Across and Along Arc Geochemical Variations in Altered Volcanic Rocks: Evidence from Mineral Chemistry of Jurassic Lavas in Northern Chile, and Tectonic Implications

Home , Carrizalillo (caldera), La Negra Formation

Lithos 239 (2015) 97–113

Contents lists available at ScienceDirect

Lithos

journal homepage: www.elsevier.com/locate/lithos

Across and along arc geochemical variations in altered volcanic rocks: Evidence from mineral chemistry of Jurassic lavas in northern Chile, and tectonic implications

Pablo Rossel a,1, Verónica Oliveros a,⁎, Mihai N. Ducea b,c, Laura Hernandez d a Departamento Ciencias de la Tierra, Universidad de Concepción, Casilla 160-C, Concepción, Chile b Department of Geosciences, University of Arizona, Tucson, AZ 85721, USA c Universitatea Bucuresti, Facultatea de Geologie Geoﬁzica, Strada N. Balcescu Nr 1, Bucuresti, Romania d Instituto de Geología Económica Aplicada (GEA), Universidad de Concepción, Chile article info abstract

Article history: Postmagmatic processes mask the original whole-rock chemistry of most Mesozoic igneous rocks from the Ande- Received 16 February 2015 an arc and back-arc units preserved in Chile. Mineral assemblages corresponding to subgreenschist metamorphic Accepted 7 October 2015 facies and/or propylitic hydrothermal alteration are ubiquitous in volcanic and plutonic rocks, suggesting Available online 23 October 2015 element mobility at macroscopic and microscopic scale. However, fresh primary phenocrysts of clinopyroxene and plagioclase do occur in some of the altered rocks. We use major and trace element chemistry of such mineral Keywords: phases to infer the geochemical variations of four Jurassic arc and four back-arc units from northern Chile. Back arc Central Andes Clinopyroxene belonging to rocks of the main arc and two units of the bark-arc are augites with low contents of Clinopyroxene HFSE and REE; they originated from melting of an asthenospheric mantle source. Clinopyroxenes from a third Mineral chemistry back-arc unit show typical OIB afﬁnities, with high Ti and trace element contents and low Si. Trace elemental Thermobarometry variations in clinopyroxenes from these arc and back-arc units suggest that olivine and clinopyroxene were the main fractionating phases during early stages of magma evolution. The last back-arc unit shows a broad spectrum of clinopyroxene compositions that includes depleted arc-like augite, high Al and high Sr–Ca diopside (adakite-like signature). The origin of these lavas is the result of melting of a mixture of depleted mantle plus Sr- rich sediments and subsequent high pressure fractionation of garnet. Thermobarometric calculations suggest that the Jurassic arc and back-arc magmatism had at least one crustal stagnation level where crystallization and fractionation took place, located at ca. ~8–15 km. The depth of this stagnation level is consistent with lower-middle crust boundary in extensional settings. Crystallization conditions calculated for high Al diopsides suggest a deeper stagnation level that is not consistent with a thinned back-arc continental crust. Thus minor garnet fractionation occurred before these magmas reached the base of the crust. The presented data support the existence of a heterogeneous sub arc mantle and complex magmatic processes in the early stages of the Andean subduction. © 2015 Elsevier B.V. All rights reserved.

1. Introduction and related volcaniclastic rocks, (chlorite–epidote–quartz–calcite– titanite–sericite–actinolite–K-feldspar–zeolite–prehnite–pumpeyllite) All the volcanic rocks that record the early Andean subduction represent either propylitic hydrothermal alteration or greenschist or system (Jurassic to Lower Cretaceous) exhibit some degree of hydro- sub-greenschist facies low-grade metamorphic events that took place thermal alteration and or low grade metamorphism, part of which during the Mesozoic (Losert, 1974; Oliveros et al., 2006; Parada et al., may have taken place in contact with seawater under submarine 2007; Robinson et al., 2004; Tristá-Aguilera et al., 2006). Considering conditions, as well as subsequent contact and regional metamorphism that under the above described metamorphic conditions large ion related to the emplacement of later arc magmas (Aguirre et al., 1989; litophile elements (LILE), low ﬁeld strength elements (LFSE) and some Levi et al., 1989; Losert, 1974; Oliveros et al., 2008; Robinson et al., high ﬁeld strength elements (HFS) can have a mobile behavior, the orig- 2004). The secondary mineral assemblages observed in lavas, dikes inal composition of the rocks may have been obliterated, which limits our ability to decipher their original elemental and isotopic characteristics, and prevents a robust petrogenetic interpretation for the Andean ⁎ Corresponding author. magmatism during the Mesozoic. E-mail address: [email protected] (V. Oliveros). 1 Present address: Facultad de Ingeniería, Geología, Universidad Andres Bello, Autopista Most of the work discussing the petrogenesis of the Jurassic to Lower Concepción-Talcahuano 7100, Concepción, Chile. Cretaceous Andean volcanics has focused on the study of the less mobile

http://dx.doi.org/10.1016/j.lithos.2015.10.002 0024-4937/© 2015 Elsevier B.V. All rights reserved. 98 P. Rossel et al. / Lithos 239 (2015) 97–113 elements (i.e.: the rare earth and high field strength elements), the Vicente, 2006). In this context, the Upper Jurassic geological units isotopic ratios of volcanic rocks and the comparison with the composi- cropping out in the Precordillera and High Andes between 26° y and tion of co-genetic less altered intrusive bodies (Kramer et al., 2005; 31°S, the Quebrada Vicuñitas Beds and the Lagunillas, Picudo and Lucassen et al., 2006; Oliveros et al., 2006, 2007; Palacios, 1978; Algarrobal formations (Fig. 1), represent a particular event in the evolu- Parada et al., 1999; Rogers and Hawkesworth, 1989; Vergara et al., tion of the back arc basins, in which basic and intermediate volcanism 1995). This geochemical comparison to contemporaneous unaltered and continental sedimentation dominated over protracted marine sed- intrusions is not possible in the case of the Upper Jurassic back arc imentary conditions (Charrier et al., 2007). volcanics (Rossel et al., 2013), since back arc plutonic rocks, if they Since the Late Cretaceous, the continuous shifting of the magmatic exist, are not exposed in this area. front to its present position in the high Andes (Charrier et al., 2007; The chemistry of primary magmatic mineral phases can be an Mpodozis and Ramos, 1989) resulted in the occurrence of numerous important tool for the characterization of magmatism. Ever since the events of hydrothermal alteration or regional low grade metamorphism definition by Bowen (1928) of the continuous and discontinuous that modified to a greater or lesser extent the geochemical composition crystallization series in igneous rocks, there are numerous examples of of the host rocks. These events are represented in all the studied units by successful methodologies that use mineral compositions to assess burial metamorphism, in the Precordillera area by the intrusion of Pa- petrogenetic conditions; for instance the tectonic discrimination leocene and mainly Eocene plutons (Cornejo et al., 1998; Iriarte et al., diagrams based on clinopyroxene composition in basalts (Leterrier 1999; Mpodozis and Cornejo, 1986; Tomilson et al., 1999) while in the et al., 1982), the characterization of magmatism based on trace and Coastal Cordillera the main thermal event affecting the arc rocks was rare earth elements (REE) compositions of clinopyroxenes and other the product of the intrusion of composite Lower Cretaceous plutons primary mineral phases (Marks et al., 2004; Vannucci et al., 1993)and (Creixell et al., 2012; Dallmeyer et al., 1996; Emparán and Calderón, several thermo-barometers based on melt(whole rock)–mineral 2010; Emparán and Pineda, 2000; Grocott et al., 1994; Welkner et al., equilibrium (Putirka, 2005, 2008; Putirka et al., 2003). Therefore, if the 2006). In addition, much of the Jurassic volcanism was developed in a chemistry of some primary igneous mineral phases is not changed submarine environment (Vergara et al., 1995; Vicente, 2005)that during secondary processes, then their major and trace elemental affected the geochemical composition of the host rocks (Kramer et al., compositions represent indirect petrogenetic proxies. 2005; Oliveros et al., 2007), hindering the clear interpretation of the Jurassic volcanic rocks cropping out in the present-day Coastal genesis of the Jurassic Andean magmatism. Cordillera, Precordillera and High Andes of north-central Chile (26°– 31°S) bear phenocrysts that are fresh or much less altered than the 3. Sample selection and analytical methods surrounding groundmass. Here we present in-situ mineral chemistry data in order to gain knowledge on the petrogenesis of altered Jurassic In order to properly characterize the magmatic sources and evolution Andean magmatic rocks. We compare these results with data obtained of the arc and back arc volcanism in northern Chile, we selected eleven from a companion study that focused on whole-rock chemistry samples for chemical analyses of fresh pyroxene and plagioclase pheno- (Rossel et al., 2013). In that paper, we showed that there is a progressive crysts (Table 1). One sample of the Agua Salada Volcanic Complex change in the geochemistry of the lavas from arc to back-arc, and linked (ASVC) represents the arc domain and two samples of the Quebrada it to variations in the magma sources. Those changes should be also Vicuñita Beds, two samples of the Lagunillas formation, four samples of traceable with the geochemistry of major and trace elements of primary the Picudo formation and two samples of the Algarrobal formation rep- minerals, such as clinopyroxene, allowing an improvement in the resent the back arc domain. Although slightly altered, the selected sam- reconstruction of the sources and primary magmatic processes of the ples are the most primitive and best preserved from the arc and back arc early stages of Andean subduction. volcanic units, they have either low contents of silica or were classified as basalt, basaltic andesite or andesite based on their mineralogy and tex- 2. Geologic background ture. Their bulk chemistry is presented in Rossel et al. (2013). Given that the olivine crystals in these rocks, when present, are completely During the Jurassic and Early Cretaceous, western South America replaced by iddingsite (mixture of Fe-rich phylosilicates), the most was characterized by an extensional and transtensional tectonic regime primitive phases in equilibrium with the original melt that have their (Charrier et al., 2007; Creixell et al., 2006; Grocott and Taylor, 2002; chemistry unaltered are the fresh clinopyroxene and to a lesser extent Scheuber and González, 1999) due to the oblique subduction of the plagioclase. dense and cold Phoenix plate (Jaillard et al., 1990; Scheuber and Electron microprobe analyses were carried out using a CAMECA González, 1999). This geodynamic regime resulted in a roll-back of the SX100 and a JEOL JXA-8600M electron microprobe at the University of slab and retreat of the trench, as well as progressive thinning of the Arizona and University of Concepción, respectively. Analytical condi- upper plate (Charrier et al., 2007; Ramos, 1999). The convergent margin tions were 15 kV, 20 nA with a 1 um beam size. Na and K were measured developed an arc region in the present-day Coastal Cordillera of south- at 15 kV, 8 nA and the remaining elements at 15 kV, 20 nA, with typical ern Perú and northern-central Chile, and a series of back arc basins to detection limits of WDS (wavelength-dispersive spectrometry) in the the east of the arc, within the modern Chilean Precordillera and High range between 0.13% (SiO2) and 0.01% (K2O). Ion microprobe analyses Andes. The arc magmatism was voluminous (Kramer et al., 2005; were performed at Arizona State University SIMS (Secondary Ion Mass Lucassen et al., 2006; Oliveros et al., 2006, 2007; Palacios, 1978; Spectrometry) facility, with a primary beam of 16O− at a current of 3 Rogers and Hawkesworth, 1989; Vergara et al., 1995), and led to the to 4 nA. For the measurements the approach of “conventional energy accumulation of volcanic sequences up to 7000 m in thickness with filtering” was used, as described by Shimizu et al. (1978), Hinton minor sedimentary intercalations, and subsequent intrusion of large (1990) and Zinner and Crozaz (1986). plutonic bodies (Buchelt and Tellez, 1988; Muñoz et al., 1988). Among Major element composition for pyroxene and plagioclase of 20 the units that represent the Jurassic arc are the La Negra Formation samples from the volcanic rocks of the Jurassic arc cropping out and its equivalents (Camarca and Oficina Viz Formations) in northern between 18°30′ and 24°00S, from Oliveros et al. (2007), are included Chile (18°–29°S), and the Agua Salada Volcanic Complex (30°–31°S). in the database for comparison. The back arc region developed in response to the progressive thinning of the continental crust (Charrier et al., 2007; Martínez et al., 2012), 4. Bulk rock petrography and geochemistry of the studied units and it is composed of up to 4000 m of marine and continental sedimentary rocks that were deposited during the Jurassic and Early Cretaceous The Jurassic to Lower Cretaceous volcano-sedimentary units in (Charrier et al., 2007; Martínez et al., 2012; Mpodozis and Ramos, 1989; the Precordillera and high Andes of northern Chile, 26°–31°S P. Rossel et al. / Lithos 239 (2015) 97–113 99

Fig. 1. Simpliﬁed geological map of the studied area including sample locations. QVB: Quebrada Vicuñita Beds. LF: Lagunillas Formation. PF: Picudo Formation. AF: Algarrobal Formation. ASVC: Agua Salada Volcanic Complex. DZ Max. age: maximum depositional age obtained through U–Pb dating of detrital zircons. IZ: crystallization age of rocks obtained through U–Pb dating of igneous zircons. Modiﬁed after SERNAGEOMIN (2003).

