Miocene to Late Quaternary Patagonian basalts (46–47°S): Geochronometric and geochemical evidence for slab tearing due to active spreading ridge subduction

www.elsevier.com/locate/jvolgeores

Journal of Volcanology and Geotherm

Miocene to Late Quaternary Patagonian basalts (46–478S):Geochronometric and geochemical evidence for slab tearing due

to active spreading ridge subduction

Christele Guivel a,*, Diego Morata b, Ewan Pelleter c,d, Felipe Espinoza b,

Rene C. Maury c, Yves Lagabrielle e, Mireille Polve f,g, Herve Bellon c, Joseph Cotten c,

Mathieu Benoit c, Manuel Suarez h, Rita de la Cruz h

a UMR 6112 bPlanetologie et GeodynamiqueQ, Universite de Nantes, 2 rue de la Houssiniere, 44322 Nantes, Franceb Departamento de Geologıa. Fac. Cs. Fısicas y Matematicas, Universidad de Chile, Plaza Ercilla 803, Santiago, Chile

c UMR 6538 bDomaines oceaniquesQ, UBO-IUEM, place Nicolas-Copernic, 29280 Plouzane, Franced CRPG-CNRS UPR A2300, BP 20, 54501 Vandoeuvre-les-Nancy, France

e UMR 5573, Dynamique de la Lithosphere, Place E. Bataillon, case 60, 34095, Montpellier Cedex 5, Francef LMTG-OMP, 14 Avenue E. Belin, 31400 Toulouse, France

g IRD-Departamento de Geologia de la Universidad de Chile, Chileh Servicio Nacional de Geologıa y Minerıa, Avda. Santa Marıa 0104, Santiago, Chile

Received 18 May 2005; received in revised form 29 August 2005; accepted 14 September 2005

Abstract

Miocene to Quaternary large basaltic plateaus occur in the back-arc domain of the Andean chain in Patagonia. They are thought

to result from the ascent of subslab asthenospheric magmas through slab windows generated from subducted segments of the South

Chile Ridge (SCR). We have investigated three volcanic centres from the Lago General Carrera–Buenos Aires area (46–478S)located above the inferred position of the slab window corresponding to a segment subducted 6 Ma ago. (1) The Quaternary Rıo

Murta transitional basalts display major, trace elements, and Sr and Nd isotopic features similar to those of oceanic basalts from the

SCR and from the Chile Triple Junction near Taitao Peninsula (e.g., (87Sr/86Sr)o=0.70396–0.70346 and qNd=+5.5�+3.0). We

consider them as derived from the melting of a Chile Ridge asthenospheric mantle source containing a weak subduction

component. (2) The Plio-Quaternary (b3.3 Ma) post-plateau basanites from Meseta del Lago Buenos Aires (MLBA), Argentina,

likely derive from small degrees of melting of OIB-type mantle sources involving the subslab asthenosphere and the enriched

subcontinental lithospheric mantle. (3) The main plateau basaltic volcanism in this region is represented by the 12.4–3.3-Ma-old

MLBA basalts and the 8.2–4.4-Ma-old basalts from Meseta Chile Chico (MCC), Chile. Two groups can be distinguished among

these main plateau basalts. The first group includes alkali basalts and trachybasalts displaying typical OIB signatures and thought to

derive from predominantly asthenospheric mantle sources similar to those of the post-plateau MLBA basalts, but through slightly

larger degrees of melting. The second one, although still dominantly alkalic, displays incompatible element signatures intermediate

between those of OIB and arc magmas (e.g., La/NbN1 and TiO2b2 wt.%). These intermediate basalts differ from their strictly

alkalic equivalents by having lower High Field Strength Element (HFSE) and higher qNd (up to +5.4). These features are

consistent with their derivation from an enriched mantle source contaminated by ca. 10% rutile-bearing restite of altered oceanic

crust. The petrogenesis of the studied Mio-Pliocene basalts from MLBA and MCC is consistent with contributions of the subslab

0377-0273/$ - s

doi:10.1016/j.jvo

* Correspondin

E-mail addre

al Research 149 (2006) 346–370

ee front matter D 2005 Elsevier B.V. All rights reserved.

lgeores.2005.09.002

g author.

ss: [email protected] (C. Guivel).

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 347

asthenosphere, the South American subcontinental lithospheric mantle and the subducted Pacific oceanic crust to their sources.

However, their chronology of emplacement is not consistent with an ascent through an asthenospheric window opened as a

consequence of the subduction of segment SCR-1, which entered the trench at 6 Ma. Indeed, magmatic activity was already

important between 12 and 8 Ma in MLBA and MCC as well as in southernmost plateaus, i.e., 6 Ma before the subduction of the

SCR-1 segment. We propose a geodynamic model in which OIB and intermediate magmas derived from deep subslab

asthenospheric mantle did uprise through a tear-in-the-slab, which formed when the southernmost segments of the SCR collided

with the Chile Trench around 15 Ma. During their ascent, they interacted with the Patagonian supraslab mantle and, locally, with

slivers of subducted Pacific oceanic crust that contributed to the geochemical signature of the intermediate basalts.

D 2005 Elsevier B.V. All rights reserved.

Keywords: slab window; slab tear; plateau basalts; alkali basalts; ridge subduction; Patagonia

1. Introduction

Neogene and Quaternary magmatic activity in the

Patagonian Andes displays numerous specific features

which can be related to the subduction of the seg-

mented South Chile Ridge (SCR) beneath the South

American plate. During the last 15 Ma, the location of

this ridge subduction (the Chile Triple Junction, CTJ)

migrated northwards as a result of the oblique colli-

sion between the Chile ridge and the South American

margin (Herron et al., 1981; Cande and Leslie, 1986;

Cande et al., 1987; Nelson et al., 1994; Bangs and

Cande, 1997; Tebbens and Cande, 1997; Tebbens et

al., 1997). The present location of the CTJ, ca. 50 km

north of the Taitao Peninsula (Fig. 1A), is marked by

near-trench magmatic activity (Forsythe and Nelson,

1985; Forsythe et al., 1986, 1995; Lagabrielle et al.,

1994, 2000; Bourgois et al., 1996; Guivel et al., 1999,

2003) and a corresponding gap in the Andean calc-

alkaline volcanic belt between the southern part of the

Southern Volcanic Zone (SSVZ, 41815V–468S) and the

Austral Volcanic Zone (AVZ, 49–548S) (Stern et al.,

1990; Ramos and Kay, 1992). East of the Andean

chain, the Patagonian back-arc domain is characterised

by numerous Neogene basaltic plateaus (Mesetas), the

emplacement of which does not seem to be connected

either with back-arc extension or with a topographic

swell or hotspot track (Ramos and Kay, 1992). Nu-

merous authors (Ramos and Kay, 1992; Kay et al.,

1993; Gorring et al., 1997, 2003; D’Orazio et al.,

2000, 2001, 2003; Gorring and Kay, 2001) have pro-

posed that these basaltic magmas were produced by

melting of subslab asthenospheric mantle upwelling

through slab windows generated from subducted

ridge segments (Dickinson and Snyder, 1979; Thor-

kelson, 1996; Murdie and Russo, 1999). Especially,

Gorring et al. (1997) and Gorring and Kay (2001)

pointed out that the spatial distribution, ages and

chemistries of the Neogene basaltic plateaus of South-

ern Argentina fit apparently with the locations of

asthenospheric windows which opened successively

when segments of the Chile ridge bounded by large

fracture zones (FZ) were subducted. Fig. 1B shows

that the subduction of these various segments, accord-

ing to their magnetic anomaly patterns, started at ca.

15–14 Ma (SCR-4, south of Desolacion FZ), 14–13

Ma (SCR-3, south of Madre de Dios FZ), 12 Ma

(SCR-2, south of Esmeralda FZ), 6 Ma (SCR-1, be-

tween Esmeralda and Tres Montes FZ), 3 Ma (SCR0,

between Tres Montes and Taitao FZ) and finally 0.3

Ma (SCR1, north of Taitao FZ), respectively (Cande

and Leslie, 1986; Forsythe et al., 1986).

In this paper, we test this model using new geochro-

nometric (K–Ar) and geochemical (major, trace element

and Sr and Nd isotopic data) on basalts from the Lago

General Carrera–Buenos Aires area (46–478S) in south-

ern Patagonia. This area is located at the latitude of the

present Chile Triple Junction position (Fig. 1B), along

the Chile–Argentina border, south of Mt. Hudson, the

southernmost active volcano of the SSVZ. As shown in

Fig. 1B, it overlies the SCR-1 slab window present

position inferred from magnetic anomalies (Cande

and Leslie, 1986; Tebbens et al., 1997; Lagabrielle et

al., 2000). Three Miocene to Quaternary basaltic com-

plexes are exposed in this area on both sides of the

Argentina/Chile border (Fig. 1C): Meseta Chile Chico

(Chile) which is capped by a basaltic pile dated back to

8.2–4.4 Ma (Espinoza et al., 2005), Meseta del Lago

Buenos Aires (Argentina) for which available K–Ar

and Ar–Ar ages range from 10.0 to 0.76 Ma (Ton-

That et al., 1999) and 10.1 Ma to b110 ka (Brown et

al., 2004), and finally Rıo Murta (Chile) subglacial

basalts, previously considered Holocene (Demant et

al., 1994, 1998; Corgne et al., 2001). We will show

that the timing and geochemistry of most of these

basaltic eruptive events do not fit with the hypothesis

of their derivation from the subslab asthenospheric

mantle from the SCR-1 fragment, and that alternative

models of opening of asthenospheric windows or tears-

in-the-slab need to be envisioned.

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370348

2. Regional geology

Most authors consider that the Miocene–Recent

evolution of the Patagonian Andes has been controlled

by the oblique northward subduction of the Chile

Ridge beneath the Andean continental margin (Fig.

1A). Plate reconstructions by Cande and Leslie

(1986) indicate that initial ridge collision started at

15–14 Ma at ca. 558S, forming a triple junction (the

Chile Triple Junction, CTJ) between South America,

Fig. 2. Plot of the ages of the southern Patagonian basalts against latitude. Where reported, errors are in 1r. The latitudes and ages of arrival to the

trench of South Chile Ridge segments SCR-3, SCR-2, SCR-1, SCRO and SCR1 are also shown. Sources of previously published ages: Pali Aike

(D’Orazio et al., 2000), Estancia Glencross (D’Orazio et al., 2001), Condor Cliff (Gorring et al., 1997), Meseta de la Muerte (Gorring et al., 1997),

Meseta Central (Gorring et al., 1997), Meseta Belgrano (Gorring et al., 1997), Northeast region (Gorring et al., 1997), Cerro Pampa (in Kay et al.,

1993), Meseta del Lago Buenos Aires (Ton-That et al., 1999, Brown et al., 2004, and this work), Meseta de Chile Chico (Espinoza et al., 2005),

Murta basalts (this work).

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 349

Nazca and Antarctic plates. Then, the CTJ migrated

northwards up to its present position at ca. 468S (Fig.

1B). On the continent, the last major compressive

phase which affected the Patagonian fold and thrust

belt started at ca. 15 Ma (Lagabrielle et al., 2004) and

is generally considered as a consequence of ridge–

trench collision. Then, in the back-arc domain, an

important Neogene magmatic event led to the emplace-

ment of large basaltic plateaus (Mesetas). It started at

ca. 12 Ma, more or less simultaneously with the em-

placement of Cerro Pampa adakites which have been

interpreted as partial melts of the young subducted

Pacific oceanic slab (Kay et al., 1993). Gorring et al.

(1997, 2003) and Gorring and Kay (2001) have dis-

tinguished two stages of building of the basaltic mese-

tas. Thick lava flows, generally tholeiitic, were erupted

during the main plateau stage. Then, after a quiescence

period sometimes several million years long, volcanic

activity resumed, emplacing smaller amounts of post-

plateau basaltic lavas (usually alkali basalts or basa-

nites) richer in incompatible elements than the main

plateau basalts. Post-plateau basalts generally crop out

Fig. 1. Geological setting of the studied volcanic rocks. (A) Simplified tecton

the location of the studied area and the sense of motion (black arrows) of the

numbers are relative velocities in cm/yr (DeMets et al., 1990); (B) simplified

segments of the South Chile Ridge (SCR) indicating the ridge collision age

active ridges (SCR0, SCR-1, SCR-2) (simplified from Guivel et al., 1999; L

South Volcanic Zone (SVZ) and northernmost volcanoes of the Austral Volc

geological map of the Lago General Carrera–Buenos Aires area in Patagonia

and Quaternary volcanic rocks are located (modified from Lagabrielle et al., 2

as strombolian cones, maars and flows filling channels

or paleolandscapes.

The ages of these basaltic plateaus, based on available

K–Ar and Ar–Ar dates, have been plotted against lati-

tude in Fig. 2. The general pattern suggests that mag-

matic activity started between 12 and 8 Ma all along the

back-arc domain of the Patagonian thrust and fold belt,

from 528S to the present position of the CTJ at 468S.Unlike Ramos and Kay (1992), Kay et al. (1993), Gor-

ring et al. (1997, 2003) and Gorring and Kay (2001), we

find no evidence for a trend towards younger ages north-

wards which might be correlated with the chronology of

the subduction of the successive Chile Ridge segments.

Young (Pliocene–Late Quaternary) or relatively young

volcanic activity is evidenced both at ca. 528S (Pali Aike

volcanic field, D’Orazio et al., 2000), 6508S (Camusu

Aike, D’Orazio et al., 2005), 49856.6V8S (Condor Cliff,

Gorring et al., 1997) and near 47–468S in Meseta del

Lago Buenos Aires (Gorring et al., 2003; Brown et al.,

2004) and Rıo Murta (Demant et al., 1998).

The studied area (Lago General Carrera–Buenos

Aires, 46–478S and 70–738W) is located within the

ic sketch map of the present-day Chile Triple Junction (CTJ) showing

Nazca and Antarctic plates with respect to the South American Plate;

map of the CTJ area and location of the fracture zones (FZ) and active

s (grey numbers) and the present-day inferred locations of subducted

agabrielle et al., 2004). Grey triangles: southernmost volcanoes of the

anic Zone (AVZ); empty triangle: Cerro Pampa adakite; (C) simplified

(46–478S) showing the location of the studied areas in which Neogene004). Local geological sketch maps of these areas are shown in Fig. 3.

