Weber 2009 Cardona Cabo de La Vela

The Cabo de la Vela Mafic–Ultramafic Complex, Northeastern

Colombian Caribbean region: a record of Multistage evolution

of a late Cretaceous intra-oceanic arc

M. B. I. WEBER1*, A. CARDONA2,3, F. PANIAGUA1, U. CORDANI4,

L. SEPULVEDA1 & R. WILSON4

1Universidad Nacional de Colombia, Calle 65 No. 78-28, Facultad de Minas,

M1-324, Medellın, Colombia2Smithsonian Tropical Research Institute, Balboa, Ancon, Panama – ECOPETROL,

Piedecuesta, Colombia3Institute of Geoscience, USP, Rua do Lago 562, Cidade Universitaria,

05508-080 Sao Paulo, Brazil4University of Leicester, Department of Geology, University Road, Leicester LE1 7RH, UK

*Corresponding author (e-mail: [email protected])

Abstract: Ophiolite-related rocks accreted to Caribbean plate margins provide insights intocomplex intra-oceanic evolution of the plate and its interaction with continental margins of theAmericas. Petrologic, geochemical and isotope (K–Ar, Sr and Nd) data were obtained in poorlyknown serpentinites, gabbros and andesite dykes of the Cabo de la Vela Mafic–UltramaficComplex from the Guajira Peninsula, north Colombia. Field relations, metasomatic alteration pat-terns and whole rock–mineral geochemistry combined with juvenile isotope signatures of thedifferent units suggest that gabbros and serpentinites formed in a slow-spreading supra-subductionzone that was brought to shallower depths and subsequently evolved to an arc setting where ande-sitic rocks formed with little sediment input. The tectonomagmatic evolution of the Cabo de la VelaMafic–Ultramafic Complex involved an intra-oceanic arc that evolved from pre-Campanian timeto 74 Ma. Relationships with other units from the Guajira Peninsula show either the existence of amature arc basement or a series of coalesced allocthonous arcs, juxtaposed before accretion ontothe passive continental margin of South American in pre-Eocene times. Correlations with ophio-lite-related rocks of the Southern Caribbean and diachronism in the accretionary process are com-patible with a WSW–ESE advancing Caribbean plate front.

Ophiolitic rocks are fundamental to understandingof the complex dynamics and evolution of oceaniccrust and its interaction with continental marginsto form so-called Cordilleran-type orogens(Shervais 2001; Beccaluva et al. 2004).

In the Caribbean area, studies of ophiolitic com-plexes have identified remnants of rift margins,ocean plateaux and arc systems formed in oceanicdomains that evolved since the Mesozoic, followingthe break-up of Pangea (Giunta et al. 2002, 2006).However, different models show lack of concensus.Testing of regional models through local relation-ships has often not been successful, possibly dueto lack of detailed information (Iturralde-Vinent &Lidiak 2006 for discusions).

A few ophiolite-related rocks have beendescribed from the Colombian Caribbean(MacDonald 1964; Lockwood 1965; Alvarez1967); however, their formation setting and their

implications for the geotectonic models of theCaribbean have not been discussed.

We present integrated petrological, mineralchemistry, whole rock geochemistry, Nd and Sr iso-topes and K–Ar geochronology data from the Cabode la Vela Mafic–Ultramafic Complex of the south-western Caribbean. They allow reconstruction of aLate Cretaceous, arc-related magmatic system thatwas accreted to the continental margin of SouthAmerica in pre-Eocene times. The implication ofthis and other available regional data is discussedwithin the framework of the Caribbean platetectonic evolution.

Geological setting

The Meso-Cenozoic history of the ColombianAndes in the Caribbean region is characterized byinteraction between the northwestern border of

From: JAMES, K. H., LORENTE, M. A. & PINDELL, J. L. (eds) The Origin and Evolution of the Caribbean Plate.Geological Society, London, Special Publications, 328, 547–566.DOI: 10.1144/SP328.22 0305-8719/09/$15.00 # The Geological Society of London 2009.

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758

South America and the Caribbean and Nazcaplates (Fig. 1). Multiple plate boundaries and Meso-Cenozoic transpressive tectonics controlled accre-tion of oceanic terranes to the continental margin.Discrete tectonic blocks record this complex evol-ution (Pindell 1993; Toussaint 1996; Montes et al.2005).

The Guajira Peninsula, in northeasternmostColombia (Fig. 2), is characterized by several iso-lated massifs with correlatable geology, surroundedby broader flat lands and Cenozoic basins(MacDonald 1964; Lockwood 1965; Alvarez1967). Within these massifs at least three main litho-tectonic belts can be identified. From oldest to mostrecent they include the following elements (Fig. 2):

1. A composite, Late Mesoproterozoicand Palaeo-zoic metamorphic domain, which includesmedium and high-grade units. This is intrudedby Jurassic magmatism, similar to the parau-tochthonous basement of the Andes (Cordaniet al. 2005; Cardona-Molina et al. (2006).

2. A weakly deformed belt of Mesozoic sedi-mentary rocks with the same depositional

characteristics and ages as the autochthonousSouth American margin (Villamil 1999).

3. A sequence of two different, Cretaceous,low-grade meta-volcano-sedimentary meta-morphic units. The northernmost unit has inter-calated mafic and ultramafic plutonic rocks,intruded by Eocene magmatism (MacDonald1964; Lockwood 1965; Alvarez 1967; Pindell1993).

Fragments of high-pressure metamafic and meta-sedimetary rocks in Miocene conglomerates on thenorthwestern fringe of the central massif (Fig. 2)may represent a probably mid-Cretaceousexhumed subduction-accretion complex (Greenet al. 1968; Zapata et al. 2005).

An association of mafic and ultramafic rocksoccurs in the Cabo de la Vela region, isolatedfrom the main massifs and close to the zone ofCaribbean–South American plate interaction. Thefollowing sections describe the geological and geo-chemical characteristics of this mafic–ultramaficcomplex and discuss their significance to interpret-ations of tectonomagmatic evolution.

400 km

Galápagos

Mexico

Colombia

CVC

Colombian Guajira

NORTH AMERICAN PLATE

SOUTH AMERICAN PLATE

NAZCA PLATE

COCOS PLATE

Fig. 1. Tectonic Framework and location of the Caribbean region. The Guajira Peninsula and the CVC are shown. Thedistribuiton of ophiolite rocks within the Caribbean region is included (data from: Lewis et al. 2006; Guinta et al. 2002).

M. B. I. WEBER ET AL.548

5960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116

Cabo de la Vela Mafic–Ultramafic Complex

The Cabo de la Vela consists of a series of isolatedsmall hills (150 m) exposing different mafic andultramafic units (Alvarez 1967). In general theyinclude a sequence of serpentinites, gabbros andmafic volcanic units. The serpentinite ultramaficunit contains gabbro and leucograbbros as lensesand pods, cross-cut by basalt dykes (Fig. 3).Miocene sediments locally overlie this sequence(Alvarez 1967).

To the SE, a lower and flat-land region of Ceno-zoic sediments extends for 5 km, separating theCabo de la Vela from the massifs of the Guajiraregion (MacDonald 1964; Lockwood 1965;Alvarez 1967; Fig. 2). Similarities between theplutonic mafic and ultramafic rocks of the Cabo dela Vela region and the Cretaceous units of themassifs suggest a geological link between them(review in Alvarez 1967).

The Bouguer gravity map of northern Colombia(Fig. 4) shows a positive anomaly centred on theCabo de la Vela, extending 100 km offshore to thenorthwest and 30 km onshore to the southeast

(Kellogg et al. 1991). This anomaly is consistentwith a NW–SE elongated dense body, with anarrow and gently sloping northwestern face and abroad and steep southeastern flank, and suggestswide extension of this mafic–ultramafic rockcomplex (Nieto & Ojeda, pers. comm., 2006).

The following section introduces the informalstratigraphic term Cabo de la Vela Mafic–Ultra-mafic Complex (CVC) and describes the field andpetrographic elements of the constituent units.

Ultramafic rocks

Ultramafic rocks of the Cabo de la Vela area aremainly serpentinites. The only exposure of a par-tially serpentinized rock (sample FP-34C) showsmeso-scale foliation, with original grains ofolivine, spinel and clinopyroxene. Bastite mineralssuggest the former presence of another pyroxene(orthopyroxene?), which indicates that the originalultramafic rock was a wherlite. Relict textures inthis rock are characterized by larger grains sur-rounded by fine-grained material and by deformed

Fig. 2. Simplified Geological map of the northern Guajira Peninsula, showing major litostratigraphic units.

THE CABO DE LA VELA MAFIC COMPLEX 549

117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174

N N W

W

S S W

W E

S E S

N E N E

N

SCALE

600 0 600 1200 km

m 5

7

Caribbean Sea

CABO DE LA VELA

Tertiary sedimentary rocks

Cenozoic cover

Cabo de la Vela Complex (CVC)

Basaltic Dykes(exaggerated scale)

x =

880

.000

y = 1,846.000

Fig. 3. Geology of the Cabo the Vela area.

