Journal of Petrology Advance Access originally published online on February 18, 2007
Journal of Petrology 2007 48(4):753-784; doi:10.1093/petrology/egl081
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Magmatic History of Somma–Vesuvius on the Basis of New Geochemical and Isotopic Data from a Deep Borehole (Camaldoli della Torre)
1Istituto Nazionale di Geofisica E Vulcanologia, Osservatorio Vesuviano, Via Diocleziano 328, Napoli, Italy
2Dipartimento Scienze Fisiche, University Federico II, Via Cinthia, Napoli, Italy
3Dipartimento Scienze Della Terra, University Federico II, L.Go S. Marcellino 10, Napoli, Italy
4Istituto di Geoscienze E Georisorse, Cnr, Via G. Maruzzi 1, Pisa, Italy
RECEIVED MARCH 13, 2006; ACCEPTED DECEMBER 22, 2006
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONA continuous-coring borehole recently drilled at Camaldoli della Torre on the southern slopes of Somma–Vesuvius provides constraints on the volcanic and magmatic history of the Vesuvian volcanic area since c. 126 ka BP. The cored sequence includes volcanic units, defined on stratigraphical, sedimentological, petrological and geochemical grounds, emitted from both local and distal vents. Some of these units are of known age, such as one Phlegraean pre-Campanian Ignimbrite, Campanian Ignimbrite (39 ka), Neapolitan Yellow Tuff (14· 9 ka) and Vesuvian Plinian deposits, which helps to constrain the relative age of the other units. The main rock types encountered are shoshonite, phonotephrite, latite, trachyte and phonolite. The sequence includes, from the base upwards: a thick succession of pyroclastic units emplaced between 126 and 39 ka, most of them attributed to eruptions that occurred in the Phlegraean area; the Campanian Ignimbrite; the products of a local tuff cone formed between 39 ka and the deposition of the products of the earliest activity of the Mt. Somma volcano; the products of the Somma–Vesuvius volcano, which include from the base upwards a thick sequence of lavas, pyroclastic rocks and the products of a local spatter cone dated between 3· 7 ka and AD 79. The data obtained from the study of the borehole show that, before the Campanian Ignimbrite eruption, low-energy explosive volcanism took place in the Vesuvian area, whereas mostly high-energy explosive eruptions characterized the Campi Flegrei activity. In the Vesuvian area, Campanian Ignimbrite deposition was followed by the eruption of a local tuff cone and a long repose time, which predated the formation of the Mt. Somma edifice. Since 18· 3 ka (Pomici di Base eruption) the activity of Somma–Vesuvius became mostly explosive with rare lava effusions. The shallowest cored deposits belong to the Camaldoli della Torre cone, formed between the Pomici di Avellino and Pomici di Pompei eruptions (3· 7 ka–AD 79). New geochemical and Sr–Nd–Pb–B-isotopic data on samples from the drilled core, together with those available from the literature, allow us to further distinguish the volcanic rocks as a function of both their provenance (i.e. Phlegraean or Vesuvian areas) and age, and to identify different magmatic processes acting through time in the Vesuvian mantle source(s) and during magma ascent towards the surface. Isotopically distinct magmas, rising from a mantle source variably contaminated by slab-derived components, stagnated at mid-crustal depths (8–10 km below sea level) where magmas differentiated and were probably contaminated. Contamination occurred either with Hercynian continental crust, mostly during the oldest stages of Vesuvian activity (from 39 to 16 ka), or with Mesozoic limestone, mostly during recent Vesuvian activity. Energy constrained assimilation and fractional crystallization (EC-AFC) modelling results show that contamination with Hercynian crust probably occurred during differentiation from shoshonite to latite. Contamination with limestone, which is not well constrained with the available data, might have occurred only during the transition from shoshonite to tephrite. From the ‘deep’ reservoir, magmas rose towards a series of shallow reservoirs, in which they differentiated further, mixed, and fed volcanic activity.
KEY WORDS: Somma–Vesuvius; crustal contamination; source heterogeneity; radiogenic and stable isotopes; magmatic system
|
INTRODUCTION |
|---|
Somma–Vesuvius, together with Campi Flegrei, Procida and Ischia, is one of the four volcanoes within the Neapolitan area (Fig. 1). Since the last eruption in AD 1944 it has been quiescent; only moderate fumarole emissions and seismicity testify to its continuing activity. The past behaviour of the volcano suggests that the present quiescence could culminate in an eruption, potentially affecting at least 600 000 people. Therefore knowledge of its magmatic structure and history is fundamental in interpreting possible variations of the volcano's dynamics, preceding an eruption. Petrological studies of Vesuvian rocks have been mostly aimed at defining the magmatic history of the volcano in the last 18· 3 kyr, since the earliest known Plinian eruption (Pomici di Base; 18· 3 ka, Bertagnini et al., 1998
). Little attention has been devoted to the activity that occurred between the eruption of the Campanian Ignimbrite (CI; 39 ka, De Vivo et al., 2001) from Campi Flegrei (Fisher et al., 1993; Orsi et al., 1996; Ort et al., 2003) and the Pomici di Base eruption, which mostly built the Mt. Somma stratovolcano. The activity older than the CI eruption, which predates the formation of the Somma–Vesuvius volcano, is also poorly studied. In this area, rocks older than 39 ka are not exposed and can, therefore, only be sampled by drilling.
