Mazzeo, Fabio Carmine (2014) CARATTERI GEOCHIMICI DEL MANTELLO SORGENTE DEL MAGMATISMO NAPOLETANO: NUOVE CONOSCENZE DALLO STUDIO DELLE OFIOLITI DEL SETTORE LUCANO DELL’APPENNINO MERIDIONALE. [Tesi di dottorato]

[img]
Preview
Text
fabio_carmine_mazzeo_26.pdf

Download (35MB) | Preview
[error in script] [error in script]
Item Type: Tesi di dottorato
Lingua: Italiano
Title: CARATTERI GEOCHIMICI DEL MANTELLO SORGENTE DEL MAGMATISMO NAPOLETANO: NUOVE CONOSCENZE DALLO STUDIO DELLE OFIOLITI DEL SETTORE LUCANO DELL’APPENNINO MERIDIONALE
Creators:
CreatorsEmail
Mazzeo, Fabio Carminefabiocarminemazzeo@libero.it
Date: 28 March 2014
Number of Pages: 380
Institution: Università degli Studi di Napoli Federico II
Department: Scienze della Terra, dell'Ambiente e delle Risorse
Scuola di dottorato: Scienze della terra
Dottorato: Scienze della Terra
Ciclo di dottorato: 26
Coordinatore del Corso di dottorato:
nomeemail
Boni, Mariaboni@unina.it
Tutor:
nomeemail
D'Antonio, MassimoUNSPECIFIED
Petrosino, PaolaUNSPECIFIED
Date: 28 March 2014
Number of Pages: 380
Uncontrolled Keywords: Ophiolites; Peridotites; Ligurian Tethys; Neapolitan Volcanism; Subduction-relates magmatism;
Settori scientifico-disciplinari del MIUR: Area 04 - Scienze della terra > GEO/08 - Geochimica e vulcanologia
Date Deposited: 07 Apr 2014 12:02
Last Modified: 14 May 2016 01:00
URI: http://www.fedoa.unina.it/id/eprint/9720

Abstract

Introduction The Mediterranean area is one of the most complex geodynamic settings of the world (e.g., Carminati et al., 2012, and references therein) as clearly illustrated by the huge variety of igneous lithologies. On the basis of trace element concentrations, and isotopic compositions, the latter being extremely variable from typical mantle to typical crustal values, sectors characterized by either anorogenic (Lustrino and Wilson, 2007) or orogenic magmatism (Harangi et al., 2006; Lustrino et al., 2011) are usually distinguished in the Central-Western Mediterranean area, including Italy. Since the Cenozoic to the Present, the Central-Western Mediterranean area has been the site of intense but discrete magmatic activity. The products show a wide compositional range, from sub-alkaline (tholeiitic and calc-alkaline) to alkaline (sodic, potassic, and ultrapotassic) and from mafic/ultramafic to felsic (Peccerillo, 2005, and references therein). Many are the hypotheses about magmagenesis and geodynamic significance of these rocks (e.g., Peccerillo and Lustrino, 2005), but most of the Mediterranean orogenic magmatism might reasonably be the result of partial melting of mantle sources modified by slab materials during and/or after subduction event(s). For the Italian magmatism in particular, this hypothesis is supported by a considerable amount of studies that highlight its post-collisional character (e.g., Peccerillo, 1999; Beccaluva et al., 1991; Conticelli et al., 2002, 2004, 2007, 2009; D’Antonio et al., 1996, 1999a, 2007, 2013; Francalanci et al., 2004, 2007; Tonarini et al., 2004; Duggen et al., 2005; Harangi et al., 2006; Avanzinelli et al., 2008, 2009; Bianchini et al., 2008; Prélevic et al., 2008; Nikogosian and van Bergen, 2010; Lustrino et al., 2011). Furthermore, the Italian magmatism has been related to the subduction of the Ionian oceanic lithosphere, being part of the wider Tethys Ocean (Gvirtzman and Nur, 1999; Faccenna et al., 2007). The genesis of magmas in the “subduction factory” is very complex because of the large number of variable factors involved, such as the original chemical and mineralogical composition of mantle, presence of fluids and their nature, depth of the source, degree of partial melting, temperature and confining pressure (e.g., Class et al., 2000; Hanyu et al., 2006; Handley et al., 2007; Leslie et al., 2009; Vigoroux et al., 2012). Nevertheless, worldwide subduction-related mafic volcanic rocks have distinctive incompatible elements patterns, characterized by strong enrichment in LILE and Pb with respect to HFSE (noticeable is the Nb-Ta-Ti depletion), and REE. This depends on the interaction between pre-enrichment mantle wedge and enriching agents coming from the slab. The mantle wedge before the enrichment is generally considered either MORB-like (e.g., Münker, 2000) or OIB-like (e.g., Peate et al., 1997). The exact nature can be inferred by investigating elements, such as some HFSE, not mobilized by any fluid phase during the subduction process. Concerning the nature of the enriching agents there are several hypotheses. Some authors consider the main role of hydrous fluids released during the dehydration of basaltic crust and pelagic sediments (e.g., Ryan et al., 1995; Bizimis et al., 2000; Dorendorf et al., 2000; Handley et al., 2007; Vigoroux et al., 2012); others invoke also the involvement of partial melting of sediments (e.g., Elliott et al., 1997; Turner et al., 1997; Elburg and Foden, 1998; Johnson and Plank, 1999; Class et al., 2000; Hanyu et al., 2006; Leslie et al., 2009). The isotopic and trace element characteristics of the basaltic and sedimentary cover in a subducting slab are generally different, though hardly quantifiable. The different mobility of some incompatible trace elements in slab-derived hydrous fluids and melts is typically used in geochemical modeling to distinguish between these two extreme cases (Brenan et al., 1995; Ayers, 1998; Johnson and Plank, 1999; Pearce et al., 1999; Becker et al., 2000; Walker et al., 2001; Hanyu et al., 2006; Handley et al., 2007; Leslie et al., 2009; Vigoroux et al., 2012). LILE are more easily mobilized from slab and transferred to the mantle if the metasomatizing agents are hydrous fluids. On the other hand, HFSE have a little mobility because of their low compatibility in hydrous fluids (Plank and Langmuir, 1998). Conversely, if the main enriching agents are melts, HFSE can be partly mobilized together with LILE. However, it is not always easy to determine which elements, and in which amounts are retained in the slab, and which transferred to the fluid phase (hydrous and/or melts) because this depends on the presence of residual mineral phases in the slab and the geothermal gradient. Anyway, if the behavior of trace elements is taken into account, it is possible to discriminate the different contributions provided by hydrous fluids and/or melts to the mantle enrichment. Current views of the enrichment of mantle beneath the Central-Western Mediterranean area involve either subduction-derived material (e.g., Peccerillo, 1999; Beccaluva et al., 1991; Conticelli et al., 2002, 2004, 2007, 2009; Francalanci et al., 2004, 2007; Tonarini et al., 2004; Duggen et al., 2005; Harangi et al., 2006; Avanzinelli et al., 2008, 2009; Bianchini et al., 2008; Prélevic et al., 2008; Nikogosian and van Bergen, 2010; Lustrino et al., 2011) or fluids metasomatizing a plume-type mantle source, calling for a within-plate geodynamic setting (Bell et al., 2013, and references therein). The latter hypothesis is simply ruled out on the basis of the major oxides content of the mafic volcanic rocks, that are typical of melts coming from restitic mantle sources, at variance with the typical fertile sources of plumes (as discussed, for instance, by Conticelli et al., 2004). In the framework of the complex magmatism of Western Mediterranean, the volcanic area of Campania (Southern Italy) is one of the most interesting. This area has been the site of intense volcanism during Plio-Quaternary times. Over the past ~40 ka, volcanism has been localized mainly at the Mt. Somma-Vesuvius complex (SV) and the Phlegrean Volcanic District (PVD), that includes the Campi Flegrei caldera as well as the Ischia and Procida islands in the Gulf of Napoli (Orsi et al., 1996). The products of all these volcanoes show geochemical and isotopic features typical of subduction-related volcanic rocks. The aim of this study is twofold. The first aim is to better identify the composition of the pre-enrichment mantle sector underlying the Neapolitan volcanic area, still poorly known due to