50

50

50

0

0

0

0

100

50

–50

73°W 72°W

12°N

Fig. 4. Bouguer gravity map of the Guajira area (Kellogg et al. 1991).


175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232

cleavages, indicating that the ultramafic protolithwas a mantle tectonite (Fig. 5) (Nicolas & Rabino-wicz 1984).

The serpentinites are mainly greenish to brown-ish rocks, in many cases mottled by the silky-looking bastite minerals. Microscopic analysisshows pseudomorph textures, such as antigorite

mesh and lizardite–antigorite hourglass. Nonpseu-domorphic textures include interpenetrating andintergrowth textures, both in the same sample andboth defined by antigorite. Other local mineralsformed during serpentinzation are bastite, magne-tite, magnesite and brucite. Relict minerals arebrown spinel and occasional amphibole. The

Fig. 5.

Colour

online=colour

hardcopy

(a) Gabbro

Q4

intrusions in serpentinite. (b) Continuous basaltic andesite dyke within serpentinite. The arrowindicates the direction of the dyke. (c) Serpentinised wehrlite (sample FP-34C) with relict olivine (Ol). Note thepresence of brown bastite (Bst), as replacement of pyroxenes. (d) Gabbro tectonite. Note the brown amphiboleinclusions in the central pyroxene (Cpx) crystal, and secondary amphibole on the recrystallised pyroxene grain margins(Hbl). The two upper pyroxenes and plagioclase (plg) have been completely recrystallised and have a granoblasticpolygonal texture. (e) Basaltic Andesite with twinned pyroxene (Cpx) phenocryst in a hornblende (Anf) andsaussuritisized plagioclase (Sauss) matrix.


233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290

presence of amphibole indicates a hydrated mantleprotolith.

Bastite may replace clinopyroxene, withopaques defining relict bend cleavages, and maybeorthopyroxene without opaques. The presence oftwo pyroxenes and the foliated texture suggeststhat the protolith was also a wehrlitic tectonite.

In addition, there are at least four generations ofveining, formed by either perpendicular or parallelto the wall opening precipitation of crysotile.

Gabbros and hornblendites

Gabbroic rocks occur in the ultramafites as small,discontinuous, irregular bodies, pods and lenseswith maximum dimensions of 1 m � 100 m. Theyare coarse-grained to pegmatitic, and contain darkgreen pyroxene (diopside) and plagioclase (labra-dorite) crystals up to 15 cm in size.

These rocks generally show well-definedmineral banding and lineation. Most show evidenceof deformation and in thin section the overall textureis granoblastic polygonal due to deformation–recrystallization of large crystals to smaller poly-gonal crystals at the edges (Fig. 5). The variousdegrees of recrystallization always involve moreplagioclase than pyroxene, which is sometimes pre-served in large crystals. These relict grains havenumerous oriented inclusions of small brownamphibole blebs that impart a characteristic lustreto hand-specimen. The euhedral shape of theblebs, their colour and the fact that they formedbefore high temperature deformation all indicatemagmatic origin (Coogan et al. 2001). Evidenceof deformation of the larger grains before recrystal-lization is conspicuous, marked by bent and kinkedcleavages and by alignment of the small amphiboleblebs that disappear with recrystallization.

The fact that pyroxene and plagioclase ductiledeformation and recrystallization annealed withgranoblastic textures, without changes in the orig-inal mineralogy assemblage, is indicative of defor-mation at granulite–amphbolite facies conditions(Seyler et al. 1998).

Hornblendites are found generally on the edgesof the gabbro dykes and sometimes as small metre-sized patches within the serpentinites. They aremade up exclusively of coarse-grained dark brownand less commonly bright green amphibole. Fieldobservations show that these rocks formed by com-plete replacement of pyroxene and plagioclase ofthe gabbros and possibly the ultramafics. Severalstages of replacement are present. Early stages areseen as coronas around pyroxenes but at moreadvanced stages the reaction front replaces all ofthe minerals (Fig. 5). Petrographically these rocksshow decussate textures and the nature of the repla-cement of these hornblendites indicates that they

formed through metasomatism during annealingand therefore after deformation.

Gabbros and hornblendites have been rodingi-tized throughout the whole complex, and transi-tional rocks are common. Massive rodingitescomprise Ca-rich minerals such as chlorite +hydrogrossular + vesuvianite+ epidote+ albite +tremolite-actinolite + prehnite. Some rodingitesare zoned, with a chlorite margin, in which therelicts of the original pyroxene can still be ident-ified. The mineral assemblages suggest that theserocks formed at greenschist facies and prehnite–pumpellyte facies conditions (Dubinska 1995;Frost 1975; Fruh-Green et al. 1996).

The transition in facies, as well as differentstyles of complex superimposed deformation andfracturing during fluid penetration, indicate thatrodingitization occurred from ductile to brittleconditions.

Mafic volcanic dykes

Dykes are continuous and undeformed and reach upto 2 m in width, cutting the previously describedlithologies. They constitute fine-grained, phaneriticto aphanitic, green to grey-green rocks, generallywith porphyritic texture.

Plagioclase (andesine) and clinopyroxene (diop-side) with intergranular to subophiitic texture areseen in thin section (Fig. 5). Subhedral crystals ofclinopyroxene and plagioclase comprise the pheno-crystal phase. Several of these rocks also containbrown amphibole as a matrix mineral and as uraliticreplacement of pyroxene. The more felsic rockscontain quartz in the matrix. All minerals are gener-ally altered to saussurite and chlorite and pumpel-lyte is common.

The suggested crystallization order would bepyroxene and plagioclase–opaques–amphibole–quartz, and petrographically may resemble basaltsor andesites.

Analytical techniques

Whole rock geochemistry

Eleven samples were analysed for major andtrace elements by X-ray fluorescence (XRF) andinductively coupled mass spectrometry (ICP-MS)at the Mineralogy and Petrology Department ofthe Institute of Geosciences of the University ofSao Paulo and at ACME analytical laboratories inCanada.

Samples were crushed with an iron steel crusherand pulverised in an agate mill. Sample preparationfor XRF included microreduction to obtain pressedpowder pellets, and fused glass discs for majorand trace element determination. Major and selected


291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348

trace element analysis were carried out in awavelength dispersive Philips PW 2400 XRF spec-trometer with detection limits generally of the orderof 1–10 ppm for trace elements, following themethodology described by Mori et al. (1999). Forother trace elements including rare earths (REE),the analyses were carried out by ICP-MS atACME analytical laboratories (Canada), afterlithium metraborate/tetraborate fusion and nitricacid digestion of a 0.2 g sample.

Mineral chemistry

Mineral analyses were obtained from carbon-coatedpolished thin sections using a Jeol 8600S electronmicroprobe at the Department of Geology, Univer-sity of Leicester, UK, using an acceleratingvoltage of 15 kV and a probe current of 30 nAwith a beam diameter of 5–10 mm. Quantitativebackground-corrected results were standardizedagainst a combination of synthetic materials andwell-characterized natural minerals and correctedfor matrix effects using a ZAF correction procedure.Minimum detection limits under the analyticalconditions used range from 0.01 wt% for Na2Oto 0.04 wt% for FeO.

K–Ar geochronology

Three whole rock samples were analysed by theK–Ar method at the Centre of GeochronologicalResearch of the University of Sao Paulo(CPGeo-USP). Two aliquots from the samesample were separated for the K and Ar analysis.Potassium analyses of each pulverized samplewere carried out in duplicate, coupled to an ultra-vacuum system. A spike of 38Ar was added andthe gas was purified in titanium and copper ovens.Final argon determinations were carried out in aReynold-type gas spectrometer. Analytical pre-cision for K is of 5% whereas for Ar it is around0.5%. Decay constants for calculation are afterSteiger & Jager (1977).

Nd–Sr isotopes

Six whole rock samples were analysed by Sm–Ndand Rb–Sr methods at the Centre for Geochronolo-gical Research of the University of Sao Paulo(CPGeo-USP). For the Sm–Nd method theanalyticalQ1 procedures followed Sato et al. (1995).Isotopic ratios 143Nd/144Nd were obtained in amulti-collector mass spectrometer Finnegan Mat,with analytical precision of 0.0014% (2s). Exper-imental error for the 147Sm–144Nd ratios is ofthe order of 0.1%. La Jolla and BCR-1 standardsyielded 143Nd/144Nd ¼ 0.511849+ 0.000025 (1s)and 0.512662 + 0.000027 (1s) respectivelyduring the period in which the analyses were

performed. The 1Nd were calculated followingDe Paolo (1988), and the constants usedinclude 143Nd/144Nd (CHUR) ¼ 0.512638 and147Sm–144Nd (CHUR)0 ¼ 0.1967.

Rb–Sr analyses followed procedures presentedby Tassinari et al. (1996). Rb and Sr values wereobtained by X-ray fluorescence, and 87Sr/86Srratios were done with mass spectrometer VG-sectormass spectrometer and corrected for isotopicfractionation during thermal ionization with a87Sr/86Sr ¼ 01194.

Geochemistry Q2

Whole rock geochemistry and mineral chemistrydata were obtained from selected samples in orderto understand its tectonic setting of the CVC.