(Fig. 1) are relatively well characterized in terms of their petrogra- Rossel et al. (2013). In spite of the variable alteration degree exhibit- phy and geochemistry. In particular, the volcanic rocks representa- ed by these rocks, the presence of fresh clinopyroxene phenocrysts is tive of the Jurassic back arc domain have been studied in detail by not uncommon. 100 P. Rossel et al. / Lithos 239 (2015) 97–113

Table 1 Main features of the analyzed samples.

Sample Unit Lithology Coordinates wt.%SiO2 #Mg Main Main First Ca Cpx Bulk-rock Bulk-rock Analysis Measured (UTM E/N) mineralogy alteration phase composition 87Sr/86Sr(i)a εNd(i)a minerals

PR-09-19 QVB BA 457666/7036398 ––Plg, Cpx, Fe–Ti Chl Plg Augite ––EMPA Cpx, Plg oxides PR-10-32 QVB BA 456968/7085732 52.33 49 Plg, Cpx, Fe–Ti Chl Plg Augite 0.7047 4.1 EMPA, Cpx, Plg oxides SIMS PR-10-72 LF AB 446046/6935452 45.26 53 Ol, Plg, Cpx, Fe–Ti Id-Cc Cpx = Ti-subsilic 0.704 2.8 EMPA, Cpx oxides Plg diopside SIMS PR-10-80 LF AB 444163/6929847 48.11 43 Ol, Plg, Cpx, Fe–Ti Id-Cc Plg? Ti-subsilic 0.7043 3.3 EMPA Cpx oxides diopside PR-09-05 PF BA 375402/6796136 57.95 45 Plg, Cpx, Opx, Fe–Ti Chl Plg Augite 0.7045 1.3 EMPA, Cpx oxides SIMS PR-10-41 PF BA 374711/6791142 55.96 47 Plg/Cpx/Fe–Ti Chl Plg Augite 0.7042 0.2 EMPA Cpx, Plg oxides PR-10-45 PF B 373840/6791219 49.74 55 Plg/Cpx/Fe–Ti Chl Cpx = Diopside 0.7042 2.7 EMPA, Cpx, Plg oxides Plg? SIMS PR-11-168 PF BA 375288/796083 ––Plg/Cpx/Fe–Ti Chl Cpx = Diopside 0.7041 3.8 EMPA, Cpx oxides Plg? SIMS M-14 AF BA 352250/6681573 53.87 41 Plg/Cpx/Fe–Ti Chl Plg Augite 0.7045 2.7 EMPA Cpx oxides M-25 AF A 345541/6681372 62.45 44 Plg/Cpx/Fe–Ti Chl Plg Augite 0.7042 0.7 EMPA Cpx oxides PR-11-153 ASVC BA 275629/6729590 54.68 56 Plg/Cpx/Fe–Ti Chl Plg Augite 0.704 2.7 EMPA, Cpx oxides SIMS

Plg = plagioclase. Cpx = clinopyroxene. Ol = olivine. Fe–Ti = Fe–Ti oxide. Chl = Chlorite. Id–Cc = Iddingsite + calcite. B = basalt. BA = basaltic andesite. A = andesite. AB = alkaline basalt. EMPA = Electronic Mircroprobe Analysis. SIMS = Secondary Ionization Mass Spectrometry a Rossel et al. (2013)

4.1. Algarrobal Formation seriate euhedral phenocrysts of plagioclase (~25% of rock volume) that can reach ~7 mm along their major axis, and minor amounts This Upper Jurassic back arc unit crops out between ~30 and 31°S. (~7%) of euhedral to subhedral clinopyroxene (Fig. 2a), ranging in size The selected samples of this formation have porphyritic textures with between 1 and 2 mm. Completely altered euhedral crystals of olivine

Fig. 2. Back-scattered electron (BSE) and second electron (SE) images of typical compositional and textural features observed in the studied rocks. BSE images of a) Large clinopyroxene (Cpx) crystal with opaque inclusions (Fe–Ti Ox.) and b) clinopyroxene phenocryst and large plagioclase (Plg) crystal with poikilitic clinopyroxene inclusions from the Algarrobal Forma- tion from sample M-14; c) SE image of clinopyroxene phenocryst from sample PR-09-19 from Quebrada Vicuñita Beds; d) BSE image of equigranular arrangement of plagioclase and clinopyroxene with interstitial olivine (Ol) from sample PR-10-80 of the Lagunillas Formation. P. Rossel et al. / Lithos 239 (2015) 97–113 101 are present but in small amounts (b3%). They range in size from 0.5 to 4.3. Quebrada Vicuñita Beds 1 mm and can occur either as phenocrysts or in the groundmass. Plagioclase and clinopyroxene mostly occur as isolated phenocrysts, This Upper Jurassic back-arc unit crops out between ~26° and 27°S. but clinopyroxene can also occur as inclusions in large plagioclase The rocks of the Quebrada Vicuñita Beds are very homogeneous in com- poikilocrysts (Fig. 2b). Alteration affecting plagioclase phenocrysts position, corresponding to porphyritic basaltic andesites, with plagio- results in the development of sericite and minor amounts of epidote, clase and clinopyroxene phenocrysts (Fig. 2c). In the selected samples but is common to find fresh crystals, or at least a fresh area inside the of this unit plagioclase phenocrysts are euhedral and up to 10 mm in phenocrysts. On the other hand, clinopyroxene phenocrysts can be size, and represent approximately 80% of coarse fraction, whereas divided in almost completely fresh and completely chloritizated clinopyroxenes are subhedral, usually less than 2 mm in size, and can crystals. No zonation is observed by optical microscope or backscattered occur grouped as glomerochrysts up to 5 mm in diameter. The electron (BSE) images into feldspars or pyroxenes. The groundmass pilotaxitic groundmass represents approximately 70–75% of the volume usually has a pilotaxitic texture and it is composed of euhedral plagio- of the rock and is composed of plagioclase, Fe–Ti oxides and minor clase, clinopyroxene, Fe–Ti oxides and minor amounts of olivine. A clinopyroxene microcrystals. Hydrothermal alteration overprint is low-grade metamorphism or hydrothermal alteration overprint is mostly recorded in the feldspars: large plagioclase phenocrysts can be evidenced by the greenish color of the groundmass under the optical partially replaced by sericite, whereas groundmass microliths are either microscope, the albitization of plagioclase microcrystals and the albitized or replaced by brownish-greenish fine clay. Clinopyroxenes presence of completely chloritized clinopyroxenes. The most common may exhibit thin chloritized rims in phenocrysts or in the groundmass. secondary phases in the matrix are chlorite and clays. A particular alteration feature recognized in one of the selected samples According to their mineralogy, with abundant of mafic constituents is the presence of amphibole crystals closely related to areas where the (mostly clinopyroxene) and the absence of primary quartz, K-feldspars replacement of the groundmass by brownish-greenish fine clay is more or hydrated minerals, the Algarrobal Formation rocks can be classified intense. Ar–Ar dating of the altered groundmass yielded an age of mainly in the range between basalts and andesites. However, due to 72.9 ± 0.3 Ma (Rossel et al., 2013), that probably reflects a thermal the intense silicic alteration affecting these lavas, the apparent compo- event due to intrusion of Upper Cretaceous to Paleocene plutons sitional range indicated by whole-rock chemistry varies from basaltic (Cornejo et al., 1998), affecting the Quebrada Vicuñita Beds rocks. andesite to rhyodacite (Rossel et al., 2013). These rocks have typical The geochemistry of Quebrada Vicuñita Beds lavas indicates a transi- geochemical features of subduction-related magmatism: calc-alkaline tional affinity for the volcanism, between calc-alkaline and tholeiitic, affinities, with no early enrichment of FeOt, Mg# ranging between 30 with a marked enrichment in FeOt, Mg# that range between 37–52 and 56, enrichment in LILE and light rare earth elements (LREE), deple- and incompatible element contents similar to those of the Jurassic arc tion in HFSE and Nb–Ta troughs similar to those observed in the Jurassic rocks (Rossel et al., 2013). Spider diagrams of these units show typical arc volcanism (Oliveros et al., 2007; Rossel et al., 2013). The whole rock features of subduction-related magmas with deep Nb–Ta troughs contents of heavy rare earth elements (HREE) with respect to chondritic (LaN/NbN =2–11), REE patterns flatter than those of the other back reservoir (Rossel et al., 2013) suggest that the magmas were generated arc units (LaN/YbN between 3 and 10), high concentrations of YbN above the garnet stabilization zone (YbN N 10). (13–23) and a flat tail in the REE diagrams normalized to the chondritic reservoir (DyN/YbN ≤ 2). 4.2. Picudo Formation 4.4. Lagunillas Formation This Upper Jurassic back arc unit crops out at ~29°S and is composed of a bimodal suite of volcanic rocks ranging from olivine and This formation is the easternmost Upper Jurassic back arc unit and clinopyroxene-bearing basalts to andesites, and rhyolitic vitric tuffs. crops out as a discontinuous belt between ~27° and 30°S. The volcanic The selected rocks have abundant fresh zoned clinopyroxene and rocks of the Lagunillas Formation are exclusively olivine basalts lava plagioclase phenocrysts, up to 2 mm in size, which represent ~30% of flows with a typical massive central part and a vesicular top; they the volume of the rock. Olivine-phenocrysts, when present, are have a fine grained equigranular texture, although minor (b5%) plagio- completely replaced by iddingsite. The observed proportion of pheno- clase and olivine phenocrysts are observed. Plagioclase is the most crysts is variable from the clinopyroxene-rich (≈30% of phenocrysts) abundant mineral phase, representing approximately ~80% of the sample PR-10-45 to the clinopyroxene-poor (b5% of phenocrysts) rocks by volume (Fig. 2d). Olivine usually represents about ~10% of sample PR-09-05. The groundmass has a pilotaxitic texture and is the volume of the rock, it can reach 2 mm in size and is always composed of euhedral plagioclases, clinopyroxenes, and minor amounts completely replaced by iddingsite (mixture of Fe-rich phyllosilicates). of olivine and Fe–Ti oxides. The selected samples show scarce alteration Plagioclase crystals are weakly altered to sericite and range in size evidences such as the greenish color of the groundmass observed under from 0.5 to 1 mm; they can also occur as isolated phenocrysts (2 mm) the optic microscope, probably related to minor degrees of propylitic in sample PR-10-80. The clinopyroxene occurs as crystals intergrown alteration, albitization of some plagioclases microcrystals and the with plagioclase; it represents less than 5 to 10% of the volume of the presence of hematite micro veins. sample and shows no evidence of alteration. Finally Fe–Ti oxides occur The volcanic rocks of this formation have transitional affinities (calc- as interstitial crystals but commonly they do not constitute more than alkaline to alkaline), with no early enrichment of FeOt, Mg# ranging 1 or 2% of the volume of the rocks. Locally, the volcanic rocks of this from 45 to 55 and geochemical features typical of subduction-related unit are pervasively altered to calcite, which replaces plagioclase and ol- magmatism (Rossel et al., 2013): they are more enriched in incompati- ivine phenocrysts and fills numerous amygdules. The strongly altered ble elements, have less pronounced Nb–Ta troughs (LaN/NbN ~2) and samples were not included in this work. higher concentrations of REE, especially LREE (LaN/YbN =8–12), than The selected rocks of Lagunillas Formation have alkaline affinities, the rocks of the Jurassic arc (Oliveros et al., 2007; Rossel et al., 2013). with no early enrichment of FeOt, Mg# that ranges between 32 and 61 Two of the six lava flows of this unit analyzed by Rossel et al. (2013) and the lowest concentrations of SiO2 of all the back-arc units, ranging have extremely high contents of Sr (2350 and 1935 ppm) which was between 45 and 49 wt.%. Spider diagrams of these lavas show patterns interpreted as the result of hydrothermal alteration processes, however, similar to those of the ocean island basalts (OIBs), with no Nb–Ta anom- a more complex petrogenetic history cannot be ruled out since they also alies (LaN/NbN b 1). The REE diagrams exhibit a marked enrichment in have low HREE content compared to chondritic reservoir (YbN ≤ 7), LREE relative to HREE (LaN/YbN =5–17) and low concentration of which may be due to the presence of garnet as residual phase in the HREE respect to chondritic reservoir (YbN ≤ 10), suggesting the presence magma source. of garnet as a residual phase in the magma source. 102 P. Rossel et al. / Lithos 239 (2015) 97–113

Table 2 Mean and range of values for EMPA and SIMS analysis of clinopyroxene crystals from the studied units and the previously published Jurassic arc samples (Oliveros et al., 2007).

