Enriched mantle source for the Central Atlantic magmatic province: New supporting evidence from...

Lithos 188 (2014) 15–32

Contents lists available at ScienceDirect

Lithos

j ourna l homepage: www.e lsev ie r .com/ locate / l i thos

Enriched mantle source for the Central Atlantic magmatic province:New supporting evidence from southwestern Europe

Sara Callegaro a,⁎, Cedric Rapaille b, Andrea Marzoli a, Hervé Bertrand c, Massimo Chiaradia b, Laurie Reisberg d,Giuliano Bellieni a, Línia Martins e, José Madeira f, João Mata e, Nasrrddine Youbi e,g, Angelo De Min h,Maria Rosário Azevedo i, Mohamed Khalil Bensalah e,g

a Università di Padova, Dipartimento di Geoscienze e CNR-IGG, via Gradenigo 6, 35100 Padova, Italyb University of Geneva, Department of Earth Sciences, 13 rue des Maraîchers, 12011 Genève, Switzerlandc Université Lyon 1 et Ecole Normale Supérieure de Lyon, Laboratoire de Géologie de Lyon, UMR CNRS 5276, 46 Allée d'Italie, 69364 Lyon Cedex 7, Franced Centre de Recherches Pétrographiques et Géochimiques (CRPG), CNRS-Université de Lorraine UMR 7358, BP 20, 54501 Vandoeuvre-les-Nancy Cedex, Francee Universidade de Lisboa, Faculdade de Ciências, Departamento de Geologia, Centro de Geologia, Portugalf Universidade de Lisboa, Faculdade de Ciências, Departamento de Geologia, Instituto Dom Luiz (LA), Portugalg Faculty of Sciences-Semlalia, Department of Geology, Cadi Ayyad University, Marrakech, Moroccoh Università degli Studi di Trieste, Dipartimento di Scienze Geologiche, via E.Weiss 8, 34127 Trieste, Italyi Universidade de Aveiro, Departamento de Geociências, GEOBIOTEC, Campus Santiago, 3810-193 Aveiro, Portugal

⁎ Corresponding author.E-mail address: [email protected] (S. Callegaro).

0024-4937/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.lithos.2013.10.021

a b s t r a c t
a r t i c l e i n f o
Article history:Received 19 May 2013Accepted 21 October 2013Available online 21 November 2013

Keywords:Central Atlantic magmatic provinceTholeiitesSr–Nd–Pb–Os isotopesEnriched mantle sourceCrustal recycling

Remnants of the Central Atlantic magmatic province (CAMP), emplaced ca. 201 Ma during the rifting phasesleading to Pangaea breakup, are still preserved in southwestern Europe (SWE) in the form of sills, dykes andlava flows. Low–Ti (TiO2 0.48–1.46 wt.%) tholeiitic basalts and basaltic andesites crop out as sills only in the Pyr-enean area, as dykes (especially the Messejana–Plasencia dyke) from central Spain to the Atlantic coast, and aslava flows within sedimentary basins in Southern Portugal. Here we present new geochemical data (major andtrace elements, mineral chemistry and combined Sr–Nd–Pb–Os analyses) on 132 samples, aiming to investigatethe mantle source of these rocks and correlate themwith magmatism from other areas of the CAMP. Crustal-likesignatures in incompatible element patterns (Nb–Ta troughs, Pb peaks, generally shared by most CAMP rocks)and the enriched Sr–Nd–Pb isotopic characters (87Sr/86Sr200 Ma 0.70529–0.70657; 143Nd/144Nd200 Ma 0.51238–0.51225; 206Pb/204Pb200 Ma 18.15–18.48;

207Pb/204Pb200 Ma 15.57–15.68;208Pb/204Pb200 Ma 37.99–38.52) appar-

ently argue in favor of crustal assimilation playing an important role in the evolution of these magmas. However,the low initial 187Os/188Os values (0.1298 ± 0.0056) as well as the restricted geochemical variations shown bySWE-CAMP rocks over such a large area limit the crustal assimilation of various Iberian lithologies to smallamounts. We thus locate this enrichment in the mantle source, in the form of upper and lower crustal materialrecycled during earlier subduction-related events. This process, while imparting crustal signatures to incompat-ible elements and Sr–Nd–Pb isotopes, would not alter the Os isotopic signature, dominated by the peridotite. Themixed contribution of 3–7% of local upper (pelitic) and lower (felsic granulitic) crust is sufficient to enrich a de-pletedmantle source, which can be either the sub-SWE lithosphere or the upper depleted asthenosphere. Similarprocesses of crustal recycling within the upper mantle have been recognized to be responsible for the mantlesource enrichment in other areas of the CAMP (Eastern North America). Geochemical correlations of the herestudied tholeiites with CAMP rocks from other areas inscribe European basalts within the main pulse of CAMPmagmatism.A subset of samples from Southern Portugal (here defined high-Sr dykes) shows different major and trace ele-ment geochemistry (e.g. Sr and CaO enrichment, SiO2 depletion) as well as more radiogenic 87Sr/86Sr200 Ma

(0.70669–0.70749) and Pb isotopic ratios (e.g., 206Pb/204Pb200 Ma 18.55) at similar 143Nd/144Nd200 Ma (0.51232–0.51224). This reflects a different magmatic evolution for these rocks, dominated by the late-stage assimilationof 10–20% local carbonates.

© 2013 Elsevier B.V. All rights reserved.

ghts reserved.

1. Introduction

Crustal recycling processes (Chauvel et al., 2009) in the lower(Willbold and Stracke, 2006) or upper (Prelević et al., 2013) mantlehave recently drawn growing attention both in the perspective of

http://dx.doi.org/10.1016/j.lithos.2013.10.021

mailto:[email protected]


http://www.sciencedirect.com/science/journal/00244937

16 S. Callegaro et al. / Lithos 188 (2014) 15–32

investigating global geodynamics and of unraveling the mantle sourceof Large Igneous Provinces (LIPs; Melluso et al., 2006; Sobolev et al.,2011). In this study we investigate the ca. 200 Ma basaltic rocks fromsouth-western Europe (France, Spain, Portugal) which are part of theCentral Atlantic magmatic province (CAMP; Marzoli et al., 1999) andpropose that they originated from amantle source enriched by recycledcrustal rocks during previous subduction processes rather than from aplume-related source (e.g. McHone, 2000; Puffer, 2001).

Following approximately 25 Ma of rifting (Schlische et al., 2003), thebreakup of supercontinent Pangaea was heralded by a widespread tho-leiiticmagmatic eventwhich occurred at the Triassic–Jurassic boundary,centered on the Central Atlantic rifting zone. Today scattered in fourcontinents, the tholeiites initially covered about 11 × 106 km2

(McHone, 2000), making the CAMP the largest known basaltic LIP onEarth.

Here we investigate the geochemistry of the European offshoot ofthe CAMP, cropping out along the Iberian Peninsula and southernFrance (southwestern Europe, SWE) as dykes, sills, lava flows and pyro-clastic deposits.We present thefirst Os isotopic data and new Sr–Nd–Pbisotopic determinations for the SWE-CAMP. These data, along withcomplementary elemental and mineral chemistry, are here used to de-fine the mantle source of these tholeiites and to compare and correlatethemwith other studied areas of the CAMP, in the aim of contributing toa better understanding of the genesis of this LIP.

2. Tectonic and lithostratigraphic evolution of southwestern Europe(late Paleozoic to late Triassic)

CAMP remnants in southwestern Europe are preserved withinthe Armorican and Aquitanian terranes (France) and the Iberiamicro-plate (Spain–Portugal), whose tectonostratigraphy results fromthe overlapping effects of the Cadomian and Variscan orogenies(e.g., Ribeiro et al., 2007 and references therein). The Permian evolutionthat followed the Variscan amalgamation was strongly constrained bylithospheric weaknesses represented by steeply dipping NNE–SSWand E–W crustal structures (leading to the development of extensionalbasins) and was marked by the intrusion of abundant S- and I-typegranitoids at 300–280 Ma and late stage Upper Permian alkaline basicmagmatism (Bea et al., 1999; Villaseca et al., 1998, 2009). The SWEarea inherited Hercynian structural patterns with a general E–W trend(e.g., the NW Pyrenean and the Gibraltar fault zones; Manspeizer,1994) and dextral strike-slip movements, while subsequent Triassic ex-tension produced subsiding basins filled by continental red sandstonesand shales of the Silves Formation (Carnian?–Norian) followed bysabkha-type evaporite deposits and terrigenous sediments of theDagorda Formation (Norian–Hettangian; Azerêdo et al., 2003; Gómezet al., 2007). The volcanic products of the Portuguese CAMP are interca-lated within the Dagorda Formation, whereas the wall rocks of thePyrenean sills are represented by Norian sediments of the Isabena For-mation (Arnal et al., 2002) or by Keuper salt diapirs. The Central Atlanticrift developed an asymmetric geometry and during the early Mesozoiccontinental stretching several distinct basins formed along thepresent-day West Iberian Margin. Extensional faults related to riftingcontrolled the geometry and subsidence history of these basins, whoseevolutionwasmarked by somewhat distinct tectonostratigraphic histo-ries (e.g., Pereira and Alves, 2011). Most of these basins, currentlyfound either off-shore or on-shore, roughly define a N–S alignment(Alentejo, Lusitanian, Peniche andGalicia). The Algarve basin developedalmost perpendicularly as a pull-apart basin related to left-lateraltranstensional shear zone separating Iberia from Africa as a conse-quence of the relative eastward drift of Africa (e.g. Terrinha et al.,2002). In this extensional scenario, ultimately leading to the openingof the Central Atlantic Ocean, such basins suffered several rifting epi-sodes. CAMPmagmas were emplaced during the second rifting episodeidentified by Pereira and Alves (2011) during the transition from conti-nental to marine influenced sedimentation.

3. CAMP in southwestern Europe (SWE-CAMP)

European CAMP crops out in a pattern typical of thismagmatic prov-ince elsewhere (McHone, 1996), where frequent dyke swarms/groupsare accompanied by a few isolated but very large dykes, and minoroccurrences of sills and volcanic extrusions (lava flows and pyroclasticdeposits), preserved in basinal volcano-sedimentary series. The samplesfor this study (n = 132) were collected in Spain, Portugal and France(sampling sites in Fig. 1; further outcrop description and samplingdetails are included in the Supplementary material).

CAMP dykes emplaced across the SWE are epitomized by the530 km long, 5–200 m thick Messejana–Plasencia dyke (Bertrand andMillot, 1987; Cebriá et al., 2003; Sebai et al., 1991), trending NE–SWacross the Iberian Peninsula from Central Spain to southwesternPortugal (61 samples; sites 17 to 32) and intruding the local metapeliticand granitic basement. Other smaller dykes, generally referred to ascoastal dykes, crop out in Brittany (north-western France; Jourdanet al., 2003; Marzoli et al., 2014; not sampled) and in SouthernPortugal where they are either roughly coast-parallel, N–S trending orparallel to the Messejana dyke (3 samples; sites 33 to 35).

Sills and lava flows are instead enclosed within more restrictedareas. Numerous sills (27 samples; sampling sites 1–16) occur alongmost of the Pyrenean chain within the Aquitanian basin (France) andwithin the Cantabrian range (Spain). The lava flows crop out involcano-sedimentary sequences (up to 8 flows, total thickness up to130 m) in the E–Wtrending Triassic–Jurassic basin of Algarve, SouthernPortugal (28 samples; sites 36 to 45), and in a narrowN–S basin of sim-ilar age (Santiago do Cacém basin, Alentejo; 2 samples; sites 46 and 47).Strongly altered lava flows were also observed in the southern edge ofthe Lusitanian Basin (Sesimbra region; not sampled).

Several thin dykelets (b1 mthick; 11 samples; sites 35–47) attributedto a late event crosscut the Portuguese extrusive sequence and were alsosampled. They trend E–Wto N110E (parallel to the orientation of the ba-sins), though their distinction from the lava flows is often difficult. Thesedykes, alongwith 2 lava flows, are grouped together in thiswork as high-Sr dykes, due to their geochemistry (Sr N 230 ppm; CaO N 12.5 wt.%; seeSections 7.2 and 7.3), as was also previously observed by Martins et al.(2008), who referred these basalts to as high-Ca rocks.

4. Methods

Major and trace element compositions of the SWE-CAMP rockswereanalyzed by X-ray fluorescence (XRF) at the Centre d'AnalysesMinéralogiques, University of Lausanne (CH; Phillips PW1400; OF, MDand AL samples), at the University of Lyon (FR; Phillips PW1400; P orE samples) or at theUniversity of Padova (IT; Philips PW2400; PIM sam-ples). Trace elements were analyzed by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) at the École Normale Supérieure of Lyon(FR) or at the University of Grenoble (FR). Mineral major elementcompositions were analyzed at Lausanne or at the IGG-CNR Padova(Italy) on a Cameca SX50 electron microprobe.

Sr–Nd–Pb isotopic compositions were determined using a Thermo-Ionization Mass Spectrometer (TIMS; Finnigan MAT262) at the Univer-sity of Geneva (CH; OF, MD, AL, E and P samples) or at the Laboratory ofIsotopic Geology at the Universidade de Aveiro (PT; PIM samples) afterleaching and cascade chemical separation from a unique sample aliquot.

Re and Os concentrations (by isotopic dilution) and Os isotopiccompositions were analyzed at the CRPG-CNRS of Nancy (France) bynegative thermal ionization mass spectrometry (NTIMS; FinniganMAT262). Re isotopic compositions for isotope dilution calculationswere measured by ICPMS (Micromass Isoprobe).

After irradiation in the CLICIT facility of the TRIGA reactor atOregon State University (U.S.A.), plagioclase separates were analyzedfor 40Ar/39Ar at the Geoscience Azur laboratory of Nice (FR) by laserstep-heating under a Coherent Innova 70-4 continuous argon-ion

Aquitanian basin

Ebro basin

Guadalquivir basin

Tajo basin

Duero basin

3536

37

38

39

40414243

44

45

34

35

33

47

46

Coastaldykes

Algarve

Santiagodo Cacém

N

FRANCa

b

c

0 25 km

12

13

12

3

4

56

1416

7

15

Lisbon

Madrid

Mes

seja

na d

yke

Pyrenees

Iberian Ranges

Betic Cordillera

Ronda

171819202122

232425

2627

2829

3132

33

34

30

Avila

0 200 km

1110

89

Pre-Mesozoic rocks

CAMP tholeiitesAlpine fold belt

Mesozoic covers

PO

RT

Fig. 1. Sampling sites plotted on a simplified geological map of the Iberian. Inset (a): overview of Iberia, centered on the Messejana dyke. Inset (b): Pyrenean area. Inset (c): SouthernPortugal. The correspondence between sampling sites and samples is defined in the Supplementary material.

17S. Callegaro et al. / Lithos 188 (2014) 15–32

laser. Extended analytical procedures and details are reported in theSupplementary material.