Mineral chemistry

Four samples, a partially serpentinized a peridotite,a serpentinite, a gabbro and a basalt, were selectedfor mineral chemical analysis. Analysed mineralswere spinel, olivine, pyroxene, plagioclase andamphibole.

Spinel

Dark brown spinel in the partially serpentinizedsample (FP-34C) and in serpentinites is armouredby magnetite grains, possibly formed during theserpentinization process. Original spinel has anallotriomorphic, interstitial texture. Preservedspinel cores have Cr/Cr þ Al composition rangesfrom 0.5 to 0.6 and Fe/Fe þ Mg around 0.5. TiO2

compositions vary from 0.07 to 0.14 wt% andAl2O3 cluster around 25 wt%. In the Cr/CrþAl v.Mg/Mg þ Fe diagram (Fig. 6a) the CVC data donot fall within the abyssal spinel peridotite data ofDick & Bullen (1984) but do overlap the spineldata from the Mariana fore-arc of hole 780,ODP-Leg 125 from Parkinson & Pearce (1998).

Pyroxene

Pyroxene is seldom preserved in the ultramaficrocks. Only relict pyroxenes were found in sampleFP-34C, where composition is Fe-rich diopside(Fig. 6b) and TiO2 was found to be very low.

Clinopyroxene from an analysed gabbroicsample is mainly diopside, falling within the samecompositional area as the relict pyroxenes fromthe serpentinites (Morimoto et al. 1988) with highMg no. between 0.90–0.91, high SiO2 and extre-mely low TiO2.

Compositional patterns of magmatic clinopyrox-ene are consistently used to infer possible tectonic


349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406

settings of ophiolitic mafic and ultramafic rocks (e.g.Nisbet & Pearce 1977; Hebert & Laurent 1987;Capedri & Venturelli 1979; Beccaluva et al. 1989).

In bivariant diagrams that use TiO2 in clinopyr-oxene as a marker of the degree of fusion and

different tectonic settings, the Cabo de La Velagabbroic rocks plot within island arc fields(Fig. 6b, c). Together with the low TiO2 content ofpyroxene from ultramafic rocks this suggests asupra-subduction setting.

0.00

0.20

0.40

0.60

0.80(a)

0.000.200.400.600.801.00

Spi

nel C

r/(C

r+A

l)

Spinel Mg/(Mg + Fe)

Mariana fore-arc

Abyssal Peridotites

Serpentinite

Sample FP-34C

(b)

Clinoenstatite Clinoferrosillite

Pigeonite

Augite

Diopside Hedenbergite

En Fs

Wo

Basaltic Andesites

Gabbro

Ultramafic rock

0

0.05

0.1

0.15

0.2

0.25

0.3(c)

0 0.06 0.12 0.18 0.24Ti

AI (

t)

MORB

IAT

BON

SiO2/100 Na2O

TiO2

MORB

IATBON

Basaltic Andesite

Gabbro

Ultramafic rock

(d)

Fig. 6. (a) Cr# [Cr/(CrþAl)] versus Mg# [Mg/(Mgþ Fe)] relation in spinels from a completely serpentinised and onepartially serpentinised rock (sample FP-34C) of the CVC. For comparison, the spinel composition from peridotites withdifferent refractory character are also plotted. Field for spinels from abyssal spinel peridotites (from Dick & Bullen1984) and Izu–Bonin–Mariana forearc (hole 780, ODP-Leg 125, Parkinson & Pearce 1998) are shown; open circlesubduction-related Oman ophiolite (Lippard et al. 1986). (b) Clinopyroxene classification for the CVC ultramafics,gabbros and basaltic dykes. (c) Co-variation diagram of Al(t) versus Ti (atomic ratios) of studied pyroxenes in ultramaficrocks, indicating their tectonic settings. (d) TiO2-Na2O-SiO2/100 in clinopyroxene (wt %) (IAT, Island Arc Tholeiite;BON, Boninite; MORB, Mid-ocean ridge basalt; after Beccaluva 1989).


407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464

Tab

le1.

Majo

rand

trace

elem

ent

analy

ses

Sam

ple

LS

-6L

S-1

8L

S-2

2L

S-2

6L

S-3

4L

S-4

4L

S-4

5L

S-5

1L

S-5

3L

S-5

9A

LS

-62

SiO

249.0

347.6

145.8

752.9

152.4

952.0

352.6

753.6

4T

iO2

1.0

11

1.0

92

0.3

02

0.7

32

0.8

92

0.3

84

0.5

21

0.3

72

Q6

Al 2

O3

7.3

110.0

914.7

514.9

516.1

516.6

13.2

34.2

9F

e 2O

37.6

55.8

33.1

11.2

110.8

96.5

9.2

23.7

Mn

O0.1

16

0.0

71

0.0

53

0.1

81

0.1

90.1

10.1

83

0.0

7M

gO

18.9

518.7

914.6

15.3

74.8

87.8

311.5

22.1

2C

aO

11.6

912.0

414.5

88.1

78.3

410.3

911.1

112.4

2N

a2O

1.0

31.4

41.1

94.3

34.7

44.4

12.1

90.7

K2O

0.0

60.0

20.0

20.7

90.5

80.0

80.0

30.0

1P

2O

50.0

16

0.0

07

0.0

06

0.0

95

0.1

44

0.0

10.0

12

0.0

07

LO

I2.5

2.2

14.3

71.8

91.3

91.9

80

2.1

7T

ota

l99.3

699.2

98.8

5100.6

3100.6

9100.3

2100.6

799.5

Typ

eH

BL

HB

LG

BR

BA

BA

BA

BA

GB

RB

AG

BR

HB

L

K498

166

166

6558

4815

––

664

–249

83

Ba

3.9

2.1

9.1

272.6

553.8

264.7

506.9

18.6

539.7

70

Rb

0.7

01.9

11

8.4

3.4

8.9

17.8

00.6

Sr

14.6

29.8

269.6

219.5

293.9

188.7

283.7

240.6

272.2

111.7

5.1

Cs

00

00.7

0.4

0.2

0.3

0.3

0.2

00

Ga

12.7

13

12.3

17.7

19.2

15.9

18.9

17.7

16.6

13.3

4.4

Ta

0.4

0.1

00.6

0.3

14.2

0.3

0.5

1.1

0.9

0.5

Nb

2.2

00

0.5

1.3

13.8

1.3

01.4

00

Hf

1.4

00

1.1

1.8

1.7

1.4

01.7

00.5

Zr

36.6

9.4

3.1

36.9

51.2

45.4

51.7

5.9

48.2

10.4

9.3

Ti

6061

6547

1810

4388

5348

––

2302

–3123

2230

Y24.9

18.8

6.3

17.5

19.3

23.7

20

9.4

22

13.8

10.9

Th

0.7

00

0.4

1.2

0.4

1.7

01.2

00

U0

00

0.2

0.5

0.3

0.6

00.3

00

Cr

733

390

132

38

40

00

74

0502

5054

Ni

129.1

91.9

68.6

16.9

17.1

7.9

16.9

73.4

13.2

53.2

26.5

Co

71.1

69

32.3

51.1

42.8

88.2

43.5

55.1

81.1

83.4

52.5

Sc

41

60

38

37

31

––

35

–40

37

V277

422

149

324

259

308

262

200

366

233

261

Cu

0.6

3.4

2117.3

91.3

108

90.7

51.1

114.9

30.7

0.4

Pb

00

00.8

1.3

2.6

1.1

00.5

00

Zn

74

365

57

54

55

10

63

43

W118.1

154.1

66.6

511.4

199.7

3671.9

212.1

392.6

333.4

901.3

651.8

Mo

0.3

0.2

0.1

0.9

0.4

50.4

0.4

0.6

0.7

0.5

Cl

1425

548

351

160

23

00

732

0279

342

(Conti

nued

)


465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522

Other minerals

Relict olivine is preserved within mesh cells fromsample FC-34C. The composition is Fo90–91.Plagioclase compositions for the gabbros weredetermined petrographically, and comprise mainlylabradorite–andesine. Chemical analyses for feld-spar in some volcanic samples indicate spilitizationfor these rocks, as plagioclase is replaced by albite.

Brown amphibole of metasomatic origin, whichreplaces minerals in the gabbros, and amphibolesfrom one of the mafic dykes were analysed, bothbeing magnesiohornblendes (Hawthorne 1981).

Major and trace element geochemistry

Five volcanic rocks, together with three gabbros andthree hornblendites representative of the main rocktypes of the CVC, were selected for major andtrace element analyses.

Element data are presented in Table 2. Elementanalyses for LILES in the gabbros are in generalbelow detection limits.

SiO2 values range from 45.87 to 52.67 wt% forgabbros, from 47.61 to 53.64 wt% for hornblendites,and cluster around 52.5 wt% for the mafic dykes.MgO wt% values are variable but are highest inthe hornblendite samples (18.79–22.12%) andlowest for the mafic rocks (4.88–5.37%).

According to the inmobile elements (Nb, Y, Zrand Ti) classification of Winchester & Floyd(1977) (not shown), all these rocks are of mafic inter-mediate composition and plot in the andesite field.