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370350

modern volcanic arc gap, south of the SSVZ and north of

the AVZ. In this zone, the Miocene to Late Quaternary

magmatic rocks investigated were emplaced over and/or

Fig. 3. Local geological sketch maps of the studied volcanic areas. (A) Rıo M

map); (B) Meseta Chile Chico (MCC) basaltic plateau (simplified from Espin

plateau (simplified from SEGEMAR 1:750,000 map). The whole rock K–A

intruded into Palaeozoic to Plio-Quaternary units (Fig.

3). Strongly deformed Palaeozoic low- to medium-grade

metasediments of the Eastern Andean Metamorphic

urta basalts (modified from SERNAGEOMIN 1:1,000,000 unpublished

oza et al., 2005); (C) Meseta del Lago Buenos Aires (MLBA) basaltic

r ages and the location of dated samples are indicated.

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 351

Complex crop out west and south of the Lago General

Carrera–Buenos Aires (Bell and Suarez, 2000). Middle

to Late Jurassic rhyolitic ignimbrites and lava flows

(with minor andesitic to basaltic intercalated flows) be-

longing to the large silicic Chon Aike Province (Pan-

khurst et al., 1998, 2000; Feraud et al., 1999)

unconformably overlie the metamorphic rocks. These

rocks are referred to as the Ibanez Group in Chile and

the El Quemado Complex in Argentina. Late Jurassic to

Lower Cretaceous marine sedimentary rocks (Coyaique

Group) and subaerial volcanics associated with conti-

nental sedimentary rocks (Divisadero Group), which

overlie diachronically the Ibanez Group (Suarez et al.,

1996), are mostly cropping out in the north of the

studied area and below the basaltic sequences of the

Meseta Chile Chico (Fig. 3B). Cenozoic sedimentary

formations are mainly exposed to the East of the Cor-

Table 1

K/Ar age data. 1: Rio Murta; 2: Meseta Chile Chico, 3: Meseta Lago Buen

Lab.

number

Sample Rock type Loc K2O

(%)

40Ar*

(�10�

6012-8 PG 01a Basaltic lava flow 1 1.00 0.291

6023-9 PG 06a Basalt, pillow lava 1 0.65 0.178

6298-9 PG 102 Basaltic lava flow 1 0.33 0.029

6315-2 PG 102 Basaltic lava flow 1 0.33 0.034

6316-3 PG 102 Basaltic lava flow 1 0.33

6046-7 PG 37a Basaltic plug 2 1.42 2.123

6340-5 PG-138a Basaltic lava flow 2 0.96 1.383

FE01-36a Basaltic lava flow 2 0.99 0.168

CC317-2a Basaltic lava flow 2 1.57 0.468

FE01-23a Basaltic plug 2 1.13 0.348

CC-313a Basaltic lava flow 2 1.33 0.426

6063-7 PG 44 Basaltic lava flow 3 2.78 0.966

6047-8 PG 52 Basaltic lava flow 3 1.20 3.968

5962-7 PG 51 Basaltic lava flow 3 1.77 1.962

6022-8 PG 65 Basaltic lava flow 3 1.23 1.979

6032-9 PG 67 Basaltic lava flow 3 0.58 1.297

6373-4 PG 69 Basaltic lava flow 3 1.14 1.589

6037-4 PG 72 Basaltic lava flow 3 1.74 2.143

6056-8 PG 75 Basaltic lava flow 3 0.85 1.319

6301-3 PG 105 Basaltic dyke 3 1.26 2.659

6297-8 PG 108 Basaltic lava flow 3 1.28 2.414

6312-7 PG 109 Basaltic lava flow 3 1.30 2.366

6287-7 PG 113 Basaltic lava flow 3 0.85 1.603

6288-8 PG 114 Basaltic neck 3 1.10 3.854

6313-8 PG 116 Basaltic lava flow 3 1.13 3.643

6302-4 PG 119 Basaltic lava flow 3 0.97 3.360

6278-6 PG 120 Basaltic lava flow 3 0.86 3.389

6314-1 PG 121 Basaltic lava flow 3 2.25 0.864

6279-7 PG 130 Basaltic lava flow 3 1.96 2.177

6280-8 PG 132 Basaltic lava flow 3 2.38 2.549

6286-6 PG 133 Basaltic lava flow 3 2.25 2.646

6296-7 PG 134 Basaltic lava flow 3 1.76 2.210

6317-4 PG 143 Teschenite 3 1.60 6.396

Analytical method is detailed in the text.a Data from Espinoza et al. (2005).

dillera front, in the Cosmeli basin and below the basal-

tic sequences of the Meseta del Lago Buenos Aires.

They are mostly fluviatile with local marine intercala-

tions and correspond to the Guadal, Galera (in Chile)

and Santa Cruz (in Argentina) Formations (Lagabrielle

et al., 2004). Plio-Quaternary moraines and glacial

deposits are mostly located in the eastern side of the

Lago General Carrera–Buenos Aires. Plutonic rocks are

represented by Meso-Cenozoic subduction-related gran-

itoids of the North Patagonian Batholith (Pankhurst et

al., 1999), and small Late Miocene and Pliocene satellite

plutons (Pankhurst et al., 1999; Suarez and De La Cruz,

2001; Thomson et al., 2001; Morata et al., 2002).

Quaternary volcanic structures belonging to the

SSVZ occur north of the studied area. The Cay,

Maca, Isla Colorada and Hudson volcanoes (Fig.

1B) are the southernmost volcanic centres of the

os Aires

7 cm3/g)

40Ar*

(%)

36Ar

(�10�9 cm3/g)

Weight

(g)

Age (Ma)

10.5 0.844 1.0069 0.90F0.08

8.0 0.694 1.0045 0.85F0.10

0.7 1.434 1.0083 0.27F0.39

1.1 1.921 1.7992 0.32F0.30

1.8001 0.21F0.27

42.7 0.971 1.0064 4.63F0.13

19.9 1.891 1.0033 4.46F0.22

89 4.4F0.8

58 7.6F0.4

59 7.9F0.4

28 8.2F0.5

26.8 0.899 1.0074 1.08F0.04

56.6 1.033 1.0012 10.23F0.26

41.8 0.945 1.0235 3.44F0.10

37.9 1.100 1.0009 4.98F0.15

37.1 0.742 1.0038 6.95F0.24

17.7 2.505 1.0025 4.32F0.23

39.6 1.116 1.0077 3.91F0.11

14.1 2.754 1.0112 4.81F0.32

26.1 2.557 1.0039 6.53F0.25

33.2 1.655 1.0077 5.80F0.19

32.0 1.702 1.0011 5.64F0.19

28.7 1.346 1.0005 5.84F0.21

50.7 1.273 1.0052 10.84F0.28

57.4 0.918 1.0016 9.97F0.25

44.1 1.442 1.0028 10.71F0.29

42.3 1.569 1.0016 12.18F0.34

14.5 1.722 1.0013 1.19F0.08

35.5 1.356 1.0133 3.44F0.11

39.2 1.389 1.0385 3.32F0.10

35.8 1.624 1.0108 3.64F0.11

27.9 1.936 1.0009 3.89F0.14

48.4 2.316 1.0045 12.36F0.33

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370352

SSVZ, and their location is controlled by the dextral

transcurrent Liquine–Ofqui fault zone (Lopez-Escobar

et al., 1995), which originated in response to the

oblique subduction of the Nazca plate beneath South

America (Cembrano et al., 1996). Previous geochemi-

cal data on these volcanoes have been published by

Lopez-Escobar et al. (1993), Demant et al. (1994) and

D’Orazio et al. (2003). The geochemical signatures of

the Cay, Maca and Isla Colorada lavas are typically

calc-alkaline, while the Hudson lavas are comparatively

Table 2

Major and trace element data for Murta basalts

Sample PG01 PG04a PG04b PG05a PG05b PG06a

Lat. 8S 46808V21 46812V58 46812V58 46812V23 46812V23 46812V2Long. 8W 72836V52 72848V23 72848V23 72848V21 72848V21 72848V2wt.%

SiO2 49.60 47.90 48.00 48.00 48.80 48.00

TiO2 2.09 1.22 1.65 1.70 1.18 1.57

Al2O3 17.70 20.00 18.55 18.30 21.30 18.25

Fe2O3* 11.40 8.63 10.28 10.50 7.60 10.22

FeO 8.72 6.60 7.86 8.03 5.81 7.82

MnO 0.19 0.13 0.16 0.17 0.12 0.16

MgO 5.20 7.35 7.22 7.00 5.80 7.50

CaO 8.00 9.50 9.80 9.62 10.40 9.85

Na2O 4.34 3.25 3.70 3.60 3.57 3.50

K2O 1.00 0.58 0.66 0.65 0.53 0.61

P2O5 0.51 0.20 0.28 0.29 0.20 0.28

LOI �0.32 1.26 �0.11 0.28 0.49 0.20

Total 99.71 100.02 100.19 100.11 99.99 100.14

Mg# 51.53 66.50 62.07 60.84 64.01 63.10

ne,- hy 0.65 �0.70 2.95 1.73 0.98 1.75

ppm

Rb 15 9.7 12 12 9 11.5

Sr 480 540 495 485 600 485

Ba 285 88 110 118 90 116

Th 2.2 0.8 1.3 1.3 0.8 0.9

Sc 25 20.8 26.2 28.5 18.7 27.5

V 221 142 190 210 138 200

Cr 56 53 60 64 48 65

Co 29 36 38 38 30 40

Ni 33 70 59 49 53 62

Y 38 19 26 29.5 19 26

Zr 185 106 146 154 101 148

Nb 11 6.5 8.8 9 5.7 8.5

La 22.5 9 11.7 12 8.7 11.2

Ce 52.5 20 29 29.5 19.5 26

Nd 30 12 18 19 11.5 17.5

Sm 6.9 3.1 4.4 4.7 3 4.1

Eu 2.25 1.16 1.5 1.57 1.14 1.48

Gd 6.8 4.2 4.9 5.2 4 4.8

Dy 6.35 3.4 4.65 5 3.3 4.5

Er 3.4 2.1 2.7 3 2 2.7

Yb 3.2 1.84 2.4 2.6 1.85 2.4

Analytical method is detailed in the text.

*Total Fe as Fe2O3; LOI: loss on ignition; Mg#=Mg/(Mg+Fe2+) assum

normative hypersthene (minus sign).

less enriched in large ion lithophile elements (D’Orazio

et al., 2003).

3. Analytical methods

One hundred samples (12 from Rıo Murta, 22

from Meseta Chile Chico, 62 from Meseta del Lago

Buenos Aires) were selected on the basis of their

petrographic freshness (macroscopic and microscop-

ic), low Loss on Ignition (LOI) values and geological

PG07v PG101 PG102 PG104a PG104b PG104c

3 46811V26 46808V41 46808V50 46812V25 46812V25 46812V251 72848V10 72837V01 72837V27 72848V17 72848V17 72848V17

48.00 48.80 48.80 48.00 48.20 47.60

1.36 1.52 1.50 1.71 1.36 1.25

18.00 17.80 17.60 18.00 19.10 19.00

9.45 9.70 9.90 10.52 9.00 9.08

7.23

0.15 0.16 0.16 0.17 0.14 0.14

7.35 7.25 7.52 6.95 7.30 8.10

9.75 10.00 10.40 9.60 10.10 10.00

3.40 3.40 3.54 3.40 3.30 3.15

0.50 0.44 0.42 0.68 0.57 0.51

0.23 0.25 0.25 0.28 0.24 0.23

1.64 0.73 �0.20 0.50 0.27 0.30

99.83 100.05 99.89 99.81 99.58 99.36

64.45 63.53 63.90 60.62 65.40 67.52

0.15 �2.28 1.10 0.37 0.46 0.51

12 12.8 7 11.4 9.7 9

435 475 480 480 522 530

100 98 165 115 101 100

0.7 0.7 0.9 1 0.95 0.7

27 32 34 30 25 22

175 210 215 230 180 165

85 145 165 65 68 57

41 37 36 38 38 42

72 74 74 51 69 84

24.5 26.8 27 28.5 23 21

140 145 150 162 130 120

5.8 5 5.1 9 7.4 6.7

9.7 10.2 10.2 12.7 10.3 9.3

25 26 25.5 30 24 21.5

15.5 15.5 16.8 19.5 15 13.5

3.9 4.1 4.2 4.7 3.8 3.5

1.35 1.54 1.54 1.65 1.4 1.26

4.5 4.8 4.7 5.5 4.4 3.9

4.2 4.7 4.7 5.15 4.1 3.65

2.5 2.7 2.7 2.8 2.3 2

2.3 2.5 2.6 2.7 2.15 1.92

ing Fe2O3/FeO=0.15; ne, -hy=wt.% normative nepheline or wt.%

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 353

position from a set of ca. 150 samples collected

during fieldwork investigations in 1998, 2001 and

2002.

Mineral analyses (available on request to the authors)

were obtained using a five spectrometer Cameca SX-50

electron microprobe (Microsonde Ouest, Brest, France).

Analytical conditions were 10–12 nA, 15 kV, counting

time 6 s. A detailed account of the procedure is given in

Defant et al. (1991). Whole rock 40K–40Ar datings of

Meseta del Lago Buenos Aires and Rıo Murta basalts

(Table 1) were performed on the 0.5- to 0.15-mm-size

fraction after crushing, sieving and cleaning with dis-

tilled water of whole-rock samples. Analyses were car-

ried out at the Laboratoire de Geochronologie,

Universite de Bretagne Occidentale (Brest, France).

