Arc Magmatic Evolution and the Construction of Continental Crust at the Central American Volcanic Arc System

Home , Ocean island basalt, Slab (geology), Volcanic arc

International Geology Review

ISSN: 0020-6814 (Print) 1938-2839 (Online) Journal homepage: http://www.tandfonline.com/loi/tigr20

Arc magmatic evolution and the construction of continental crust at the Central American Volcanic Arc system

Scott A. Whattam & Robert J. Stern

To cite this article: Scott A. Whattam & Robert J. Stern (2015): Arc magmatic evolution and the construction of continental crust at the Central American Volcanic Arc system, International Geology Review, DOI: 10.1080/00206814.2015.1103668

To link to this article: http://dx.doi.org/10.1080/00206814.2015.1103668

View supplementary material

Published online: 14 Dec 2015.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tigr20

Download by: [The University of Texas at Dallas] Date: 15 December 2015, At: 18:30 INTERNATIONAL GEOLOGY REVIEW, 2015 http://dx.doi.org/10.1080/00206814.2015.1103668

Arc magmatic evolution and the construction of continental crust at the Central American Volcanic Arc system Scott A. Whattama and Robert J. Sternb aDepartment of Earth and Environmental Sciences, Korea University, Seoul, Republic of Korea; bGeosciences Department, University of Texas at Dallas, Richardson, TX 75083-0688, USA

ABSTRACT ARTICLE HISTORY Whether or not magmatic arcs evolve compositionally with time and the processes responsible Received 30 September 2015 remain controversial. Resolution of this question requires the reconstruction of arc geochemical Accepted 1 October 2015 evolution at the level of a discrete arc system. Here, we address this problem using the well- KEYWORDS studied Central American Volcanic Arc System (CAVAS) as an example. Geochemical and isotopic Volcanic arc; subduction; data were compiled for 1031 samples of lavas and intrusive rocks from the ~1100 km-long continental crust; tectonics; segment of oceanic CAVAS (Panama, Costa Rica, Nicaragua) built on thickened oceanic crust Central America; Caribbean; over its 75 million year lifespan. We used available age constraints to subdivide this data set into Galapagos Plume six magmatic phases: 75–39 Ma (Phase I or PI); 35–16 Ma (PII); 16–6 Ma (PIII); 6–3 Ma (PIV); 5.9–0.01 Ma (PVa arc alkaline and PVb adakitic); and 2.6–0 Ma (PVI, Quaternary to modern magmatism, predominantly ≪ 1 Ma). To correct for magmatic fractionation, selected major and trace element abundances were linearly regressed to 55 wt.% SiO2. The most striking observation is the overall evolution of the CAVAS to more incompatible element enriched and ultimately continental-like compositions with time, although magmatic evolution took on a more regional character in the youngest rocks, with magmatic rocks of Nicaragua becoming increasingly distinguishable from those of Costa Rica and Panama with time. Models entailing progressive arc magmatic enrichment are generally supported by the CAVAS record. Progressive enrichment of the oceanic CAVAS with time reflects changes in mantle wedge composition and decreased melting due to arc crust thickening, which was kick-started by the involvement of enriched plume mantle in the formation of the CAVAS. Progressive crustal thickening and associated changes in the sub-arc thermal regime resulted in decreasing degrees of partial melting over time, which allowed for progressive enrichment of the CAVAS and ultimately the production of continental- like crust in Panama and Costa Rica by ~16–10 Ma.

1. Introduction and their definitions). These characteristics are largely due to the fluid-mediated nature of convergent margin Subduction zone magmatism results primarily from the magmatism and to the fact that, in contrast to igneous dehydration of subducted oceanic crust and sediment activity at mid-ocean ridges and hotspots, arc magmatic melting and the subsequent transfer of these liquids to activity stays in the same place relative to the under- the overlying mantle wedge where partial melting lying crust for tens of millions of years. occurs (White and Patchett 1984; McCulloch and Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 Study of arc igneous rocks must also consider the role Gamble 1991; Plank and Langmuir 1993; Hawkesworth of the underlying crust, because this crust can be et al. 1993a, 1993b; Pearce and Peate 1995; Ishikawa involved in magmagenesis, obscuring the geochemical and Tera 1997; Kimura et al. 2014). The diagnostic che- and isotopic signature of mantle-derived magmas. Thick mical signatures of subduction zone magmatism granitic continental crust favours the establishment of include: (1) abundant felsic rocks; (2) a tendency to MASH (melting, assimilation, storage, and homogeniza- minimize Fe-enrichment during magmatic fractionation; tion, Hildreth and Moorbath 1988) zones, with massive (3) elevated abundances of large ion lithophile elements involvement of especially the lower crust in the resultant (LILEs) relative to the light rare earth elements (LREEs); magmas. Intra-oceanic arc (IOA) systems (see review of and (4) depletion of high field strength elements Stern 2010 and references therein) – where the crust is (HFSEs) (e.g. Arculus 1994) (see Table 1 for a list of the thinner, more mafic, and more refractory – are sites most common abbreviations and acronyms used here

Table 1. Abbreviations and definitions of commonly used terms magmatic arcs argued for evolution from early low-K (Whattam and Stern 2015). tholeiitic magmas to later incompatible element- Abbreviation/ enriched, high-K calc-alkaline and shoshonitic magma- acronym Definition tism. Jakeš and White (1972) suggested that the most BAB Back arc basin BCC Bulk continental crust important chemotemporal (chemical changes with time) CAVAS Central American Volcanic Arc system trends exhibited by magmatic arcs include: a switch from CLIP Caribbean Large Igneous Province (an OP) GAA Greater Antilles Arc the eruption of early tholeiites followed by later calc- HFSE High-field strength element (e.g., Nb, Zr, Ti) alkaline and finally shoshonitic magmas; progressive HREE Heavy REE IBM Izu–Bonin–Mariana (a convergent margin in the enrichments in K and other fluid-mobile LILE elements western Pacific) such as Rb, Ba, and Sr and other large cations (Th, U, Pb) IOA Intra-oceanic arc (or magmatic arc) IODP International Oceanic Drilling Program (now and LREE; increases in K2O/Na2O ratios; and decreases in International Ocean Discovery Program) iron enrichment and K/Rb ratios. Arculus and Johnson LILE Large ion lithophile element (e.g., Rb, Ba) (1978) challenged this interpretation by pointing out LREE Light REE MASH Melting, assimilation, storage, and homogenization several exceptions including a decrease in incompatible MORB Mid-ocean ridge basalt (pure asthenospheric melt) elements with time for the Cascades and Lesser Antilles. OIB Ocean island basalt (tholeiitic and alkalic basalts of within-plate oceanic volcanoes) In a similar vein, recent studies of stratigraphically con- OPB Oceanic plateau basalt (plume basalt) strained tephra in IODP cores indicate that the composi- PI, PII. . .PVI (Temporal) Phase I, Phase II. . .Phase VI REE Rare earth element tion of Izu-Bonin-Mariana arc magmas has changed very THI Tholeiitic index: tholeiitic suites have THI > 1; calc- little over the past ~40 Ma (Lee et al. 1995; Bryant et al. alkaline suites have THI < 1 et al VAB Volcanic arc basalt (subduction-modified basalt) 2003;Straub2003;Straub . 2015). Resolving the controversy as to why some convergent margin magmatic systems evolve with time whereas others do not is important for understanding convergent where contributions from the crust of the overriding margin processes and how continents form. The first step plate are minimized. IOAs are thus preferred for inferring is to reconstruct the magmatic history of the arc; the subduction-related magmatic processes. IOAs represent second step is to understand what this tells us about the the most important sites of juvenile, mantle-derived, processes controlling magma evolution, which could continental crust formation and arc–continent collision reflect variations in slab contributions, mantle contribu- and the subsequent accretion of arc-related terranes is tions, crustal contributions, local tectonics, or all four. Our believed to be key for the growth of continental crust chemotemporal study of the Central American Volcanic (Taylor and McLennan 1985;Rudnick1995;Rudnickand Arc system (CAVAS) is restricted to the ~75–0Maarc Fountain 1995). Approximately 85–95% of the mass of segment constructed upon oceanic crust in Nicaragua, continental crust is estimated to have formed at mag- Costa Rica, and Panama; we do not consider the part of matic arcs above subduction zones (Rudnick 1995;Barth the arc in El Salvador and Guatemala, which may be built et al. 2000). on the continental crust of the Chortis Block. Our studied It has been recognized since the earliest discussions of time interval is identical to that of the study of Gazel et al. Plate Tectonics that convergent margin magmatism (2015), which also documents the physical and chemical shows strong spatial controls, which are a function of evolution of the arc in Panama and Costa Rica. The study slab depth, i.e. from the production of depleted tholeiites of Gazel et al.(2015) differs spatially from ours as theirs Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 above shallow subduction zones to the generation of does not encompass Nicaragua or the Late Cretaceous enriched alkali basalts over deep subduction zones Golfito Complex of southernmost Costa Rica. Moreover, (Kuno 1966; Dickinson and Hatherton 1967;Sugimura the study of Gazel et al.(2015)doesnotincludekey 1968;Gill1970;Ringwood1974). More recent studies geochemical data from the studies of Lissinna (2005)and demonstrated fundamental relations between subduc- Buchs et al.(2010), which provide important constraints tion zone chemical systematics and variations in the on the geochemical composition of earliest arc magmas mantle wedge melting regime (Plank and Langmuir erupted in Panama and southernmost Costa Rica. 1988) and chemical variability as a function of slab or Nevertheless, our results mostly support their conclusions. wedge processes (Turner and Langmuir 2015). It is less Although a number of studies have documented certain whether or not arc magmas evolve composition- the chemical variability in CAVAS magmas (e.g. ally with time. Some early (Jakeš and White 1969, 1972; Patino et al. 2000;Planket al. 2002;Hoernleet al. Jakeš and Gill 1970) and more recent (Jolly et al. 1998a, 2008; Heydolph et al. 2012), all of these deal with 1998b, 2001; duBray and John 2011; Zernack et al. 2012; magmatic activity, which spans a maximum of ~30 mil- Gazel et al. 2015) studies of the chemical evolution of lion years duration (30 Ma to the present); some of INTERNATIONAL GEOLOGY REVIEW 3

et al o o o o o these (e.g. Heydolph . 2012) encompass only 100 W 90 W 80 W 70 W 60 W NORTH Nicaragua and segments of the CAVAS to the west of AMERICAN N Hispaniola PLATE Cuba Virgin Nicaragua constructed upon continental crust. Other Yucatan 20o N Islands Jamaica

studies have dealt with longer-duration studies of L GAA PR A Fig. 1b CARIBBEAN system A et al CAVAS evolution but only of Panama (Wegner . s PLATE y bbean LIP s NIC ri t

a 2011). To date, no studies have explicitly considered C m Fig. 2 Curacao 400 km CR o the chemotemporal evolution of the entire segment of PAN 10 N the CAVAS constructed on oceanic crust in Panama, Venezuela Costa Rica, and Nicaragua. Here, we present the first COCOS PLATE synergistic chemotemporal treatment of the CAVAS Colombia SOUTH

from establishment by ~75 Ma to the present. We o Galapagos NAZCA AMERICAN 0N first demonstrate that many of the chemotemporal a Islands PLATE PLATE trends outlined by Jakeš and White (1972)holdtrue for the oceanic CAVAS system between 75 and 16 Ma 87o W (sometimes between 75 and 6 Ma), and show how this GUAT Nicaragua Rise is reflected in trace element and isotope systematics. Second, we explain CAVAS chemotemporal evolution MFZ HONDURAS Central Chortis Terrane in terms of varying degrees and modes of melting; the e rran Te ‘ ’ o sou is nature and relative depletedness of the source; rela- 14 N thern rt C o ho h ES rt C tive contributions of fluids and sediments; subducted i s nr ~ COBB T te Siuna CARIBBEAN e s seamount and mantle plume contributions; and the r a ra E Terrane PLATE 100 km n role of major tectonic, tectonomagmatic, and oceano- e graphic events over the course of CAVAS evolution. oceanic NICARAGUA basement s t Our goal is to understand what the CAVAS teaches (of the SCT & es en H m Siuna Terrane) rp ca about arc magmatic evolution. This compilation thus b es serves two purposes: it represents a report on CAVAS arc evolution and it illustrates the challenges facing Figure 1. (a) Present-day tectonic configuration of the Circum- any effort to capture the long-term magmatic evolu- Caribbean region (modified from Meschede and Frisch 1998). tion of convergent plate margin igneous activity. Abbreviations: CR, Costa Rica; LAA, Lesser Antilles Arc; GAA, Greater Antilles Arc; NIC, Nicaragua; PAN, Panama. (b) Sketch of Middle America showing the distribution of Chortis Block 2. Synopsis of CAVAS magmatic history: terranes and the Siuna Terrane of Nicaragua, El Salvador, and et al distribution in space and time Honduras (modified from Rogers . 2007b). Abbreviations: COBB, Continental–oceanic basement boundary; ES, El Salvador; The modern CAVAS volcanic front stretches ~1100 km GUAT, Guatemala; MFZ, Motagua Fault Zone. along the western margin of the Caribbean plate from Costa Rica through Nicaragua, El Salvador, and Guatemala to the Guatemala–Mexico border at the originally interpreted as a CLIP segment (Hauff et al. southern margin of the North American plate (Figure 1). 2000b), but more recently as a 75–66 Ma arc segment CAVAS also extended into Panama, but this part of the (Buchs et al. 2010), an interpretation that we follow Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 arc shut down within the last few millions of years. here on the basis of geochemical considerations Unequivocal CAVAS magmatic activity began ~75 Ma (Section 4, see also Whattam and Stern 2015). (Buchs et al. 2010) (however see also Whattam and Oligocene and early Miocene igneous rocks are best Stern 2015) when the Farallon Plate (now the Cocos exposed in Costa Rica, western Nicaragua, and iso- Plate) began to subduct beneath thickened Caribbean lated regions in Panama (Figure 2). Middle and late Large Igneous Province (CLIP) oceanic plateau (e.g. Hauff Miocene arc sequences are broadly exposed west of et al. 2000b and references therein; Whattam and Stern the Canal Zone in central Panama, whereas late 2015). Today, the CAVAS reflects the eastward subduc- Miocene–Pliocene CAVAS sequences are best exposed tion of the Cocos plate at ~70–85 mm/year beneath the in Costa Rica (Figure 2). Quaternary bimodal lavas and western edge of the Caribbean plate (Carr et al. 2003). adakitic intrusions are concentrated in SE Costa Rica The oldest CAVAS sequences (Phase I, PI) are best and western Panama and behind the volcanic front in exposed in the 73–39 Ma Sona–Azuero Arc of western NW Costa Rica (Figure 2). A detailed treatment of the Panama and the 70–39 Ma Chagres–Bayano Arc of chemotemporal evolution of the CAVAS over its eastern Panama (Figure 2). The Golfito Complex was 75 Ma history is provided in the Supplementary data 4 S. A. WHATTAM AND R. J. STERN

Distribution of CAVAS Other relevant subduction-related NICARAGUA samples used (this study) features & materials Phase, interval (Ma) TM F CY Chortis Block PI (75-39, n = 139) arc basement of apparent Arc PII (35-16, n = 67) 89-85 Ma CLIP plateau affinity1 m PII ALI accreted Cretaceous-? oceanic ~SCT (SW)/Siuna 1 12° N PIII (16-6, n = 122) seamounts and plateaus Terrane (NE) boundary PIV (6-3 , n = 19) Neogene arc (i.e., undifferentiated MN PV (5.9-0.02) magmatism spanning PII-PVI)1, 2 PVa arc alks (n = 15) CR-PAN mafic adakites3 PVb adakites (n = 86) CR-PAN alkali basalts3 SP Arc PVI (2.6-0, n = 583) rear arc alkali basalts3 Cor TD NC, HD d CARIBBEAN PLATE & TG de COSTA Caribbean Large Igneous Province Tal RICA 10° N AG (CLIP) N Arc Canal Basin SJ Chag BDT res LC Fm TR PANAMA -B a y DM a EB PQ n Golfito EV CB o GF LY PC MJ Arc Co Arc rd de Pan 08° N PI Sona-Azuero Arc CI Sona Azuero BP