Since most of the samples show signs of altera-tion, chemical interpretation is based on high fieldstrength elements and transitional metals consideredto be immobile (Pearce & Cann 1973; Winchester &Floyd 1977).

In the V v. Ti/1000 of Shervais (1982), allsamples plot towards the arc-related rocks field,whereas in the Ti v. Zr plot after Pearce (1982),the basaltic andesite dyke samples fall within thevolcanic arc field (figures not shown). Gabbrosand hornblendites have particularly low Zr.

Additional constraints on the geotectonic settingcan be obtained from REE and multi-element dia-grams presented in Figure 7b, c. Gabbros show astrongly depleted LREE pattern compared withHREE when plotted on the chondrite-normalizedREE diagram (Fig. 7b), with Nd–Lu valuesranging from 6.67 to 11.00. They have a notablepositive Eu anomaly, indicating that plagioclasefractionation was an important factor in thegenesis of these rocks. Gabbro REE patternsroughly resemble modern MORB-type plutonicrocks. The multi-element diagram shows negativeZr and Ti anomalies as well as positive Sr and BaT

ab

le1.

Conti

nued

Sam

ple

LS

-6L

S-1

8L

S-2

2L

S-2

6L

S-3

4L

S-4

4L

S-4

5L

S-5

1L

S-5

3L

S-5

9A

LS

-62

La

2.5

00

2.9

6.8

3.8

6.4

0.5

40

0.5

Ce

9.3

1.2

17.6

14.2

8.6

13.5

1.4

10.1

1.4

1.7

Pr

1.5

50.3

40.1

51.2

41.9

51.4

42.0

30.2

61.4

90.3

0.3

1N

d7.8

1.6

0.8

6.6

9.3

7.3

10.3

1.3

7.2

2.2

1.9

Sm

2.7

1.3

0.4

2.1

2.9

2.5

2.5

0.8

2.6

10.9

Eu

0.8

10.3

90.2

80.6

60.8

60.7

50.9

10.5

50.7

70.5

80.4

2G

d3.6

72.2

10.6

92.8

13.1

43.6

33.2

81.1

3.5

31.6

51.4

3T

b0.7

30.4

10.1

50.4

80.5

50.6

0.5

30.2

50.5

90.3

40.2

8D

y4.4

33.0

61.0

92.8

63.3

23.5

93.0

91.4

3.0

92.1

81.6

Ho

0.9

20.6

90.2

30.6

90.7

30.8

80.7

20.3

90.8

40.5

60.3

8E

r2.2

51.8

40.6

41.7

1.7

82.0

81.9

21.0

72.1

81.4

11.1

Tm

0.4

10.2

50.1

0.2

70.2

60.3

60.2

90.1

50.3

30.2

40.1

9Y

b2.9

51.9

60.6

41.9

62.1

62.5

32.1

21.0

92.3

81.5

51.1

6L

u0.4

30.2

40.1

10.3

20.3

10.3

60.2

90.1

40.3

50.2

0.1

5

HB

L¼

horn

ble

ndit

e,G

BR¼

gab

bro

,B

A¼

bas

alti

can

des

ite.


523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580

Tab

le2.

Rep

rese

nta

tive

min

eral

analy

ses

Spin

elO

livin

eP

yro

xen

eA

mphib

ole

Sam

ple

LS

-32

FP

-34C

Sam

ple

FP

-34C

FP

-34C

Sam

ple

FP

-34C

LS

22

LS

11

Sam

ple

LS

18

LS

22

LS

11

Type

SE

RP

ER

PE

Rbord

erP

ER

centr

eP

ER

GB

RA

ND

HB

LG

BR

AN

D

SiO

20.0

00.0

3S

iO2

40.1

640.4

8S

iO2

53.2

653.5

151.7

4S

iO2

47.4

649.4

051.5

1T

iO2

0.1

10.0

5T

iO2

0.0

00.0

0T

iO2

0.0

00.0

80.0

8T

iO2

1.0

40.7

00.4

2A

l 2O

325.5

226.2

8A

l 2O

30.0

10.0

0A

l 2O

31.5

01.5

91.4

7A

l 2O

311.5

59.6

35.6

3C

r 2O

339.0

541.7

5C

r 2O

30.0

00.0

0C

r 2O

30.3

63.2

31

2.5

9C

r 2O

30.0

20.0

40.0

0F

eO23.3

622.3

9F

eO8.6

99.3

7F

eO1.8

80.0

00.0

0F

eO5.1

55.0

812.1

8M

nO

0.3

20.1

3M

nO

0.1

10.1

1M

nO

0.0

60.1

00.4

0M

nO

0.1

10.0

60.4

0M

gO

10.8

611.1

3M

gO

49.8

850.0

3M

gO

17.9

516.7

09.3

1M

gO

18.4

618.7

87.1

1C

aO

0.0

00.0

0C

aO

0.0

20.0

3C

aO

24.3

424.0

723.4

5C

aO

12.3

912.8

518.7

4N

iO0.1

1N

iO0.0

50.0

1N

iO0.1

90.4

40.8

9N

iO2.0

61.7

52.5

8K

2O

0.0

0K

2O

0.0

00.0

2K

2O

0.0

00.0

00.0

0K

2O

0.0

60.0

50.0

1N

iO0.4

20.4

3N

iO0.0

80.1

10.0

0T

ota

l99.3

2101.8

0T

ota

l99.3

3100.4

8T

ota

l99.5

699.7

599.9

2T

ota

l98.3

798.4

498.5

8

Si

0.0

00.0

1S

i0.9

90.9

9S

i1.9

41.9

61.9

8S

i6.6

56.9

07.5

3T

i0.0

20.0

1T

i0.0

00.0

0A

l0.0

60.0

40.0

2T

i0.1

10.0

70.0

5A

l7.4

47.4

7A

l0.0

00.0

0A

l0.0

00.0

30.0

4A

l1.9

11.5

80.9

7C

r7.6

47.9

6C

r0.0

00.0

0F

e(ii

i)0.0

30.0

60.0

6C

r0.0

00.0

00.0

0V

0.0

00.0

0F

e(ii

)0.1

80.1

9C

r0.0

50.0

00.0

0F

e 20.6

00.5

91.4

9F

e(ii

i)0.8

80.5

3M

n0.0

00.0

0T

i0.0

00.0

00.0

0M

n0.0

10.0

10.0

5F

e(ii

)3.9

53.9

9M

g1.8

31.8

2F

e(ii

)2

0.0

10.0

40.3

4M

g3.8

53.9

11.5

5M

n0.0

70.0

3N

i0.0

10.0

1M

n0.0

00.0

00.0

1C

a1.8

61.9

22.9

4M

g4.0

04.0

0C

a0.0

00.0

0M

g0.9

70.9

10.5

3N

a0.5

60.4

70.7

3C

a0.0

00.0

0T

ota

l3.0

13.0

1C

a0.9

50.9

40.9

6K

0.0

10.0

10.0

0Z

n0.0

00.0

0N

a0.0

10.0

30.0

7N

i0.0

10.0

10.0

0T

ota

l24.0

024.0

0K

0.0

00.0

00.0

0T

ota

l15.5

715.4

815.3

0A

lto

t0.0

60.0

00.0

0T

ota

l4.0

94.0

24.0

2

Cr/

Cr

1A

l0.5

10.5

2F

o91.0

090.3

9W

o48.6

847.4

848.7

4F

e/F

e1

Mg

0.5

50.5

3F

a8.8

99.5

0E

n49.9

645.8

426.9

3M

g/

Mg

1F

e0.4

50.4

7T

p0.1

10.1

2F

s0.6

65.1

020.9

8C

r/C

r1

Al

0.5

10.5

2A

c0.7

01.5

73.3

5

SE

R¼

Ser

pen

tinit

e,P

ER¼

serp

enti

nis

edper

idoti

te,

HB

L¼

horn

ble

ndit

e,G

BR¼

gab

bro

,A

ND¼

andes

ite.


581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638

spikes, indicative of a subduction-related com-ponent in the source.

Volcanic dykes show flat patterns, with a smallnegative Eu anomaly, when plotted on thechondrite-normalized REE diagram. Apparentslight variations in the LREE, with La–Lu rangingfrom 9.06 to 21.94, indicate small degrees of differ-entiation. This is in agreement with the petrographicobservations, where the LREE enriched samplescontain amphibole and quartz in addition to pyrox-ene and plagioclase.

The multi-element diagram for basaltic andesitesshows a decrease from LILE-enriched to HFSE-depleted and contrasts with the gabbros. The LILEenrichment is generally attributed to element mobi-lity during alteration; however, the increased con-tents of relatively immobile elements like Th andLa suggest that this pattern is inherited from theoriginal magmatic source. The basaltic rocks alsoshow Nb, Zr and a weak Ti negative anomaly,

which together with positive Sr and Ba anomaliesand the already mentioned pattern, characterizesubduction-related magmas.

Hawkesworth et al. (1993a, b), subdividedisland arc basalts into two groups on the basis ofLREE/HREE, using La–Yb ratios to discriminatebetween predominantly intra-oceanic arcs (La–Yb , 5) and arcs developed near continentalmargins (La–Yb . 5). The CVC basaltic rocksfall within the low La–Yb island arc group andoverlap the data from the Mariana arc.

