Geochemical Journal, Vol. 47, pp. 149 to 165, 2013

Shallow-depth melt eduction due to ridge : LA-ICPMS U–Pb igneous and detrital zircon ages from the Triple Junction and the Taitao Peninsula, Chilean

RYO ANMA1* and YUJI ORIHASHI2

1Faculty of Life and Environmental Sciences, University of Tsukuba, Ten-nodai 1-1-1, Tsukuba, Ibaraki 305-8572, Japan 2Earthquake Research Institute, The University of Tokyo, Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-0032, Japan

(Received April 25, 2012; Accepted January 26, 2013)

To understand the processes of melt eduction in a ridge subduction zone, we performed U–Pb dating on zircons sepa- rated from igneous and sedimentary rocks that were newly dredged from the area and from volcanic rocks collected from the Taitao peninsula, southern Chile. The youngest fraction of the U–Pb age population was used to estimate the age of magmatism or sedimentation. Our new results indicate that the fore-arc region became volcanically active over a period of ~0.4 m.y., after obduction of the (~5.7 to 5.2 Ma) from the west and after granite intrusions related to ridge subduction at ~6 Ma. Fore-arc volcanism produced ejecta of basaltic to dacitic compositions and migrated from offshore (~5.3 Ma) to inland (~4.6 Ma) along the Chile Margin Unit that trends northeast–southwest. The volcanism further extended east to produce the dacitic volcanic plug of Pan de Azucar (~4.3 Ma) and lavas in Fjord San Pedro (~2.9 Ma). The migration took place at a rate of ~2.3 cm/y to ~5.3 cm/y. Another intrusion of a granite pluton, widely distributed offshore of the Taitao ophiolite, took place at ~4.0 Ma. Distributions of old detrital zircon and zircon xenocryst ages were used to evaluate, respectively, the influence of subducted sediments and igneous crustal material. Our results indicate that crustal material influenced only Pan de Azucar and Fjord San Pedro dacites; other acidic magmatism shows moderate evidence for incorporation of subducted sediments. Therefore, melts were formed at shallow depths near the triple junction, due mainly to partial melting of the subducted slab and sediments, and then ejected instantaneously. Depleted oceanic materials became anhydrous, and a volcanic gap was formed along the Andean arc.

Keywords: ridge subduction, slab melting, subducted sediment, U–Pb ages, zircon, Chile, LA-ICPMS

to understand not only the processes of magma genesis INTRODUCTION but also the formation of initial continental crusts (Nel- Several authors have suggested that ridge subduction son and Forsythe, 1989). is important for the genesis of large granite batholiths and/ Among current ridge subduction zones (e.g., McCrory or paired metamorphic belts (Iwamori, 2000; Iwamori et et al., 2009), the Chile Triple Junction (CTJ) (Fig. 1a) is al., 2007; Aoya et al., 2009) because of its high heat flow expected to be the best example for a case study because: (DeLong et al., 1979). It has been suggested that igneous (1) the subducting ridge system is a topographically rocks with adakitic compositions are a product of partial prominent feature; (2) the half-spread-rate is ~4 cm/y, and melting of a young and hot subducted slab under garnet– a significant portion of the world’s oceanic crust has amphibolite to eclogite conditions (Rapp et al., 1991; formed under such a rate; and (3) the mid axes of the Defant and Drummond, 1990; Kay et al., 1993; Peacock Chile Ridge system are sub-parallel to the continental et al., 1994; Drummond et al., 1996; Springer and Seck, margin, and thus along-arc and across-arc variation is 1997), and hence, are related to ridge subduction. Fur- expected. ther, compared with the genesis of trondhjemite–tonalite– A chain of active volcanoes in the Southern Volcanic granodiorite (TTG) rocks, they are recognized widely to Zone (SVZ) extends from low latitudes along the Andean form initial continental crusts during Archean (Martin, Mountains (Fig. 1a) and is disrupted at Hudson Volcano 1993, 1996). Therefore, it is important to study the (H in Fig. 1a) near the latitude of CTJ. To the south from magmatism occurring in recent ridge subduction zones CJT, active volcanoes of the Austral Volcanic Zone (AVZ) are sparsely distributed (e.g., Stern et al., 1976; Ramos, 1989; Stern, 2004 and references therein). Volcanoes of *Corresponding author (e-mail: [email protected]) SVZ typically have basaltic to andesitic compositions with Copyright © 2013 by The Geochemical Society of Japan. arc-signatures whereas those of AVZ have adakitic com-

149 Fig. 1. Tectonic settings (a) and geology (b) of the study area. The Patagonian volcanic gap lies in between the Hudson (H) and Lautaro (L) volcanoes. See text for explanations.

positions (Martin, 1993; Lopez-Escobar et al., 1995; Stern extends into the fore-arc region, and another chain of and Kilian, 1996; Stern, 2004; Shinjoe et al., 2013). The Middle to Late Miocene plutonism and Miocene adakitic gap between Hudson Volcano and Lautaro Volcano (L in magmatism (Kay et al., 1993; Ramos et al., 2004; Fig. 1a) at the northern tip of the AVZ is known as the Orihashi et al., 2013) appears behind the active volca- Patagonian Volcanic Gap (Orihashi et al., 2004; Stern, noes of the AVZ. Together with the Patagonian Volcanic 2004). The contrast in rock compositions between the two Fields in the extra back-arc side, this area provides a good volcanic zones and the presence of a volcanic gap are the opportunity to study both along-arc and across-arc varia- first-order along-arc variations observed around the Chile tion of magmas that are spatially and temporally con- Ridge subduction zone. strained (Stern et al., 1990; Ramos and Kay, 1992; Kay Recent geochronological studies reveal the presence et al., 1993; Gorring et al., 1997). of Miocene to Pliocene granites in the Late Jurassic to Finally, the presence of the Taitao ophiolite, exposed Eocene Patagonian Batholiths that extend from 37°S to ~50 km southeast of the current CTJ, and related activi- the southern tip of the South American continent ties of acidic magmas in the westernmost part of the Taitao (Pankhurst et al., 1999; Suarez and De la Cruz, 2001; Peninsula (Fig. 1b) provide a unique opportunity to com- Thomson et al., 2001; Thomson, 2002; Herve et al., 2007). pare the compositions of incoming materials (altered oce- The distribution of such young granite plutons almost anic crust and sediments), magmatic and hydrothermal overlaps with the SVZ volcanoes between latitudes 40°S processes occurring in the sea at present and in the time and 45°S. Hudson Volcano in the southernmost SVZ of the ophiolite obduction, and compositions of magmas marks the location of the back-arc side of the Miocene– as outcomes of these processes (see Section “Geology of Pliocene magmatism. To the south, Miocene plutonism the Taitao peninsula”).