|
Whole-rock geochemical and isotopic data (Sr, Nd, Pb, B) on samples from pyroclastic deposits and lava flows cored in a 240 m deep borehole at Camaldoli della Torre, along the southern slopes of Somma–Vesuvius, provide a unique opportunity to reconstruct the history of the volcanic system before the deposition of the CI and during the first phase of the building of the Somma–Vesuvius stratovolcano (39–19 ka).
|
GEOLOGICAL, GEOPHYSICAL AND PETROLOGICAL BACKGROUND |
|---|
Volcanism in the Vesuvian area commenced about 400 ka BP, according to 40Ar–39Ar ages on lavas cored at 1125 m below sea level (b.s.l.) in the Trecase area (Fig. 1; Brocchini et al., 2001
), although the present Somma–Vesuvius volcano formed largely in the last 39 kyr. Somma–Vesuvius [1180 m above sea level (a.s.l.)] includes the older Mt. Somma stratovolcano and the recent Vesuvius cone, which has grown within the Mt. Somma caldera. The oldest period of activity (39–19 ka) was dominated by lava flow extrusion and low-energy explosive eruptions, which built up the Somma stratovolcano (Santacroce, 1987; Santacroce et al., 2005). Since 19 ka, Somma–Vesuvius activity has been characterized by a series of Plinian and sub-Plinian eruptions preceded by quiescent periods lasting from centuries to millennia, followed, at least in the past 4 kyr, by periods of open-conduit, semi-persistent activity characterized by lava effusions and low-energy explosive eruptions (Santacroce, 1987). The most recent period of semi-persistent volcanism began after the 1631 sub-Plinian eruption and lasted until the last eruption in 1944. Since then the volcano has been quiescent, showing no signs of unrest.
The volcano is located at the intersection of NW–SE- and NE–SW-trending fault systems (Bianco et al., 1998, and references therein). Results of geophysical and geological investigations (Bruno et al., 1998; Zollo et al., 1998, 2002; Brocchini et al., 2001; Scarpa et al., 2003, and references therein) have shown that the shallow structure of Somma–Vesuvius includes 1· 5–2 km of interbedded lavas, volcanoclastic, marine, and fluvial sedimentary rocks of Pleistocene age, overlying Mesozoic limestone basement at 2· 5–3 km depth b.s.l. Seismic tomography studies and the pattern of local volcano-tectonic seismicity (Lomax et al., 2001; Scarpa et al., 2003) exclude the presence of magmatic reservoirs with a diameter in excess of 0· 5–1 km within the upper 0–5 km of the crust. However, seismic tomography studies (Zollo et al., 1996; Auger et al., 2001) have indicated the presence of a widely distributed (at least 400 km2) low-velocity layer, with a flat top at about 8 km b.s.l. beneath the volcano, interpreted as the top of the present-day magma reservoir. The inferred S-wave (c. 0· 6–1· 0 km/s) and P-wave (c. 2· 0 km/s) velocities indicate the presence of a partially crystallized, sill-like magma body. This depth coincides with the top of a thick zone (10–20 km depth) of clinopyroxene crystallization, inferred from studies of the trapping pressure of fluid inclusions (e.g. Belkin et al., 1985; Marianelli et al., 1995). The Moho discontinuity in the Vesuvian area is at about 30 km (Capuano et al., 2003), significantly shallower than the Moho below the Apennines, as a consequence of back-arc extension in the Tyrrhenian Sea (Patacca & Scandone, 1989).
A general consensus exists that the mantle source(s) of magmas beneath Somma–Vesuvius, as well as beneath the Campi Flegrei caldera, and Procida and Ischia islands, is (are) enriched in incompatible elements, radiogenic Sr and Pb, and unradiogenic Nd relative to mid-ocean ridge basalt (MORB)-source mantle. Most researchers agree that the enrichment process is related to mantle contamination by subducting slab-derived materials (e.g. Hawkesworth & Vollmer, 1979; Rogers et al., 1985; Serri, 1990; Beccaluva et al., 1991; Ayuso et al., 1998; D'Antonio et al., 1999a, 2007; Peccerillo, 1999; Civetta et al., 2004a, 2004b; Tonarini et al., 2004; Piochi et al., 2006). Only a few workers have proposed the involvement of fluids of deep origin in an intraplate tectonic setting (e.g. Cundari, 1980; Vollmer, 1989; Bell et al., 2004). The first hypothesis is strongly supported by geological and geophysical data, which suggest a NW-subducting slab beneath the Tyrrhenian Sea (Selvaggi & Chiarabba, 1995). Geochemical and isotopic variations in Vesuvian magmas have been further attributed, at least in part, to crustal contamination of mantle-derived magmas (e.g. Civetta et al., 2004a, 2004b; Piochi et al., 2006).
In the last 19 kyr, Somma–Vesuvius has erupted about 50 km3 of magma of variable composition (e.g. Joron et al., 1987), ranging from slightly silica-undersaturated (K-basalt to K-trachyte) to highly silica-undersaturated (K-tephrite to K-phonolite). Three magmatic periods have been recognized (Santacroce et al., 2005): the first (19–10 ka) is characterized by emission of slightly undersaturated lavas (K-basalt to K-latite) and pyroclastic deposits (K-latite to K-trachyte), the second (10 ka–AD 79) by phonotephrite to phonolite magmas, and the third (AD 79–1944) by magma compositions from leucititic phonotephrite to leucititic phonolite. In the last 19 kyr the activity of Somma–Vesuvius has alternated between open- and closed-conduit conditions (Santacroce, 1987; Santacroce et al., 2005). During the open-conduit regime, the volcano was characterized by semi-persistent low-energy activity. The closed-conduit regime has probably favoured the formation of shallow magma chambers, the activity of which culminated in Plinian or sub-Plinian eruptions (Santacroce et al., 1993).