One aliquot of sample was powdered for K analysis

by atomic absorption after HF chemical attack and

0.5–0.15-mm grains were used for argon isotopic anal-

yses. Argon extraction was performed by the direct

technique under high vacuum (10�5–10�7 hPa) using

induction heating of a molybdenum crucible. The argon

content was measured by isotope dilution and argon

isotopes were analysed in a 1808 stainless steel mass

spectrometer, according the original procedure de-

scribed by Bellon et al. (1981). Age calculations, fol-

lowing the equation of Mahood and Drake (1982) and

using the Steiger and Jager’s (1977) recommended con-

stants, are given, with 1r error, in Table 1. The K–Ar

ages on Meseta de Chile Chico basalts discussed in this

paper are taken from Espinoza et al. (2005).

Major and trace-element analyses (Tables 2 and 3)

were conducted on agate-ground powders by inductive-

ly coupled plasma–atomic emission spectroscopy (ICP-

AES) except Rb which was determined with flame

atomic emission spectroscopy, at the Universite de

Bretagne Occidentale (Brest, France) and checked

against IWG-GIT standards BE-N, AC-E, PM-S and

WS-E. Relative standard deviation is ca. 1% for SiO2

and 2% for the other major elements except for low

values (b0.50% oxide) for which the absolute standard

deviation is 0.01%. For trace elements, relative standard

deviation is ca. 5% except for concentrations below six

times the detection limit, for which the absolute stan-

dard deviation is about one third of the detection limit.

Detection limits are 2 ppm for Ba, V, Cr, Co, Ni, Zr and

Ce; 1 ppm for Nd, Gd and Er; 0.5 ppm for Rb, Sr, Nb,

La and Sm; 0.3 ppm for Y, Dy and Th; 0.15 ppm for Sc,

Eu and Yb. Specific details for the analytical methods

and sample preparation can be found in Cotten et al.

(1995). The major and trace-element data on Meseta

Chile Chico basalts discussed in this paper are taken

from Espinoza et al. (2005).

Sr and Nd isotopic data were measured on a single

batch of HCl 2N leached whole rock powder; 70 mg of

powder were dissolved in a HNO3–HF mixture from

which Sr and Nd were eluted. Nd was run on a double

Re filament and Sr on a W filament with Ta activator.

Measurements were performed on a Finnigan MAT 261

and Thermo Triton T1mass spectrometers (Universite de

Bretagne Occidentale, Brest). Sr and Nd isotopic ratios

were calculated at t=0 using the 40K–40Ar determined

ages. Nd initial ratios are expressed as qNd. The errors on87Sr/86Sr and 143Nd/144Nd ratios are reported in Table 4.

4. Field data and K–Ar geochronometry

4.1. Rıo Murta basalts (Chile)

These basalts, previously investigated by Demant et

al. (1994, 1998) and Corgne et al. (2001), crop out in

the bottom of the glacial valley of Rıo Murta, dug into

the North Patagonian Batholith granitoids some 30 km

SSE of Hudson volcano (Fig. 3A). Their total preserved

volume is small (b1 km3, Demant et al., 1998). They

occur either as columnar jointed basaltic flows in the

Rıo Murta river bed, eroded down to a few metres by

the stream, or as subglacial and sublacustrine volcanics.

These include pillow lavas and lava tubes up to 3 m in

diameter with glassy chilled margins, as well as hyalo-

clastic breccias associated with varved clays, moraines

and tills. On the basis of field evidence for local sub-

glacial emplacement followed by moderate erosion,

former authors have considered them as Holocene.

However, two out of three K–Ar dates given in Table

1 point out to an emplacement at ca. 0.85–0.9 Ma,

while the third one (b0.5 Ma) obtained on the upper-

most part of a flow occupying the main Rıo Murta

stream bed might be consistent with an Holocene age.

4.2. Meseta Chile Chico (MCC, Chile)

The MCC flood basalt cover (Fig. 3B), up to 900–

1000 m thick, is composed of two sequences (Espinoza

et al., 2005). The basal Lower Basaltic Sequence

(LBS), which unconformably overlies the Mesozoic

Ibanez Group and the Neocomian Cerro Colorado For-

mation, is formed by a 500–550-m-thick pile of Eocene

basaltic lava flows (57–40 Ma, Charrier et al., 1979;

Baker et al., 1981; Petford et al., 1996; Espinoza et al.,

2005) crosscut by several basanitic necks and diatrems.

These Eocene basalts can be correlated with the 57–45

Ma Posadas Basalt (Baker et al., 1981; Kay et al., 2002)

and with the 42F6 Ma Balmaceda Basalts (Baker et

al., 1981; Demant et al., 1996), which crop out east and

Table 3

Selected major and trace element data for Meseta del Lago Buenos Aires volcanic rocks

Sample PG143 PG65 PG67 PG75 PG105 PG113 PG114 PG132 PG133 PG51 PG52 PG69 PG72 PG108

Age 12.4 4.98 6.95 4.81 6.53 5.84 10.84 3.32 3.64 3.45 10.23 4.32 3.91 5.8

Lat. 8S 46838V41 47804V 47804V 47803.602 46843V33 46846V47 46846V33.6 47809V28 47809V30 47810V14 47820V15 47804V 47804.08 294948 UTM

Long. 8W 71828V08 71848V 71848V 71849.141 71842V06 71842V07 71842V47,9VV 71833’25,6VV 71833V19 71832V22 70848V18 71848V 71847.995 4820953 UTM

Type MP-inter MP-inter MP-inter MP-inter MP-inter MP-inter MP-inter MP-inter MP-inter MP-alk MP-alk MP-alk MP-alk MP-alk

wt.%

SiO2 48.50 47.00 47.35 50.00 47.90 48.30 47.15 55.00 51.80 48.80 48.70 47.75 47.80 47.50

TiO2 1.76 1.48 1.43 1.54 1.64 1.61 1.25 1.40 1.68 2.22 2.44 2.16 2.63 2.06

Al2O3 16.55 15.85 15.70 17.20 16.25 16.70 15.60 18.40 16.50 16.00 15.82 17.60 17.75 16.10

Fe2O3* 10.30 11.55 11.42 10.66 10.34 11.60 10.36 8.65 13.10 10.80 12.91 10.77 11.92 11.90

MnO 0.15 0.17 0.18 0.17 0.16 0.17 0.18 0.19 0.23 0.16 0.18 0.16 0.17 0.17

MgO 6.42 8.37 9.40 5.66 8.54 7.15 10.40 2.07 2.63 7.03 4.92 6.28 4.78 7.12

CaO 8.45 9.65 9.60 9.10 8.65 9.60 10.00 5.30 5.75 8.20 8.60 9.05 8.00 9.85

Na2O 3.97 2.80 3.28 3.42 3.70 3.38 2.47 5.85 4.58 4.14 3.56 3.68 3.76 3.54

K2O 1.60 1.21 0.61 0.89 1.48 1.03 1.28 2.60 2.33 1.80 1.16 1.41 1.66 1.42

P2O5 0.43 0.40 0.38 0.29 0.66 0.44 0.41 0.91 1.28 0.68 0.54 0.46 0.53 0.54

LOI 1.41 0.75 �0.11 0.75 0.28 �0.27 0.57 �0.26 �0.52 �0.50 0.39 0.43 0.64 �0.23

Total 99.54 99.23 99.24 99.68 99.60 99.72 99.67 100.11 99.36 99.33 99.22 99.75 99.64 99.97

Mg# 59.23 62.81 65.73 55.30 65.81 58.96 70.06 35.80 31.87 60.27 47.04 57.61 48.31 58.24

ne,-hy 3.68 0.74 1.66 �10.49 3.78 1.09 0.77 1.13 �10.39 4.40 �6.71 3.17 1.40 4.61

ppm

Rb 40.5 27.5 22 17.5 34.5 21.5 39.5 46.5 41 31.5 20.5 23.5 27 23.5

Sr 685 725 708 444 870 555 700 718 534 742 595 688 682 682

Ba 345 310 287 190 455 260 340 965 600 470 440 315 415 320

Th 4.95 3.5 3.6 2.5 5.2 2.5 4 5.8 6.7 3.25 2.8 3.05 3.35 3

Sc 23 29 29 26 22 28 30 10 12 19 23.5 22 21 26

V 200 215 213 198 190 225 240 46 54 174 240 210 240 235

Cr 167 286 320 150 220 181 485 2 2 200 28 42 12 185

Co 35 43 45 35 39 41 45 11 20 38 42 38 36 43

Ni 68 162 182 59 180 95 212 1 2 115 32 50 26 107

Y 23 26 23.7 25 25 25 23 33 44 30 30 24.5 29.5 245

Zr 164 157 149 141 210 149 124 294 500 240 205 205 232 180

Nb 20 11.6 11.4 12 24.5 15.3 11 45 53 37 29 31 36.5 30

La 28 24.5 24 14.5 38.5 21 22 55 61 37 26 28.5 31 28

Ce 56 55 53 34 75 45 47.5 109 124 74 53 60 68 57

Nd 29 34 30 18 36.5 24 26 47 52 36 32 28 33 30.5

Sm 6.3 6.85 6.3 4.4 7 5.6 5.5 9.2 11.7 7.5 7 6.1 7 6.7

Eu 1.86 1.95 1.81 1.42 2.07 1.76 1.64 3.2 3.28 2.26 2.22 1.92 2.15 2.07

Gd 6.2 5.7 5.45 4.3 5.8 5 4.7 7.7 9.9 6.9 7.2 5.9 6.8 6

Dy 4.4 4.6 4.35 4.35 4.7 4.5 4 6.2 8 5.5 5.4 4.6 5.45 4.8

Er 2.2 2.3 2.2 2.25 2.4 2.3 2 3 4 2.7 2.8 2.1 2.6 2.2

Yb 1.9 2.18 2.12 2.12 2.1 2.16 2 2.76 3.8 2.6 2.3 1.95 2.32 1.94

Analytical methods detailed in the text. MP: main plateau; PP: post-plateau; Alk.: alkali; Int.: intermediate. K–Ar ages in Ma.

*Total Fe as Fe2O3; LOI: Loss on ignition; Mg#=Mg/(Mg+Fe2+) assuming Fe2O3/FeO=0,15; ne,- hy=wt.% normative nepheline or wt.% normative hypersthene (minus sign).

C.Guivel

etal./JournalofVolca

nologyandGeotherm

alResea

rch149(2006)346–370

354

Sample PG109 PG116 PG119 PG120 PG130 PG134 PG44 PG121 PG41 PG46 PG50 PG123 PG126 PG127

Age 5.64 9.97 10.71 12.18 3.44 3.89 1.08 1.19

Lat. 8S 294962 UTM 47804V29,2VV 47806V08 47806V13,6VV 47809V32 47810V27 46852V24 47806V28 46841V24 47803V35 47807V42 47803V05 47803V13 47803V35Long. 8W 4820912 UTM 71801V21,1VV 70859V59,6VV 70859V34,5VV 71833V15,3VV 71832V29,4VV 70844V08 70859V16,5VV 70849V48 70846V54 70851V58 71802V55,4VV 71802V50,6 71801V44,4VVType MP-alk MP-alk MP-alk MP-alk MP-alk MP-alk PP-alk PP-alk PP-alk PP-alk PP-alk PP-alk PP-alk PP-alk

SiO2 47.00 47.10 47.20 47.20 49.00 48.50 45.90 46.85 43.50 44.50 46.00 47.80 47.60 48.10

TiO2 2.05 2.36 2.38 2.34 2.71 2.31 2.29 2.56 2.64 2.34 2.45 2.03 2.43 2.42

Al2O3 16.00 16.20 16.25 16.10 17.00 16.75 14.50 17.25 13.15 14.35 14.40 16.65 17.45 16.85

Fe2O3* 11.95 11.95 12.04 12.00 13.20 10.75 10.80 10.75 12.08 12.15 11.50 10.40 11.76 11.35

MnO 0.17 0.16 0.16 0.17 0.20 0.16 0.17 0.16 0.18 0.18 0.17 0.15 0.17 0.15

MgO 7.45 7.03 7.40 7.40 3.85 7.01 8.10 5.37 10.30 10.14 8.80 7.05 4.54 5.87

CaO 9.60 9.55 9.32 9.90 6.85 7.72 8.15 9.70 10.60 9.40 9.42 7.42 9.55 6.70

Na2O 3.48 3.54 3.80 3.63 4.35 4.16 5.00 4.04 3.50 3.55 3.65 4.65 3.67 5.19

K2O 1.35 1.14 1.17 1.10 2.04 1.85 2.90 2.22 2.31 1.90 2.28 2.13 1.27 2.33

P2O5 0.50 0.44 0.43 0.42 0.86 0.62 1.50 0.74 1.05 0.69 0.86 0.79 0.41 0.83

LOI 0.13 0.54 �0.59 �0.26 �0.67 0.06 0.29 �0.19 �0.05 0.08 �0.26 0.51 1.14 �0.27

Total 99.68 100.01 99.56 100.00 99.39 99.89 99.60 99.45 99.26 99.28 99.27 99.58 99.99 99.52

