM, Valera JL, Doglioni C. Subduction-related intermediate-depth and deep seismicity in Italy: insights from thermal and rheological modelling. Physics of the Earth and Planetary Interiors (2005) 149:65–79.[CrossRef][Web of Science]
Catalano R, Doglioni C, Merlini S. On the Mesozoic Ionian Basin. Geophysical Journal International (2001) 144:49–64.[CrossRef][Web of Science]
Cavazza W, Wezel FC. The Mediterranean region—a geological primer. Episodes (2003) 26:160–168.[Web of Science]
Class C, Miller DM, Goldstein SL, Langmuir CH. Distinguishing melt and fluid subduction components in Umnak Volcanics, Aleutian Arc. Geochemistry, Geophysics, Geosystems (2000) 1. paper number 1999GC000010.
Condomines M, Sigmarsson O. Why are so many magmas close to 238U–230Th radioactive equilibrium? Geochimica et Cosmochimica Acta (1993) 57:4491–4497.[CrossRef][Web of Science]
Conticelli S, Peccerillo A. Petrology and geochemistry of potassic and ultrapotassic volcanism in central Italy: petrogenesis and inferences on the evolution of the mantle sources. Lithos (1992) 28:221–240.[CrossRef][Web of Science]
Conticelli S, D’Antonio M, Pinarelli L, Civetta L. Source contamination and mantle heterogeneity in the genesis of Italian potassic and ultrapotassic volcanic Rocks: Sr–Nd–Pb isotope data from Roman Province and Southern Tuscany. Mineralogy and Petrology (2002) 74:189–222.[CrossRef][Web of Science]
Davies JH, Stevenson DJ. Physical model of source region of subduction zone volcanics. Journal of Geophysical Research (1992) 97(B2):2037–2070.
De Astis G, Peccerillo A, Kempton PD, La Volpe L, Wu TW. Transition from calc-alkaline to potassium-rich magmatism in subduction enviroments: geochemical and Sr, Nd, Pb isotopic constraints from the island of Vulcano (Aeolian Arc). Contributions to Mineralogy and Petrology (2000) 139:684–703.
Del Moro A, Fornelli A, Piccarreta G. Disequilibrium melting in granulite-facies metasedimentary rocks of the Northern Serre (Calabria–Southern Italy). Mineralogy and Petrology (2000) 70:89–104.[CrossRef][Web of Science]
DePaolo DJ, Wasserburg GJ. Petrogenetic mixing models and Nd–Sr iotopic patterns. Geochimica et Cosmochimica Acta (1979) 43:615–627.[CrossRef][Web of Science]
De Rosa R, Guillou H, Mazzuoli R, Ventura G. New unspiked K–Ar ages of volcanic rocks of the central and western sector of the Aeolian Islands: reconstruction of the volcanic stages. Journal of Volcanological and Geothermal Research (2003) 120:161–178.[CrossRef]
Dewey JF, Helman ML. Kinematiscs of the Western Mediterranean. Alpine Tectonics, Geological Society, London, Special Publications—Coward MP, Dietrich D, Park RG, eds. (1989) 45:265–283.
Di Battistini G, Montanini A, Vernia L, Venturelli G, Tonarini S. Petrology of melilite-bearing rocks from the Montefiascone Volcanic Complex, Roman Magmatic Province: new insights into the ultrapotassic volcanism of Central Italy. Lithos (2001) 59:1–24.[CrossRef][Web of Science]
Di Leo P, Dinelli E, Mongelli G, Schiattarella M. Geology and geochemistry of Jurassic pelagic sediments, Scisti silicei Formation, southern Apennines, Italy. Sedimentary Geology (2002) 150:229–246.[CrossRef][Web of Science]
Di Stefano R, Chiarabba C, Lucente F, Amato A. Crustal and uppermost mantle structure in Italy from the inversion of P-wave arrival times: geodynamic implications. Geophysical Journal International (1999) 139:483–498.[CrossRef][Web of Science]
Ellam RM, Hawkesworth CJ, Menzies MA, Rogers NW. The volcanism of Southern Italy: role of subduction and the relationship between potassic and sodic alkaline magmatism. Journal of Geophysical Research (1989) 94:4589–4601.