Lagunillas Formation Quebrada Vicuñita Beds Picudo Formation

PR-10-72 PR-10-80 PR-09-19 PR-10-32 PR-09-05 PR-10-41 PR-10-45

(N = 11) (N = 11) (N = 29) (N = 32) (N = 4) (N = 20) (N = 33)

Mean Range Mean Range Mean Range Mean Range Mean Range Mean Range Mean

SiO2 45.3 43.4/46.4 48.2 44.7/50.5 50.8 49.9/51.6 51.7 50.3/52.4 50.2 48.2/52.8 51.3 50.1/52.2 49.4

TiO2 4.40 3.8/5.1 2.98 1.7/3.5 0.71 0.6/0.9 0.6 0.45/0.8 0.86 0.6/1.2 0.85 0.6/1.2 0.96

Al2O3 6.64 5.6/7.8 3.99 1.9/4.6 3.02 1.7/2.7 2.18 1.8/3.4 3.86 2.1/5.4 3.16 2.4/4.1 7.01

Cr2O3 0.03 0/0.05 0.02 0/0.04 0.11 0/0.24 0.13 0/0.24 0.03 0.01/0.07 0.03 0/0.15 0.15 FeO 8.22 7.9/8.7 9.08 8.8/9.5 10.2 9.5/15.8 9.78 8.5/11.9 9.65 8.8/10.3 8.78 7.6/11.2 7.11 MnO 0.14 0.1/0.2 0.18 0.1/0.2 0.24 0.2/0.4 0.29 0.2/0.4 0.26 0.2/0.3 0.25 0.2/0.3 0.16 MgO 12.1 11.5/12.5 12.7 12.1/13.8 15.1 12.6/16.1 15.8 15.7/16.5 15.4 13.9/17.4 14.9 13.9/16.2 14.4 CaO 21.9 22.2/21.6 21.6 21.2/21.9 18.1 16.8/19.2 18.5 16.7/19.4 19 16.8/20.2 20.1 18.7/20.8 20.4

Na2O 0.64 0.6/0.7 0.64 0.6/0.8 0.3 0.26/0.36 0.3 0.26/0.34 0.36 0.3/0.5 0.44 0.3/0.6 0.57 Si 1.72 1.66/1.75 1.83 1.8/1.91 1.92 1.90/1.93 1.93 1.90/1.95 1.88 1.83/1.94 1.91 1.88/1.93 1.82 Ti 0.13 1.1/1.5 0.09 0.05/0.1 0.02 0.2/0.3 0.02 0.1/0.2 0.02 0.02/0.03 0.02 0.02/0.03 0.03 Al 0.3 0.25/0.35 0.18 0.09/0.2 0.1 0.07/0.12 0.1 0.08/0.15 0.17 0.09/0.24 0.14 0.10/0.18 0.31 Cr 0 0/0.002 0.00 0/0.001 0 0/0.007 0 0/0.007 0 0/0.002 0 0/0.004 0 Fe2+ 0.26 0.25/0.27 0.29 0.28/0.3 0.33 0.3/0.51 0.31 0.26/0.38 0.3 0.27/0.32 0.27 0.24/0.35 0.22 Mn 0.01 ±0.001 0.01 -0.002 0.01 0.006/0.02 0.01 0.007/0.01 0.01 0.006/0.01 0.01 0.005/0.01 0.01 Mg 0.68 0.66/0.71 0.72 0.68/0.78 0.88 0.73/0.91 0.88 0.83/0.92 0.86 0.78/0.95 0.83 0.79/0.89 0.76 Ca 0.89 0.88/0.91 0.88 0.86/0.93 0.74 0.7/0.77 0.74 0.67/0.77 0.76 0.66/0.82 0.8 0.75/0.82 0.81 Na 0.05 0.40/0.53 0.05 0.03/0.06 0.02 0.02/0.03 0.02 0.02/0.025 0.03 0.02/0.03 0.03 0.02/0.04 0.04 Mg# 72.4 70.7/73.7 71.3 69.7/72.9 72.5 58.7/74.4 74.2 68.8/77.2 73.9 71.2/76.9 75.2 69.6/78.0 78.3 En 37.2 36.2/38.2 38 37.1/40.4 44.9 37.6/46.3 45.7 42.8/47.6 44.7 40.8/49.4 43.5 41.6/46.2 43.5 Fs 14.2 13.7/15.0 15.3 14.7/16.1 17 15.5/26.4 15.9 13.7/19.6 15.7 14.1/16. 14.3 12.4/18.4 12.1 Wo 48.6 48.1/49.1 46.7 44.6/47.5 38 36/39.9 38.5 34.8/40.0 39.6 34.2/42.7 42.2 39.3/43.4 44.4

4.5. Agua Salada Volcanic Complex Mg#2 diagram (Fig. 3), with Ti contents over or very close to 0.1 a.p.f.u. Clinopyroxene Mg# ranges between 50 and 85, with those This Upper Jurassic arc unit crops out between ~30° and 31°S in from back arc samples having a more restricted range, mostly between Coastal Cordillera of northern Chile. The rocks of this unit are mostly 67 and 85 (Fig. 3). The highest Mg# are from sample PR-10-45 of the volcanic and subvolcanic basaltic andesites strongly affected by contact Picudo Formation, the lowest are from crystals of the Jurassic arc metamorphism related to the intrusion of large granitoids of Lower Cre- rocks, and the more restricted range is observed in clinopyroxenes of taceous age (Emparán and Pineda, 2000). The rocks have porphyritic Lagunillas Formation (69–74). Clinopyroxenes in basaltic samples PR- textures with up to 40% of phenocrysts. Plagioclase is the main primary 10-45 from Picudo and PR-10-72 and PR-10-80 from Lagunillas forma- mineral phase representing up to 80% of the volume of the rocks; the tions, and PR-11-168 from Picudo Formation are diopsides, whereas in crystals are strongly but not completely altered to sericite and epidote. the other arc and back arc samples the pyroxenes are augites (Fig. 4a). Clinopyroxene is a subordinate phase representing no more than 5% of Three low-Wo clinopyroxenes from the Jurassic arc rocks can be classi- the volume of the rock; the crystals are replaced by chlorite, although fied as pigeonites. Clinopyroxene have Mg# ranging between 65 and 78, fresh cores can be found. Amphibole and quartz also occur, especially the most primitives are those from sample PR-10-41 (Figs. 3, 4a). A few in fine-grained samples, but these phases likely formed during contact analyzed phenocrysts from the Agua Salada Volcanic Complex are metamorphism. augites with Mg# that ranges between 67 and 70; their composition is The volcanic rocks from this complex have calc-alkaline/transitional very similar to the clinopyroxene phenocrysts of the Jurassic arc rocks to tholeiitic affinities, like most of the Jurassic arc rocks, but with no (Oliveros et al., 2007). clear early enrichment of FeOt. HREE contents respect to chondritic res- Scarce orthopyroxene crystals were found in evolved samples of the ervoir are generally lower (YbN ≈ 5–11) than other arc rocks. The #Mg Picudo Formation. (PR-09-05 and PR-10-41) and in the Jurassic arc ranges from 41 to 68, suggesting that these samples do not represent samples (Oliveros et al., 2007) and they correspond to enstatites. primary melts.

5.2. Trace element chemistry 5. Pyroxene chemistry Trace element chemistry of the clinopyroxenes is listed in the 5.1. Major elements Table B (Suppl. content); REE concentration and ratios were normalized to the chondrite composition of Sun and McDonough (1989). The REE Major element chemistry of the clinopyroxenes is listed in Table 2 patterns display a convex-upward shape with the highest enrichments and detailed in the Table A (Suppl. content). Pyroxene description and between Nd and Eu. Unlike the clinopyroxenes from the Agua Salada classification are based on atoms per formula unit (a.p.f.u.) normalized Volcanic Complex (arc domain), the minerals from back arc samples to 6 oxygens (Morimoto et al., 1988) and include data from this work show no negative Eu anomalies (Fig. 5), suggesting that there was no and previously analyzed pyroxenes from Oliveros et al. (2007). major previous fractionation of plagioclase in those melts, even when Most of the clinopyroxenes from the Picudo and Lagunillas forma- petrographic analysis suggest that plagioclase is commonly the first tions are aluminous. They have higher contents of Ca, Na, Ti and lower phase to crystallize. The CeN/SmN (LREE) and GdN/YbN (HREE) ratios contents of Si, Fe, Mg compared to pyroxenes from the Algarrobal For- vary significantly between 0.1–1.5 and 0.2–3.2, but most of the samples mation, Quebrada Vicuñita Beds and those of the Jurassic arc rocks. In are between 0.3 to 0.6 and 0.9 to 1.9. However, the ratios obtained in particular, clinopyroxenes from sample PR-10-72 (Lagunillas Forma- one alkali basalt from Lagunillas Formation (0.7 to 0.9 and 2.5 to 3.2 in tion) can be classified as subsilicic and titanian, defining a completely 2 different fractionation trend compared to the other units in the Ti vs Mg/(Mg + Fet)abundancesina.p.f.u. P. Rossel et al. / Lithos 239 (2015) 97–113 103

Picudo Formation Algarrobal Formation Jurassic Arc

PR-10-45 PR-11-168 All Samples M-14 M-24 PR-11-153 Oliveros et al., 2007

(N = 33) (N = 17) (N = 4*) (N = 9) (N = 8) (N = 5) (N = 179)