Mg# 59.23 57.82 58.89 58.97 40.47 60.31 63.61 53.79 66.52 66.04 64.07 61.24 47.36 54.65

ne,-hy 4.43 3.47 5.02 4.79 1.59 4.50 16.26 9.67 14.71 11.22 9.67 9.02 2.79 10.55

Rb 22 14.3 13.8 12.9 36 33 46.5 40 37 33 43 31.5 15.2 35

Sr 685 600 590 582 578 761 1290 855 1020 790 870 810 615 885

Ba 285 235 215 220 480 442 715 560 755 525 670 390 250 460

Th 3 1.75 1.8 1.6 4.6 3.7 9.4 6.1 7.4 5 6.15 3.3 2.15 4.8

Sc 25 23 24 25 20 19 14.5 25 28 22.5 24 17 25 13.3

V 225 232 235 255 175 165 153 240 265 225 240 150 200 130

Cr 190 205 220 240 2 175 225 83 345 270 240 140 100 98

Co 42 43 44 48 27 37 42 32 52 51 49 37 42 35

Ni 124 75 86 91 6 114 190 42 215 215 170 134 44 87

Y 24 23.5 22.5 22.5 37.5 29 28 27.5 27 22.5 25 25 25.5 25

Zr 185 159 152 158 395 252 420 267 300 225 275 252 167 268

Nb 29 24 24 23.5 48 38 98 61 76 62 67 38 21 60

La 27.5 19 18.4 17.5 47 36 90 47 67 45 54 32 18.3 42

Ce 54 39 40 38 94 73 153 89 117 85 97 65 40 81

Nd 29.5 22.5 22.5 22 45.5 34 63 40.5 56 39.5 46 32 22.5 38.5

Sm 6.55 5.5 5.25 5.2 10 7.3 11 8.15 9.4 7.5 8.5 6.8 5.15 7.9

Eu 2.05 1.82 1.8 1.78 2.92 2.25 3.2 2.41 2.91 2.23 2.45 2.14 1.83 2.53

Gd 5.6 5.1 5.2 5.1 9 6.7 8.25 7 7.4 6.4 6.8 5.9 5.8 6.9

Dy 4.7 4.3 4.35 4.2 7.15 5.6 5.6 5.25 5.35 4.6 5.05 4.75 4.8 5.15

Er 2.1 2.1 2 2 3.5 2.7 2.6 2.4 2.5 2.2 2.25 2.2 2.2 2

Yb 1.88 1.75 1.74 1.69 3.25 2.46 1.8 2.04 1.87 1.6 1.8 1.97 2.09 1.81

C.Guivel

etal./JournalofVolca

nologyandGeotherm

alResea

rch149(2006)346–370

355

Table 4

Sr and Nd isotopic data

Sample Loc. SiO2 Age Mg# TiO2 La/Nb (87Sr/86Sr)0 2 s (143Nd/144Nd) 2r qNd

PG102 1 48.80 0.27 63.90 1.50 2.00 0.703532 4 0.512920 10 5.50

PG01a 1 49.60 0.90 51.53 2.09 2.05 0.703958 3 0.512792 8 3.00

PG06a 1 48.00 0.85 63.10 1.57 1.32 0.703460 5 0.512916 7 5.42

IBA47a 1 46.75 65.00 1.51 0.703590 0.512900 5.11

FE01-36b 2 46.01 4.40 1.42 2.03 0.704140 0.512873 4.58

PG65 3 47.00 4.98 62.81 1.48 2.11 0.704109 11 0.512910 10 5.31

PG75 3 50.00 4.81 55.30 1.54 1.21 0.704260 10 0.512751 11 2.20

PG105 3 47.90 6.53 65.81 1.64 1.57 0.704364 3 0.512795 10 3.06

PG113 3 48.30 5.84 58.96 1.61 1.37 0.704362 4 0.512730 9 1.79

PG114 3 47.15 10.84 70.06 1.25 2.00 0.704215 6 0.512821 7 3.57

PG108 3 47.50 5.80 58.24 2.06 0.93 0.704367 5 0.512736 8 1.91

PG109 3 47.00 5.64 59.23 2.05 0.95 0.704394 5 0.512724 7 1.68

PG116 3 47.10 9.97 57.82 2.36 0.79 0.703996 7 0.512799 9 3.14

Analytical methods are detailed in the text. International standards (NBS 987, La Jolla) are run regularly on both mass spectrometers. Typical values

are (1) Triton T1 87Sr/86Sr=0.710250F12, 143Nd/144Nd=0.511850F6 and (2) 87Sr/86Sr=0.710251F16 143Nd/144Nd=0.512104F6 (JNdi).

Average blanks for this study are 0.2 for Nd and 0.4 ng for Sr.

Loc.: location in Fig. 1 (1=Rıo Murta; 2=Meseta Chile Chico; 3=Meseta del Lago Buenos Aires).

Ages from Table 1; geochemical values from Tables 2 and 3.a Data from Demant et al. (1998).b Data from Espinoza et al. (2005).

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370356

north of the MCC area, respectively. The rather flat

upper surface of the MCC, which covers ca. 300 km2,

corresponds to the 400-m-thick Upper Basaltic Se-

quence (UBS) of Miocene–Pliocene tabular basaltic

lava flows and necks which provided K–Ar ages of

8.2, 7.9, 7.6, 4.6, 4.5 and 4.4 Ma (Espinoza et al.,

2005). They are underlain by two rhyolitic flows

dated at 13.1 and 9.8 Ma, respectively. On the basis

of their age range and field features, they can be

considered equivalent to the main plateau basaltic se-

quence (Gorring et al., 2003) of the Meseta del Lago

Buenos Aires (MLBA) documented below. Very young

cones and associated flows do not occur on the MCC,

which therefore has apparently not undergone any mag-

matic event equivalent to the post-plateau volcanic

phases of MLBA and other Patagonian plateaus.

4.3. Meseta del Lago Buenos Aires (MLBA, Argentina)

This Meseta (Fig. 3C) is one of the largest (ca. 6000

km2) basaltic plateaus in the Patagonian back-arc do-

main. Its main plateau sequence (Gorring et al., 1997,

2003) is composed of an up to 300-m-thick pile of

tabular basaltic lava flows overlying the Miocene

molasse sediments of the Rıo Zeballos Group (Santa

Cruz Formation). Previous K–Ar and 40Ar/39Ar ages of

these main plateau basalts (Sinito, 1980; Baker et al.,

1981; Mercer and Sutter, 1982; Ton-That et al., 1999)

range from 10 to 4.5 Ma and more recently Brown et al.

(2004) obtained three 40Ar/39Ar isochron ages of 10.12,

7.86 and 7.71 Ma on lavas from this sequence. Some of

these flows are interbedded with glacial tills (Ton-That

et al., 1999). The MLBA post-plateau lavas have been

studied in detail by Gorring et al. (2003) and dated by

Brown et al. (2004) and Singer et al. (2004). They

erupted from more than 150 small monogenetic volca-

nic centres (strombolian and spatter cones and very

well-preserved maars). The flows form a volcanic pile

usually ca. 100 m thick topping the MLBA, and many

of them poured down its eastern slopes. Brown et al.

(2004) published 31 isochron 40Ar/39Ar ages ranging

from 3.3 to less than 0.1 Ma for these post-plateau

MLBA basalts. Different volcanic pulses have been

recognised by these authors at 3.2–3.0 Ma, 2.4 Ma, 1.7

Ma, 1.35 Ma, 1.0 Ma, 750 ka, 430–330 ka, and finally,

b110 ka. Other 40Ar/39Ar and K–Ar ages obtained by

Singer et al. (2004) from lavas interbedded with moraine

deposits range from 1016F10 ka to 109F3 ka.

We have measured 22 new K–Ar ages (Table 1 and

Fig. 3C) on MLBA basalts: 2 on the post-plateau lavas

(1.19F0.08 Ma and 1.08F0.04 Ma) and 20 on the

main plateau sequence. These latter ages allow to ex-

tend the known range of the main plateau activity from

12.18F0.34 Ma (PG 120) to 3.32F0.10 Ma (PG 132)

and to identify a quiescence period between ca. 10 and

7 Ma.

The oldest volcanic pile crops out, as noticed by

previous authors, along the southeastern border of the

plateau, and especially along the trail from Estancia La

Vizcaina to Laguna del Sello. Three samples from

tabular lava flows from this cross section gave ages

of 12.18F0.34 Ma (PG-120, 1000 m), 10.71F0.29

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 357

Ma (PG-119, 1050 m) and finally 9.97F0.25 Ma (PG-

116, 1170 m) from bottom to top, respectively. This

sequence is unconformably overlain by younger lava

flows (sample PG-121, 940 m, 1.19F0.08 Ma old) of

the post-plateau sequence (Fig. 3C) which poured out

down its slopes towards the plain near Estancias La

Vizcaina and Casa de Piedra.

A good cross section of the top of the main plateau

sequence is exposed in the southern border of the

Meseta along the Hacienda El Ghio horse-trail near

Rıo Torrentoso. Four samples of tabular flows from

this cross section provided K–Ar ages ranging from

3.89F0.14 Ma (sample PG-134, 980 m) to 3.32F0.10

Ma (sample PG-132, 1455 m). A 40-m-thick glacial till

unit is interbedded between this latter sample and two

slightly older flows dated at 3.44F0.11 Ma and

3.64F0.11 Ma, respectively. These data suggest that

the main plateau stage of the MLGA volcanism ended

at 3.3 Ma, and thus, that there was no significant time

gap separating it from the post-plateau stage which

started at 3.3 Ma according to Brown et al. (2004).

In addition, they provide a rather precise dating for one

of the Pliocene glacial events already documented in

the area by Ton-That et al. (1999).

On the western border of the Meseta, near Estancia

Los Corrales located 38 km south of Los Antiguos

Fig. 4. Total alkali–silica classification diagram (Le Maitre et al., 1989) for th

anhydrous basis. The heavy line represents the boundary between alkaline a

Murta basalts. Open triangles: intermediate main plateau lavas from Meseta

MCC; Open squares: intermediate main plateau lavas from Meseta del Lag

from MLBA. Black diamonds: post-plateau lavas from MLBA. Data on M

Espinoza et al. (2005).

along the Paso Roballos road, the lowest lava flows of

the main plateau sequence gave ages of 5.84F0.21

Ma (PG 113), 5.80F0.19 Ma (PG 108) and

5.64F0.19 Ma (PG 109), respectively. These flows

overlie the Rıo Zeballos Group Miocene molasse

which is crosscut by an older basaltic neck dated at

10.84F0.28 Ma (PG 114).

Finally, a few mafic rocks cropping out away from

MLBA also provided K–Ar ages consistent with those

of the main plateau building stage. One of them, a

columnar jointed hypovolcanic intrusion locally re-

ferred to as a btescheniteQ body, located north of the

MLBA near Estancia Las Chicas, has been dated back

to 12.40F0.33 Ma (sample PG 143) and a basaltic lava

flow 5 km east of Bajo Caracoles, near the SE edge of

the MLBA, at 10.23F0.26 Ma (sample PG 52).

5. Petrologic and geochemical data

5.1. Sample classification

A large majority of the studied rocks are petrograph-

ically fresh and their Loss On Ignition (LOI) values

range from slightly negative to ca. 1 wt.%. According

to the TAS diagram shown in Fig. 4, they are mostly

basaltic (basalts, basanites and trachybasalts) although a

e Miocene to Late Quaternary igneous rocks recalculated to 100 wt. %,

nd subalkaline series (Irvine and Baragar, 1971). Open diamonds: Rıo

Chile Chico (MCC). Black triangles: alkaline main plateau lavas from

o Buenos Aires (MLBA). Black squares: alkaline main plateau lavas

eseta Chile Chico (MCC) intermediate and primitive lavas are from

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370358

few samples plot in the fields of basaltic trachyandesite

and trachyandesite. We have classified MLBA samples

from our data set according to their main plateau or

post-plateau position, based on their field relationships

and K–Ar ages. For the latter, we have postulated that

the transition between the main plateau and post-pla-

teau stages occurred abruptly at 3.3 Ma (Brown et al.,

2004 and our data as discussed above). As discussed

above, no post-plateau activity can be identified in

MCC, where all the Mio-Pliocene lavas crop out either

as a tabular flow pile or as necks and dykes crosscutting

them (Espinoza et al., 2005).

A further discrimination has been operated within

our data set. Indeed, basaltic lavas from both the

MLBA and the MCC display chemical features very

similar to those considered typical of Ocean Island

Basalts (OIB), as already shown by previous authors

(Hawkesworth et al., 1979; Baker et al., 1981; Ramos

and Kay, 1992). However, Stern et al. (1990), Gorring

et al. (1997, 2003), Gorring and Kay (2001) and Espi-

noza et al. (2005) have shown that some Patagonian

main plateau and post-plateau basalts display weak to

moderate bsubduction-relatedQ geochemical imprints

traduced by relative depletions in High Field Strength

Elements (HFSE) vs. Large Ion Lithophile Elements

Fig. 5. Selected plots of major (recalculated to 100 wt.%, anhydrous basis)

TiO2–Mg#; (B) TiO2–La/Nb; (C) Nb–Mg#; (D) normative nepheline–Mg#.

(LILE) and Light Rare-Earth Elements (LREE), and

sometimes as well by specific isotopic signatures.

These features are clearly observed in our MLBA and

MCC data set, in which about 30% of the samples

depart from the usual compositional range of OIB by

displaying La/Nb ratios greater than unity (up to 3.7)

and TiO2 contents usually lower than 2 wt.% (Fig. 5A

and B). Their relative depletion in Nb is unlikely to

result from fractionation of Ti-magnetite during differ-

entiation because their Nb contents increase with de-

creasing Mg numbers (Mg#, Fig. 5C). Consequently,

we have identified by specific symbols in all the geo-

chemical diagrams the MLBA and MCC samples char-

acterised by La/NbN1. Stern et al. (1990) termed these

btransitionalQ basalts, but we will rather use this word todescribe the Rıo Murta basalts which plot in between

the fields of alkalic and subalkalic basalts in most

geochemical diagrams. Therefore, the btransitionalQbasalts of Stern et al. (1990) will be referred to here

as bintermediateQ samples (i.e., intermediate between

OIB and subduction-related lavas), as opposed to the

other samples, termed bcratonicQ basalts by Stern et al.

(1990) and which display typical OIB characteristics

(including La/Nbb1). The petrogenesis of the

bintermediateQ lavas will be discussed separately below.

elements, major element parameters, and trace elements (in ppm). (A)

Symbols as in Fig. 4.

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 359

5.2. Major elements and petrographic types

Most of the studied basalts are silica-undersaturated

and contain up to 22 wt.% normative nepheline (Fig.