COCOS RIDGE SEAMOUNT PROV 86° W 84° W 82° W 80° W 250 km

Figure 2. Detail of the study area (boxed region in Figure 1) showing the distribution of PI to PVI (75–0 Ma) magmatism in Panama, Costa Rica, and Nicaragua as bracketed by this study. Note that the present-day CAVAS volcanic front trends to the northwest from northwest Nicaragua through Honduras, El Salvador, and Guatemala to the southwest margin of the North American Plate as shown in Figure 1. An approximate boundary between the Southern Chortis Terrane (SCT) to the southwest and the Siuna Terrane to the northeast is shown for southern Nicaragua (see text and Rogers et al. 2007a, 2007b). Distribution of PI–PIV magmatism in Panama is based on the studies of Lissinna (2005); Buchs et al.(2010); Wegner et al.(2011); Rooney et al.(2010); Farris et al.(2011); Montes et al.(2012a, 2012b); and Whattam et al.(2012 and references therein). Distribution of PI–PIV magmatism in Costa Rica is based on the studies of MacMillan et al.(2004); Gazel et al.(2005, 2009); and Buchs et al.(2010). Distribution of PII–PIV magmatism in Nicaragua is based on the studies of Ehrenborg (1996); Elming et al.(2001); Plank et al.(2002); and Saginor et al.(2011). Distribution of PV magmatism in Panama and Costa Rica is based on the studies of Defant et al.(1991a, 1991b, 1992); Drummond et al.(1995); MacMillan et al.(2004); and Lissinna (2005); see also Gazel et al.(2009 and references therein). Distribution and extent of the Neogene arc (i.e. light green-blue shade that encompasses our PII–IV magmatism) are from Elming et al.(2001) and Buchs et al. (2010). Phases I–V magmatic products shown as circles as opposed to larger shaded regions indicate either a smaller extent of magmatism or situations in which in the extent of magmatic products is uncertain. In the case of Phase IV magmatism, each red circle represents a discrete Quaternary (2.6–0 Ma) volcano (locations and distribution taken from Mann et al. 2007). Distribution of the Seamount Province to the immediate west of the Cocos Ridge is from Gazel et al.(2009) and location and distribution of adakite localities are as shown in Wegner et al.(2011). Abbreviations of localities, discrete volcanoes, and oceanic features: BDT, Bocas del Toro; BH, Bahia Pina; Cord de Pan, Cordillera de Panama; Cord de Tal, Cordillera de Talamanca; CI, Coiba Island; EB, El Baru (volcano); EV, El Valle (volcano); LY, La Yeguada (volcano); MJ, Maje; MN, Managua; PC, Panama City; PI, Pearl Islands; PQ, Petaquilla; PROV, Province; SJ, San Jose. Abbreviations of arcs and arc-related units and formations: AG, Aguacate; C-B, Chagres-Bayano; CY, Coyol; DM, Dominical (unit); GF, Golfito; LC Fm, La Cruz Formation; PR, Paso Real; S-A, Sona-Azuero; SP, Sarapiqui; TD, Trinidad; TM Fm,

Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 Tamarindo Formation; TR, Talamanca Range. Abbreviations of units interpreted as CLIP oceanic plateau: HD, Herradura; NC, Nicoya Complex; TG, Tortugal. Abbreviations (legend): ALI, adakitic-like intrusives (Whattam et al. 2012); CLIP, Caribbean Large Igneous Province (oceanic plateau). Superscripts: 1, Buchs et al.(2010); 2, Elming et al.(2001); 3, Hoernle et al.(2008).

(see http://dx.doi.org/10.1080/00206814.2015. Central America is a collage of several terranes, con- 1103668). tinental in the north (Chortis Block) and oceanic (CLIP) in the south (Figure 2) (e.g. Mann et al. 2007). To avoid biasing our reconstruction of magmatic evolution because of the interaction of arc magmas with pre- 3. Geochemical, geochronological, and isotope existing continental crust, we limited our investigation data set compilation, manipulation, and to the ~1100 km-long portion of the CAVAS constructed sources on oceanic crust in Nicaragua, Costa Rica, and Panama. Detailed methods are provided in the Supplementary We note here that the exact nature (continental vs. data. Below we summarize our methodology for data oceanic) of segments of the basement of Nicaragua compilation, manipulation, and filtering. and the contact between continental and oceanic INTERNATIONAL GEOLOGY REVIEW 5

basement units are controversial and we use the inter- ) pretations of Venable (1994) and Rogers et al.(2007a, 2007b) for westernmost Nicaragua. These workers con- Continued

clude that both the Siuna Terrane of western Nicaragua ( 0

and the westernmost Chortis Block (southern Chortis – 583 2.6 1.3 ± 1.3 Terrane, SCT, Figure 1b) comprise Early Cretaceous vol- Phase VI

canic arc fragments constructed on oceanic basement regressed to 55 wt. % that accreted to the eastern Chortis Terrane and the Central Chortis Block, respectively, in the Early Cretaceous (e.g. see Table 2 of Rogers et al. 2007b; see also Baumgartner et al. 2008). CAVAS samples from 0.15 –

Nicaragua used in our study (Ehrenborg 1996; Elming 86 PhaseVb 2.2 ± 2.0 et al. 2001; Plank et al. 2002; Gazel et al. 2011; Saginor 4.20 et al. 2011) are limited to those from northwestern, southwestern, and eastern Nicaragua, which comprise the Siuna Terrane and the SCT (Figure 2) and hence are the ones interpreted as having been constructed upon an oceanic basement. We do not consider CAVAS sam- 0.01 – ples from the north of the Siuna Terrane in Nicaragua, 15 1.94 0.07 1.98 0.06 1.11 0.03 0.48 0.02 0.62 0.02 0.37 0.01 3 ± 3 48.55 1.60 39.94 1.42 23.91 0.75 694.10 53.10 1400.52 35.03 581.14 6.94 PhaseVa or the ones from El Salvador and Guatemala, which may 5.90 be underlain by Palaeozoic and older continental crust of the Chortis Block. Conversely, there is little dispute as to the nature of the basement beneath Panama and Costa Rica, which is universally considered as a CLIP oceanic plateau (e.g. Hauff et al. 2000a, 2000b). 3 – We compiled all relevant geochemical, geochrono- 19 0.62 0.05 2.28 0.15 1.24 0.03 1.28 0.02 1.82 0.02 0.28 0.03 0.69 0.04 0.32 0.81 0.20 0.00 3.07 0.07 4.05 0.04 3.18 0.04 2.99 0.03 0.787.95 0.02 0.130.77 1.56 7.78 0.09 0.12 0.10 0.86 NA 6.92 0.01 NA 0.03 0.81 0.81 8.29 0.01 0.10 0.08 0.83 0.18 3.48 0.22 6.29 0.11 5.41 0.08 4.55 0.04 logical, and isotopic data available in the literature for CAVAS samples from Panama, Costa Rica, and Nicaragua. Collective (i.e. Panama plus Costa Rica,

Panama plus Costa Rica plus Nicaragua, or Costa Rica 66 – 122 0.28 0.45 2.06 3.20 0.92 8.28 0.81 4.02 16 plus Nicaragua) linearly regressed data (see Section 11 ± 5 4.5 ± 1.5 Phase III Phase IV 3.3. below) for each temporal phase is provided in Table 2; Supplementary Table S1 provides an expanded version of Table 2 at the level of specific 0.01 0.02 0.07 0.06 0.03 0.15 0.20 region. Raw geochemical data and the sources for 0.13 these data are provided in Supplementary Table S2. Sources for the isotope data sets are provided in the caption for relevant figures in Section 5 and also in 16 – 67 0.21 0.33 2.00 3.05 0.92 8.47 0.82 4.24

Supplementary Table S3. We amassed 1031 pertinent 35 Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 Phase II samples with a full set of major element chemistry 25.5 ± 9.5 analyses but with varying completeness of trace element and isotopic data. Both volcanic (n = 922) and plutonic (n = 109) samples are included. (In cases where it was not specified whether the sample was volcanic or plutonic, we assume these to be volcanic

in our calculation of relative abundance of each. 39 – 139 0.60 0.05 1.01 0.07 1.45 1.90 0.16 0.14 0.01 0.19 0.02 7.81 0.76 22.16 2.31 30.53 38.24 4.34 1.61 0.05 3.21 0.09 0.848.37 0.03 0.15 0.92 0.28 5.21 0.13 75 258.44 11.63 447.90 20.99 612.86 726.16 91.04 Phase I Samples described as basaltic dikes are considered as 57 ± 18 (PAN+CR) SE (PAN+CR+NIC) SE (PAN+CR+NIC) (CR+NIC) SE (arc alks) SE (adak) SE (CR+NIC) SE plutonic). As it is difficult to determine where the (ppm) 1

magmatic front was over the 75 million year lifespan (wt.%) 55

of the CAVAS, we have not distinguished between 55 Selected major and trace element data of CAVAS lavas and intrusives of PI, PII, PIII, PIV, PV and PVI and of these phases discriminated by region linearly O

magmas emplaced at the magmatic front from those 2 /MgO t t 5 . O 2 2 emplaced behind the magmatic front except for 2 O O/Na O (no. samples) 2 2 2 P FeO Trace elements Rb TiO Age range (Ma) Mean age (Ma) Tholeiitic Index K FeO MgO K Sr YTi 22.40 5052.40 197.54 0.87 5508.99 23.55 200.60 1.20 5515.25 23.80 4682.45 117.12 21.53 9331.73 1.46 712.58 5170.84 18.13 60.19 0.54 4845.56 13.42 462.05 0.26 20.43 0.28 n Major oxides Na Table 2. Quaternary lavas. We do not think this significantly SiO Downloaded by [The University of Texas at Dallas] at 18:30 15 December 2015 6

Table 2. (Continued). Phase I Phase II Phase III Phase IV PhaseVa PhaseVb Phase VI .A HTA N .J STERN J. R. AND WHATTAM A. S. (PAN+CR) SE (PAN+CR+NIC) SE (PAN+CR+NIC) (CR+NIC) SE (arc alks) SE (adak) SE (CR+NIC) SE Zr 68.44 3.95 86.16 3.69 100.33 94.18 12.66 129.88 11.63 137.38 3.45 96.24 2.67 Nb 2.45 0.20 4.74 0.40 5.65 4.43 1.12 36.95 2.56 7.79 0.56 8.31 0.43 K 5021.37 388.52 8402.51 606.44 12066.77 15732.12 1331.53 16063.74 543.75 16474.73 521.82 9220.40 219.38 Ba 241.11 18.69 468.84 30.44 773.85 876.82 65.93 916.46 50.31 981.16 24.48 638.45 9.45 La 6.34 0.50 10.29 0.62 14.94 19.24 1.51 28.56 2.96 32.06 1.36 18.13 0.77 Ce 14.70 1.09 22.23 1.07 29.84 35.70 9.04 50.55 6.22 60.86 2.51 37.79 1.51 Pr 2.12 0.14 3.22 0.14 4.31 4.46 1.00 6.67 0.78 7.40 0.31 4.76 0.17 Nd 10.00 0.60 14.67 0.62 18.46 19.23 3.85 26.29 3.10 28.77 1.06 19.52 0.59 Sm 2.81 0.14 3.18 0.16 4.27 4.09 0.57 5.35 0.55 4.86 0.15 4.24 0.10 Eu 0.93 0.04 1.69 0.20 1.31 1.22 0.15 1.83 0.15 1.35 0.04 1.25 0.02 Gd 3.33 0.14 4.05 0.18 4.43 3.90 0.40 5.40 0.52 3.94 0.09 3.94 0.07 Tb 0.57 0.02 0.68 0.03 0.70 0.58 0.04 0.74 0.06 0.49 0.01 0.60 0.01 Dy 3.77 0.16 4.15 0.21 4.13 3.51 0.21 3.78 0.27 2.60 0.04 3.48 0.05 Ho 0.81 0.03 0.86 0.05 0.84 0.70 0.03 0.67 0.04 0.48 0.01 0.70 0.01 Er 2.36 0.09 2.51 0.13 2.41 2.02 0.09 1.67 0.10 1.27 0.02 1.98 0.03 Tm 0.35 0.01 0.36 0.04 0.32 0.42 0.00 0.22 0.01 0.18 0.00 0.29 0.00 Yb 2.34 0.09 2.46 0.13 2.32 2.01 0.11 1.29 0.07 1.15 0.02 1.93 0.03 Lu 0.36 0.01 0.38 0.02 0.36 0.31 0.02 0.19 0.01 0.17 0.00 0.30 0.00 ∑ REE2 50.81 1.38 70.73 1.45 88.61 97.39 10.02 133.22 7.63 145.58 3.06 98.92 1.81 Pb 1.44 0.09 2.88 0.20 4.08 2.88 0.25 4.67 0.21 6.73 0.26 3.59 0.09 Th 0.86 0.07 1.44 0.15 2.18 4.51 1.37 4.99 0.42 6.53 0.38 3.00 0.20 U 0.29 0.02 0.61 0.06 0.79 1.58 0.42 2.10 0.12 1.96 0.09 1.08 0.06 V 269.67 10.49 245.21 8.10 234.85 217.42 13.43 NA NA 239.66 3.65 192.50 2.35 K/Rb 642.76 79.61 379.22 48.12 395.30 411.43 58.24 330.88 15.63 412.53 19.61 385.58 15.16 Ba/La 38.04 4.22 45.56 4.02 51.81 45.57 4.95 32.08 3.76 30.61 1.51 35.22 1.59 Rb/Zr 0.11 0.01 0.26 0.03 0.30 0.41 0.07 0.37 0.04 0.29 0.01 0.25 0.01 Ba/Zr 3.52 0.34 5.44 0.42 7.71 9.31 1.43 7.06 0.74 7.14 0.25 6.63 1.37 U/Th 0.34 0.04 0.42 0.07 0.36 0.35 0.14 0.42 0.04 0.30 0.02 0.36 0.03 Pb/Ce 0.10 0.01 0.13 0.01 0.14 0.08 0.02 0.09 0.01 0.11 0.01 0.10 0.00 Sr/Nd 25.83 1.93 30.54 1.93 33.21 37.77 8.93 26.40 3.71 48.69 2.17 29.77 0.97 Ba/Th 281.71 32.25 325.74 39.72 354.79 194.32 60.59 183.79 18.47 150.23 9.50 212.83 14.58 Ba/Nb 98.58 10.97 98.93 10.50 137.02 197.77 52.21 24.81 2.19 125.94 9.63 76.82 4.12 Th/Nb 0.35 0.04 0.30 0.04 0.39 1.02 0.40 0.13 0.01 0.84 0.08 0.36 0.03 Zr/Nb 27.98 2.76 18.18 1.71 17.76 21.24 6.09 3.52 0.40 17.63 1.35 11.58 0.68 Ba/Ti 0.05 0.00 0.09 0.01 0.14 0.19 0.01 0.10 0.01 0.19 0.01 0.13 0.00 Zr/Y 3.05 0.21 3.66 0.24 4.22 4.37 0.66 7.16 0.68 10.23 0.33 4.71 0.15 Nb/Yb 1.05 0.09 1.93 0.19 2.44 2.21 0.57 28.56 2.44 6.79 0.50 4.31 0.23 La/Nb 2.59 0.29 2.17 0.22 2.64 4.34 1.15 0.77 0.10 4.12 0.34 2.18 0.15 La/Sm 2.26 0.21 3.24 0.25 3.50 4.70 0.75 5.34 0.77 6.60 0.34 4.27 0.21 La/Yb 2.71 0.24 4.18 0.33 6.45 9.59 0.93 22.08 2.54 27.95 1.26 9.40 0.42 Ti/V 18.74 1.03 23.97 1.15 23.48 21.54 1.43 NA NA 21.58 0.41 25.17 0.45 Bold text of some linearly regressed values indicate increasing values of the listed incompatible major and trace elements and ratios that increase from the previous phase; PI data are also in bold text for visual aid. See Supplementary data for methods of linear regression.Abbreviations: NA, not analyzed or not applicable; NC (for THI), not calculable where total number of FeO4.0 and/or FeO8.0 was < 1. Superscripts: 1, 2 for THI and ΣREE are to indicate that errors for these are STD (for all other element and element ratios, uncertainties are standard error, SE, of the mean). Notes: (1) This table represents a condensed version of Supplementary Table 1, which provides the same regressed data in addition to region specific regressed data (i.e. regressed data provided at the level of region – Panama, Costa Rica, etc.). (2) The data in this table are represented as large, light grey ‘x’s in our various geochemical plots. The coloured symbols represent distinct regions and the data for these are provided in t Supplementary Table 1 (see Note 1 above). (3) FeO4.0 and FeO8.0 are based on non-normalized abundances of FeO and MgO. (4) References for datasets are provided in Section 3.2. (5) Raw data from which values were calculated are provided Supplementary Table S2. INTERNATIONAL GEOLOGY REVIEW 7

biases the data as there is no evidence to suggest Region Phase c ‘behind arc’ activity prior to Quaternary time. 12 Va Vb VI PAN CR 10 NIC 3.1. Temporal subdivisions 8 trachy- basalt Based on radiometric and biostratigraphic ages 6 reported in the literature, we temporally subdivide – 4 CAVAS magmatism into six phases at 75 39 Ma (Phase dacite – andesite I, PI, Panama, Costa Rica), 35 16 Ma (PII, Panama, Costa 2 bas andesite 5.9-0 Ma Rica, Nicaragua), 16–6 Ma (PIII, Panama, Costa Rica, basalt Nicaragua), 6–3 Ma (PIV, Costa Rica, Nicaragua), 5.9– 14 Region Phase b 0.02 Ma (PVa, arc alkaline in Costa Rica and western 12 III IV PAN Panama, and PVb adakites, in Panama and Costa Rica), CR NIC and 2.6–0 Ma (PVI, the Quaternary and modern volcanic 10 TR-TIS front, Costa Rica, Nicaragua) (Figure 2). Further details 8 O wt. % BDT on age bracketing are provided in the Supplementary 2 6 data. We are unsatisfied with the coarse temporal reso-