150 R. Anma and Y. Orihashi To understand the processes of melt eduction in a ridge in the eastern sections of the batholith have Late Paleozoic subduction zone and to evaluate the influence of ages (354 to 253 Ma), whereas basement rocks distrib- subducted sediments and lower crustal components on uted in the fore-arc region (including the Taitao penin- magmatism in ridge subduction zones, we performed U– sula) have Late Permian to Jurassic ages (258 to 207 Ma). Pb dating on zircons that were separated from igneous The basement rocks also contain detrital zircons younger rocks and clastic sediments distributed near the CTJ. than 600 Ma, based on U–Pb ages (Thomson and Herve, 2002). The Chile Margin Unit is a mixture of lavas, TECTONIC SETTINGS pyroclastic rocks and semi-consolidated sediments and The study area is characterized by two oceanic plates, was supposedly distributed in the fore-arc area of the the in the north and the in the Chilean margin during the ridge subduction events south. These plates are separated by spreading centers of (Bourgois et al., 1993; Guivel et al., 1999). Mpodozis et the Chile Ridge system and subduct beneath the South al. (1985) reported six whole-rock K–Ar ages ranging American plate. The convergent rates of the two plates, from 4.4 to 2.5 Ma from the Chile Margin Unit. with respect to the , are 9 cm/y and The Taitao ophiolite consists of a completed sequence 2 cm/y, respectively (Fig. 1a) (Herron et al., 1981; Cande typical of a Penrose-type oceanic lithosphere (Anony- et al., 1982; Cande and Leslie, 1986; Tebbens and Cande, mous, 1972); pillow lavas and sheet flows interbedded 1997; Tebbens et al., 1997). The Chile Ridge system has by clastic deposits, sheeted dike complexes, gabbros and a central axis that trends north–northwest (NNW) and ultramafic rocks from top to bottom (Mpodozis et al., consists of several segments separated by fracture zones 1985; Forsythe et al., 1986; Bourgois et al., 1992, 1993; that trend east–northeast (ENE). Nelson et al., 1993) and show evidence of hydrothermal The current plate configuration (Fig. 1a) and plate alteration that is typical of ocean-floor metamorphism kinematic data (Cande and Leslie, 1986; Breitsprecher (Shibuya et al., 2007). The ultramafic rocks provide evi- and Thorkelson, 2009) allow us to deduce that a ~700 km dence for at least two stages of melting. The first stage segment of the spreading Chile Ridge started to subduct took place at ~1.6 Ga and the other stage is recent and at high latitudes approximately 14 Ma. The position of related to magmatism in the Chile Ridge (Schulte et al., the CTJ migrated north until it reached the southern end 2009). The gabbros have compositions similar to normal of the Taitao peninsula at ca. 10 Ma. A series of Mid-Oceanic Ridge basalts (N-MORB) whereas pillow of three short spreading centers (<100 km basalts have enriched compositions (Kaeding et al., 1990; long) took place offshore of the Taitao peninsula at al- Lagabrielle et al., 1994; Le Moigne et al., 1996; Guivel most the same latitude at ca. 6 Ma and ca. 3 Ma. These et al., 1999). U–Pb ages of gabbros and sheeted dikes are subductions occurred because the NNW-trending central ~5.6 Ma and ~5.2 Ma, respectively (Anma et al., 2006). axis of Chile Ridge is oblique to the north–south (NS) The sheeted dikes were thermally metamorphosed trend of the continental margin. The position of the CTJ (Shibuya et al., 2007) by the intrusion of the Seno migrates northward when a spreading ridge subducts Hoppner pluton (SH in Fig. 1b) (Anma et al., 2009a). (Ramos, 1989). In contrast, the triple junction migrated Anma et al. (2009a) reported U–Pb ages of zircons sepa- southward when a fracture zone subducted. Lagabrielle rated from clastic sediments that were interbedded in the et al. (2000) estimated that temperatures of 800 to 900°C volcanic rocks and concluded that their deposition post- were attained at depths of 10 to 20 km below the current dates ~4.9 Ma. Only five zircon grains out of 24 yielded CTJ area because ridge subduction events were persist- ages older than Late Miocene, and all were younger than ent. Spreading ridge subduction may have removed sev- 290 Ma. eral hundred square-kilometers of rock from the fore-arc The term “Taitao granites” is collectively used to in- region due to subduction erosion (Bourgois et al., 1996). dicate granitic rocks distributed around the Taitao ophiolite (Anma et al., 2009a). The Tres Montes pluton (TM in Fig. 1b) has granodioritic to trondhjemitic com- GEOLOGY OF THE TAITAO PENINSULA positions and a U–Pb age of ~5.7 Ma (Anma et al., 2009a). There are four main rock bodies exposed in the The Seno Hoppner pluton has a granitic composition and westernmost part of the Taitao peninsula (Fig. 1b). The U–Pb age of ~5.1 Ma (Anma et al., 2009a). The age of Chonos metamorphic rocks of mainly sedimentary origin the Cabo Raper pluton (CR in Fig. 1b) is ~4.0 Ma (Herve form the basement of the study area. Thomson and Herve et al., 2003) and has granodioritic to trondhjemitic com- (2002) determined both fission-track and U–Pb ages us- positions (Anma et al., 2009a; Kon et al., 2013). The Tres ing zircons separated from basement rocks that are widely Montes and Cabo Raper plutons are distributed in the west distributed in the west and east of the South Patagonian and the Seno Hoppner pluton in the east of the Taitao Batholith. They reported that basement rocks distributed ophiolite (Fig. 1b). The Tres Montes granodiorite includes

Shallow melt eduction due to Chile ridge subduction 151 many inherited zircons (7 out of 15 measurements) with slopes of the tip of the Taitao peninsula. U–Pb ages ranging from 695 to 97 Ma, whereas the Seno The dredging of sediments aimed to obtain composi- Hoppner granite includes only one inherited zircon (405 tions of materials that subduct and eventually melt to form Ma) out of 31 measurements (Anma et al., 2009a). The arc magmas at deep sections of the ridge subduction zone. Cabo Raper granodiorite rarely includes inherited zircon Though volumetrically small, subducted sediments play of Cretaceous age (Herve et al., 2003). an important role in enhancing partial melting of In addition, there are two dacitic volcanic centers with subducted materials. Dredges around the Taitao penin- unknown age in the east of the Taitao ophiolite; Fjordo sula aimed to understand the distribution of the submerged San Pedro-Cupula San Pablo (FSP in Fig. 1b) and Pan de extension of the Taitao ophiolite and granites, the prod- Azucar (PA) (Guivel et al., 1999). Rocks of the Pan de ucts of a ~6 Ma ridge subduction event. Three dredge Azucar dacite have a higher Sr/Y ratio (~25) and lower Y operations (D01, D02 and D06 in Fig. 1b) were conducted content (12 ppm) and lay between the adakitic and arc- for this purpose. volcanic fields on an Sr/Y-Y diagram. Other acidic rocks, Dredge D01 was operated at the basal part of the north- including the Taitao granites, have lower Sr/Y ratios and ern slope of a ridge that trends east–west (EW) and is higher Y contents and plot on the arc dacite-andesite field. located offshore of the southwestern end of the Taitao peninsula. We refer to this ridge as the South Taitao Ridge (STR) in Fig. 1b. From STR, we recovered tuff breccia, OFFSHORE GEOLOGY lapilli tuff and lapilli stones of dacitic composition with A series of dredge operations was previously con- clasts and breccias of andesitic to basaltic compositions ducted along the mid axis of the spreading centers of the (~54 kg), and also recovered consolidated sandstones and subducting Chile Ridge system (Klein and Karsten, 1995; siltstones (~245 kg). Recovered rocks are angular to sub- Karsten et al., 1996; Sherman et al., 1997). Geochemical rounded in shape. Pyroclastic rocks have only sporadic variations were reported along the ridge with enriched Mn-coating, whereas a thin Mn-coating is observed on MORBs near the segment boundaries. Thus, the compo- most sandstones and siltstones. Several structures are sition of the subducting oceanic plate is also variable near observed in the sedimentary rocks: a thin alternation of the CTJ. Rocks from the fore-arc regions were recovered sandstone and mudstone, graded bedding in sandstones, by an Ocean Drilling Program (ODP) (Behrmann et al., load casts, bioturbation in siltstones, micro-normal faults 1994; Guivel et al., 1999) and by a series of dredge op- and vein structures. The thin alternation of sandstone and erations (Guivel et al., 2003). Rock locations covered the mudstone could be interpreted as evidence of distal Taitao Ridge (Fig. 1b) and the fore-arc region to the north turbidites. The recovered pyroclastic and sedimentary of the Taitao Ridge where ridge subduction is ongoing. rocks resemble those of the Chile Margin Unit of the Dacites and basaltic andesites of recent ages were dredged Taitao ophiolite (Fig. 1b), suggesting that the Chile Mar- from the surface of the fore-arc region, and they carry a gin Unit is continuous to the southwest with a general subduction-related geochemical imprint (Guivel et al., northeast–southwest (NE–SW) trend in the Taitao penin- 2003). In contrast, cores obtained by the ODP Leg 141 sula. This general trend swings to EW in the STR. (Sites 859–861) record sedimentation episodes down to Dredge D02 was conducted along the westward- Early Pliocene and imply that subduction accretion had facing slope beneath the Taitao ophiolite. The dredging ceased in the Late Pliocene (Behrmann et al., 1994). recovered granites (~6 kg), consolidated sandstones and The Taitao Ridge consists of both Late Pliocene to siltstones (~46 kg) and a pyroclastic rock (0.12 kg). The Pleistocene basaltic rocks with subalkalic to tholeiitic pyroclastic rock includes breccias that consist of sparsely- compositions and silicic rocks overlain by a 20 m se- pyroxene-phyric basalt to basaltic andesite with a thin quence of silty clays with ages ranging Late Pliocene to altered glassy rim. Structures observed in the sedimen- Pleistocene (Behrmann et al., 1994; Guivel et al., 2003). tary rocks include an alternation of sandstone and Behrmann et al. (1994) suggested that the volcanic rocks mudstone, imbrication of clasts, bioturbation, black seams were probably formed at a ridge-transform intersection. (layer-parallel coherent faults), vein structures and cal- We also conducted a series of dredge operations off cite veins (with and without slicken lines). Foraminifers Taitao peninsula during the MR-06-08 cruise of R/V Mirai were recognized in a few specimens. The sedimentary operated by JAMSTEC. Our goal was to collect igneous rocks have a thin Mn-coating. Granitic rocks resemble to rock and sediment samples from an offshore area that can those of the Cabo Raper intrusion (CR in Fig. 1b) based help us to understand solid earth recycling processes that on mineral assemblage and texture. They are angular to occur today or occurred during the last 6 m.y. subduction sub-angular blocks and have no Mn-coating. We suggest of the Chile Ridge system. The target rocks for dredge that they fell off from a nearby exposure of the Cabo Raper operations were sediments distributed near the ridge sub- pluton. Neither ultramafic nor metamorphic rocks were duction zone, and rocks distributed along the western recovered.