5D), while a minority (including 8 over 12 Rıo Murta

samples) is hypersthene-normative. In the TAS dia-

gram, the MLBA and MCC samples plot consistently

above Irvine and Baragar’s (1971) limit between alka-

lic and subalkalic compositions while Murta samples

spread around this limit. The Murta basalts also fit

Middlemost’s (1975) requirements for the definition of

transitional basalts, i.e., they plot in the subalkalic

field in the K2O–SiO2 diagram and in the alkalic

field in the Na2O–SiO2 diagram. They will thus be

termed transitional. All MCC Mio-Pliocene mafic

lavas are alkali basalts, trachybasalts or basaltic tra-

chyandesites (Espinoza et al., 2005; Figs. 4 and 5)

whether or not they display bintermediateQ La/Nb

ratios (N1). MLBA post-plateau basalts contain usual-

ly 8–22 wt.% normative nepheline (Fig. 5D) and plot

in the fields of basanites and tephrites in the TAS

diagram (Fig. 4). MLBA main plateau basalts, inter-

mediate or not, are merely alkali basalts, the CIPW

norms of which contain either small amounts of neph-

eline (less than 5%) or of hypersthene (Fig. 5D and

Table 3). The TAS diagram of Fig. 4 also shows that

the MLBA-MCC alkali basaltic/basanitic samples dis-

playing La/Nbb1 and TiO2N2 wt.% plot consistently

in the basalt/basanite/trachybasalt fields while the in-

termediate samples are basalts, trachybasalts or basal-

tic trachyandesites. The basalts range from primitive

(Mg#N65, MgON8 wt.%) to evolved (Fig. 5A, C, D)

and the patterns of major element variations vs. Mg#

(not shown) suggest the occurrence of fractionation

effects involving olivine, clinopyroxene, plagioclase

and titanomagnetite. Normative nepheline contents

tend to decrease with Mg# (Fig. 5D), a pattern often

observed in alkali basalt series.

In short, the distribution of petrographic types within

the studied sample set is relatively simple: Rıo Murta

lavas are transitional basalts while all the other ones are

of dominant alkali affinity. Among the latter, two

groups can be distinguished: a genuine alkali one

with typical OIB geochemical signatures (La/Nbb1

and TiO2N2 wt.%), which will be referred to as the

balkali groupQ and a second one (the bintermediate

groupQ) displaying incompatible element signatures in-

termediate between those of OIB and arc lavas (La/

NbN1 and TiO2b2 wt.%). The alkali group includes

the post-plateau MLBA lavas, all of which are strongly

silica-undersaturated (basanites and tephrites), together

with alkali basalts and trachybasalts from the main

plateau MLBA stage and from MCC. The intermediate

group is only represented in MLBA (main plateau) and

MCC by alkali basalts, trachybasalts, basaltic trachyan-

desites and trachyandesites. It is noteworthy that main

plateau lavas from both Meseta Chile Chico and Lago

Buenos Aires are silica-undersaturated (alkali basalts

and basanites), while most of other Neogene Patago-

nian main plateaus are principally made up of tholeiitic

basalts (Gorring and Kay, 2001).

5.3. Mineral chemistry

The Murta basalts are porphyritic, with plagioclase

(An73–48Ab26–50Or1–2), olivine (Fo87–72) and minor au-

gite (Wo46En42Fs12) phenocrysts set in a microcrystal-

line groundmass of plagioclase (An68Ab31Or1), olivine

(Fo82–73), Ti-magnetite and acicular quenched Ti-rich

augite (Wo50En25Fs25). Megacrysts of plagioclase (up

to 2 cm in size) and clinopyroxene are frequent in some

samples.

The Mio-Pliocene MCC basaltic lava flows, plugs

and dykes (Espinoza et al., 2005) bear low amounts

of olivine (Fo83–65) and minor clinopyroxene (Wo45En40–46Fs8–15) phenocrysts set into a microlitic to

subdoleritic groundmass containing olivine (Fo75–54),

augite (Wo49–43En44–37Fs13–14), plagioclase (An68–66)

and Ti-magnetite.

The MLBA main plateau basalts contain plagioclase

(An55–70), clinopyroxene (Wo49En42Fs9) and olivine

(Fo86–62) phenocrysts set into a microlitic groundmass

containing plagioclase (An66), olivine (Fo81–60), rare

clinopyroxene, Ti-magnetite and glass. The MLBA

post-plateau basalts are very fresh vesicular aphyric to

moderately porphyritic lava flows, with occasional pla-

gioclase (An66–72), clinopyroxene (Wo52En38Fs10) and

olivine (Fo83–73) phenocrysts. Their glassy to microlitic

groundmass contains plagioclase, olivine (Fo69), clin-

opyroxene (Wo51En40Fs9), Ti-magnetite and glass.

Corroded quartz xenocrysts, always rimmed by small

clinopyroxene aggregates, are often present in basaltic

lava flows from MCC and MLBA, indicating that they

experienced some extent of crustal contamination.

The compositions of calcic clinopyroxene pheno-

crysts from MCC and MLBA are typical of those

from alkaline intraplate basalts. They are Ti- and Al-

rich and plot consistently within the alkalic fields in

Leterrier et al.’s (1982) diagrams (not shown).

5.4. Trace-element features

Compatible element contents of the Murta basalts

are consistently low (Table 2). Those of MLBA lavas

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370360

(Table 3), as well as those from MCC basalts, range

from concentrations close to those of near-primitive

basaltic magmas (Co=45–55 ppm, Ni=200–220 ppm,

Cr=300–450 ppm) to very low ones. They decrease

rather abruptly with Mg# (diagrams not shown), a

feature consistent with the occurrence of olivine and

clinopyroxene fractionation effects. Incompatible trace-

element abundance patterns normalised to the compo-

Fig. 6. Primitive mantle normalised (Sun and McDonough, 1989) incompa

Ridge (bold patterns, samples D34-1 and D42-4; Klein and Karsten, 1995) an

Lopez-Escobar et al., 1993) are shown for comparison; (B) intermediate MC

main plateau lavas; (D) alkaline MCC lavas (analyses in Espinoza et al., 20

MLBA lavas.

sition of Primitive Mantle (Sun and McDonough, 1989)

are shown in Fig. 6.

Patterns of selected Murta transitional basalts have

been plotted in Fig. 6a together with those of represen-

tative basaltic samples from the Chile Ridge (Klein and

Karsten, 1995) and from Hudson volcano (Lopez-Esco-

bar et al., 1993). Murta basalts display slightly LILE-

and LREE-enriched patterns (average primitive mantle-

tible multi-element patterns. (A) Rıo Murta transitional basalts, Chile

d Hudson volcano basalts (dashed patterns, samples Hud-1 and Hud-3;

C lavas (analyses in Espinoza et al., 2005); (C) intermediate MLBA

05); (E) alkaline main plateau MLBA lavas; (F) alkaline post-plateau

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 361

normalised (La/Sm)N=1.75) and are slightly but signif-

icantly depleted in Nb relative to K and La. These

features are very similar to those of Chile Ridge seg-

ment 3 sample D42-4 (Klein and Karsten, 1995) dis-

playing subduction-related geochemical affinities.

Sample PG01a is slightly richer in all trace elements

but Sr than other less differentiated Murta samples.

However, its incompatible trace-element pattern is par-

allel to the others, a feature consistent with olivine and

plagioclase fractionation.

Incompatible trace-element patterns of intermediate

main plateau basalts from MCC and MLBA are shown

is Fig. 6B and C, respectively. Both groups of interme-

diate basalts share the same trace-element characteris-

tics. They display LILE- and LREE-enriched patterns

(average primitive-mantle normalised (La/Sm)N=2.72

for MCC basalts and (La/Sm)N=3.00 for MLBA

basalts) and are noticeably but variably depleted in

Nb. This depletion in Nb seems to be attenuated with

increasing differentiation, as Nb contents in intermedi-

ate basalts increase when Mg# decreases (Fig. 5C). This

feature implies that the more or less pronounced Nb

depletion of the intermediate basalts is unlikely to be

Fig. 7. (A) Plot of (87Sr/86Sr)o against qNd for the studied lavas. White di

Demant et al., 1998); white squares: intermediate basalts from MLBA; blac

qNd for the studied lavas and isotopic ratios from other Patagonian magmatic

Buenos Aires (Gorring et al., 2003) and the main plateau sequences from Pata

Pampa adakites from Kay et al. (1993); Pacific MORB from Peate et al. (199

(CTR) (Guivel et al., 2003); and white circles: heterogeneous South Chil

modelling, two mixing curves have been calculated (ticks every 10%). [A]: m

and adakitic melt derived from slightly altered oceanic crust (black star); [B

slightly altered mid-ocean ridge basalt (black star); (see Table 5 for model

related to differentiation processes (e.g., fractionation of

Ti-magnetite) but is pristine and linked to the nature of

their source. It is worth to note that the range of La/Nb

ratios found in intermediate main plateau basalts (MCC

and MLBA) is far greater (Fig. 5B) than the one (La/

Nb*b1.3; Nb*=17�Ta) found by Gorring et al.

(2003) in some post-plateau alkali basalts.

Main plateau alkali basalts from both MCC and

MLBA show trace-element patterns typical of OIB

(Fig. 6D and E). They display levels of LILE and

LREE enrichment similar to those of intermediate

basalts (average (La/Sm)N ratios of 3.09 and 2.90,

respectively) but lack the negative Nb anomalies typical

of the latter.

Post-plateau alkali basalts from MLBA display

very homogeneous trace element compositions. Like

the MLBA main plateau alkali basalts, they show

OIB-like trace element patterns (Fig. 6F) with an

average (La/Sm)N=3.93, slightly higher than that of

the main plateau lavas. They display a wide range of

LREE concentrations ((La/Yb)N from 6.3 to 37) with

almost constant HREE contents (YbN ranging from

3.25 to 4.26).

amonds: basalts (b1 Ma) from Rıo Murta (grey diamond, data from

k squares: alkali basalts from MLBA. (B) Plot of (87Sr/86Sr)o against

rocks: fields of the alkaline post-plateau (b1 Ma) basalts from Meseta

gonian Basaltic Field as defined by Gorring and Kay (2001), the Cerro

7); black circles: basaltic andesites from the Chile Trench Taitao Ridge

e Ridge (SCR) basalts (Klein and Karsten, 1995). For geochemical

ixing model between mantle source similar to that of the alkali basalts

]: mixing model between similar mantle source than in model [A] and

parameters and see text for details).

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370362

5.5. Sr and Nd isotopic data

The isotopic compositions of the studied Mio-Plio-

cene magmatic rocks are listed in Table 4 and plotted

in Fig. 7A and B. Published data from Murta basalts

as well as additional isotopic ratios from other Pata-

gonian magmatic rocks are also plotted on this figure.

Initial strontium isotopic ratios (87Sr/86Sr)o are scat-

tered between 0.70346 and 0.70439 and corresponding

(143Nd/144Nd)o between 0.51291 and 0.51272. The

isotopic compositions closest to Mid-Oceanic Ridge

Basalts (MORB) are those from Murta transitional ba-

salts ((87Sr/86Sr)o=0.70346–0.70353; (143Nd/144Nd)o=

0.512916�0.512920; qNd=+5.4�+5.5), with the ex-

ception of sample PG-01a ((87Sr/86Sr)o=0.70396;

(143Nd/144Nd)o=0.512792; qNd=+3.0) which is more

differentiated as shown by its Mg#=51.5 (Table 2) and

is also characterised by a strong negative Nb anomaly

(La/Nb=2, Table 2). The Murta samples plot within

the mantle array. Their signature is clearly different

from those of Cerro Pampa adakites, but rather

similar to that of basaltic andesites from the Chile

Trench Taitao Ridge (Guivel et al., 2003). Main

plateau MLBA alkali basalts plot in the same

field of Fig. 7B that MLBA post-plateau basalts

(Gorring et al., 2003). In this diagram, MLBA

intermediate basalts define a sub-vertical trend root-

ed in the former field and evolving outside this

field towards higher 143Nd/144Nd at rather constant87Sr/86Sr ratios.

6. Discussion

6.1. A subslab asthenospheric origin for the Rıo Murta

basalts

Rıo Murta transitional basalts have trace elements

patterns very similar to some Hudson volcano lavas

(Lopez-Escobar et al., 1993) and also to those of

Enriched MORB (E-MORB) from the anomalous

segment 3 of the SCR (Klein and Karsten, 1995)

(Fig. 6A). These E-MORB from segment 3 of the

SCR show trace-element patterns very similar to

those of convergent margin magmas. This feature

has been interpreted as reflecting contamination of

the SCR mantle source by various amounts of either

oceanic sediments and altered oceanic crust or melts/

fluids derived from (Klein and Karsten, 1995). The

Sr/Nd isotopic ratios of Rıo Murta transitional basalts

overlap those of the very heterogeneous SCR basalts

(Fig. 7B). These features allow us to infer that their

source may be identical to the mantle source of the

SCR basalts. As the SCR-1 ridge segment which

entered the trench 6 Ma ago is thought to be pres-

ently located beneath the studied area (Fig. 1B), the

Murta basalts may represent melts from the subslab

SCR-1 asthenospheric mantle passing through the

slab window, as envisioned by previous authors

(Demant et al., 1998; Corgne et al., 2001). Their

chemical similarities with Hudson magmas might

suggest that this window (or its zone of influence)

extended northwards to the edge of the SSVZ, as

already pointed out by D’Orazio et al. (2003). The

specific geochemical features of sample PG-01a, i.e.,

lower Mg#, strong negative Nb anomaly and respec-

tively more (Sr) and less (Nd) radiogenic isotopic

signature compared to the others, might reflect crust-

al contamination effects of subslab magmas, likely

by the Patagonian Batholith through which they

ascended.

6.2. Origin of the MLBA and MCC main and post-

plateau alkali basalts and basanites

The petrogenesis of Plio-Pleistocene basalts from

MLBA has been previously studied by Hawkes-

worth et al. (1979), Baker et al. (1981) and Gor-

ring et al. (2003). These works have demonstrated

the highly alkaline affinity of these basalts (nephe-

line-normative basanites and alkali basalts) with a

strong OIB-like geochemical signature and relatively

enriched Sr/Nd isotopic ratios (87Sr/86Sr= 0.7041–

0.7049; 143Nd/144Nd=0.51264–0.51279). These fea-

tures have been interpreted by Gorring et al. (2003)

as consistent with their derivation from an OIB-like

source involving the deep subslab asthenospheric

mantle, together with a contribution of the enriched

subcontinental lithospheric mantle (predominantly EM

I type). The study of peridotitic xenoliths within Pata-

gonian basalts (Rivalenti et al., 2004) has documented

a lithospheric mantle metasomatised by asthenosphere-

derived alkali basaltic melts. Our new isotopic data on

main plateau MLBA alkali basalts plot within the

previously determined field of Patagonian basalts

(Fig. 7B). Although we basically agree with the inter-

pretations of Gorring et al. (2003), we did not find any

unquestionable evidence for a specific geochemical

imprint of the subslab asthenosphere opposed to that

the supraslab (i.e., mantle wedge) enriched astheno-

sphere proposed by Stern et al. (1990). Moreover, our

data do not allow us to ascribe the origin of the EM I

signature to either the Patagonian lithospheric mantle

or to an heterogenous subslab or supraslab astheno-

spheric mantle.