5. Petrography

Despite their wide geographic distribution and their contrastingmodes of emplacement, as intrusive (Pyrenean sills and Iberian dykes)or extrusive (Portuguese lava flows) bodies, the SWE-CAMP tholeiitesshow similar textures and mineralogy. The mineral assemblage ofthe Pyrenean basaltic sills and of the Iberian dykes mainly consists ofplagioclase, Ca-rich (augite) and Ca-poor (pigeonite) clinopyroxene,olivine and Fe–Ti oxides (magnetite with exsolved ilmenite). As usualin tholeiitic rocks, there is textural evidence for plagioclase crystalliza-tion preceding that of pyroxene. Late stage crystallization is representedby quartz and/or quartz and alkali-feldspar intergrowths (forminggranophyric textures in the coarse-grained Messejana dyke facies)

plus rare apatite, primary biotite and brown amphibole. Three texturaltypes are observed within the sills, i.e. a) a fine-grained chilled facies,porphyritic with intersertal to intergranular groundmass, found up toa few tens of cm from the contact with the sedimentary wall-rock;b) an intermediate ophitic to sub-ophitic facies, collected within a fewmeters from the margins, and c) a coarse-grained (almost gabbroic)ophitic facies, with poikilitic clinopyroxene including laths of plagio-clase found in the inner part of the sills. The lava flows from thePortugal basins (Algarve and Santiago do Cacém) are in general finergrained than the SWE-CAMP dykes and sills, due to faster cooling ofthe relatively thin lava flows. Observed textures vary from porphyritic–intersertal to subophitic and the mineralogy is similar to that describedfor the subvolcanic bodies.

The petrographic characteristics of the high-Sr rocks from theAlgarve basin differ from those generally observed in the SWE-CAMP.They show finer-grained porphyritic textures (see also high-Ca rocks


descriptions by Martins et al., 2008) with phenocrysts of plagioclase,olivine and clinopyroxene (augite). Frequently these clinopyroxenesare strongly zoned showing augite cores rimmed by a Ca–Al-richclinopyroxene similar to those from thematrix, where plagioclase, oliv-ine and oxides are also present. Notably, one of these rocks (AL6)contained a dark green, centimeter-sized xenolith, showing a granular(amphibolitic) texture with prismatic green amphibole and late crystal-lized poikilitic plagioclase, accompanied by zeolites, quartz and zircon.

Alteration is moderate in the SWE-CAMP tholeiites and mainly af-fects the coarser grained rocks. Samples showing pervasive alterationwere excluded from further consideration. Sheet silicates (on interstitialglass), sericite (on plagioclase) and iddingsite (after olivine) are themain alteration minerals. In some samples rare pyrite and chalcopyriteof very small size (ca. 0.01 mm) was identified.

6. 40Ar/39Ar dating

40Ar/39Ar radio-isotopic dating was performed on a plagioclase(An63–48) separate from a Messejana dyke sample (MD25, collectednear Odemira, Portugal). It yielded a plateau age, though with a saddle-shaped spectrum suggestive of excess argon and with a large analyticalerror due to the low K content of the analyzed plagioclase (detaileddata chart and diagram are included in the Supplementary material).Recalculated after Renne et al. (2010), MD25a yields an age of206.5 ± 2.7 Ma defined by 73.4% of the released Ar. Plateau steps yieldCa/K (calculated from 37Ar/39Ar) consistent with the composition ofMessejana dyke plagioclase phenocrysts (Ca/K varying from ~45 in thecore to ~10 at the rim). Due to the observed Ar excess (40Ar/36Ar inter-cept on the inverse isochron: 378.6 ± 29.4), the inverse isochronage of sample MD25 is preferred (203.1 ± 5.6 Ma; MSWD 11.2). This40Ar/39Ar age for the Messejana dyke is similar, within error, to thosepreviously published 40Ar/39Ar plateau ages for the Algarve and Santiagodo Cacém lava flows by Verati et al. (2007; 199.7 ± 1.4 Ma, after recal-culation for the decay constant following Renne et al., 2010) and tothose from Sebai et al. (1991), for the Messejan–Plasencia dyke. Theselatter authors did not obtain 40Ar/39Ar plateau ages on plagioclaseseparates, but combining data with those from biotite, amphibole andpyroxene, they determined an age of ca. 200 Ma for the dyke, withoutvariations along its 530 km, supporting emplacement in a single briefepisode. Pyrenean sill samples from this study did not yield plateauages. So far only K–Ar ages are available (195 ± 10 Ma and191 ± 10 Ma on whole-rock; Walgenwitz et al., 1976; 197 ± 7 Maand 195 ± 8 Ma on plagioclase separates, Montigny et al., 1982;recalculated for the 40 K decay constant of Steiger and Jäger, 1977),except for a U/Pb zircon age of 198.7 ± 2.1 Ma (Rossi et al., 2003) on aPyrenean weathered rock, interpreted as a CAMP tuff.

7. Geochemistry

7.1. Mineral chemistry and thermobarometry

Olivine crystals were analyzed for the Pyrenean sills, the Messejanaand the high-Sr dykes, whereas fresh olivine was not found withinthe Portuguese lava flows. The olivine cores from the intrusive rockshave rather uniform and relatively forsterite (Fo)-rich compositions(Fo82–71) in the dykes, whereas in the Pyrenean sills, even for distinctcrystals of the same rock, olivine cores display a larger range(Fo82–55). Normal zoning (more fayalitic rims) is moderate in olivineenclosed in augite and more evident for the phenocrysts in contactwith the groundmass. In general, olivine cores are too low in Fo tobe in equilibrium with the whole-rock composition considering aKDFe/Mg = (Fe/Mg)ol / (Fe/Mg)liq of 0.30 ± 0.03 (Roeder and Emslie,

1970). The chemical disequilibrium between olivine cores and whole-rocks is not consistent with textural evidence, since olivine crystalliza-tion in all SWE-CAMP magmas is not apparently preceded by thecrystallization of other Fe–Mg phases (e.g., pyroxene). Accumulation

of Mg-rich minerals could explain an increase of the Mg# of thewhole-rock or else the low-Fo olivine core compositions reflect a pro-gressive diffusive Fe–Mg re-equilibration between the crystallizingmineral and the slowly cooling and evolving residual liquid (mostly indykes and sills).

Augite and pigeonite display various compositions in both intrusiverocks and lava flows. Some of the slow cooled intrusive rocks presentexsolved lamellae of orthopyroxene in pigeonite (or inverted pigeon-ite). Pyroxenes are excellent recorders of the liquid evolution withtheir Mg# (100 × Mg / (Mg + Fe2+)), varying from 89–67 in the sillsto 88–49 in theMessejana dyke and 90–39 in the lava flows. Pyroxenesfrom the chilled margins and pyroxene cores from sub-ophitic rockscrystallized early, and show relatively homogeneous augitic composi-tions (Wo39–41En49–51Fs7–9). Pigeonite (often Mg-rich; Wo9En70Fs21)and augite display two parallel trends of increasing TiO2 with decreas-ing Mg# (not shown), confirming the co-precipitation of these phasesbefore Ti-oxide saturation. Plagioclase phenocryst cores range incomposition from bytownite to labradorite for the Pyrenean sills(An86–64Ab14–35Or0.2–1.1), the Iberian dykes (An86–56Ab13–42Or0.3–1.4),and the Portuguese lava flows (An85–66Ab15–33Or0.3–1.0). High-Sr dykesfrom the Portuguese basins display a peculiar mineral assemblage ex-pressing an abrupt change in magma composition, towards higherCaO contents with differentiation. In fact, these rocks show relativelylower SiO2 (45.2–50.0 wt.%) and higher CaO (12.5–18.5 wt.%) contentswhen compared to the dominant ophitic/subophitic facies rocks (SiO2

49.5–53. 3 wt.%; CaO 7.7–13.3 wt.%). They are also silica–undersaturated(slightly nepheline normative, with the exception of 4 samples that areolivine-hypersthene normative) in contrast to the typical saturated/oversaturated character of the CAMP magmatism. The high sensitivityof pyroxene to magma composition variations is best exemplifiedby the abrupt changes in crystal chemistry between phenocryst cores(augites;Wo40En50Fs10) and their Ca–Al-rich rims (Wo50En35Fs15). Crys-tallization of pyroxene cores is contemporaneouswith that of plagioclase(An55) and olivine (Fo63) phenocrysts. These are set in a fine- to veryfine-grained granular matrix, which also includes (Ca, Al-rich)clinopyroxene, Ca-rich plagioclase (An80), Fe-rich olivine (Fo50) and Ti-magnetite, indicating a significant Ca-enrichment in the magma beforematrix crystallization.

The augite–pigeonite geothermometer (QUILF; Andersen et al.,1993) yields equilibrium temperatures ranging from 1170 °C to1081 °C in some SWE-CAMP rocks. An orthopyroxene–augite couple(from sample MD13) yields a temperature of 1197 °C. Thesegeothermometric estimates suggest initial clinopyroxene crystallizationat 1230 ± 20 °C for the Algarve lava flows (see Martins et al., 2008).The clinopyroxene geobarometer of Nimis and Ulmer (1998) indicatesupper crustal crystallization pressures (from 0.2 to 0.35 GPa) for corecompositions of augites of both the Iberian dykes and the Portugueselava flows. Rim compositions of the same augites yield pressures rang-ing from 0.05 to 0.18 GPa, suggesting shallow subsurface conditionsfor the final crystallization steps. Clinopyroxene crystallization mayhave started at pressures of 0.69 ± 0.11 GPa as calculated by Martinset al. (2008).

7.2. Major element whole rock geochemistry

According to the total alkali versus silica (TAS) classification (LeMaitre et al., 2002; Fig. 2), the SWE-CAMP samples are mainly sub-alkaline basalts with rare basaltic andesites (the latter correspondingonly to lava flows and Messejana dyke samples). The majority is repre-sented by quartz-normative rocks, but one lava flow (AL20), 3 samplesfrom the Messejana dyke (MD13, MD15, MD37) and one from thePyrenean sills (OF34) are olivine/hypersthene-normative. In contrast,high-Sr dykes are mostly nepheline normative. The SWE-CAMP tholei-ites show low TiO2 contents (0.48–1.46 wt.%), as domost CAMP basalts,which are mainly low-Ti (TiO2 strictly lower than 2.0 wt.%) continentaltholeiites (e.g. De Min et al., 2003; Marzoli et al., 2004). As a whole,

Pyrenean sills

Messejana dyke

Coastal dykes

Portuguese lava flows

High-Sr dykesBasalt

BasalticTrachyandesite

BasalticAndesite

SiO2 (wt%)

Na 2O

+K

2O (

wt%

)

1

2

3

4

5

6

4544 46 47 48 49 50 51 52 53 54

Fig. 2. Total Alkali versus Silica (TAS) classification diagram (LeMaitre et al., 2002) of the 132 sampled SWE-CAMP sills, dykes and flows.


SWE-CAMP rocks have moderately evolved compositions, with MgOcontent mostly varying between 9.4 and 4.1 wt.%. Slightly higher MgOcontents (11.0–10.5 wt.%) are displayed by a cluster of 4 samples ofPyrenean sills, which may however be affected by accumulation ofmafic minerals, as suggested by olivine compositions (see Section 7.1).Globally, the data range is consistent with those of previous studies(e.g., Alibert, 1985; Béziat et al., 1991; Cebriá et al., 2003; Martins andKerrich, 1998; Martins et al., 2008). Major element compositions andvariations (Fig. 3) are broadly similar for sills, dykes and lava flows,and they do not vary with the sampling location or stratigraphic height(for the lava piles). Notably also, theMessejana dyke does not show anysystematic variation along its ca. 530 km length. In general, SiO2, TiO2,Na2O, K2O, and P2O5 increase and CaO decreases with decreasing MgO.Al2O3 is nearly constant for the Portuguese lava flows and the high-Srdykes (ca. 13.2–15.2 wt.%), whereas the other intrusive rocks (especial-ly from theMessejana dyke) show a steep increase in Al2O3 with differ-entiation (from 14 to 18 wt.%). SiO2 and CaO contents (and variations)mark the geochemical peculiarity of high-Sr dykes, which, at ca. 8.0 to7.0 wt.% MgO plot in a near-vertical array towards low SiO2 (47.9–49.5 wt.%) and high CaO (18.5–13.6 wt.%) contents (Fig. 3). Theseranges are within those presented in previous studies (i.e. Martins andKerrich, 1998; Martins et al., 2008).

7.3. Trace element whole rock geochemistry

Primitive mantle normalized (Mc Donough and Sun, 1995) multi-element diagrams (Fig. 4a) highlight (a) the similar incompatible ele-ment contents and ratios displayed by all the groups, except for twoMessejana dyke samples (P11 and MD13) with distinctly lower incom-patible element contents, but also relatively higher MgO contents(9.4 and 8.3 wt.%) compared to the other Messejana samples; (b) thegenerally high contents for the most incompatible elements (Rb, Ba,Th, U, K, La, Ce, Pb, Sr) in SWE-CAMP rocks (around 10–30 times higherthan the primitive mantle) as compared with the lower enrichment inthe less incompatible ones (Nd, P, Zr, Hf, Sm, Eu, Gd, Tb, Ti, Dy, Ho, Y,Er, Yb, Lu; less than 10 times primitive mantle values); (c) the couplingof positive Pbwith negative Nb–Ta anomalies, which is a common char-acteristic of CAMP basalts (cf. Callegaro et al., 2013; Merle et al., 2011;Puffer, 2001). Most rocks also display a through at Ti and P and a peakat K. Similar arc-like incompatible element patterns have been observedalso for Late Carboniferous (300–250 Ma) gabbros from the same area

(Orejana et al., 2009); (d) the high Sr contents (348–1046 ppm) ofthe accordingly named Portuguese rocks, showing up to 4 times the Srcontent of other SWE-CAMP rocks at similar MgO (140–260 ppm) andrelatively low Rb (2–16 ppm; cf. also Martins and Kerrich, 1998).

In a chondrite normalized (CH;McDonough and Sun, 1995) diagram(Fig. 4b), most of the SWE-CAMP rocks display similar Rare EarthElement (REE) contents and sub-parallel REE patterns, with a moderateenrichment of light (LREE) vs. heavy REE (HREE; LaCH/YbCH = 2.3–3.6).In general, REE concentrations are more variable for Iberian dykes andsills than for the Portuguese lava flows and high-Sr dykes. In particular,LREE/HREE is slightly higher (LaCH/YbCH = 3.1; PIM22) in basaltssampled at the base of the volcanic piles in Algarve (Rocha dos Soidossection) with respect to the stratigraphically higher flows (LaCH/YbCH = 2.6; PIM28). For most of the samples the enrichment in LREErelative to HREE is correlated with REE and SiO2 concentrations. Theleast and the most enriched tholeiites show slightly positive (up to1.15) and negative (down to 0.85) Eu anomalies, respectively (calculatedas: Eu/Eu* = EuCH / [(GdCH × SmCH)^0.5]).