the scarcity of suitable primitive mafic rocks; the second aim is to characterize the nature of the subduction-related components (fluids and/or melts from sediments and/or altered oceanic crust) that modified the pre-enriched mantle sector. In previous attempts to model the mantle enrichment for the PVD, average subducted slab material compositions from literature have been utilized (D’Antonio et al., 1999a, 2007, 2013; Tonarini et al., 2004; Piochi et al., 2004). For the purposes of the present work, the igneous rocks and associated sedimentary cover of the Mt. Pollino ophiolitic sequences, have been studied as possible contaminants that more realistically enriched the mantle source region of Campania during subduction. They represent ocean-derived terrains obducted on the western margin of the Adria continental micro-plate during the Apennine orogenesis (Vitale and Ciarcia, 2013, and references therein), and should be the best representative of the material that was subducted into the Western Mediterranean mantle during the closure of the Ligurian branch of Tethys (Bortolotti and Principi, 2005 and references therein). Abyssal peridotites and Alpine-type peridotite ophiolite massifs, together with mantle xenoliths associated with alkaline rocks, provide a good opportunity to investigate the physical state, chemical composition and mineralogy of the upper mantle. As a result of both alteration and metamorphism, ophiolite rocks commonly occur on Earth surface with chemical and mineralogical modifications. However, they have a wider distribution with respect to mantle xenoliths. Also, knowledge of ophiolites and modern oceanic lithosphere were fundamental for the development of a conceptual model of the oceanic crust. According to this model, the oceanic crust consists of a 4 to 6 km thick igneous crust (pillow basalts, sheeted dykes and gabbroic layer, from top downward) produced by melts formed by decompression melting of ascending asthenospheric mantle, overlying a peridotite basement. Moreover, the igneous pile may be overlaid by a few hundred meters of pelagic/terrigenous sediments. During the last thirty years, the study of peridotites from the Mediterranean Area has provided a wealth of information, constraining the composition and, partially, the evolution of the Mediterranean Lithospheric Mantle. In Italy, ophiolites occur in scattered outcrops located mainly in the Alps and Northern Apennine (Piccardo et al., 2009 and references therein). In Southern Apennine, ophiolite outcrops are very rare, occurring only in southern Basilicata and northern Calabria (Beccaluva et al., 1983; Spadea, 1994) and like as olistoliths in Miocene turbidites of Cilento Group at Monte Centauino (Di Girolamo et al., 1992) in Southern Campania. Ophiolites from the Alpine and Apennine orogenic terranes are believed to represent fragments of oceanic lithosphere of the Ligurian-Piedmontese (Ligurian Tethys) basin that formed during Late Jurassic between the Europe and Adria continental blocks, following extension driven by far field tectonic forces and that were obducted on continental crust during the closure of this ocean (Bortolotti and Principi, 2005). More information on the mineralogy and geochemistry of the Mediterranean upper mantle has been derived especially from the largely studied Northern Apennine ophiolite sequences. In the past decades, the Southern Apennine ophiolites have been largely studied (Lanzafame et al., 1978, 1979a, 1979b; Beccaluva et al., 1983; Di Girolamo et al., 1992; Spadea, 1994; Piluso et al., 2000; Liberi et al., 2006; Cristi Sansone et al., 2011) but no significant petrographic, geochemical and isotopic inferences on the upper mantle they represent have been reported. The use of Tethyan sediments as a contaminant for the mantle could better constrain the enrichment event(s) of the mantle sector underlying the PVD, allowing a more complete understanding of the complex geochemical and isotopic features of erupted magmas. To achieve this goal, new geochemical and Sr-Nd-isotopic data are presented on representative samples of the pillow-lavas and meta-sedimentary cover of the Timpa delle Murge Formation ophiolites, and of the Crete Nere Formation, in order to fully characterize the material subducted during the orogenesis. Using the acquired data a geochemical and isotopic model is proposed to highlight a possible link between the oceanic crust subducted during the Tethys closure and the geochemical features of the Plio-Quaternary subduction-related Neapolitan Volcanic Area, as these sedimentary and igneous rocks, or other similar, may have changed the composition of the upper mantle underlying a large region of the Mediterranean Area. PART 1: Petrological characterization of Mt. Pollino ophiolites Geological setting In Italy, ophiolites occur in scattered outcrops located mainly in the Alps and Northern Apennine (Robertson, 2002; Bortolotti and Principi, 2005). In Southern Apennine, ophiolite outcrops are very rare, occurring only in Southern Campania, Basilicata and Northern Calabria (Beccaluva et al., 1983; Di Girolamo et al., 1992; Spadea, 1994; Tortorici et al., 2009; Vignaroli et al., 2009). Southern Apennine ophiolites are in many ways different from tipical ophiolite sequences. In fact, the Alpine-Apennine ophiolites consist of a serpentinized peridotite basement (direct exposure of mantle peridotites on the sea-floor) and a reduced crustal sequence characterized by lack of sheeted-dyke complexes, relatively small volumes of gabbros intruded in the peridotite basement, and a discontinuous basaltic and oceanic sediments cover. They are believed to represent fragments of Tethys oceanic crust that were obducted on continental crust during the closure of the Ligurian branch of Tethys ocean prior to 35 Ma ocean (Bortolotti and Principi, 2005; Liberi et al., 2006). Bonardi et al. (1988) ascribed these fragments to the Liguride Complex, a Cretaceous-Oligocene accretionary complex representing the highest structural element of the Apennine Chain, thrust over the Mesozoic carbonate units. In the Mt. Pollino area, the Liguride Complex covers the carbonate terrains of the Alburno-Cervati Unit, and includes three distinct tectonic units named Frido Unit, Episcopia-San Severino Mélange, and North Calabrian Unit. The Episcopia-San Severino Mélange includes dark green cataclastic serpentinite exposed in scattered outcrops, likely representing fragments of an upper mantle portion, associated with garnet gneisses and amphibolites (Spadea, 1982), recently attributed to old crust, both oceanic and continental (Frido Unit; Vitale et al., 2013). The North Calabrian Unit, divided into Timpa delle Murge Formation, Crete Nere Formation, and Saraceno Formation from base upwards, crops out widely on Timpa delle Murge Hill, along the boundary between Basilicata and Calabria, close to Mt. Pollino. In particular, the Timpa delle Murge formation is an ophiolite sequence including small gabbroid bodies, abundant and well-preserved pillow lavas, and a pelagic sedimentary cover. The latter is made up of silicified red-green shales affected by a pencil cleavage, containing several layers of quartz-arenites, followed by green-red radiolarites with nodular structures, with thin intercalations of calpionella marly limestones. Upwards, the succession continues with intercalated quartz-arenites and black shales that mark the transition to the Crete Nere Formation. The latter covers a large time span, from Late Cretaceous to Late Eocene, although most deposition occurred in Middle Eocene (Bonardi et al., 1988b). The radiolarian cherts of the Timpa delle Murge Formation were paleontologically dated at ~160 Ma (Marcucci et al., 1987). In our opinion, this age can be confidentially considered to define the end of oceanic crust generation in that area. Given the marked temporal hiatus between the two formations, it is clear that the Timpa delle Murge ophiolites were tectonically embedded in the much younger Crete Nere Formation