Geochemical and isotopic data (Civetta et al., 2004b) show that in the ‘deep’ magma reservoir, whose top is located at a depth of about 8 km b.s.l., mantle-derived melts stagnated, differentiated and were probably contaminated by continental crust, as a likely consequence of the high temperatures reached by crustal rocks because of repeated intrusion of magma into the storage reservoir (de Lorenzo et al., 2006). From the ‘deep’ reservoir, magma rises to form shallow magma chambers at variable depths, in which it experiences low-P differentiation and mixing, before feeding volcanic surface activity (e.g. Civetta et al., 1991a; Marianelli et al., 1995; Cioni et al., 1998).
|
THE CAMALDOLI DELLA TORRE SEQUENCE |
|---|
To extend back in time our knowledge of the structrure and volcanic and magmatic history of the Vesuvian area, we have studied in detail a 240 m stratigraphic sequence (Fig. 2), composed of lava flows and pyroclastic deposits, drilled at Camaldoli della Torre (CdT), close to the town of Torre del Greco (Fig. 1). The borehole was continuously drilled with a core diameter variable between 8 and 12 cm. It was drilled to instal a strainmeter of the Italian Istituto Nazionale di Geofisica e Vulcanologia (INGV) surveillance network. The studied sequence also includes the Campanian Ignimbrite, erupted from Campi Flegrei at 39 ka BP, which is used as a marker horizon in the Neapolitan area because of its widespread dispersal (Fisher et al., 1993
; Ort et al., 2003).
|
The rock units of the CdT sequence, described in Fig. 2 and in detail in the Electronic Appendix 1 (available for downloading from http://petrology.oxfordjournals.org), have been correlated with units known from the literature, based on their lithological, sedimentological and petrological characteristics. The sequence, from the base (240 m from the ground surface and 121· 6 m b.s.l.) upwards, includes the following.
- A succession of 10 pyroclastic units separated by palaeosols and lacustrine deposits. These units are older than the CI and younger than Tyrrhenian age deposits (126 ka; Romano et al., 1994) widely distributed in the Campanian plain (Ferranti et al., 2006) but not penetrated by the CdT borehole. Unit CdT-a is interpreted as the product of an eruption that occurred in the Vesuvian area on the basis of thickness and sedimentological features. Units CdT-b, CdT-c, CdT-d and CdT-e are interpreted as distal pyroclastic fallout, not correlated to any known unit, although they are geochemically similar (see discussion below) to the older than CI Phlegraean units. Units CdT-f, CdT-g, CdT-h, CdT-i, and CdT-j are correlated with the PRc, PRd, PRi, TLc (emplaced between 58 and 46· 5 ka) and PRm (44· 3 ka) units of Campi Flegrei (Pappalardo et al., 1999), on the basis of their litholology, sedimentology, geochemistry and Sr isotope composition.
- A palaeosol.
- The Campanian Ignimbrite (39 ka), including the basal fallout deposit and pyroclastic-current deposits, as described by Fisher et al. (1993), Civetta et al. (1997), Rosi et al. (1999) and Pappalardo et al. (2002a).
- The pyroclastic-current deposits CdT-k1 (3 m thick) and CdT-k2 (2· 6 m thick). Sedimentological and geochemical features suggest an origin in the Vesuvian area for CdT-k1 and in the Phlegraean area for CdT-k2.
- The pyroclastic unit CdT-l (67 m thick), composed of a succession of ash surge beds and subordinate coarse pumice fallout deposits. It was probably produced by a local tuff cone, at present completely buried under younger products and previously unrecorded.
- The pyroclastic-current unit CdT-m (1· 7 m thick), with sedimentological and lithological features that do not allow us to discriminate between Phlegraean and Vesuvian origin.
- A thick sequence (3 m) of reworked deposits and mature palaeosols.
- A thick succession (36 m thick) of at least 15 lava flows (CdT-n01 to CdT-n15), produced by Mt. Somma activity. The lavas are rarely separated by palaeosols and reworked deposits and do not include pyroclastic deposits.
- The Pomici di Base Tephra (18· 3 ka; Bertagnini et al., 1998), clearly recognizable on the basis of lithology and sedimentology.
- Two lava flows units, CdT-o01 and CdT-o02, emplaced between 18· 3 and 16 ka and separated by reworked deposits. The latter lava flow unit is stratigraphically and compositionally relatable to the lava and scoria formation that crops out on the NE slope of Mt. Somma (Vallone S. Severino lava and scoriae formation) (Santacroce & Sbrana, 2003).
- The Pomici Verdoline Tephra (16 ka; Cioni et al., 2003), confidently recognized on the basis of lithology and sedimentology.
- A succession of pyroclastic deposits with intercalated palaeosols, emplaced between 16 and 8 ka, including the Phlegraean Neapolitan Yellow Tuff (14· 9 ka, Deino et al., 2004) and the Pomici Principali Tephra (10· 3 ka; Di Vito et al., 1999).
- The Pomici di Mercato Tephra (8 ka; Arnò et al., 1987), clearly recognizable on the basis of lithology and sedimentology.
- A palaeosol containing fragments of the Agnano–Monte Spina Tephra (4· 1 ka; de Vita et al., 1999).