Elliott TR. Fractionation of U and Th during mantle melting: a reprise. Chemical Geology (1997) 139:165–183.[CrossRef][Web of Science]
Elliott TR, Plank T, Zindler A, White W, Bourdon B. Element transport from slab to volcanic front at the Mariana arc. Journal of Geophysical Research (1997) 102:14991–15019.[CrossRef]
Faccenna C, Funiciello F, Giardini D, Lucente P. Episodic back-arc extension during restricted mantle convection in the Central Mediterranean. Earth and Planetary Science Letters (2001) 187:105–116.[CrossRef][Web of Science]
Faccenna C, Jolivet L, Piromallo C, Morelli A. Subduction and the depth of convection in the Mediterranean. Journal of Geophysical Research (2003) 108(B2):2099. doi:10.1029/2001JB001690.[CrossRef]
Faccenna C, Funiciello C, Civetta L, D’Antonio M, Moroni M. Slab disruption, mantle circulation, and the opening of the Tyrrhenian basins. Cenozoic Volcanism in the Mediterranean Area. Geological Society of America, Special Papers—Beccaluva L, Bianchini G, Wilson M, eds. (2007) 418:153–169.
Falsaperla S, Lanzafame G, Longo V, Spampinato S. Regional stress field in the area of Stromboli (Italy): insights into structural data and crustal tectonic earthquakes. Journal of Volcanological and Geothermal Research (1999) 88:147–166.[CrossRef]
Finetti IR. Ionian and Alpine Neotethyan Ocean opening. In: Atlases in Geosciences. CROP PROJECT: Deep Seismic Exploration of the Central Mediterranean and Italy—Finetti IR, ed. (2005) Amsterdam: Elsevier. 103–107.
Fornelli A, Piccarreta G, Del Moro A, Acquafredda P. Multi-stage melting in the lower crust of the Serre (Southern Italy). Journal of Petrology (2002) 43:2191–2217.
Francalanci L, Barbieri M, Manetti P, Peccerillo A, Tolomeo L. Sr isotopic systematics in volcanic rocks from the island of Stromboli (Aeolian arc). Chemical Geology (1988) 73:164–180.
Francalanci L, Manetti P, Peccerillo A. Volcanological and magmatological evolution of Stromboli volcano (Aeolian Islands): the roles of fractional crystallisation, magma mixing, crustal contamination and source heterogeneity. Bullettin of Volcanology (1989) 51:355–378.[CrossRef]
Francalanci L, Taylor SR, McCulloch MT, Woodhead DJ. Geochemical and isotopic variations in the calc-alkaline rocks of Aeolian Arc (Southern Tyrrhenian Sea, Italy): constraints on the magma genesis. Contributions to Mineralogy and Petrology (1993) 113:300–313.[CrossRef][Web of Science]
Francalanci L, Tommasini S, Conticelli S. The volcanic activity of Stromboli in the 1906–1998 A.D. period: mineralogical, geochemical and isotope data relevant to the understanding of the plumbing system. Journal of Volcanology and Geothermal Research (2004) 131:179–211.[CrossRef][Web of Science]
Francalanci L, Avanzinelli R, Tommasini S. A west–east geochemical and isotopic traverse along the subaerial volcanism of the Aeolian Island arc, Southern Tyrrhenian Sea, Italy: Inferences on mantle source processes. Cenozoic Volcanism in the Mediterranean Area. Geological Society of America, Special Papers—Beccaluva L, Bianchini G, Wilson M, eds. (2007) 418:235–263.