Range Mean Range Mean Range Mean Range Mean Range Mean Range Mean Range

48.3/51.9 49.7 48.1/51.9 50.1 49.4/51.3 52.1 50.7/53.2 52.6 51.9/53.3 51.2 50.7/51.9 #### 44.9/53.4 0.5/1.5 1.07 0.7/1.7 0.94 0.85/1.1 0.71 0.6/0.8 0.80 0.7/0.9 0.91 0.65/1.83 0.74 0.3/2.6 4.6/8.1 3.97 1.7/5.7 4.5 3.16/4.6 1.83 1.5/2.0 1.67 1.4/2.2 1.93 1.86/2.04 2.20 1.0/8.1 0/0.51 0.01 0/0.03 0.06 0.001/0.15 0.04 0/0.07 0.04 0.004/0.1 0.04 0.02/0.05 0.08 0.0/0.7 5.1/9.2 7.43 6.2/8.4 8.24 7.1/9.65 11.9 11.0/14.0 12.1 11.4/12.8 11.6 11.2/11.91 #### 6.1/16.5 0.1/0.24 0.26 0.2/0.5 0.23 0.16/0.26 0.32 0.3/0.5 0.35 0.3/0.4 0.33 0.27/0.4 0.34 0.2/0.6 13.4/16.1 14.9 14/16.7 14.9 14.4/15.43 15.4 14.0/15.9 14.9 14.6/15.1 15.5 15.4/15.7 #### 6.0/17.0 19.0/21.5 21.8 20.6/22 20.3 18.98/21.81 17.6 17.2/18.0 18 17.6/18.6 18.3 18.16/18.4 #### 13.0/21.3 0.5/0.7 0.49 0.4/0.6 0.47 0.36/0.57 0.26 0.2/0.3 0.31 0.3/0.4 0.3 0.26/0.31 0.30 0.18/0.93 1.79/1.89 1.86 1.82/1.94 1.87 1.82/1.91 1.94 1.92/1.96 1.95 1.93/1.97 1.95 1.93/1.97 #### 1.78/1.96 0.01/0.04 0.03 0.02/0.05 0.02 0.024/0.03 0.02 0.015/0.02 0.02 0.02/0.03 0.02 0.02/0.03 #### 0.01/0.08 0.20/0.35 0.18 0.06/0.23 0.14 0.14/0.31 0.08 0.07/0.09 0.07 0.06/0.10 0.07 0.06/0.10 #### 0.04/0.38 0/0.014 0 0/0.001 0 0.01/0.15 0 0/0.002 0 0/0.003 0 0/0.003 #### 0.00/0.02 0.15/0.28 0.23 0.19/0.26 0.26 0.22/0.302 0.37 0.34/0.45 0.38 0.35/0.40 0.38 0.35/0.40 #### 0.19/0.54 0.003/0.01 0.01 0.01/0.014 0.01 0.005/0.008 #### 0.008/0.01 0.01 0.01/0.013 0.01 0.01/0.013 #### 0.01/0.02 0.74/0.87 0.83 0.78/0.92 0.83 0.76/0.86 0.86 0.80/0.88 0.82 0.81/0.84 0.82 0.81/0.84 #### 0.35/0.94 0.76/0.85 0.87 0.82/0.9 0.81 0.76/0.874 0.7 0.69/0.72 0.72 0.70/0.74 0.72 0.70/0.74 #### 0.51/0.84 0.03/0.05 0.04 0.03/0.04 0.03 0.026/0.041 0.02 0.017/0.02 0.02 0.02/0.03 0.02 0.02/0.03 #### 0.01/0.07 72.6/85.0 78.1 74.8/82.4 76.4 73.9/78.3 69.7 63.9/71.7 68.7 67.1/69.9 68.7 67.1/69.9 69.4 39.4/83.0 41.3/47.1 42.9 40.5/47.6 43.7 42.9/44.7 44.3 40.8/45.5 43.0 42.1/43.7 43.0 42.1/43.7 42.9 20.8/48.1 8.3/15.7 12 9.9/13.7 13.5 12./15.7 19.3 45.5 19.6 18.5/20.6 19.6 18.5/20.6 18.9 9.6/32.1 42.1/46.6 45.1 42.2/45.8 42.8 39.6/45.1 36.5 17.7/23.0 37.5 36.6/38.6 37.5 36.6/38.6 38.1 30.0/47.0 sample PR-10-72) and a basaltic andesite from Picudo Formation (1.0 to Feldspars from the other units show a restricted range of composi-

1.5 and 2.0 to 2.6 in sample PR-11-168) are much higher than those of tion (An64–50), with a few isolated bytownite and andesine crystals the other samples. The pyroxenes of Quebrada Vicuñita Beds and Agua from Quebrada Vicuñita Beds and Algarrobal Formation lavas and

Salada Volcanic Complex have the flattest HREE patterns (GdN/ oligoclase from the Lagunillas Formation lavas. On the other hand, the YbN ≈ 1), but the arc unit is slightly more enriched in these elements plagioclases in the Jurassic arc lavas (Camaraca, Oficina Viz, La Negra (YbN ≈ 12 versus ≈ 22, respectively). Formations and Agua Salada Volcanic Complex) are mainly labradorite, Sr concentrations for all clinopyroxenes are high compared to although they exhibit a broad range of composition, ranging from An74 clinopyroxenes from mid ocean ridge basalts (MORB, Fig. 6a), especially to An01. in the clinopyroxenes from Lagunillas and Picudo Formations. A Alkali feldspars are minor constituents of the studied rocks and they particularly high enrichment in this element is observed in one basaltic can be found as microlites in a few evolved arc andesites (VO175 and andesite of the Picudo Formation (PR-11-168; Fig. 6a), with values VO169 from Camaraca and Oficina Viz formations respectively), a ranging between 86 and 265 ppm, which is four to ten times higher sample of the Picudo Formation (PR-10-41; not shown) and in the alkali than the mean of the other subduction related clinopyroxenes included basalts of the Lagunillas Formation (Fig. 4b). In the latter, they corre- in this work. Contents of Nb, Zr and Hf for most samples are in the range spond mainly to anorthoclase crystals, with a few sanidines (Fig 4b). of MORB clinopyroxenes and together with the analyzed samples form an evolutionary trend similar to that observed in arc clinopyroxenes of 7. Thermobarometry of clinopyroxenes and plagioclases the Aleutians (Yogodzinski and Kelemen, 2007; Fig. 6b). The Lagunillas Formation clinopyroxenes have Zr contents that are two to five times Thermobarometry calculations using various equations (Putirka, greater than MORB (Fig. 6a). 2005, 2008; Putirka et al., 2003) for plagioclase/whole rock, The orthopyroxenes of Picudo Formation basaltic andesite (PR-09- clinopyroxene/whole rock, clinopyroxene only and two pyroxenes 05) display a distinct pattern in the REE diagrams, with depletion in equilibrium were performed on minerals from the Jurassic arc and

LREE relative to HREE (LaN/YbN ≈ 0.04), and lack of a negative Eu back arc rocks (see Fig. 7a for location and unit names). anomaly (Fig. 5e). Estimated crystallization temperatures and depth (pressures) for plagioclase/whole rock equilibrium in arc samples mostly fall (~50%) 6. Feldspar chemistry within a relative narrow range of 1160 to 1200 °C (Fig. 7b) and 18 to 26 km (Fig. 7c), with mean values of 1183 °C and 21.6 km. 6.1. Major elements Clinopyroxene/whole rock equilibrium of the same units yielded lower crystallization temperatures, mostly between 1110 and 1150 °C, Major element chemistry of feldspars is listed in Table 3 and detailed and depths between 8 and 18 km, with mean values of 1120 °C and in Table C (Suppl. content). Description and classiﬁcation is in atoms per 15.6 km. Otherwise, single clinopyroxene equilibration of arc samples, formula unit (a.p.f.u.) normalized to eight oxygens. Previously analyzed suggest a narrow range of temperature mostly between 1150 and samples from the Camaraca, Oﬁcina Viz, La Negra Formations (Jurassic 1175 °C, with a mean value of 1162 °C, with the exception of pyroxenes arc, Oliveros et al., 2007) are included in the calculations and diagrams. in Camaraca Formation rocks that show lower values generally between Most of the studied feldspars are bytownite to andesine plagioclase 1130 and 1155 °C, but with a clear overlap of the upper 25% of the values (Fig. 4b). Some anorthite-rich bytownites are found in lavas of the of this unit respect to other arc units. Calculated crystallization depths Picudo Formation and these samples exhibit a rather continuous using this method range mostly between 6 and 12 km, with a mean fractionation trend from An85 to An37. The sample with the most anor- value of 9 km. Uncommon ortho/clinopyroxene equilibrium pairs thitic plagioclase (An85–64) is the Picudo Formation basalt (PR-10-45) from lavas of the Camaraca Formation (Fig. 7b) shown equilibration and that with the least anorthitic plagioclase is PR-11-168 (An70–37). temperatures of ~1025 °C. 104 P. Rossel et al. / Lithos 239 (2015) 97–113

Fig. 3. Mg# versus major cation abundances for the studied pyroxenes. Pyroxene composition of the Jurassic arc samples are after Oliveros et al. (2007) and this work. Dashed lines group different populations of measurements in Picudo and Lagunillas formations.

Crystallization temperatures estimated for back arc units (Fig. 7d,f) from the basaltic sample PR-10-45, and a lower peak at ~1080 °C group- show that most of the used algorithms overlap between 1150 and ing the rest of the analysis of this unit. The clinopyroxene only calcula- 1200 °C, with higher mean values for plagioclase/whole rock equilibri- tion for the Lagunillas Formation basalts yield temperatures (1050 to um (1172 °C), followed by single clinopyroxene equilibration 1080 °C) which are anomalously lower than the rest of the back arc (1165 °C), and lower mean values for clinopyroxene/whole rock calcu- samples (Fig. 4d). Finally, a few ortho/clinopyroxene pairs from a lations (1150 °C). Anomalous behavior is observed in the plagioclase/ basaltic andesite (PR-10-41) from Picudo Formation yielded a range of whole rock equilibrium of the Picudo Formation samples. The obtained equilibration temperatures between ~975 and 1030 °C (Fig. 4d). values range mainly between 1210 and 1240 °C, with the lower temper- Crystallization pressures calculated for minerals from the back arc atures slightly overlapping the mean values calculated for the other units (Fig. 7c,e) have a more complex behavior compared to tempera- back-arc samples (Fig. 7f). The clinopyroxene/whole rock equilibrium ture plots, but the results using the three different equations overlap calculation yielded a wide range of temperatures following a bimodal in a range of depths between ~8 and 20 km. The most consistent results distribution (Fig. 7f) with a peak at ~1200 °C grouping the results are those obtained on the basis of clinopyroxene composition solely, P. Rossel et al. / Lithos 239 (2015) 97–113 105

Fig. 4. a) Pyroxene quadrilateral diagram after Morimoto et al. (1988) and b) feldspar classiﬁcation diagram for the studied minerals. Jurassic arc data are from Oliveros et al. (2007).ASVC: Agua Salada Volcanic Complex.