7.4. Sr–Nd–Pb isotopes

From the 132 SWE-CAMP samples, 59were selected for 87Sr/86Sr, 51for 143Nd/144Nd, 41 for Pb and 13 for 187Os/188Os isotopic analyses(Tables 1 and 2). Isotopic values were age-corrected to 200 Ma takinginto account trace element contents obtained by ICPMS and, in thecase of Re–Os, by isotope dilution.Most samples show rather high initial87Sr/86Sr200 Ma (0.70529–0.70657) and low initial 143Nd/144Nd200 Ma

(0.51238–0.51225; 0.1 N εNd N −2.5; Fig. 5; Table 1) whereas high-Srdykes extend to more radiogenic Sr isotopic compositions (0.70669–0.70749). Thus, if we exclude high-Sr samples, the whole isotopic dataset displays a broad negative trend, parallel to the mantle array, inthe 87Sr/86Sr200 Ma versus 143Nd/144Nd200 Ma isotopic space (Fig. 5a).Previously published (Alibert, 1985; Cebriá et al., 2003; Martins andKerrich, 1998; Martins et al., 2008) Sr and Nd isotopic ratios for theMessejana dyke and the Pyrenean sills yielded a very similar range of87Sr/86Sr200 Ma (0.70522–0.70747) and 143Nd/144Nd200 Ma isotopic com-positions (0.51227–0.51248; 0.85 N εNd N −3.4). In general, despitethe wide scatter of the SWE-CAMP data, the most evolved rocks (lowMgO) have also lower 143Nd/144Nd200 Ma and higher 87Sr/86Sr200 Ma.

The age-corrected 206Pb/204Pb200 Ma (18.15–18.54), 207Pb/204Pb200 Ma

(15.57–15.67), and 208Pb/204Pb200 Ma (37.99–38.54) of the SWE-CAMP

Pyrenean sills

Messejana dyke

Coastal dykes


High-Sr dykes

SiO

2

54

52

50

48

46

44

18

16

14

12

20

16

12

8

CaO

Al 2O

3

1.5

1.0

0.5

0.0

K2O

1.6

1.4

1.2

1.0

0.8

0.6

0.4

TiO

2

13

12

11

10

9

8

7

Na 2O

Fe 2O

3 to

t

3.0

2.6

2.2

1.8

1.4

1.0

0.3

0.2

0.1

0.0

P2O

5

3 6 9 123 6 9 12

MgO MgO

FC from AL3 (MELTS)

FC from MD33 (MELTS)

Fig. 3.Major element variation diagrams for 132 SWE-CAMP tholeiites. Dashed and solid lines represent liquid lines of descent calculated with MELTS (Ghiorso and Sack, 1995) startingfrom a dyke (MD33) and a Portuguese lava flow (AL3) sample, respectively. Fractional crystallization was modeled at 0.1 GPa for near-anhydrous compositions (b2 wt.% H2O in startingmagma),with oxygen fugacity at theQuartz–Fayalite–Magnetite buffer. Thickmarks are plotted every 20 °C drop in T from the liquidus (1175° and 1221° forMD33 andAL3 compositions,respectively).


samples correlate positively. They plot above the NHRL (Hart, 1984) andcompared to it they define a sub-parallel array in the 208Pb/204Pb200 Ma

vs. 206Pb/204Pb200 Ma space and a slightly oblique one in the 207Pb/

204Pb200 Ma vs. 206Pb/204Pb200 Ma space (Fig. 5c–d). In this latter diagram,the Pyrenean sills (18.22–18.48; 15.60–15.68) and the Portuguese lavaflows (18.15–18.34; 15.58–15.63) have the most and least radiogenic

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

Rb Ba Th U K Nb Ta La Ce Pb PSrNd Zr Hf

Sm Eu Gd Tb TiDy Ho Y Er

Yb Lu

10

100

1

10

100

Roc

k/P

rimiti

ve m

antle

Roc

k/C

hond

rite

1

Pyrenean sills

Messejana + coastal dykes


High-Sr dykes

Incompatible Elements

Rare Earth Elements

a

bPyrenean sills

Messejana + coastal dykes

High-Sr dykes


Fig. 4. (a) Primitive mantle-normalized (McDonough and Sun, 1995) incompatible ele-ment and (b) chondrite-normalized REE contents of 59 SWE-CAMP tholeiites. Composi-tional fields are outlined for the 4 groups of samples.


compositions, respectively. The analyzed samples from the Iberian (bothMessejana and Coastal) dykes (18.16–18.39; 15.59–15.65) plot betweenthe two groups above. Even with a large overlap, the Pyrenean sills tendto have higher 208Pb/204Pb200 Ma (38.00–38.52) compared to the Iberiandykes (38.09–38.29) and the lava flows (38.01–38.15). Only one sample(P40) from the Portuguese high-Sr dykes was analyzed for Pb isotopes,and it yielded relatively high 206Pb/204Pb200 Ma (18.54) and 208Pb/204Pb200 Ma (38.54) values, plotting out of the cluster formed by theother samples, yet falling within the range of the other samples for207Pb/204Pb200 Ma (15.64).

Plots of Pb isotopes vs. Sr and Nd isotopic compositions are generallyscattered for the SWE-CAMP basalts. However, the samples with themost radiogenic initial Pb isotopic compositions have also high 87Sr/86Sr200 Ma values. Also, for samples with 143Nd/144Nd200 Ma b 0.51233(εNd N −1.0), theNd versus Pb initial isotopic ratios are negatively cor-related, while this correlation is not observed for samples characterizedby higher 143Nd/144Nd200 Ma values.

7.5. Os isotopes

The analyzed SWE-CAMP basalts (9 intrusive and 4 lava flows)display moderate Os (32–188 ppt) and Re (213–678 ppt; Table 2) con-tents which are not correlated. Though the sample most depleted in Re(213 ppt) is a little evolved sample from the Messejana dyke (MD13;MgO 9.4 wt.%; Os 48 ppt), the Portuguese lava flows are in generalmore depleted in Re (326–396 ppt; except the outlier AL22, yielding537 ppt) than the other tholeiites and display the largest range of Oscontents, up to 188 ppt (AL20). Only the Coastal dyke MD37 yields asimilarly high Os content (172 ppt) whereas the other dykes and sills

all show Os b 99 ppt. While no relationship exists between Os contentand MgO, a rough anti-correlation is observed between Re and MgO(not shown). Positive correlations are visible in plots of Re versus REEand HFSE (not shown), coherent with the fact that all these elementsbehaved incompatibly.

Measured 187Os/188Os ratios in the 13 analyzed whole-rock samplesrange from 0.159 to 0.405, while 187Re/188Os ranges from 9.1 to 95.2.Eleven of these samples display a linear 187Os/188Os vs. 187Re/188Os cor-relation, suggesting an apparent errorchron age of 199 ± 8 Ma(MSWD = 7.7; calculated with Isoplot3; Ludwig, 2003) with a 187Os/188Os intercept of 0.1298 ± 0.0056 (Fig. 7; all uncertainties 2σ).

In this diagram, samples MD42 and AL19 plot at higher 187Re/188Osvalues compared to the apparent errorchron, and were thus excludedfrom the age calculation. Also, the calculated initial ratios for samplesMD42 and AL19 are not realistic, indicating that these samples experi-enced Re-loss or Os-gain substantially after emplacement. As the Recontents of these samples are not elevated compared to those of othersamples with similar MgO contents, loss of Os seems more likely, espe-cially as these samples have among the lowest Os concentrations of allthe samples analyzed. Sample E20was also excluded, because it yieldedan anomalously high Re content (1698 ppt). Since it was crushed (un-like all other analyzed samples) in a Ni–Cr grinder this may resultfrom contamination. SampleMD18was analyzed twice.While themea-sured 187Os/188Os (0.248 and 0.199) and 187Re/188Os (37.7 and 21.5) ra-tios differ substantially between the two powder aliquots, the initial187Os/188Os (0.123 ± 0.004 and 0.128 ± 0.003) agree within uncer-tainty. This agreement suggests that the difference in the measured ra-tios is due to the “nugget effect”, i.e. the presence ofmicrophases such assulfides rich in Os. These can impart Re/Os heterogeneity to an igneousrock, which with time develops into Os isotopic heterogeneity.

As a whole, the initial 187Os/188Os ratios of the 11 samples definingthe ~198 Ma errorchron range from 0.123 ± 0.004 to 0.138 ± 0.009,and thus display only a limited amount of variation despite the largegeographic area considered. Such initial Os isotopic values are similarto those of CAMP basalts from North and South America (Callegaroet al., 2013; Merle et al., 2011, accepted for publication).

8. Discussion

8.1. Closed system magmatic differentiation

Taking into account the Mg# and compatible element compositions(see data table in Supplementarymaterial) of the studied rocks, none ofthem can be considered to be a primitive magma. However, in order toconstrain the role that closed-systemdifferentiationmight play in creat-ing the relatively wide compositional range of the SWE-CAMP rocks, wechose 2 samples (MD33 and AL3;MgO8.0 and 9.1 wt.%, respectively) torepresent parental magmas for closed-system differentiation modelingof the intrusive tholeiites and the lava flows, respectively. A third sam-ple (OF12)was chosen tomodel fractional crystallization of the Pyrene-an sills, but it yielded similar results asMD33, thus itwas not included inthefigure and discussion. These 3 samples are all ratherfine grained andare consistentwith a near-liquidusmelt.MoreMg-rich samples, notablyfrom the Pyrenean sills, are instead crystal-rich and are probably cumu-lative rocks (see Section 7.2), not suitable for modeling of liquid lines ofdescent.

We calculated liquid lines of descent for fractional crystallization(FC) paths using the software MELTS (Ghiorso and Sack, 1995; Fig. 3),considering fO2 conditions (at the QFM buffer) which are typical oflow-Ti CAMP basalt (e.g. De Min et al., 2003), pressure conditions(0.3 to 0.1 GPa) consistent with our plagioclase compositions andclinopyroxene geobarometry (see Section 7.1) and initial water con-tents ranging between 0 and 2 wt.%. Anhydrous conditions do not re-produce the major element evolution of the intrusive samples, andonly poorlymatch those of the lava flows. On the contrary, low pressure(b0.3 GPa) FC paths of slightly hydrous (b2 wt.% H2O)magmas overlap

Table 1Sr–Nd–Pb isotopic data for SWE-CAMP tholeiites.

Sample Group Rb Sr Sm Nd U Th Pb Re Os87Rb86Sr

147Sm144Nd

238U204Pb

235U204Pb

OF 1 Pyrenean sill 13.4 174 2.67 10.03 0.25 1.08 2.4 0.223 0.161 6.59 0.048OF 2 Pyrenean sill 12.4 205 2.78 10.27 0.22 1.19 2.1 0.658 0.039 0.175 0.164 6.79 0.050OF 3 Pyrenean sill 9.1 246 2.81 10.19 0.18 1.27 1.8 0.107 0.166 0.08 0.001OF 4 Pyrenean sill 12.4 175 2.86 10.55 0.29 1.25 2.0 0.205 0.164 9.02 0.066OF 5 Pyrenean sill 12.6 185 2.65 9.90 0.27 1.14 2.9 0.612 0.056 0.197 0.162 5.85 0.043OF 6 Pyrenean sill 13.8 182 2.79 10.47 0.24 1.22 2.1 0.219 0.161 7.08 0.052OF 12 Pyrenean sill 10.4 165 2.32 8.56 0.24 1.04 2.0 0.182 0.144 7.51 0.055OF 15 Pyrenean sill 11.5 171 2.78 10.34 0.20 1.20 1.3 0.195 0.163 9.92 0.073OF 17 Pyrenean sill 18.0 194 2.57 9.61 0.27 1.06 2.0 0.268 0.162 8.57 0.063OF 22 Pyrenean sill 14.4 208 2.87 11.07 0.29 1.21 2.0 0.200 0.156 9.19 0.068OF 23 Pyrenean sill 11.9 236 2.76 10.40 0.34 1.27 6.1 0.146 0.161 3.46 0.025OF 24 Pyrenean sill 21.8 191 2.97 11.40 0.26 1.27 1.8 0.330 0.158 9.04 0.066OF 30 Pyrenean sill 18.7 171 2.71 9.82 0.17 1.16 1.5 0.316 0.167 7.10 0.052OF 35 Pyrenean sill 20.2 201 2.64 9.95 0.24 1.09 7.7 0.291 0.160 1.98 0.015OF 9 Pyrenean sill 7.7 142 1.94 7.01 0.17 0.75 1.6 0.157 0.147 0.09 0.001OF 34 Pyrenean sill 12.5 145 2.03 7.31 0.17 0.74 2.4 0.489 0.072 0.249 0.168 4.62 0.034MD 2 Messejana dyke 21.8 193 3.14 11.99 0.38 1.65 3.6 0.327 0.159 6.61 0.049MD 4 Messejana dyke 21.7 186 2.91 10.87 0.15 1.23 2.5 0.338 0.162 3.65 0.027MD 10 Messejana dyke 28.0 206 0.393MD 13 Messejana dyke 7.0 171 1.19 4.29 0.12 0.53 1.0 0.213 0.048 0.118 0.167 0.10 0.001MD 16 Messejana dyke 17.2 176 2.83 10.41 0.19 1.19 2.5 0.283 0.164 4.77 0.035MD 18 Messejana dyke 14.0 177 2.78 10.29 0.26 1.12 1.9 0.678 0.088 0.229 0.163 8.32 0.061MD 18b Messejana dyke 0.639 0.144MD 20 Messejana dyke 14.0 174 0.233MD 24 Messejana dyke 24.3 187 3.46 13.61 0.48 1.97 4.8 0.376 0.154 6.19 0.045MD 36 Messejana dyke 27.8 179 2.80 10.55 0.26 1.17 2.4 0.449 0.160 6.86 0.050MD 41 Messejana dyke 11.5 168 2.89 10.69 0.24 1.25 2.1 0.198 0.163 7.04 0.052MD 42 Messejana dyke 13.1 169 2.85 10.58 0.23 1.22 2.1 0.606 0.032 0.224 0.163 6.67 0.049MD 46 Messejana dyke 17.4 206 2.88 11.07 0.30 1.57 4.2 0.244 0.157 4.41 0.032MD 47 Messejana dyke 18.9 172 0.318MD 50 Messejana dyke 23.6 213 0.321E 6' Messejana dyke 19.1 219 2.52 9.74 0.31 1.29 3.2 0.252 0.156 6.08 0.045E 9 Messejana dyke 14.0 161 2.89 10.69 0.23 1.20 2.4 0.252 0.143 6.00 0.044E 20 Messejana dyke 17.9 166 2.65 9.89 0.27 1.12 2.0 1.698 0.075 0.312 0.162 8.25 0.061E 27 Messejana dyke 19.4 191 0.294P 8 Messejana dyke 13.7 169 2.72 10.06 0.25 1.12 5.6 0.235 0.163 0.04 0.000P 11 Messejana dyke 18.1 167 1.96 7.39 0.22 0.95 2.1 0.314 0.161 6.61 0.049MD 33 Coastal dyke 11.3 166 2.65 9.65 0.14 1.09 3.1 0.610 0.041 0.197 0.166 2.87 0.021MD 37 Coastal dyke 15.2 165 2.40 8.83 0.14 0.96 1.0 0.532 0.172 0.266 0.165 0.12 0.001P 23 Coastal dyke 15.2 164 2.71 10.09 0.21 1.15 3.1 0.268 0.162 4.15 0.030AL 2 Portuguese lava flow 21.1 205 3.58 13.99 0.48 2.07 2.9 0.298 0.155 10.47 0.077AL 3 Portuguese lava flow 13.6 176 2.71 10.19 0.28 1.20 1.6 0.224 0.161 11.44 0.084AL 5 Portuguese lava flow 15.4 179 2.69 10.23 0.40 1.26 1.9 0.249 0.159 13.35 0.098AL 10 Portuguese lava flow 16.8 173 3.43 13.17 0.32 1.70 1.6 0.281 0.158 12.45 0.091AL 11 Portuguese lava flow 16.8 176 3.26 12.27 0.35 1.47 2.1 0.276 0.160 10.11 0.074AL 14 Portuguese lava flow 15.7 174 2.81 11.03 0.33 1.54 2.5 0.261 0.154 8.27 0.061AL 19 Portuguese lava flow 9.2 165 2.25 8.35 0.16 0.98 1.2 0.396 0.040 0.161 0.163 8.33 0.061AL 20 Portuguese lava flow 2.4 175 1.82 6.58 0.20 0.72 0.9 0.356 0.188 0.040 0.168 13.60 0.100AL 21 Portuguese lava flow 6.5 183 3.10 11.78 0.33 1.43 1.8 0.103 0.159 11.69 0.086AL 22 Portuguese lava flow 11.3 186 3.25 12.36 0.37 1.59 1.8 0.537 0.038 0.176 0.159 12.85 0.094P 32 Portuguese lava flow 16.6 168 2.92 10.99 0.24 1.53 2.3 0.326 0.099 0.286 0.160 6.56 0.048PIM 4 Portuguese lava flow 21.4 168 3.89 14.47 0.46 1.85 2.5 0.369 0.163 0.15 0.001PIM 8 Portuguese lava flow 15.2 167 2.74 10.08 0.31 1.28 2.0 0.263 0.164 0.13 0.001PIM 22 Portuguese lava flow 20.0 186 3.99 15.31 0.55 2.26 3.4 0.311 0.158 0.14 0.001AL 9 High-Sr dyke 2.6 759 0.010PIM 2 High-Sr dyke 9.5 1046 3.09 11.50 0.46 1.40 1.3 0.026 0.162 0.30 0.002PIM 19r High-Sr dyke 4.0 620 23.00 0.019PIM 24 High-Sr dyke 4.7 953 2.88 11.07 0.39 1.26 0.9 0.014 0.157 0.37 0.003PIM 26 High-Sr dyke 15.6 446 3.00 11.08 0.39 1.40 1.8 0.101 0.164 0.18 0.001PIM 28 High-Sr dyke 5.5 815 2.80 10.06 0.36 1.15 1.4 0.020 0.168 0.22 0.002P 40 High-Sr dyke 2.0 348 2.79 10.55 0.34 1.30 1.6 0.017 0.160 13.66 0.100