152 R. Anma and Y. Orihashi Fig. 2. Tera-Wasserburg plots for zircons separated from volcanic rocks of the Chile Margin Unit define lower transect ages (left). Weighted mean ages (right) are calculated independently using only grains with concordant ages.

Dredge D06 was conducted in the Taitao Ridge lo- consolidation may be attributed to sedimentation before cated northwest of the Taitao ophiolite (Fig. 1b). Dredge and after the subduction of a segment of the Chile Ridge D06 recovered 194 kg of fine-grained sandstones, system (Anma et al., 2009b). siltstones and claystones. The degree of consolidation of this material is bimodal: one is weakly consolidated, and METHODOLOGY the others are well consolidated. Consolidated sedimen- tary rocks are angular to sub-rounded blocks and have a We separated zircon crystals using a conventional sporadic and thin Mn-coating or none at all. Structures method. The separated grains were mounted on a Teflon observed in consolidated sedimentary rocks include a thin sheet that is free from Pb to allow age determination of a lamina of sandstone, bioturbation, trace fossils, load casts, grain smaller than the diameter of a laser beam. The wavy depositional surfaces, vein structures, orthogonal ICPMS used in this study was a Thermo Elemental brittle fractures, and brecciation. These structures could PlasmaQuad3 (PQ3) installed at ERI, The University of be interpreted as evidence of distal turbidites. In contrast, Tokyo. To obtain a high sensitivity, an S-option interface the weakly consolidated sediments consist mainly of (Hirata and Nesbitt, 1995) and VG CHICANE ion lens siltstones and claystones with no Mn-coating and signifi- (Hirata, 2000) were applied to the PQ3 instrument. We cant deformation structures. The contrasting degree of used a New Wave UP-213 laser ablation system with a

Shallow melt eduction due to Chile ridge subduction 153 Fig. 3. Tera-Wasserburg plots (left) and weighted mean ages (right) for zircons included in dacites from Pan de Azucar (TPD125) and Fjord San Pedro (TPD129).

frequency-quadrupled Nd-YAG UV laser having a 213 resent the age of zircon crystallization (Supplementary nm wavelength. Details of the analytical procedure, in- Table S1). In this study, we calculated both weighted mean cluding its precision and accuracy, are described in ages and lower intercept ages (Stacy and Kramers, 1975) Orihashi et al. (2008). During our runs, age-standard zir- for comparison with previous age data (Herve et al., 2003; con 91500 was measured repeatedly for monitoring. The Anma et al., 2006, 2009a). Only concordant ages were obtained weighted mean of 238U–206Pb ages (N = 45) was used to calculate and compare the weighted mean ages of 1049.3 ± 3.9 Ma (95% conf.), slightly younger than the different rock types (igneous and sedimentary rocks) and recommended value of 1062.4 ± 0.3 Ma (Wiedenbeck et to evaluate the influence of the subducted sediments and al., 1995). Since the difference was only 1.2% and within lower crust on the composition of magmas. the range of analytical accuracy (±2%), we did not apply any calibration. We used the Isoplot program (Ludwig, RESULTS 2001) for calculations of concordia and intercept ages, statistics, and plots. When obtained 238U–206Pb and 235U– Chile Margin Unit (TPD45 and TPD85) 207Pb ages overlapped within the range of analytical er- Out of 30 measurements obtained from zircons in the rors (i.e., degree of discordance (%) was closer to zero matrix of a pyroclastic rock of the Chile Margin Unit than –1), they were considered to be concordant and rep- (TPD45) (Table S1), 15 are concordant and Early Pliocene

154 R. Anma and Y. Orihashi (4%)

(8%)

(6%)

(8%)

(17%)

(17%)

(10%)

694 Ma

903 Ma

Precambrian

784~1108 Ma

587~1000 Ma

590 and 1290 Ma

752 and 1387 Ma

(8%)

(8%)

(3%) n.d. 676~1226 Ma

(13%)

(12%)

(40%)

476 Ma

417~561 Ma

519~590 Ma

469~519 Ma

Early Paleozoic

455 and 561 Ma

467 and 515 Ma

(3%)

(12%)

(17%)

(22%)

(12%)

(10%)

(20%)

255~377 Ma

248~324 Ma

complex

n.d. n.d. n.d. n.d.

n.d. n.d. n.d. n.d.

n.d. 259~404 Ma

n.d. 278~400 Ma

n.d. 275 Ma

n.d. 358 Ma

n.d. 265~394 Ma

(8%)

(4%)

pre-Jurassic accretionary

187 Ma

Mesozoic Late Paleozoic

211 and 232 Ma

(6%)

(7%)

(24%)

(17%)

(10%)

(33%) 29 Ma (10%)

(77%)

(49%)

109 Ma

85~144 Ma

94~146 Ma

67~132 Ma

31~124 Ma

Patagonian batholith

(6%) n.d. n.d. n.d. n.d. n.d. n.d.

n.d. 29 and 90~136 Ma

n.d. 90~128 Ma

n.d. n.d. n.d. n.d. n.d. n.d.

n.d. 59 and 63 Ma

(9%)

group

13 Ma

(12%)

(10%)

(11%)

(10%)

0.28 Ma.

6.6 Ma

Miocene ±

10~17 Ma

14~19 Ma

11~19 Ma

11~16 Ma

0.08 Ma

0.11 Ma

0.16 Ma

0.16 Ma

0.51 Ma*

0.19 Ma

0.10 Ma

0.19 Ma

0.58 Ma

0.76 Ma

±

±

±

±

±

±

±

±

±

±

(88%)

(44%)

(37%)

(90%)

(32%)

(23%)

(30%)

(13%)

(100%)

(100%)

Weighted mean age of

youngest age fraction (%)

24 2.85

26 5.27

17 4.59

25 4.44

21 4.77

35

39 4.43

Number of

concordant ages

ystallization age of D02-G01 granite: i.e., 4.06

dant ages

e 46 5.22

sandatone 10 4.29

Sample/rock description

TPD45 CMU pyroclastics

TPD85 CMU dacite clast

TPD159 PA dacite

TPD129 FSP dacite

D02-G01 granite 4 4.58

D02-G03 granite 21 4.15

D01-V02 volcanic conglomerat

D01-V25 dacite clast

D01-S17 fine-grained

D06-S07 laminated fine-graind sandstone

D06-S20 deformed siltstone

Table 1. Distribution of concor Table age was taken as a cr *Lower intercept

Shallow melt eduction due to Chile ridge subduction 155 Fig. 4. Tera-Wasserburg plots (left) and weighted mean ages (right) for igneous zircons separated from granitic rocks dredged at D02 site (Fig. 1b).