with a mean value of ~9.6 km (excluding the pyroxenes from the Picudo homogeneous mantle source, iii) compositional heterogeneities in the Formation basaltic sample PR-10-45, see Section 8.3). On the other magma source, iv) previous fractionation of mineral phases, or a combi- hand, the plagioclase/whole rock barometer show higher values for nation of the above. the Lagunillas and Picudo formations crystals, but for the latter there The Mg# of clinopyroxene increases with increasing crystallization is a large dispersion in the data (Fig. 7e,g). Particularly consistent are temperature (Trigila and DeBenedetti, 1993) and is independent of the results obtained in the Picudo Formation for clinopyroxene/whole pressure (Putirka et al., 1996). On the other hand, Si in clinopyroxene rock and clinopyroxene only barometers, with matching range of is generally replaced by Al in the equilibrium exchange of Jd values, mean value and bimodal distribution of the data (Fig. 7d,f). (NaAlSi2O6)–DiHd (Ca(Mg,Fe)Si2O6) and CaTs (CaAlAlSiO6)–DiHd, which is sensitive to pressure. Experimental results show that Al in 8. Discussion clinopyroxenes increases with decreasing temperature and increasing pressure (Putirka et al., 1996; Trigila and DeBenedetti, 1993), whereas 8.1. Compositional variations in clinopyroxenes as proxies of magma sources Si increases with increasing temperature and decreasing pressure. The Mg#, Al and Si contents in the clinopyroxenes of the first group (pyrox- The major element composition of clinopyroxenes from the Jurassic enes from samples of the Lagunillas and Picudo formations) suggest that arc and back arc units can be divided in two groups that follow distinct these minerals crystallized from a melt that was either colder or deeper- fractionation patterns. The first group, represented by pyroxenes from storage than that from which the second group of pyroxenes crystal- two samples of Picudo Formation (PR-10-45, PR-11-168) and the lized. However, equilibration conditions calculated for the basaltic sam- clinopyroxenes in Lagunillas formation lavas, is characterized by higher ple from the Picudo Formation (PR-10-45; Fig. 7e,f) suggest that the contents of Al, Ti, Ca, Na and lower contents of Si (Fig. 3). The second stagnation level of this back arc melt was at greater depths and higher group, which is represented by pyroxene from samples of the Quebrada temperatures than that of the magmas represented by other samples Vicuñita Beds, Algarrobal Formation, Picudo Formation (PR-09-05) and from arc and back arc (see Section 8.2). Thus, the high content of Al in Jurassic arc rocks, has low Al, Ti, Ca, Na and higher contents of Si. the pyroxenes from the Picudo Formation basalt could be the result of Clinopyroxenes in the remaining sample of Picudo Formation (PR-10- higher pressure, not lower temperature, of crystallization. 41) show features that resemble the second group, but they are slightly Furthermore, the most evolved samples of the Picudo Formation enriched in Al and Na. (PR-09-05 and PR-10-41) have low Al contents (Fig. 3) suggesting crys- According to the discrimination diagrams of Leterrier et al. (1982) tallization pressures similar to the pyroxenes from the Quebrada the clinopyroxenes show a broad range of compositions, typical of Vicuñita Beds, the Algarrobal Formation and the Jurassic arc rocks (de- subduction zones (Fig. 8). The Lagunillas Formation crystals define a pleted second group). This is consistent with the two populations of particular pattern due to their high Ti contents, with clear alkaline temperatures calculated for the Picudo clinopyroxenes (Fig. 7e) and affinities (Fig. 8a), while pyroxenes from the other units have composi- the crystallization order of clinopyroxene and plagioclase observed in tions that resemble the clinopyroxenes from the MORB and the basalts each sample, which suggest that the magmas of the Picudo Formation from spreading centers such as back-arc basins (Fig. 8b), but with a clear may have experienced complex transport and storage histories prior fingerprint of subduction-related magmas (Fig. 8c). The clinopyroxenes to their emplacement during late Jurassic. of the Picudo Formation exhibit a significant alkaline component The observed variations in Na and other incompatible elements in (Fig. 8a inset), but plot near to or within the field of calc-alkaline affinity the clinopyroxenes could be due to different degrees of melting of a (Fig. 8c). Thus, there are slight differences in the genesis of this unit with common magma source, such as a depleted sub-arc mantle. The high respect to the Algarrobal Formation, the Quebrada Vicuñita Beds and contents of Na and REE, and the high Sm/Yb ratios in the clinopyroxenes the Jurassic arc rocks for which the alkaline component is negligible of the Picudo and Lagunillas formations, along with the restricted (Fig. 8a). The likely factors that control the observed differences volume of volcanism in these two units compared to the Jurassic arc between the groups of studied pyroxenes are: i) variations in the rocks and Algarrobal Formation, are in agreement with a lower degree pressure/temperature crystallization conditions and/or composition of of partial melting as proposed by Rossel et al. (2013). The high contents a common parental magma, ii) different degrees of melting of a of Ti, Zr, Nb and Hf in pyroxene crystals from the Lagunillas Formation, 106 P. Rossel et al. / Lithos 239 (2015) 97–113

Fig. 5. Chondrite normalized REE patterns for the clinopyroxenes of the Picudo Formation, Quebrada Vicuñita Beds, Lagunillas Formation and Agua Salada Volcanic Complex. Two analyzed orthopyroxenes in sample PR-09-05 (basalt from the Picudo Formation) show typical LREE depleted pattern. Normalizing values are from Sun and McDonough (1989).

compared to the other arc and back arc units studied in this work, similar proportions of clinopyroxene and olivine and minor amounts of suggest that the parental magmas of this formation derive from an plagioclase (Fig. 9). The compositional pattern of pyroxenes from the OIB-type source (Figs. 3, 6a,b, 8a). Picudo Formation basalt (PR-10-45) suggests the same fractionation Clinopyroxene in OIBs (Bohrson and Clague, 1988; Clague, 1988; assemblage as the second group, but the Na and Al enrichment in these Chen et al., 1992; Fedor and Galar, 1997; Tracy, 1980) and continental crystals must be related to differences in the source. The high-Sr pyrox- alkali basalt (Aoki, 1968, 1970; Arai et al., 2000, 2001; Ho et al., 2000; enes from one basaltic andesite in the same unit (PR-11-168) suggest a Kovács et al., 2004) usually have high contents of Na and Ti. For a source that was either Sr-rich or lacked of plagioclase fractionation (see given concentration of SiO2, the continental alkali basalt clinopyroxenes discussion in Section 8.3). The compositional variation trends of pyrox- have lower Mg# (70–86) than the clinopyroxene in the OIB (79–89) or enes from this sample suggest that fractionation of clinopyroxene and volcanic arcs (N85). The Mg# in the Lagunillas Formation pyroxenes minor plagioclase is the most likely process for magma evolution ranges between 69 and 74, suggesting a close relationship of these (Fig. 9b). lavas to a continental intraplate mantle source. This is consistent and conﬁrms the results of Rossel et al. (2013), who propose the existence 8.2. Mineral–liquid (whole rock) equilibrium and thermobarometry of an OIB-like mantle as the source of the Lagunillas Formation lavas implications based on whole rock geochemistry and radiogenic isotopes. The Zr vs Sr and Sr vs Yb variations among clinopyroxenes of the The thermobarometric results on clinopyroxene for each studied second group (low Al, Ti, Ca, Na) can be explained by fractionation of unit, calculated with the same equation, display a Gaussian distribution P. Rossel et al. / Lithos 239 (2015) 97–113 107

with a well-defined peak for both P and T. This behavior can be interpreted as indicative of preferred or protracted crystallization conditions for the clinopyroxenes, suggesting the existence of levels of stagnation for the magmas before eruption, which is consistent with the porphyritic nature of most of the analyzed samples. The magma stagnations levels could have been favored by compositional disconti- nuities in the mantle and crust during the Jurassic. The inferred temperature and depth of these apparent stagnation levels (i.e., the crystallization conditions for the clinopyroxene and plagioclase) seem to be highly dependent on the equation used (Fig. 10), especially in the arc samples. In general, calculations based on plagioclase chemistry show higher values than those based on pyroxene, which is consistent with petrographic observations that suggest an early crystallization of olivine, when present, followed by plagioclase and finally pyroxene. Otherwise, the gap between the P–T conditions calculated for plagioclase or pyroxene could be in part due to the porphyritic nature of the lavas or to the alteration degree of the rocks. Given that the results based on clinopyroxene only geothermobarometry show more narrow and concordant ranges than the rock/mineral methods, it is likely that the alteration processes affecting the bulk-rock chemistry biased the thermobarometric calculations. Furthermore, few measurements of the studied minerals plot on the equilibrium line, which is the method proposed by Putirka et al. (2003) for testing the validity of the thermobarometric results (Fig. 10a). Only four samples plot on the equilibrium line which matches measured and predicted composition for the clinopyroxene, and therefore the calculated pressures and temperatures probably approach their crystallization conditions. These are two lavas from the La Negra Formation in the arc domain, VO04 and VO69 (Oliveros et al., 2007), and two lavas from the back arc domain, PR-10-45 (Picudo Formation) and PR-10-72 (Lagunillas formation). In order to evaluate the reliability of the remaining calculations, we performed a comparison between the measured REE contents on back arc samples and the predicted REE chemistry of these rocks based on clinopyroxene REE composition (Fig. 10b). The basalt and andesite of the Picudo Formation (PR-10-45 and PR-09-05, respectively) along with one sample of the Quebrada Vicuñita Beds (PR- 10-32) show a good fit between measured whole-rock and clinopyroxene based REE contents, suggesting lower degrees of alteration for these samples. For the other analyzed samples and for the plagioclase/whole rock equation, which yielded systematically higher P–T values, there is Fig. 6. a) Sr versus Zr and b) Sr/Zr versus Nd/Yb diagrams for clinopyroxenes. Aleutian arc no further means to assess the reliability of the obtained results. trend after Yogodzinski and Kelemen (2007). MORB data are from Dick and Naslund The obtained temperatures and depths of equilibrated samples from (1996). the arc are ~1100 °C and ~6–18 km (Fig. 7b,c); the mean temperature for all samples and measurements calculated using the clinopyroxene/

Table 3 Mean and range of values for EMP analysis of feldspar crystals from the studied units and the previously published Jurassic arc samples (Oliveros et al., 2007).

Lagunillas Formation Quebrada Vicuñita Beds Picudo Formation Algarrobal Formation Jurassic Arc

N=9 N=19 N=38 N=8 N=220

Mean Range Mean Range Mean Range Mean Range Mean Range

SiO2 58.13 53.27/64.93 53.66 48.06/59.24 52.05 47.04/63.89 53.39 53.00/53.94 54.79 41.27/69.66

Al2O3 25.71 20.42/29.56 28.76 25.13/32.95 30.42 21.30/34.24 28.24 27.45/28.83 27.94 18.48/32.45 FeO 0.48 0.28/0.64 0.87 0.55/0.98 0.74 0.40/1.22 1.01 0.91.1.17 0.89 0.00/15.63 CaO 7.39 2.42/11.55 11.8 6.73/16.52 12.87 2.76/17.34 11.51 10.91/12.12 11.03 0.00/20.42

Na2O 6.13 4.93/7.74 4.74 2.18/7.53 4.04 1.88/6.81 4.81 4.56/5.13 4.76 0.22/11.77

Ka2O 2.09 0.34/6.07 0.32 0.11/0.79 0.39 0.17/5.82 0.37 0.34/0.41 0.75 0.02/18.39 Si 2.62 2.41/2.91 2.43 2.20/2.66 2.36 2.15/2.86 2.44 2.41/2.47 2.48 2.09/3.00 Al 1.37 1.08/1.58 1.54 1.33/1.78 1.62 1.12/1.84 1.52 1.48/1.55 1.49 1.01/1.91 Fe2+ 0.16 0.009/0.022 0.03 0.02/0.04 0.03 0.02/0.05 0.04 0.03/0.04 0.03 0.00/0.61 Ca 0.36 0.07/0.56 0.57 0.32/0.81 0.63 0.13/0.85 0.56 0.54/0.59 0.55 0.00/1.10 Na 0.54 0.43/0.67 0.42 0.19/0.66 0.36 0.14/0.59 0.43 0.40/0.46 0.41 0.02/0.98 K 0.12 0.022/0.35 0.02 0.01/0.05 0.02 0.01/0.33 0.02 0.02/0.024 0.04 0.00/1.09 Ab 52.8 42.7/65.2 41.2 19.2/64.0 35.4 13.8/59.1 42.2 39.7/44.9 41.88 1.83/98.95 An 35.4 6.7/65.3 56.9 31.6/80.2 62.4 13.0/85.8 55.7 52.9/58.3 53.75 0.01/90.96 Or 11.8 2.0/34.1 1.8 0.6/4.4 2.2 1.0/32.6 2.1 1.9/2.4 4.37 0.12/98.01

Structural formulae on the basis of 6 oxygens. N = number of measurements per sample. 108 P. Rossel et al. / Lithos 239 (2015) 97–113