well with major element evolution trends of most SWE-CAMP tholei-ites. The most evolved compositions can be reached starting from thechosen parental magmas through 20 wt.% FC. Moreover, most observedmineral compositions are reproduced by this modeling, confirming thatmoderately hydrous conditions favor early olivine saturation. As is typ-ical in tholeiitic magmas, Ca-rich plagioclase is an early crystallizingphase preceding augitic clinopyroxene (as observed in the phenocrystcores). The calculated (MELTS) anorthite content in fractionated plagio-clase increases with decreasing pressure, and this possibly explainssome inverse zoning observed in this mineral. Therefore, crystallizationof olivine, plagioclase and clinopyroxene assemblages dominated the

differentiation of the SWE-CAMP tholeiites, and polybaric crystallizationmainly occurred in the upper crust (0.3–0.1 GPa) even though it mayhave started at the mantle–crust boundary (0.69 ± 0.11 GPa; seeSection 7.1).

The compositions of some samples are not obtained by the calcu-lated liquid lines of descent, namely the high-Sr dykes, (significantlyricher in CaO and poor in SiO2) and some evolved samples withhigher SiO2, Fe2O3tot and lower Al2O3 contents than those predictedby MELTS FC models. To investigate the evolution of these latterrocks and to account for the variations in incompatible elements(e.g. Th and Nb) which are not explained by simple FC processes, we

Table 1Sr–Nd–Pb isotopic data for SWE-CAMP tholeiites.

232Th204Pb

187Re188Os

87Sr86Sr

143Nd144Nd

206Pb204Pb

207Pb204Pb

208Pb204Pb

187Os188Os

87Sri86Sr

143Ndi144Nd εNdi

206Pbi204Pb

207Pbi204Pb

208Pbi204Pb

187Osi188Os

29.3 0.70616 0.51257 18.537 15.641 38.652 0.70552 0.51236 −0.35 18.329 15.630 38.35837.5 84.22 0.70677 0.51255 18.648 15.660 38.830 0.4049 0.70627 0.51233 −0.97 18.433 15.649 38.454 0.12660.6 0.70653 0.51255 0.70622 0.51233 −0.91

39.9 0.70621 0.51257 18.510 15.619 38.486 0.70563 0.51235 −0.53 18.225 15.605 38.08625.6 53.38 0.70624 0.51259 18.489 15.631 38.424 0.3115 0.70568 0.51237 −0.15 18.304 15.622 38.167 0.135137.2 0.70621 0.51255 18.483 15.631 38.487 0.70559 0.51234 −0.81 18.259 15.620 38.11333.7 0.70581 18.556 15.635 38.594 0.70529 18.319 15.623 38.25661.1 0.70600 0.51254 18.598 15.639 38.611 0.70545 0.51233 −0.97 18.285 15.623 37.99835.1 0.70659 0.51256 18.748 15.683 38.833 0.70583 0.51235 −0.58 18.477 15.669 38.48139.7 0.70667 0.51253 18.514 15.625 38.529 0.70610 0.51232 −1.18 18.224 15.610 38.13013.4 0.70669 0.51255 18.424 15.658 38.431 0.70627 0.51234 −0.75 18.315 15.652 38.29746.1 0.70659 0.51256 18.608 15.629 38.640 0.70565 0.51236 −0.47 18.322 15.615 38.17849.0 0.70642 0.51259 18.588 15.643 38.625 0.70552 0.51237 −0.27 18.364 15.632 38.1339.2 0.70688 0.51255 18.544 15.682 38.615 0.70606 0.51234 −0.84 18.481 15.679 38.5230.4 0.70585 0.70541

20.2 32.98 0.70676 0.51256 18.551 15.651 38.550 0.2375 0.70605 0.51234 −0.81 18.405 15.644 38.347 0.128629.6 0.70730 0.51254 18.580 15.646 38.586 0.70637 0.51234 −0.88 18.371 15.635 38.28931.4 0.70711 0.51259 18.503 15.622 38.503 0.70615 0.51237 −0.11 18.388 15.616 38.188

0.70750 0.51250 18.468 15.618 38.404 0.706380.5 21.60 0.70618 0.51253 0.2082 0.70584 0.51231 −1.32 0.1368

31.1 0.70709 0.51258 18.355 15.597 38.403 0.70628 0.51236 −0.36 18.204 15.589 38.09137.4 37.72 0.70638 0.51257 18.423 15.633 38.521 0.2478 0.70572 0.51235 −0.52 18.160 15.620 38.145 0.1231

21.53 0.1990 0.12780.70634 0.51258 18.494 15.614 38.443 0.70567

26.4 0.70760 0.51246 18.470 15.631 38.491 0.70653 0.51225 −2.48 18.274 15.621 38.22632.2 0.70711 0.51256 18.556 15.638 38.608 0.70583 0.51235 −0.54 18.339 15.627 38.28538.2 0.70589 0.51256 18.503 15.628 38.525 0.70532 0.51234 −0.75 18.281 15.617 38.14137.4 95.24 0.70606 0.51256 18.491 15.623 38.504 0.70542 0.51235 −0.68 18.280 15.612 38.129 0.077724.2 0.70674 0.51248 18.496 15.656 38.509 0.70604 0.51227 −2.08 18.357 15.649 38.267

0.70731 0.51250 18.587 15.642 38.552 0.706410.70696 0.51254 18.519 15.646 38.567 0.70605

26.1 0.70711 0.51254 18.498 15.621 38.449 0.70639 0.51233 −0.96 18.306 15.611 38.18832.0 0.70623 18.424 15.608 38.411 0.70551 18.234 15.598 38.09035.5 110.82 0.70631 0.51256 18.534 15.629 38.583 0.2813 0.70542 0.51235 −0.56 18.273 15.616 38.227 −0.0849

0.70695 0.51273 0.706110.2 0.70639 0.51258 0.70572 0.51237 −0.30

29.5 0.70656 0.51251 18.512 15.647 38.473 0.70567 0.51229 −1.67 18.303 15.636 38.17722.3 73.99 0.70588 0.51257 18.396 15.628 38.423 0.3822 0.70532 0.51235 −0.58 18.305 15.623 38.199 0.13770.8 14.94 0.70636 0.51258 0.1792 0.70560 0.51236 −0.33 0.1298

23.8 0.70651 0.51260 18.359 15.622 38.440 0.70574 0.51238 0.06 18.228 15.615 38.20146.1 0.70700 0.51250 18.553 15.621 38.491 0.70615 0.51230 −1.60 18.222 15.604 38.02850.1 0.70646 0.51256 18.635 15.608 38.586 0.70582 0.51235 −0.66 18.274 15.590 38.08443.0 0.70617 0.51253 18.576 15.634 38.530 0.70546 0.51233 −1.08 18.154 15.613 38.09867.9 0.70641 0.51255 18.669 15.646 38.693 0.70561 0.51234 −0.80 18.276 15.626 38.01244.6 0.70643 0.51258 18.529 15.592 38.457 0.70565 0.51237 −0.26 18.210 15.576 38.01039.7 0.70630 0.51255 18.513 15.608 38.450 0.70555 0.51234 −0.73 18.252 15.595 38.05251.6 48.11 0.70603 0.51257 18.568 15.620 38.608 0.2526 0.70557 0.51235 −0.57 18.305 15.607 38.091 0.093651.0 9.13 0.70570 0.51259 18.600 15.633 38.659 0.1589 0.70559 0.51237 −0.13 18.170 15.611 38.147 0.128851.8 0.70581 0.51252 18.634 15.624 38.650 0.70552 0.51231 −1.38 18.265 15.605 38.13156.6 70.25 0.70612 0.51252 18.601 15.627 38.562 0.3592 0.70562 0.51232 −1.26 18.195 15.606 37.995 0.127142.9 15.90 0.70660 0.51256 18.546 15.623 38.506 0.1833 0.70579 0.51235 −0.59 18.339 15.613 38.076 0.13070.6 0.70672 0.51256 0.70567 0.51234 −0.730.6 0.70661 0.51252 0.70586 0.51231 −1.430.6 0.70746 0.51247 0.70658 0.51227 −2.21

0.70725 0.51252 19.304 15.641 39.135 0.707220.9 0.70757 0.51250 0.70749 0.51229 −1.77

0.70730 0.51252 0.707251.2 0.70748 0.51244 0.70743 0.51224 −2.770.7 0.70739 0.51253 0.70710 0.51232 −1.200.7 0.70719 0.51255 0.70713 0.51233 −1.07

53.4 0.70672 0.51250 18.976 15.659 39.074 0.70667 0.51229 −1.75 18.545 15.637 38.538


need to take into account open-system processes for the magmaticdifferentiation.

8.2. Open system magmatic differentiation

Though most SWE-CAMP rocks cluster in a tight portion of the Sr–Nd isotopic space, those having the lowest 143Nd/144Nd and the highest87Sr/86Sr isotopic ratios (which we will refer to as “enriched” isotopiccompositions; see Fig. 5a) might reflect some crustal assimilation(cf. Cebriá et al., 2003). Notably, some of the most differentiated rocks(i.e. high SiO2, low MgO) show the most enriched Sr–Nd isotopic

compositions (Fig. 6), suggesting that they underwent combined crustalcontamination and fractional crystallization. Since high-Sr dykes do notshare this feature (i.e., they are SiO2-poor and 87Sr/86Sr-rich), they mayhave assimilated a different crustal component, and their contamina-tion process is therefore modeled separately (Fig. 9). Modeling of thecrustal contamination process for most SWE-CAMP rocks thus calls foran overview of the Iberian crust (Fig. 8a) to assess the plausible compo-sition(s) of crustal contaminant(s), although, since the tholeiites cropout in a very large region, it is impossible to define a single contaminantvalid for all of them. We focused our attention on the Central Iberianzone that has been studied in great detail (e.g., Bea et al., 1999; Nägler

Table 2Re–Os concentrations and isotopic data of SWE-CAMP rocks.

Sample Group Re PPt Os ppt 187Re/188Os 2σ uncertainty187Re/188Os

187Os/188Os 2σ uncertainty187Os/188Os

Initial 187Os/188Os(198 Ma)

2σ uncertainty187Os/188Os

OF 2 Pyrenean sill 658 39 84.22 2.53 0.40493 0.00078 0.1266 0.009OF 5 Pyrenean sill 612 56 53.38 1.60 0.31151 0.00089 0.1351 0.006OF 34 Pyrenean sill 489 72 32.98 0.99 0.23754 0.00057 0.1286 0.004MD 13 Messejana dyke 213 48 21.60 0.65 0.20821 0.00054 0.1368 0.002MD18 Messejana dyke 678 88 37.72 1.13 0.24777 0.00053 0.1231 0.004MD 18b Messejana dyke 639 144 21.53 0.65 0.19900 0.00057 0.1278 0.003MD 33 Coastal dyke 610 41 73.99 2.22 0.38223 0.00185 0.1377 0.009MD 37 Coastal dyke 532 172 14.94 0.45 0.17919 0.00080 0.1298 0.002P 32 Portuguese lava flow 326 99 15.90 0.48 0.18326 0.00047 0.1307 0.002AL 20 Portuguese lava flow 356 188 9.13 0.27 0.15892 0.00051 0.1288 0.001AL 22 Portuguese lava flow 537 38 70.25 2.11 0.35919 0.00052 0.1271 0.007

Data not included in isochron calculationMD 42 Messejana dyke 606 32 95.24 2.86 0.39234 0.00051 0.0777 0.010E 20 Messejana dyke 1698 75 110.82 3.32 0.28130 0.00114 −0.0849 0.012AL 19 Portuguese lava flow 396 40 48.11 1.44 0.25259 0.00065 0.0936 0.005


et al., 1995; Ugidos et al., 1997; Villaseca et al., 1998, 1999, 2007, 2009,2012), being aware that a modeling involving these lithologies mightnot be applicable to the Algarve lava flows, which were emplaced on acrustal block with a different geological history (e.g., it was notaccreted to Iberia before the Upper Paleozoic) and thus with a possiblydifferent isotopic composition.