- The Pomici di Avellino Tephra (3· 7 ka; Cioni et al., 2000), confidently recognized on the basis of lithology and sedimentology.
- A succession of lavas and coarse scoriae beds of the Camaldoli della Torre cone (CdT-cone), overlain in the CdT area by the Pomici di Pompei deposits. Consequently, its age is between 3· 7 ka and AD 79. The cone-forming eruption was previously attributed to an activity occurred between 18· 3 and 16· 0 ka and fed by eccentric structures (Santacroce & Sbrana, 2003).
|
ANALYTICAL TECHNIQUES AND SAMPLE CLASSIFICATION |
|---|
We have analysed 32 lava samples, from 0 to 109 m depth (CdT-cone to CdT-n01), and 28 pyroclastic rock samples, from 110 to 240 m depth in the sequence (CdT-m to CdT-a) (Fig. 2). From the pyroclastic deposits we have analysed pumice fragments.
Major and trace element contents (Table 1) were determined by inductively coupled plasma atomic emission spectrometry (ICP-AES; major oxides) and inductively coupled plasma mass spectrometry (ICP-MS; trace elements) at the Centre de Récherches Pétrographiques et Géochimiques, Cedex, Nancy (France). Accuracy was checked against International Rock Standards (Govindaraju, 1994); precision was 1–2% (1
) for major and 3–5% (1) for trace elements. Sr- and Nd-isotope ratios of whole-rock and clinopyroxene separates were determined by thermal ionization mass spectrometry (TIMS) with a multicollector ThermoFinnigan Triton TI mass spectrometer at INGV-Osservatorio Vesuviano, Napoli. Table 2 reports all Sr and Nd isotopic measurements with their error (2m, number of measured ratios = 180). The mean measured value of 87Sr/86Sr for NIST SRM 987 was 0· 710250 ± 0· 000014 (2, number of standard analyses = 56) and that of 143Nd/144Nd for La Jolla is 0· 511850 ± 0· 000015 (2, number of standard analyses = 25). The Sr and Nd blanks were negligible for the analysed samples during the period of measurement. Pb isotope ratios were determined by TIMS with a multicollector Finnigan Mat 262 mass spectrometer at Istituto di Geoscienze e Georisorse, CNR Pisa. The Pb isotope ratios were corrected for instrumental mass fractionation with a factor of 0· 15% per mass unit using NIST SRM 981 as reference. The external reproducibility of NIST SRM 981 was better than 0· 1% during the course of the measurements. B isotope ratios were measured on a VG54E single-collector mass spectrometer at Istituto di Geoscienze e Georisorse, CNR Pisa. The results are reported as 
11B (i.e. the permil deviation of the 11B/10B ratio measured in the sample from that measured in the chemically processed NIST SRM 951 standard). Based on replicate analyses of standards and many samples, the precision and accuracy of both B elemental and isotopic composition are estimated to be within ±0· 5
(Tonarini et al., 2003).
|
|
According to the TAS classification diagram (total alkali vs SiO2; Le Maitre et al., 1989
; Fig. 3) the rocks vary from shoshonite to latite, phonotephrite, trachyte and phonolite. They have been divided, according to their age, into three groups. Group 1 includes rocks older than 39 ka (from CdT-a to CdT-j), which are mostly trachytes, with subordinate phonolites. Only one unit is of local origin (CdT-a); all the others are of Phlegraean origin, recognized on the basis of their geochemistry and Sr isotope composition as described below. Group 2 includes rocks dated between 39 and 16 ka (from CdT-k1 to CdT-o02); that is, between the two marker beds represented by the CI and Pomici Verdoline Tephra. This group includes, from the base upwards, at least three trachytic pyroclastic units (CdT-k1, CdT-k2 and CdT-l), two originating in the Vesuvian area and one (CdT-k2) in the Phlegraean area, and a sequence of lavas varying from shoshonite to phonotephrite, shoshonite and latite. Phonotephritic rocks of this group are older than 18· 3 ka and are the oldest rocks of this composition ever found at Somma–Vesuvius with an age within the last 39 kyr. Group 3 includes only phonotephritic lavas, emplaced between the Pomici di Avellino (3· 7 ka) and Pomici di Pompei (AD 79) Tephra and belonging to the CdT-cone. We have not analysed samples from the identified Plinian and sub-Plinian units, for which compositional data are available in the literature (Civetta et al., 1991a; Bertagnini et al., 1998; Cioni et al., 1999, 2003).
|
CIPW norms have been calculated by assigning Fe2O3/FeO values to the rocks according to their TAS classification, following Middlemost (1989
). The pyroclastic rocks of Groups 1 and 2 (phonotephrite, trachyte and phonolite) range from silica-undersaturated to silica-oversaturated, with normative nepheline (0–7 wt %) or quartz (0–8 wt %). All the shoshonitic and latitic lavas of Group 2 are nepheline-normative, whereas the phonotephritic lavas of Groups 2 and 3 have both normative nepheline and leucite. In summary, the oldest analysed Vesuvian lava is the least differentiated shoshonite (CdT-n01), whereas phonotephritic lavas were emplaced during two different periods: (1) before 18· 3 ka (CdT-n07, CdT-n08, CdT-n09 and CdT-n13); (2) between 3· 7 ka and AD 79 (CdT-cone). Latitic lavas (CdT-o02) are bracketed between 18· 3 ka (Pomici di Base Tephra) and 16 ka (Pomici Verdoline Tephra).