terranes, that crop out both below and over them. The ophiolite sequences of Basilicata and Calabria were affected by subduction-related HP/LT metamorphism, marking a burying episode followed by exhumation occurred during the Late Oligocene-Early Miocene Apennine orogenesis (Piluso et al., 2000; Liberi et al., 2006; Cristi Sansone et al., 2011). However, the Timpa delle Murge ophiolites suffered such metamorphism to a much lesser extent than all other ophiolite sequences in Southern Italy, attaining the green schist facies only, due to either ocean-floor metamorphism or accretionary wedge tectonic evolution (Cristi Sansone et al., 2011). Thus, the original geochemical and isotopic features of these rocks should be quite well preserved. However, despite the large wealth of studies, no petrographic, geochemical and isotopic data are available for the sedimentary terrains associated with these ophiolites. Field and Laboratory activities Ninety samples of either mafic or ultramafic igneous rocks, and some sedimentary rocks of the ophiolitic sequence that crop out in a few Southern Lucanian localities (Episcopia, San Severino Lucano, Timpa Pietrasasso and Timpa delle Murge) have been collected. At the University of Napoli Federico II, the most representative samples were cut with a saw and crushed in a jaw crusher, and powders were produced in a low-blank agate mortar from clean chips first washed in distilled water. Major oxides and some trace elements (Sc, V, Cr, Ni, Rb, Sr, Ba, Y, Zr and Nb) were analyzed by X-ray fluorescence using a sequential X-ray spectrometer Philips PW2400 at the Centres Cientìfics i Tecnològics de la Universitat de Barcelona (CCiTUB), Spain. The volatile content (LOI) was measured using standard thermogravimetric methods at the CCiTUB. Other trace elements including the Rare Earth Elements (REE) were analyzed by Inductively Coupled Plasma - Mass Spectrometry (ICP-MS) using a Perkin Elmer Elan 6000 at the CCiTUB. Whole rock powders were dissolved with high-purity HF + HNO3 + HClO4 mixtures. Relative precision was generally better than 1–2% for major oxides and better than 5–10% for trace elements. Sr- and Nd-isotopic compositions of Mt. Pollino volcanic and sedimentary rocks were determined on whole rocks (~0.1 g). Whole rock powders were leached with warm HCl for 10 minutes and dissolved with high-purity HF + HNO3 + HCl mixtures. Sr and Nd were separated using standard column chromatographic methods, using Dowex AG50W X-8 (200-400 mesh) and Ln Spec cation exchange resins for Sr and Nd, respectively. Sr- and Nd-isotopic ratios were measured by thermal ionization mass-spectrometry (TIMS) using a ThermoFinnigan Triton TI at the Istituto Nazionale di Geofisica e Vulcanologia, sezione di Napoli, Osservatorio Vesuviano. The internal precision of each measurement is expressed as ± 2 times the standard error (2se), where se = /√n. In the period of sample measurements, the NIST-SRM 987 standard gave a mean value of 87Sr/86Sr = 0.710249 (2 = 1.47 x 10-5; N = 67), and the La Jolla standard gave a mean value of 143Nd/144Nd = 0.511833 (2 = 7.65 x 10-6; N = 21). The Nd-isotopic ratios measured on the samples were normalized to the accepted values of the La Jolla standard (143Nd/144Nd = 0.51185). The mineral compositions were obtained with a SEM-WDS Cameca SX100 microprobe at CCiTUB, equipped with four INCA X-act detectors, operating at a 15kV beam voltage, with a 50–100mA filament current, variable spot size and 50s net acquisition time. Precision and accuracy were controlled using an internal standard. Trace element contents in clinopyroxene and host glassy matrix were determined by laser ablation-inductively coupled plasma-sector field mass spectrometry (LA-ICPSFMS) at CNR–Istituto di Geoscienze e Georisorse (IGG, Pavia, Italy). The microprobe has a double-focusing sector field analyzer (Finnigan Mat, Element I) coupled with a Q-switched Nd:YAG laser source (Quantel Brilliant). The fundamental emission of the laser source (1,064 nm, in the near-IR region) was converted to 213 nm by three harmonic generators. Spot diameter varied in the range of 40–60 m. Precision and accuracy (both better than 10% for concentrations at ppm level) were assessed by means of repeated analyses of NIST SRM 612 and BCR-2g standards. The modal composition of gabbros was determined with an optical microscope, coupled with a Leica DFC-280 camera and software Leica QWin at University of Naples Federico II. Results Ultramafic Rocks All peridotite samples contain large amount of serpentine and other phyllosilicates such as chlorite and pumpellyte. The mineralogical composition of residual primary paragenesis is very homogeneous, so that the original composition must have been peridotite for all samples. Three samples are characterized by millimeter-sized porphyroclasts of olivine and orthopyroxene, varying from anhedral to subhedral. Both olivine and orthopyroxene show internal deformation (deformation lamellae along the slip planes, kink banding and wavy extinction). Clinopyroxene is present as small crystals or as exsolution lamellae in orthopyroxene. Spinels are typically anhedral. Rare is the presence of anhedral crystals of amphibole. The remaining samples show porphyroblastic textures with millimeter-sized porphyroclasts of brecciated olivine, and kink-banded and plastically deformed orthopyroxene. Olivine and orthopyroxene grains are commonly elongated. The elongated coarse grains (up to a few centimeters) give a well-developed foliation to the rock. Spinels occur as disseminated, completely anhedral or amoeboid interstitial grains and sometimes as inclusions in olivine. Clinopyroxene occur as intercumulus small crystals or as exsolution lamellae in orthopyroxene. Olivine is generally uniform in composition and its forsterite (Fo) content is variable in the ranges 90.1–90.6 in peridotites with amphibole, and 91.2–91.9 in peridotites without amphibole (Fig. 7.1). NiO wt% of olivine composition ranges between 0.22 and 0.42 in peridotites with amphibole, and between 0.42 and 0.69 in peridotites without amphibole, so there is a decrease in Fo and NiO values of the olivines from peridotites without amphibole to peridotites with amphibole (Fig. 7.2). The concentrations of Mn, Al, Ca and Ti are negligible. Orthopyroxene (Fig. 7.6) is generally enstatite in composition (En = 88-92 mol%), with Cr2O3 contents up to 1.2 wt% and Al2O3 variable in the range 1.5-2.8 wt.%. Clinopyroxenes are diopside in composition (Fig. 7.6) and show Mg# in the range 88-93, high Cr2O3 content (1-3 wt.%), very high CaO contents (20-24 wt.%) and low TiO2 (< 0.5 wt.%). REE contents of clinopyroxene (Fig 7.36) are about 10xChondrite in MREE and HREE region where the patterns are almost flat, while the LREE contents are considerably depleted, with a significant difference between peridotites with amphibole (LaN = 1-4xChondrite) and peridotites without amphibole (LaN = 0.05-0.5xChondrite). Spinel of peridotites (Fig. 7.9) is chromiferous (Cr# = 19-26 in peridotites with amphibole, and 42-59 in peridotites without amphibole) and aluminiferous (37-44 wt.%). TiO2 contents of spinel are up to 0.11 wt%. The core of grains is usually not affected by alteration and has a low Fe3+ content. Amphiboles are pargasite (Fig. 7.10) with Mg# = 87-90, FeO = 2.3-4.4 wt.%, TiO2 = 3.4-3.9 wt.%, Na2O = 3.5-4.3 wt.%, and Cr2O3 = 1.1-1.8 wt.%. Important evidence of post-partial melting processes are present in serpentinized peridotite rocks of Mt. Pollino. Clinopyroxene exsolution lamellae in orthopyroxene porphyroclasts are a clear effect of a low T sub-solidus re-equilibration (Parkinson and Pearce, 1998). Conversely, the presence of secondary minerals on the edge of larger porphyroclasts, and of interstitial minerals (predominantly clinopyroxene) showing magmatic textures, are interpreted as evidence of melt-rock reaction (Suhr and Edwards, 2000) both in abyssal peridotites (Hellebrand et al., 2002) and in ophiolite massifs (Piccardo et al., 2007a). In addition, thermo-barometric estimates indicate that typical values are rarely preserved in mantle spinel facies (T = 1100-1200 °C and P = 10-11kbar). The LOI values of the serpentinites vary in the range 9-12 wt.