The upper crust of the Central Iberian and the Catabrian Zones ofthe Iberian massif is composed mainly (ca. 70% of the outcrops;Villaseca et al., 1999) of post-kinematic S- and I-typeVariscan granitoids

a

c

NHRL

Pyrenean sills

Messejana dyke

Coastal dykes


High-Sr dykes

Fig. 5. Isotopic diagrams showing combined initial (200 Ma) Sr–Nd–Pb data for the SWE-CAMPSr and Nd isotopic data and measured Pb data (Cebriá et al., 2003) for the Pyrenean sills and tPb isotopic compositions of the Northern Hemisphere Reference Line (NHRL) are from Hart (1

(Barbero et al., 1995; Bea et al., 1999; Moreno-Ventas et al., 1995; Pérez-Soba andVillaseca, 2010; Pinarelli andRottura, 1995; Villaseca et al., 1998,2009) intruded into Cambrian metaigneous upper crust (orthogneisses;Bea et al., 1999; Villaseca et al., 1998) and in Precambrian–early Cambrianmetasedimentary rocks (Nägler et al., 1995; Ugidos et al., 1997).The Variscan granitoids have low Sr (ca. 40 ppm) and Eu contents(e.g. Pérez-Soba and Villaseca, 2010), show variable but generallyenriched 87Sr/86Sr200 Ma and 143Nd/144Nd200 Ma compositions and areinterpreted as partial melts of metaigneous lower crustal sources

NHRL

b

d

NHRL

tholeiites. Previously published (Alibert, 1985; Cebriá et al., 2003) age-corrected (200 Ma)he Messejana dyke are plotted along with our data.984).

Pyrenean sills

Messejana dyke

Coastal dykes


High-Sr dykes

a b

Fig. 6. Initial 87Sr/86Sr (a) and 143Nd/144Nd (b) values for the SWE-CAMP samples plotted as a function of MgO.


(Pérez-Soba and Villaseca, 2010). Unmetamorphosed pelites andmetapelitic shales show rather enriched Sr–Nd isotopic signatures andstrongly variable Pb isotopic ratios (Nägler et al., 1995; Ugidos et al.,1997). Felsic intrusive rocks are bordered by (volumetrically insignifi-cant) occurrences of mafic and intermediate igneous bodies (gabbro-tonalites and camptonitic lamprophyres; Bea et al., 1999; Orejana et al.,2006; 2009; Villaseca et al., 1998), showing a much more restricted andless enriched range of isotopic compositions than the local felsic uppercrust. Frequent xenoliths sampled by alkaline magmas (camptoniticlamprophyres) document an Iberian lower crust dominated byfelsic peraluminous with rarer metapelitic and charnockitic granulites(Villaseca et al., 1999). These lower crustal rocks are enriched in Sr(ca. 260 ppm; Villaseca et al., 1999), but are less isotopically enrichedthan the local upper crust. Compositions of the six contaminants usedfor the modeling (i.e., two granitoids, two orthogneisses, one pelite andone felsic granulite from the lower crust) are reported in Table 3 andshown in Fig. 8a–b and are considered to represent the extreme poles ofthe compositional spectrum of the local crust.

An EC-AFC (Energy Constrained Assimilation and FractionalCrystallization; DePaolo, 1981) modeling (not shown). For the uncon-taminated starting composition, we used a relatively low 87Sr/86Sr200 Ma value (close to the least radiogenic sample), towards whichthe SWE-CAMP samples tend to converge (0.7050; Sr 150 ppm; cf.also data from Cebriá et al., 2003; Fig. 5), coupled with a rather

Table 3Sr–Nd–Pb trace element and isotopic compositions selected as crustal contaminants and mant

Reservoir 87Sr/86Sr200 Ma [Sr] ppm 143Nd/144Nd200 Ma [N

EC AFC modelingStarting composition 0.7050 150 0.51243 6Granitoids min 0.7122 102 0.51224 43Granitoids max 0.7258 84 0.51198 36Orthogneiss min 0.7105 176 0.51220 48Orthogneiss max 0.7308 107 0.51199 25Pelite 0.7266 140 0.51170 39Lower crust (felsic granulite) 0.7139 258 0.51192 31

Mantle source modelingDMM 0.7020 9.80 0.51300 0CAP 0.7030 21.10 0.51290 1SCLM 0.7028 11.30 0.51280 1EMI 0.7049 12.14 0.51202 1EMII 0.7078 12.34 0.51188 1LC (granulitic) 0.7090 220 0.51204 30LC (metapelitic) 0.7130 150 0.51170 40LC (granulitic min) 0.7075 440 0.51190 27pelite min 0.7218 129 0.51180 32

radiogenic Nd isotopic signature (143Nd/144Nd200 Ma 0.51243; εNd0.97;Nd6.5 ppm). This Sr–Nd composition is reasonably representativeof primitive CAMP magma since low-Ti basalts from the entire CAMPseem to converge towards these values (cf. Alibert 1985; Cebriá et al.,2003; Deckart et al., 2005; De Min et al., 2003; Verati et al., 2007), in-cluding low 187Os/188Os CAMP basalts, which are likely to have experi-enced little contamination (Callegaro et al., 2013; Merle et al.,accepted for publication). The Pb isotopic signatures of the startingcomposition were selected to resemble the less radiogenic end ofSWE-CAMP and typical low-Ti CAMP trends (206Pb/204Pb200 Ma 18.2;207Pb/204Pb200 Ma 15.6; 208Pb/204Pb200 Ma 38.0; Pb 2 ppm; e.g. Merleet al., 2011 and references therein).

The 187Os/188Os199 Ma composition (ca. 0.1298 ± 0.0056; Fig. 7b) ofthe SWE-CAMP basalts constrains the maximum contribution of crustalassimilation during the evolution of these magmas. Depending on theOs isotopic composition of themantle source, the amount of crustal con-tribution to the SWE-CAMP basalts ranges from near-zero to a maxi-mum of ca. 7%. Such maximum value is calculated considering aprimarymagma generated by amantle sourcewith the lowest plausible187Os/188Os199 Ma ratio (ca. 0.1243), which is compatible with an off-craton sub continental lithospheric mantle (SCLM; Carlson, 2005). Ifthis parental magma with ca. 250 ppt Os (a conservatively high esti-mate of the Os content in the most primitive CAMP basalts; Callegaroet al., 2013; Merle et al., accepted for publication) was contaminated

le source end-members.

d] ppm 206Pb/204Pb200 Ma207Pb/204Pb200 Ma

208Pb/204Pb200 Ma [Pb] ppm

.5 18.20 15.60 38.00 218.13 15.65 38.61 28

.3 18.10 15.62 38.39 4918.00 15.62 38.00 2018.30 15.72 38.50 20

.70 18.75 15.75 39.12 35

.10 18.40 15.63 38.50 30

.7 17.93 15.47 36.97

.4 19.90 15.65 39.30

.2

.7 17.02 15.44 37.98

.5 18.85 15.68 39.31

18.40 15.63 38.5018.75 15.75 39.12

aPyrenean sills

Messejana dyke

Coastal dykes


b

Age = 198.6+8.1 MaInitial 187Os/188Os = 0.1298+0.0056

11/13 samples - MSWD = 7.7

0 20 40 60 80 100 120 0 50 100 150 2000.1

0.2

0.3

0.4

0.5 0.150

0.145

0.140

0.135

0.130

0.125

0.120

0.115

0.110

187Re/188Os

187 O

s/18

8 Os

199M

a

[Os] ppt

187 O

s/18

8 Os

data-point error bars are 2o

PUM

Fig. 7. (a) In a Re/Os vs Os/Os diagram 11 SWE-CAMP rocks (over 13 analyzed samples) form an errorchron. Age and uncertainties were calculated with Isoplot (Ludwig, 2003). (b) InitialOs ratios plotted as a function of Os contents (ppt). PUM composition (primitive upper mantle; recalculated to 199 Ma) is after Meisel et al. (2001). Uncertainties (2σ) and include in-runerrors, blank and weighing uncertainties. Uncertainties on 187Os/188Os199 Ma range between 0.001 and 0.009 (error bars smaller than symbols when below 0.003).


by high 187Os/188Os continental crust (Os 31 ± 9 ppt; 187Os/188Osca.1.04, Peucker-Ehrenbrik and Jahn, 2001) in a non-replenishedmagma chamber, it would readily (i.e., at N7% crustal assimilation)reach 187Os/188Os values N 0.14, higher than those observed in SWE-CAMP samples. Also, high Nd/Th ratios (7–10), typical of mantle-derived rocks (monazite-bearing crustal rocks have Nd/Th ca. at 2.6;e.g., Bea et al., 1999) argue against important crustal hybridization ofthese magmas during emplacement, instead suggesting an enrichedcharacter of the mantle source. We therefore modeled EC-AFC pathsfor the above mentioned crustal compositions to check if up to 7% as-similation of any of them can reproduce Sr–Nd–Pb isotopic variationsof SWE-CAMP.

For such low (≤7%) extents of crustal assimilation, none of theconsidered crustal rocks reproduces the observed Sr–Nd–Pb isotopiccompositions of SWE-CAMP successfully. Contamination trends calcu-lated for the less isotopically enriched granitoid and orthogneisswould encompass the entire range of data for the SWE-CAMP inSr–Nd isotopic space (explaining trends towards more radiogenic Sr),but only if 25–30% of contamination is considered. Such high extentsof contamination are not only incompatible with the Os isotope con-straints, but are also inconsistent with major element compositions.At ≤ 7% Os-constrained contamination, the shifts of Sr–Nd–Pb isotopiccompositions would be smaller than the observed variation for SWE-CAMP basalts. On the contrary, assimilation of up to 7% of the isotopical-ly most enriched upper crust (the orthogneiss max and granitoids maxof Fig. 8) has a significant effect on Nd isotopic compositions and canexplain the evolution of SWE-CAMP basalts towards low 143Nd/144Nd.Yet fails to reproduce the trend to high 87Sr/86Sr at relatively high143Nd/144Nd displayed by other SWE-CAMP rocks, and would drive Pbisotopic compositions (207Pb/204Pb, in particular) towards values moreradiogenic than those observed for the studied rocks. Similar observa-tions are valid for assimilation of lower crustal felsic granulites.

Hence, the following observations arise from the EC-AFC modeling:a) we cannot reproduce the Sr–Nd–Pb isotopic characteristics ofSWE-CAMP samples through b7% assimilation of the most plausibleupper and lower crustal lithologies; b) the effects produced by crustalassimilation of a certain lithology in the Sr–Nd isotopic space are oftenopposite or at best negligible in the Pb–Pb isotopic systems; c) sinceSWE-CAMP rocks crop out over a large area, we cannot define a singlecontaminant composition, due to the presence of very different litholo-gies in the crust of the Iberian peninsula.

Notably, while emplaced within areas of very different crustal com-positions, all the SWE-CAMP rocks share a rather uniform and slightly

enriched isotopic flavor. Such a common character is unlikely to be con-trolled and imparted by assimilation of various amounts of differentcrustal lithologies. We can therefore ascribe the minor isotopic hetero-geneities within the SWE-CAMP tholeiites to crustal assimilation, butthis process is probably not the fundamental conveyor of the generallyenriched isotopic character of these rocks.

8.2.1. The high-Sr dykesA different approach was followed to model crustal contamination

for the high-Sr dykes from Algarve and Santiago do Cacém basins. Thecomposition of the metaigneous amphibolite xenolith from AL6(Sr 1032 ppm; 87Sr/86Sr200 Ma ~ 0.70735, as extrapolated from the re-gression of all the high-Sr samples; further details in the Supplementarymaterial) or local carbonate rocks of the Carboniferous BordeiraFormation (Sr 2000–9000 ppm; 87Sr/86Sr200 Ma ~ 0.70771; Martinsand Kerrich, 1998; Martins et al., 2008) were considered as plausiblecontaminants.

The fine grain size of hypocrystalline high-Sr rocks (supportive ofrapid cooling) and the abrupt changes in mineral chemistry and theanomalous trends shown in whole-rock SiO2, CaO, Na2O and Sr vs.MgO (Fig. 3) rule out a progressive contamination process during frac-tional crystallization, and rather suggest an abrupt assimilation event,occurring shortly before emplacement of the magmas. Therefore, wecalculated a simple mixing curve (Fig. 9) between a Ca- (21.9 wt.%)and Sr-rich (ca. 1000 ppm) crustal rock, similar in composition to thexenolith, and a typical SWE-CAMP tholeiitic basalt from the Algarvebasin (lava flow AL10 was chosen). While compositions of someSr-rich rocks (e.g. P40 or similar samples from Martins and Kerrich,1998) are reached through conspicuous but plausible (20–30%)addition of this crustal contaminant, those of the most extremelySr-rich basalts would require an unreasonable (40 to 100%) amount ofassimilation, and in some cases (e.g. PIM2) they are never reached.Therefore, we argue against assimilation of this lithology, and adhereto the model of Martins et al. (2008). According to these authors,up to 20% contribution of the carbonate lithology (Sr 6200 ppm;87Sr/86Sr200 Ma 0.70771; Visean shallow water carbonates of theBordeira Formation cropping out in SW Portugal; Manuppella et al.,1992) is sufficient to explain the Sr and Ca enrichment trends of allour Sr-rich basalts (Fig. 9). This model also offers an explanation forthe very low SiO2 contents (Fig. 3) and for the high δ18O (up to8.73‰; Martins et al., 2008) characterizing these high-Sr dykes. Supportfor this model comes from Jolis et al., (2013), who recently demonstrat-ed experimentally that a basalt can easily and very rapidly assimilate

206Pb/204Pb 200Ma

206Pb/204Pb 200Ma

87Sr/86Sr 200Ma

207 P

b/20

4 Pb

200M

a

c

b

d

Pyrenean sills

Messejana dyke

Coastal dykes


High-Sr dykes

SWE-CAMP

granitoids min

granitoids max

orthogneiss max

orthogneiss min

lower crust

pelite

granitoids min

orthogneiss min

orthogneiss max

orthogneiss max

lower crust

pelite

143 N

d/14

4 Nd

200M

a

207 P

b/20

4 Pb

200M

a

SWE-CAMP

87Sr/86Sr 200Ma

143 N

d/14

4 Nd

200M

a

a

SWE-CAMPMafic and alkaline rocksS- and I-type Granitoids Orthogneiss and paragneissMetapelitesMigmatitesLower crust

EC AFC paths

Iberian crust

4%

4%

30%

26%

6%

12%

20%

4%

2%

6%

4%

4%

4%8%

NHRL

Geo

chro

n

Fig. 8. Previously published (Bea et al., 1999; Nägler et al., 1995; Orejana et al., 2009; Ugidos et al., 1997; Villaseca et al., 1998, 1999, 2009) Sr–Nd–Pb isotopic compositions of differentlithologies composing the Iberian (upper and lower) crust are reported on 87Sr/86Sr vs. 143Nd/144Nd (a) and 206Pb/204Pb200 Ma vs. 207Pb/204Pb200 Ma (b) diagrams. Circled data pointswere used as contaminants. Evolution paths calculated with EC-AFC modeling (Spera and Bohrson, 2001) for various compositions of the Iberian continental crust are plotted alongwith isotopic data of the SWE-CAMP tholeiites on 87Sr/86Sr200 Ma vs. 143Nd/144Nd200 Ma (c) and 206Pb/204Pb200 Ma vs. 207Pb/204Pb200 Ma isotopic spaces (d). Tickmarks every 2% assimilation(Ma* parameter of Spera and Bohrson, 2001). Compositions of the contaminants are listed in Table 3.


carbonate lithologies, and that this melt–carbonate interaction canexplain CaO enrichment and SiO2 depletion in the melt, as well as vari-ations in Sr isotopic signatures. Also, they stress the importance ofphysicalmixing andmingling as driven by volatiles exsolved during car-bonate assimilation. Therefore, this process not only clarifies the chem-ical variations observed in Portuguese CAMP high-Sr dykes, but it alsoserves to explain their peculiar petrographic features (glass-rich,hypocrystalline, highly vesicular), typical of volatile-rich magmas.