– – – O+Na 4

lution in PI (75 39 Ma), PII (35 16 Ma), and PIII (16 2 K 6 Ma); one positive outcome of this study would be to 2 stimulate future integrated geochronologic studies of 16-3 Ma CAVAS PI, II, and III. Geochemical analyses were com- 14 Region Phase a piled for lavas and intrusives of the aforementioned ne 12 I II li e trachy- ka in PAN al al temporal phases PI (n = 139), PII (n = 67), PIII dacite lk CR ba 10 su (n = 122), PIV (n = 19), PV (PVa, n = 15, PVb, n = 86), trachy- and PVI (n = 583) (Table 2 provides mean data linearly basaltic trachy- andesite rhyolite 8 andesite regressed to 55 wt.% SiO2, see Section 3.3. below). We trachy- include chemical data only for those samples that are 6 basalt

confidently assigned to a particular formation and thus 4 dacite age range. 2 basaltic andesite 75-16 Ma basalt andesite 0 3.2. Geochemical data set compilations 40 4550 55 60 65 70 75 80 SiO2 wt. % Details of geochemical data set compilations employed including references are provided in the Supplementary Figure 3. Total alkali-silica (TAS) (Le Bas et al. 1986) classifica- data. tion of (a) PI and PII (75–39 Ma, 35–16 Ma), (b) PIII and PIV (16– 6 Ma, 6–3 Ma) and (c) PV and PVI (5.9–0.02 Ma, 2.6–0 Ma) lavas and intrusives. Abbreviations in (b): BDT, Bocas del Toro; TR-TIS, 3.3. Geochemical data manipulation Talamanca Range-Talamancas Intrusive Suite (see Figure 2 for locations). See Table 2 for the number of samples of each phase It would be optimal to correct all data to primitive plotted here and in subsequent plots. References for all phases Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 basalt, with Mg# = 65, but the paucity of these sam- here and in succeeding figures are provided in the ples makes this impractical. To compare the compiled Supplementary data. suites of geochemical data, selected major and trace element abundances were linearly regressed to large ranges in MgO at a given SiO2 content, e.g. PI

55 wt.% SiO2 (Table 2). This composition lies near lavas and intrusives that range from ~3.8 to 7.5 wt.% the mafic end of our sample suite, for which SiO2 MgO at 55 wt.% SiO2 further justifies our normaliza- contents range from 45 to 78 wt.% (Figure S1, tion to silica. The regressed oxide and trace element

Figures 3 and 4) and close to the mean SiO2 content concentrations are expressed as X55, where X repre- of 55.8 wt.%. Although SiO2 content can vary as a sents the linearly regressed oxide or trace element. function of melt generation mechanisms and pres- This parameter must be interpreted thoughtfully, sures associated with crystal fractionation, the vast given the processes that could contribute to CAVAS

majority of samples have 60 or less wt.% SiO2 with lava compositional diversity. We discuss the implica- mean compositions of each temporal phase ranging tions of interpreting the chemotemporal trends of from ~51 to 59 wt.%. Furthermore, the existence of regressed elemental abundances in Section 6.1. 8 S. A. WHATTAM AND R. J. STERN

In all but one case, the data sets employed provide shoshonite series non-normalized major element compositions; in these tholeiitic and calc-alkaline series cases we have filtered the data such that only samples basalt/ b-a and/ dacite & rhyolite/ gabbro g-d dior granodior & gran that yield major element concentrations that sum to 97–102 wt.% (excluding volatiles) are used, in an iden-

Phase, interval (Ma) 2 tical manner to the CentAm Database (Jordan et al. K255 O R PI (75-39) 0.61 0.09 2012). In the one case where data adjusted to sum to PII (35-16) 1.01 0.51 PIII (16-6) 1.45 0.45 100% is presented (Plank et al. 2002), we filtered the PIV (6-3) 1.89 0.18 PVI (2.6-0) 1.11 0.45 data such that only those with <3% loss on ignition 5 were used (46 of 48 samples). An exception to the cal c-alkMa 4 -6 – – gh- K 16 97 102 wt.% rule is our use of three Phase IV (6 3 Ma) hi III P a 3 16 M Nicaraguan samples (CO-Nic-6, CO-Nic-17, C51, Saginor 35- PII et al 2 Ma . 2011), which have sums between 96.42 and 6-3 med-K calc-alk PIV 96.75 wt.%. As there is only one Phase IV Nicaraguan 1 PI 75 39-Ma d low-K thol sample (C-06-Nic-3) with 97–102 wt.% oxides, we also use these three other samples. Samples with trace ele- Region Phase 6 Va Vb VI PAN ment data only were not used because these could not 5 CR NIC be normalized to 55 wt.% SiO2. Geochemical data from 4 the modern volcanic front in the CentAm Database is 5.9-0 Ma 3 subjected to various other filters (see Jordan et al. 2012,

2 at http://www.earthchem.org/grl). In our geochemical

1 plots, oxides are shown in wt.% and sums are recalcu- O wt.%

2 c lated to 100% anhydrous. Trace element concentrations K are expressed in ppm. 6 Region Phase III IV IS PAN TR-T 5 CR NIC BDT 4 4. Results 16-3 Ma 3 4.1. Chemotemporal trends 2 A major concern is whether or not it is useful to treat 1 b the chemotemporal evolution of an arc as a whole, or subdivide it further. To address this concern, we con- 6 Region Phase I II e sider both collective CAVAS chemotemporal trends PAN honit hos 5 CR s (changes in magma chemistry with time irrespective of K 4 high- k geographic location) and region-specific chemotem- calc-al poral trends. We emphasize that some chemotemporal 3 75-16 Ma med-K lc-alk trends, particularly for PI magmatism, which was the 2 ca longest of all phases (~36 million years), must reflect 1 hol aA low-K t an aggregate of shorter episodes and trends that Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 0 40 45 50 55 60 65 70 75 80 require more radiometric ages to be resolved. For exam-

SiO2 wt.% ple, PI magmatism appears to have begun at about 75 Ma in western Panama in the Sona–Azuero Arc and versus Figure 4. SiO2 K2O plot (Peccerillo and Taylor 1976) of (a) easternmost Costa Rica in the Golfito Arc. Magmatism – – – PI and PII (75 39 Ma, 35 16 Ma), (b) PIII and PIV (16 6 Ma, continued until about 39 Ma in the Sona–Azuero Arc 6–3 Ma), and (c) PV and PVI (5.9–0.02 Ma, 2.6–0 Ma) lavas and intrusives. In (d) the lines represent best-fit linear trend lines of (according to Lissinna 2005), but the lifespan of the collective data (i.e. all geographic regions within a given phase) Golfito Arc was much shorter, terminating by ~66 Ma of phases I, II, III, IV, and VI. The relatively low R2 values of PI (see Buchs et al. 2010 and references therein). Similarly, data are likely the result of element mobility and alteration and Chagres–Bayano magmatism in eastern Panama did not in general for all phases because the data is collective, e.g. begin until about 70 Ma and ended at about the same linearly regressed best-fit lines for discrete regions generally time (39 Ma) (Wegner et al. 2011) as the Sona–Azuero yield much higher R2 values (see Supplementary Figure S1). Abbreviations in top box: bas/gabb, basalt, gabbro; b-a/ g-a, Arc. Thus, PI magmatism reflects complex aggregations basaltic- and gabbroic-andesite; and/dior, andesite, diorite; of trends characteristic of three chemically and tempo- grandior and gran, granodiorite, granite. rally discrete arc segments. For this reason, we have INTERNATIONAL GEOLOGY REVIEW 9

parsed PI data in some instances into these discrete arc first three stages for Panama from 56.3, 58, and segments. 58.4 wt.%, whereas it dropped from 52.3 to 51.4 wt.% between PI and PII for Costa Rica, before climbing to

59 wt.% during PIII (Nicaragua mean SiO2 remained the 4.2. Synopsis of major element chemical evolution same during PII (55.5 wt.%) and PIII (55.1 wt.%)) (Figure Below we summarize the main features identified from S1). The first three phases span ~75 to 6 Ma, or ~92% of our compilations to delineate CAVAS major element the CAVAS history. The overall trend towards increas- evolution (Figures 3–6). We provide a more detailed ingly fractionated magmas reversed in the last 6 Ma, as treatment of major element chemotemporal evolution the arc erupted more mafic lavas during PIV, with a in the Supplementary data. mean of 51.4 wt.% SiO2 (Costa Rica and Nicaragua CAVAS evolved from an early, primitive, mafic con- record mean SiO2 of 51 and 53 wt.%, respectively). struct to an increasingly fractionated and enriched arc CAVAS erupted increasingly enriched lavas over the up until the end of PIII (Nicaragua) or PIV (Costa Rica) first ~69 million years of its history, from low-K lavas

(Figures 3–6). Mean SiO2 increased only slightly over the during PI to medium-K lava during PII to high-K lavas

a b FeOt c 8

7 Region Phase 5.9-0 Ma Va VI Fe 0 PAN -enr PVI 6 CR NIC ichme ol h lk 5 PIV 5 t -a high-Fe lc ca nt 4 med-Fe 10 PIII 3 increasing 15 2 TH 1 arc maturity (from 1-4) low-Fe 2 1 3 20 4 8 C-A 25 PII 7 Region Phase 16-3 Ma III IV PAN 6 CR 30 NIC l 5 ho k t al c- 35 high-Fe al 4 c calc-alk thol t med-Fe 40 3 Time (Ma) Time

FeO /MgO 2 low-Fe 45 1 50 8

7 Region Phase 75-16 Ma 55 I II PI PAN 6 CR 60 ol 5 th lk -a lc high-Fe ca 65 4 med-Fe 3 70 Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 2 low-Fe 75 1

40 45 50 55 60 65 70 75 80 0.6 0.8 1.0 1.2 1.4

SiO2 wt. % Na22 O+K O MgO Tholeiitic Index

t/ Figure 5. (a) Subalkaline affinity discrimination diagram on the basis of SiO2 versus FeO MgO (Miyashiro 1974) superimposed with the high-, medium-, and low-Fe subdivisions of Arculus (2003) of lavas and intrusives of (from bottom to top) PI and PII, PIII and PIV, and PV and PVI magamatism. Two PI PAN, one PIII NIC, one PVa, and two PVI CR with high silica plot outside the plot with FeOt/MgO > 8.). (b) Subalkaline affinity discrimination diagram on the basis of total alkalis-FeOt-MgO (AFM, Irvine and Baragar 1971) superimposed with the trends of increasing arc maturity (from 1 to 4, as labelled in the top plot) from Brown (1982) of (from bottom to top) PI and PII; PIII and PIV; and PVa and PVI lavas and intrusives. Arc maturity trends are from the following arc systems: 1: Tonga-Marianas, South Sandwich; 2: Aleutians-Lesser Antilles; 3: New Zealand, Mexico, Japan; and 4; Cascades, northern Chile, New t Guinea. (c) Tholeiitic index (Fe4.0/Fe8.0 where Fe4.0 and Fe8.0 represent the average FeO of lavas and intrusives with 3–5 and 7–9 wt. % MgO, respectively) (Zimmer et al. 2010) versus time of phases in which THI was calculable (see Supplementary data). The light grey ‘X’s represent mean THI of both or all regions for a particular (temporal) phase. 10 S. A. WHATTAM AND R. J. STERN

1.6

Region Phase PAN 1.4 I II III IV Va, VbVI CR NIC PAN all MORB 1.2 CR

55 NIC )

2 1.0 IBM (TiO 0.8 a

0.4 PVa µ=55 0.69

0.3 55 )

5 0.2 MORB O 2

(P 0.1 IBMVAB b 2.8 2.4 2.0 1.6 OIB

55 1.2 O

2 0.8 IBMVAB K 0.4 c MORB 0.90 0.75 55

O 0.60 2

/Na 0.45

55 0.30 O 2 K 0.15 d 0.00 70 60 50 40 30 20 10 0 Time (Ma)

Figure 6. Concentrations of (a) TiO2, (b) P2O5, and (c) K2O and the ratios of (d) K2O/Na2O of magmatic products of PI (75–39 Ma), PII (35–16 Ma), PIII (16–6 Ma), PIV (6–3 Ma), PV (5.9–0.02 Ma), and PVI (2.6–0 Ma) and phases further discriminated into regions linearly regressed to 55 wt.% SiO2 versus time. MORB and OIB values are from Sun and McDonough (1989) and mean Izu–Bonin–Mariana volcanic arc basalt (IBMVAB) data is calculated from data as compiled by Jordan et al. (2012, N = 517). In (a) and (b), the upper limit (mean plus standard error of the mean, SE) of TiO2 of PVa arc alkalic basalts is 1.68 and the linearly regressed P2O5 of the PVa arc alkalic basalts is 0.69 ± 0.04 (SE). Note also that in (a), the TiO2 of PII PAN and CR overlap and in (b–d), P2O5,K2O, and K2O/Na2O almost completely overlap in PII PAN, CR, and NIC, thereby making symbol discrimination difficult. The light grey ‘X’s here and in subsequent plots represent the mean of both or all three regions of a particular phase. In most instances, the vertical uncertainty (SE

Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 of the mean of regressed compositions) is smaller than the symbol size. To avoid clutter, the horizontal SE (age) is shown for phases only (i.e. this is not shown for regional within-phase data).

during PIII (Figure 4). During the late Miocene (PIV), the Collectively, the CAVAS evolved over its 75 million CAVAS began to erupt less-enriched magmas, earlier in year lifespan from an initial low-K, tholeiitic to a weakly Nicaragua than in Costa Rica and PIV through PVI was calc-alkaline system in its infancy during PI (75–39 Ma) characterized by medium-K calc-alkaline magmatism. A to a medium-K calc-alkaline system in PII (35–16 Ma) to plot of tholeiitic index (THI, Zimmer et al. 2010, see a high-K calc-alkaline system during PIII (16–6 Ma). After Supplementary data for further details of THI) demon- normal arc magmatism ended in Panama by 6 Ma, mag- strates that the overall enrichment of the CAVAS was matic activity was dominated by the production of accompanied by a trend from early tholeiitic and calc- adakites and arc alkaline basalts in Panama and Costa alkaline affinities to stronger calc-alkaline affinities apart Rica and a return to medium-K calc-alkaline igneous from PIII Nicaragua, which exhibits a weakly tholeiitic activity in Costa Rica and Nicaragua thereafter. PVI affinity (Figure 5c). lavas in particular returned to greater iron enrichment. INTERNATIONAL GEOLOGY REVIEW 11