% with an average MgO > 41 wt%, but with a visible gap between samples with and without amphibole (Fig 10.21). The three samples with amphibole have the highest values of Al2O3, CaO, SiO2, Sc and V, but lower Ni and Co contents, while the other samples show lower contents of Al2O3, SiO2, CaO, Sc, and V, but higher Ni and Co contents. Excellent correlations were observed in major oxides and some trace elements plotted against MgO. Incompatible elements, Al, Ca, Si, Sc, and V show negative covariance with MgO. In contrast, the compatible elements Ni and Co show positive covariance with MgO. These trends, according to Parkinson and Pearce (1998) are related to increase of the partial melting degree. Compared with peridotites from modern oceanic settings (Fig. 10.26), most of the chromite grains in peridotites with amphibole plot within the field of abyssal peridotites (low Cr# and high Mg#), whereas spinels of peridotites without amphibole fall in the field of fore-arc peridotites (high Cr# and low Mg#). The inverse correlation between Cr# and Mg# of spinel-group minerals in the investigated rocks is consistent with increase in degree of partial melting. According to this trend, the studied serpentinites show a wide variation in degree of partial melting. Regarding the relationship between Fo content of olivine and Cr# of spinel (Fig. 10.27), all rock types fall within the olivine-spinel mantle array (OSMA) of Arai (1994), which is regarded as evidence for their residual origin, thus confirming that the trend is related to partial melting (8% for peridotites with amphiboles and about 20% for peridotites without amphibole). However, the peridotites from Mt. Pollino fall at significantly lower SiO2 contents and at significantly higher FeO contents with respect to the corresponding MgO values on the refractory trends calculated by Niu (1997), indicating that a simple partial melting and melt extraction process cannot be responsible of these compositional features (Fig. 10.23). Indeed, the degree of partial melting based on the LREE concentrations of clinopyroxenes is much lower than that recorded by spinel (<3% for peridotites with amphibole; <6% for peridotites without amphiboles), and this is even lower than those recorded by HREE in clinopyroxenes (Fig. 10.27). On the basis of this geochemical evidence, the depleted spinel peridotites of Mt. Pollino have been interpreted as the result of a reactive melt/rock interaction with the percolating melt during open system migration in spinel-facies condition after melt extraction. Information on the composition of the percolating melts in depleted spinel peridotites of Mt. Pollino have been obtained by the clinopyroxene trace element compositions. Most likely, re-equilibration of clinopyroxenes has occurred by interaction with alkaline liquids. Indeed, the La/Yb ratios of Mt. Pollino peridotites follow the trend of partial melting of a clinopyroxene in equilibrium with DMM-type mantle but, in contrast, the Ti/Eu ratios deviate from this trend, moving toward the compositions shown by clinopyroxenes of alkaline and carbonatitic melts. Summarizing, field, petrographic and compositional features of Mt. Pollino peridotites allow us to interpret these rocks as deriving from an internal portion of the Ligurian Tethys ocean basin, similarly to the ophiolites of the Northern Apennines and the Alps. Mafic rocks Basalts from Mt. Pollino show a very fine-grained, phorphyritic texture, with phenocrysts of plagioclase, often in aggregates and albitized, olivine, and rare pyroxene. In the groundmass, olivine, orthopyroxene and magnetite are recognizable in addition to plagioclase and clinopyroxene. On the basis of the mineral paragenesis they are tholeiitic basalts. Olivines (Fig. 7.26) show a fairly restricted range of composition (Fo77–79) and are characterized by low MnO (<0.15 wt%) and CaO (< 0.1 wt.%). Calculated distribution coefficients (Fig. 10.10) indicate that most olivines are not in equilibrium with host rock (Fe/MgKdOl-liq = 0.30–0.33; Roeder and Emslie, 1970). Temperature values calculated using the Putirka (2008) geothermometer range from 1081 °C to 1254 °C. Clinopyroxene (Fig. 7.27) is augite (Wo35-41En47-53Fs9-16) with Mg# = 71-84, and has low TiO2 content (<1 wt%). Al2O3 tends to increase with decreasing Mg#. Calculated distribution coefficients (Fig. 10.11) for the majority of basaltic samples fall outside the range of clinopyroxene-host rock (Fe/MgKdcpx-liq = 0.24-0.30; Grove and Bryan, 1983). The compositional range of pigeonite is Wo6-12En66-71Fs19-25 with Mg# = 19–40. Temperature values calculated using the Ishii (1975) geothermometer range from 1078 °C to 1156 °C. Plagioclase in the tholeiitic basalts (Fig. 7.31) is mainly labradorite (An83–47). It typically shows normal zoning. Only a few compositions are andesine (An44–33). Calculated distribution coefficients indicate that most plagioclase phenocrysts are in equilibrium with the host rock (Ca/NaKdplag-melt = XCaplag * XNaliq/XNaplag * XCaliq = 0.80–1.85; Kinzler and Grove, 1992). Plagioclase-liquid geothermometer (Thy et al., 2013) yields values close to 1200 °C. Ti-magnetite is the dominant spinel of the tholeiitic rocks (rare Cr-spinels have been found in the most Fo-rich olivines), and has Al2O3 <2 wt%, and high ulvöspinel contents (91–58 mol%). Ilmenite has low MgO content (0.9–3.4 wt%). Calculated equilibration temperatures and oxygen fugacity using the magnetite-ilmenite equilibrium pairs (using ILMAT Excel worksheet by Lepage, 2003) give temperatures and oxygen of 1335 to 650 °C and from -8.8 to -13.6 log units, respectively, and plot close to the synthetic quartz-fayalite-magnetite (QFM) buffer. The LOI values of the mafic rocks vary in the range 2-8%. On the basis of their CIPW norms, they are quartz and hyperstene tholeiites. The original chemical composition has been strongly affected by secondary processes, as the tholeiitic character typical of MORB is not more recognizable, and only the samples with LOI < 3% are “typical” MORB with low K2O (< 0.6wt.%) and low TiO2 (< 2wt.%), with Mg# value variable in the range 50-65. The rocks with LOI values >4% are affected by chemical and mineralogical alteration processes. REE distribution patterns (Fig. 9.13) are flat with LREE slightly enriched compared with HREE. In the PM-normalized spiderdiagram (Fig. 9.19), the samples show a strong enrichment in Rb, Ba, K and, to a lesser extent Sr, due to chemical weathering. Conversely, REE and HFSE show a general pattern typical of MORB, with slight depletion in LREE, and flat HREE. The similarity between these rocks and MORB is confirmed by trace element contents and ratios. Indeed, in the triangular diagrams Zr-3Y-Ti/100 (Fig 10.14; Pearce and Cann, 1973) and Zr/4-Y-2Nb (Meschede, 1986) the mafic rocks plot in the MORB field. One pillow lava sample has initial (to 160 Ma) 143Nd/144Nd = 0.512775 and 87Sr/86Sr = 0.707329, quite different from the typical MORB values, due to chemical weathering. The basaltic pillow lavas of Mt. Pollino crystallized from evolved magmas as indicated by low MgO, Ni and Cr concentrations. The mineralogical assemblage and geochemical variations suggest that fractional crystallization of olivine, plagioclase and augite from more primitive magma compositions was responsible for the evolution of the basaltic rocks. Pillow lava basalts are characterized by high Zr/Nb (25–56), Zr/Hf (24–34) and low Ba/Nb (3.4–10.5) ratios. These values are well within the range of normal-MORB worldwide (Sun and McDonough, 1989; Gale et al., 2013). The depleted geochemical characteristics of these basalts are evident from contents of high field strength elements (Zr = 29-35 ppm and Nb = 0.6-1.2 ppm), typical of normal-MORB. The chemical composition of Mt. Pollino pillow lavas is consistent with magmas generated by varying degrees of non-modal fractional partial melting (totaling not more than 15%) of a four phases MORB-type asthenospheric spinel-facies mantle (Fig. 10.18). Gabbros have a typical ophitic texture with large phenocrysts of olivine and plagioclase. In all samples plagioclase is rarely preserved. The observed crystallization order is olivine, spinel, plagioclase, pyroxene, and it is typical of MORB (olivine tholeiite basalts). Olivine shows a restricted range of composition (Fig. 7.16) from Fo81 to Fo66. MnO ranges from 0.3 to 0.6 wt.