8.3. Insights on the SWE-CAMP mantle source

From the above discussion itwas concluded that (with the exceptionof the high-Sr dykes) crustal contamination was in most cases negligi-ble, with the more radiogenic Sr compositions probably reflectingless than 7% of crustal assimilation. Taking these factors into accountit may be considered that the enriched Sr–Nd–Pb isotopic and incom-patible element signatures of SWE-CAMP basalts are indicative of along-term enriched mantle source. In terms of Sr–Nd–Pb isotopic com-positions, SWE-CAMP tholeiites are rather enriched (Fig. 10), plottingfar from a depleted mantle source (DMM; Salters and Stracke, 2004)but rather close to the EMII pole (Willbold and Stracke, 2006; Zindlerand Hart, 1986), whereas the initial Os isotopic composition is similarto the primitive upper mantle187Os/188Os ratio (currently 0.1296;

~0.1282 at 200 Ma) estimated by Meisel et al. (2001). If compositionsfor the DMM and the EMII are recalculated at 200 Ma, a ~ 1:1 mixtureof these two sources matches perfectly the SWE-CAMP Sr–Nd isotopiccompositions. On the other hand, Pb isotopic compositions would de-mand a ca. 90% DMM and 10% EMII source mixture. Given thisdecoupling, different end members for the mantle composition mustbe considered. Previous authors investigating the SWE-CAMP mantlesource (Alibert, 1985; Cebrià et al., 2003; Martins et al., 2008) proposeda lithospheric origin for this part of the province, suggesting decompres-sion during passive rifting as the cause of mantle melting, followed byan increasing asthenospheric contribution. In this sense,metasomatizedportions of the SWE lithospheric mantle can represent the enrichedend-member required by our isotopic data. The presence of theseenriched components in the SWE lithospheric mantle is testified bya) alkaline magmatism preceding SWE-CAMP emplacement (cf. Beaet al., 1999) and b) enriched geochemistry of lithospheric mantle xeno-liths recorded from Western Europe (Bianchini et al., 2007, 2011;Chazot, 2005). Furthermore, mantle xenoliths record several episodesof enrichment and depletion in the geodynamic history of the litho-spheric mantle under the Iberian area (Bianchini et al., 2007). Mantlexenoliths and peridotitic massifs from Western Europe (see Fig. 4 inthe Supplementary material) show isotopic signatures ranging fromdepleted to enriched, which overlap the most primitive signatures

0.5130

0.5128

0.5126

0.5124

0.5122

0.5120

143 N

d/14

4 Nd 20

0Ma

87Sr/86Sr200Ma

206Pb/204Pb200Ma

207 P

b/20

4 Pb 20

0Ma

0.7020 0.7040 0.7060 0.7080

17.70 18.20 18.70 19.20 19.70

15.70

15.60

15.50

15.45

15.40NHRL

Geo

chro

n

a

b

0.5132

Pitcairn

Society

Samoa

Canary

EMI

Cape Verde

Society

Samoa

Pitcairn

15.55

15.65

EMII

DMM

Cape Verde

Canary

EMI

Pyrenean sills

Messejana dyke

Coastal dykes


High-Sr dykes

low-Ti CAMP

high-Ti CAMP

Tristan

EMII

DMM

Tristan

high-Ti CAMP

low-Ti CAMP

Fig. 10. Sr–Nd (a) and Pb–Pb (b) isotopic diagrams showing combined initial (200 Ma)Sr–Nd–Pb data for the SWE-CAMP tholeiites, plotted along with several OIBs representa-tive of Common Component (C; Hanan and Graham, 1996; Canary, Cape Verde), EMI(Tristan da Cunha, Pitcairn) or EMII-type (Society, Samoa) mantle poles (Gibson et al.,2005; Hofmann, 2003; Willbold and Stracke, 2010; Zindler and Hart, 1986). Low-Ti andhigh-Ti CAMP fields are taken from Merle et al. (2011) and references therein. DMM,EMI and EMII mantle poles age-corrected to 200 Ma.

High-Sr dykes (this study)

High-Sr dykes (previous work)

Portuguese lava flows (this study)

Mixing lines

CAMP basalts

CAMP magma+

amphibolitic xenolith

CAMP magma+

carbonate

20%

20%

40%

Carbonate

Amphiboliticxenolith

Fig. 9. 1000/Sr versus 87Sr/86Sr200 Ma diagram representing a mixing model for the crustalcontamination of high-Sr dykes (white circles) with either an amphibolitic (black line) ora carbonatic lithology (dashed line), starting from the composition of an uncontaminatedPortuguese lava flow (AL10; gray circles). Tick marks on the mixing lines are every 10%assimilation. Diamonds: compositions of high-Sr dykes fromMartins and Kerrich (1998).


of SWE-CAMP tholeiites. Sm–Nd and Re–Os model ages for Ronda(southern Spain) or the Pyrenean peridotites (Reisberg and Lorand,1995) range between 2.3 and 1.3 Ga (Early–Middle Proterozoic) andare concordant with Nd model ages (2.2–1.2 Ga) of crustal rocks fromFrance and Iberia (Nägler et al., 1995; Villaseca et al., 1998), suggestingthe preservation in the lithospheric mantle under Iberia of domainsolder than 2 Ga. Such long-term persistence of lithospheric mantle do-mains allowed subduction-related processes (melt or fluid metasoma-tism) during various pulses of Precambrian and Paleozoic continentalaccretion (Bianchini et al., 2011; Gonzalez-Jimenez et al., 2013) to im-print the SWE lithosphericmantle. The high Rb/Sr and lowSm/Nd ratiosleft by these events produced enriched isotopic signatures by radiogenicingrowth through time. Studies on the late Paleozoic alkaline and gab-broic rocks from Spain (Orejana et al., 2009) suggested that the mantleunder the Iberia peninsula was enriched by recycled pelitic sedimentsand lower crust, which may have imparted an EMII-like signature

The Central Iberian Zone has experienced several episodes ofPaleozoic, mainly post-collisional basic magmatism (Bea et al., 1999;Moreno-Ventas et al., 1995; Orejana et al., 2006, 2009), evolving fromcalc-alkaline (gabbroic intrusions), to alkaline (lamprophyric andcamptonitic dykes), to tholeiitic (SWE-CAMP dykes, sills and flows).The most primitive of these rocks (in particular the Variscan gabbroicintrusions from the Spanish Central System; Orejana et al., 2009)show incompatible element patterns (Nb–Ta–P–(Ti) troughs, Pb peak)and Sr–Nd–Pb enriched isotopic characteristics remarkably similar tothose of the SWE CAMP, arguing in favor of a common source forthese magmatic suites. Orejana et al. (2009) showed that a mixingmodel involving 1–2% of local felsic and metapelitic granulitic lowercrust (Orejana et al., 2006; Villaseca et al., 1999, 2007) recycled withina depleted mantle (143Nd/144Nd305 Ma 0.5128; 87Sr/86Sr305 Ma 0.7028)fits well with the isotopic variation of the late Variscan gabbros, whichencompasses that of our SWE-CAMP tholeiites (cf. Fig. 8).

We modeled the isotopic composition of a ternary mixed source,using several combinations of one depleted and two enriched end-members (compositions are listed in Table 3). Of the various combina-tions, we report in Fig. 11 the two that best fit the observed isotopicvariations for SWE-CAMP samples. If the compositions reported inOrejana et al. (2009) are considered (i.e. SCLM, metapelitic and granu-litic lower crust; Fig. 11a), a 3–7% addition of recycled crustal materialto the upper mantle drives its compositions towards those of

SWE-CAMP rocks. We then modeled the mixing between a depletedend-member with the chemical composition of the DMM (Fig. 11b),enriched by felsic granulites (representing recycled lower crustalmaterial) and pelites (representing recycled sediments from theupper crust). In this second case, 2.5–5% contribution of these crustalmaterials in the mantle source would be sufficient to explain the ob-served enrichment. It is important to stress that our recycled peliticcomponent is representing the average composition of the local uppercrust. Moreover, we consider the depleted component to bemost likelythe upper asthenospheric mantle, which underwent decompressionsince the end of the Variscan orogeny, rather than the colder andmore refractory SCLM. Pb isotopes are not useful in constraining this

a

b

2%

3%

4%

5%

3%

4%

5%

7% Pyrenean sills

Messejana dyke

Coastal dykes


High-Sr dykes

SCLM

GRANULITIC LC

PELITIC LC

DMM

LCUC

META

Fig. 11. Sr–Nd isotopic plots of pseudo-binary mixing calculations between (a) SCLM,lower crustal granulites and metapelites (after Orejana et al., 2009) and (b) DMM, lowercrust and upper crust. Black lines mark mixing paths, with tick marks every 10% stepwhere not elsewise indicated. Input data in Table 3. All the end-members are recalculatedto 200 Ma.


mixing model, since a) the triplets of compositions all plot in a line inPb–Pb diagrams, and b) Pb isotopes are totally controlled by theenriched component(s), due to their Pb contents, which are threeorders of magnitude higher than that of the depleted reservoir (ca. 30vs. 0.02 ppm), and c) as already pointed out in Section 8.2, Pb isotopicsignatures of the Iberian crustal lithologies broadly overlap those ofSWE-CAMP basalts. Following all these observations, we argue infavor of a dominantly DMM-like source enriched by a 3–5% mixedlower and upper crustal components introduced into the upper mantleduring subduction-related recycling episodes. This model was also ap-plied to study the mantle source of the Eastern North American CAMPdykes and flows (Callegaro et al., 2013; Merle et al., accepted forpublication). We note that this process would have little effect on theOs isotopic composition of the source, because of the very large contrastin Os concentrations between mantle peridotites (~3000–5000 ppb)and most crustal rocks (~10–50 ppt).

8.4. Mantle plume contribution?

Sr–Nd–Pb signatures of SWE-CAMPbasalts trend towards EMII com-positions, which could be interpreted as evidence for the involvement ofa deep mantle plume. However, several observations argue against thishypothesis. As already observed, mixing models between a depletedmantle source and an EMII-flavored end-member show discrepanciesbetween Sr–Nd and Pb isotopic systems. Furthermore, with the soleexception of Fernando de Noronha (Gerlach et al., 1987), OIBs fromthe Atlantic generally do not show EMII-type isotopic signatures. More-over, as argued by Martins et al. (2008), a mantle plume impingingunder thepost-Variscan thick Iberian lithospherewould undergo partialmelting at high pressures, within the garnet stability field.

Cebriá et al. (2003) proposed a different composition for a hypothet-ical mantle plume acting as the source for SWE-CAMP basalts, namelythe CAP (Central Atlantic Plume). Assuming that channeled CAPmateri-al could be at the origin of the Low Velocity Component or the Astheno-spheric Reservoir under Europe (Hoernle et al., 1995), they used theisotopic compositions (age-corrected to 200 Ma) of melts (Cenozoicundersaturated alkali basalts) issued from these reservoirs as proxiesfor CAP chemistry, coupled with trace element concentrations ofplume-related tholeiites from Loihi Seamount (Hawaii). If we use thesuggested CAP composition (143Nd/144Nd 0.5129; Nd 15.5 ppm;87Sr/86Sr 0.730; Sr 300 ppm) to model the ternary mixing ofSWE-CAMPmantle source along with the previously mentioned valuesof granulitic lower crust and pelitic upper crust, 40% contribution ofenriched domains is required to sufficiently enrich the mantle source.However, if less extreme trace element contents (Nd 1.04 ppm; Sr21.1 ppm, corresponding to the primitive mantle; McDonough andSun, 1995) are used for the ternarymixing, Sr andNd isotopic character-istics of the SWE-CAMP rocks are reproduced by a mantle sourcedominated by the CAP and enriched by 4–7% of mixed lower andupper crustal material. Nevertheless, in both cases, Pb isotopic ratiosof the CAP component are very high (206Pb/204Pb 19.90; 207Pb/204Pb15.65; 208Pb/204Pb 39.30) and would convey highly radiogenic Pbisotopic signatures, which are unobserved in SWE-CAMP tholeiites.We therefore chemically rule out a significant involvement of deepmantle-plume material for SWE-CAMP basalts. Nevertheless, even if achemical contribution from amantle plume can be excluded, plume im-pingement may provide the heat input needed for melting (e.g. Hill,1991; Wilson, 1997). Alternatively, several non-plume scenarios,including thermal heat incubation (Coltice et al., 2007), edge-drivenconvection (McHone et al., 2005), and linear delamination alongmajor shear zones (Liégeois et al., 2013), could explain why the mantleunderwent the high degrees of melting required to originate the CAMPbasalts.

9. Correlations with other CAMP occurrences

Correlating CAMP rocks fromdifferent areas is of paramount interestfor outlining the timing and the modalities of emplacement of thismagmatic province. Due to the uncertainties of the 40Ar/39Ar datingmethod, it is impossible to rely solely on absolute ages for this goal.Thus it is necessary to consider geochemical correlations as well. In par-ticular, previous studies on lava piles from Moroccan basins (Bertrandet al., 1982; Marzoli et al., 2004) highlighted a temporal evolution ofmajor and trace elements in CAMP basalts, which, supported bymagnetostratigraphy allowed their correlation with Eastern NorthAmerican CAMP occurrences (e.g. Marzoli et al., 2011; Tollo andGottfried, 1992). Marzoli et al. (2004) observed systematic time-related major element variations for the Central High Atlas Moroccanbasalts from the SiO2- and TiO2-rich lower basalts, to the SiO2- andTiO2-poor upper basalts, whereas the last flows (recurrent basalts)are characterized by the highest TiO2 and by low SiO2 contents. Themajor element evolution of the Moroccan basalts is accompanied by aprogressive decrease of incompatible element contents (e.g., Rb, Ba, Sr,


Nb) and of the LREE/HREE ratio from the lower to the recurrentbasalts. Similar major and trace element variations are observedfor the North American CAMP flows and, combined with bio- andmagnetostratigraphic correlations suggest a slightly earlier onset ofCAMP volcanism in Morocco and synchrony between the MoroccanIntermediate flows with the U.S.A. lower (Orange Mt.) flows (Marzoliet al., 2011).