Fig. 7. Calculated crystallization conditions for plagioclases and pyroxenes. Transformation from pressures (kbar) to depth (Km) were made considering a mean density for the crust of 2.9 g/cm3, except for calculations in pyroxenes of sample PR-10-45, in which was used an upper mantle density of 3.3 g/cm3 which induces a variation in calculated crystallization depth of approximately 2.7 to 3.8 km more for cpx/wr equilibrium, and 2.0 to 3.6 km more for cpx only equilibrium than predicted crust densities. a) Location of the studied units. b) Equilibration temperature calculated in arc units. c) Equilibration pressures calculated in arc units. d) Equilibration temperatures calculated in back-arc units. e) Equilibration pressures calculated in back-arc units. The heavy line represents the mean value for all samples using the clinopyroxene only equation, the dashed line represents the mean value of all samples using the clinopyroxene/whole rock equation, the dashed/dotted line represents the mean value of all samples using the plagioclase/whole rock equation. Mean values of equilibrated measurements according to Putirka et al. (2003) are plotted as black lines in corresponding boxes. Heterogeneous results for the Picudo Formation are plotted additionally as histograms of calculated: f) equilibration temperatures (dashed line represents the equilibration temperatures obtained with the clinopyroxene only equation), and g) equilibration pressures (dashed line represents the equilibration pressure obtained with the clinopyroxene only equation). Small circles and stars above and under box plots represent anomalous and extreme measurements respectively, beyond the 95% of confidence. All calculations are after Putirka (2005, 2008) and Putirka et al. (2003). whole rock equation is only slightly lower and the pressure ranges Other samples from the Picudo Formation show no equilibration be- calculated using the clinopyroxene/whole rock and clinopyroxene tween predicted and measured clinopyroxene composition. They only equations are very similar. The consistency among the calculations yielded lower temperatures than the equilibrated sample (PR-10-45), using equilibrated and non-equilibrated samples suggests that the especially with the clinopyroxene/whole rock equation (Fig. 7f). The obtained P–T conditions probably represent a real stagnation level estimated crystallization depths are also lower than the equilibrated during magma ascent in the arc domain. Furthermore, the calculated samples, whether using the clinopyroxene/whole rock or clinopyroxene pressure conditions are in agreement with the estimated depths of the only equation. These results are in agreement with the more evolved lower-middle crust boundary in extensional convergent margins nature of the samples and overlap the calculated P–T conditions for (Christensen and Mooney, 1995; Greene et al., 2006; Rudnick and other arc and back arc samples. However, given the lack of evidence Fountain, 1995) and similar to the emplacement depths for Jurassic for equilibrium in the samples, with the exception of PR-09-05 (good plutonic bodies in Coastal Cordillera of northern Chile (González, match between measured and calculated REE, Fig. 10b), the data should 1999) that are commonly proposed as feeders of the Jurassic arc volca- be taken with caution. nism (Lucassen and Franz, 1994). The Lagunillas Formation alkaline basalts are equigranular, fine For the back-arc domain, the results using the different equations in grained and have a restricted range of Mg#, suggesting a history of equilibrated and non-equilibrated samples from the Quebrada Vicuñita fast cooling and very little fractionation for the magmas. However, Beds and Algarrobal Formation show narrow and overlapping ranges they show a wide range of thermobarometric data, with no concordance for crystallization temperatures and depths (~1150–1200 °C and 8 to between different equations. The calculations could then be geologically 20 km) (Fig. 7d,e) which suggests that the data are valid. The Picudo meaningless, even though a few pyroxenes show apparent equilibrium Formation measurements show a wide range of equilibration condi- in measured versus predicted clinopyroxene composition (Fig. 10a). tions (Fig. 7f,g). Crystals in equilibrium with whole rock composition Taking into account that the crystallization depths estimated from the Picudo Formation basalt (PR-10-45) indicate equilibration from different geothermobarometers are similar for samples of the temperatures higher than 1200 °C and depths of 35–36 km. The same arc and back arc domains, it is possible to infer the existence of results are obtained with clinopyroxene only equation. These conditions magma stagnation levels (or protracted P–T conditions for the crys- are somewhat higher than expected for mantle–lower crust boundary tallization of the phenocrysts) consistent with the existence of a in the back arc region of extensional subduction systems. Thus, the thinned continental crust as result of the transtensional/extensional calculated depth of crystallization of ~35–36 km of this sample could regime that dominated the Andean margin during the Jurassic reflect another process. (Creixell et al., 2011; Grocott and Taylor, 2002; Scheuber and P. Rossel et al. / Lithos 239 (2015) 97–113 109

Fig. 9. a) Sr vs Zr and b) Yb vs Sr diagrams with the proposed fractionation patterns for the analyzed clinopyroxenes. Gray areas show the distribution of the Picudo Formation samples. Black arrows indicate the crystallization patterns for the given fractionated mineral phases. Cpx: clinopyroxenes. Plg: plagioclase. Ol: olivine. Gt: garnet.

back arc domains, which probably corresponded to the lower- middle crust boundary. The existence of different reservoirs or stagnation levels has also been proposed for subduction-related middle Jurassic dyke swarms in central Chile (Creixell et al., 2009).

8.3. Petrogenesis of the Picudo Formation lavas and implications for the tectonics of the early Andean subduction

Fig. 8. a) Ca + Na versus Ti, b) Ca versus Ti + Cr and c) AlT versus Ti discrimination dia- Conditions predicted for the Picudo Formation basalts seems grams after Leterrier et al. (1982). Dashed lines group different populations of measure- ﬁ ments in Picudo Formation. dif cult to be explained by a buoyancy driven stagnation level in a extensional back arc framework. In fact, the high diversity of geochemical compositions observed in this single unit over a restricted geo- González, 1999). Furthermore, based on the obtained results we sug- graphical area, which include typical arc related calc-alkaline gest that the architecture of the Jurassic arc crust had at least one andesites (PR-09-05), slightly Nb-enriched basalts (PR-10-45) and main stagnation level between ~8 to 15 km, in both the arc and high-Sr basaltic andesites (PR-09-06 and PR-11-168), is surprising. 110 P. Rossel et al. / Lithos 239 (2015) 97–113

setting proposed for the Jurassic and Lower Cretaceous in northern and central Chile. Melting of a subducted eclogitized slab, which produces senso stricto adakites (Defant and Drummond, 1990), is not a likely scenario given that (a) the subducted slab was probably old and cold during late Juras- sic (Charrier et al., 2007) and (b) the depth of magma generation for the Picudo Formation volcanism was at least 100 km beneath the volcanic centers, if they were located in the back-arc, and for normal thermal conditions, melting of the slab is not possible at these depths (Mibe et al., 2011). Equally, melting of eclogitized rocks under a thickened continental margin, which produces sensu lato adakites, is unrealistic given the protracted extensional setting observed in the Andean margin since the Early Triassic (Charrier et al., 2007; Mpodozis and Ramos, 1989). Recent works of Goss and Kay (2009) and Goss et al. (2013) have shown the existence of adakitic-like volcanics in the transition between the Chilean flat-slab and the Central Volcanic Zone, apparently related to the tectonic erosion and subsequent assimilation of fore-arc crust. Since no flat slab is observed in the area during Jurassic times this hypothesis could be hard to apply. MacPherson et al. (2006) have proposed another mechanism to produce adakite-like signature in subduction-related magmas that may apply to the samples of the Picudo Formation: high pressure fractionation of basaltic magmas beneath an island arc. Such a mechanism can explain the high variation in the La/Yb (HREE) ratios, without significant changes in the isotopic composition, of the Picudo formation lavas, but fails to reproduce their high Sr contents. Based on the available data published by Rossel et al. (2013) and Lucassen et al. (2001) we constructed isotope mixing curves between MORB and two other end members: the Paleozoic granitic rocks and

Fig. 10. a) Schematic presentation of predicted versus measured clinopyroxene compo- nents, after Putirka et al. (2003). Black dots that plot “on the equiline” represent clinopyroxene calculations from samples (VO04 and VO69 from the arc and PR-10-45 and PR-10-72 from the back-arc) that probably are closer to real crystallization conditions, compared to measurements that fall in shaded area. b) Modeled REE composition of the melts in equilibrium with the analyzed clinopyroxenes (full and dashed lines). Gray lines are the result of the use of partition coefficients for clinopyroxenes in basaltic melts and black lines are the result of the use of partition coefficients for clinopyroxenes in andesitic melts. Dashed lines in sample PR-10-72 are extreme partition coefficients (High and Low) for clinopyroxenes in alkaline basaltic systems. The gray squares represent whole rock REE contents. Fig. 11. Sr/Y vs Y diagram for the basaltic to andesitic samples from the arc and back-arc included in this work. Bulk-rock data are from Rossel et al. (2013). A and B are partial melting curves of a mixture of MORB and subducted sediments in a proportion of 8/2 leaving a residue with modal composition of eclogite (clinopyroxene:garnet:rutile = 85:14:1 in A and 90:9:1 in B). C, equilibrium crystallization curve of a basaltic melt with the composi- These heterogeneities were attributed to secondary alteration processes tion of sample PR-10-45(*) and fractionating minerals in these proportions and remobilization of LILE by Rossel et al. (2013).However,thehighSr garnet:olivine:clinopyroxene = 5:55:40. D, partial melting curve of a mixture of depleted content in the fresh clinopyroxene phenocrysts is undoubtedly a mantle and subducted sediments in a proportion of 8/2 leaving a residue of harzburgite (olivine:orthopyroxene:clinopyroxene:garnet = 59:30:9:2). Sr and Y values for MORB magmatic feature. Furthermore, the high Sr lavas also have low contents and depleted mantle MORB are from Workman and Hart (2005) and for subducted sedi- of HREE and MgO, high Al2O3 concentrations and intermediate SiO2 ments are from Plank and Langmuir (1998). Mineral melt partition coefficients are 0.073 contents (Rossel et al., 2013), that represent, with the exception of the (Sr) and 1.63 for (Y) for clinopyroxene, 0.03 (Sr) and 9.3 (Y) for garnet, 0.036 (Sr) and low silica content, a typical adakitic signature (Defant and Drummond, 0.0076 (Y) for rutile, 0.01 (Sr) and 0.036 (Y) for olivine and 0.03 (Sr) and 0.01 (Y) for orthopyroxene. Clinopyroxene, orthopyroxene and garnet coefficients are from Barth 1990), as shown in the Sr/Y vs Y diagram (Fig. 11). In order to explain et al. (2002); rutile coefficients are from Foley et al. (2000) and olivine coefficients are this geochemical signature, it is necessary to compare the geological from Philpotts and Schnetzler (1970) and Nielsen et al. (1992). BADR: “normal” arc ba- frameworks required to produce adakitic-like lavas to the geological salt–andesite–dacite–rhyolite field. P. Rossel et al. / Lithos 239 (2015) 97–113 111 the Lagunillas Formation average (Fig. A, Supl. content). Our results from the Lagunillas Formation, have enrichments in Ti, Zr, Hf, and Nb, show that the best way to achieve the isotopic composition for most and steeper REE patterns compared to the former group. The composi- of the Picudo Formation lavas is a mixture of 80% of MORB and 20% of tional features observed in the Picudo Formation clinopyroxenes are in- Paleozoic rocks. dicative of complex petrogenetic processes that involve subduction/ We consider unlikely that the upper plate Paleozoic rocks were erosion of enriched sediments or forearc rocks and subsequent high introduced via crustal contamination of the magmas because there are pressure fractionation that led to the formation of adakitic melts, no significant variations in the Sr isotopic composition of the Picudo For- whereas the compositional features observed in clinopyroxenes of mation in spite of their range of silica concentrations (SiO2 ~45to53%). Lagunillas Formation suggests anhydrous melting of an enriched source Sediment subduction is thus a more plausible mechanism to achieve the such as the subcontinental mantle. composition of the Picudo Formation lavas, because it implies that the iso- Trace element variations in the studied samples suggest that topic enrichment takes place before fractionation of the magmas. clinopyroxene and olivine extraction are responsible for the fraction- Forward modeling was performed in order to explain the high Sr/Y ation patterns observed in the back-arc units Quebrada Vicuñita Beds and low Y contents of the Picudo Formation samples (Fig. 11). Based and Algarrobal Formation, whereas in the Picudo Formation magmas, on the most likely magma sources for this unit derived from the isotopic garnet is fractionated. composition of its volcanic rocks (see above and Fig. A of the suppl. The petrographic analysis and the thermobarometry calculations content), two types of mixture were used in the modeling: MORB/sed- suggest that plagioclase was the first Ca phase to crystallize in most of iments and Average Depleted Mantle/sediments. The best fit for the cal- the studied samples. Clinopyroxene crystallization conditions indicate culated curves is achieved using the composition of the present-day slightly lower but overlapping temperatures and pressures for the high Sr sediments from the Peru trench (Plank and Langmuir, 1998). crystals belonging to arc units, relative to the back-arc ones. The results Thus, the Sr and Y contents of the Picudo formation adakite-like lavas suggest the presence of at least one main stagnation level for the can be attained through 1) one stage melting (3–5%) of a mixture of magmas, both in the arc and back arc domains, located at ~8–15 km sediments and oceanic basalts or 2) a two stage process that starts (and 1100 °C) below the surface. This depth could represent the with melting (~25%) of a mixture of depleted mantle and sediments lower-middle crust boundary. A sub-group of analysis from the Picudo to achieve the composition of the enriched basaltic sample PR-10-45, Formation pyroxenes suggest deeper crystallization levels at ~35– and subsequent high pressure fractionation of olivine, clinopyroxene 36 km (and ~1200 °C) below the back arc area, which could be related and minor garnet. to high pressure fractionation and resulted in the adakitic-like signature