K–Ar geochronology

Previous age constraints from this region wererestricted to stratigraphical relationships withMiocene sediments. Three andesitic dykes weredated by the K–Ar whole rock method. Obtainedages (Table 3a) overlap within error. Their

0.01

0.1

1

10

100

1000

CsRb

BaTh

UK

NbLa

CeSr

NdZr

SmEu

TiY

ErYb

Lu

Sam

ple/

MO

RB

0.1

1

10

100(a)

LaCe

PrNd Sm

EuGd

TbDy

HoEr

TmYb

Lu

Sam

ple/

Nak

amur

a, 1

977

1

10

100(b)

LaCe

PrNd Sm

EuGd

TbDy

HoEr

TmYb

Lu

Sam

ple/

Nak

amur

a, 1

977

0.01

0.1

1

10

100

1000

CsRb

BaTh

UK

NbLa

CeSr

NdZr

SmEu

TiY

ErYb

Lu

Sam

ple/

MO

RB

Fig. 7. Geochemistry of gabbros

Q5

from the CVC. Normalizing values are from Nakamura (1977) and Pearce (1983).(a) REE patterns normalized to chondrite and multi-element patterns normalized to MORB for CVC gabbros. Opencircles represent hornblendites, closed circles represent gabbros. (b) REE patterns normalized to chondrite and multi-element patterns normalized to MORB for CVC basaltic andesite dykes. The grey fields outline the compositionalrange of the southern Mariana Trough (Gribble et al. 1996).


639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696

differences may be related to minor hydrothermalalteration shown by saussuritization of plagioclase.We consider c.74 Ma to be the age of dyke intrusioninto the mafic–ultramafic unit based on the moreprecise analytical quality. Low K contents in thehornblendites preclude reliable K–Ar data to con-strain the pre-dyke intrusion history. However,cross-cutting relations and depth of formation indi-cate that the gabbroic and ultramafic rocks musthave been exhumed before intrusion of the dykes,and therefore a significant lapse of time must havepassed between the formation of these twodifferent units.

Sr–Nd isotopes

Andesitic rocks and gabbros were analysed forRb–Sr and Sm–Nd isotopic ratios. Initial ratioswere calculated for the 74 Ma K–Ar age. Resultsare presented in Table 3b and 3c and Figure 9.The 1Nd and initial 87Sr/86Sr values for the basalticandesite dykes range between 4.1 and 7.3 and0.7038 and 0.7041 respectively. Gabbros have 1Nd

of 9.9 and initial 87Sr/86Sr of 0.7029 and 0.7031.When compared with the basaltic dykes, thegabbros overlap the MORB field (Fig. 9), indicatingderivation from a depleted source, whereas thebasaltic dykes plot similar to primitive island arc

rocks found in other suites in the circum-Caribbeanrealm. Isotopic variation in the basalt dykes could beexplained by different input of sedimentary com-ponents into the subduction-zone magmas. Interest-ingly, Mariana back-arc samples plot between thegabbros and the basaltic dykes from the CVC,which could be interpreted as a higher input of thesedimentary component for the basaltic dykes inthe CVC. This is further seen by slight off-settowards higher Sr ratios of the CVC and the Carib-bean plutons when compared with the Marianaback-arc basalts.

When compared with the highly negative 1Nd

data from Pre-Mesozoic basement and Jurassicplutons from the Guajira region (Cordani et al.2005; Cardona-Molina et al. 2006), the data fallaway from older crustal signatures. Thereforeassimilation of older crust is precluded. Unpub-lished isotopic data from the Eocene continental-arcmagmatic rocks intruding Cretaceous metamorphiccomplexes of the Guajira region, possibly related tothe CVC, show mixing with older basement rocks orsediments, but this is not seen in the CVC rocks.

These isotopic characteristics clearly indicate aprimitive mantle source and a minor sedimentinput in the subduction-related magma source ofthe basaltic rocks. The differences between the iso-topic ratios of gabbros and basaltic dykes may be

Table 3. Geochronological and isotopical data

K-ArSample Rock % K K error

(%)Ar40 RadccSTP/g(�1026)

Ar40Atm(%)

Age(Ma)

+ (Ma)

LS-34 Basalt 0.5013 0.5000 1.46 71.94 73.5 4.2LS-44 Basalt 0.1711 0.5000 0.47 68.79 69.5 4.9LS-53 Basalt 0.5162 1.7499 1.58 77.84 77.3 5.4

Sm-NdSample Rock Sm

(ppm)Nd

(ppm)

147Sm/144Nd

Error 143Nd/144Nd

Error fSm/Nd 1 (74Ma)

LS-51 Gabro 0.661 1.459 0.2739 9 0.513183 17 0.39 9.90LS-22 Gabro 0.761 0.513228 12 2 1.00LS-34 Basalt 0.680 8.892 0.0462 2 0.512774 41 2 0.76 4.05LS-44 Basalt 2.462 7.340 0.2028 7 0.512950 8 0.03 6.02LS-26 Basalt 1.986 6.057 0.1983 7 0.513011 13 0.01 7.25

Rb-SrSample Rock Rb

(ppm)Sr

(ppm)Rb87/Sr86

(X)Error Sr87/Sr86

(Y)Error 87Sr/86Sr

74 Ma

LS-51 Gabro 0.76 254.29 0.0087 1 0.70291 6 0.7029LS-22 Gabro 15.28 228.79 0.1932 125 0.70338 7 0.7032LS-34 Basalt 8.12 272.99 0.0861 7 0.70386 10 0.7038LS-44 Basalt 3.00 166.68 0.0520 4 0.70396 7 0.7039LS-26 Basalt 10.58 211.38 0.1448 12 0.70406 1 0.7041


697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754

explained by different input at source. The gabbrosformed by partial fusion of mantle in asuprasubduction-zone environment, with limitedinput from the subducted slab. Input from the

subducting plate was greater in basaltic andesitesand shifts the isotopic ratios towards lower 1Nd

and higher Sr, probably reflecting a changingtectonic configuration.

0 1 2 3 4 50

5

10

15

20

25

30

Yb

La

La/Yb = 5

N-MORB

Grenada

Aleutians Is.

Mariana. Sandwich Is.

Basaltic Andesites

Gabbros

Fig. 8. La v. Yb in basalts (SiO2 , 55%) Q6in intraoceanic and continental margin island arc basalts, after Hawkesworthet al. (1993a, b). Other data fields from Jolly et al. (2006) and Taylor & Martinez (2003).

MORB Caribbean PIA

Guajira Eocene arc

–10.0

–5.0

0.0

5.0

eNd

10.0

15.0

0.701 0.702 0.703 0.704 0.705 0.706 0.707 0.708 0.709 0.71

Ocean Sediments

Mariana arc

87Sr/86Sr

Basalts Gabbros

Basaltic Andesites Gabbros

Fig. 9. 1Nd v. 87Sr/86Sr for gabbros and basalts of the CVC. Other fields depicted are from the Mariana arc (Gribbleet al. 1996 and sources listed therein), MORB (White & Hofmann 1982) and Caribbean PIA (Jolly et al. 2006).Also shown is the unpublished data field for the Eocene arc magmatism.


755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812

Tectonomagmatic setting of the

Cabo de la Vela Complex

Based on the lack of units typical of complete ophio-lite sequences and the unclear relation with themargin where it was accreted, the Cabo de la Velacomplex represents a Cordilleran-type ophiolite, asdo many other Caribbean ophiolites (Beccaluvaet al. 1996, 2004).

Field and petrologic evidence indicate thatthe Cabo the la Vela rocks followed a complexsuccession of events, recording the dynamicoceanic tectonic cycles common of ophiolite rocks(Shervais 2001).

The oldest unit comprises mantle rock of wherli-tic composition. Coarse-grained gabbros and trocto-lites intrude this peridotite, recording continuousuplift in a slow spreading setting. Both units showevidence of correlatable high temperature defor-mation as well as hydration. Mineral chemistry ofspinel and pyroxene (Figs 6a, b) from peridotitesand gabbros indicates that these units correspondto a tectonite formed in a supra-subduction zoneenvironment (Kenemetsky et al. 2001; Okamuraet al. 2006). The MORB geochemical and morejuvenile isotopic signature from the gabbroscombined with the supra-subduction signaturefrom the mineral chemistry are more akin to aback-arc tectonic setting where both MORB or sub-duction zone signatures are common (Saunders &Tarney 1984).

The presence of hornblendites and different gen-erations of pervasive serpentinization and rodingiti-zation events are indicative of several hydrothermaloverprints ranging from deep to shallow crustaldepths and recording a continuous tectonic exhuma-tion history in an oceanic setting. This, together withlateral heterogeneity of hydrothermal alteration,may be taken as evidence of a slow spreadingridge environment (Cannat 1996; Cannat et al.1992). Absence of the other typical ophioliticunits, like cumulate gabbros and intrusive basalticsheeted dykes, may be explained by erosion or tec-tonic removal after emplacement onto the margin,but nevertheless fits nicely into the evidence of aslow spreading ridge environment for these rocks.