Such correlations can be applied also to the SWE CAMP rocks(Fig. 12). In particular, the TiO2 vs. La/Yb compositions of the vastmajor-ity of the SWE CAMP rocks (TiO2 0.5–1.2 wt.%; La/Yb 3.4–4.6) overlapthose of basaltic flows from the Moroccan Upper Unit (Marzoli et al.,2004) and in part those from the Orange Mt. basalt group (Marzoliet al., 2011) from Eastern North America. A few samples (3 from theMessejana Dyke, 3 from Portuguese lava flows and 1 from a Pyreneansill) extend towards higher TiO2 (1.0–1.3 wt.%) and La/Yb (4.5–5.3)compositions, typical of the Moroccan Intermediate Unit (and of sparseOrange Mt. samples). Notably, SWE CAMP rocks are lacking the high-TiO2, high-La/Yb compositions of the first Moroccan CAMP occurrences(Lower Unit), as well as the high-TiO2, low-La/Yb signature of the latestages of Moroccan and U.S.A. CAMP magmatism (Recurrent Unit andHookMt. basalt, respectively). Also, Sr–Nd–Pb (andOs) isotopic compo-sitions of the SWE-CAMP tholeiites totally overlap, and are thereforegeochemically correlated, with those of Eastern North American CAMPlava flows (cf. Merle et al., accepted for publication). Hence, CAMPmagmatism in the SWE area appears to have occurred during themain magmatic pulse of the province and appears to be issued from amantle source similar to that of that of the Moroccan Intermediate-Upper basalts and North American Orange Mt. basalts.

10. Conclusions

Southwestern Europe, from the Pyrenees to the Atlantic coast is con-stellated by remnants of the Central Atlantic magmatic province, in theformof sills (Pyrenean area), dykes (in south-western Iberia), lavaflowsand pyroclastic deposits (Southern Portugal). All the tholeiites sharesimilar major and trace element compositions, with the exception ofthe few CaO-enriched and SiO2-depleted high-Sr dykes from SouthernPortugal. Major element variations are supportive of an evolutionof these magmas controlled by closed system fractional crystallizationof up to 20% of the typical mineralogical assemblage of CFBs (olivine +plagioclase + augite ± pigeonite + Fe–Ti oxides) at upper crustal

Recurrent Unit

Hook Mt. Group

Orange Mt. Group

Upper Unit

Intermediate Unit

Lower Unit

Preakness Group

Pyrenean sills

Messejana dyke

Coastal dykes


High-Sr dykes

Fig. 12. TiO2 (wt.%) vs. La/Yb of the SWE-CAMP rocks are plotted alongwith compositionalfields for previously published data for other areas of the CAMP for comparison. Datasources: Marzoli et al. (2004) and Merle et al. (2013).

pressures (based on clinopyroxene geobarometry). Incompatibleelements show positive Pb anomalies and negative Nb–Ta spikeswhich are typically interpreted as crustal signatures. All the samplesshare similar and slightly enriched Sr–Nd–Pb isotopic characteristics(87Sr/86Sr200 Ma 0.70529–0.70657; 143Nd/144Nd200 Ma 0.51238–0.51225;206Pb/204Pb200 Ma 18.15–18.48; 207Pb/204Pb200 Ma 15.57–15.68;208Pb/204Pb200 Ma 37.99–38.52), along with rather unradiogenic187Os/188Os199 Ma (0.123–0.138). This geochemical similarity isconsistent with the possibility of the Messejana–Plasencia dykerepresenting a feeder structure for the Southern Portugal lava flows,though no field evidence for this relationship has ever been observed.The low, mantle-like Os isotopic signatures of the SWE-CAMP rocksconstrain the crustal contamination of these magmas with upper andlower crustal lithologies of the Iberian area (Variscan granitoids, pelites,felsic granulites) to amaximumof 7%. Therefore, the enriched characterand the crustal-like signaturesmust be derived from themantle source.Isotopic similarities with Hercyinian gabbros from the same area sug-gest a common, enriched mantle source for these rocks. We envisage,similarly to what was proposed for other areas of the CAMP (EasternNorth America), a depleted upper mantle (either asthenospheric orlithospheric) as the dominant source of these magmas, enriched by 3to 7% upper and lower crustal domains recycled during the Paleozoicsubduction events that closed the Rheic ocean and assembled thePangaea supercontinent. Crustal recyclingwithin the uppermantle con-veyed the enriched isotopic flavors observed by the SWE-CAMP and thecrustal-like incompatible element patterns, but did not modify theOs isotopic compositions, due to the high Os content of the ambientperidotite compared to that of crustal rocks. The contribution of a man-tle plume (both EMII-flavored or showing CAP composition; Cebriáet al., 2003) as source of thesemagmas is unlikely based on geochemicalarguments.

High-Sr dykes from the Southern Portugal basins underwent aslightly different evolution, controlled, in the late stages of theiremplacement, by the assimilation of (10–20%) carbonate lithologies.In particular, this assimilation had the effect of raising the Sr and CaOcontents along with lowering SiO2 levels in these rocks, and it wasaccompanied by a deviation towards more radiogenic Sr isotopic signa-tures (0.70669–0.70749) at similar Nd isotopic characters (0.51232–0.51224). The only high-Sr samples analyzed for Pb isotopes yieldedradiogenic Pb ratios (18.55; 15.64; 38.54), which is in agreement withthe proposed contamination process.

Correlation with other CAMP occurrences through absolute agescomparison has limitations, since the only available 40Ar/39Ar plateauage for the Messejana dyke (206.5 ± 2.6 Ma) presented here suffersfrom excess Ar, and is thus slightly older than previously publishedage (199.7 ± 1.4 Ma) from coeval lava flows from the Algarve basin(Verati et al., 2007). However, the inverse Ar isochron age on theMessejana–Plasencia dyke sample (203.1 ± 5.6 Ma) and the Re–Os iso-chron age obtained for all the analyzed samples (198.6 ± 8.1 Ma) agreewithin error with previously published results and with the main peakof CAMP activity (ca. 201 Ma; Marzoli et al., 2011). SWE-CAMP can begeochemically correlated (chiefly based on incompatible elementratios) with the magmatic activity of the Moroccan Intermediate andUpper Unit (Marzoli et al., 2004) as well as with the Eastern NorthAmerican Orange Mt. flows (Marzoli et al., 2011). This includes SWE-CAMP magmatism within the main pulse of CAMP activity.

Acknowledgments

This contribution is in memory of our colleague Enzo M. Piccirillowho recently passed away.

G. Féraud (Ar analyses), F. Bussy and R. Carampin (EMP analyses,Lausanne and Padova), D. Fontignie (Sr–Nd–Pb isotopes, Geneva),C. Lavanchy, P. Capiez, F. Capponi (XRF, Lausanne, Lyon and Geneve),C. Douchet and F. Keller (ICP-MS, Lyon, Grenoble) are thanked for ana-lytical assistance. We are grateful to M. Dorais, J.-P. Liégeois and


L. Melluso (Ed.) for careful and constructive reviews that stronglyimproved the manuscript. The Fonds National Suisse (FNRS2000-064580.01) and Fondazione Cariparo (project: MARZECCE09CA.RI.PA.RO. — PI: A. Marzoli) supported this research. Additionalsupport was provided by CNRS (France), CNRI (Italy), CNRST(Morocco) and FCT (Portugal).

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.lithos.2013.10.021.

References

Alibert, C., 1985. A Sr–Nd isotope and REE study of late Triassic dolerites from thePyrenees (France) and the Messejana dyke (Spain and Portugal). Earth and PlanetaryScience Letters 73, 81–90.

Andersen, D.J., Lindsley, D.H., Davidson, P.M., 1993. QUILF: a Pascal program to assessequilibria among Fe–Mg–Ti oxides, pyroxenes, olivine and quartz. Computers andGeosciences 19, 1333–1350.

Arnal, I., Calvet, F., Márquez-Aliaga, A., Solé de Porta, N., 2002. La plataforma carbonatadaepeírica (Formaciones Imón e Isábena) del Triásico superior del Noreste de laPenínsula Ibérica. Acta Geologica Hispánica 37, 299–328.

Azerêdo, A.C., Duarte, L.V., Henriques, M.H., Manuppella, G., 2003. Da dinâmica continen-tal no Triásico aos mares do Jurássico Inferior e Médio. Cadernos de Geologia dePortugal, Instituto Geológico e Mineiro (43 pp.).

Barbero, L., Villaseca, C., Rogers, G., Brown, P.E., 1995. Geochemical and isotopic disequi-librium in crustal melting: an insight from the anatectic granitoids from Toledo,Spain. Journal of Geophysical Research 100, 15745–15765.

Bea, F., Montero, P., Molina, J.F., 1999. Mafic precursors, peraluminous granitoids, and latelamprophyres in the Avila batholith; a model for the generation of Variscan batho-liths in Iberia. Journal of Geology 107, 399–419.

Bertrand, H., 1987. Le magmatisme tholéitique continental de la marge Ibérique,précurseur de l'ouverture de l'Atlantique Central: les dolérites du dyke deMessejana–Plasencia (Portugal–Espagne). Comptes Rendus de l'Académie desSciences 304, 215–220.

Bertrand, H., Dostal, J., Dupuy, C., 1982. Geochemistry of early Mesozoic tholeiites fromMorocco. Earth and Planetary Science Letters 58, 225–239.

Béziat, D., Joron, J.L., Monchoux, P., Treuil, M., Walgenwitz, F., 1991. Geodynamic implica-tions of geochemical data for the Pyrenean ophites (Spain–France). Chemical Geology89, 243–262.

Bianchini, G., Beccaluva, L., Bonadiman, C., Nowell, G., Pearson, G., Siena, F., Wilson, M.,2007. Evidence of diverse depletion and metasomatic events in harzburgite–lherzolite mantle xenoliths from the Iberian plate (Olot, NE Spain): implications forlithosphere accretionary processes. Lithos 94, 25–45.

Bianchini, G., Beccaluva, L., Nowell, G.M., Pearson, D.G., Siena, F., 2011. Mantle xenolithsfrom Tallante (Betic Cordillera): insights into the multi-stage evolution of the southIberian lithosphere. Lithos 124, 308–318.

Callegaro, S., Marzoli, A., Bertrand, H., Chiaradia, M., Reisberg, L., Meyzen, C., Bellieni, G.,Weems, R., Merle, R., 2013. Upper and lower crust recycling in the source of CAMPbasaltic dykes from southeastern North America. Earth and Planetary Science Letters376, 186–199.

Carlson, R.W., Pearson, D.G., James, D.E., 2005. Physical, chemical, and chronologicalcharacteristics of continental mantle. Reviews of Geophysics 43, RG1001.http://dx.doi.org/10.1029/2004RG000156.

Cebriá, J.M., Lopez-Ruiz, J., Doblas, M., Martins, L.T., Munha, J., 2003. Geochemistry of theEarly Jurassic Messejana–Plasencia dyke (Portugal–Spain); implications on the originof the Central Atlantic Magmatic Province. Journal of Petrology 44, 547–568.

Chauvel, C., Marini, J.C., Plank, T., Ludden, J.N., 2009. Hf–Nd input flux in the Izu–Marianasubduction zone and recycling of sub ducted material in the mantle. Geochemistry,Geophysics, Geosystems 10. http://dx.doi.org/10.1029/2008GC002101.

Chazot, G., Charpentier, S., Kornprobst, J., Vannucci, R., Luais, B., 2005. Lithospheric mantleevolution during continental break-up: the west Iberia non-volcanic passive margin.Journal of Petrology 46, 2527–2568.

Coltice, N., Phillips, B.R., Bertrand, H., Ricard, Y., Rey, P., 2007. Global warming of themantle at the origin of flood basalts over supercontinents. Geology 35, 391–394.

De Min, A., Piccirillo, E.M., Marzoli, A., Bellieni, G., Renne, P.R., Ernesto, M., Marques, L.,2003. The Central Atlantic Magmatic Province (CAMP) in Brazil: Petrology, geochem-istry, 40Ar/39Ar ages, paleomagnetism and geodynamic implications. In: Hames, W.E.,McHone, J.G., Renne, P.R., Ruppel, C. (Eds.), The Central Atlantic Magmatic Province:Insights from Fragments of Pangea. American Geophysical Union Geophysical Mono-graphs, 136, pp. 209–226.

De Paolo, D.J., 1981. Trace element and isotopic effects on combinedwallrock assimilationand fractional crystallization. Earth and Planetary Science Letters 53, 189–202.

Deckart, K., Bertrand, H., Liegeois, J.-P., 2005. Geochemistry and Sr, Nd, Pb isotopic compo-sition of the Central Atlantic Magmatic Province (CAMP) in Guyana and Guinea.Lithos 82, 289–314.

Gerlach, D.C., Stormer, J.C., Mueller, P.A., 1987. Isotopic geochemistry of Fernando deNoronha. Earth and Planetary Science Letters 85, 129–144.

Ghiorso, M.S., Sack, R.O., 1995. Chemical Mass Transfer in Magmatic Processes IV. ARevised and Internally Consistent Thermodynamic Model for the Interpolation

and Extrapolation of Liquid–Solid Equilibria in Magmatic Systems at ElevatedTemperatures and Pressures. Contributions to Mineralogy and Petrology 119,197–212.

Gibson, S.A., Thompson, R.N., Day, J.A., Humphris, S.E., Dickin, A.P., 2005. Melt-generationprocesses associated with the Tristan mantle plume: Constraints on the origin ofEM-1. Earth and Planetary Science Letters 237, 744–767.

Gómez, J.J., Goy, A., Barrón, E., 2007. Events around the Triassic–Jurassic boundary innorthern and eastern Spain: a review. Palaeogeography, Palaeoclimatology, Palaeo-ecology 244, 89–110.

González-Jiménez, J.M., Villaseca, C., Griffin, W.L., Belousova, E., Konc, Z., Ancochea, E.,O'Reilly, S.Y., Pearson, N.J., Garrido, C.J., Gervilla, F., 2013. The architecture of theEuropean–Mediterranean lithosphere: a synthesis of the Re–Os evidence. Geology41, 547–550.

Hanan, B.B., Graham, D.W., 1996. Lead and helium isotope evidence from oceanic basaltsfor a common deep source of mantle plumes. Science 272, 991–995.

Hart, S.R., 1984. A large-scale isotope anomaly in the Southern Hemisphere mantle.Nature 309, 753–757.

Hill, R.I., 1991. Starting plumes and continental break-up. Earth and Planetary ScienceLetters 104, 398–416.

Hoernle, K., Zhang, Y.-S., Graham, D., 1995. Seismic and geochemical evidence for large-scale mantle upwelling beneath the eastern Atlantic and western and centralEurope. Nature 374, 34–39.

Hofmann, A.W., 2003. Sampling mantle heterogeneity through oceanic basalts: isotopesand trace elements. In: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry.The Mantle and Core, Oxford, UK, pp. 61–101.

Jolis, E.M., Freda, C., Troll, V.R., Deegan, F.M., Blythe, L.S., McLeod, C.L., Davidson, J.P., 2013.Experimental simulation of magma–carbonate interaction beneath Mt. Vesuvius,Italy. Contributions to Mineralogy and Petrology. http://dx.doi.org/10.1007/s00410-013-0931-0.