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 363

The alkali basalts from MLBA (main plateau) and

the basanites from MLBA (post-plateau; Gorring et al.,

2003) display rather similar Sr and Nd isotopic ratios

(Fig. 7A and B and Table 4), suggesting their derivation

from a single (or similar) type(s) of enriched mantle

source(s). However, the variable slopes of their incom-

patible trace element patterns (Fig. 6D–F) are consistent

with variable degrees of partial melting of such a source

(Gorring et al., 2003), the lowest ones corresponding to

the post-plateau MLBA basanites which show the high-

est La/Yb ratios. Luhr et al. (1995) have plotted near-

primitive basalts in a La/Yb vs. Yb diagram to docu-

ment partial melting degrees of an enriched (lherzolitic)

mantle source with variable contributions of spinel and

garnet. In Fig. 8, where we have used the source

composition proposed by these authors (La=1.79

ppm and Yb=0.31 ppm), the position of post-plateau

MLBA basanites is consistent with 1.5–5% melting of a

source in which garnet is slightly more abundant than

spinel. Main plateau MCC and MLBA alkali lavas

would derive from somewhat larger (5–10%) melting

degrees of a less garnet-rich source. However, these

calculations are rather dependent from the assumed

composition of the source. For instance, using the

source composition proposed by Gorring and Kay

(2001), i.e., La=0.885 ppm and Yb=0.423 ppm,

leads to obtain very low melting degrees (0.1–2%) for

post-plateau MLBA basanites, and lower ones (2–5%)

for the main plateau lavas (diagram not shown).

Fig. 8. Plot of La/Yb against Yb for primitive samples (Mg# N63) of our data

1970) of garnet and spinel lherzolite sources (modified from Luhr et al., 199

ppm. The mode of the garnet lherzolite is taken as ol/opx/cpx/gt=60:25:9

proportions entering the melt are taken as ol/opx/cpx/gt or sp=10:20:65:5. P

as 0.0002:0.002:0.069:0.01:0.002 and 0.0015:0.049:0.28:4.1:0.007 for the m

6.3. Origin of the MLBA and MCC intermediate lavas

Alkali basalts and intermediate lavas from both

MCC and MLBA (main plateau) display roughly sim-

ilar chemical characteristics and tend to plot along the

same trends in some diagrams. Because of the present

lack of isotopic data on MCC basalts, the following

discussion will be focused on MLBA (main plateau)

intermediate lavas.

The MLBA intermediate lavas erupted synchro-

nously with the MLBA alkali basalts, and both types

have roughly similar compositions in major elements

(with the exception of TiO2) and trace elements (with

the exception of Nb). The intermediate lavas display-

ing the highest Nd isotopic ratios are also charac-

terised by high Mg# and strong depletions in Nb

and Ti (Fig. 9).

These features, together with the trend they define

with MLBA alkali basalts in Fig. 7, could suggest that

their genesis was controlled by a mixing process be-

tween a component related to the alkali basalts (or their

mantle source) and a bcontaminantQ characterised by a

relatively unradiogenic Sr isotopic signature similar to

that of the alkali basalts but with higher Nd isotopic

ratios (above the mantle array) and a selective depletion

in Ti and Nb.

Mature continental crust and oceanic sediments, al-

though depleted in Ti and Nb, have Sr isotopic signa-

tures much more radiogenic than required for the

set. Symbols as in Fig. 4. Results of non-modal batch melting (Shaw,

5) are shown. Composition of the source: La=1.79 ppm and Yb=0.31

:6 and that of spinel lherzolite as ol/opx/cpx/sp=58:30:10:2. Phase

artition coefficients for La and Yb were selected from literature values

inerals ol/opx/cpx/gt/sp, respectively.

Fig. 9. Plots of Mg#, TiO2, La/Nb and Nb against qNd. Symbols as in Fig. 4.

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370364

bcontaminantQ (Andean continental crust and Leg ODP

site 141 Chile Trench sediments, 87Sr/86Sr=0.715 and

=0.708, respectively; Stern and Kilian, 1996) and can

thus be discarded as potential candidates. Contamina-

tion by adakitic magmas may also be envisioned, as

their composition matches the required trace element

and isotopic features, providing the adakitic component

derived from the melting of slightly altered oceanic

crust with a Sr isotopic signature slightly higher than

that of MORB. The addition of up to 10% adakitic melt

(isotopic and trace-element compositions given in Table

5) to the mantle source of MLBA alkali basalts does not

fit the Sr, Nd isotopic trend of the intermediate basalts

(curve A in Fig. 7B). Moreover, the trace-element

patterns of basaltic melts derived from such a blend

are inconsistent with the observed ones (Fig. 10A and

Model A in Table 5), as they display strong positive Sr

anomalies and Yb depletion typical of adakites, which

do not exist in the intermediate basalts.

Thus, we have tested another model (model B)

involving the mixing between up to 10% of a

slightly altered oceanic crust (87Sr/86Sr=0.70375;143Nd/144Nd=0.5131) and an alkali basalt mantle

source similar to that considered in the former model.

This mixing model (parameters and results given in

Table 5), similar to the former one but for the compo-

sition of the contaminant, accounts for the isotopic

signature of intermediate MLBA lavas (curve B in

Fig. 7). In addition, the trace-element features of the

basaltic magma derived from such a mix fits with the

trace elements patterns of the intermediate lavas except

for the lack of negative Nb anomalies (Fig. 10B and

Model B in Table 5). Thus, other processes must be

envisioned to explain these anomalies.

The negative correlation between Nb and Mg#

observed for intermediate lavas (Fig. 5C) suggests

that the negative Nb anomaly is attenuated when

differentiation progresses. This feature implies that

Nb depletion with respect to adjacent incompatible

elements is pristine and linked to the mantle source

of the intermediate basalts. It could correspond to the

bweak subduction componentQ detected in ultramafic

xenoliths of southern Patagonian basalts by Rivalenti

et al. (2004). However, this Nb depletion is unlikely to

represent an overall feature of the asthenospheric

mantle, as it does not occur in the MLBA post-plateau

basanites and main plateau alkali basalts. The origin of

this signature could thus be either lithospheric or

Table 5

Results of mixing calculations and compositions of the corresponding end-members

1 2 3 4 5

Mantle source Adakitic melt MORB Model A Model B

Rb 2.20 – 4.44 – 23.95

Ba 28.50 306.00 27.80 560.95 283.51

Th 0.30 4.90 0.49 7.59 3.18

Nb 2.90 11.73* 2.30 37.12 27.87

La 2.75 26.60 3.55 48.97 26.99

Ce 5.40 60.90 10.45 101.51 54.74

Sr 68.50 1,886.00 131.00 2,232.71 666.91

Nd 2.95 30.30 9.26 47.58 29.97

Sm 0.66 4.41 3.06 7.74 6.73

Eu 0.21 1.16 1.10 2.12 2.08

Gd 0.56 – 4.27 – 6.21

Dy 0.47 – 4.95 – 5.33

Y 2.40 – 29.30 – 29.49

Er 0.21 – 2.97 – 2.48

Yb 0.19 0.72 2.90 1.06 2.0187Sr/86Sr 0.704394 0.703636 0.703636 0.703823 0.704261143Nd/144Nd 0.512724 0.513069 0.513069 0.5129079 0.5128132

qNd 1.68 8.41 8.41 5.26 3.42

(1) Mantle source composition (PG109/10); (2) adakitic melt (RB5 from Cerro Pampa, Nb*=17�Ta (Kay et al., 1993); (3) MORB composition

(sample D20-1) from SCR1 (Klein and Karsten, 1995); (4) model A: 10% batch partial melt of 90% of (1) mixed with 10% of (2); (5) model B: 10%

batch partial melt of 90% of (1) mixed with 10% of (3). A constant bulk source mode of 0.58 olivine, 0.275 opx, 0.095 cpx, 0.015 gt and 0.035

spinel was used. Partition coefficients used in calculations are from Gorring and Kay (2001). Isotopic compositions for adakitic melt and MORB are

those of slightly altered oceanic crust (DSDP/ODP sites 417/418, flow 300; Staudigel et al., 1995).

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 365

shallow asthenospheric, and related to a blocalQ—i.e.,

not widespread—component derived from slightly al-

tered oceanic crust. Thorkelson (1996) and Thorkelson

and Breitsprecher (2005) have shown that the slab

edges of an asthenospheric window are able to melt

at depth, generating adakitic magmas and leaving

restite fragments which may become long-term resi-

dents of the continental lithospheric mantle. However,

if the restite becomes entrained in the asthenosphere, it

may then undergo partial melting. Furthermore, as Nb

concentrations in intermediate basalts are in average

lower than in the alkali basalts (26 vs. 38 ppm), Nb

should be retained in some residual mineral during

these processes. Rutile is the best candidate as it

concentrates only Ti, Nb and Ta and is commonly

observed as a residual phase during partial melting of

oceanic basalts under P–T conditions consistent with

those of hot subduction zones (Ringwood, 1990;

Foley et al., 2000; Schmidt et al., 2004). Amphibole

and/or phlogopite could also be considered, but their

occurrence in the restite should be detectable from the

behaviour of incompatible elements other than Nb and

Ti. Thus, we propose that the origin of the Ti- and

Nb-depleted intermediate MLBA basalts could be

linked to the contribution to their source of rutile-

bearing restites of partially melted oceanic crust

from the edges of the asthenospheric window, incor-

porated either in the shallow asthenospheric or the

deep lithospheric Patagonian mantle.

6.4. Tectonic setting of MLBA and MCC: successive

opening of ridge-derived asthenospheric windows?

As discussed above, the petrogenetic features of the

studied Mio-Pliocene basalts from MLBA (main pla-

teau) and MCC are consistent with possible contribu-

tions of the deep subslab asthenosphere, the South

American subcontinental lithospheric and astheno-

spheric mantle and the subducted oceanic crust to

their magma sources. In the slab window opening

model previously developed by Ramos and Kay

(1992), Kay et al. (1993), Gorring et al. (1997, 2003)

and Gorring and Kay (2001), melting may occur at the

boundary between the ascending subslab asthenosphere

and the overlying subcontinental lithosphere, with oc-

casional contributions of the downgoing basaltic crust

or of its melting products (Cerro Pampa adakites).

Alternatively, subslab-derived melts may interact with

the subcontinental mantle or with the oceanic crust at

the slab window edges during their ascent towards the

surface.

However, a critical aspect of the slab window model

of Gorring, Kay and co-authors is that they consider

that several windows developed successively beneath

Fig. 10. Primitive mantle normalised (Sun and McDonough, 1989)

incompatible multi-element patterns of compositions obtained from

batch partial melting (Shaw, 1970) of mantle sources derived from

mixing calculations (Table 5); (A) model A and (B) model B. A

constant bulk source mode of 0.58 olivine, 0.275 opx, 0.095 cpx,

0.0.15 gt and 0.035 spinel was used. Patterns of main plateau MLBA

intermediate samples with Mg# N59 are shown for comparison. See

text for details.

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370366

Patagonia during the last 15 Ma, each segment of the

SCR stopping its activity and thus developing its own

slab window after colliding with the Chile trench. Such

a process should result into a younging northward

pattern of plateau building (Fig. 1) that Gorring et al.

(1997) claim to have identified from their age data.

Presently available literature ages, combined with our

own K–Ar results, are plotted against latitude in Fig. 2.

No clear age decrease is observed from South to North,

especially regarding the onset of magmatic activity, and

no connection between the ages and the timing of

arrival of the various SCR segments to the Chile trench

is easily identifiable. Magmatic activity seems to begin

between ca. 12 and 8 Ma for nearly all the dated

volcanic centres including Estancia Glencross (528S),Mesetas Belgrano, Central and de la Muerte, the North-

east volcanic region and finally MLBA and MCC (46–

478S). Moreover, a phase of relative paucity of volcanic

activity seems to have occurred around 7 Ma (except

possibly in MLBA), followed by a new pulse starting at

ca. 5–4 Ma in many volcanic centres. Thus, the chro-

nology of emplacement of the Patagonian plateau and

post-plateau lavas seems hardly consistent with their

ascent through asthenospheric windows derived from

the successive segments of the SCR. Especially, the

early magmas of MLBA (main plateau stage) and

MCC cannot have ascended through the asthenospheric

window generated by the segment SCR-1 which en-

tered the trench at 6 Ma only and is now thought to be

located beneath the studied area. Indeed, the post-pla-

teau lavas of MLBA (b3.3 Ma) and the Quaternary

basalts of Rıo Murta may have ascended through this

window (Gorring et al., 1997, 2003; Demant et al.,

1998; Corgne et al., 2001). In short, discrepancies

between the ages of emplacement of Patagonian basal-

tic plateaus and the age of the subduction of ridge

segments lead us to reconsider the modalities and tim-

ing of slab window opening beneath Patagonia.

6.5. Proposition of a new tectonic model involving slab

tearing linked to spreading ridge collision

As discussed above, emplacement of thick alkali

basalt sequences started at ca. 12 Ma, i.e., after the

tectonic phase that built up the Cordillera of the south-

ern Patagonian Andes. Indeed, the Late Miocene basal-

tic flows are roughly horizontal and always appear to

post-date the main folding and thrusting event. This

contractional phase ended with a major tectonic event

recognized at the scale of the entire southern Patagonia,

characterised by thrusting of the pre-Cenozoic rocks of

the Cordillera over the frontal Oligocene–Miocene ma-

rine to continental molasse (Lagabrielle et al., 2004).