% and tends to increase at decreasing MgO. NiO is variable and ranges from 0.11 to 0.25 wt.%. CaO and Al2O3 are negligible. Clinopyroxene (Fig. 7.17) is both augite (Mg# = 76-87) and diopside (Mg# = 84-93) in composition. Al2O3 varies from 2.2 to 9.6 wt.% in diopsides and from 2.2 to 6.4 wt.% in augites and is positively correlated with Mg#, whereas TiO2 shows a negative correlation with Mg#(Fig. 7.19). Cr2O3 contents range from 1.6 to 0.5 wt.% in diopsides and from 0.9 to 0.3 in augites. The relatively low Al2O3 contents of the augites indicate a low-pressure environment of crystallization. Clinopyroxenes show slight LREE depletion (LaN/SmN=0.16-0.22). They have convex upward chondrite-normalized REE patterns (Fig. 7.34) with a slight negative slope from the MREE to the HREE. Small negative Eu anomalies are noted in some clinopyroxenes (Eu/Eu*=0.6-0.9). The chondrite-normalized incompatible trace element patterns of the clinopyroxenes show that Ba, Nb, Ta, Pb, Zr and Hf are depleted relative to the REE. Th and U are variably enriched relative to Ba, Nb, Ta and Pb. Plagioclase in the gabbros (Fig. 7.23) ranges from bytownite to labradorite (An = 88-50 mol.%) Compositional zoning is generally normal. FeOt content ranges from 0.1 to 0.2 wt.%. Plagioclase is enriched in the LREE (LaN/LuN=62-67) and shows marked positive Eu anomalies (Eu/Eu*=18.3-34.4). HREE concentrations are low (below those of chondrite). Sr concentrations range from 495 to 610 ppm. Chromium-rich spinel has high Cr2O3 (42-45 wt.%) and FeO (32-36 wt.%) contents. The TiO2 contents range from 1.4 to 1.7 wt.%. MnO rarely exceeds 0.5 wt.%. Cr# (where Cr# = Cr/(Cr + Al)*100) ranges from 68 to 73. Ti-magnetites have low Al2O3 (< 0.5 wt.%) and variable ulvospinel contents (37-79 mol.%). Gabbros have MgO = 4.8-6.5 wt.%, TiO2 = 0.36-0.54 wt.%, Al2O3 < 17.5-19 wt.%, CaO = 12.9-14.5 wt.%, and high Ni (about 900 ppm) and Cr (about 1000 ppm). They have generally low concentrations of incompatible trace elements (Rb = 3-5 ppm; Zr = 15-25 ppm; Y = 10-13 ppm. Their high LOI values (3-4 wt.%) reflect alteration of the primary minerals (plagioclase altered to albite) and could have affected concentrations of mobile elements such as K, Rb, Ba and Sr. Gabbros have low REE contents (around 2-10 times chondritic values) and the REE patterns show moderate HREE enrichment (LaN/YbN ~ 0.8) and positive europium anomalies (Fig. 9.12). The high Al2O3, Sr and Eu concentrations reflect the high (>60 vol.%) modal plagioclase content. Calculated distribution coefficients for olivines (Fe/MgKdOl-liq = 0.30–0.33; Roeder and Emslie, 1970; Fig. 10.10) and clinopyroxes (Fe/MgKdcpx-liq = 0.24-0.30; Grove and Bryan, 1983; Fig. 10.11) indicate that some mafic minerals are not in equilibrium with the host rock. The trace element concentrations of clinopyroxenes of gabbros have been used to calculate the trace element composition of the parental liquid. The equilibrium liquids calculated from clinopyroxene compositions are moderately HREE enriched and LREE depleted (Fig. 10.19; LaN/YbN = 8.8-0.9). Their Nb, Ti and Sr contents are low and produce negative anomalies in the incompatible trace element patterns, and can be the result of removal of Fe-Ti oxides and plagioclase. The trace element concentrations of the parental liquids closely match those of magmas produced by 5% partial melting of DMM spinel-facies mantle. Sedimentary rocks In the log(SiO2/Al2O3) vs. log(Fe2O3/K2O) diagram (Fig. 6.6), the meta-pelites from the Timpa delle Murge Formation and two more from the Crete Nere Formation are classified as shales, whereas the other two meta-pelites from the Crete Nere Formation are in the wacke field. All meta-pelites are characterized by a very fine grain size, and in some a very fine lamination is present. They all consist mainly of a red-green clay and mica matrix with rare and scattered grains of quartz and K-feldspar. The Primordial Mantle (PM)-normalized trace element distributions of shales and wackes are very similar to the GLOSS (Global Subduction Sediment; Plank and Langmuir, 1998) values, with the exception of Rb and K, which are more enriched, whereas Sr and Ba are significantly depleted (Fig. 9.22). The Timpa delle Murge limestone shows a PM-normalized trace element distribution distinct from those of meta-pelites, with LILE generally more abundant than HFSE (Fig. 9.22). Relevant are the positive spikes of Ba and Pb, and negative of Nb, Zr, Hf and Ti. The initial isotopic composition of Timpa delle Murge shales, recalculated to the age of overlying radiolarian cherts (160 Ma), is characterized by low 143Nd/144Nd (0.512044-0.512006) and high 87Sr/86Sr (0.740696-0.740005) ratios. The initial isotopic composition of Crete Nere meta-pelites, recalculated to an age averaging the main deposition interval of the formation (50 Ma), is characterized by low 143Nd/144Nd (0.511983-0.512010) and high and very variable 87Sr/86Sr (0.715754-0.753295) ratios. PART 2: Subduction-related enrichment of the Neapolitan volcanoes mantle source: new constraints on the characteristics of the slab-derived components. Concerning the Neapolitan volcanic area, previous investigations on the nature of both pre-enrichment mantle and subduction-derived components (D’Antonio et al., 1999a, 2007, 2013; Paone, 2004, 2006; Piochi et al., 2004; Tonarini et al., 2004) have not taken into account all possible geological, petrological, geochemical and isotopic constraints. In order to shed light on the nature of the pre-enrichment mantle, the most mafic rocks of the Neapolitan volcanic area have been selected. The Quaternary magmatic rocks of Procida include high-Mg, K-basaltic lithic lava fragments, dispersed in hydromagmatic tuff of the Solchiaro eruption, that have geochemical features typical of near-primary magmas (D’Antonio and Di Girolamo, 1994; D’Antonio et al., 1996, 1999a, 2007; De Astis et al., 2004). Thus, they provide one of the few opportunities to investigate the mantle source of magmas on the basis of their trace element concentrations and isotopic composition. In particular, having the lowest 87Sr/86Sr, and the highest 143Nd/144Nd and δ11B values among the Italian Plio-Quaternary volcanics, these rocks represent magmas unaffected by open-system evolution processes, such as continental crust assimilation and concomitant fractional crystallization (AFC), very common at the Neapolitan volcanoes (e.g., Tonarini et al., 2004, 2009; D’Antonio et al., 2007; Di Renzo et al., 2007, 2011; Arienzo et al., 2009; Tomlinson et al., 2012), that could mask the source signature. Furthermore, no rocks with such primitive composition are known for the SV, where only some olivine-hosted melt inclusions (MIs) approach the composition of a primary magma (Schiano et al., 2004). Nature of the pre-enrichment mantle wedge The nature of the pre-enrichment mantle source in Central-Southern Italy, including the Southern Tyrrhenian basin, is strongly debated. On the basis of isotopic and geochemical features of mafic volcanic rocks, some authors have proposed an OIB-like mantle (e.g., Ellam et al., 1989; Peccerillo and Panza, 1999; Peccerillo et al., 2008), others a MORB-like mantle (e.g., Serri, 1990; D’Antonio et al., 1999a; Conticelli et al., 2004; Schiano et al., 2004; Paone, 2006; Francalanci et al., 2007), or a MORB-OIB mixture (Beccaluva et al., 1991; Gasperini et al., 2002; De Astis et al., 2006). In all the cases, it is postulated that the original mantle has been