CR PIV µ= 58 CAVAS lavas show similar increases in other incompati- 55 50 BCC ble major elements in addition to potassium between PI Region Phase 40 I II III IV Va, VbVI and PIII, especially P2O5 and K2O/Na2O for Costa Rica PAN 30 CR OIB 55 and Panama, but less so for K2O and K2O/Na2O for NIC

Rb 20 Nicaragua between PII and PIII (Figure 7). A clear diver- OPB (grey, lined) 10 gence between Costa Rica and Nicaragua is seen during a IBMVAB MORB PIV (6–3 Ma), whereas Costa Rica continues to more 1000

800 b CR PIV µ=55 1080 enriched P2O5,K2O, and K2O/Na2O, and Nicaraguan – 600 magmas decrease to lower concentrations (Figures 6b 55 BCC 400 6d). The behaviour of TiO2 is more complex because it is Ba OIB 200 incompatible until magnetite precipitates. Apart from IBM MORB Nicaragua exhibiting anomalously high TiO2 relative to 1500 Panama and Costa Rica, other normalized major incom- 1250 c 1000 OIB patible element concentrations of Panama, Costa Rica, 55 750 and Nicaragua were nearly identical during PIII. Sr BCC 500 250 IBMVAB MORB 4.3. Trace element chemical evolution 8 CR PIV µ= 8.76 d 55 6 4.3.1. Incompatible element trends BCC 55 Mean abundances of all LILEs, Th, U, Pb, Zr, LREEs, 4 OIB Σ Th LREEs/HREEs, and REEs (normalized to 55 wt.% SiO2) 2 IBMVAB increase between PI (75–39 Ma) and PIII (16–6 Ma) and MORB 2.5 usually until 3 Ma (PIV, 6–3 Ma), with a maximum CR PIV µ= 2.82 e 55 exhibited by PVb adakites before decreasing slightly in 2.0 – 55 1.5 Quaternary lavas (PIV, 2.6 0 Ma) (Figures 7 and 8, U BCC 1.0 OIB Table 2). For example, Ba55 rises from PI (~240 ppm) IBMVAB to PII (~470 ppm), PIII (~770 ppm), and PIV (~880 ppm) 0.5 MORB before reaching a maximum in PVa arc alkalic basalts 7 BCC (~900 ppm) and PVb adakites (~980 ppm); subse- f (11) 5 quently, PVI arc basalts record a Ba OIB 55 of 640 ppm, inter- 55 IBMVAB

mediate to that of PII and PIII (Figure 7b, Table 2). The Pb 3 remaining fluid-mobile LILEs (Rb, Sr; Figures 7a and 7c) MORB 1 and K demonstrate similar trends. Trends for LILEs could 150 OIB (280) partially reflect greenschist-facies alteration-derived BCC 120 g mobility (e.g. K, Rb, Ba) and the effects of plagioclase 90 fractionation (Sr); however, the fact that alteration-resis- 55 MORB Zr 60 tant incompatible element concentrations also increase 30 IBMVAB with time suggests that the trends mostly reflect an 0 overall progressive enrichment of incompatible ele- 70 60 50 40 30 20 10 0 Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 ments in CAVAS magmas. Time (Ma) Regressed Th, U, Pb, and Zr concentrations show Figure 7. Concentrations of (a) Rb, (b) Ba, (c) Sr, (d) Th, (e) U, (f) evolutionary trends that are similar to those of the Pb, and (g) Zr of magmatic products of PI, PII, PIII, PIV, PV, and LILEs, with progressive increases between PI and PIV PVI linearly regressed to 55 wt.% SiO2 versus time. References (Th55,U55)(Figures 7d and 7e) or PI and PIII (Pb55, for MORB, OIB, and mean IBM (arc basalt) and other relevant Zr55)(Figures 7f and 7g), with a maximum reached in details are given in the caption for Figure 8 and the value for the PVa adakites or PVb arc alkali lavas in the case bulk continental crust (BCC) is from Rudnick and Gao (2003). Mean oceanic plateau basalt (OPB, thick grey line) is calculated of U . 55 from data of references provided in the Supplementary LREE55 (La55-Nd55) (Ce55 shown only in Figure 8a), Document LREE55 fractionations (La55/Sm55,La55/Yb55)(Figures 8b and 8c), and ΣREE55 (Figure 8d) collectively increase The aforementioned trends with time are summar- similarly from PI to PIV with a maximum exhibited by ized in collective and region-specific chondrite-normal- the PVa adakites followed by a drop back to approxi- ized REE and N-MORB-normalized plots on Figures 9 mately PIV abundances during PVI. and 10, respectively. The most striking feature of the 12 S. A. WHATTAM AND R. J. STERN

80 OIB and PIII; LREE fractionations generally stay the same Region Phase 60 I II III IV Va, VbVI during this interval apart from PII Panama, which exhi- PAN 55 CR BCC bits anomalously high La/Sm (Figure 8). Only slightly 40 NIC

Ce increased LREE fractionations and ΣREEs are shown by OPB (grey, lined) IBMVAB 20 Nicaragua between PII and PIII, before moderate to a MORB significant drops in Ce, LREE fractionations, and ΣREEs 7 6 b are recorded during PIV. In contrast, Ce, LREE fractiona-

55 5 BCC tions, and ΣREEs show significant positive spikes during 4 OIB /Sm 3 PIV in Costa Rica before dropping to values similar to

55 IBMVAB 2 Nicaragua during PVI and Costa Rica and Panama dur- La 1 MORB ing PIII. 30 Similarly, when PI-IV and PVI data are parsed by 25 c OIB

55 region and chondrite-normalized REEs and N-MORB- 20 BCC normalized incompatible plots are considered /Yb 15

55 (Figure 10), a number of salient features are evident: 10 IBMVAB La 5 (1) Regressed, mean chondrite-normalized REEs and MORB 150 N-MORB-normalized trace element patterns of all 120 d OIB (199 ppm) three PI arc segments (Golfito Complex, Sona-Azuero, BCC 55 90 and Chagres-Bayano) fall within the range of composi- IBMVAB – REE 60 tions of 98 82 Ma western Costa Rica units interpreted MORB 30 as CLIP (Figures 10a and 10b). This demonstrates simi-

0 lar sources for PI arc segments and the CLIP (as verified 70 60 50 40 30 20 10 0 by Pb and Nd isotopes, Section 4). (2) The similarity of Time (Ma) the Golfito N-MORB-normalized incompatible element Figure 8. Concentrations and ratios of (a) Ce, (b) La/Sm, (c) La/ signature with that of the Sona-Azuero and Chagres- Yb, and (d) ƩREEs of magmatic products of PI, PII, PIII, PIV, PV, Bayano segments and its prominent negative Nb and PVI linearly regressed to 55 wt.% SiO2 versus time. anomaly coupled with LILE enrichment (Figures 10a and 10b) clearly favour its interpretation as an arc collective plots is the progressive enrichments in line- segment (Buchs et al. 2010) as opposed to a plateau arly regressed LREEs (La-Nd), LREE fractionations (La/Sm, segment (Hauff et al. 2000b). (3) Apart from minor La/Yb), ΣREE, and all but the least incompatible ele- exceptions, the chondrite-normalized REEs and ments (Figures 9a and 8c) as also demonstrated in N-MORB-normalized patterns of PII Panama, Costa Figures 6–8. Compositions of the CAVAS ultimately Rica, and Nicaragua are remarkably similar (Figures became ‘continental-like’ by PIII or certainly by PIV (6– 10c and 10d). This suggests that magmas for all three 3 Ma) (Figures 9, and 9d). The plots in Figure 9 also arc segments were likely derived from similar sources demonstrate that the progression to bulk continental during PII. (4) BCC-like compositions are achieved in crust (BCC) compositions was gradual and not sudden. Panama and Costa Rica arguably during PIII (16–6Ma) It is important to note, however, that when data are and certainly by PIV (Figures 10e–10h). (5) Nicaragua parsed into discrete regions, although Panama and chemotemporal evolution began to diverge from Costa Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 Costa Rica usually follow progressive enrichments with Rica and Panama during PIII (Figures 10e and 10f). time as described above, Nicaragua does not, as is read- Whereas Costa Rica and Panama PIII lavas became ily apparent in Figure 10. For example, concentrations of more enriched than PII, Nicaragua PIII lavas changed Th, U, and Zr decrease between PII and PIII Nicaragua, little from PII compositions. (6) Chemotemporal diver- whereas Pb remains unchanged during this interval, in gence between Costa Rica and Nicaragua is most contrast to Costa Rica and Panama, which (apart from U, obvious in PIV when Costa Rica again continued to which remains unchanged in Costa Rica between PI and more enriched compositions and Nicaragua again did PII) show increases in these elements between PI and not (Figures 10g and 10h). (7) N-MORB-normalized PIII (Figure 7). Similarly, when parsed into region, LREEs, incompatible element patterns of PVI (2.6–0Ma) LREE fractionations, and ΣREE deviate from overall Costa Rica and Nicaragua are similar (Figures S2i, j); trends, suggesting progressive enrichment. For exam- PVI thus marks the ‘resetting’ of production of magmas

ple, only Ce55 and ΣREEs exhibit higher concentrations with more depleted compositions similar to those gen- with time in both Costa Rica and Panama between PI erated during PIII or even earlier. INTERNATIONAL GEOLOGY REVIEW 13

200 a PVa 100 PV (5.9-0.02 Ma) arc alks (inset) Bulk Continental Crust (60.2 wt.% SiO ) PVb 2 adaks PI PIV

10 55

Phase, interval (Ma) mean Izu-Bonin Mariana PI (75-39) PIV (6-3) mean Volcanic Arc basalt PII (35-16) PVI (2.6-0) PIII (16-6)

Phase /Chondrites b 2 La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu 1000 c

100

10 55

Phase /N-MORB d .1 Cs Ba U K Ce Pr P Zr Eu Dy Yb Cs Ba U K Ce Pr P Zr Eu Dy Yb Rb Th Nb La Pb Sr Nd Sm Ti Y Lu Rb Th Nb La Pb Sr Nd Sm Ti Y Lu

Figure 9. (a, c) Chondrite-normalized REEs and (b, d) N-MORB-normalized incompatible element plots of incompatible and REE concentrations of collective (a, b) PI, PII, PIII, PIV, PV, and PVI magmatic products linearly regressed to 55 wt.% SiO2 and (c, d) PI and PIV from (a) and (b) versus Bulk Continental Crust (BCC, 60.2 wt.% SiO2, Rudnick and Gao 2003) and mean Izu–Bonin–Mariana volcanic arc basalt composition (as compiled by Jordan et al. 2012, where mean IBM VAB is based on the number of samples with a full suite of REE). References for these tectonomagmatic suites are the same in succeeding figures. Chondrite and N-MORB abundances are from Nakamura (1974) and Sun and McDonough (1989), respectively.

The REE and N-MORB-normalized plots discriminated mantle melting; degrees of crustal melting; or a combi- by region and discrete PI arc units (Figure 10) show nation of these processes. Incompatible element ratios subtle intra-phase differences not listed above. For that gauge source fertility such as Zr/Y and Nb/Yb are example, Figures 10a and b illustrate that the useful for elucidating source evolution, i.e. changes in 68–39 Ma Chagres–Bayano Arc lavas of eastern the degree of partial melting or fertility of the source.

Panama more closely resemble average Izu–Bonin– Zr55/Y55 (Figure 11a) and Nb55/Yb55 (Figure 11c) both Mariana (IBM) volcanic arc basalt (VAB) than do lavas increase with time between PI and the end of PIV in from the ~75–39 Ma Sona-Azuero and 75–66 Ma Golfito magmatic rocks of Panama and Costa Rica. When

Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 segments. Despite this and other minor region-specific parsed into region however, increases in both Zr/Y differences, the regressed incompatible trace element and Nb/Yb between PI and PII Costa Rica are slight data demonstrate that CAVAS magmas were initially and the collective shift to higher Nb/Yb during PII is generated from a depleted mantle source, which was due to Panama, which exhibits anomalously high Nb/Yb also strongly influenced by plume contributions (see (~3). The most significant shifts to higher Zr/Y and Nb/ below). The CAVAS changed little over the first half of Yb occurred between PIII and PIV in Costa Rica. As there its lifespan (during PI, 75–39 Ma); segments in Panama, is no evidence of the presence of garnet in the source of Costa Rica, and Nicaragua were all similar during PII (35– any CAVAS igneous rocks except for PV adakites and arc 16 Ma), but magma compositions diverged thereafter. alkali basalts, higher Zr/Y and Nb/Yb probably represent either a shift to more enriched sources or lesser degrees of partial melting; conversely, as these ratios decrease in 4.3.2. Trends in magma enrichment/depletion and Nicaragua between PII and PIII, this suggests either subduction additions shifts to higher degrees of partial melting or tapping Arc magmatic evolution can reflect changing: slab con- of a more depleted source or both. tributions; mantle wedge compositions; degrees of 14 S. A. WHATTAM AND R. J. STERN

200 500 PI 75-39 Ma Golfito Arc (CR, n = 8) 100 Sona-Azuero Arc (PAN, n = 57) Chagres-Bayano Arc (PAN, n = 68) 100 all (PAN+CR, n = 139) 98-82 Ma western Costa Rica igneous units interpreted as CLIP (N = 37) 10

10 1 mean Izu-Bonin Mariana mean Volcanic Arc basalt (n = 177) a b 200 500

PII 35-16 Ma PAN (ALI+non-ALI, n = 35 ) 100 PAN (ALI, n = 13) CR (n = 15) 100 NIC (n = 17) all (PAN+CR+NIC, n = 67) Phase I (all)

10 1

c d 200 500

PIII 16-6 Ma PAN (n = 54) 100 CR (n = 38) NIC (n = 31) 100 all (PAN+CR+NIC, n = 123) Phase I (all) Phase II (all)

10 /N-MORB 55 55 10 Bulk Continental Crust e

(BCC, 60.2 wt.% SiO2 ) 1 has hase /Chondrites P

P e f 200 500

CR (n = 15) 100 NIC (n = 4) all (CR+NIC, n = 19) 100 Phase I (all) Phase II (all) BCC Phase III (all)

10 1

PIV 6-3 Ma g h 200 500

PVI 2.6-0 Ma CR (n = 405) 100 NIC (n = 178) all (CR+NIC, n = 573) 100 PI (all) PII (all) PIII (all) PIV (all)

10 BCC 1 Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 i j 2 .1 La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs Ba U K Ce Pr P Zr Eu Dy Yb Rb Th Nb La Pb Sr Nd Sm Ti Y Lu

Figure 10. (a, c, e, g, i) Chondrite-normalized REE and (b, d, f, h, j) N-MORB-normalized incompatible plots of incompatible and REE concentrations (a, b) PI, (c, d) PII, (e, f) PIII, (g, h) PIV and (i, j) PVI magmatic products of Panama, Costa Rica, and Nicaragua parsed into regions and in (a) distinct arc segments. In (a, b) the 98–82 Ma western Costa Rica units interpreted as CLIP comprise the Nicoya Complex (Sinton et al., 1997; Hauff et al. 2000b), and the Herradura and Tortugal complexes (tholeiitic basalts and diabases only, n = number of samples with a complete suite of REE analyses, Hauff et al. 2000b). In (c, d) ALI represents PII ~32–19 Ma adakitic-like intrusives (Whattam et al. 2012) that formed subsequent to the shutdown of PI and via partial melting of mafic arc substrate (see text). Chondrite and N-MORB abundances are from Nakamura (1974) and Sun and McDonough (1989), respectively.