Gasperini D, Blichert-Toft J, Bosh D, Del Moro A, Macera P, Albarède F. Upwelling of deep mantle material through a plate window: evidence from geochemistry of Italian basaltic volcanics. Journal of Geophysical Research (2002) 107(B2):2367–2386.[CrossRef]
Gillot PY, Keller J. Radiochronological dating of Stromboli. Acta Vulcanologica (1993) 3:69–77.
Gueguen E, Doglioni C, Fernandez M. On the post-25 Ma geodynamic evolution of the western Mediterranean. Tectonophysics (1998) 298:259–269.[CrossRef][Web of Science]
Gvirtzman Z, Nur A. The formation of Mount Etna as the consequence of slab rollback. Nature (1999) 401:782–785.[CrossRef]
Gvirtzman Z, Nur A. Residual topography, lithospheric structure and sunken slabs in the central Mediterranean. Earth and Planetary Science Letters (2001) 187:117–130.[CrossRef][Web of Science]
Hawkesworth CJ, Vollmer R. Crustal contamination versus enriched mantle: 143Nd/144Nd and 87Sr/86Sr evidence from the Italian volcanics. Contributions to Mineralogy and Petrology (1979) 69:151–165.[CrossRef][Web of Science]
Hawkesworth CJ, Hergt JM, Ellam RM, McDermott F. Element fluxes associated with subduction related magmatism. Philosophical Transactions of the Royal Society of London, Series A (1991) 335:393–405.[CrossRef]
Hawkesworth CJ, Turner SP, Peate DW, McDermott F, van Calsteren P. U–Th isotopes in arc magmas: implications for element transfer from the subducted crust. Science (1997) 276:551–555.[CrossRef][Web of Science][Medline]
Heumann A, Davies GR, Elliott T. Crystallization history of rhyolites at Long Valley, California, inferred from combined U-series and Rb–Sr isotope systematics. Geochimica and Cosmochimica Acta (2002) 66:1821–1837.[CrossRef]
Hochstaedter A, Gill J, Peters R, Broughton P, Holden P. Across-arc geochemical trends in the Izu–Bonin arc: Contributions from the subducting slab. Geochemistry, Geophysics, Geosystems (2001) 2. paper number 2000GC000105.
Hornig-Kjarsgaard I, Keller J, Koberski U, Stadlbauer E, Francalanci L, Lenhart R. Geology, stratigraphy and volcanological evolution of the island of Stromboli. Aeolian Arc, Italy. Acta Vulcanologica (1993) 3:21–68.
Iwamori H. Phase relations of peridotites under H2O-saturated conditions and ability of subducting plates for transportation of H2O. Earth and Planetary Science Letters (2004) 227:57–71.[CrossRef][Web of Science]
Johnson MC, Plank T. Dehydration and melting experiments constrain the fate of subducted sediments. Geochemistry, Geophysics, Geosystems (1999) 1. paper number 1999GC000014.
Keppler H. Constraints from partitioning experiments on the composition of subduction-zone fluids. Nature (1996) 380:237–240.[CrossRef]
Kessel R, Schmidt MW, Ulmer P, Pettke T. Trace element signature of subduction-zone fluids, melts and supercritical liquids at 120–180 km depth. Nature (2005) 437:724–727.[CrossRef][Medline]
Le Maitre RW, Streickeisen A, Zanettin B, Le Bas MJ, Bonin B, Bateman P, Bellieni G, Dudek A, Efremova S, Keller J, Lameyre J, Sabine PA, Schmid R, Sørensen H, Woolley AR. Igneous Rocks. A Classification and Glossary of Terms (2002) Cambridge: Cambridge University Press. 236.
Luais B. Mantle mixing and crustal contamination as the origin of the high-Sr radiogenic magmatism of Stromboli (Aeolian Arc). Earth and Planetary Science Letters (1988) 95:411–416.[Web of Science]
Mattei M, D’Agostino N, Faccenna C, Piromallo C, Rossetti F. Some remarks on the geodynamics of the Italian region. Periodico di Mineralogia (2004) 74:7–27.