modified by subduction-derived components. In our attempt to identify the nature of the pre-enrichment mantle source of the Neapolitan volcanic area, the composition of the mafic rocks of Procida will be compared to the average composition of MORB, both N- and E-type (Gale et al., 2013) and Canary Islands OIB (Lustrino and Wilson, 2007). The latter authors claim that the Canary Islands OIB can most likely represent the mantle of the Western Mediterranean area before the Alpine-Apennine orogenesis. The K-basalts of Procida show REE patterns similar to those of Canary Islands OIB, although with lower LREE and MREE contents; furthermore, their HREE trends are comparable to those of Canary Islands OIB, although flatter (Fig. 11.8b). Compared to MORB, the Procida mafic rocks are quite similar to the average E-type in terms of LREE and MREE, whereas HREE contents are significantly lower. Conversely, these REE patterns are distinct from that of the average N-MORB, that they cross with a quite regular, steep slope (Fig. 11.8b). In the PM-normalized spiderdiagrams (Fig. 11.8a), LILE, Th and U of Procida mafic rocks are in the range of typical Canary Islands OIB, although more enriched with respect to those of typical N- and E-MORB. Among HFSE, Nb, Ta, Zr, Hf and Ti are depleted with respect to those of Canary Islands OIB, and are comparable to those of a typical E-MORB except Ti, which is lower. Notably, Pb is much more enriched in Procida K-basalts than any MORB and Canary Islands OIB. Again, their LREE and MREE are less abundant than those of Canary Islands OIB, and more similar to those of E-MORB. Overall, the trace elements distribution pattern of the Procida mafic rocks shows the typical features of subduction-related magmas. In order to infer the composition of the PVD mantle wedge before the enrichment, let us consider those HFSE that, according to experimental data, are immobile in hydrous fluids and melts, and therefore, should best represent the original source. According to Johnson and Plank (1999), Ti, Gd, Nb, Ta, Y and Yb satisfy this requirement. When considering relevant ratios of the latter HFSE, the Procida K-basalts have the following values (Fig. 11.10): Nb/Ta is ~15, the same as that of MORB (Hofmann, 1988; Sun and McDonough, 1989; Salters and Strake, 2004; Gale et al., 2013); Nb/Yb is ~6.5, the same as that of E-MORB, and distinct from that of N-MORB (7 vs. 1.5; Gale et al., 2013); Nb/Y is 0.5 – 0.6, close to that of E-MORB (~0.64; Gale et al., 2013); Y/Yb is ~11.5, very similar to that of MORB (~11; Hofmann, 1988; Sun and McDonough, 1989; Salters and Strake, 2004; Gale et al., 2013); Zr/Hf is ~43, very similar to that of E-MORB (~43.4; Gale et al., 2013). Only the Ti/Gd ratio of Procida mafic rocks is slightly lower than that of MORB (~1600 vs. 1800 – 2000; Salters and Strake, 2004; Gale et al., 2013). On the other hand, the Nb/Y and Nb/Yb of the Procida mafic rocks are very far from those of Canary Island OIB. Thus, the HFSE ratios would seem to suggest an E-MORB-like pre-enrichment mantle source (EMM) for the Procida mafic magmas (see also D’Antonio et al., 1999a) and, by inference, for the PVD and the entire Neapolitan volcanic area. Conversely, the Oligo-Miocene calc-alkaline mafic rocks of Sardinia (Franciosi et al., 2003) are very close to the average composition of N-MORB. The rocks of the Somma-Vesuvius complex plot middle-way between the Procida K-basalts and the Canary Islands OIB. As a matter of fact, no SV mafic volcanic rocks is really representative of a primitive magma, because they are strongly porphyritic (15 – 65 vol%; Piochi et al., 2006; Di Renzo et al., 2007). Moreover, at SV open-system magma evolution processes are very common, as suggested by trace elements and isotope relationships (e.g., Di Renzo et al., 2007, and references therein), and this casts serious doubts on any interpretation of the meaning of trace element ratios. However, a few MgO-rich MIs hosted in SV olivines have Nb/Yb ~9 (Schiano et al., 2004), thus in the range of PVD mafic rocks, suggesting substantially similar sources. Nature of the subduction-derived components In order to obtain information on the subduction-derived components that enriched the PVD mantle wedge, trace elements not significantly fractionated during melting of a typical MORB mantle (Workman and Hart, 2005) have been considered, because their ratios in the melt are similar to those in the source material. Evaluating the possible contributions from subducted sediments first, they can be both a partial melt and a hydrous fluid from a sedimentary rock similar to those of Mt. Pollino. According to Class et al. (2000), Th and Nd are excellent indicators of the type of sedimentary component because they are not mobilized by hydrous fluid, whereas Th is more incompatible than Nd during melting of pelagic sediments (Johnson and Plank, 1999). In the 143Nd/144Nd vs. Th/Nd diagram (Fig. 11.11a), the mafic rocks of Procida show a Th/Nd ratio which could be explained with a binary mixing between EMM and a sediment-derived component, either melt or hydrous fluid. However, the lower 143Nd/144Nd values of Procida K-basalts relative to EMM are better explained through metasoma¬tism by sediment melts, because hydrous fluids are poor in Nd and cannot significantly change the isotopic signature of the source. A key control factor for the geochemical composition of melts from sediments is the presence of one or more residual mineral phases. The composition of the Procida mafic rocks is compatible with the presence of residual garnet during melting of subducted sediments. This has been proposed by Avanzinelli et al. (2008) and Conticelli et al. (2009) based on the high Th concentration, Th/U ratios and LREE/HREE ratios, features typical of circum-Tyrrhenian potassic and ultrapotassic rocks, including those of the Neapolitan volcanic area. As highlighted by experimental studies (e.g., Kerrick and Connolly, 2001), garnet stabilizes during subduction of alumina- (i.e., pelitic) and carbonate-rich oceanic sediments. For low degrees of partial melting of such sediments, garnet retains HREE in the slab, producing a melt relatively enriched in LREE: this might explain well the overall REE pattern of Procida K-basalts. According to Johnson and Plank (1999), Th, Yb and La are immobile in hydrous fluids; however, during melting of a sediment in equilibrium with residual garnet, Th is mobilized (Dsediment-melt = 0.7) while La and Yb are retained. In this process, Yb (Dsediment-melt = 4.3) is retained much more than La (Dsediment-melt = 1.8), and this can explain the low Yb/Th (~0.6) and high La/Th (~8) ratios of Procida mafic rocks compared to the most primitive calc-alkaline Sardinia basalts (Yb/Th = 1.2 – 2.7; La/Th = 4.4 – 7.8) considered as the partial melting product of a mantle source enriched by hydrous fluids only (Franciosi et al., 2003). As a matter of fact, a small contribution from aqueous fluids cannot be ruled out for the Procida K-basalts, taking into account their δ11B data and overall compositional similarity with some Aeolian Arc rocks, whose isotopic compositions is attributed to both sediment melts and aqueous fluids in the mantle source (Tonarini et al., 2001, 2004; D’Antonio et al., 2007; Francalanci et al., 2007, and references therein). However, it must be pointed out that the enrichment by melts from pelagic sediments cannot explain all trace element contents, in particular those of Sr and Ba, both enriched in the Procida mafic rocks relative to E-MORB. Moreover, their Ba/Th ratios are much higher than those obtained for both sediment melt and hydrous fluids (Fig. 11.11c). This suggests a contribution from a third enriching component to the source. This can be provided by limestone, which has high concentrations of Ba and Sr, being poor in other elements such as Th, U and REE (Plank and Langmuir, 1998). As pointed out by Avanzinelli et al. (2009), moving from Northwest to Southeast along the Italian peninsula, the negative Sr anomaly