CAVAS PI La55/Sm55 and La55/Yb55 are similar to OPB Zr55/Y55 of PII Panama is equivalent to that of mean in PI; by PII both ratios meet or exceed those of IBM OPB, whereas PII Costa Rica and Nicaragua are slightly VAB, although the more LREE-enriched nature of less enriched with values intermediate to that of MORB Panama PII lavas is apparent (Figure 8). Similarly, and OPB (Figure 11a). Apart from the PVa alkalic INTERNATIONAL GEOLOGY REVIEW 15

PVb µ= 10.2 55 Costa Rica and Nicaragua, the possible causes of which 8 OIB (9.7) Region Phase are discussed in Section 6.3.3. 7 I II III IV Va, VbVI BCC

55 6 PAN Also plotted on Figure 11 (d) is La55/Nb55, which CR /Y 5 NIC gauges the ‘depth’ of the negative Nb anomaly and is

55 OPB (grey, lined) 4 Zr ~1 or less for magmas formed away from subduction 3 a IBM (1.9) MORB zones, e.g. La/Nb of OIB and MORB are 0.8 and 1.1, 60 respectively (Figure11d). Collectively, La /Nb varies NIC PIV µ=55 60.6 55 55 50 IBMVAB 55 little from PI to PIII (2.1–2.6), but jumps to higher values 40 /Nb 30 MORB in PIV Costa Rica (4.1), similar to the PVb adakites (4.2) 55

Zr 20 and Nicaragua (5.4), before falling back to values similar BCC – 10 b OIB to PI PIII in PVI Costa Rica and Nicaragua (2.2, collec- 5 OIB (22.2) tively) (Figure 11d, Table 2). PVa alkali basalts record a PVa µ= 28.6 55 BCC 4 PVb µ=55 6.8 55 very low OIB-like La55/Nb55 of 0.8. 3 Ratios of specific fluid-mobile element to HSFE (Sr/ /Yb

55 2 Nd, Pb/Ce, U/Th, Ba/La, Ba/Th, Ba/Nb, and Th/Nb) are

Nb 1 useful for inferring changes in subducted slab contribu- c MORB IBMVAB tions and are plotted on Figure 12 to illustrate how this 6 IBMVAB 5 d (7.71) influence on the CAVAS source has changed over time. 55 4 – BCC These plots generally show increases between PI PIII /Nb 3 and PI–PIV, but when considered on the basis of dis- 55 2 MORB

La crete region it is apparent that there is little change in 1 OIB 0 Panama and Costa Rica and the most significant 70 60 50 40 30 20 10 0 changes are in Nicaragua, which generally show large Figure 11. Ratios of (a) Zr/Y, (b) Zr/Nb, (c) Nb/Yb, and (d) La/Nb jumps in these ratios between PII and PIV or PII and PIII. of magmatic products of PI, PII, PIII, PIV, PV, and PVI linearly Sr55/Nd55 collectively increases from PI (0.11, 3.39, 25.8) regressed to 55 wt.% SiO2. In (a) the Zr55/Y55 of PVb adakites to PII (0.26, 5.45, 31.1) to PIII (0.31, 7.74, 33.2) to PIV plots outside of the plot 10.23 ± 0.33 and in (c) the PVa arc (0.41, 9.33, 37.8) and is the only ratio that also increases alkalic basalts and PVb adakites plot outside the plot with Nb55/ when parsed into regions (Figure 12a). CAVAS U/Th in Y of 28.56 ± 2.44 and 6.79 ± 0.50, respectively. 55 Panama and Costa Rica scatters with no hint of a trend and is broadly higher than mean IBM VAB even in PI

basalts and PVb adakites, which have higher regressed (Figure 12c). However, Pb55/Ce55 increases from PI (0.10) Zr/Y than BCC (~7) and even ocean island basalt (OIB) to PII (0.13) to PIII (0.14) before dropping to a value (9.7) in the case of the adakites (Zr55/Y55 of 10.2), (0.08) lower than that of PI during PVI (Figure12b) and CAVAS Zr55/Y55 is much less than BCC. Nevertheless, enrichment is dominated by Nicaragua with much it is evident that both Panama and Costa Rica evolved higher Pb/Ce than Panama and Costa Rica during PII-

towards BCC-like compositions during PIII with Zr55/Y55 PI. CAVAS Pb55/Ce55 never reaches as high a mean as intermediate to OPB and BCC, whereas Nicaragua Zr55/ IBM VAB (Figure 12b), but by PIII Nicaragua Pb/Ce is Y55 decreased slightly in PIII. A similar trend is seen in identical to that of mean IBM VAB and by PIV it is Zr55/Nb55 and Nb55/Yb55, which reflect source deple- identical to BCC. Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 tion and/or the degree of partial melting. Overall, these Fluid-mobile Ba and melt-mobile Th are good tracers trends demonstrate a shift to strong enrichment of total and shallow subducted slab inputs, especially beneath Costa Rica, but only slight enrichments in when compared with La and Nb (Pearce and Peate

Panama and Nicaragua between 35 and 16 Ma. 1995). Ba55/La55,Ba55/Th55,Ba55/Nb55, and Th55/Nb55 During PIV, dramatic changes are recorded with diver- versus time are also plotted in Figure 12. Variations in gence between Costa Rica and Nicaragua. Whereas PIV Ba/Th in the CAVAS have been interpreted as reflecting Costa Rica records low, BCC-like concentrations of Zr55/ differences in the amount or composition of sediments Nb55 (16) akin to the PVb adakites (18), PIV Nicaragua over time (Patino et al. 2000). Ba/Th in general reflects alternatively records by far the highest Zr55/Nb55 of shallow additions of subduction-related fluids to the any phase or region (61). An identical pattern is seen mantle source (Pearce et al. 2005)(Figure 12e). CAVAS

in Nb55/Yb55 with PIV Costa Rica magmas recording the Ba55/Th55 increases from PI (280) to PII (326) to PIII (355) highest ratios (5) and Nicaragua the lowest ratios before subsequently dropping to values near those of (0.42). This further demonstrates a dramatic composi- IBM VAB during PIV, PV, and PVI (Figure 12e, Table 2). tional divergence between 6 and 3 Ma igneous rocks in However, and similar to other slab contributions 16 S. A. WHATTAM AND R. J. STERN

60 discussed above, this collective progression in Ba/Th 50 Region Phase a I II III IV Va, VbVI (and Ba/La) is the influence of Nicaragua (Figures 12d 55 PAN 40 CR and 12e). For example, Ba/Th remains essentially the

/Nd NIC IBMVAB 30 – – 55 OIB same in Panama (~200 300) during PI PIII and in

Sr 20 BCC Costa Rica, although Ba/Th increases slightly between 10 MORB PI and PII (from ~425–450), this ratio drops suddenly in OPB (grey, lined) 0.30 IBMVAB PIII to a value (~250) similar to that of Panama. (Ba/Nb) 0.25 b BCC et al

55 and deep (Th/Nb) subduction additions (Elliot . 0.20 IBMVAB 1997; Pearce et al. 2005)(Figures 12f and 12g) show /Ce 0.15 55 0.10 no increases between PI and PII and only slight Pb 0.05 MORB increases in PIII before jumping to high values in PIV. & OIB 0.9 Whereas PIV Ba/Nb is unchanged from PIII in Costa Rica, CR µ= 1.11 0.8 c Ba/Nb jumps to an extremely high value in PIV in

55 0.7 0.6 Nicaragua (702). Th/Nb is much higher in both PIV

/Th MORB 0.5 55 IBMVAB Costa Rica and Nicaragua than in PIII Costa Rica and

U 0.4 OIB Nicaragua. Similar to other trends described above, PVI 0.3 BCC 140 Ba/Nb and Th/Nb drop, similar to or slightly less than 130 d those of PI-PIII. PVa arc alkalic basalts show low, OIB-like 90 Ba55/Nb55 and Th55/Nb55 and the PVb adakites show Ba/ 80 Nb similar to PIII and Th/Nb similar to PIV. The signifi-

55 70 60 cance in the difference in trends displayed by Ba/Th /La 50 versus

55 Ba/Nb and Th/Nb is explored in Section 6. 40 Ba 30 BCC 20 IBMVAB 10 5. Radiogenic isotope trends MORB 1200 We compiled available data for radiogenic isotopes Sr, 1100 e Nd, and Pb. We do not emphasize Sr isotopic composi- 1000 tions as these are vulnerable to alteration. As Nd and Pb 900 55 isotopes are much less affected by alteration and record 500 shallow subduction additions /Th the most obvious trends in CAVAS isotopes source evo-

55 400 lution (Figures 13–15), we restrict our discussion here to Ba 300 IBMVAB these isotopes, but provide details of Sr isotopes in the 200 OPB (grey, line) OIB BCC Supplementary data. There are no Hf isotope data for 100 MORB pre-PVI CAVAS sequences, but these should start to 400 350 appear as LA-ICP-MS zircon geochronology of the f NIC PIV µ= 702 55 300 total subduction additions region advances. Sources for the compiled isotope 250 IBMVAB /Nb 200 data sets are listed in the captions for Figures 13–15. 55 250 BCC

Ba 100 OIB

Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 50 1.4 MORB 5.1. Pb isotopes 1.2 g 55 1.0 BCC Raw isotope data (Sr, Nd, Pb) (and the sources for these 0.8 IBMVAB /Nb deep subduction additions 55 0.6 data, in addition to being provided in the captions of 0.4 Th – OIB Figures 13 15) is provided in Supplementary Table S3. 0.2 206 204 208 204 MORB We show the initial Pb/ Pb versus Pb/ Pb plots 0.0 70 60 50 40 30 20 10 0 of PI–PVI in Figure 13. Similar to other isotope plots Time (Ma) presented (Figures 14 and 15, Supplementary Figure Figure 12. Ratios of (a) Sr/Nd, (b) Pb/Ce, (c) U/Th, (d) Ba/La, (e) S4), we compare the composition of PI lavas and intru- Ba/Th, (f) Ba/Nb, and (g) Th/Nb of magmatic products of PI, PII, sives with those of the 98–82 Ma western Costa Rica PIII, PIV, PV, and PVI linearly regressed to 55 wt.% SiO2 versus igneous units (at Nicoya, Herradura, and Tortugal, see time. Figure 2 for locations) (Sinton et al. 1997; Hauff et al. INTERNATIONAL GEOLOGY REVIEW 17

GP at 2.8-0 Ma other western CR igneous units PVa & PVb (from d) 39.4 c 39.4 98-82 Ma CLIP 5.9-0.01 Ma 80-60 Ma accreted OIB 39.2 39.2

PII & PIII Panama Region Phase 39.0 (from b, light green) 44.8 39.0 Va Vb 65.9 61.5 PAN PI 75-39 Ma CR 38.8 49.4 38.8 Region Phase I 46.9 central, southern & PAN SAA CBA 38.6 38.6 49.9 52.3 CR GA SE Costa Rica BVF & 40.0 VF alkaline (pale yellow) 54.3 CBA (eastern 38.4 38.4 67.5 68.5 (Panama, light pink) 60.9 64.6 38.2 38.2 DM PII Costa Rica (from b, light blue) 72.3 SAA (central Panama, light yellow) a 4 4 38.0 38.0 39.6 PII-PIV b PVI 39.4 SP 39.4 Pb/ Pb 35-3 Ma bductingGD) Pb/ Pb 2.6-0 Ma su (N ) 98-82 Ma CLIP SP D 08 20 208 20 G 39.2 Region Phase (from a, white) (C R 39.2 Region Phase VI II III IV C 21.0 C PAN ting CR NW central 8.6 CCR c 39.0 CR bdu 39.0 NIC NW SW 19.2 14.7 su eastern 21.9 CR, NIC 38.8 10.7 38.8 CR, NIC 6.8 Cocos Plate 16.5 21.7 CR CR 16-6 Ma NE CR sediments 8.8 20.1 35-16 Ma central CR 14.3 CR 6-3 Ma 38.6 20.0 7.1 38.6 NW CR 34.4 CR Miocene 32.4 18.5 CR 75-66 Ma NW NIC SW NIC 38.4 16.7 Miocene NIC Miocene 38.4 a Costa Rica PAN M 17.0 9-6 Region Miocene Nicaragua 1 38.2 Ma 38.2 9 Galapagos Islands (i.e., GP) at 2.8-0 Ma Cocos/Nazca Plate 75-3 d 38.0 38.0 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 206Pb/ 204 Pb 206Pb/ 204 Pb

Figure 13. 206Pb/204Pb versus 208Pb/206Pb isotope data for (a) PI (75–39 Ma) CAVAS magmatic products of Panama and Costa Rica versus lavas from 98–82 Ma western Costa Rica igneous units that include Nicoya, Herradura, and Tortugal (Sinton et al. 1997;Hauffet al. 2000a, 2000b) (locations are shown on Figure 2), which are interpreted by most as CLIP or ‘plume- and arc-related’ (PAR, by Whattam and Stern 2015, see text); 80–60 Ma accreted OIB in western Costa Rica (Hauff et al. 2000b); and 2.8 Ma lavas of the Galapagos Islands (i.e. Galapagos Plume, GP, White et al. 1998); and CAVAS magmatic products of (b) PII–PIV (35–3 Ma), (c) PVa (arc alkaline basalts, 5.90– 0.01 Ma) and PVb (adakites, 4.20–0.15 Ma), and (d) PVI (2.6–0 Ma) in Panama, Costa Rica, and Nicaragua. In (a) the PI Sona-Azuero Arc (SAA) samples from central Panama are discriminated from all others from the Chagres-Bayano Arc in eastern Panama to highlight the compositional differences. Also in (a) two accreted OIB samples from Osa (OSA6 and OSA16, Hauff et al. 2000b) exhibit very low 208Pb/204Pb of 37.946 and 38.025 and plot of the diagram. Note that in (d), as the PVI samples from northwest Nicaragua might have been constructed on continental as opposed to oceanic basement (see Section 1), these have been discriminated (open triangles) from samples of southwest Nicaragua constructed on oceanic basement. Moreover, for visual clarity to see clearer the compositional differences between central versus northwest Costa Rica during PVI, samples from Central Costa Rica (pink squares) are discriminated from those of northwest Costa Rica. Geographical segmentation of PVI magmatism in Costa Rica and Nicaragua (e.g. central Costa Rica, NW Nicaragua) is based on the geographical segmentations of Hoernle et al.(2008, their Figure 1); an exception is our use of ‘eastern Nicaragua’, which comprises samples from the eastern margin of Nicaragua between ~12.3 to 12.6° N and –83.7 to 84.0° E (e.g. Cukra Hill, Pearl Lagoon from Gazel et al. 2011). Isotope data from the CAVAS are from Gazel et al.(2009,PII–PIV, Costa Rica); Gazel et al.(2011, PIV, PVa, PVb, PVI, Costa Rica); Hoernle et al.(2008, PVa, PVb, Panama; PVa southern and central Costa Rica BVF, PVb southern Costa Rica; PVI NW, central and southern Costa Rica VF; PVI, southwest, and northwest Nicaragua; and the Miocene Costa Rica and Miocene et al –

Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 Nicaragua fields); Wegner .(2011)(PI PIII, PVb, Panama); the GEOROC data base at: http://georoc.mpchainz.gwdg.de/georoc/ (PVI, Costa Rica, and Nicaragua; only dated samples that yield ages of Quaternary are plotted); and the CentAm Database version 1.02 at: http://www.earthchem.org/grl.databases (PVI, Costa Rica and Nicaragua, as compiled by Jordan et al. 2012). Fields for the subducting Cocos and Coiba ridges (CCR, or Central Galapagos Domain, CGD), Seamount Province (SP, or Northern Galapagos Domain, NGD), and Cocos/Nazca Plate are from Hoernle et al.(2000) and Werner et al.(2003) (as presented in Hoernle et al. 2008). The compositions of (subducting) Cocos sediments are from Feigenson et al.(2004) and the depleted mantle (DM) composition is from Werner et al.(2003; see Gazel et al. 2009 for details). The boundaries of the stippled field with horizontal lines represent mixing lines that connect the three compositional end-members (DM, depleted mantle, SP, and CCR, Cocos and Coiba ridges) required to explain the isotopic variations. In (a) and (b) the numbers beside some PI, PII, and PIII Panama samples represent ages in Ma (from Wegner et al. 2011). The boxes inset in (b) and (d) represent the summaries of temporal isotopic changes without the symbols (the box in (b) represents the summary of (a) and (b) and the box in (d) represents the summary of (d) only). Abbreviations: BVF, behind volcanic front; CCR, Cocos-Coiba ridges; CGD, Central Galapagos Domain; NGD, Northern Galapagos Domain; SP, Seamount Province; VF, volcanic front.