McDermott F, Defant MJ, Hawkesworth CJ, Maury RC, Joron JL. Isotope and trace element evidence for three component mixing in the genesis of the North Luzon lavas (Philippines). Contributions to Mineralogy and Petrology (1993) 113:9–23.[CrossRef][Web of Science]
McDonough WF. Constraints on the composition of the continental lithospheric mantle. Earth and Planetary Science Letters (1990) 101:1–18.[CrossRef][Web of Science]
Meletti C, Patacca E, Scandone P. Construction of a seismotectonic model: the case of Italy. Pure and Applied Geophysics (2000) 157:11–35.[CrossRef][Web of Science]
Melluso L, Conticelli S, D’Antonio M, Mirco NP, Saccani E. Petrology and mineralogy of wollastonite- and melilite-bearing paralavas from the Central Apennines, Italy. American Mineralogist (2003) 88:1287–1299.
Montuori C, Cimini GB, Favali P. Teleseismic tomography of the southern Tyrrhenian subduction zone: New results from seafloor and land recordings. Journal of Geophysical Research (2007) 112:B03311. doi:10.1029/2005JB004114.[CrossRef]
Morelli C, Giese P, Cassinis R, Colombi B, Guerra I, Luongo G, Scarascia S, Schutte KG. Crustal structure of Southern Italy. A seismic refraction profile between Puglia–Calabria–Sicily. Bollettino di Geofisica Teorica e Applicata (1975) 18:183–210.
Pasquale V, Verdoya M, Chiozzi P. Thermal structure of the Ionian Slab. Pure and Applied Geophysics (2005) 162:967–986.[CrossRef][Web of Science]
Panza GF. Lateral variations in the lithosphere in correspondence of the southern segment of EGT. In: Second EGT Workshop: the Southern Segment—Galson DA, Mueller S, eds. (1985) Strasbourg: European Science Foundation. 47–51.
Panza GF, Pontevivo A, Chimera G, Raykova R, Aoudia A. The lithosphere–asthenosphere: Italy and surroundings. Episodes (2003) 26:169–174.[Web of Science]
Peacock SM, Rushmer T, Thompson AB. Partial melting of subducting oceanic crust. Earth and Planetary Science Letters (1994) 121:227–244.[CrossRef][Web of Science]
Peccerillo A. Geochemical similarities between the Vesuvio, Phlegrean Fields and Stromboli volcanoes: petrogenetic, geodynamic and volcanological implications. Mineralogy and Petrology (2001) 73:93–105.[CrossRef][Web of Science]
Peccerillo A. Plio-Quaternary magmatism in Italy. Episodes (2003) 26:222–226.[Web of Science]
Peccerillo A, Panza GF. Upper mantle domains beneath central–southern Italy: petrological, geochemical and geophysical constraints. Pure and Applied Geophysics (1999) 156:421–443.[CrossRef][Web of Science]
Peccerillo A, Kempton PD, Harmon RS, Wu TW, Santo AP, Boyce AJ, Tripodo A. Petrological and geochemical characteristics of the Alicudi volcano, Aeolian Islands, Italy: implications for magma genesis and evolution. Acta Vulcanologica (1993) 3:235–249.