This phase occurred necessarily after 16.3 Ma (youn-

gest known age of the mammal fauna of the continental

molasse) and before 12–10 Ma (the age of the basal

flows of the alkali plateau basalts). It is also recorded

by fission track analysis east of the present-day topo-

graphic divide where rapid cooling and denudation

ceased between 12 and 8 Ma (Thomson et al., 2001).

It must be noted that the initiation of the subduction of

the Chile Ridge at 15–14 Ma in southern Patagonia

coincides with this last main contractional phase that

affected the entire Cordillera. A period of very rapid

erosion and peneplanation followed the tectonic uplift,

producing a relatively flat surface on which the alkali

basalts were emplaced.

Considering the ages at which segments SCR1,

SCR0, SCR-1 and SCR-2 entered the trench (0.3, 3, 6

and 12 Ma Fig. 1B), and assuming that the Patagonian

plateau and post-plateau basaltic magmas originated

from a mantle that ascended through a slab window,

it becomes clear that this (or these) window(s) opened

well before the subduction of the corresponding ridge

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 367

segments. This has to be taken into account when

attempting to link plateau basalt emplacement to the

opening of such a window. For this reason, we favor a

model based on the process of slab tearing at depth

when collision starts at the trench (van den Beukel,

1990) and leading ultimately to slab breakoff (e.g.,

Davies and von Blanckenburg, 1995; von Blancken-

burg and Davies, 1995; Mason et al., 1998). In such a

model, slab tear would start when strong tectonic cou-

pling in the forearc occurs before the subduction of the

ridge axis itself. This situation may occur when a series

of large spreading ridge segments approaches the

trench, that is for the segments SCR-4 to SCR-2, all

of which arrived in less than 3 Ma within the subduc-

tion zone (15–14 Ma for SCR-4, 14–13 Ma for SCR-3,

and 12 Ma for SCR-2). Slab tear may then have prop-

agated towards the north into slightly older lithosphere

emplaced at segment SCR-1. The lack of correlation

between latitude and onset of volcanic activity (Fig. 2)

suggests that this propagation was very rapid: indeed,

Fig. 11. Cartoon showing the main stages of the proposed tectonic

model of slab tearing during ridge collision at the trench.

volcanic activity started earlier in MLBA than in Estan-

cia Glencross (528S).A tentative sketch of our proposed tectono-magmatic

model of Patagonian plateau basalt emplacement is

shown in Fig. 11. As compression occurs in the Cordil-

lera leading to active orogenesis, tension forces are

applied to the descending slab that will break off be-

neath the continental plate. A slab tearing all along the

active Patagonian margin results ultimately into the

detachment and sinking of the deep part of the sub-

ducted plate. OIB-type magmas would be generated

by the partial melting of the subslab asthenospheric

mantle uprising through the tear-in-the-slab, possibly

near its boundary zone with the overlying continental

lithospheric mantle (Coulon et al., 2002). Alternatively,

heat transfer through the tear might have induced partial

melting of the supraslab mantle (Davies and von

Blanckenburg, 1995). Then, the ascending magmas

would interact with the Patagonian lithospheric mantle

and locally with altered Pacific crust from the edges of

the slab tear (intermediate magmas) and ultimately be

emplaced in the back-arc domain between ca. 528 and468S. Finally, for all segments, after slab tear, the

spreading axis will enter the trench. The final stages of

this evolution will correspond to the opening of btrueQ(ridge-related) slab windows. The latter are responsible

for the genesis and ascent of the most recent basaltic

magmas, i.e., in the northern and eastern Mesetas (in-

cluding MLBA post-plateau phase) and Rıo Murta dur-

ing a second magmatic pulse starting at ca. 5–4 Ma.

7. Conclusions

1. The Quaternary Rıo Murta transitional basalts dis-

play obvious geochemical similarities to the SCR

and CTJ oceanic basalts. We consider them as de-

rived from the melting of a Chile Ridge astheno-

spheric mantle source containing a weak subduction

component. Their position above the inferred loca-

tion of the slab window corresponding to the SCR-1

segment subducted 6 Ma ago is consistent with a

slab window opening model previously developed

by Ramos and Kay (1992), Kay et al. (1993), Gor-

ring et al. (1997, 2003) and Gorring and Kay (2001).

2. Two groups may be identified among the main

plateau basalts of MLBA and MCC. The first one

includes alkali basalts and trachybasalts displaying

typical OIB signatures and thought to derive from

the melting of OIB-type mantle sources involving

the deep subslab asthenosphere and the subcontinen-

tal mantle, as previously shown by Gorring et al.

(1997, 2003).

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370368

3. The second group of samples, although dominantly

alkalic, displays incompatible element signatures

intermediate between those of OIB and arc magmas

(e.g., La/NbN1 and TiO2b2 wt.%). These interme-

diate basalts differ from their alkalic equivalents by

their HFSE-depleted character and their higher qNd(up to +5.4). We ascribe these specific features to

their derivation from an enriched mantle source

contaminated by ca. 10% rutile-bearing restite of

altered oceanic crust, likely derived from the edges

of a slab window or slab tear.

4. The chronology of emplacement of main plateau

basalts from MLBA (12.4–3.3 Ma) and MCC

(8.2–4.4 Ma) is inconsistent with their origin from

an asthenospheric window opened as a consequence

of the subduction of the Chile Ridge segment SCR-1

which entered the trench at 6 Ma. This fact allows us

to question the model developed by Gorring et al.

(1997) and Gorring and Kay (2001), in which the

Neogene basaltic magmas of Southern Argentina

plateaus ascended through asthenospheric windows

which opened successively when segments SCR-4,

SCR-3, SCR-2 and finally SCR-1 of the Chile

ridge were subducted. In our preferred geodynamic

model, OIB and intermediate magmas of MLBA

and MCC, as well as those of other Patagonian

plateaus (Mesetas Belgrano, Central, de la Muerte

and the Northeast volcanic region) originated from

deep asthenospheric mantle uprising through a tear-

in-the-slab subparallel to the trench, which formed

when the southernmost segments of the SCR col-

lided with the Chile Trench around 15 Ma.

Acknowledgements

This research was supported by the cooperation

program ECOS-Sud ACU01 and was part of the Chi-

lean FONDECYT Project 1000125 and French DyETI

project 2004–2005. We thank Drs. M. D’Orazio and C.

Stern for their pertinent and helpful reviews of the

manuscript. Fieldwork assistance of Leonardo Zuniga

(Pituso) is acknowledged.

References

Baker, P.E., Rea, W.J., Skarmeta, J., Caminos, R., Rex, D.C.,

1981. Igneous history of the Andean Cordillera and Patago-

nian Plateau around latitude 468S. Philos. Trans. R. Soc. Lond.,A 303, 105–149.

Bangs, N.L., Cande, C., 1997. Episodic development of a convergent

margin inferred from structures and processes along southern

Chile margin. Tectonics 16, 489–503.

Bell, M., Suarez, M., 2000. The Rıo Lacteo formation of southern

Chile. Late Paleozoic orogeny in the Andes of southernmost

South America. J. South Am. Earth Sci. 13, 133–145.

Bellon, H., Quoc Buu, N., Chaumont, J., Philippet, J.C., 1981.

Implantation ionique d’argon dans une cible support: application

au tracage isotopique de l’argon contenu dans les mineraux et les

roches. C. R. Acad. Sci., Paris 292, 977–980.

Bourgois, J., Martin, H., Lagabrielle, Y., Le Moigne, J., Frutos Jara,

J., 1996. Subduction erosion related to spreading-ridge subduc-

tion: Taitao Peninsula (Chile margin triple junction area). Geology

24, 723–726.

Brown, L.L., Singer, B.S., Gorring, M.L., 2004. Paleomagnetism and40Ar/39Ar chronology of lavas from Meseta del Lago Buenos

Aires, Patagonia. Geochem. Geophys. Geosyst. 5 (1), Q01H04.

doi:10.1029/2003GC000526.

Cande, S.C., Leslie, R.B., 1986. Late Cenozoic tectonics of the

Southern Chile Trench. J. Geophys. Res. 91, 471–496.

Cande, S.C., Leslie, R.B., Parra, J.C., Hobart, M., 1987. Interaction

between the Chile ridge and the Chile trench: geophysical and

geothermal evidence. J. Geophys. Res. 92, 495–520.

Cembrano, J., Herve, F., Lavenu, A., 1996. The Liquine–Ofqui fault

zone: a long-lived intra-arc fault system in southern Chile. Tecto-

nophysics 259, 55–66.

Charrier, R., Linares, E., Niemeyer, H., Skarmeta, J., 1979. K–Ar ages

of basalt flows of the Meseta Buenos Aires in southern Chile and

their relation to the southeast Pacific triple junction. Geology 7,

436–439.

Corgne, A., Maury, R., Lagabrielle, Y., Bourgois, J., Suarez, M.,

Cotten, J., Bellon, H., 2001. La diversite des basaltes de Patagonie

a la latitude du point triple du Chili (468–478 lat. S): donnees

complementaires et implications sur les conditions de la subduc-

tion. C. R. Acad. Sci., Paris 333, 363–371.

Cotten, J., Le Dez, A., Bau, M., Caroff, M., Maury, R.C., Dulski, P.,

Fourcade, S., Bohn, M., Brousse, R., 1995. Origin of anomalous

rare-earth element and yttrium enrichments in subaerially ex-

posed basalts: evidence from French Polynesia. Chem. Geol. 119,

115–138.

Coulon, C., Megartsi, M., Fourcade, S., Maury, R.C., Bellon, H.,

Louni-Hacini, A., Cotten, J., Coutelle, A., Hermitte, D., 2002.

Post-collisional transition from calc-alkaline to alkaline volcanism

during the Neogene in Oranie (Algeria): magmatic expression of a

slab breakoff. Lithos 62, 87–110.

Davies, J.H., von Blanckenburg, F., 1995. Slab breakoff: a model

of lithosphere detachment and its test in the magmatism and

deformation of collisional orogens. Earth Planet. Sci. Lett. 129,

85–102.

Defant, M.J., Richerson, M., de Boer, J.Z., Stewart, R.H., Maury,

R.C., Bellon, H., Drummond, M.S., Feigenson, M.D., Jackson,

T.E., 1991. Dacite genesis via both slab melting and differentia-

tion: Petrogenesis of La Yeguada Volcanic Complex, Panama.

J. Petrol. 32, 1101–1142.

Demant, A., Herve, F., Pankhurst, R.J., Magnette, B., 1994. Alkaline

and calc-alkaline Holocene basalts from minor volcanic centres in

the Andes of Aysen, Southern Chile. VII Congreso Geologico

Chileno, Concepcion, Actas vol. II, pp. 1326–1330.

Demant, A., Herve, F., Pankhurst, R., Suarez, M., 1996. Geochemis-

try of early Tertiary back-arc basalts from Aysen, southern Chile

(44–468S): geodynamic implications. Third International Sympo-

sium on Andean Geology (ISAG), St-Malo, pp. 17–19.

Demant, A., Belmar, M., Herve, F., Pankhurst, R.J., Suarez, M., 1998.

Petrologie et geochimie des basaltes de Murta: une eruption sous-

glaciaire dans les Andes patagoniennes (468 lat. S.). Relation avec

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370 369

la subduction de la ride du Chili. C. R. Acad. Sci., Paris 327,

795–801.

DeMets, C., Gordon, R.G., Argus, D.F., Stein, S., 1990. Current plate

motions. Geophys. J. Int. 101, 425–478.

Dickinson, W.R., Snyder, W.S., 1979. Geometry of triple junctions

related to San Andreas transform. J. Geophys. Res. 84 (B2),

561–572.

D’Orazio, M., Agostini, S., Mazzarini, F., Innocenti, F., Manetti, P.,

Haller, M., Lahsen, A., 2000. The Pali Aike Volcanic Field,

Patagonia: slab-window magmatism near the tip of South Amer-

ica. Tectonophysics 321, 407–427.

D’Orazio, M., Agostini, S., Innocenti, F., Haller, M., Manetti, P.,

Mazzarini, F., 2001. Slab window-related magmatism from south-

ernmost South America: the Late Miocene mafic volcanics from

the Estancia Glencross Area (6528S, Argentina–Chile). Lithos57, 67–89.

D’Orazio, M., Innocenti, F., Manetti, P., Tamponi, M., Tonarini, S.,

Gonzalez-Ferran, O., Lahsen, A., Omarini, R., 2003. The Quater-

nary calc-alkaline volcanism of the Patagonian Andes close to the

triple junction: geochemistry and petrogenesis of volcanic rocks

from the Cay and Maca volcanoes (6458S, Chile). J. South Am.

Earth Sci. 16, 219–242.

D’Orazio, M., Innocenti, F., Manetti, P., Haller, M., Di Vincenzo,

Tonarini, S., 2005. The Late Pliocene mafic lavas from the

Camusu Aike Volcanic Field (~508S, Argentina): evidences for

geochemical variability in slab window magmatism. J. South Am.

Earth Sci. 18, 107–124.

Espinoza, F., Morata, D., Pelleter, E., Maury, R.C., Suarez, M.,

Lagabrielle, Y., Polve, M., Bellon, H., Cotten, J., de la Cruz, R.,

Guivel, C., 2005. Petrogenesis of the Eocene and Mio-Pliocene

alkaline basaltic magmatism in Meseta Chile Chico, Southern

Patagonia, Chile: evidence for the participation of two slab win-

dows. Lithos 82 (3–4), 315–343.

Feraud, G., Alric, V., Fornari, M., Bertrand, H., Haller, M., 1999.40Ar–39Ar dating or the Jurassic volcanic province of Patagonia:

migrating magmatism related to Gondwana break-up and subduc-

tion. Earth Planet. Sci. Lett. 172, 83–96.

Foley, S.F., Barth, M.G., Jenner, G.A., 2000. Rutile/melt partition

coefficients for trace elements and an assessment of the influence

of rutile on the trace element characteristics of subduction zone

magmas. Geochim. Cosmochim. Acta 64, 933–938.