2000a, 2000b), interpreted by most as CLIP oceanic CAVAS via melting of a mixed, plume- and subduction- plateau fragments or as (hybrid) plume- and arc-related modified source (Whattam and Stern 2015). We do this units generated soon after subduction initiation at the particularly as trace element chemistry of PI magmas 18 S. A. WHATTAM AND R. J. STERN

GP at 2.8 Ma (from d) DM PI 75-39 Ma Region Phase PVa & PVb 0.51310 0.51310 Va Vb 5.9-0.01 Ma Region Phase I PAN GP at CR 0.51305 60 Ma PAN SAA CBA 0.51305 BUR CR GA

0.51300 other western CR igneous units 0.51300 40.0 54.3 98-82 Ma CLIP 61.5 0.51295 80-60 Ma accreted OIB 0.51295

0.51290 60.9 0.51290 Seamousub 72.3 49.9-44.8 d 52.3 uc all regions nt tin g 0.51285 67.5 0.51285 (N P of Costa Rica G rov SAA 65.9 D 68.5 BVF & VF alkaline ) in (central c GP at e 0.51280 Panama, 90 Ma 0.51280 (yellow) TAP CBA (eastern Panama, light pink) a c 0.51275 0.51275 0.51315 Cocos/Nazca Plate Region Phase PII- PIV PVI 0.51310 II III IV 0.51310

Nd/ Nd PAN 35-3 Ma Nd/ Nd 2.6-0 Ma 26-6 Ma CRVFH CR 1430.51305 144 20.0 1430.51305 144 17.0 19.2 21.9 subducting Cocos 18.5 7.3-6.8 & Coiba ridges 0.51300 21.0 0.51300 CCR (CGD) 0.51295 14.2 8.6 0.51295 Region 10.7 Galapagos Islands 34.4 16.7 20.1 SP 32.4 21.7 see (c) (i.e., GP) at 2.8-0 Ma 0.51290 98-82 Ma CLIP 0.51290 14.7 DM CR, PAN NW (from a) DM NIC CR 35-6 Ma central Region Phase VI CR 6-3 Ma CR E 0.51285 0.51285 NW CR NW central CCR SW CCR a NIC NW SW eastern CR N 0.51280 75-66 Ma 0.51280 Cocos Plate sediments PA 6 M 5- CR, NIC b (0.5124-0.5127) 7 d 0.51275 0.51275 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 206Pb/ 204 Pb 206Pb/ 204 Pb

Figure 14. 206Pb/204Pb versus 143Nd/144Nd isotope data for (a) PI (75–39 Ma) CAVAS magmatic products of Panama and Costa Rica versus 98–82 Ma western Costa Rica units, which include Nicoya, Herradura, and Tortugal (Sinton et al. 1997; Hauff et al. 2000b)and which are interpreted by most as CLIP or as plume- and arc-related units generated soon after subduction initiation at the CAVAS (Whattam and Stern 2015); 80–60 Ma accreted oceanic island basalt (OIB) units in western Costa Rica that comprise the Burica, Osa, and Burica units (Hauff et al. 2000b; see also Hoernle et al. 2002; Hoernle and Hauff 2007; Buchs et al. 2009, 2011); and the Galapagos Plume (GP) at 90 Ma, 60 Ma (fields at 90 and 60 Ma from Hauff et al. 2000a), and 2.8 Ma (i.e. the Galapagos Islands, White et al. 1998); and CAVAS magmatic products of (b) PII–IV (35–3 Ma), (c) PVa (arc alkaline basalts, 5.90–0.01 Ma) and PVb (adakites, 4.20–0.15 Ma), and (d) PVI (2.6–0 Ma) in Panama, Costa Rica, and Nicaragua. In (a): the PI Sona-Azuero Arc (SAA) samples from central Panama are differentiated from the Chagres-Bayano Arc in eastern Panama to highlight the compositional differences; BUR and TAP stand for Burica (an accreted OIB unit in western Costa Rica) and Tortugal alkali picrites, repectively. In (a) and (b) the numbers beside some PI, PII, and PIII Panama samples represent ages in Ma (from Wegner et al. 2011). In (b) the 26–6 Ma CRVF field demarcates samples from the southern Costa Rica Miocene volcanic front (superscript ‘H’ for Hoernle et al. 2008). The boxes inset of (b) and (d) represent summaries of temporal isotopic changes without the symbols (the box in (b) represents the summary of (a) and (b) and the box in (d) represents the summary of (d) only). In (d) the geographical segmentation of PVI magmatism in Costa Rica and Nicaragua (e.g. central Costa Rica, northwest Nicaragua) is based on the geographical segmentations of Hoernle et al.(2008,theirFigure 1); an exception is our use of ‘eastern Nicaragua’, which comprise samples from the eastern margin of Nicaragua between ~12.3 to 12.6° N and –83.7 to 84.0° E (e.g. Cukra Hill, Pearl Lagoon from Gazel et al. 2011). Isotope data from the CAVAS are from Gazel et al.(2009,PII–PIV, Costa Rica); Gazel et al.(2011, PIV, PVa, PVb, PVI, Costa Rica); Hoernle et al.(2008, PVa, PVb, Panama; PVa southern and central Costa Rica BVF, PVb southern Costa Rica; PVI NW, central and southern Costa Rica VF; PVI, southwest and northwest Nicaragua; and the Miocene Costa Rica and Miocene Nicaragua fields); Wegner et al.(2011)(PI–PIII, PVb, Panama); the GEOROC data base at: http://georoc.mpchainz.gwdg.de/ Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 georoc/ (PVI, Costa Rica, and Nicaragua; only dated samples that yield ages of Quaternary are plotted); and the CentAm Database version 1.02 at: http://www.earthchem.org/grl.databases (PVI, Costa Rica, and Nicaragua, as compiled by Jordan et al. 2012). Fields for the subducting Cocos and Coiba ridges (CCR, or Central Galapagos Domain, CGD), Seamount Province (SP, or Northern Galapagos Domain, NGD), and Cocos/Nazca Plate are from Hoernle et al.(2000) and Werner et al.(2003) (as presented in Hoernle et al. 2008). The compositions of (subducting) Cocos sediments are from Feigenson et al.(2004) and the depleted mantle (DM) composition is from Werner et al.(2003; see Gazel et al. 2009 for details). Other abbreviations not defined above: CBA, Chagres-Bayano Arc; CCR, Cocos– Coiba ridges; CGD, Central Galapagos Domain; NGD, Northern Galapagos Domain; SAA, Sona–Azuero Arc.

exhibits a clear affinity for 98–82 Ma units interpreted as complexes of Panama (which encompass PI–PIII of this CLIP (see Section 4, main text). study) and comparison of their Pb isotope compositions Based on isotope and radiometric data amassed for with those of CLIP (from data of Kerr et al. 1997, 2002; the ~73–69 Ma Sona-Azuero (based on the radiometric Hauff et al. 2000a), Wegner et al.(2011) demonstrated: (1) age data of Wegner et al. 2011 only), ~70–39 Ma an overlap in initial 206Pb/Pb204 versus initial 208Pb/204Pb Chagres–Bayano, and the ~20–7MaCordilleranarc of the older arc systems with the CLIP; and (2) a shift to INTERNATIONAL GEOLOGY REVIEW 19

PV PIV PVI CLIP PI PII PIII 0.51315 other western CR igneous units onset of collision of Cocos Ridge 0.51310 98-82 Ma CLIP with CAVAS trench? 80-60 Ma accreted OIB Osa SW 0.51305 NW NIC Burica NIC NW CR (N = 5) eastern 0.51300 NIC central CR 0.51295 Galapagos Islands

Nd/0.51290 Nd (i.e., GP) at 2.8-0 Ma (grey bar) 143 144 0.51285 Region Phase I PII PIII PIV PVa PVb PVI PAN SAA CBA 0.51280 age of accreted CR GA NW central NIC NW SW eastern OIB in western CR 0.51275 100 90 80 70 60 50 40 30 20 10 0 Time (Ma)

Figure 15. Initial 143Nd/144Nd versus time for CAVAS magmatic products of PI–PVI in Panama, Costa Rica, and Nicaragua. Large symbols represent undated samples and are plotted at the mean age of the associated unit; small symbols represent dated samples. Colour code schemes with their age brackets (age uncertainties) and encompassed units from left to right are as follows: light grey, 98–82 Ma, western Costa Rica igneous units interpreted as CLIP; light purple, PI (75–39 Ma); light blue, PII (35–16 Ma); light green, PIII (16–6 Ma); PIV (6–3 Ma), light yellow; and PVI (2.6–0 Ma), light pink. The 80–60 Ma accreted OIB field is demarcated by the region between the two solid grey vertical lines and the PV (PVa arc alkalics and PVb adakites, 5.9–0.15 Ma) field is demarcated by the region between the two dotted vertical lines. For Costa Rica and Nicaragua PVI samples, locations were plotted with coordinates provided in each relevant paper based on the way Hoernle et al.(2008, Figure 1) subdivided Costa Rica into northwest central, and southwest and Nicaragua into southwest and northwest. In most cases the vertical error (compositional uncertainty) is smaller than the symbol size and is not plotted. References for data sets are as given in the Figure 18 caption.

more radiogenic Pb (higher 206Pb/204Pb and 208Pb/204Pb) (Hoernle et al. 2008;Gazelet al. 2009, 2011). These work- with time. This plateau-like affinity of the oldest arc ers also demonstrated the low 206Pb/204Pb and samples (Figure 13a) is not surprising as plume emplace- 208Pb/204Pb nature of Nicaragua arc magmas, the com- ment was likely the catalyst for the initiation of the positions of which have changed little from the Miocene CAVAS (plume-induced subduction initiation, or PISI, (e.g. compare Figures 13b and 13d). Whattam and Stern 2015) and hence early subduction Figure 13a also shows the 40Ar/39Ar ages of tapped a strongly plume-modified mantle source. Initial Panamanian arc samples (Wegner et al. 2011), which Pb isotopic compositions evolve with time during PII to further underscores the compositional similarities of PI PIV, but show increasingly distinctive regional variations, arc samples with those of 98–82 Ma western Costa Rica with Nicaragua always being less radiogenic than Costa CLIP sequences (Sinton et al. 1997; Hauff et al. 2000a, Rica–Panama. Similar to the trend to enriched sources 2000b). However, Figure 13 also demonstrates that with time explained above, Hoernle et al.(2008)and samples from the 73–69 Ma Sona–Azuero Arc Gazel et al.(2009, 2011) demonstrated a similar trend in (Wegner et al. 2011) of central Panama and the ~75– Costa Rica after 6 Ma, which is location dependent 66 Ma Golfito Arc (Hauff et al. 2000a, 2000b;Buchs

Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 (Feigenson et al. 2004; Goss and Kay 2006; Hoernle et al. 2010) of easternmost Costa Rica are markedly et al. 2008;Gazelet al. 2009, 2011)(Figures 13c and less radiogenic than lavas of the CLIP and lavas of the 13d). For example, whereas northwest PVI Costa Rica younger (~70–39 Ma) Chagres–Bayano Arc (Wegner lavas plot similarly to older PIV Costa Rica magmas (com- et al. 2011) of eastern Panama (at 77.2–79.9° W). pare Figures 13b and 13d), with relatively low Wegner et al.(2011) noted that the Sona–Azuero Arc 206Pb/204Pb (~18.6–19.1) and 208Pb/204Pb (~38.2–38.8), was less radiogenic with respect to Pb isotopes than PVI lavas from central and northeast Costa Rica exhibit the Chagres-Bayano Arc, but did not correlate this higher 206Pb/204Pb (~19.0–19.3) and 208Pb/204Pb (~38.7– observation with data from lavas of the Golfito 39.1), with the most enriched samples overlapping the Complex or other arc-related complexes in Costa Rica compositional field defined by the subducting Seamount and Nicaragua. Whereas the two samples from the Province (Figure 13d, see location of province on Sona-Azuero Arc (72.3 and 67.5 Ma) exhibit Figure 2). This along-arc isotopic provinciality is one of 206Pb/204Pb of 18.53–18.65 and 208Pb/204Pb of 38.11– the main pieces of evidence given that subducted sea- 38.41 similar to that of the Golfito Arc with 206Pb/204Pb mounts modified the Costa Rica mantle source after 6 Ma and 208Pb/204Pb of 18.57–18.77 and 38.20–38.33 (Hauff 20 S. A. WHATTAM AND R. J. STERN

et al. 2000b), respectively, the oldest Chagres-Bayano The data for Pb isotopes discussed above demon- Arc sample (68.5 Ma) exhibits 206Pb/204Pb of 18.82 and strates that while sources for Panama and Costa Rica 208Pb/204Pb of 38.42; lavas of the entire suite of lavas could be argued as distinct (from each other) Chagres-Bayano Arc samples display a range in throughout PI–PIII, overall the sources can be consid- 206Pb/204Pb of 18.75–19.07 and 208Pb/204Pb of 38.33– ered as relatively similar throughout PI–PIV. Apart from 38.85, which encompasses the composition of the old- the relatively slight increases in 208Pb/204Pb between PI est sample and which are higher than those of the and PIII, Costa Rica magmas remained essentially Golfito and Sona–Azuero arcs. The fact that Chagres– unchanged with respect to Pb isotope composition Bayano Arc samples are typically more radiogenic than between 75 and 6 Ma. Similarly, Panama magmas similarly aged Sona–Azuero and Golfito Arc lavas and recorded only slight increases in 206Pb/204Pb and that some relatively old Chagres–Bayano Arc samples 208Pb/204Pb in PII and PIII relative to PI in Panama, but exhibit relatively very high 206Pb/204Pb and 208Pb/204Pb showed elevated values relative to Costa Rica until the (e.g. a 65.9 Ma sample with 206Pb/204Pb and end of PIII (16–6 Ma). It is not until PIV (6–3 Ma) that the 208Pb/204Pb of 19.07 and 38.83, respectively) suggest source of Costa Rica magmas ‘catch up’ to the more not a temporal evolution with time, but rather a loca- radiogenic Pb isotope compositions of PII and PIII mag- tion-dependent trend at least in Panama, for reasons mas in Panama (Figure 13b) with 206Pb/204Pb of 18.86– that are uncertain; sources to the west were clearly less 19.12 and 208Pb/204Pb of 38.54–38.86, which fall (almost) radiogenic with respect to Pb than sources to the east completely within the range of PII and PIII Panama during PI (Figure 13a)andremainedsoduringPIIand magmas (206Pb/204Pb of 18.79–19.12 and 208Pb/204Pb PIII (see below). of 38.55–38.92). This observation of an apparent lack of radiogenic Several studies have focused on the isotopic nature enrichment of Pb with time in Panama is largely sup- and evolution of post 6 Ma lavas of the CAVAS (e.g. ported by the fact that PII and PIII (36–6 Ma) Panama Feigenson et al. 2004; Goss and Kay 2006; Hoernle et al. samples plot almost entirely within the range of PI 2008;Gazelet al. 2009, 2011). A major finding of these Panama with 206Pb/204Pb of 18.79–19.12 and aforementioned studies was the location- (along the vol- 208Pb/204Pb of 38.55–38.92; however, the broad E–W canic arc) dependent isotopic composition of Quaternary distribution of these samples from the Panama–Costa to present lavas. Although PVI Costa Rica magmas span Rica border in the west to the Panama Canal near the the entire gamut of 206Pb/204Pb and 208Pb/204Pb compo- westernmost border of the Chagres–Bayano Arc in the sitions exhibited by older (PI–PVI) lavas (Figure 13d)and east makes it difficult to ascertain whether the PII and extend to even more radiogenic compositions, their loca- PIII sources were spatially associated with those of the tion in 206Pb/204Pb versus 208Pb/204Pb space is conditional Sona–Azuero Arc or the Chagres–Bayano Arc. Perusal of upon their geographic location along the volcanic arc. Figure 2 of Wegner et al.(2011) suggests that the Whereas Costa Rica lavas from NW Costa Rica exhibit a majority of samples with isotope data (those with cor- wide range of Pb isotopes extending from relatively responding radiometric age data) are from regions to unradiogenic compositions (e.g. with 206Pb/204Pb of the immediate north and west of the Sona and Azuero ~38.26 and 208Pb/204Pb of ~18.63) to moderately radio- peninsulas in the Cordillera de Panama (see also our genic compositions (with 206Pb/204Pb of ~38.84 and Figure 2). Hence, based on this spatial distribution, it 208Pb/204Pb of ~19.10) similar to maximum values exhib- may then be reasonable to assume a temporal shift to ited by older PI (Chagres-Bayano Arc)–PIII Panama mag- Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 more radiogenic Pb, assuming a Sona-Azuero proximal mas and PII–PIV Costa Rica magmas (e.g. compare source for PII and PIII magmas in Panama. Figures 13a and 13b with d), PVI Nicaraguan magmas During PII in Costa Rica, 206Pb/204Pb (18.55–18.69) did are unradiogenic and uniquely range to very depleted not rise from PI values (18.57–18.77), but 208Pb/204Pb compositions akin to N-MORB (Figure S5d). In contrast, (38.19–38.45) ranged to higher concentrations (than PI magmas from NE and Central Costa Rica exhibit much with 208Pb/204Pb 38.20–38.33). Similarly, throughout PIII, higher 206Pb/204Pb and 208Pb/204Pb of ~38.76–39.10 and 206Pb/204Pb (18.60–18.80) changed little from initial PI 19.08–19.29, respectively. Similarly, the PV arc alkalic values, but 208Pb/204Pb (38.20–38.48) continued to rise or basalts and adakites exhibit a similar highly radiogenic at least range to higher values. A dramatic change to nature (with respect to Pb) as lavas from central and NE unequivocally more radiogenic Pb compositions began in Costa Rica (Figure 13d). PIV (6–3 Ma) in Costa Rica, with 206Pb/204Pb and 208Pb/204Pb Feigenson et al.(2004) demonstrated the exception- reaching values (18.86–19.12 and 38.54–38.86 respectively) ally unradiogenic nature of local marine sediments (with similar to (a) Panama during PIII and (b) NW Costa Rica respect to Pb) and concluded that the high radiogenic during PVI (2.6–0 Ma, see below). nature of central Costa Rica magmas did not require a INTERNATIONAL GEOLOGY REVIEW 21