|
PETROGRAPHY OF THE SAMPLES |
|---|
Lava flows and pyroclastic deposits of the CdT sequence show variable degrees of vesicularity and phenocryst content (details are given in Electronic Appendix 1). The trachytic pumice fragments of Group 1 have between about 35 and 50 vol. % vesicles. Their texture is generally weakly porphyritic, with phenocryst contents between 1 and 3 vol. %. The phenocryst assemblage is dominated by alkali feldspar, subordinate green clinopyroxene, plagioclase, biotite and magnetite. The groundmass is mainly glassy, with scarce alkali feldspar and clinopyroxene microlites, often arranged in a pilotaxitic texture. Alteration has generated halos at the margins of the pumice clasts. The trachytic pumice fragments of Group 2, generally similar in texture and phenocryst assemblage to those of Group 1, do not contain phenocrysts of biotite and are more vesiculated (45–50 vol. %). The shoshonitic and phonotephritic lavas of Group 2 have generally porphyritic and glomeroporphyritic textures; the phenocryst content varies between 15 and 35 vol. %. Shoshonites generally include phenocrysts of clinopyroxene > leucite > olivine > plagioclase, whereas plagioclase and olivine phenocrysts are scarcer or absent in phonotephrite; biotite, magnetite and apatite are accessory phases. The groundmass is intergranular, with abundant microlites of plagioclase, leucite and clinopyroxene, and subordinate olivine, magnetite and biotite. Lavas of Group 2 at a depth between 90 and 108 m (CdT-n01 to CdT-n06) show evidence of low-T alteration, such as partial transformation of olivine phenocrysts to iddingsite (Fe-oxyhydroxides), and the occurrence of patches of carbonate in the groundmass. Lavas at depths between 90 and 49 m (from CdT-n07 to CdT-o02) are fresh. The latitic lavas of Group 2 contain about 15 vol. % of plagioclase > clinopyroxene > biotite > magnetite phenocrysts, in a felty to intergranular groundmass, characterized by microlites of the same minerals occurring as phenocrysts plus minor amounts of leucite and olivine. The phonotephrites of Group 3 include about 15–20 vol. % of clinopyroxene > leucite > biotite phenocrysts in an intergranular groundmass, with microlites of those phases plus plagioclase, biotite and magnetite. Alteration has produced partial transformation of olivine to iddingsite and sometimes patches of haematite in the groundmass.
|
MAJOR AND TRACE ELEMENT GEOCHEMISTRY |
|---|
Major and trace element variations (Table 1; Figs 4 and 5) show distinct trends in part consistent with crystal–liquid differentiation. CaO, MgO, TiO2, Fe2O3tot and P2O5 (the last two not shown in Fig. 4) contents decrease with increasing SiO2 from the shoshonites of Group 1 and the phonotephrites of Groups 2 and 3, to the latites of Group 2, and to the trachytes and phonolites of Group 1. Na2O and K2O contents generally increase with differentiation, although Na2O decreases in the trachytes of Group 2, probably as an effect of alteration. Al2O3 increases regularly up to the least differentiated trachytes of Group 1, and then decreases in the most differentiated trachytes and phonolites of Groups 1 and 2. Light rare earth element (LREE), Th, U, Hf, Ta, Zr and Pb contents generally increase with increasing differentiation. Nb, Y, middle REE (MREE; with the exception of Eu) and heavy REE (HREE) contents increase slightly with increasing differentiation until the trachytic composition is reached, when they decrease. Eu is fairly constant at increasing SiO2 up to 60 wt %, and then abruptly decreases. Rb content is more irregular, with a decrease in the phonotephrites, an increase in the latites and again a decrease in the trachytes and phonolites. Sr, Ba, Cr, Co and Ni contents decrease with increasing differentiation. A careful investigation of the diagrams of Fig. 5 (e.g. the Zr vs SiO2 plot) allows distinctions to be made within the different age groups of volcanic rocks as a function both of their provenance (i.e. Phlegraean or Vesuvian areas) and composition (i.e. shoshonite to latite, phonotephrite and trachyte). We can thus distinguish the following in the Zr vs SiO2 plot: the phonotephrites of Groups 2 and 3 (‘a’ in Fig. 5); the shoshonites and latites of Group 2 (‘b’ in Fig. 5); the trachytes of local origin of Group 1 (‘c’ in Fig. 5); the pre CI and post CI trachytes and phonolites of Phlegraean origin of Groups 1 and 2 (‘d’ in Fig. 5); the trachytes of local origin younger than 39 ka (‘e’ in Fig. 5); the CI products. A main consideration is that the phonolites and trachytes define a single trend, whereas the Vesuvian pre CI and post CI trachytes have different major and trace element composition; moreover, the Vesuvian shoshonites to latites and phonotephrites describe different geochemical trends, the phonotephrites being more enriched in incompatible elements with respect to the shoshonites and latites.
|
|
Chondrite-normalized REE patterns (Fig. 6) show that both the shoshonites and phonotephrites of Groups 2 and 3 are characterized by similar highly fractionated patterns, with a negative Eu anomaly (Eu/Eu* 0· 75–0· 96). The trachytes and phonolites of Groups 1 and 2 are characterized by highly LREE-enriched patterns, a negative Eu anomaly (Eu/Eu* 0· 6–0· 91) and flat HREE patterns.