Forsythe, R.D., Nelson, E., 1985. Geological manifestation of ridge

collision: evidence from the Golfo de Penas–Taitao basin, south-

ern Chile. Tectonics 4, 477–495.

Forsythe, R.D., Nelson, E.P., Carr, M.J., Kaeding, M.E., Herve, M.,

Mpodozis, C., Soffia, J.M., Harambour, S., 1986. Pliocene near-

trench magmatism in southern Chile: a possible manifestation of

ridge collision. Geology 14, 23–27.

Forsythe, R.D., Meen, J.K., Bender, J.F., Elthon, D., 1995. Geochem-

ical Data of Volcanic Rocks and Glasses Recovered from Site 862:

Implications for the Origin of the Taitao Ridge, Chile Triple

Junction Region, Proc. ODP, Sci. Results, vol. 141. Ocean Dril-

ling Program, College Station, TX, pp. 331–348.

Gorring, M., Kay, S., 2001. Mantle processes and sources of Neogene

slab window magmas from Southern Patagonia, Argentina.

J. Petrol. 42, 1067–1094.

Gorring, M., Kay, S., Zeitler, P., Ramos, V., Rubiolo, D., Fernandez,

M., Panza, J., 1997. Neogene Patagonian plateau lavas: continen-

tal magmas associated with ridge collision at the Chile Triple

Junction. Tectonics 16, 1–17.

Gorring, M., Singer, B., Gowers, J., Kay, S., 2003. Plio-Pleistocene

basalts from the Meseta del Lago Buenos Aires, Argentina: evi-

dence for asthenosphere–lithosphere interactions during slab win-

dow magmatism. Chem. Geol. 193, 215–235.

Guivel, C., Lagabrielle, Y., Bourgois, J., Maury, R., Fourcade, S.,

Martin, H., Arnaud, N., 1999. New geochemical constraints for

the origin of ridge-subduction-related plutonic and volcanic suites

from the Chile triple Junction (Taitao Peninsula and Site 862,

LEG ODP141 on the Taitao Ridge). Tectonophysics 311, 83–111.

Guivel, C., Lagabrielle, Y., Bourgois, J., Martin, H., Arnaud, N.,

Fourcade, S., Cotten, J., Maury, R., 2003. Very shallow melting

of oceanic crust during spreading ridge subduction: origin of near-

trench Quaternary volcanism at the Chile triple junction. J. Geo-

phys. Res. 108 (B7), 2345. doi:10.1029/2002JB002119.

Hawkesworth, C.J., Norry, M.J., Roddick, J.C., Baker, P.E., Francis,

P.W., Thorpe, R.S., 1979. 143Nd/144Nd, 87Sr/86Sr and incompati-

ble trace element variations in calc-alkaline andesitic and plateau

lavas from South America. Earth Planet. Sci. Lett. 42, 45–57.

Herron, E.M., Cande, S.C., Hall, B.R., 1981. An active spreading

center collides with a subduction zone: a geophysical survey of

the Chile margin triple junction. Mem. - Geol. Soc. Am 154,

683–701.

Irvine, T.N., Baragar, W.R.A., 1971. A guide to the chemical

classification of the common volcanic rocks. Can. J. Earth Sci.

8, 523–548.

Kay, S., Ramos, V., Marquez, M., 1993. Evidence in Cerro Pampa

volcanic rocks for slab-melting prior to ridge collision in southern

South America. J. Geol. 101, 703–714.

Kay, S.M., Ramos, V.A., Gorring,M.L., 2002. Geochemistry of Eocene

Plateau basalts related to ridge collision in southern Patagonian.

XV8 Congreso Geologico Argentino (El Calafate), CD-rom.

Klein, E., Karsten, 1995. Ocean–ridge basalts with convergent-margin

affinities from the Chile Ridge. Nature 374, 52–57.

Lagabrielle, Y., Le Moigne, J., Maury, R.C., Cotten, J., Bourgois, J.,

1994. Volcanic record of the subduction of an active spreading

ridge, Taitao Peninsula (southern Chile). Geology 22, 515–518.

Lagabrielle, Y., Guivel, C., Maury, R., Bourgois, J., Fourcade, S.,

Martin, H., 2000. Magmatic–tectonic effects of high thermal

regime at the site of active ridge subduction: the Chile Triple

Junction model. Tectonophysics 326, 255–268.

Lagabrielle, Y., Suarez, M., Rossello, E.A., Herail, G., Martinod, J.,

Regnier, M., de la Cruz, R., 2004. Neogene to Quaternary tectonic

evolution of the Patagonian Andes at the latitude of the Chile

triple junction. Tectonophysics 385, 211–241.

Le Maitre, R.W., Bateman, P., Dudek, A., Keller, J., Lameyre, J., Le

Bas, M.J., Sabine, P.A., Schmid, R., Sorensen, H., Streckeisen, A.,

Wolley, A.R., Zanettin, B., 1989. A Classification of Igneous

Rocks and Glossary of Terms: Recommendation of the IOU

Subcommission of the Systematics of Igneous Rocks. Blackwell,

Oxford. 193 pp.

Leterrier, J., Maury, R.C., Thonon, P., Girard, D., Marchal, M., 1982.

Clinopyroxene composition as a method of identification of the

magmatic affinities of paleo-volcanic series. Earth Planet. Sci.

Lett. 59, 139–154.

Lopez-Escobar, L., Kilian, R., Kempton, P., Tagiri, M., 1993. Petrog-

raphy and geochemistry of Quaternary rocks from the Southern

Volcanic Zone of the Andes between 41830V and 46V00VS, Chile.Rev. Geol. Chile 20, 33–55.

Lopez-Escobar, L., Cembrano, J., Moreno, H., 1995. Geochemistry

and tectonics of the Chilean Southern Andes basaltic Quaternary

volcanism (37–468S). Rev. Geol. Chile 22, 219–234.

Luhr, J.F., Aranda-Gomez, J.J., Housh, T.B., 1995. San Quintin

volcanic field, Baja California Norte, Mexico: geology, petrology

and geochemistry. J. Geophys. Res. 100, 10,353–10,380.

C. Guivel et al. / Journal of Volcanology and Geothermal Research 149 (2006) 346–370370

Mahood, G., Drake, R.E., 1982. K–Ar dating young rhyolitic rocks: a

case study for the Sierra La Primavera, Jalisco, Mexico. Geol.

Soc. Amer. Bull. 93, 1232–1241.

Mason, P.R.D., Seghedi, I., Szakacs, Downes, H., 1998. Magmatic

constraints on geodynamic models of subduction in the East

Carpathians, Romania. Tectonophysics 297 (1–4), 157–176.

Mercer, J., Sutter, J.F., 1982. Late Miocene–Earliest Pliocene

glaciation in southern Argentina: implications for global ice-

sheet history. Palaeogeogr. Palaeoclimatol. Palaeoecol. 38,

185–206.

Middlemost, E.A.K., 1975. The basalt clan. Earth-Sci. Rev. 11,

337–364.

Morata, D., Barbero, L., Suarez, M., De la Cruz, R., 2002. Early

Pliocene magmatism and high exhumation rates in the Patagonian

Cordillera (46840VS): K–Ar and fission track data. Fifth Interna-

tional Symposium on Andean Geology. ISAG, Toulouse, France,

pp. 433–436.

Murdie, R.E., Russo, R.M., 1999. Seismic anisotropy in the region of

the Chile margin triple junction. J. South Am. Earth Sci. 12 (3),

261–270.

Nelson, E., Forsythe, R., Arit, I., 1994. Ridge collision tectonics in

terrane development. J. South Am. Earth Sci. 7, 271–278.

Pankhurst, R.J., Leat, P.T., Sruoga, P., Rapela, C.W., Marquez, M.,

Storey, B.C., Riley, T.R., 1998. The Chon-Aike province of

Patagonia and related rocks in West Antarctica: a silicic large

igneous province. J. Volcanol. Geotherm. Res. 81, 113–136.

Pankhurst, R.J., Weaver, S.D., Herve, F., Larrondo, P., 1999.

Mesozoic–Cenozoic evolution of the North Patagonian Bath-

olith in Aysen, southern Chile. J. Geol. Soc. (Lond.) 156,

673–694.

Pankhurst, R.J., Riley, T.R., Fanning, C.M., Kelley, S.P., 2000. Epi-

sodic silicic volcanism in Patagonia and the Antarctic Peninsula:

chronology of magmatism associated with the break-up of Gond-

wana. J. Petrol. 41, 605–625.

Peate, D.W., Pearce, J.A., Hawkesworth, C.J., Colley, H., Edwards,

C.M.H., Hirose, K., 1997. Geochemical variations in Vanuatu Arc

lavas: the role of subducted material and a variable mantle wedge

composition. J. Petrol. 38, 1331–1358.

Petford, N., Cheadle, M., Barreiro, B., 1996. Age and origin of

southern flood basalts, Chile Chico region (46845VS). Third Inter-

national Symposium on Andean Geology. ISAG, St Malo, France,

pp. 629–632.

Ramos, V.A., Kay, S.M., 1992. Southern Patagonian plateau basalts

and deformation: back-arc testimony of ridge collisions. Tectono-

physics 205, 261–282.

Ringwood, A.E., 1990. Slab–mantle interactions: 3. Petrogenesis of

intraplate magmas and structure of the upper mantle. Chem. Geol.

82, 187–207.

Rivalenti, G., Mazzucchelli, M., Laurora, A., Ciuffi, S.I.A., Zanutti,

A., Vannucci, R., Cingolani, C.A., 2004. The back-arc mantle

lithosphere in Patagonia, South America. J. South Am. Earth Sci.

17, 121–152.

Schmidt, M.W., Dardon, A., Chazot, G., Vanucci, R., 2004. The

dependance of Nb and Ta rutile-melt partitioning on melt com-

position and Nb/Ta fractionation during subduction processes.

Earth Planet. Sci. Lett. 226, 415–432.

Shaw, D.W., 1970. Trace element fractionation during anatexis. Geo-

chim. Cosmochim. Acta 34, 237–243.

Singer, B.S., Ackert Jr., R.P., Guillou, H., 2004. 40Ar/39Ar and K–Ar

chronology of Pleistocene glaciations in Patagonia. Geol. Soc.

Amer. Bull. 116, 434–450. doi:10.1130/B25177.1.

Sinito, A.M., 1980. Edades geologicas, radimetricas y magneticas de

algunas vulcanitas cenozoicas de las provincias de Santa Cruz y

Chubut. Rev. Asoc. Geol. Argent. 35, 332–339.

Staudigel, H., Davies, G.R., Hart, S.R., Marchant, K.M., Smith, B.M.,

1995. Large scale isotopic Sr, Nd and O isotopic anatomy of

altered oceanic crust: DSDP/ODP sites 417/418. Earth Planet. Sci.

Lett. 130, 169–185.

Steiger, R.H., Jager, E., 1977. Subcommission on geochronology:

convention on the use of decay constants in geo- and cosmochro-

nology. Earth Planet. Sci. Lett. 36, 359–362.

Stern, C.R., Kilian, R., 1996. Role of the subducted slab, mantle

wedge and continental crust in the generation of adakites from the

Andean Austral Volcanic Zone. Contrib. Mineral. Petrol. 123,

263–281.

Stern, C.R., Zartman, F.A., Futa, K., Zartman, R.E., Peng, Z., Kyser,

T.K., 1990. Trace-element and Sr, Nd, Pb, and O isotopic com-

position of Pliocene and Quaternary alkali basalts of the Patago-

nian Plateau lavas of southernmost South America. Contrib.

Mineral. Petrol. 104, 294–308.

Suarez, M., De La Cruz, R., 2001. Jurassic to Miocene K–Ar dates

from esatern central Patagonian Cordillera plutons, Chile (458–488S). J. Geol. Soc. (Lond.) 157, 995–1001.

Suarez, M., De La Cruz, R., Bell, M., 1996. Estratigrafıa de la region

de Coyhaique (latitud 458–468S); Cordillera Patagonica, Chile.

XIII Congreso Geologico Argentino y III Congreso de Explora-

cion de Hidrocarburos I, pp. 575–590.

Sun, S.,McDonough,W.F., 1989. Chemical and isotopic systematics of

oceanic basalts; implications for mantle composition and processes.

In: Saunders, A.D, Norry, J.M. (Eds.), Magmatism in the Ocean

Basins, Spec. Publ. - Geol. Soc. Lond., vol. 42, pp. 313–345.

Tebbens, S.F., Cande, S.C., 1997. Southeast Pacific tectonic evolution

from the early Oligocene to present. J. Geophys. Res. 102,

12061–12084.

Tebbens, S.F., Cande, S.C., Kovacs, L., Parra, J.C., LaBrecque, J.L.,

Vergara, H., 1997. The Chile ridge: a tectonic framework.

J. Geophys. Res. 102, 12035–12059.

Thomson, S.F., Herve, F., Stockhert, B., 2001. Mesozoic–Cenozoic

denudation history of the Patagonian Andes (southern Chile) and

its correlation to different subduction processes. Tectonics 20,

693–711.

Thorkelson, D.J., 1996. Subduction of diverging plates and the prin-

ciples of slab window formation. Tectonophysics 255, 47–63.

Thorkelson, D.J., Breitsprecher, K., 2005. Partial melting of slab

window margins: gensis of adakitic and non-adakitic magmas.

Lithos 79, 25–41.

Ton-That, T., Singer, B., Morner, N.A., Rabassa, J., 1999. Datacion de

lavas basalticas por 40Ar/39Ar y geologıa glacial de la region del

lago Bueno Aires, provincia de Santa Cruz, Argentina. Rev. Asoc.

Geol. Argent. 54, 333–352.

van den Beukel, J., 1990. Breakup of young oceanic lithosphere in the

upper part of a subduction zone: implications for the emplacement

of ophiolites. Tectonics 9 (4), 825–844.

von Blanckenburg, F., Davies, J.H., 1995. Slab breakoff: a model for

syncollisional magmatism and tectonics in the Alps. Tectonics 14

(1), 120–131.