in age. The weighted mean age is 4.59 ± 0.08 Ma. A lower Pan de Azucar dacite (TPD159) intercept age of 4.59 ± 0.12 Ma was obtained using 27 Twenty-eight spot ages were obtained from zircons concordant and discordant Pliocene spot ages (Fig. 2). separated from the Pan de Azucar dacite, and 25 spot ages This specimen also includes zircons with older concord- are concordant among them (Table S1). Eleven concord- ant ages (13 Ma and 109 Ma) and a discordant age (303 ant spot ages are of Late Miocene to Early Pliocene. When Ma). we calculated the weighted mean age, two old spot ages Out of 30 spot ages obtained from zircons in a dacitic were statistically rejected. These two old ages coincide block that was included in the volcani-clastic rocks of with the age of the Taitao ophiolite and the Tres Montes the Chile Margin Unit (TPD85), 21 spot ages are con- pluton, within the error range. We suggest that the older cordant (Table S1). All concordant spot ages are of Late zircons are exotic and related to the Taitao ophiolite/gran- Miocene to Early Pliocene and yield a weighted mean ite. The remaining zircons yield a lower intercept age of age of 4.77 ± 0.11 Ma. A lower intercept age was also 4.30 ± 0.17 Ma and a weighted mean age of 4.44 ± 0.16 calculated to be 4.63 ± 0.09 Ma using 27 concordant and Ma (Fig. 3). discordant Miocene–Pliocene spot ages (Fig. 2). There Other older concordant ages (N = 14) range from 29 were only three zircons with older discordant ages; they to 694 Ma. Among them, six grains (29~136 Ma) coin- range from 13 to 36 Ma. cide with the age of the South Patagonian batholith, two

156 R. Anma and Y. Orihashi Fig. 5. Tera-Wasserburg plots (left) and weighted mean ages (right) for zircons separated from volcani-clastic rocks dredged at D01 site near the crest of the South Taitao Ridge (Fig. 1b).

grains have Triassic ages and three grains have Late one Mesozoic grain and four Late Paleozoic grains coin- Paleozoic ages. The latter five ages are equivalent to the cide with the age of the pre-Jurassic metamorphic rocks age of the pre-Jurassic metamorphic rocks (Table 1). (Table 1). Jurassic–Triassic ages were obtained only from the Pan de Azucar and Fjord San Pedro dacites. Fjord San Pedro dacite (TPD129) Thirty-six spot ages were obtained from zircons sepa- Dredged granites (D02-G01 and D02-G03) rated from the Fjord San Pedro dacite, and 24 spot ages Zircons were separated from two blocks of granite are concordant among them (Table S1). Nine concordant collected during the D02 dredge operation of the MR08- ages are of Late Pliocene and yield a weighted mean age 06 cruise (Fig. 1b). Thirty-one spot ages out of 33 meas- of 2.85 ± 0.16 Ma (Fig. 3). However, no lower intercept urements of the D02-G01 zircons are of Late Miocene to age was obtained from the Late Pliocene fraction due to Early Pliocene and yield a lower intercept age of 4.06 ± an unclear common Pb trend (Watson et al., 1997). 0.28 Ma (Fig. 4). The lower intercept age coincides with Other older concordant ages (N = 15) range from 90 the SHRIMP U–Pb ages of the Cabo Raper pluton (3.97 to 1108 Ma. Among them four grains (90~128 Ma) coin- ± 0.14 Ma to 3.84 ± 0.09 Ma: Herve et al., 2003; Anma et cide with the age of the South Patagonian batholith, and al., 2009a). However, only four concordant ages were

Shallow melt eduction due to Chile ridge subduction 157 Fig. 6. Tera-Wasserburg plots (left) and weighted mean ages (right) for clastic zircons dredged at D01 and D06 sites (Fig. 1b).

obtained, and a weighted mean was calculated to be 4.58 Dredged volcani-clastic rocks (D01-V02 and D01-V25) ± 0.51 Ma, which is slightly older than the age of the We separated zircons from volcani-clastic rocks col- pluton. This result arises because of the limited number lected during the D01 dredge that are identical to the rocks of concordant ages, and we interpret the lower intercept of the Chile Margin Unit found in the on-land exposures age to be the age of the crystallization. This sample also (Fig. 1b). Sample D01-V02 consists of polymictic vol- contains zircon with older discordant ages of 37 Ma and canic conglomerates and D01-V25 is a dacite clast in 189 Ma. volcani-clastic rocks. Zircons separated from another block of dredged gran- Seventeen spot ages among 53 measurements of the ite (D02-G03) yielded a weighted mean age of 4.15 ± 0.19 zircons from sample D01-V02 are of Late Miocene to Ma using 19 concordant spot ages out of 30 measurements Early Pliocene and yield a lower intercept age of 5.00 ± (Fig. 4). A lower intercept age was also calculated using 0.21 Ma. A weighted mean age calculated using 14 con- 28 concordant and discordant ages to be 3.89 ± 0.18 Ma. cordant ages was 5.16 ± 0.17 Ma (Fig. 5). One zircon Both ages coincide with the age of the Cabo Raper pluton grain is significantly young and was statistically excluded within the error range. The sample also contains zircons from the calculation. This young grain appears because with concordant ages of 59 Ma and 63 Ma. The Cabo the D01-V02 sample consists of polymictic conglomer- Raper pluton thus extends NNW from the on-land expo- ates and we expect an incorporation of younger detrital sures in the westernmost tip of the Taitao peninsula. zircon during reworking. Instead, the younger age (~4.2

158 R. Anma and Y. Orihashi Ma) may limit the age of sedimentation. This sample con- tains 31 zircons with older concordant ages ranging from 10 to 1000 Ma. Among them, four grains are of Middle Miocene (10 to 17 Ma), three grains are the age of the Cretaceous South Patagonian batholith (85 to 144 Ma), and 10 grains have Late Paleozoic ages (259 to 404 Ma). Only six Late Miocene to Early Pliocene concordant ages were obtained from 33 zircons found in the D01- V25 dacite clast. The calculated weighted mean age was 5.27 ± 0.19 Ma, and no lower intercept age was obtained (Fig. 5). The dacite clast contains 20 zircons with old concordant ages: three grains have Miocene ages (14 to 19 Ma), nine grains are the age of the Cretaceous South Patagonian batholith (94 to 146 Ma), three grains are of Late Paleozoic (278 to 400 Ma) and the other five grains range from 519 to 1290 Ma.

Dredged sediments We separated zircons from three dredged blocks of Fig. 7. Age distributions of clastic zircons and zircon xenocrysts sediments. Sample D01-S17 is a weakly consolidated fine- included in igneous rocks. Grains with concordant ages were grained massive sandstone. Sample D06-S07 is a well- used for the plot. consolidated siltstone with thin upward-fining laminas that consist of grains ranging in size from very fine to clay-equivalent. The sample is interpreted to be a distal turbidite (Anma et al., 2009b). Sample D06-S20 is a This sample also includes one 2.88 Ma discordant zircon moderately consolidated siltstone with deformation struc- grain. Thirty-four spot ages older than Late Miocene were tures that include faults with breccia and vein structure. also obtained: four grains are Middle Miocene (11 to 16 Young detrital zircons separated from marine Ma), 19 grains are the age of the South Patagonian sediments tend to have discordant ages. We obtained 27 batholith (31 to 124 Ma), eight grains are Late Paleozoic spot ages from D01-S17. Nine concordant and discord- (265 to 394 Ma), and three grains have Precambrian ages ant ages define a Pliocene lower intercept age of 4.23 ± (676 Ma, 724 Ma and 1226 Ma). 0.29 Ma (Fig. 6). Among them, 3 spot ages are concord- ant and yield a weighted mean of 4.29 ± 0.58 Ma. Seven DISCUSSION older concordant ages are obtained as well: three ages are Early Paleozoic (469 to 519 Ma) while the remaining Incoming material and its incorporation into magma four are Late Miocene (6.6 Ma), Paleogene (28.9), Late Oceanic crust rarely contains zircon crystals. Thus, Paleozoic (275) and Precambrian (903 Ma) ages. our method only allows us to evaluate the incorporation Fifty spot ages were obtained from the distal turbidite of sediments into magmas, as sediments subducted with of D06-S07. They include three discordant ages that sug- a sinking oceanic plate, and/or the contamination of gest a deposition after Late Miocene to Early Pliocene magma by the crust, as magma ascended through crustal (4.0~6.1 Ma). However, these ages do not provide a reli- material. Compositions of oceanic crust are therefore ex- able lower intercept age or weighted mean age. The sam- cluded from the following argument. We omit discordant ple includes 35 detrital zircons with older concordant ages. age data both to minimize the influence of old recycled A majority of the zircons (N = 27) are Cretaceous and zircons included in the Patagonian batholith and the base- range from 67 to 132 Ma, four grains exhibit Miocene ment rocks (e.g., Pankhurst et al., 2006) and to establish ages (11 to 19 Ma), and the remaining grains have the primary age distributions of igneous rock. Paleozoic (358 Ma and 476 Ma) and Precambrian (751 Distributions of detrital zircon ages obtained from the Ma and 1387 Ma) ages. dredged sediments are shown in Fig. 7 and Table 1. The The deformed siltstone of D06-S20 contains four out youngest detrital zircon ages from the sediments limit the of 50 measured zircons with Late Miocene to Early age of deposition to less than ~4 Ma for D01-S17 sand- Pliocene concordant ages. A weighted mean age was cal- stone and D06-S20 siltstone. The sedimentation postdates culated using the two youngest zircon grains plus one the obduction of the Taitao ophiolite. For the area around overlapped discordant age; the weighted mean age is 4.43 the D06 dredge site, the sediments deposited after the Last ± 0.76 Ma (Fig. 6). No lower intercept age was obtained. Glacial Maximum probably originated from the Rio