in the spiderdiagrams of mafic rocks progressively disappears, and this suggests an increase in the subducted carbonate component. Experimental studies suggest that partial melting of marls can produce highly reactive, potassic/ultrapotassic melts that can act as a key enriching agent of the mantle wedge (Thomsen and Schmidt, 2008; Grassi et al., 2012). The altered basalt (and gabbro) of the oceanic crust should also be considered as a further slab-derived component. During subduction, the altered MORB experiences dehydration and burial metamorphism, releasing hydrous fluids and possibly melts. At large depth, the rock becomes an eclogite, with stabilization of garnet (e.g., Kessel et al., 2005; Klimm et al., 2008). The analyzed Mt. Pollino pillow lava sample has been used as representative of altered oceanic crust. A small contribution from fluids derived from such a component could explain the Th/Nd and Ba/Th ratios of Procida mafic rocks. Conversely, their low Yb/Th (~0.65) cannot result from binary mixing between EMM (Yb/Th ~24) and fluids from altered MORB (Yb/Th ~4). For this reason, if there has been any contribution related to the oceanic crust, this had little influence compared to the contribution deriving from the sediments. Mantle enrichment, timing of mantle enrichment, and genesis of magmas in the Neapolitan volcanic area In order to model how, and in which amount, the mantle source was enriched by subduction slab-derived components to generate the magmas that feed the Neapolitan volcanoes, the incompatible trace element contents and Sr-Nd isotopic compositions must be considered all together. All constraints presented in the previous discussion have been taken into account. The pre-enrichment mantle has been hypothesized to be EMM-type, and such a source has been assumed as starting material. The composition of one shale sample, and that of the limestone sample from the Timpa delle Murge Formation have been considered as representative of the sedimentary material that covered the Tethys oceanic crust; the latter can be represented by the pillow lava sample. During the subduction event, both melt and fluids derived from all these variable materials might have caused the enrichment of the mantle wedge, while leaving garnet as a residual phase within the slab. The geochemical modeling includes the following steps: i) addition of variable amounts of fluid from shale (FS), melt from shale (MS), melt from limestone (ML), and fluid from altered MORB (FAM) to the EMM source; ii) non-modal fractional partial melting of the enriched source; iii) comparison between the composition of calculated partial melts with the K-basalts of Procida, representative of the PVD, and olivine (Fo90)-hosted mafic MIs, representative of SV (Schiano et al., 2004). The calculated mantle sources, and partial melts that match better than any others the PVD and SV mafic rocks are illustrated through PM-normalized spiderdiagrams named Model 1, Model 2 and Model 3 (Fig. 11.12). Model 1 is the result of 1% melt from shale + 1% fluid from shale added to 98% of the EMM source. Model 2 is the result of 1% fluid from altered MORB + 2% fluid from shale added to 97% of the same EMM source. Model 3 is the result of 0.5% melt from limestone, 1.5% melt from shale, and 0.3% fluid from shale added to 97.7% of the same EMM source. Thus, each model is based on a diverse combination of the possible slab-derived components; other combinations are possible, of course, but the modeled sources would generate too strange melts in the following steps of the modeling. As an effect of addition of slab-derived components during the enrichment event, each of the resulting three mantle sources should have changed to an amphibole-bearing spinel-peridotite (Ol58Opx27Cpx10Sp3Amph2). The latter, according to experimental petrology constraints, is the best candidate to generate mildly potassic arc magmas (e.g., Wendlandt and Eggler, 1980a; Foley, 1992; Melzer and Foley, 2000; Conceição and Green, 2004). Then, each enriched source has been put through 2.5% non-modal, fractional partial melting to generate a primary magma. The calculated amounts of minerals participating to the melt are 0% olivine, 6% orthopyroxene, 12% clinopyroxene, 37% spinel and 45% amphibole. The spiderdiagram of the melts resulting from the three calculations, and particularly that of Model 3, are very similar to the primitive mafic rocks of Procida, and hence PVD. However, these primary melts do not match at all the SV mafic MIs. The higher content of K and related elements of SV with respect to PVD rocks might require a different source mineralogy, and/or a different partial melting degree, despite the proximity between the two volcanic complexes (Peccerillo, 2005 and references therein). For this reason, the genesis of SV primary magmas has been modeled starting from the same modified mantle sources as those used for PVD, but assuming a phlogopite-bearing spinel peridotite (Ol57Opx18Cpx15Sp8Phl2), the best candidate to generate potassic arc magmas (e.g., Wendlandt and Eggler, 1980b; Foley, 1992; Melzer and Foley, 2000; Conceição and Green, 2004). This source has been molten through 1% non-modal, fractional partial melting, with the following amounts of minerals participating to the melt: 0% olivine, 12% orthopyroxene, 17% clinopyroxene, 23% spinel and 48% phlogopite. The calculated primary melts are quite similar to the SV mafic MIs, with some difference among the three models. The spiderdiagram of Model 3, that matches better the SV mafic MIs, shows a lower enrichment of Th and U. A key test for the validity of the modeled mantle enrichment processes would be represented by a good match between the Sr- and Nd-isotopic compositions of the most mafic rocks from Procida (PVD) and SV, and those of the calculated sources and primary melts. In order to model the isotopic contamination of the source, all the following constraints must be taken into account: i) the hypothetical Rb, Sr, Sm and Nd contents, and Sr-Nd isotopic composition of EMM (Workman and Hart, 2005); ii) the results of the geochemical models, and in particular the Rb, Sr, Sm and Nd content of each slab-derived enriching agent, for each of the three sources modeled; iii) the Sr-Nd isotopic composition of each enriching agent: that of sample MD07-01 has been used for the shale (both melt and fluid), that of sample MD07-04 for the fluid from altered oceanic crust, whereas Sr- and Nd-values reported in the literature for another Italian limestone (Conticelli et al., 2007) have been adopted for the Timpa delle Murge limestone. The Sr-Nd-isotopic composition of the enriched source(s) would be given by the weighted sum of the contribution of the pre-enrichment mantle and those of the assumed enriching agents, taking into account the relative Sr and Nd contents of each. Another important constraint is given by time. Two scenarios are possible. If the enrichment of the source were occurred recently, namely within say 1 Ma ago, the present-day isotopic composition of the enriching agents would directly affect that of the source and of its partial melts (scenario 1). On the other hand, if the enrichment were occurred several tens of million years in the past, then the source would have been modified by enriching agents with a Sr-Nd isotopic composition different from that measured today, that must be re-calculated back to the time of enrichment (scenario 2). Furthermore, if such source enriched in the past underwent partial melting in recent times, the time evolution of its Sr-Nd isotopic composition should be considered, due to radioactive decay of 87Rb and 147Sm, in order to infer that of the partial melts. These scenarios have been considered in order to carry out calculations of the Sr-Nd isotopic composition of the enriched sources, and thus primary magmas. The scenario 1 results in an enriched source (that of Model 1) that would have 87Sr/86Sr too high, and 143Nd/144Nd too low (0.70545 and 0.51265, respectively, relative