The one-stage model fails to explain the low SiO2 and Mg contents observed in two samples. The interpretation of the thermobarometric (~52% and ~3% respectively) of the adakite-like lavas, and it is not results in Lagunillas Formation lavas is not straightforward, but the realistic considering the expected thermal conditions of the slab equigranular textures and limited range of Mg# suggest rapid ascent below the back-arc domain (Mibe et al., 2011). The basaltic sample of the magmas to the surface through the lithosphere and very little (PR-10-45) follows the melting curve of MORB and sediments but the fractionation. contents of Ni and Cr are higher whereas SiO2 is significantly lower The obtained geochemical and thermobarometric results are consis- than the average MORB, which make a genetic relationship with this tent with the main extensional–transtensional framework proposed for raw endmember questionable. The two stage modeling on the other the Andean margin during Jurassic and Early Cretaceous. hand, can explain both the elemental heterogeneity and the isotopic ho- Supplementary data to this article can be found online at http://dx. mogeneity of the Picudo Formation lavas. Moreover, the crystallization doi.org/10.1016/j.lithos.2015.10.002. conditions calculated in clinopyroxenes of Picudo Basalt (PR-10-45) are in agreement with the high pressure fractionation model proposed Acknowledgments by MacPherson et al. (2006). Otherwise, if no sediments are involved, it is possible to infer the This study was funded by the FONDECYT No. 11080040 (Verónica subduction erosion of the Jurassic fore-arc as the source of the high-Sr Oliveros) and Romanian National Sciences Foundation (UEFISCDI) component, as proposed by Goss and Kay (2009) and Goss et al. grant PHII-ID-PCE-2011-3-0217, and a research grant from Exxon- (2013) for Miocene adakites in the flat slab region at 27°–30°S. If this Mobil to the University of Arizona (Mihai Ducea). G. Saldías and C. is the case, a contractional pulse during late Jurassic is needed in order Marquardt (Antofagasta Minerals), L. Serrano and M. Ramírez (Lumina to trigger subduction erosion. Scheubert and González (1999), Creixell Copper), C. Jara (Barrick Exploraciones) and C. Cornejo (Goldcorp) are et al. (2011) and Ring et al. (2012), propose the existence of a thanked for granting access to outcrops and providing logistic support transpressional event, recorded in the Middle to Upper Jurassic arc that was very important for the development of this study. D. R. Hervig units from northern and central Chile. The transpression could have of and K. Domanik are thanked for assistance in analytical work. R. caused a minor event of subduction erosion that fed the magmatic Merino and M. Labbé are thanked for assistance in the field work. P. system and mixed with mantle melts beneath the back-arc area, pro- Vásquez and M. Roden are thanked for the thorough and detailed ducing the adakite-like lavas of the Picudo Formation lavas. reviews that significantly improved earlier versions of this manuscript.

9. Conclusions References

Aguirre, G., Levi, B., Nyström, J.O., 1989. The link between metamorphism, volcanism and Major and trace element chemistry of clinopyroxene phenocrysts geotectonic setting during the evolution of the Andes. In: Daly, J.S., Cliff, R.A., Yardley, from altered Jurassic volcanic rocks cropping out in northern Chile B.W.D. (Eds.), Evolution of Metamorphic Belts. Geological Society Special Publication were studied in order to gain petrogenetic information otherwise diffi- 43, pp. 223–232. cult to obtain via a conventional whole-rock study. Clinopyroxenes are Aoki, K., 1968. Petrogenesis of ultrabasic and basic inclusions in alkali basalts, Iki island, Japan. American Mineralogist 53, 241–256. classified in two groups: the first are augite phenocyrsts in rocks of Aoki, K., 1970. Petrology of Kaersutite-bearing ultramaficandmafic inclusions in Iki the main Jurassic arc, Algarrobal Formation and Quebrada Vicuñita island, Japan. Contributions to Mineralogy and Petrology 25, 270–283. Beds. These pyroxenes are depleted in Ti, trace elements such as Zr, Hf Arai, S., Hirai, H., Uto, K., 2000. Mantle peridotite xenoliths from the Southwest Japan arc: fl a model for the sub arc upper mantle structure and composition of the Western and have at REE patterns, suggesting that the parental melts originated Pacific rim. Journal of Mineralogical and Petrological Sciences 95, 9–23. from depleted mantle sources typical for the supra subduction astheno- Arai, S., Abe, N., Hirai, H., Shimizu, Y., 2001. Geological, petrographical and petrological sphere. The second group comprises aluminous diopsides from the characteristics of ultramafic–mafic xenoliths in Kurose and Takashima, northern Kyu- shu, Southwestern Japan. Science Reports of Kanazawa University. 46 pp. 9–38. Picudo Formation and aluminous to aluminous-subsilic-titanian diop- Barth, M.G., Foley, S.F., Horn, I., 2002. Partial melting in Archean subduction zones: sides from the Lagunillas Formation. These pyroxenes, especially those constraints from experimentally determined trace element partition coefficients 112 P. Rossel et al. / Lithos 239 (2015) 97–113