Shallow melt eduction due to Chile ridge subduction 159 Fig. 8. Tectonic evolution of Chile ridge subduction system. (a) Eastward migration of the ridge system (Cande et al., 1987) and the current configuration of the slab window (Murdie et al., 1993; Breitsprecher and Thorkelson, 2009; Russo et al., 2010). H and L mark locations of Hudson and Lautaro volcanoes, respectively. Position of the Chile ridge system and slab window configura- tion at 4 Ma (b) and 8 Ma (c).

Baker, Rio Bravo and Rio Pascua drainage systems in the xenocrysts included in the Pan de Azucar and Fjord through Golfo de Penas (Fig. 1b). For the area around the San Pedro dacites. Thus, the crust underneath these vol- D01 site, they probably originated directly from the Taitao canoes consists of the Late Permian to Jurassic basement, peninsula through the Mornington Submarine Canyon a conclusion that could be readily derived from the geo- north of the Taitao Ridge (Bourgois et al., 2000; Anma et logical map in Fig. 1b. It is noteworthy, however, that al., 2009b). Although Golfo de Penas may have formed there is no evidence for the contamination of crust in the after the ridge collision (Forsythe and Nelson, 1985) and other igneous rocks in the study area, including the Taitao the high topographic relief of the southern Patagonian granites (Anma et al., 2009a). Instead, all acidic igneous Andes formed as a consequence of the ~6 Ma ridge colli- rocks in this area show moderate evidence for the incor- sion event (Ramos, 2005), the current drainage system poration of subducted sediments. that incised the Andean Arc may have been established The observed age distribution of clastic zircons also during the Holocene epoch. suggests that the drainage system that crosscut the An- Ages of all detrital zircons are younger than 1387 Ma, dean arc must have been established during sedimenta- indicating that the crust exposed in the watershed formed tion (post 4 Ma), since most clastic zircons with the age after Middle Proterozoic. Dredged volcani-clastic rocks of the metamorphic basement must have been derived and dacitic rocks of Pan de Azucar and Fjord San Pedro from the back-arc side. contain detrital zircons or xenocrysts with a similar age distribution. This feature indicates that subducted Magmatism in the Chile margin and subduction of frac- sediments were incorporated into magmas. ture zones Although, the age of the fore-arc basement is Late Zircons separated from on-land outcrops of the Chile Permian to Jurassic (Thomson and Herve, 2002), the sam- Margin Unit indicate that the fore-arc region became pled sediments do not contain detrital zircons with volcanically active at 4.8 to 4.6 Ma. Ages of the sub- Mesozoic ages. This result is partly because of the lim- merged extension of the Chile Margin Unit were obtained ited data. Anma et al. (2009a) reported detrital zircon ages from dredged samples and indicate that volcanic activi- from a sedimentary rock interbedded within volcanic ties started in the west at ~5.3 Ma. Thus, the volcanic rocks in the Taitao ophiolite (Veloso et al., 2007), but did activity in the Chile margin started with the initiation of not find Late Permian to Jurassic zircon. subduction of the Tres Montes Fracture Zone (Fig. 8b). The Late Permian to Jurassic zircons were only found Subduction began just 0.4 m.y. after the obduction of the

160 R. Anma and Y. Orihashi Fig. 9. (a) Age distribution of volcanic activities in the Chile margin. Plots (b) and (c) show the relationship between obtained U–Pb ages and orthogonal distance from the reference line in (a); the line extends from the tip of the Taitao ridge and parallels the axis of the Chile ridge, segment 1.

Taitao ophiolite (5.7 to 5.2 Ma: Anma et al., 2006) and (Fig. 9a) were used to calculate the migration rate of the after synchronous intrusions of Tres Montes (TM in Fig. igneous activity. Results yield a rate of 2.3 cm/y with 1) and Seno Hoppner-Estero Cono granite plutons (SH in magmatism in Fjord San Pedro (Fig. 9b) included and Fig. 1) (5.7 Ma and 5.2 to 5.1 Ma, respectively: Anma et 5.3 cm/y with magmatism excluded (Fig. 9c). Both rates al., 2009a) that related to a ridge subduction event at ~6 are less than the subduction rate of the Nazca plate (9 Ma (Segment-2 in Fig. 8c). cm/y) and close to the half-spread rate of the Chile ridge Anma et al. (2009a) reported a ~4.9 Ma Bahia system (3.5 cm/y). We suspect that the lavas in Fjord San Barrientos tonalitic intrusion distributed in small outcrops Pedro represent a different stage of magmatism because (perhaps as a thick dike) along the Chile Margin Unit. there is an age gap of ~1.6 m.y. between the Pan de Azucar They interpreted this intrusion to be a member of a gra- and Fjord San Pedro activities, and migration rate statis- nitic intrusion related to the emplacement of the Taitao tics are better (correlation coefficient R = 0.96) when ophiolite. Current results, however, indicate that the Bahia excluding the Fjord San Pedro data. Perhaps, lavas in Barrientos intrusion has a similar age as the volcanic rocks Fjord San Pedro were related to magmatism at a slab win- of the Chile Margin Unit and is spatially related to them. dow edge formed along the western extension of the Tres Thus, we conclude that the Bahia Barrientos tonalite is Montes Fracture Zone (Fig. 8b). related to magmatism in the Chile Margin Unit. The general EW orientation of the South Taitao Ridge The volcanic activities in the fore-arc region of the swings to a NE–SW orientation at the on-land exposures Chile margin migrated from the South Taitao Ridge to of the Chile Margin Unit. Inflection in the migration rate the on-land Chile Margin Unit (west to east) (Fig. 9). was also observed in between the magmatism of the South Acidic volcanism further extended east to the dacitic vol- Taitao Ridge and Pan de Azucar (Fig. 9c). This inflection canic plug of Pan de Azucar (~4.3 Ma) and to the lavas in could be attributed to the block rotation of the Taitao Fjord San Pedro (~2.9 Ma). The age distribution of the ophiolite during its final emplacement (Veloso et al., 2005, Chile Margin magmatism and the distance from a refer- 2009). If so, the inflection implies that the original trend ence line that parallels the spreading Chile Ridge axis of the Chile Margin Unit magmatism was EW, sub-