|
Modelling major element variations
Table 3 reports the results of geochemical modelling of major element variations performed by least-square calculations (Stormer & Nicholls, 1978
) for the shoshonite–latite–trachyte range, using as a starting composition the least differentiated shoshonite (sample s108.3 from CdT-n01) and mineral chemistry data from Santacroce (1987) and the authors' unpublished data. The results suggest that it is possible to make the transition from shoshonite to latite with c. 39–45% fractionation, and from latite to trachyte of Group 2 (i.e. the tuff cone of local origin) with 88% fractionation (Table 3). In agreement with petrographic observations, the calculated mineral assemblage is clinopyroxene, plagioclase, olivine and leucite, with subordinate magnetite and apatite for the shoshonite–latite transition; sanidine and biotite substitute for leucite in the assemblage subtracted for the latite–trachyte transition. The plausibility of these models is suggested by the very low values of the sum of the squares of residuals, between 0· 28 and 0· 06. The result of single-stage modelling starting from the shoshonite s108.3 (CdT-n01) to the most evolved trachyte s143.8 (CdT-l) also confirms the plausibility of a genesis by fractional crystallization (sum of the squares of residuals < 0· 001; Table 4). Conversely, certain aspects of the trace element data cannot be explained by fractional crystallization of the observed liquidus phases. For instance, the three-fold Th enrichment in the trachyte with respect to the shoshonite is not matched by Zr, Ta, Y and LREE variations, which show only 1· 5–2-fold enrichment. As these are all incompatible elements, a simple Rayleigh fractionation model cannot reproduce the observed variation with c. 88% fractional crystallization. Furthermore, modelling of the shoshonite–phonotephrite transition gave poor results (sum of the squares of residuals, in the best case, 2· 61), requiring accumulation of a significant amount of leucite (>8%), thus excluding any possibility of deriving the phonotephritic magma from the shoshonitic magma by closed-system fractional crystallization (see Belkin et al., 1993; Wood & Trigila, 2001). Thus, open-system processes could better explain the geochemical variations observed between the different groups of rocks.
|
|
|
ISOTOPE GEOCHEMISTRY |
|---|
The Sr, Nd, Pb and B isotopic compositions of the analysed samples are reported in Table 2. Sr, Nd and Pb show a range of variation: 87Sr/86Sr 0· 70666–0· 70758, 143Nd/144Nd 0· 51255–0· 51245, 206Pb/204Pb 19· 04–19· 10, 207Pb/204Pb 15· 69–15· 74, 208Pb/204Pb 39· 20–39· 36. These ranges are slightly smaller than those reported in the literature for Vesuvian rocks (87Sr/86Sr 0· 7062–0· 7081; 143Nd/144Nd 0· 51261–0· 51238; 206Pb/204Pb 18· 94–19· 12; 207Pb/204Pb 15· 38–15· 72, 208Pb/204Pb 38· 92–39· 28) (Hawkesworth & Vollmer, 1979
; Cortini & Hermes, 1981; Civetta et al., 1987, 1991a, 2004a; Civetta & Santacroce, 1992; Caprarelli et al., 1993; Santacroce et al., 1993; Cioni et al., 1995; Ayuso et al., 1998; Somma et al., 2001; Cortini et al., 2004; Piochi et al., 2006). 11B values (Table 2) are fairly similar in all the samples, around –7.
A plot of Sr isotopic composition and SiO2 wt % vs stratigraphic height (Fig. 7) shows a complex pattern from the base upwards. The trachytes of local origin of Group 1 have 87Sr/86Sr of c. 0· 70715, whereas the trachytes and phonolites of Phlegraean origin show an increase upsection of 87Sr/86Sr from c. 0· 7068 to 0· 70712. The CI rocks have 87Sr/86Sr of c. 0· 7074, which is typical of the most evolved CI products (Civetta et al., 1997; D'Antonio et al., 2007). The 87Sr/86Sr of Group 2 trachytes slightly increases upsection from 0· 707413 to 0· 707490, then it changes again in the oldest shoshonites, which are characterized by the least radiogenic Sr isotope composition (c. 0· 7066). The 87Sr/86Sr increases again from shoshonites to phonotephrites to latites, then decreases in the phonotephrites of Group 3, younger than 4 ka. The isotopic differences observed between samples of different provenance and composition reflect the major and trace element differences described above, with the exception of phonotephrites of Groups 2 and 3 that, although they are characterized by a geochemical trend distinct from that of shoshonites–latites, fall in the same Sr isotopic range. 87Sr/86Sr ratios of cpx separated from some of the shoshonites and phonotephrites of Group 2 are also reported. The mineral isotopic data are similar within analytical error to the whole-rock values, with the exception of cpx separates from one shoshonite and one phonotephrite that are more radiogenic than the whole-rock. The silica variation through time well illustrates the different compositions of the products emplaced in the Vesuvian area.
|
The range of 87Sr/86Sr, as measured by the CdT sequence, is large, from 0· 7066 to 0· 7076. It is noteworthy that this range becomes even larger by considering the Vesuvian products of the past 4 kyr (Civetta & Santacroce, 1992
; Santacroce et al., 1993; Ayuso et al., 1998; Civetta et al., 2004a; Piochi et al., 2006).