subducted component (addition). Various enrichment 208Pb/204Pb, Figure 13), beginning in PIII and reaching models for central Costa Rica were proffered by a pre-2.6 Ma peak in PIV Costa Rica. Feigenson et al.(2004), but favoured the ones entailing The PVa arc alkalic basalts and PVb adakites con- either the melting of enriched veined mantle or the re- tinue the isotopic shift to lower 143Nd/144Nd (Figures melting of Galapagos Plume-influenced mantle. In con- 14c)andhigher206Pb/204Pb and 208Pb/204Pb and over- trast, more recent models to explain enrichment include lap the fields of Cocos-Coiba Ridges and Seamount forearc erosion (Goss and Kay 2006) and Seamount Province fields. Similarly, the PVI lavas from central Province–subduction interaction (Hoernle et al. 2008; Costa Rica (with 143Nd/144Nd of 0.512918–0.512979) Gazel et al. 2009, 2011). Models are scrutinized in (Figures 14d and 15), which range to the lowest Section 6. 143Nd/144Nd, also overlap the Cocos–Coiba Ridges and Seamount Province fields. Analogous to the situation with Pb isotopes where PVI northwest Costa Rica 5.2. Nd isotopes lavas formed from depleted sources relative to those In contrast to Pb isotopes (206Pb/204Pb and 208Pb/204Pb), of PVI central and northeast Costa Rica by virtue of which are clearly less radiogenic in Golfito and Sona- exhibiting lower 206Pb/204Pb and 208Pb/204Pb, the PVI Azuero compared with units interpreted as CLIP and northwest Costa Rica lavas have more radiogenic Nd the Chagres–Bayano Arc (Figure 13), the 143Nd/144Nd than PVI lavas in central and NE Costa Rica with values of all PI arc segments fall within the range of 143Nd/144Nd of 0.512922–0.513040. Also similar to the units interpreted as CLIP (Figures 14 and 15). PI lavas situation with Pb isotope compositions, which chan- are strikingly similar to those of 98–82 Ma western Costa ged little in Nicaragua between the Miocene and pre- Rica units interpreted as CLIP in terms of 143Nd/144Nd sent, there is little change in this interval in Nd isotope (Figures 14a and 15,); only alkali basalts of the Tortugal compositions (Figure 14d). An exception is three PVI Complex exhibit contrasting (lower) 143Nd/144Nd (2.6–0 Ma) behind the volcanic front alkalic basalts (as (0.512740–0.512798) (Hauff et al. 2000b)(Figure 14a). In described in the Georoc dataset where these samples terms of Nd isotopes, PI arc lavas in the west (i.e. Golfito were listed) in eastern Nicaragua and within the and Sona–Azuero) are similar to those in the east Caribbean Sea (~12.3°N/83.8°W), which plot similarly (Chagres–Bayano) (Figures 14a and 15). In detail, the in the 206Pb/204Pb versus 143Nd/144Nd space as PIII Golfito Complex records significantly higher 143Nd/144Nd Panama and PIV Costa Rica with ~ 18.87–19.90 (0.512922–0.512959) than the Sona–Azuero Arc 206Pb/204Pb and 0.512999–0.515303 143Nd/144Nd. The (0.512864–0.0512866). remaining (NW and SW) Nicaragua samples plot clo- Plots of 206Pb/204Pb versus 143Nd/144Nd and sest to depleted MORB with high 143Nd/144Nd 143Nd/144Nd versus time (Figures 14 and 15) also show (~0.51299–0.51311) and low 206Pb/Pb204 (Figure 14d). a temporal trend to more radiogenic Nd relative to Pb, as magma sources beneath both Costa Rica and 6. Discussion Panama start to differentiate from CLIP-like 143Nd/144Nd during PII by becoming more radiogenic The general evolution of the CAVAS towards more (Figure 15). Whereas the oldest radiometrically dated enriched compositions suggests that one or more of Sona-Azuero and Chagres–Bayano samples (72.3– the several processes became increasingly important 65.9 Ma) (Wegner et al. 2011) exhibit the lowest with time: a greater depth of melting so that garnet Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 143Nd/144Nd and plot along the lower cusp of the com- played a role; an increase in sediment fluid and melt positional field defined by the Galapagos plume at additions; an increase in plume or OIB contributions to 90 Ma, all younger radiometrically dated PI Panama the mantle source; a decrease in the degrees of partial samples from Chagres–Bayano with ages of 65.9– melting in the mantle wedge; changes in the thermal 44.9 Ma display higher 143Nd/144Nd and plot near the structure of the slab or wedge; or a combination of two lower cusp of the Galapagos plume compositional field or more of these processes. A noteworthy feature is that at 60 Ma (Figures 14a and b). Furthermore, PII (35– incompatible trace element enrichment is accompanied 16 Ma) samples from both Panama and Costa Rica by isotopic evolution to more depleted sources (higher record higher 143Nd/144Nd than PI products, but plot 143Nd/144Nd, lower 206Pb/204Pb) with time (Figure 14). within the field of the Galapagos Plume at 60 Ma This anti-correlation may reflect that the asthenospheric (Figure 14b). PIII lavas return to slightly lower mantle flowing into the mantle wedge changed with 143Nd/144Nd than older PII. These data demonstrate a time, from enriched (plume-like) during PI to more switch to less radiogenic Nd, which was accompanied depleted (more MORB-like) in Neogene. Strong along- by more radiogenic Pb (higher 206Pb/204Pb and arc isotopic gradients are found in Quaternary lavas, 22 S. A. WHATTAM AND R. J. STERN

with MORB-like sources for Nicaragua and plume-like Basalt 4 sources for Costa Rica. Basalt t3 CLIP We explain below why and how a dual process mantle t2 source crust enrichment entailing a first-order mechanism of progressively 2 t1 K % wt. O decreasing degrees of partial melting as a result of arc a 0 50 70 Mantle crustal thickening and associated changes in the mantle SiO2 wt. % melting thermal structure possibly coupled with subducted sea- Andesite Rhyolite mounts and participation of more depleted mantle Basalt (t3) represents the one most likely responsible for incompa- K55 FC generates (t2) tio n tible enrichment in the CAVAS. K55 a intermediate & K (t1) felsic melts; no Below, we first discuss how to interpret CAVAS che- 55 fractio n crustal melting motemporal trends. Next, we examine the chemotemporal record to determine whether the roles of crust b 55 interaction and tectonics were important in controlling Andesite Rhyolite Basalt CAVAS magmagenetic evolution. Finally, on the basis of crustal melt

(t3) ng Lower crust melting xi our observations and interpretations of CAVAS chemo- K55 i

m generates felsic (t2) K55 ing melts; mixing temporal evolution, we discuss the larger challenge of generates K (t1) mix determining why some arc magmatic systems evolve 55 intermediate melts

with time whereas others do not. c 55

Figure 16. Illustration of how element concentrations regressed 6.1. How to interpret CAVAS X55 chemotemporal to 55 wt.% SiO2 (X55) can be interpreted using potassium (K55) trends as an example. (a) Mantle source is enriched with time and melts to generate progressively more enriched basalts, from Lavas erupted from long-lived arc systems – like the most depleted (t1) to most enriched (t3). Mantle-generated – CAVAS are likely to reflect more complex magma basalts traverse the crust quickly and do not heat the crust evolution processes than those erupted from the other sufficiently to melt it. No differentiated magmas are produced two major magmagenetic systems of divergent plate and it is easy to infer progressive mantle enrichment. (b) margins (mid-ocean ridges) and hotspots, particularly Progressively enriched mantle output (basalt) experiences frac- those on fast-moving oceanic plates. Such complex- tional crystallization to form andesite and rhyolite. Regression of compositions to K55 allows mantle evolution to be clearly ities include ponding of mantle-derived mafic melts in identified. (c) Progressively enriched mantle output heats the the crust where fractionation to form intermediate and CLIP lower crust, causing it to melt and generate felsic liquids felsic melts can occur. Such ponding will heat the crust (rhyolite). These mix with basaltic magmas to generate ande- and cause melting to generate felsic melts, and the sites. Regression to K55 allows most of the progressive mantle mixing of mafic and felsic melts can generate inter- enrichment to be identified. mediate melts (e.g. Stamatelopoulou-Seymour et al. 1990). Thinner mafic oceanic crust like that of CLIP is melting occurred to generate all CAVAS basalts – differ-

less likely to re-melt than thicker continental crust, but ences in X55 largely reflect the variations in mantle how much crustal melting occurs in each case is enrichment, presumably due to the addition of subduc- unclear; in both cases, there is likely to be more melt- tion components. On the other hand, if crustal melting Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 ing with time as the lower crust is warmed by the to form felsic melts and the resultant hybridization is passage and ponding of magma, but this depends the most important controlling compositional diversity, on magma flux (Annen et al. 2006).Theroleofthe then the compositional spectrum is more difficult to

crustal filter must be kept in mind when interpreting interpret. As magmas with 55 wt.% SiO2 are more akin CAVAS chemotemporal evolution. to the ones recording mantle input (~50 wt.% SiO2)as Mantle processes must also be considered when opposed to crustal input (~70 wt.% SiO2), abundances interpreting CAVAS chemotemporal trends. If the man- normalized to 55 wt.% SiO2 (X55) emphasize mantle tle source of basalt magmas evolves compositionally, input. the resultant complex magmatic evolution can be difficult to interpret, as summarized in Figure 16. If the most important process is fractional crystallization, compari- 6.2. Role of the crust son of X55 allows different mantle influxes of basalt to be compared at similar stages of evolution. In this case Below we use two approaches to evaluate the extent to – and assuming that similar degrees of mantle wedge which CAVAS chemotemporal trends could reflect INTERNATIONAL GEOLOGY REVIEW 23

crustal interactions: trends in K2O and other incompati- (Supplementary Figure S4), some of which had specific ble elements versus SiO2 and Nd isotopic data. magmatic expressions for the CAVAS. Such localized tectonic events and their magmatic responses provide

6.2.1. Incompatible elements versus SiO2 trends few insights into the larger question of overall mag- One perspective on the issue of crustal interaction can matic evolution, but the extent to which these have be gleaned from plots showing best-fit lines through affected magma compositions provides fundamental

incompatible elements versus SiO2 (e.g. K2O vs. SiO2, insights into what are the key controls on arc magma- Figure 4d, see also Supplementary Figure S2). Mixing genesis in the CAVAS, provided it is possible to remove of mantle-derived basalts with crustal melt is likely to this ‘tectonic noise’ from the ‘true’ magmatic signals. It define magmatic trends that converge on the crustal is thus important to identify these so we can disentan- melt, as shown in Figure 16c. CAVAS magmatic trends gle which aspects of the CAVAS record are source evo- do not converge on a likely crustal melt composition, lution signal and which are local tectonic noise. and instead show increasing slopes with enrichment, Three outstanding aspects of CAVAS magmatic evo- more consistent with magmatic differentiation from dif- lution that reflect local tectonics are: (1) the divergence ferent mafic parents. The slope of the PIV trend is the in compositions between Costa Rica plus Panama versus only exception, and it could be that the lower slope for Nicaragua, which began between PIII (16–6 Ma) and this sequence reflects the mixing between the mafic peaked during PIV (6–3 Ma); (2) the production of PV melts of enriched mantle and the crustal melt. adakites and arc alkaline magmas in Panama and Costa A lack of crustal interaction is also readily apparent Rica after 6 Ma; and (3) the ‘resetting’ of the CAVAS

on various incompatible elements versus SiO2 plots source back to more depleted compositions after 3 Ma, when considering the best line of fit through the data which, apart from Nicaragua erupting moderately more (not shown). For example, when considering PI Ba, Rb, mafic lavas than Costa Rica, are otherwise broadly simi- 2 Th, U, and Nb versus SiO2, the low R values (0.07, 0.01, lar in terms of incompatible trace element systematics 0.1, 0.1, and 0.03, respectively) demonstrate that crustal in both regions. These three issues are addressed interaction was not a major factor in magmagenesis. further below.