Shallow melt eduction due to Chile ridge subduction 161 parallel to transform faults (fracture zones) that separate tually erupt to form the Andean arc volcanic rocks of SVZ segments of the Chile Ridge system. Subduction of frac- and AVZ (Stern, 1991; Kilian and Behrmann, 2003). The ture zones must play an important role on the magmatism primary factor that accounts for a lack of arc magmatism in the Chile Margin Unit. (Patagonian Volcanic Gap) seems to be the presence of a slab window (Fig. 8a). If there is no slab, then there is no Eduction of melts at shallow depths arc magmatism at the volcanic front. Intrusion of Cabo Raper pluton took place at ~4.0 Ma in the fore-arc region of the Taitao ophiolite and became CONCLUSIONS widely distributed in the offshore area. All acidic magmatism occurred very close, in space, to the Chile We performed U–Pb dating of zircons separated from Triple Junction and coincided, in time, with the ridge sub- igneous rocks and sediments collected from the Chile duction at ~6 Ma. Thus, we conclude that melts were Triple Junction and the Taitao peninsula, southern Chile. formed at shallow depths near the triple junction due The youngest fractions of the U–Pb age population were mainly to partial melting of the subducted slab and used to estimate the age of magmatism. Our results indi- sediments. The melts were then expelled instantaneously cate that the fore-arc region became volcanically active from the interface between the subducting slab and over- ~0.4 m.y. after obduction of the Taitao ophiolite (~5.7 to riding wedge. Distributions of magmatism in the Chile 5.2 Ma) and emplacement of synchronous granites both Margin Unit indicate that an enhanced melt generation related to the ridge subduction at ~6 Ma. The fore-arc occurred along a slab window edge beside an extension volcanism migrated from offshore (~5.3 Ma) to inland of the fracture zone during its subduction. Composition (~4.6 Ma) along the NE–SW-trending Chile Margin Unit. of the melts, though it varies, is non-adakitic (Anma et The volcanism further extended to the east to produce al., 2009a; Kon et al., 2013) except for the melt compo- the dacitic volcanic plug of Pan de Azucar (~4.4 Ma). sition associated with the Pan de Azucar dacite which The magmatism migrated from west to east at a rate of includes transitional geochemical features. Thorkelson 5.3 cm/y and was related to the subduction of a fracture and Breitsprecher (2005) demonstrated that non-adakitic zone. Another intrusion of a granite pluton, widely dis- melts of granodioritic to tonalitic composition can be tributed offshore of the Taitao ophiolite, took place at ~4.0 generated along plate edges at depths of 5 to 65 km due Ma. Lavas in Fjord San Pedro erupted at another slab to partial melting of slab window margins, whereas window margin at ~2.8 Ma. Distributions of detrital zir- adakitic melts form proximal to plate edges at depths of con and zircon xenocryst ages indicate that the influence 25 to 90 km. The range in the depths of magmatism that of the lower crust is only seen in the Pan de Azucar and they considered agrees with our observations. The acidic Fjord San Pedro dacites; other acidic magmatisms show magmatism of the Chile Margin Unit was related to the moderate evidence for incorporation of subducted subduction of fracture zones. The oceanic crust along the sediments. We conclude that melts were formed at shal- fracture zones may have been re-hydrated because water low depths near the triple junction due mainly to partial penetrated to great depths during strike-slip deformation. melting of the subducted slab and sediments. These melts The re-hydrated crust may have enhanced melt genera- then ejected from the subduction interface instantane- tion at shallow depths. ously. Depleted oceanic materials became barren, and the Our results also support a shallow melt eduction sce- Patagonian Volcanic Gap was formed along the Andean nario that is predicted by numerical models (e.g., Iwamori, arc. 2000; Iwamori et al., 2007; Maruyama and Okamoto, 2007). When slab melting occurs, as exemplified by Kon Acknowledgments—We thank members of Shipboard Scien- tific Party of the MR08-06 cruise lead by Drs. Natsue Abe and et al. (2013), H2O and other volatiles are incorporated into the melts (Iwamori, 1998). Melts expelled at shal- Naomi Harada for their support. Dr. S. Machida helped us with low depths carry subducted sediments and volatiles up- dredging. Captain Akamine, his crew, and the Technical Sup- ward. The oceanic slab subducts further and becomes port Staffs of the R/V Mirai are appreciated for their dedicated work. On-land fieldwork was supported by JSPS KAKENHI barren as melts are produced, even under high tempera- Grant No. 13373004 to RA. Our thanks go to the members of ture conditions at deeper depths (~100 km: Tatsumi and the CHRISTMASSY (CHile RIdge Subduction To MAgma Eggins, 1995). This mechanism explains the presence of Supply SYstem) project for their field assistance. Measurements the Patagonian Volcanic Gap along the Andean arc in the were supported by JSPS KAKENHI Grant No. 21403012 to east of the Chile Ridge subduction zone. In contrast, for YO and a grant-in-aid from the Earthquake Research Institute zones where the subducting oceanic plate is older and cooperative research program (2011-G-07), The University of cooler, the subducted sediments or eroded crustal materi- Tokyo to RA. We thank Profs. Yildirim Dilek and Victor A. als may sink with the slab to depths of ~100 km, and then Ramos for their critical comments that helped to improve the dehydration processes enhance melting and melts even- manuscript.

162 R. Anma and Y. Orihashi J. Geophys. Res 92 REFERENCES physical and geothermal evidence. . (B1), 495–520. Anma, R., Armstrong, R., Danhara, T., Orihashi, Y. and Iwano, Defant, M. J. and Drummond, M. S. (1990) Derivation of some H. (2006) Zircon sensitive high mass-resolution ion micro- modern arc magmas by melting of young subducted probe U–Pb and fission-track ages for gabbros and sheeted lithosphere. Nature 347, 662–665. dykes of the Taitao ophiolite, Southern Chile, and their tec- DeLong, S. E., Schwarz, W. M. and Anderson, R. N. (1979) tonic implications. Island Arc 15, 130–142. Thermal effects of ridge subduction. Earth Planet. Sci. Lett. Anma, R., Armstrong, R., Orihashi, Y., Ike, S., Shin, K.-C., 44, 239–246. Kon, Y., Komiya, T., Ota, T., Kagashima, S., Shibuya, T., Drummond, M. S., Defant, M. J. and Kepezhinskas, P. K. (1996) Yamamoto, S., Veloso, E. E., Fannin, M. and Herve, F. Petrogenesis of slab-derived trondhjemite–tonalite–dacite/ (2009a) Are the Taitao granites formed due to subduction adakite magmas. Trans. R. Soc. Edinburgh, Earth Sci. 87, of the Chile ridge? Lithos 113, 246–258. 205–215. Anma, R., Orihashi, Y., Veloso, E. E., Kawamura, K., Motoki, Forsythe, R. and Nelson, E. (1985) Geological manifestation A., Jalowski, T., de Souza, R., Ortega, C., Martini, S., Aniya, of ridge collision: Evidence from the Golfo de Penas–Taitao M. and Abe, N. (2009b) Rock sampling at the Taitao area. basin, southern Chile. Tectonics 4, 477–495. MIRAI Cruise Report MR08-06 Leg 1, JAMSTEC (http:// Forsythe, R., Nelson, E., Carr, M., Kaeding, M. E., Herve, M., www.godac.jamstec.go.jp/cruisedata/mirai/e/MR08- Mpodozis, C., Soffia, J. M. and Harambour, S. (1986) 06_leg1.html). Pliocene near-trench magmatism in southern Chile: A pos- Anonymous (1972) Penrose field conference on ophiolites. sible manifestation of the ridge collision. Geology 14, 23– Geotimes 17, 24–25. 27. Aoya, M., Mizukami, T., Uehara, S. and Wallis, S. R. (2009) Gorring, M. L., Kay, S. M., Zeitler, P. K., Ramos, V. A., Rubiolo, High-P metamorphism, pattern of induced flow in the man- D., Fernandez, M. I. and Panza, J. L. (1997) Neogene tle wedge, and the link with plutonism in paired metamor- Patagonian plateau lavas: Continental magmas associated phic belts. Terra Nova 21, 67–73. with ridge collision at the Chile Triple Junction. Tectonics Behrmann, J. H., Lewis, S. D., Cande, S. C. and ODP Leg 141 16, 1–17. Scientific Party (1994) Tectonic and geology of spreading Guivel, C., Lagabrielle, Y., Bourgois, J., Maury, R. C., Fourcade, ridge subduction at Chile Triple Junction: a synthesis of S., Martin, H. and Arnaud, N. (1999) New geochemical results from Leg 141 of the Ocean Drilling Program. Geol. constraints for the origin of ridge-subduction-related plu- Rundsch. 83, 832–852. tonic and volcanic suites from the Chile triple junction Bourgois, J., Lagabrielle, Y., Maury, R., Le Moigne, J., Vidal, (Taitao peninsula and Site 862, LEG ODP141 on the Taitao P., Cantagrel, J. M. and Urbina, O. (1992) Geology of the ridge). Tectonophysics 311, 83–111. Taitao peninsula (Chile margin triple junction area, 46– Guivel, C., Lagabrielle, Y., Bourgois, J., Martin, H., Arnaud, 47°S); Miocene to Pleistocene obduction of the Bahia N., Foucade, S., Cotton, J. and Maury, R. C. (2003) Very Barrientos Ophiolite. EOS 73(43), 592. shallow melting of oceanic crust during spreading ridge Bourgois, J., Lagabrielle, Y., Le Moigne, J., Urbina, O., Janin, subduction: Origin of near-trench Quaternary volcanism at M.-C. and Beuzart, P. (1993) Preliminary results of a field the Chile Triple Junction. J. Geophys. Res. 108(B7), 2345, study of the Taitao ophiolite (southern Chile): Implications doi:10.1029/2002JB002119. for the evolution of the Chile triple junction. Ofioliti 18, Herron, E. M., Cande, S. C. and Hall, B. R. (1981) An active 113–129. spreading center collides with a subduction zone: A geo- Bourgois, J., Martin, H., Lagabrielle, Y., Le Moigne, J. and physical survey of the Chile Margin triple junction. Geol. Frutos Hara, J. (1996) Subduction erosion related to Soc. Am. Memoir 154, 683–701. spreading-ridge subduction: Taitao peninsula (Chile mar- Herve, F., Fanning, M. C., Thomson, S. N., Pankhurst, R. J., gin triple junction area). Geology 24, 723–726. Anma, R., Veloso, E. E. and Herrera, C. (2003) SHRIMP Bourgois, J., Guivel, C., Lagabrielle, Y., Calmus, T., Boulegue, U–Pb and FT Pliocene ages of near-trench granites in Taitao J. and Daux, V. (2000) Glacial-interglacial trench supply peninsula, Southern Chile. Short Papers—IV South Ameri- variation, spreading-ridge subduction, and feedback con- can Symposium on Isotope Geology, 190–193. trols on the Andean margin development at the Chile triple Herve, F., Pankhurst, R. J., Fanning, M. C., Calderon, M. and junction area (45–48°S). J. Geophys. Res. 105, 8355–8386. Yaxley, G. M. (2007) The South Patagonian batholith: 150 Breitsprecher, K. and Thorkelson, D. J. (2009) Neogene kin- my of granite magmatism on a plate margin. Lithos 97, 373– ematic history of Nazca–Antarctica–Phoenix slab windows 394. beneath Patagonia and the Antarctic Peninsula. Hirata, T. (2000) Development of a flushing spray chamber for Tectonophysics 464, 10–20. inductively coupled plasma-mass spectrometry. J. Anal. At. Cande, S. C. and Leslie, B. (1986) Late Cenozoic tectonics of Spectrom. 15, 1447–1450. southern Chile trench. J. Geophys. Res. 91(B1), 471–496. Hirata, T. and Nesbitt, R. W. (1995) U–Pb isotope Cande, S. C., Herron, E. and Hall, B. (1982) The early Cenozoic geochronology and zircon: Evaluation of the laser probe- tectonic history of the southeast Pacific. Earth Planet. Sci. inductively coupled plasma mass spectrometry technique. Lett. 57, 63–74. Geochim. Cosmochim. Acta 59, 2491–2500.