The history of the andesitic dykes differs fromthe previously discussed ultramafics–gabbro unit.The dykes show neither deformation nor pervasivehydrothermal alteration, and must therefore corre-spond to another stage of the evolution history. Geo-chemical data show that these rocks may representan intraoceanic island arc with poor sedimentinput or older crust contamination. Sr and Nd isoto-pic comparison between the andesites and thegabbros neatly shows the differences betweenmagma sources and confirms that the CVC plutonicand volcanic rocks are two separate units, formed at

different stages within an intraoceanic environment.A possible explanation for such magmatic vari-ations could include the presence of differentmantle sources (MORB and supra-subduction)related to different phases of migration of the arcin a long term subduction environment (Stern 2002).

These events are constrained by the 74 MaK–Ar crystallization age obtained for the volcanicdykes and imply that the ultramafic rocks andthe gabbros of the CVC were emplaced beforethe Campanian.

Modern-day analogues for slow-spreadingsupra-subduction zone environments are theMariana Trough and the Lau back-arc (Gribbleet al. 1996; Taylor & Martinez 2003). Basaltsfrom both back-arc basins include arc-like com-ponents and MORB-like end-members. Mantleflow and convection induce mixing of previouslydepleted mantle sources and produce a range ofcompositions that can vary through time betweenend-members (Taylor & Martinez 2003). Thiscould well be the case in the CVC, whereby theultramafic and gabbroic units represent of a moreMORB-like end-member source and the andesiticdykes unit a later, more subduction-related end-member source of the same arc.

As described in the geological setting, two diff-erent Cretaceous low-grade vulcano-sedimentarymetamorphic units are exposed in the GuajiraPeninsula, the Jarara Formation to the SE and theEtpana Formation to the NW, described in greatdetail by Alvarez (1967) and Lookwood (1965).These are temporally and geologically linked tothe Cabo de la Vela Mafic–Ultramafic Complexof this paper (Fig. 2).

The lithostratigraphic characteristics of themetamorphic protoliths from these two unitsinclude a sequence of mainly siliclastic sediments(pelites to rudites) with intercalations of mafictuffs and lavas indicative of a mature arc-relatedgeological setting. Differences between the unitsare that the Jarara Fm. includes marbles and moreabundant volcanic rocks, whereas the Etpana Fm.has intermixed serpentinites and gabbros. Thesemetasediments are part of the same basin, with theEtpana Fm. representing the deeper sedimentaryenvironment (Lockwood 1965). Fossil ages rangefrom Turonian to Maastrichtian. Nearby Cretaceousunits of the autochthonous margin of South Americalack the volcanic and predominant silicicalstic com-ponents of the Etpana and Jarara Fms. (McDonald1964; Villamil 1999). It is therefore also possibleto assume that the Jarara–Etpana sequence formedin an allochtonous arc position.

The contact between these Cretaceous units andbasement rocks has been described as a shear-zone(MacDonald 1964; Lockwood 1965; Alvarez1967) that we interpret as the possible suture


813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870

between authoctonous South America and theallochtonous Jarara–Etpana arc. The presence ofundeformed Eocene continental arc plutonism inthis region and the regionally correlatable plutonsof the Santa Marta massif, that clearly intrude theSouth American margin (Tschanz et al. 1974),also suggest that accretion occurred before thismagmatic event.

The relation between the Jarara–Etpanasequence and the CVC is hidden. However, asAlvarez (1967) pointed out, the ultramafic and gab-broic rocks found in the CVC resemble the interca-lated mafic and ultramafic rocks of the Etpana Fm.Reconnaissance field and petrographic observationsof the Etpana Formation indicate a strong similarityto the CVC in the nature and distribution of serpen-tinization and rodingization (Arredondo et al.2005). The map-pattern shows that the expectedtectonic pattern of intermixing of these units

(clearly aligned lens shapes in the low graderocks) is not evident, and therefore we suggest thatthe vulcano-sedimentary protoliths of the EtpanaFm. could have been deposited over a substratumcorrelatable with the CVC ultramafic andgabbroic units.

This interpretation has two corollaries. The firstis that the mafic dykes from the CVC may be con-temporaneous with the deposition of both Jararaand Etpana Fms. The second is that deformationmay be related to a subsequent tectonic episode.

A tentative model for the early Cretaceous toEocene evolution of the CVC derived from theabove data is presented in Figure 10. It includesthree main stages:

1. Initiation of an ocean – ocean subduction zone,forming a slow-spreading back-arc basin inpre-Campanian times. This back-arc,

Fig. 10. Three-stage model for the formation of the Cabo de la Vela Mafic-Ultramafic Complex. SAM represents theSouth American margin (Modified from Giunta et al. 2006).


871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928

represented by the CVC ultramafic and gab-broic units, was progressively exhumed as aconsequence of the slow spreading dynamics.

2. Subduction zone configuration changed andarc-like magmatism developed upon amigrating spreading centre, represented by theandesitic dykes of the CVC. Possibly simul-taneously the arc evolved and sedimentarydeposits formed. The distal sedimentarysequence, Eptana Fm., developed upon thepreviously rifted basin, whereas the more shelf-like deposits of the Jarara Fm. formed assediments near the arc. Continuous oceansubduction carried the arc towards the passivecontinental margin of South America.

3. The units were accreted onto the SouthAmerican margin. Following the accretion theCaribbean plate began to subduct under SouthAmerica. This event is constrained by the ageof Eocene magmatism that intruded thealready deformed Etpana and Jarara For-mations at 47 Ma (Lockwood 1965). Cardona-Molina et al. (2006) deduced that this couldhave occurred between the late Cretaceousand the Palaeocene, based on Ar/Ar spectraof basement rocks.

Caribbean realm

The tectonic evolution of the Caribbean is related toJurassic–Early Cretaceous formation of oceanicProto-Caribbean crust following the separation ofNorth and South America. Subsequent developmentof a multistage intraoceanic-arc or several intra-oceanic arcs in either a near mid American or aPacific position happened from late Cretaceous toRecent times. Thickened oceanic plateau crustmigrated from the west between the Americas,leaving fragments on the continental margin(Pindell 1993; Pindell & Kennan 2001; reviews inGiunta et al. 2002, 2006; James 2006).

The Cabo de la Vela Mafic–Ultramafic Complexand the associated Etpana–Jarara Fms of theGuajira region record the tectonic evolution of anisland arc of Campanian and older age (wholerock K–Ar age of 74 Ma for the basaltic andesites,and the older ultramafic and gabbroic units). Thisformed on already mature arc basement or as partof coalesced arcs juxtaposed before their accretiononto a passive continental margin in pre-Eocenetimes. High-pressure rocks in Miocene conglomer-ates that show notable similarities to rocks of theVenezuelan Cordillera de la Costa and Margarita,dated at 90–110 Ma (Sisson et al. 1997; Stockhertet al. 1995; Zapata et al. 2005), quartz and granitoidclasts in metaconglomerates from the Jarara Fm.may be vestiges of earlier subduction. This implies

prolonged intra-oceanic plate convergence for thisintra-oceanic arc setting. Other remnants of mag-matic arcs and subduction complexes recordingocean–ocean convergence in the southern Carib-bean are the early Cretaceous Villa de Cura andDos Hermanas units from Venezuela, the SantaMarta Schists of Colombia and the WashikembaFormation in Bonaire (MacDonald et al. 1971;Beccaluva et al. 1996; Giunta et al. 2002;Thompson et al. 2004). Petrographical and geo-chemical comparison, as well as age constrains,show similarities which indicate that the DosHermanas unit or the Washikemba Formation(Giunta et al. 2002; Thompson et al. 2004) couldbe correlated to the rocks of the Jarara–Etpanaformations and the CVC.

Ages similar to those obtained for the CVChave been recorded throughout the DutchAntilles. In Curacao, Sinton et al. (1998) report ac. 76 Ma Ar/Ar age for a dolerite sill thatintrudes the Albian–Turonian Curacao lava suc-cession. Late Cretaceous turbidites overlying thissequence contain a significant population ofeuhedral zircons in the range of 70–87 Ma(Wright 2004). In Aruba, Priem et al. (1986),based on various Rb–Sr and K–Ar age determi-nations, suggest that a c. 72 Ma thermal eventwas responsible for the isotopic resetting of theAruba Batholith. These data suggest an importantc. 70–76 Ma magmatic event that correlates withthe CVC.

The segmented distribution of correlatable Cre-taceous arc-related and other intra-oceanic units(including the CVC) along the continental marginof Northern South America and the Caribbean,and SW- to east- younging accretion, becomingyounger from Ecuador to Colombia and Vene-zuela, are compatible with the northeasterlymigration of the Caribbean plate with its arcfronts between the Americas, and both the accre-tion and subsequent disruption within an obliquecontinental margin (Pindell 1993; Toussaint1996; Ave-Lallement & Sisson 2005; Vallejoet al. 2006).

This project was partially funded by the National Univer-sity of Colombia, Medellin.