to the measured values of 0.7051 and 0.5127 of the Procida K-basalts. The enriched source of Model 3 would have 87Sr/86Sr slightly lower (0.70526) and the same 143Nd/144Nd, whereas the enriched source of Model 2 would have 87Sr/86Sr much higher (0.70718). These results suggest that the enrichment event of the mantle source might have occurred in the past (scenario 2). Thus, variable calculations have been carried out for the sources 1, 2 and 3 (obtained from Models 1, 2 and 3), in order to obtain the enrichment time in the past such that each source would have had a Sr-Nd isotopic composition that, after evolution through time, would match that of partial melts generated in recent times, to be compared to the present-day 87Sr/86Sr and 143Nd/144Nd of Procida K-basalts. A very good match has been obtained for Source 1 and Source 3, for an enrichment event occurred 45 Ma ago. This age is compatible with the hypothesis that the subduction of Tethys ocean beneath the continental margin of the Corsica-Sardinia-Calabria block may have initiated during the Eocene (Lustrino et al., 2009; Shimabukuro et al., 2012). Since then, that enriched source (or other sources enriched in a similar manner) has generated subduction-related magmas spatially displaced in a large portion of Central-Western Mediterranean area: the Oligo-Miocene calc-alkaline magmas of Sardinia (32-12 Ma); the calc-alkaline and K-alkaline magmas of Tyrrhenian Seamounts (Cornacya, Anchise, Marsili; 12 – <0.1 Ma); the calc-alkaline and K-alkaline magmas of Pontine islands and Campania Plain (4.2 – 0.13 Ma); the K-alkaline magmas of Neapolitan volcanic area (0.4 Ma – Present). Of course, it cannot be ruled out that more the one events might have occurred through time, with enriching agents of variable geochemical and isotopic composition. It is worth noting that, despite a large number of geochemical and isotopic data suggest that the main evolutionary process of Somma-Vesuvius magmas is fractional crystallization (e.g., Civetta et al., 1991; Cioni et al., 1995; Ayuso et al., 1998; Santacroce et al., 2007), open-system processes such as magma mixing and/or contamination at mid-lower crustal depth might have played a significant role (Civetta et al., 2004; Di Renzo et al., 2007). The latter process may justify the Sr-Nd isotopic composition of SV rocks, different from that of Procida mafic rocks. In order to model crustal contamination of SV magmas, the EC-AFC (energy constrained assimilation and fractional crystallization) model of Spera and Bohrson (2001), based on the energy and mass balance between assimilant and magma, has been employed. A good candidate for the assimilant could have been granulite-facies meta-igneous rocks (granodioritic, tonalitic and gabbroic) exposed in Southern Calabria, that are representative of the intermediate-lower crust remaining after the Hercynian Orogeny and exhumed during the Alpine Orogeny (Caggianelli et al., 1991; Rottura et al., 1991). Those rocks have been supposed to compose the crustal basement underneath the Campania region, although there is no geological evidence for that, and used as an assimilant in crustal contamination modeling (Pappalardo et al., 2002; Paone, 2006; Di Renzo et al., 2007; D’Antonio et al., 2007, 2013). Alternatively, the crystalline basement may be similar to the African continental crust (e.g., Davidson and Wilson, 1989). EC-AFC modeling using these two types of continental crust have been carried out. The Sr-Nd isotopic composition of the present-day modeled enriched sources of Model 1 and Model 3 have been assumed as that of the starting magma. The temperature of magma and the enthalpy were calculated using MELTS (http://melts.ofm-research.org), starting from the chemical composition of MIs reported in Schiano et al. (2004), while the temperature of the assimilant (670 °C) was assumed according to Wyllie (1977). The bulk partition coefficients for Sr and Nd in the primary SV magma were calculated for the transition shoshonite-latite, according to the mineral phenocrysts and their abundance in the MELTS mass balance calculation. EC-AFC modeling shows that the difference in Sr-Nd-isotopic composition between the modeled source and the most mafic rocks of SV can be explained by the assimilation of about 1.4% of Calabrian continental crust or, alternatively, 1.7% of African continental crust. The latter modeling, however, does not match well the SV Nd-isotopes (Fig. 11.13). PART 3: Conclusions The Neapolitan volcanic area (Southern Italy), that includes the Phlegrean Volcanic District and the Somma-Vesuvius complex, has been the site of an intense Plio-Quaternary magmatic activity, producing volcanic rocks showing a subduction-related geochemical and isotopic signature. The high-Mg, K-basaltic lithic lava fragments dispersed in hydromagmatic tuff of the Solchiaro eruption (22 ka, Procida Island) are the least evolved volcanic rocks of the Neapolitan volcanic area and provide constraints on the nature and role of both pre-enrichment mantle and subduction-related components. Their average values of HFSE ratios such as Nb/Yb, Nb/Y, and Zr/Hf, relative to typical EMM and Canary Islands OIB, would seem to indicate an EMM-like, spinel peridotite pre-enrichment source for the Procida and, perhaps Somma-Vesuvius, mafic magmas. In order to constrain the characteristics of subduction-slab derived components added to this mantle sector, new geochemical and Sr-Nd-isotopic data have been acquired on meta-sediments and pillow lavas from Timpa delle Murge ophiolites, representing fragments of Tethyan oceanic crust obducted during the Apennine orogenesis, and that may be similar to sediments subducted during the closure of Tethys ocean. Based on trace elements compositions (e.g., Th/Nd, Yb/Th, Ba/Th) the addition of three distinct subduction components to the mantle wedge can be hypothesized: partial melts from shales, aqueous fluids from shales, and partial melts from limestone. Moreover, trace elements and Sr-Nd-isotopic ratios suggest a greater role for melts from pelitic sediments, relative to melts from limestones and aqueous fluids, whereas seem to rule out a significant role for altered oceanic crust. Some trace element characteristics, such as the relatively low HREE contents point to residual garnet during partial melting in the slab. Modeling based on variations of trace elements and isotopic ratios indicates that the pre-subduction mantle source of the Phlegrean Volcanic District and Somma-Vesuvius was enriched by 2-3% of subduction slab-derived components, in agreement with previous modeling based on Sr-O isotopes for Ischia island (D’Antonio et al., 2013). The enrichment event might have stabilized amphibole and/or phlogopite in the mantle source. The partial melting degree of this enriched source (amphibole-bearing spinel peridotite) should have been 2.5% to generate the Procida primary magmas. Modeling based on trace element contents, and 87Sr/86Sr and 143Nd/144Nd values constrains the age of the enrichment event to about 45 Ma ago, confirming that the Plio-Quaternary magmatism of the Neapolitan area is post-collisional. This suggests that the origin of the subduction-derived enriching agents can be related to the closure of Tethys, which remnants are represented by the Calabria-Lucanian ophiolites, such as that of Timpa delle Murge Formation. A model of partial melting of a source with the same chemical composition as that inferred for the PVD, but with different mineralogy (phlogopite-bearing spinel peridotite) reproduces the same trace elements distribution pattern of the olivine-hosted shoshonitic melt inclusions from the 1906 AD lava flow of Somma-Vesuvius under 1% partial melting degree. Primitive melts from this enriched mantle were subsequently modified by both FC and AFC processes at mid-lower crust depth, with assimilation of 1.4% of continental crust of composition likely similar to the Hercynian meta-igneous rocks of Calabria.

Actions (login required)

View Item View Item