between eclogitic minerals and tonalitic melts under upper mantle conditions. Kovács, I., Jajacz, Z., Szabó, C., 2004. Type-II xenoliths and related metasomatism from the Precambrian Research 113, 323–340. Nograd–Gomor Volcanic Field, Carpathian–Pannonian region (northern Hungry– Bohrson, W., Clague, D., 1988. Origin of ultramafic xenoliths containing exsolved pyrox- southern Slovakia). Tectonophysics 393, 139–161. enes from Hualalai volcano, Hawaii. Contributions to Mineralogy and Petrology Kramer, W., Siebel, W., Romer, R., Haase, G., Zimmer, M., Ehrlichmann, R., 2005. Geochem- 100, 139–155. ical and isotopic characteristics and evolution of the Jurassic volcanic arc between Bowen, N.L., 1928. The Evolution of Igneous Rocks. Princeton University Press, New Jersey. Arica (18°30′S) and Tocopilla (22°S), North Chilean Coastal Cordillera. Chemie der Buchelt, M., Tellez, C., 1988. The Jurassic La Negra Formation in the area of Antofagasta, Erde 65, 47–78. northern Chile (lithology, petrography, geochemistry). In: Bahlburg, H., Breitkreuz, Leterrier, J., Maury, R.C., Thonon, P., Girard, D., Marchal, M., 1982. Clinopyroxene compo- C., Giese, P. (Eds.), The Southern Central AndesLecture Notes in Earth Sciences 17. sition as a method of identification of the magmatic affinities of paleo-volcanic series. Springer, Heidelberg, pp. 171–182. Earth and Planetary Science Letters 59, 139–154. Charrier, R., Pinto, L., Rodriguez, M.P., 2007. Tectonoestratigraphic evolution of the Ande- Levi, B., Aguirre, L., Nyström, J.O., Padilla, H., Vergara, M., 1989. Low-grade regional meta- an Orogen in Chile. In: Moreno, T., Gibbons, W. (Eds.), The Geology of Chile. The Geo- morphism in the Mesozoic-Cenozoic volcanic sequences of the Central Andes. Journal logical Society, London, pp. 21–144. of Metamorphic Geology 7, 487–495. Chen, C., Presnall, D., Stern, R., 1992. Petrogenesis of ultramafic xenoliths from the 1800 Losert, J., 1974. The formation of stratiform copper deposits in relation to alteration of vol- Kaupulehu flow, Hualalai volcano, Hawaii. Journal of Petrology 33, 163–202. canic series (on north Chilean examples). Rezpravy Éeskolovenské Akad. Vid. Rocnik. Christensen, N., Mooney, W., 1995. Seismic velocity structure and composition of the con- 84 pp. 1–77. tinental crust: a global view. Journal of Geophysical Research 100, 9761–9788. Lucassen, F., Franz, G., 1994. Arc related Jurassic igneous and meta-igneous rocks in the Clague, D., 1988. Petrology and ultramafic xenoliths from Loihi Seamount, Hawaii. Journal Coastal Cordillera of northern Chile/Region Antofagasta. Lithos 32, 273–298. of Petrology 29, 1161–1186. Lucassen, F., Becchio, R., Harmon, R., Kasemann, S., Franz, G., Trumbull, R., Wilke, H.G., Cornejo, P., Mpodozis, C., Tomlinson, A.J. 1998. Hoja Salar de Maricunga, Región de Romer, R.L., Dulsky, P., 2001. Composition and density model of the continental Atacama. Servicio Nacional de Geología y Minería. Mapas Geológicos 7: 1 mapa 1: crust at an active continental margin — the Central Andes between 21° and 27°S. 100,000. Santiago. Tectonophysics 341, 195–223. Creixell, C., Parada, M.A., Roperch, P., Morata, D., Arriagada, C., Pérez de Arce, C., 2006. Lucassen, F., Kramer, W., Bartsch, V., Wilke, H.G., Franz, G., Romer, R.L., Dulski, P., 2006. Nd, Syntectonic emplacement of the Middle Jurassic Concón MaficDikeSwarm,Coastal Pb and Sr isotope composition of juvenile magmatism in the Mesozoic large magmat- Range, central Chile (33°S). Tectonophysics 425, 101–122. ic province of northern Chile (18°–27°S): indications for a uniform subarc mantle. Creixell, C., Parada, M.A., Morata, D., Roperch, P., Arriagada, C., 2009. The genetic relation- Contributions to Mineralogy and Petrology 152, 571–589. ship between mafic dike swarms and plutonic reservoirs in the mesozoic of central MacPherson, C.G., Dreher, S.T., Thirlwall, M.F., 2006. Adakites without slab melting: high chile (30°–33°45′S): insights from AMS and geochemistry. International Journal of pressure differentiation of island arc magma, Mindanao, the Philippines. Earth and Earth Sciences 98, 177–201. Planetary Science Letters 243, 581–593. Creixell, C., Parada, M.A., Morata, D., Vásquez, P., Pérez de Arce, C., Arriagada, C., 2011. Marks, M., Halama, R., Wenzel, T., Markl, G., 2004. Trace element variations in clinopyroxene Middle-Late Jurassic to Early Cretaceous transtension and transpression during arc and amphibole from alkaline to peralkaline syenites and granites: implications for building in central Chile: evidence from mafic dike swarms. Andean Geology 38 mineral–melt trace-element partitioning. Chemical Geology 211, 185–215. (1), 37–63. http://dx.doi.org/10.5027/andgeoV38n1-a04. Martínez, F., Arriagada, C., Mpodozis, C., Peña, M., 2012. The Lautaro basin: a record of Creixell, C., Ortiz, M., Arévalo, C., 2012. Geología del área Carrizalillo-El Tofo, Regiones de inversion tectonics in northern Chile. Andean Geology 39 (2), 258–278. Atacama y Coquimbo. Servicio Nacional de Geología y Minería. Serie Geología Básica Mibe, K., Kawamoto, T., Matsukage, K.N., Fei, Y., Ono, S., 2011. Slab melting versus slab de- N 133 p. 134 Santiago. hydration in subduction zone magmatism. Proceedings of the National Academy of Dallmeyer, D., Brown, M., Grocott, J., Taylor, G., Treloar, P., 1996. Mesozoic magmatic and Sciences 108 (20), 8177– 8182. tectonic events within the andean plate boundary zone, 26°–27°30′, North Chile: Mpodozis, C., Cornejo, P. 1986. Hoja Pisco Elqui. Servicio Nacional de Geología y Minería, constrains from 40Ar/39Ar mineral ages. Journal of Geology 104, 19–40. Carta Geológica de Chile. escala 1: 250,000. Defant, M.J., Drummond, M.S., 1990. Derivation of some modern arc magmas by melting Mpodozis, C., Ramos, V., 1989. The Andes of Chile and Argentina. In: Ericksen, G.E., Cañas, of young subducted lithosphere. Nature 347, 662–665. M.T., Reinemund, J.A. (Eds.), Geology of the Andes and Its Relation to Hydrocarbon Dick, H., Naslund, H., 1996. 1996. Late-stage melt evolution and transport in the shallow and Energy Resources. Circum Pacific Council for Energy and Hydrothermal Re- mantle beneath the East Pacific Rise. In: Mevel, C., Gillis, K.M., Allan, J.F., Meyer, P.S. sources, Huston, Earth Science Series 11, pp. 59–90. (Eds.), Proc. ODP Sci. Results vol. 147. Ocean Drilling Program, College Station, TX, Morimoto, N., Fabrics, J., Ferguson, A.K., Ginzburg, I.V., Ross, N., Seifert, F.A., Zussman, J., pp. 103–134. Aoki, K., Gottardi, G., 1988. Nomenclature of pyroxenes. Mineralogical Magazine 52, Emparán, C., Calderón, M. 2010. Geología área Ovalle-Peña Blanca, Región de Coquimbo. 535–550. Servicio Nacional de Geología y Minería, Serie Geología Básica xx escala 1:100,000 Muñoz, N., Venegas, R., Telles, C., 1988. La Formacion La Negra: Nuevos antecedentes Emparán C, Pineda G. 2000. Área La Serena–La Iguera. Región de Coquimbo. Servicio estratigráficos en la Cordillera de la Costa de Antofagasta. In Congreso Geológico Nacional de Geología y Minería. Mapa Geológico, No. 18, 1 mapa 1:100.000, Santiago. Chileno, No. 5. Actas 1, A283–A311. Fedor, R., Galar, P., 1997. A view into the subsurface of Mauna Kea volcano, Hawaii: crys- Nielsen, R.L., Gallahan, W.E., Newberger, F., 1992. Experimentally determined mineral– tallization processes interpreted through the petrology and petrography of gabbroic melt partition coefficients for Sc, Y and REE for olivine, orthopyroxene, pigeonite, and ultramafic xenoliths. Journal of Petrology 38, 581–624. magnetite and ilmenite. Contributions to Mineralogy and Petrology 110, 488–499. Foley, S.F., Barth, M.G., Jenner, A., 2000. Rutile/melt partition coefficients for trace ele- Oliveros, V., Féraud, G., Aguirre, L., Fornari, M., Morata, D., 2006. The Early Andean Mag- ments and an assessment of the influence of rutile on the trace element characteris- matic Province (EAMP): 40Ar/39Ar dating on Mesozoic volcanic and plutonic rocks tics of subduction zone magmas. Geochimica et Cosmochimica Acta 64, 933–938. from the Coastal Cordillera, Northern Chile. Journal of Volcanology and Geothermal González, G., 1999. Mecanismo y profundidad de emplazamiento del Pluton de Cerro Research 157, 311–330. Cristales, Cordillera de la Costa, Antofagasta, Chile. Revista Geologica de Chile 26 Oliveros, V., Morata, D., Aguirre, L., Féraud, G., Fornari, M., 2007. Jurassic to Early Cretaceous (1), 43–66. subduction-related magmatism in the Coastal Cordillera of northern Chile (18°30′– Goss, A.R., Kay, S.M., 2009. Extreme high field strength elements (HFSE) depletion and 24°S): geochemistry and petrogenesis. Revista Geologica de Chile 34, 209–232. near-chondritic Nb/Ta ratios in Central Andean adakite-like lavas (~27°S, ~68°W). Oliveros, V., Simonetti, A., Morata, D., Vergara, M., Aguirre, L., Belmar, M., Calderon, S., Earth and Planetary Science Letters 270, 97–109. 2008. In-Situ U–Pb dating of very low-grade metamorphic titanite in Upper Goss, A.R., Kay, S.M., Mpodozis, C., 2013. Andean adakite-like high-Mg andesites on the Jurassic-Lower Cretaceous volcanic rocks of central Chile, using laser ablation–MC– northern margin of the Chilean–Pampean flat-slab (27–28.5°S) associated with fron- ICP-MS. VI South American Symposium on Isotope Geology. tal arc migration and fore-arc subduction erosion. Journal of Petrology 54 (11), Palacios, C., 1978. The Jurassic Paleovolcanism in Northern Chile. Tubingen University, 2193–2234. Germany, p. 99 (PhD thesis). Greene, A., DeBari, S., Kelemen, P., Blusztajn, J., Clift, P., 2006. A detailed geochemical study Parada, M.A., Nyström, J.O., Levi, B., 1999. Multiple sources for the Coastal Batholith of of island arc crust: the Talkeetna arc section, South–Central Alaska. Journal of Petrol- central Chile (31–34°S): geochemical and Sr–Nd isotopic evidence and tectonic ogy 47 (6), 1051–1093. implications. Lithos 46, 505–521. Grocott, J., Brown, M., Dallmeyer, G., Taylor, G., Treloar, P., 1994. Mechanism of continen- Parada, M.A., López-Escobar, L., Oliveros, V., Fuentes, F., Morata, D., 2007. Andean tal grow in extensional arcs: an example from the Andean Plate Boundary Zone. Ge- magmatism. In: Moreno, T., Gibson, W. (Eds.), The Geology of Chile. The Geological ology 22, 391–394. Society of London, pp. 115–146. Grocott, J., Taylor, G.K., 2002. Magmatic arc fault systems, deformation partitioning and Philpotts, J., Schnetzler, C., 1970. Phenocrysts-matrix partitioning coefficients for K, Rb, Sr emplacement of granitic complexes in the Coastal Cordillera, north Chilean Andes and Ba with applications to anorthosite and basalt genesis. Geochimica et (25°30′Sto27°00′S). Journal of the Geological Society 159 (4), 425–442. Cosmochimica Acta 34, 307–322. Hinton, R.W., 1990. Ion microprobe trace-element analysis of silicates: measurement of Plank, T., Langmuir, C.H., 1998. The geochemical composition of subducting sediment and multi-element glasses. Chemical Geology 83, 11–25. its consequences for the crust and mantle. Chemical Geology 145, 325–394. Ho, K., Chen, J., Smith, A., Juang, W., 2000. Petrogenesis of two groups of pyroxenite from Putirka, K., Johnson, M., Kinzler, R., Walker, D., 1996. Thermobarometry of maficigneous Tungchihsu Penghu Island, Taiwan Strait: implications for mantle metasomatism be- rocks based on clinopyroxene-liquid equilibria, 0–30 kbar. Contributions to Mineral- neath SE China. Chemical Geology 167, 355–372. ogyandPetrology123,92–108. Iriarte, S., Arévalo, C., Mpodozis, C. 1999. Mapa Geológico de la Hoja La Guardia, Región de Putirka, K., Mikaelian, H., Ryerson, F., Shaw, H., 2003. New clinopyroxene-liquid Atacama. Servicio Nacional de Geología y Minería. Mapa Geológico 13, escala 1: thermobarometers for mafic, evolved, and volatile-bearing lava compositions, with 100.000. applications to lavas from Tibet and the Snake River Plain, Idaho. American Mineral- Jaillard, E., Soler, P., Carlier, G., Mourier, T., 1990. Geodynamic evolution of the northern ogist 88, 1542–1554. and central Andes during early to middle Mesozoic times: a Tethyan model. Journal Putirka, K., 2005. Igneous thermometers and barometers based on plagioclase + liquid of the Geological Society 147 (9), 1009–1022. http://dx.doi.org/10.1144/gsjgs.147.6. equilibria: tests of some existing models and new calibrations. American Mineralogist 1009. 90, 336–346. P. Rossel et al. / Lithos 239 (2015) 97–113 113

Putirka, K., 2008. Thermometers and barometers for volcanic systems. In: Putirka, K., Trigila, R., DeBenedetti, A., 1993. Petrogenesis of Vesuvius historical lavas constrained by Tepley, F. (Eds.), Reviews in Mineralogy and Geochemistry 69, pp. 61–120. Pearce element ratio analysis and experimental phase equilibria. Journal of Volcanol- Ramos, V.A., 1999. Plate tectonic setting of the Andean Cordillera. Episodes 22 (3), ogy and Geothermal Research 58, 315–343. 183–190. Tristá-Aguilera, D., Barra, F., Ruiz, J., Morata, D., Talavera-Mendoza, O., Kojima, S., Ferraris, Robinson, D., Bevins, R.E., Aguirre, L., Vergara, M., 2004. A reappraisal of episodic burial F., 2006. Re–Os isotope systematics for the Lince–Estefanía deposit: constraints on metamorphism in the Andes of central Chile. Contributions to Mineralogy and Petrol- the timing and source of copper mineralization in a stratabound copper deposit, ogy 146, 513–528. Coastal Cordillera of Northern Chile. Mineralium Deposita 41, 99–105. Ring, U., Willner, A., Layer, P., Richter, P., 2012. Jurassic to Early Cretaceous postaccretional Vannucci, R., Rampone, E., Picardo, G.B., Ottolini, L., Bottazzi, P., 1993. Ophiolitic sinistral transpression in northcentral Chile (latitudes 31–32°S). Geological Magazine magmatism in the Ligurian Tethys: an ion microprobe study of basaltic 149 (2), 202–220. clinopyroxenes. Contributions to Mineralogy and Petrology 115, 123–137. Rogers, G., Hawkesworth, C.J., 1989. A geochemical traverse across the North Chilean Vergara, M., Levi, B., Nystrom, J., Cancino, A., 1995. Jurassic and Early Cretaceous island arc Andes: evidence for crust generation from the mantle wedge. Earth and Planetary volcanism, extension, and subsidence in the Coast Range of central Chile. Geological Science Letters 91, 271–285. Society of America Bulletin 107, 1427–1440. Rossel, P., Oliveros, V., Ducea, M., Charrier, R., Scaillet, S., Retamal, L., Figueroa, O., 2013. Vicente, J.C., 2005. Dynamic paleogeography of the Jurassic Andean Basin: pattern of The Early Andean subduction system as an analogue to island arcs: evidence from transgression and localisation of main straits through the magmatic arc. Revista de across-arc geochemical variations in northern Chile. Lithos 179 (2), 211–230. la Asociación Geológica Argentina 60, 221–250. Rudnick, R., Fountain, D., 1995. Nature and composition of the continental crust: a lower Vicente, J.C., 2006. Dynamic Paleogeography of the Jurassic Andean Basin: pattern of re- crustal perspective. Reviews of Geophysics 33 (3), 297–309. gression and general considerations on main features. Revista de la Asociación Scheuber, E., González, G., 1999. Tectonics of the Jurassic-Early Cretaceous magmatic arc Geológica Argentina 61, 408–437. of the north Chilean Coastal Cordillera (22°–26°S): a story of crustal deformation Welkner, D., Arevalo, C., Godoy, E. 2006. Geología del área Freirina-El Morado, Región de along a convergent plate boundary. Tectonics 18, 895–910. Atacama. Servicio Nacional de Geología y Minería. Carta geológica de Chile. Serie Shimizu, N., Semet, M.P., Allegre, C.J., 1978. Geochemical applications of quantitative ion- Geología Básica, No 100. 1 mapa escala 1:100,000. Santiago. microprobe analysis. Geochimica et Cosmochimica Acta 42 (9), 1321–1334. Workman, R.K., Hart, S.R., 2005. Major and trace element composition of the depleted SERNAGEOMIN, 2003. Mapa Geológico de Chile: versión digital vol. 4. Publicación MORB mantle (DMM). Earth and Planetary Science Letters 231, 53–72. Geológica Digital, Santiago, Servicio Nacional de Geología y Minería. Yogodzinski, G., Kelemen, P., 2007. Trace elements in clinopyroxenes from Aleutian xeno- Sun, S.S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts: liths: implications for primitive subduction magmatism in an island arc. Earth and implications for mantle composition and processes. Geological Society Special Publi- Planetary Science Letters 257, 617–632. cation 42, 313–345. Zinner, E., Crozaz, G., 1986. A method for the quantitative measurement of rare earth el- Tomilson, A., Cornejo, P., Mpodozis, C. 1999. Hoja Potrerillos, Región de Atacama, vol. 14. ements in the ion microprobe. International Journal of Mass Spectrometry and Ion Servicio Nacional de Geología y Minería. Mapa Geológico. escala 1:100,000. Processes 69, 17–38. Tracy, R., 1980. Petrology and genetic signiﬁcance of an ultramaﬁc xenolith suite from Tahiti. Earth and Planetary Science Letters 48, 80–96.