Cande, S. C., Leslie, R. B., Parra, J. C. and Hobart, M. (1987) Iwamori, H. (1998) Transportation of H2O and melting in sub- Interaction between the Chile ridge and Chile trench: Geo- duction zones. Earth Planet. Sci. Lett. 160, 65–80.

Shallow melt eduction due to Chile ridge subduction 163 Iwamori, H. (2000) Thermal effects of ridge subduction and its McCrory, P. A., Wilson, D. S. and Stanley, R. G. (2009) Con- implications for the origin of granitic batholith and paired tinuing evolution of the Pacific–Juan de Fuca–North metamorphic belts. Earth Planet. Sci. Lett. 181, 131–144. America slab window system—A trench–ridge–transform Iwamori, H., Richardson, C. and Maruyama, S. (2007) Numeri- example from the Pacific rim. Tectonophysics 464, 30–42.

cal modeling of thermal structure, circulation of H2O, and Mpodozis, C., Herve, M., Nasi, C., Soffia, J., Forsythe, R. and magmatism–metamorphism in subduction zones: Implica- Nelson, E. (1985) El magmatismo plioceno de la península tions for evolution of arcs. Gondowana Res. 11, 109–119. Tres Montes y su relación con la evolución del punto triple Kaeding, M., Forsythe, R. and Nelson, E. (1990) Geochemistry de Chile. Rev. Geol. Chile 25–26, 13–28. of the Taitao ophiolite and near-trench intrusions from the Murdie, R. E., Prior, D. J., Styles, P., Flint, S. S., Pearce, R. G. Chile margin triple junction. J. S. Am. Earth Sci. 3, 161– and Agar, S. M. (1993) Seismic response to ridge-transform 177. subduction: Chile triple junction. Geology 21, 1095–1098. Karsten, J. L., Klein, E. M. and Sherman, S. B. (1996) Subduc- Nelson, E. and Forsythe, R. (1989) Ridge collision at conver- tion zone geochemical characteristics in ocean ridge basalts gent margins: implications for Archean and post-Archean from the southern Chile Ridge: implications of modern ridge crustal growth. Tectonophysics 161, 307–315. subduction systems for the Archean. Lithos 37, 143–161. Nelson, E., Forsythe, R., Diemer, J., Allen, M. and Urbina, O. Kay, S. M., Ramos, V. A. and Márquez, M. (1993) Evidence in (1993) Taitao ophiolite: A ridge collision ophiolite in the Cerro Pampa volcanic rocks for slab-melting prior to ridge- forearc of southern Chile (46°S). Rev. Geol. Chile 20, 137– trench collision in Southern South America. J. Geol. 101, 166. 703–714. Orihashi, Y., Naranjo, J. A., Motoki, A., Sumino, H., Hirata, Kilian, R. and Behrmann, J. H. (2003) Geochemical constraints D., Anma, R. and Nagao, K. (2004) The Quaternary vol- on the sources of Southern Chile Trench sediments and their canic activities of Hudson and Lautaro volcanoes, Chilean recycling in arc magmas of the Southern Andes. J. Geol. Patagonia: new constraints from K–Ar ages. Rev. Geol. Soc., London 160, 57–70. Chile 31, 207–224. Klein, E. M. and Karsten, J. L. (1995) Ocean-ridge basalts with Orihashi, Y., Nakai, S. and Hirata, T. (2008) U–Pb age convergent-margin geochemical affinities from the Chile determinations for seven standard zircons using inductively ridge. Nature 374, 52–57. coupled plasma-mass spectrometry coupled with frequency Kon, Y., Komiya, T., Anma, R., Hirata, T., Shibuya, T., quintupled Nd-YAG (λ = 213 nm) laser ablation system: Yamamoto, S. and Maruyama, S. (2013) Petrogenesis of the comparison with LA-ICP-MS zircon analyses with a NIST ridge subduction-related granitoids from the Taitao Penin- glass reference material. Resour. Geol. 58, 101–123. sula, Chile Triple Junction Area. Geochem. J. 47, this is- Orihashi, Y., Anma, R., Motoki, A., Haller, M. J., Hirata, D., sue, 167–183. Iwano, H., Sumino, H. and Ramos, V. A. (2013) Evolution Lagabrielle, Y., Le Moigne, J., Maury, R., Cotton, J. and history of the crust underlying Cerro Pampa, Argentine Bourgois, J. (1994) Volcanic record of the subduction of an Patagonia: Constraint from LA-ICPMS U–Pb ages for ex- active spreading ridge, Taitao peninsula (southern Chile). otic zircons in the Mid-Miocene adakite. Geochem. J. 47, Geology 22, 515–518. this issue, 235–247. Lagabrielle, Y., Guivel, C., Maury, R., Bourgois, J., Fourcade, Pankhurst, R. J., Weaver, S. D., Herve, F. and Larrondo, P. S. and Martin, H. (2000) Magmatic-tectonic effects of high (1999) Mesozoic–Cenozoic evolution of the North thermal regime at the site of active ridge subduction: The Patagonian Batholith in Aysen, southern Chile. J. Geol. Soc., Chile triple junction model. Tectonophysics 326, 255–268. London 156, 673–694. Le Moigne, J., Lagabrielle, Y., Whitechurch, H., Girardeau, J., Pankhurst, R. J., Rapela, C. W., Fanning, C. M. and Marquez, Bourgois, J. and Maury, R. C. (1996) Petrology and M. (2006) Gondowanide continental collision and the ori- geochemistry of the ophiolitic and volcanic suites of the gin of Patagonia. Earth-Sci. Rev. 76, 235–257. Taitao peninsula–Chile triple junction area. J. S. Am. Earth Peacock, S. M., Rushmer, T. and Thompson, A. B. (1994) Par- Sci. 9, 43–58. tial melting of subducting oceanic crust. Earth Planet. Sci. Lopez-Escobar, L., Cembrano, J. and Moreno, H. (1995) Lett. 121, 227–244. Geochemistry and tectonics of the Chilean Southern Andes Ramos, V. A. (1989) Andean foothills structures in the north- basaltic Quaternary volcanism (37–46°S). Rev. Geol. Chile ern Magallanes basin, Argentina. A.A.P.G. Bull. 73, 887– 22, 219–234. 903. Ludwig, R. K. (2001) User Manual for Isoplot/Ex rev. 2.49. Ramos, V. A. (2005) Seismic ridge subduction and topography: Berkeley Geochronology Center Special Publication 1a, Foreland deformation in the Patagonian Andes. Berkeley Geochronology Center, Berkeley, CA. Tectonophysics 399, 73–86. Martin, H. (1993) The mechanisms of petrogenesis of the Ramos, V. A. and Kay, S. M. (1992) Southern Patagonian pla- Archean continental crust-Comparison with modern proc- teau basalts and deformation: Back-arc testimony of ridge esses. Lithos 30, 373–388. collisions. Tectonophysics 205, 261–282. Martin, H. (1996) Adakitic magmas: modern analogues of Ramos, V. A., Kay, S. M. and Singer, B. S. (2004) Las adakitas Archean granitoids. Lithos 46, 411–429. de la Cordillera Patagónica: Nuevas evidencias geoquímicas Maruyama, S. and Okamoto, K. (2007) Water transportation y geocronológicas. Rev. Assoc. Geol. Argentina 59, 693– from the subducting slab into the mantle transition zone. 706. Gondowana Res. 11, 148–165. Rapp, R. P., Watson, E. B. and Miller, C. F. (1991) Partial melt-