The authors wish to acknowledge the contributions ofJorge Gomez from the Colombian Geological Survey(INGEOMINAS) and the staff of the GeochronologicalResearch Centre (CP-Geo) and geochemical laboratoriesof the University of Sao Paulo. We also thankG. Y. Ojeda and M. Nieto from the ICP for their helpwith gravimetric data and C. Jaramillo for constructivecomments on the initial draft of the paper.

We are also grateful to K. H. James for inviting andencouraging us to present this paper and G. Giunta forthe constructive review and suggestions to improve themanuscript.


929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986

References

ALVAREZ, W. 1967. Geology of the Simarua and Carpin-tero area, Guajira Peninsula, Colombia. PhD thesis,Princeton University.

ARREDONDO, L. F., WEBER, M. ET AL. 2005. Petrografıade las ultramafitas y rodingitas de la Serranıa de Jarara,Penınsula de la Guajira-Colombia. X Congreso Colom-biano de Geologıa – Simposio de Geologıa Regional.

AVE-LALLEMENT, H. G. & SISSON, V. B. 2005. Exhuma-tion of eclogites and blueschist in northern Venezuela:constraints from kinematic analysis of deformationstructures. In: AVE LALLEMENT, J. G. & SISSON,V. B. (eds) Caribbean–South American Plate Inter-actions, Venezuela. Geological Society of America,Boulder, CO, Special Papers, 394, 193–306.

BECCALUVA, L., COLTORTI, M. ET AL. 1996. Cross sec-tions through the ophiolitic units of th Southern andNorthern Margins of the Caribbean Plate, in Venezuela(Northern Cordilleras) and Central Cuba. Ofioliti, 21,85–103.

BECCALUVA, L., COLTORTI, M., GIUNTA, G. & SIENA,F. 2004. Tethyan vs. Cordilleran ophiolites: a reapprai-sal of distinctive tectono-magmatic features of supra-subduction complexes in relation to the subductionmode. Tectonophysics, 393, 163–174.

BECCALUVA, L., MACCCIOTTA, G., PICCARDO, G. B. &ZEDA, O. 1989. Clinopyroxene composition of ophio-lite basalts as petrogenetic indicator. ChemicalGeology, 77, 165–182.

CANNAT, M. 1996. How thick is the magmatic crust atshow spreading oceanic ridges? Journal of Geophysi-cal Research, 101, 2847–2857.

CANNAT, M., BIDEAU, D. & BOGAULT, H. 1992. Ser-pentnized peridotites and gabbros in the Mid-AtlanticRidge axial valley at 158370N and 168520N. Earthand Planetary Science Letters, 109, 87–106.

CAPEDRI, S. & VENTURELLI, G. 1979. Clinopyroxenecomposition of ophiolitic metabasalts in the Mediterra-nean area. Earth and Planetary Science Letters, 43,61–73.

CARDONA-MOLINA, A., CORDANI, U. & MACDONALD,W. 2006. Tectonic correlations of pre-Mesozoic crustfrom the northern termination of the ColombianAndes, Caribbean region. Journal of South AmericanEarth Sciences, 21, 337–354.

COOGAN, L. A., MACLEOD, C. J. ET AL. 2001. Whole-rock geochemistry of gabbros from the SouthwestIndian Ridge: constraints on geochemical fractiona-tions between the upper and lower oceanic crust andmagma chamber processes at (very) slow-spreadingridges. Chemical Geology, 178, 1–22.

CORDANI, U. G., CARDONA, A., JIMENEZ, D., LIU, D. &NUTMAN, A. P. 2005. Geochronology of Proterozoicbasement inliers from the Colombian Andes: tectonichistory of remmants from a fragmented grenvillebelt. In: VAUGHAN, A. P. M., LEAT, P. T. & PAN-

KHURST, R. J. (eds) Terrane Processes at theMargins of Gondwana. Geological Society, London,Special Publications, 246, 329–346.

DE PAOLO, D. J. 1988. Age dependence of the compo-sition of continental crust: evidence from Nd isotopicvariations in granitic rocks. Earth and PlanetaryScience Letters, 90, 263–271.

DICK, H. J. B. & BULLEN, T. 1984. Chromian spinel as apetrogenetic indicator in abyssal and alpine-type peri-dotites and spatially associated lavas. Contributions toMineralogy and Petrology, 86, 54–76.

DUBINSKA, E. 1995. Rodingites of the Eastern Part of theJordanow–Gogolow Serpentinite Massif, LowerSilesia, Poland. The Canadian Mineralogist, 33,585–608.

FROST, B. R. 1975. Contact metamorphism of serpenti-nites: chloritic black-wall and rodingite at Paddy-go-easy, Central Cascades, Washington. Journal ofPetrology, 16, 272–313.

FRUH-GREEN, G. L., PLASS, A. & LECUYER, C. 1996.Petrologic and stable isotope constraints on hydrother-mal alteration and serpentinization of the EPR shallowmantle at Hess Depp (Site 895). In: MEVEL, C. ET AL.(eds) Proceedings of the Ocean Drilling Program,Scientific Results, 147. Ocean Drilling Program,Texas A&M University, 255–291.

GIUNTA, G., BECCALUVA, L., COLTORTI, M., SIENA, F.& VACCARO, C. 2002. The southern margin of theCaribbean Plate in Venezuela: tectono-magmaticsetting of the ophiolite units and kinematic evolution.Lithos, 63, 19–40.

GIUNTA, G., BECCALUVA, L. & SIENA, F. 2006. Carib-bean Plate margin evolution: constraints and currentproblems. Geologica Acta, 4, 265–277.

GREEN, D. H., LOCKWOOD, J. P. & KISS, E. 1968. Eclo-gite and almandine–jadeite–quartz rock from theGuajira Peninsula, Colombia, South America. Ameri-can Mineralogist, 53, 1320–1335.

GRIBBLE, R. F., STERN, R. J., BLOOMER, S. H., STUBEN,D., O’HEARN, T. & NEWMAN, S. 1996. MORBmantle subduction components interact to generatebasalts in the southern Mariana Trough back-arcbasin. Geochimica et Cosmchimica Acta, 60,2153–2166.

HAWKESWORTH, C. J., GALLAGHER, K., HERGT, J. M. &MCDERMOTT, F. 1993a. Mantle slab contributions inarc magmas. Annual Reiews in Earth and PlanetaryScience, 21, 175–204.

HAWKESWORTH, C. J., GALLAGHER, K., HERGT, J. M. &MCDERMOTT, F. 1993b. Trace element fractionationprocesses in the generation of island arc basalts. In:TARNEY, J., PICKERING, K. T., KNIPE, R. J. &DEWEY, J. F. (eds) Melting and Melt Movement inthe Earth. Clarendon Press/The Royal Society,London, 393–405.

HAWTHORNE, F. C. 1981. Crystal chemistry of the amphi-boles. Reviews in Mineralogy and Geochemistry, 9A,1–102.

HEBERT, R. & LAURENT, R. 1987. Mineral chemistry ofthe plutonic section of the Troodos ophiolite: newconstraints for genesis of arc-related ophiolites. In:MALPAS, G. J. & MOORES, E. M. (eds). Troodos1987 Ophiolites and Oceanic Lithosphere, 149–163.

ITURRALDE-VINENT, M. A. & LIDIAK, E. G. (eds) 2006.Caribbean plate tectonics: stratigraphic, magmatic,metamorphic and tectonic events (UNESCO/IUGSIGCP Project 433). Geologica Acta, 4, 1–341.

JAMES, K. H. 2006. Arguments for and against thePacific origin of the Caribbean Plate: discussion,finding for an inter-American origin. Geologica Acta,4, 279–302.


987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044

JOLLY, W. T., LIDIAK, E. G. & DICKIN, A. P. 2006.Cretaceous to Mid-Eocene pelagic sediment budgetin Puerto Rico and the Virgin Islands (northeastAntilles Island arc). Geologica Acta, 4, 35–62.Q7

KELLOGG, J. N., GODLEY, V. M., ROPAIN, C., BERMU-

DEZ, A. & AIKEN, C. L. V. 1991. Gravity Field ofColombia, Eastern Panama and Adjacent MarineAreas. The Geological Society of America, Boulder,CO, Map and Chart Series MCH 070.

KENEMETSKY, V. S., CRAWFORD, A. J. & MEFFRE, S.2001. Factors controlling chemistry of magmaticspinel: an empirical stud of associated olivine,Cr-spinel and melt inclusions from primitive rocks.Journal of Petrology, 42, 345–352.

LEWIS, J. F., DRAPER, G., PROENZA, J. A., ESPAILLAT,J. & JIMENEZ 2006. Ophiolite-related ultramaficrocks (serpentinites) in the caribbean region: a reviewof their occurrence, composition, origin, emplacementand ni-laterite soil formation. Geologica Acta, 4,237–263.

LIPPARD, S. J., SHELTON, A. W. & GASS, I. G. 1986. TheOphiolite of Northern Oman. Geological SocietyMemoirs, 11. Blackwell, Oxford, 178.Q8

LOCKWOOD, J. P. 1965. Geology of the Serranıa de Jararaarea, Guajira Penınsula, Colombia. PhD thesis,Princeton University.