Jourdan, F., Marzoli, A., Bertrand, H., Cosca, M., Fontignie, D., 2003. The northernmostCAMP: 40Ar/39Ar age, petrology and Sr–Nd–Pb isotope geochemistry of the Kerfornedike, Brittany, France. In: Hames,W.E., McHone, J.G., Renne, P.R., Ruppel, C. (Eds.), TheCentral Atlantic Magmatic Province: Insights from Fragments of Pangea. AmericanGeophysical Union, Geophysical Monograph, 136, pp. 209–226.

Igneous rocks: a classification and glossary of terms. In: Le Maitre, R.W., et al. (Eds.),Recommendations of the International Union of Geological Sciences, Subcommissionof the Systematics of Igneous Rocks. Cambridge University Press, UK.

Liégeois, J.-P., Abdesalam, M.G., Ennih, N., Ouabadi, A., 2013. Metacraton: nature, genesisand behavior. Gondwana Research 23, 220–237.

Ludwig, K.R., 2003. User's manual for Isoplot 3.00. Berkeley Geochronology Center SpecialPublication. (74 pp.).

Manspeizer, W., 1994. The breakup of Pangea and its impact on climate: consequences ofVariscan–Alleghanide orogenic collapse. In: Klein, G.D. (Ed.), Pangea: Paleoclimates,Tectonics, and Sedimentation during Accretion, Zenith, and Breakup of a Superconti-nent. Geological Society of America Special Paper, 288, pp. 169–185.

Manuppella, G., 1992. Carta Geológica da Região do Algarve. Serviços Geológicos dePortugal; 2 sheets at the 1:100.000 scale and Explanatory Note.

Martins, L.T., Kerrich, R., 1998. Magmatismo toleítico continental no Algarve (Sul dePortugal): Um exemplo de contaminação crustal “in situ”. Comunicações do InstitutoGeológico e Mineiro 85, 99–116.

Martins, L.T., Madeira, J., Youbi, N., Munhá, J., Mata, J., Kerrich, R., 2008. Rift-relatedmagmatism of the Central Atlantic magmatic province in Algarve, SouthernPortugal. Lithos 101, 102–124.

Marzoli, A., Renne, P.R., Piccirillo, E.M., Ernesto, M., Bellieni, G., DeMin, A., 1999. Extensive200-million-year-old continental flood basalts of the Central Atlantic magmatic prov-ince. Science 284, 616–618.

Marzoli, A., Bertrand, H., Knight, K.B., Cirilli, S., Buratti, N., Vérati, C., Nomade, S., Renne,P.R., Nassrddine, Y., Martini, R., Allenbach, K., Neuwerth, R., Rapaille, C., Zaninetti, L.,Bellieni, G., 2004. Synchrony of the Central Atlantic magmatic province and theTriassic–Jurassic boundary climatic and biotic crisis. Geology 32, 973–976.

Marzoli, A., Jourdan, F., Puffer, J.H., Cuppone, T., Tanner, L.H., Weems, R.E., Bertrand,H., Cirilli, S., Bellieni, G., De Min, A., 2011. Timing and duration of the CentralAtlantic magmatic province in the Newark and Culpeper basins, eastern U.S.A.Lithos 122, 175–188.

Marzoli, A., Jourdan, F., Bussy, F., Chiaradia, M., Costa, F., 2014. Petrogenesis of tholeiiticbasalts from the Central Atlantic magmatic province as revealed by mineral majorand trace elements and Sr isotopes. Lithos 188, 44–59.

McDonough, W.F., Sun, S.-S., 1995. The composition of the Earth. Chemical Geology 120,223–254.

McHone, J.G., 1996. Broad-terrane Jurassic flood basalts across northeastern NorthAmerica. Geology 24, 319–322.

McHone, J.G., 2000. Non-plume magmatism and rifting during the opening of the centralAtlantic Ocean. Tectonophysics 316, 287–296.

McHone, J.G., Anderson, D.L., Beutel, E.K., Fialko, Y.A., 2005. Giant dikes, rifts, flood basaltsand plate tectonics: a contention of mantle models. In: Foulger, G.R., Natland, J.H.,Presnall, D.C., Anderson, D.L. (Eds.), Plates, Plumes and Paradigms. GSA SpecialVolume, 388, pp. 401–420.

Meisel, T., Walker, R.J., Irving, A.J., Lorand, J.-P., 2001. Osmium isotopic compositions ofmantle xenoliths: a global perspective. Geochimica Cosmochimica Acta 65,1311–1323.

Melluso, L., Mahoney, J.J., Dallai, L., 2006. Mantle sources and crustal input as recorded inhigh-Mg Deccan Traps basalts of Gujarat (India). Lithos 89, 259–274.

Merle, R., Marzoli, A., Bertrand, H., Reisberg, L., Verati, C., Zimmermann, C., Chiaradia, M.,Bellieni, G., Ernesto, M., 2011. 40Ar/39Ar ages and Sr–Nd–Pb–Os geochemistryof CAMP tholeiites from Western Maranhão basin (NE Brazil). Lithos 122,137–151.



http://refhub.elsevier.com/S0024-4937(13)00337-X/rf0005




































http://dx.doi.org/10.1029/2004RG000156




http://dx.doi.org/10.1029/2008GC002101














































http://dx.doi.org/10.1007/s00410-013-0931-0

http://dx.doi.org/10.1007/s00410-013-0931-0






















































Merle, R., Marzoli, A., Reisberg, L., Bertrand, H., Nemchin, A., Chiaradia, M., Callegaro, S.,Jourdan, F., Bellieni, G., Kontak, D., Puffer, J., McHone, J.G., 2013. Sr, Nd, Pb and Os iso-tope systematics of CAMP tholeiites from Eastern North America (ENA): evidence of asubduction-enriched mantle source. Journal of Petrology http://dx.doi.org/10.1093/petrology/egt063 (in press).

Montigny, R., Azambre, B., Rossy, M., Thuizat, R., 1982. Etude K/Ar du magmatismebasique lié au Trias supérieur des Pyrénées Conséquences méthodologiques etpaléogéographiques. Bulletin de Mineralogie 105, 673–680.

Moreno-Ventas, I., Rogers, G., Castro, A., 1995. The role of hybridization in the genesis ofHercynian granitoids in the Gredos Massif, Spain: inferences from Sr–Nd isotopes.Contributions to Mineralogy and Petrology 120, 137–149.

Nägler, T.F., Schäfer, H.J., Gebauer, D., 1995. Evolution of theWestern European continen-tal crust: implication from Nd and Pb isotopes in Iberian sediments. ChemicalGeology 121, 345–357.

Nimis, P., Ulmer, P., 1998. Clinopyroxene geobarometry of magmatic rocks. Part 1. Anexpanded structural geobarometer for anhydrous and hydrous basic and ultrabasicsystems. Contributions to Mineralogy and Petrology 133, 122–135.

Orejana, D., Villaseca, C., Paterson, B.A., 2006. Geochemistry of pyroxenitic andhornblenditic xenoliths in alkaline lamprophyres from the Spanish Central System.Lithos 86, 167–196.

Orejana, D., Villaseca, C., Pérez-Soba, C., López-García, J.A., Billström, K., 2009. The Variscangabbros from the Spanish Central System: a case for crustal recycling in the sub-continental lithospheric mantle? Lithos 110, 262–276.

Pereira, R., Alves, T.M., 2011. Margin segmentation prior to continental break-up: aseismic–stratigraphic record of multiphased rifting in the North Atlantic (SouthwestIberia). Tectonophysics 505, 17–34.

Perez-Soba, C., Villaseca, C., 2010. Petrogenesis of highly fractionated I-type peraluminousgranites: La Pedriza pluton (Spanish Central System). Geologica Acta 8, 131–144.

Peucker-Ehrenbrink, B., Jahn, B.M., 2001. Rhenium-osmium isotope systematics andplatinum group element concentrations: loess and the upper continental crust.Geochemistry, Geophysics, Geosystems 2 (2001GC000172).

Pinarelli, L., Rottura, A., 1995. Sr and Nd isotopic study and Rb–Sr geochronology of theBejar granites, Iberian Massif, Spain. European Journal of Mineralogy 7, 577–589.

Prelević, D., Jacob, D.E., Foley, S.F., 2013. Recycling plus: a new recipe for the formation ofAlpine–Himalayan orogenic mantle lithosphere. Earth and Planetary Science Letters362, 187.197.

Puffer, J.H., 2001. Contrasting high field strength element contents of continental floodbasalts from plume versus reactivated-arc sources. Geology 29, 675–678.

Reisberg, L., Lorand, J.-P., 1995. Longevity of sub-continental mantle lithosphere from os-mium isotope systematics in orogenic peridotite massifs. Nature 376, 159–162.

Renne, P.R., Mundil, R., Balco, G., Min, K., Ludwig, K.R., 2010. Joint determination of40 K decay constants and 40Ar*/40K for the Fish Canyon sanidine standard, andimproved accuracy for 40Ar/39Ar geochronology. Geochimica et CosmochimicaActa 74, 5349–5367.

Ribeiro, A., Munhá, J., Dias, R., Mateus, A., Pereira, E., Ribeiro, L., Fonseca, P., Araújo, A.,Oliveira, T., Romão, J., Chaminé, H., Coke, C., Pedro, J., 2007. Geodynamic evolutionof the SW Europe Variscides. Tectonics 26, TC6009.

Roeder, P.L., Emslie, R.F., 1970. Olivine-liquid equilibrium. Contributions to Mineralogyand Petrology 29, 275–289.

Rossi, Ph., Cocherie, A., Mark-Fanning, C., Ternet, Y., 2003. Datation U–Pb sur zircons desdolérites tholéiitiques pyrénéennes (ophites) à la limite Trias–Jurassique et relationsavec les tufs volcaniques dits «infra-liasiques» nord-pyrénéens. Comptes RendusGeoscience 335, 1071–1080.

Salters, V.J.M., Stracke, A., 2004. Composition of the depleted mantle. Geochemistry,Geophysics, Geosystems 5. http://dx.doi.org/10.1029/2003GC000597.

Schlische, R.W., Withjack, M.O., Olsen, P.E., 2003. Relative timing of CAMP, rifting, conti-nental breakup, and basin inversion; tectonic significance. In: Hames, W.E.,McHone, J.G., Renne, P.R., Ruppel, C.A. (Eds.), The Central AtlanticMagmatic Province:

Insights from fragments of Pangea. American Geophysical Union, Geophysical Mono-graph, 136, pp. 33–60.

Sebai, A., Feraud, G., Bertrand, H., Hanes, J., 1991. 40Ar–39Ar dating and geochemistry oftholeiitic magamatism related to the early opening of the Central Atlantic rift. Earthand Planetary Science Letters 194, 455–472.

Sobolev, S.V., Sobolev, A.V., Kuzmin, D.V., Krivolutskaya, N.A., Petrunin, A.G., Arndt, N.T.,Radko, V.A., Vasiliev, Y.R., 2011. Linking mantle plumes, large igneous provincesand environmental catastrophes. Nature 477, 312–316.

Spera, F.J., Bohrson, W.A., 2001. Energy-constrained open-system magmatic processes I:general model and energy-constrained assimilation and fractional crystallization(EC-AFC) formulation. Journal of Petrology 42, 999–1018.

Steiger, R.H., Jäger, E., 1977. Subcommission on geochronology: convention on the use ofdecay constants in geo- and cosmochronology. Earth and Planetary Science Letters 36,359–362.

Terrinha, P., Ribeiro, C., Kullberg, J.C., Lopes, C., Rocha, R., Ribeiro, A., 2002. Compressiveepisodes and faunal isolation during rifting, Southwest Iberia. Journal of Geology110, 101–113.

Tollo, R.P., Gottfried, D., 1992. Petrochemistry of Jurassic basalt from eight cores,Newark Basin, New Jersey: implications for the volcanic petrogenesis of theNewark Supergroup. In: Puffer, J.H., Ragland, P. (Eds.), Eastern North AmericanMesozoic Magmatism. Geological Society of America Special Paper, 268,pp. 233–260.

Ugidos, J.M., Valladares, M.I., Recio, C., Rogers, G., Fallick, A.E., Stephens, W.E., 1997. Prov-enance of Upper Precambrian–Lower Cambrian shales in the Central Iberian Zone,Spain: evidence from chemical and isotopic study. Chemical Geology 136, 55–70.

Verati, C., Rapaille, C., Féraud, G., Marzoli, A., Bertrand, H., Youbi, N., 2007. 40Ar/39Ar agesand duration of the Central Atlantic Magmatic Province volcanism in Morocco andPortugal and its relation to the Triassic–Jurassic boundary. Palaeogeography,Palaeoclimatology, Palaeoecology 244, 308–325.

Villaseca, C., Barbero, L., Rogers, G., 1998. Crustal origin of Hercynian peraluminous granit-ic batholiths of central Spain: petrological, geochemical and isotopic (Sr, Nd) argu-ments. Lithos 43, 55–79.

Villaseca, C., Downes, H., Pin, C., Barbero, L., 1999. Nature and composition of the lowercontinental crust in central Spain and the granulite–granite linkage: inferencesfrom granulitic xenoliths. Journal of Petrology 40, 1465–1496.

Villaseca, C., Orejana, D., Paterson, B.A., Billstrom, K., Perez-Soba, C., 2007. Metaluminouspyroxene-bearing granulite xenoliths from the lower continental crust in centralSpain: their role in the genesis of Hercynian I-type granites. European Journal ofMineralogy 19, 463–477.

Villaseca, C., Bellido, F., Pérez-Soba, C., Billström, K., 2009. Multiple crustal sources forpost-tectonic I-type granites in the Hercynian Iberian Belt. Mineralogy and Petrology96, 197–211.

Villaseca, C., Orejana, D., Belousova, E.A., 2012. Recycled metaigneous crustal sources forS- and I-type Variscan granitoids from the Spanish Central System batholith:Constraints from Hf isotope zircon composition. Lithos 153, 84–93.

Walgenwitz, F., 1976. Etude pétrologique des roches intrusives triasiques, des écailles dusocle profond et des gîtes de chlorite de la région d'Elizondo (Nacarre espagnole).Université de Besançon, France (PhD thesis).

Willbold, M., Stracke, A., 2006. Trace element composition of mantle end-members:implications for recycling of oceanic and upper and lower continental crust. Geo-chemistry, Geophysics, Geosystems 7. http://dx.doi.org/10.1029/2005GC001005.

Willbold, M., Stracke, A., 2010. Formation of enriched mantle components byrecycling of upper and lower continental crust. Chemical Geology 276, 188–197.

Wilson, M., 1997. Thermal evolution of the Central Atlantic passive margins: continentalbreak-up above a Mesozoic superplume. Journal of Geological Society of London 154,491–495.

Zindler, A., Hart, S.R., 1986. Chemical geodynamics. Annual Reviews of Earth and Plane-tary Sciences 14, 493–571.






















































http://dx.doi.org/10.1029/2003GC000597























































http://dx.doi.org/10.1029/2005GC001005








Enriched mantle source for the Central Atlantic magmatic province: New supporting evidence from...

Documents

Transcript of Enriched mantle source for the Central Atlantic magmatic province: New supporting evidence from...