164 R. Anma and Y. Orihashi ing of amphibolite/eclogite and the origin of Archean Blackwell Science, 211 pp. trondhjemites and tonalites. Precambrian Res. 51, 1–25. Tebbens, S. F. and Cande, S. C. (1997) Southeast pacific tec- Russo, R. M., VanDecar, J. C., Comte, D., Mocanu, V. I., tonic evolution from early Oligocene to Present. J. Geophys. Gallego, A. and Murdie, R. E. (2010) Subduction of the Res. 102, 12061–12084. Chile Ridge: Upper mantle structure and flow. GSA Today Tebbens, S. F., Cande, S. C., Kovacs, L., Parra, J. C., LaBrecque, 20, doi:10.1130/GSATG61A.1. J. L. and Vergara, H. (1997) The Chile ridge: A tectonic Schulte, R. F., Schilling, M., Anma, R., Farquhar, J., Horan, framework. J. Geophys. Res. 102, 12035–12059. M., Komiya, T., Piccoli, P. M., Pitcher, L. and Walker, R. Thomson, S. N. (2002) Late Cenozoic geomorphic and tectonic (2009) Chemical and chronologic complexity in the con- evolution of the Patagonian Andes between latitudes 42°S vecting upper mantle: Evidence from the Taitao ophiolite, and 46°S: An appraisal based on fission-track results from southern Chile. Geochim. Cosmochim. Acta 73, 5793–5819. the transpressional intra-arc Liquine–Ofqui fault zone. Geol. Sherman, S. B., Karsten, J. L. and Klein, E. M. (1997) Soc. Am. Bull. 114, 1159–1173. Petrogenesis of axial lavas from the southern Chile ridge: Thomson, S. N. and Herve, F. (2002) New time constraints for major element constraints. J. Geophys. Res. 102(B7), the age of metamorphism at the ancestral Pacific 14963–14990. Gondowana margin of southern Chile (42°–52°S). Rev. Shibuya, T., Komiya, T., Anma, R., Ota, T., Omori, S., Kon, Y., Geol. Chile 29, 255–271. Yamamoto, S. and Maruyama, S. (2007) Progressive meta- Thomson, S. N., Herve, F. and Stockhert, B. (2001) Mesozoic– morphism of the Taitao ophiolite; Evidence for axial and Cenozoic denudation history of the Patagonian Andes off-axis hydrothermal alterations. Lithos 98, 233–260. (Southern Chile) and its correlation to different subduction Shinjoe, H., Orihashi, Y., Naranjo, J. A., Hirata, D., Hasenaka, processes. Tectonics 20, 693–711. T., Fukuoka, T., Sano, T. and Anma, R. (2013) Boron and Thorkelson, D. J. and Breitsprecher, K. (2005) Partial melting other trace element constraints on the slab-derived compo- of slab window margins: genesis of adakitic and non- nent in Quaternary volcanic rocks from the Southern Vol- adakitic magmas. Lithos 79, 25–41. canic Zone of the Andes. Geochem. J. 47, this issue, 185– Veloso, E. E., Anma, R. and Yamazaki, T. (2005) Tectonic rota- 199. tions during the Chile ridge collision and obduction of the Springer, W. and Seck, H. A. (1997) Partial fusion of basic Taitao ophiolite (Southern Chile). Island Arc 14, 599–615. granulites at 5 to 15 kbar: implications for the origin of Veloso, E. E., Anma, R., Ota, T., Komiya, T., Kagashima, S. TTG magmas. Contrib. Mineral. Petrol. 127, 30–45. and Yamazaki, T. (2007) Paleocurrent patterns of the sedi- Stacy, J. S. and Kramers, J. D. (1975) Approximation of terres- mentary sequence of the Taitao ophiolite constrained by trial lead isotope evolution by a two-stage model. Earth anisotropy of magnetic susceptibility and paleomagnetic Planet. Sci. Lett. 26, 207–221. analyses. Sediment. Geol. 201, 446–460. Stern, C. R. (1991) Role of subduction erosion in the genera- Veloso, E. E., Anma, R. and Yamaji, A. (2009) Ophiolite em- tion of Andean magmas. Geology 19, 78–81. placement and the effects of the subduction of the Chile Stern, C. R. (2004) Active Andean volcanism: its geologic and Ridge System: Heterogeneous paleostress regimes recorded tectonic setting. Rev. Geol. Chile 31, 161–206. in the Taitao ophiolite (Southern Chile). Andean Geol. 36, Stern, C. R. and Kilian, R. (1996) Role of the subducted slab, 3–16. mantle wedge and continental crust in the generation of Watson, E. B., Cherniak, D. J., Hancher, J. M., Harrison, T. M. adakites from the Andean Austral Volcanic Zone. Contrib. and Wark, D. A. (1997) The incorporation of Pb into zir- Mineral. Petrol. 123, 263–281. con. Chem. Geol. 141, 19–31. Stern, C. R., Skewes, M. A. and Durán, M. (1976) Volcanismo Wiedenbeck, M., Alle, P., Corfu, F., Griffin, W. L., Meier, M., Orogénico en Chile Austral. Congreso Geológico de Chile, Oberli, F., Von Quadt, A., Roddick, J. C. and Spiegel, W. Actas, No. 1, Vol. 2, 195–212, . (1995) Three natural zircon standards for U–Th–Pb, Lu– Stern, C. R., Frey, F. A., Futa, K., Zartman, R. E., Peng, Z. and Hf, trace element and REE analyses. Geostand. Newslett. Kyser, T. K. (1990) Trace element and Sr, Nd, Pb and O 19, 1–23. isotopic composition of Pliocene to Quaternary alkali basalts of the Patagonian Plateau Lavas of southernmost South America. Contrib. Mineral. Petrol. 104, 294–308. SUPPLEMENTARY MATERIALS Suarez, M. and De la Cruz, R. (2001) Jurassic to Miocene K– URL (http://www.terrapub.co.jp/journals/GJ/archives/ Ar dates from eastern central Patagonian Cordillera plutons, data/47/MS243.pdf) Chile (45°–48°S). Geol. Mag. 138, 53–66. Table S1 Tatsumi, Y. and Eggins, S. (1995) Subduction Zone Magmatism.

Shallow melt eduction due to Chile ridge subduction 165