MACDONALD, W. D. 1964. Geology of the Serranıa deMacuira area, Guajira Penınsula, Colombia. PhDthesis, Princeton University.

MACDONALD, W., DOOLAN, B. & CORDANI, U. 1971.Cretaceous–Early Tertiary metamorphic age valuesfrom the South Caribbean. Geological Society ofAmerica Bulletin, 82, 1381–1388.

MESCHEDE, M. 1986. A method of discriminatingbetween different types of mid-ocean ridge basaltsand continental tholeiites with the Nb–Zr–Y dia-grams. Chemical Geology, 56, 207–218.Q9

MONTES, C., HATCHER, R. D. & RESTREPO-PACE, P.2005. Tectonic reconstruction of the northern Andeanblocks: oblique convergence and rotations derivedfrom the kinematics of the Piedras–Girardot area,Colombia. Tectonophysics, 399, 221–250.

MORI, P. E., REEVES, S., TEIXEIRA, C. & HAUKKA, M.1999. Development of a fused glass disc XRF facilityand comparison with the pressed powder pellettechnique at Instituto de Geociencias, Sao PauloUniversity. Revista Brasileira de Geociencias, 29,441–446.

MORIMOTO, N., FABRIES, J. ET AL. 1988. Nomenclatureof pyroxenes. American Mineralogist, 62, 53–62.

NAKAMURA, N. 1974. Determination of REE, Ba, Fe,Mg, Na and K in carbonaceous and ordinary chon-drites. Geochimica et Cosmochimica Acta, 38,75–775.Q3

NICOLAS, A. & RABINOWICZ, M. 1984. Mantle flowpattern at oceanic spreading centres: relation withophiolitic and oceanic structures. In: GASS, I. G.,LIPPARD, S. J. & SHELTON, A. W. (eds) Ophiolitesand Oceanic Lithosphere. Geological Society,London, Special Publications, 13, 147–151.

NISBET, E. G. & PEARCE, J. A. 1977. Clinopyroxene com-position in mafic lavas from different tectonic settings.Contributions to Mineralogy and Petrology, 63,149–160.

OKAMURA, H., ARAI, S. & KIM, Y.-U. 2006. Petrology offorearc peridotite from the Hahajima Seamount, theIzu-Bonin arc, with special reference to chemicalcharacteristics of chromian spinel. MineralogicalMagazine, 70, 15–26.

PARKINSON, I. J. & PEARCE, J. A. 1998. Peridotites fromthe Izu–Bonin–Mariana forearc (ODP Leg 125): evi-dence for mantle melting and melt–mantle interactionin a supra-subduction zone setting. Journal of Petrol-ogy, 39, 1577–1618.

PEARCE, J. A. 1982. Trace element characteristics of lavasfrom destructive plate boundaries. In: THORPE, R. S.(ed.) Andesites. Wiley, Chichester, 525–548.

PEARCE, J. A. & CANN, J. R. 1973. Tectonic setting ofbasic volcanic rocks determined using trace elementanalyses. Earth and Planetary Science Letters, 19,290–300.

PINDELL, J. L. 1993. Evolution of the Gulf of Mexico andthe Caribbean. In: DONOVAN, S. K. & JACKSON, T. A.(eds) Caribbean Geology: An Introduction. Universityof the West Indies Publisher’s Association, Kingston,13–39.

PINDELL, J. L. & KENNAN, L. 2001. Kinematic evolutionof the Gulf of Mexico and Caribbean. Transactions,Petroleum Systems of Deep-water Basins: Globaland Gulf of Mexico Experience. GCSSEPM 1stAnnual Research Conference, Houston, TX. GulfCoast Section Society of Economic Paleontologistsand Mineralogists Foundation, 159–192.

PINDELL, J. L., KENNAN, L., STANEK, K. P., MARESCH,W. V. & DRAPER, G. 2006. Foundations of Gulf ofMexico and Caribbean evolution: eight controversiesresolved. Geologica Acta, 4, 303–341. Q10

PRIEM, H. N. A., BEETS, D. J., BOELRIJK, N. A. I. M. &HEBEDA, E. H., 1986. On the age of the late Cretac-eous tonalitic/gabbroic batholith on Aruba. Nether-lands Antilles (southern Cribbean borderland).Geologie en Mijnbouw, 65, 247–256.

SAUNDERS, A. D. & TARNEY, J. 1984. Geochemicalcharacteristics of basaltic volcanism within back-arcbasin. In: KOKELAAR, B. P. & HOWELLS, M. F.(eds) Marginal Basin Geology: Volcanic AssociatedSedimentary and Tectonic Processes in Modern andAncient Marginal Basins. Geological Society,London, 16, 59–76.

SEYLER, M., PAQUETTE, J.-L., CEULENEER, G.,KIENAST, J.-K. & LOUBET, M. 1998. Magmaticunderplating, metamorphic evolution, and ductileshearing in a Mesozoic lower crustal–upper mantleunit (Tinaquillo, Venezuela) of the Caribbean Belt.Journal of Geology, 106, 35–58.

SHERVAIS, J. W. 1982. Ti–V plots and the petrogenesis ofmodern and ophiolitic lavas. Earth and PlanetaryScience Letters, 59, 101–118.

SHERVAIS, J. W. 2001. Birth, death, and resurrection: thelife cycle of suprasubduction zone ophiolites. Geo-chemistry, Geophysics, Geosystems, 2 (paper number2000GC000080).

SINTON, C. W., DUNCAN, R. A., STOREY, M., LEWIS, J.& ESTARADA, J. J. 1998. An oceanic flood basalt pro-vince within the Caribbean plate. Earth and PlanetaryScience Letters, 155, 221–235.

SISSON, V. B., ERTAN, I. F. & AVE LALLEMANT,H. G. 1997. High pressure (�2000 MPa)


1045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102

glaucophane-bearing pelitic schist and eclogite fromCordillera de la Costa belt, Venezuela. Journal ofPetrology, 38, 65–83.

STEIGER, R. H. & JAGER, E. 1977. Subcommission ongeochronology: convention on the use of decayconstants in geo- and cosmochronology. Earth andPlanetary Science Letters, 36, 359–362.

STERN, R. J. 2002. Subduction zones. Reviews of Geo-physics, 10; doi: 1029/2001RG000108.

STOCKHERT, B., MARESCH, W. V. ET AL. 1995. CrustalHistory of Margarita Island (Venezuela) in detail: con-straint on the Caribbean plate-tectonic scenario.Geology, 23, 787–790.

TASSINARI, C. C. G., MEDINA, J. G. C. & PINTO, M. C. S.1996. Rb–Sr and Sm–Nd geochronology and isotopegeochemistry of central Iberian metassedimentaryrocks (Portugal). Geologie en Mijnbouw, 75, 69–79.

TAYLOR, B. & MARTINEZ, F. 2003. Back-arc basin basaltsystematics. Earth and Planetary Science Letters, 210,280–297.

THOMPSON, P. M. E., KEMPTON, P. D. ET AL. 2004.Elemental, Hf–Nd isotopic and Geochronological con-straints on an island arc sequence associated with theCretaceous Caribbean plateau: Bonaire, Dutch Antil-les. Lithos, 74, 91–116.

TOUSSAINT, J. F. 1996. Evolucion geologica de Colom-bia. Cretacico. Universidad Nacional de Colombia,Medellın.

TRENKAMP, R., KELLOGG, J. N., FREYMUELLER, J. T. &MORA, H. P. 2002. Wide plate margin

deformation, southern Central America and northwes-tern South America. CASA GPS Observations, 15,157–171. Q11

TSCHANZ, C., MARVIN, R., CRUZ, J., MENNERT, H. &CEBULA, E. 1974. Geologic evolution of the SierraNevada de Santa Marta. Geological Society ofAmerica Bulletin, 85, 269–276.

VALLEJO, C., SPIKINGS, R. A., LUZIEUX, L., WINKLER,W., CHEW, D. & PAGE, L. 2006. The early interactionbetween the Caribbean Plateau and the NW SouthAmerican Plate. Terra Nova, 18, 264–269.

VILLAMIL, T. 1999. Campanian–Miocene tectono-stratigraphy, depocenter evolution and basin develop-ment of Colombia and Western Venezuela.Paloegeography, Palaeoclimatology, Paleoecology,153, 239–275.

WINCHESTER, J. A. & FLOYD, P. A. 1977. Geochemicaldiscrimination of different magma series and theirdifferentiation products using immobile elements.Chemical Geology, 20, 325–343.

WRIGHT, J. E. 2004. Aruba and Curacao: remnants of acollided Pacific oceanic plateau? Initial geologicresults from the BOLIVAR Project. American Geophy-sical Union, Fall Meeting 2004, abstract no.T33B-1367.

ZAPATA, G., WEBER, M. ET AL. 2005. Analisis petrogra-fico de las rocas de alta presion de la Serranıa de Jarara,La Guajira y sus implicaciones tectonicas. X CongresoColombiano de Geologıa – Simposio de GeologıaRegional, 68–69.


1103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160

Documents

Weber 2009 Cardona Cabo de La Vela