Peccerillo A, Dallai L, Frezzotti ML, Kempton PD. Sr–Nd–Pb–O isotopic evidence for decreasing crustal contamination with ongoing magma evolution at Alicudi volcano (Aeolian arc, Italy): implications for style of magma–crust interaction and for mantle source compositions. Lithos (2004) 78:217–233.[CrossRef][Web of Science]
Plank T, Langmuir CH. The chemical composition of subducting sediment and its consequences for the crust and mantle. Chemical Geology (1998) 145:325–394.[CrossRef][Web of Science]
Poli S, Schmidt MW. H2O transport and release in subduction zones: experimental constraints on basaltic and andesitic systems. Journal of Geophysical Research (1995) 100:22299–22314.[CrossRef]
Pontevivo A, Panza GF. The lithosphere–asthenosphere system in the Calabrian Arc and surrounding seas—Southern Italy. Pure and Applied Geophysics (2006) 163:1617–1659.[CrossRef][Web of Science]
Roselle GT, Ring MT, Engi M. Melonpit: a finite element code for simulating tectonic movement and heat flow within subduction zones. American Journal of Science (2002) 302:381–409.
Rüpke LH, Phipps Morgan J, Hort M, Connolly JAD. Serpentine and the subduction zone water cycle. Earth and Planetary Science Letters (2004) 223:17–34.[CrossRef][Web of Science]
Santo AP, Chen Y, Clark AH, Farrar E, Abebe T. 40Ar/39Ar ages of the Filicudi island volcanics: implications for the volcanological history of the Aeolian Arc, Italy. Acta Vulcanologica (1995) 7:7–12.
Schmidt MW, Vielzeuf D, Auzanneau E. Melting and dissolution of subducting crust at high pressures: the key role of white mica. Earth and Planetary Science Letters (2004) 228:65–84.[CrossRef][Web of Science]
Scrocca D, Doglioni C, Innocenti F. Constraints for an interpretation of the Italian geodynamics: a review. Memorie Descrittive della Carta Geologica d’Italia (2003) LXII:15–46.
Sigmarsson O, Martin H, Knowles J. Melting of a subducting oceanic crust from U–Th disequilibria in austral Andean lavas. Nature (1998) 394:566–569.[CrossRef]
Stalder R, Ulmer P, Thompson AB, Günter D. High pressure fluids in the system MgO–SiO2–H2O under upper mantle conditions. Contributions to Mineralogy and Petrology (2001) 140:607–618.[Web of Science]
Stampfli G. Plate tectonics of the Apulia–Adria microcontinents. In: Atlases in Geosciences. CROP PROJECT: Deep Seismic Exploration of the Central Mediterranean and Italy—Finetti IR, ed. (2005) Amsterdam: Elsevier. 747–766.
Staudigel H, Davies GR, Hart SR, Marchant KM, Smith BM. Large scale isotopic Sr, Nd and O isotopic anatomy of altered oceanic crust: DSDP/ODP sites 417/418. Earth and Planetary Science Letters (1995) 130:169–185.[CrossRef][Web of Science]
Stracke A, Hofmann AW, Hart SR. FOZO, HIMU, and the rest of the mantle zoo. Geochemistry, Geophysics, Geosystems (2005) 6. Q05007 doi:10.1029/2004GC000824.
Sun SS, McDonough WF. Chemical and isotopic systematics of ocean basalts: implication for mantle composition and processes. Magmatism in the Ocean Basins. Geological Society, London, Special Publications—Saunders AD, Norry MJ, eds. (1989) 42:313–345.
Tatsumi Y. Migration of fluid phases and genesis of basalt magmas in subduction zones. Journal of Geophysical Research (1989) 94:4697–4707.
Taylor SR, McLennan SM. The Continental Crust: its Composition and Evolution (1985) Oxford: Blackwell Scientific Publications. 312.