The Campi Flegrei origin of most of the trachytic pyroclastic units older than the CI (CdT-f, CdT-g, CdT-h, CdT-i, and CdT-j), suggested on the basis of lithological and sedimentological characteristics (Fig. 2), is corroborated by the similarities of their Sr isotope compositions and geochemistry to the Campi Flegrei trachytes and phonolites older than CI (Pappalardo et al., 1999). Figure 8, modified after Pappalardo et al. (1999), indicates the 87Sr/86Sr composition of pre CI Campi Flegrei and CdT trachytes and phonolites. With the exception of four trachyte samples from the CdT-a unit, of local origin (circled in the figure), all other CdT trachytes and phonolites are geochemically similar to the Campi Flegrei trachytes and phonolites older than the CI. Arrows in Fig. 8 link the chemical and isotopic composition of the CdT rocks and equivalent Campi Flegrei rocks.
|
The 87Sr/86Sr ratios of the rocks studied here, and those from the literature, plotted vs SiO2 content (Fig. 9) show three different trends, referred to as A, B and C. Trend A, mostly defined by shoshonites and phonotephrites, is characterized by a large range of Sr isotope composition (0· 7068–0· 7080) at relatively constant SiO2 content (48–49 wt %). Trend B, from shoshonite to latite and phonolite, is characterized by an increase of 87Sr/86Sr (0· 7062–0· 7078), over a SiO2 increase from 50 to 54 wt %. Trend C, mostly described by trachytic pyroclastic rocks of different ages and provenance, both from Vesuvian and Campi Flegrei areas, is characterized by a slight increase of 87Sr/86Sr (0· 7068–0· 7075), with SiO2 increasing from 58 to 65 wt %. A similar trend, although reversed with respect to that of Sr and less evident, is also shown by Nd isotopes (Table 2).
|
The increase of 87Sr/86Sr and the decrease of 143Nd/144Nd with increasing differentiation, in particular for increasing silica from 50 to 54 wt % (shoshonite–latite, trend B of Fig. 9) and, to a lesser extent, from 58 to 65 wt % (trend C, Fig. 9), suggest that crustal contamination has a role in explaining the isotopic variability of Vesuvian magmas. A Sr–Nd isotopic plot of CdT rocks (Fig. 10a) shows an overall inverse correlation between Sr and Nd isotopes. The diagram also shows the fields of the Campi Flegrei (>39 ka), Ischia and Procida rocks, and the CI and Campi Flegrei rocks younger than 39 ka (Civetta et al., 1991b
; D'Antonio et al., 1999a, 1999b, 2007; Pappalardo et al., 1999, 2002a, 2002b; De Astis et al., 2004; Tonarini et al., 2004). The trachytes of Phlegraean origin (Group 1) plot in the field of the Campi Flegrei >39 ka, Ischia and Procida rocks, whereas the trachytes of local origin of Groups 1 and 2 plot in the field of the CI and Campi Flegrei rocks younger than 39 ka. The shoshonites, phonotephrites and latites of Group 2 plot in both fields, showing a good negative isotopic correlation trend that points to isotopic values of both Hercynian continental crust (87Sr/86Sr 0· 710242–0· 71770, 143Nd/144Nd 0· 51203–0· 51266; Rottura et al., 1991) and Mesozoic limestone (87Sr/86Sr 0· 70729–0· 70924, 143Nd/144Nd 0· 51201–0· 51214; Iannace, 1991; Stille et al., 1996).
|
|
DISCUSSION |
|---|
The interpretation of stratigraphical, chemical and isotopical data for the CdT sequence has allowed us to contribute to the reconstuction of both the geological and volcanic history of the Vesuvian area and the evolution of the magmatic feeding system, before and during the growth of the Somma–Vesuvius volcano. Furthermore, the role of variable evolution processes acting on the Vesuvian magmas during their ascent to the surface has been investigated using the chemical and isotopic variations of the CdT sequence. These variations can be attributed to three different processes, crustal contamination, source heterogeneity and magma mixing, whose roles are discussed in the following sections.
Geological and volcanic history of the CdT area
The geological and volcanological investigations on the CdT sequence have shed light on the evolution of the area over the past 130 kyr. The CdT sequence shows that before the CI eruption, low-energy explosive volcanism took place in the Vesuvian area, whereas mostly high-energy explosive eruptions were generated in the Campi Flegrei area. A sequence of marine deposits was discontinuously drilled down to 375 m in the Trecase 1 borehole at the same stratigraphic height (Brocchini et al., 2001). The fossil content of these deposits suggests that they were emplaced in a marginal marine environment, which evolved into a shore environment and then into a clearly subaerial environment. The absence of volcanic deposits in the discontinously drilled Trecase 1 borehole was interpreted by Brocchini et al. (2001) as evidence of lack of volcanic activity in the Vesuvian area. Intercalation of subaerial pyroclastic deposits to thin lacustrine sediments at comparable stratigraphic height within the CdT sequence suggests that for a long time interval the CdT area was located at sea level and was affected by a slow subsidence counterbalanced by tephra accumulation. Emplacement of the CI was preceded by 5 kyr of non-deposition during which a thick, mature palaeosol formed. The elevation of this palaeosol in the CdT borehole (85 m b.s.l.) is incompatible with the sea level at the time of the CI eruption, about 70 m lower than at present (Chappell et al., 1996). This implies that the area has been affected by a net subsidence of at least 15 m in the past 39 kyr. This minimum value increases if the elevation of the top of the CI in the plains around Vesuvius (between 0 and 10 m b.s.l.; authors' unpublished data) and in the CdT borehole (68 m b.s.l.) are compared. On the basis of the available data we can infer that after CI deposition the area was affected by a net displacement of 15–60 m.
CI deposition was followed by the eruption of a tuff cone and a long repose time, which predated the extrusion of a thick sequence of lavas from Mt. Somma.
Since 18· 3 ka (Pomici di Base eruption), the activity of Somma–Vesuvius was mainly explosive. Deposits of four Vesuvian Plinian eruptions were emplaced in the CdT area, interbedded with palaeosols and deposits of the Campi Flegrei high-energy eruptions. This succession sheds light on the time relationships between Vesuvius and Campi Flegrei activity. The shallowest cored deposits belong to the CdT-cone, although they were previously attributed to activity that occurred between 18· 3 and 16· 0 ka (Santacroce & Sbrana, 2003). Their stratigraphic position clearly indicates that the eruption occurred between the Pomici di Avellino and Pomici di Pompei eruptions (3· 7 ka–AD 79).
Crustal contamination
Petrological and geophysical data, as discussed in the Introduction, have indicated the existence beneath Vesuvius of a large ‘deep’ crustal magma chamber, whose top is at c. 8 km b.s.l., and a series of shallower chambers, located in the Mesozoic limestone basement. Therefore crustal contamination could potentially have occurred during magma storage at different depths.
As regards the shallow reservoirs, Civetta et al. (1991a), Santacroce et al. (1993), Cioni et al. (1995) and Del Moro et al. (2001) have concluded that contamination with limestone did not occur during magma storage and differentiation. This conclusion was mostly based on the observation that most Vesuvian rocks are characterized by the same Sr isotopic range as the limestone xenoliths included in Plinian and strombolian products and considered representative of the Mesozoic basement (0· 70740–0· 70793; Cortini & Hermes, 1981; Civetta et al., 1991a; Del Moro et al., 2001; Fulignati et al., 2005). Furthermore, the Sr isotopic composition of the Plinian eruption products shows that in some cases the first extruded low-Sr content, more differentiated magma is more enriched in radiogenic Sr, whereas in others the first erupted low-Sr content, more differentiated magma is less enriched in radiogenic Sr (Civetta et al., 1991a; Civetta & Santacroce, 1992; Cioni et al., 1995). These arguments are good evidence against contamination with limestone in Plinian and strombolian magma chambers, because contamination with more radiogenic material should give in all cases a positive correlation between 87Sr/86Sr and degree of differentiation. According to Civetta et al. (1991a), Santacroce et al. (1993) and Cioni et al. (1995), isotopic variation during the course of the Plinian and strombolian eruptions and the mineral isotopic disequilibria result from mixing of at least two isotopically distinct magma batches, one a residue of previous eruptions, the other rising from the ‘deep’ reservoir. In contrast, limestone contamination has recently been proposed by Piochi et al. (2006) to explain the Sr isotopic range (0· 7073–0· 7078), shown mainly by the Plinian and strombolian products younger than 9 ka, which vary from phonotephrite to phonolite.
As regards magmatic processes in the ‘deep’ reservoir, hosted either in the Hercynian crystalline crust or in the Mesozoic limestone, crustal contamination has recently been proposed by Civetta et al. (2004b), mostly based on thermal modelling. According to their model, repeated arrivals of mantle-derived magma batches since 400 ka, at a depth of about 10 km, generated a large, laterally extensive sill-like magma reservoir, which progressively heated the surrounding crustal rocks and thus allowed contamination.
Crustal contamination could explain the isotopic trend B from shoshonite to latite (Fig. 9), involving an increase in 87Sr/86Sr and a decrease in 143Nd/144Nd with increasing SiO2 (Table 2). To test this, we assume that the least radiogenic and least differentiated shoshonite from unit CdT-n01 (s107.8, 87Sr/86Sr = 0· 706661 ± 8, SiO2 50· 35 wt %), is representative of the uncontaminated or least contaminated magma, whereas the most radiogenic latite from unit CdT-o02 (s50.9, 87Sr/86Sr = 0· 707584 ± 9, SiO2 53· 69 wt %), is representative of the contaminated magma. The sample assumed to represent the parental magma was erupted during the initial stages of volcano growth, after the CI eruption. To simulate contamination, we have used the EC-AFC (energy constrained assimilation and fractional crystallization) model of Spera & Bohrson (2001), which is based on the energy and mass balance between contaminant and magma. In the simulation we have assumed that contamination occurs in the ‘deep’ reservoir, in which fractionating shoshonitic magma could be contaminated either by Hercynian crust (model a) or by limestone (model a').
Using the Hercynian crust as contaminant (model a), we have assumed an initial temperature of 1180°C (T0m) for the shoshonitic magma, similar to that obtained from melt inclusion geothermometric data for Vesuvian phonotephrites (Marianelli et al., 1995), and a temperature of 650°C (T0a) for the crustal rocks, according to the thermal modelling of Civetta et al. (2004b). We have assumed a liquidus temperature (Tla) for these crustal rocks of 840°C, and a solidus temperature (Ts) of 720°C (Wyllie, 1977). We have calculated the crystallization enthalpy for a shoshonitic magma in the temperature range 1180–1000°C using the MELTS algorithm (Ghiorso & Sacks, 1995). The composition of the Hercynian crust, mainly granodioritic and tonalitic, is based on the work of Rottura et al. (1991). The likely bulk partition coefficients for Sr and Nd in the shoshonite have been evaluated according to the mineral phenocrysts and their abundance in the mass balance calculation results for the transition shoshonite–latite (Table 3), using the partition coefficients data of Villemant (1988) for Campanian shoshonite and trachybasalt. The modelling results for Sr and Nd (element abundances and isotopes) are reported in Fig. 11a and b, and the parameters of the simulations are given in Table 4. Modelling results confirm the plausibility of an AFC process acting on the Somma–Vesuvius </