Tonarini S, Leeman WP, Ferrara G. Boron isotopic variations in lavas of the Aeolian volcanic arc, South Italy. Journal of Volcanological and Geothermal Research (2001) 110:155–170.[CrossRef]
Trua T, Serri G, Marani M, Renzulli A, Gamberi F. Volcanological and petrological evolution of Marsili seamount (southern Tyrrhenian Sea). Journal of Volcanological and Geothermal Research (2002) 114:441–464.[CrossRef]
Turner SP, Foden J. U, Th, Ra disequilibria, Sr, Nd, Pb isotope and trace element variation in Sunda arc lavas: predominance of a subducted sediment component. Contributions to Mineralogy and Petrology (2001) 142:43–57.[Web of Science]
Turner SP, Hawkesworth CJ, van Calsteren P, Heath E, Macdonald R, Black S. U-series isotopes and destructive plate margin magma genesis in the Lesser Antilles. Earth and Planetary Science Letters (1996) 142:191–207.[CrossRef][Web of Science]
Turner SP, Hawkesworth CJ, Rogers N, Bartlett J, Worthington T, Hergt J, Pearce J, Smith I. 238U–230Th disequilibria, magma petrogenesis and flux rates beneath the depleted Tonga–Kermadec Island arc. Geochimica et Cosmochimica Acta (1997) 61:4855–4884.[CrossRef][Web of Science]
Turner SP, Bourdon B. Insights into magma genesis at convergent margins from U-series isotopes. Uranium Series Geochemistry. Mineralogical Society of America, Reviews in Mineralogy and Geochemistry—Bourdon B, Henderson GM, Lundstrom CC, Turner SP, eds. (2003) 52:255–315.
Vroon PZ, van Bergen MJ, White WM, Varekamp JG. Sr–Nd–Pb isotope systematics of the Banda Arc, Indonesia: combined subduction and assimilation of continental material. Journal of Geophysical Research (1993) 98(B12):22349–22366.[CrossRef]
Vroon PZ, van Bergen MJ, Klaver GJ, White WM. Strontium, neodymium, and lead isotopic and trace-element signatures of the East Indonesian sediments: provenance and implications for Banda arc magma genesis. Geochimica et Cosmochimica Acta (1995) 59:2573–2598.[CrossRef][Web of Science]
Weldeab S, Emeis KC, Hemleben C, Siebel W. Provenance of lithogenic surface sediments and pathways of riverine suspended matter in the Eastern Mediterranean Sea: evidence from 143Nd/144Nd and 87Sr/86Sr ratios. Chemical Geology (2002) 186:139–149.[CrossRef][Web of Science]
Workman RH, Hart SR. Major and trace element composition of the depleted MORB mantle (DMM). Earth and Planetary Science Letters (2005) 231:53–72.[CrossRef][Web of Science]
Wortel MJR, Spakman W. Subduction and slab detachment in the Mediterranean–Carpathian Region. Science (2000) 290:1910–1917.[CrossRef][Web of Science][Medline]
Zindler A, Hart SR. Chemical geodynamics. Annual Review of Earth and Planetary Sciences (1986) 14:493–571.[CrossRef][Web of Science]
Zou H, Zindler A. Theoretical studies of 238U–230Th–226Ra and 235U–231Pa disequilibria in young lavas produced by mantle melting. Geochimica et Cosmochimica Acta (2000) 64:1809–1817.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. Boari, S. Tommasini, M. A. Laurenzi, and S. Conticelli Transition from Ultrapotassic Kamafugitic to Sub-alkaline Magmas: Sr, Nd, and Pb Isotope, Trace Element and 40Ar-39Ar Age Data from the Middle Latin Valley Volcanic Field, Roman Magmatic Province, Central Italy J. Petrology, February 19, 2009; (2009) egp003v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Avanzinelli, T. Elliott, S. Tommasini, and S. Conticelli Constraints on the Genesis of Potassium-rich Italian Volcanic Rocks from U/Th Disequilibrium J. Petrology, February 1, 2008; 49(2): 195 - 223. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Boari and S. Conticelli MINERALOGY AND PETROLOGY OF ASSOCIATED Mg-RICH ULTRAPOTASSIC, SHOSHONITIC, AND CALC-ALKALINE ROCKS: THE MIDDLE LATIN VALLEY MONOGENETIC VOLCANOS, ROMAN MAGMATIC PROVINCE, SOUTHERN ITALY Can Mineral, December 1, 2007; 45(6): 1443 - 1469. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||