6.2.2. Nd isotopes 6.3.1. Neogene Nicaragua–Costa Rica compositional The Nd isotopic data also record no evidence of divergence increased crustal participation with time. CAVAS mag- According to our present data set, CAVAS magma che- mas have very CLIP-like compositions during PI, but the mistries in each region appear to have been similar source region evolved to more depleted composition before beginning to diverge during PIII (16–6 Ma) and with time, as shown by higher 143Nd/144Nd (Figure 15). dramatically diverging during PIV (6–3 Ma) (e.g. Figures This isotopic evolution is most easily explained by the 6–12). Whereas Costa Rica PIV lavas continued an participation of more depleted mantle with time, per- enrichment trend relative to PIII compositions, PIV haps by the replacement of the original Galapagos/CLIP Nicaragua lavas reverted to derivation from a more plume-like mantle by a more depleted normal depleted mantle source that was more modified by asthenosphere. slab contributions (e.g. Figures 11 and 12). These dis- We conclude that CAVAS chemotemporal evolution similarities are especially clear for PIV Costa Rica lavas,

does not reflect increased crustal interactions with time, which exhibit strong LREE fractionations (La55/Sm55 and Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 a conclusion also reached by Gazel et al.(2015). Instead, La55/Yb55) in contrast to the only slightly enriched nat- CAVAS chemotemporal variations mostly reflect the ure of PIV Nicaragua lavas (Figure 8, Table 2). Other fractionation of mantle-derived mafic magmas. In this differences apparent on the N-MORB normalized plot case, the approach we outlined in Section 6.1 captures (Figure 10j) are the lower Th (and higher U/Th) and mantle source variations, similar to what is shown in higher Ba/Th, Ba/La, and Nb/La of PIV Nicaraguan lavas Figure 16b. relative to Costa Rican PIV lavas, which partly reflect the greater sediment additions and, in the case of elevated U/Th in Nicaragua lavas, a change in composition of the 6.3. Role of local tectonic events subducted sediments (Plank et al. 2002). It is important to determine the extent to which long- What was the cause of compositional divergence term arc magmatic evolution is affected by regional between Costa Rica and Nicaragua CAVAS beginning tectonic events such as changing slab configurations ~16–6 Ma? There are several factors related to either and upper plate extension. Several major tectonic changes in what the subducted slab was doing (steeper events occurred in and around CAVAS during its lifetime slab dip), what was coming off the subducted slab 24 S. A. WHATTAM AND R. J. STERN

(change in fluid and/or melt compositions due to chan- data suggests that Ba/La has always been lower in Costa

ging sediment composition, the onset of OIB seamount Rica than in Nicaragua (Ba55/La55 of 14–31 in Costa Rica, subduction beneath central Costa Rica, or higher fluid which peaked in PIII, and 61–130 in Nicaragua, which flux), or what the upper plate was doing (changes in peaked in PIV, Table 2). Similarly, Plank et al.(2002) mantle wedge composition and upper plate extension). established that high Ba/La (>70) in the Nicaragua vol- These possibilities are further described next. canic front is a long-lived feature. Irrespective of the With respect to changes in subducted sediment com- cause of lower Ba/La in Costa Rica, it has been operating positions, the U and Ba enrichment in Nicaraguan mag- over the entire CAVAS lifetime, although the discre- mas has been explained as reflecting a ‘carbonate crash’ pancy is the largest during PIV. If Ba/La is indeed related ~10 Ma (event 7, Figure S4) when dominantly carbonate to the subduction and interaction of hotspot tracks, the sedimentation on the Cocos Plate was replaced by long-lived nature of the signal suggests that plume– hemipelagic ooze sedimentation as the result of the subduction interactions began with initial oceanic island shallowing of the carbonate compensation depth by accretion and subduction ~60 Ma (e.g. Hoernle et al. ~800 m; this occurred because the Panama ridge cut 2002; Hoernle and Hauff 2007; Buchs et al. 2009; 2011; off deep water flowing westwards out of the Caribbean see also Gazel et al. 2015) or even earlier (see Whattam (Plank et al. 2002). Subduction erosion of forearc and Stern 2015). igneous units and trench sediments and the addition The inferred higher degrees of partial melting in of changed sediment components to the mantle source Nicaragua versus Costa Rica (Saginor et al. 2013) are (Goss and Kay 2006) have been posited as an alternative ascribed to the steeper slab dip beneath Nicaragua to seamount subduction to explain enrichment. versus Costa Rica (Syracuse et al. 2008). This resulted in However, Hoernle et al.(2008) note that the isotopic slab-derived fluids being released across a smaller width composition of forearc units does not match those of of the arc beneath Nicaragua, which enhanced melting the Seamount Province in contrast to post-6 Ma (Carr et al. 1990). Melting beneath Nicaragua was further enriched lavas in (mostly central) Costa Rica, which are enhanced by extension (see below). Furthermore, the isotopically similar to the Seamount Province. Moreover, slab beneath Nicaragua is more serpentinized (~10–20% we showed in Section 4 that enrichment increased extending some 20–28 km beneath the slab surface) dramatically over a short interval in PIV (6–3 Ma), than that beneath Costa Rica (Syracuse et al. 2008; Van which suggests a rapid, short-lived enrichment process Avendonk et al. 2011; see also Heydolph et al. 2012); that contrasts with that expected from forearc erosion, thus, more fluid might have been released into the which should be modest and prolonged. asthenospheric mantle wedge beneath Nicaragua rela- The subduction of Cocos Plate seamounts and their tive to Costa Rica. Greater fluid flux would cause more metasomatic interaction with the mantle source melting beneath Nicaragua than Costa Rica and would beneath central Costa Rica (event 9, Figure S4) have also deliver different subduction-related metasomatic been proposed to explain the distinctive enrichments components. We do not know when these three differ- recorded by PVa alkaline basalts and PVb adakites (e.g. ences – slab dip, upper plate extension, and subduction Hoernle et al. 2008; Gazel et al. 2009, 2011) as commen- of more serpentinized slab – between Nicaragua and cing about the same time (10–8 Ma) as the arrival of the Costa Rica convergent margins first appeared, but one Cocos Ridge and the carbonate crash. However, inspec- or more of them starting ~16–6 Ma could explain the tion of Figure 13 (206Pb/204Pb vs. 208Pb/204Pb) and observed compositional divergence. Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 Figure 14 (206Pb/204Pb vs. 143Nd/144Nd) shows that Pb A possible cause of compositional divergence and Nd isotopic compositions similar to that of the between Costa Rica and Nicaragua lava compositions subducting Seamount Province occurred as early as was the formation of the Nicaragua depression, which 20 Ma in Panama (blue circle, with age of 19.2 Ma). began during PII at about 23 Ma and continues until Two other PII (16–6 Ma) samples from Panama also today (Funk et al . 2009). Extension thinned the litho- plot within the subducting Seamount Province field in sphere and lengthened the melting columns beneath terms of both Pb isotopes (206Pb/204Pb and 208Pb/204Pb) Nicaragua, allowing for higher degrees of melting, and Nd isotopes (Wegner et al. 2011). If seamount sub- which was further aided by the enhanced water flux duction was important for enrichment, this might have from a more steeply dipping, moderately serpentinized started much earlier than 8–10 Ma. Seamount subduc- slab. Plank and Langmuir (1988) showed how the height tion thus may explain the presence of OIB-like signa- of the mantle column available for melting beneath arcs tures in ~6 Ma and younger central Costa Rica CAVAS is a function of arc crust thickness. If melting begins sequences (Reagan and Gill 1989; Hoernle et al. 2008; beneath arcs at similar depths, then the column of Gazel et al. 2009, 2011, 2015); however, our regressed mantle that undergoes decompression melting is INTERNATIONAL GEOLOGY REVIEW 25

shorter beneath the thicker arc crust. Hence, the longer apparent reversal of magmatic evolution? As La and Nb mantle column for arcs built on rifted crust will lead to have similar, low Kd values (e.g. ~0.05 and 0.005 in higher degrees of partial melting. However, Turner and basalt clinoproxene, GERM data base) and Zr/Nb can Langmuir (2015) concluded that the length of the melt- be fractionated only by extremely low degrees of partial ing column cannot be the determining factor in con- melting (F < 0.5%, Sun and McDonough 1989), the trolling the extent of partial melting and that rather a higher La/Nb of PIV Nicaragua relative to Costa Rica is thicker arc edifice would depress isotherms deeper (see more likely to reflect greater additions of La relative to also Karlstrom et al. 2014). Hence, if the lithosphere Nb from the subducted slab, consistent with other indi- beneath Nicaragua was thinner than that beneath cators as discussed previously. However, the higher Zr/ Costa Rica as a result of forming the Nicaragua depres- Nb and the much lower Nb/Yb of Nicaragua PIV mag- sion, a taller melting column accompanied by steeper mas relative to Costa Rica demonstrate either the higher dipping slab and more water flux and the migration degrees of partial melting or the tapping of a more upwards of hotter isotherms (Turner and Langmuir depleted source or both (Pearce and Peate 1995). 2015) would be expected to lead to higher degrees of We suggest that the higher degrees of partial melt- partial melting beneath Nicaragua relative to Costa Rica. ing in Nicaragua, which began about 20 Ma, coincident with the initial formation of the Nicaragua depression, 6.3.2. Production of PV adakites and arc alkaline was the result of associated lithosphere thinning and magmas in Panama and Costa Rica after 6 Ma changes in the mantle thermal structure. As lithospheric Subduction terminated beneath Costa Rica and Panama thickening would displace colder isotherms deeper soon after 8 Ma as the result of collision and attempted resulting in a lower degree of partial melting, litho- subduction of the Cocos Ridge beneath Central America spheric thinning would have an opposite effect, i.e. (Abratis and Wörner 2001); this was followed soon after the displacement of hotter isotherms upwards resulting by the production of adakites and arc-alkaline basalts. It in a higher degree of partial melting (Turner and has long been recognized that the end of subduction Langmuir 2015). It may be that lithospheric thinning may be marked by the eruption of alkaline basalts (e.g. propagated eastwards to beneath Costa Rica by the Jakeš and White 1969), and sometimes a bimodal arc time the ‘modern-day’ CAVAS began to erupt at alkaline basalt and adakite association accompanies the 2.6 Ma. Alternatively, along-arc, trench-parallel mantle subduction of a hot slab (e.g. Kimura et al. 2014). The flow may be partly responsible for the recent changes in formation of alkaline-adakite bimodal magmatic asso- magma compositions (Hoernle et al. 2008) ciations have been explained by varying percentages of sediments and altered crust comprising the slab, differ- 6.4. Long-term chemotemporal evolution of the ing fractions of slab flux, and differing P–T conditions of CAVAS mantle melting (e.g. Martin et al. 2005). Gazel et al. (2009, 2011) suggest that Costa Rica adakites may There is strong evidence that, despite significant effects form as a result of seamount subduction, but this is due to regional tectonic changes, the CAVAS system not a common explanation for adakites. Nevertheless, evolved chemically over its 75 million year history, in the case of the CAVAS, the isotope evidence provides from depleted to enriched magmas. In this section we strong support for seamount subduction and interac- first consider what processes were responsible, the evo- tion with mantle sources (Gazel et al. 2011). According lution of CAVAS towards continental crust composi- Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 to the model of Gazel et al.(2011) for instance, post- tions, the role of mantle plumes in arc evolution, and 6 Ma alkali basalt and adakite formation in western finally present a synthesis and model for this evolution. Panama and Costa Rica was related to terminal subduction, slab break-off, and subsequent asthenospheric 6.4.1. Causes of progressive CAVAS magma upwelling. In either case, the eruption of adakites and enrichment alkali basalts during CAVAS PV reflects local tectonics In Section 6.2, we showed that increasing crustal inter- related to subduction termination and does not reveal action was probably not responsible for source enrich- much about the overall CAVAS magmatic evolution. ment over time and we conclude that the long-term enrichments reflect systematic changes in magma 6.3.3. ‘Resetting’ of the CAVAS source back to more sources and/or magmagenetic processes with time. depleted compositions after 3 Ma Such systematic changes could include the participation PVI marks the resetting of CAVAS magma compositions of progressively more enriched mantle, progressively to compositions similar to more depleted compositions greater slab contributions, or progressively lower such as those of PI–PIII. What was responsible for this degrees of melting. Our conclusions thus echo those 26 S. A. WHATTAM AND R. J. STERN

of Turner and Langmuir (2015), who concluded that two cumulative processes. In the first case, increased effi- distinct explanations could account for the global varia- cacy of fluid transfer from slab to mantle would be tions in arc magma compositions: different extents of responsible for increasing the fluid-mobile to fluid- mantle wedge melting due to differing (mantle) thermal immobile incompatible element ratios; fluid-mobile structures, or varying contributions from the subducted elements released from the slab would immediately slab. We can track mantle evolution with Zr/Y affect melt compositions, but would have no long- (Figure 11a) and Nb/Yb (Figure 11c), both of which term effect on the mantle wedge composition. In the increase with mantle enrichment, but can also vary as second case, the cumulate effect of fluids released a function of the degree of partial melting (or if garnet from the subducted slab would metasomatically enrich is involved, which except for PV adakites and arc alka- themantlesourceinLILEsovertime.Asweseeno line basalts does not seem to be the case for CAVAS). reason as to why subduction processes should more Zr/Y increases over the CAVAS lifespan, suggesting efficiently release fluid-mobile elements with time, we either that the mantle feeding the early CAVAS was prefer the second explanation that the cumulate effect more depleted than that feeding it now or that the of fluids released from the slab metasomatically degree of partial melting has decreased over time. Nb/ enriched the mantle source over time and conclude Yb ratios diverge strongly, with higher values for Costa that the residence time of circulating asthenosphere in Rica and Panama (indicating mantle source enrichment the mantle wedge is long relative to subduction- or lesser melting) and lower values for Nicaragua (indi- related enrichment processes. We note that the cating mantle source depletion or higher melting). The observed anti-correlation of trace element enrichment trace element evidence of progressive enrichment con- and especially Nd isotopic evidence for the increas- trasts with the isotopic evidence that more depleted ingly depleted mantle source region may reflect that mantle was involved; however, isotopic compositions the asthenospheric mantle source flowing into the reflect time-integrated parent/daughter ratios and are mantle wedge changed with time, from enriched insensitive to recent changes in source composition. We (plume-like) during PI to more depleted (more MORB- conclude from the trace element and isotopic data that like) in Neogene, especially for Nicaragua lavas. the CAVAS mantle source changed significantly over the A change in mantle or slab thermal structure (Turner 75 Ma life of CAVAS, becoming somewhat enriched with and Langmuir 2015) may represent alternative mechan- time, at least for Costa Rica and Panama. isms for varying slab-mantle fluxes over time. In parti- Another possible explanation is that the subduction cular, the driving of cold isotherms to progressively input has increased with time, or that its nature has deeper depths associated with arc edifice thickening changed. Subduction input can be subdivided into could displace mantle melting to higher pressures and those elements that are mobile in hydrous fluids (Rb, lower temperatures and hence lower degrees of melt- Ba,Sr,U,Pb,LREEs)andthosethatareonlytrans- ing. This change in mantle thermal structure could thus ported by melts (Zr, Th, Nb, HREEs). Absolute concen- result in more enriched magmas than those generated trations of linearly regressed fluid-mobile trace beneath a relatively thinner lithosphere. We expand on

elements K2O, Rb, Ba, Sr, U, and Pb increase over the this idea in Section 6.4.4. life of the arc (Figures 6–8), although chemotemporal evolution is more complex when the CAVAS is subdi- 6.4.2. Continental crust formation at the CAVAS vided into regions. Part of the subduction input can be It is generally agreed that continental crust today Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 isolated by focusing on the ratios of fluid-mobile to mostly forms above subduction zones. Despite difficul- fluid-immobile incompatible trace elements, which are ties in resolving the differences between the mafic, generally acknowledged to reflect the transfer of fluid- LREE-depleted nature of igneous rocks generated at mobile elements from the subducted slab to the man- IOCs versus those of more siliceous, LREE-enriched con- tle wedge. The only such ratio that increases with time tinental arcs and their roles in continental crustal gen- in Panama and Costa Rica between PI and PIII is Sr/Nd; esis (Kay 1985; Ellam and Hawkesworth 1988), the U/Th increases between PI and PII only and Pb/Ce accretion of arc-related terranes due to collision is con- remainsmoreorlessconstantthroughtime sidered to have been fundamental in the growth and (Figure 12). Nonetheless, the patterns of some of development of continental crust throughout at least these ratios suggest that part of the explanation for the Phanerozoic time (Taylor and McLennan 1985; the long-term source enrichment of CAVAS magmas Rudnick 1995; Rudnick and Fountain 1995). might reflect subduction-related metasomatism. Based on ~0.3–6 Ma silicic ignimbrites from the Costa Subduction-related progressive enrichment of the Rica volcanic front, Vogel et al.(2004) suggest that CAVAS source could reflect either instantaneous or continental crust formed there as the result of the INTERNATIONAL GEOLOGY REVIEW 27

addition of silicic magmas to the subduction-modified, enriched arcs in general, would be further intensified oceanic plateau crust, possibly aided by the foundering by the lack of backarc formation behind the arc, a of mafic/ultramafic residues and cumulates from the feature of which characterizes, for example, the base of the crust. Vogel et al.(2004) suggested that enriched Greater Antilles Arc (GAA) system. Enrichment the origin of these silicic magmas was the result of at the well-documented GAA (Jolly et al. 1998a, 1998b, reprocessing initially juvenile, mantle-derived, subduc- 2001) was not sudden but gradual over tens of millions tion-related magmas that ponded in the crust similar to of years, similar to that of the CAVAS. It seems more the process proposed for converting basaltic crust to reasonable to us that an overarching cause of enrich- continental crust in Costa Rica by Pichler and Weyl ment would be common in different arc systems as (1975). Although this might have been important late opposed to unique in each arc. We conclude that gra- in the evolution of the CAVAS in Costa Rica, this does dual enrichment with time demonstrated by the CAVAS not explain the general enrichment in the CAVAS that is the result of gradual thickening of the arc substrate began during PIII (35–16 Ma). As shown in Figure 10, and a change in the wedge thermal structure, which continental crust-like compositions were achieved in allowed for the driving of isotherms downwards to the CAVAS by at least 6 Ma. Furthermore, the transfor- higher pressures and colder temperatures and hence mation from mafic to continental-like compositions lesser degrees of partial melting (Turner and Langmuir started much earlier (by ~35 Ma) and was gradual. 2015). Gazel et al.(2015) concluded that partial melting of enriched subducting Galapagos hotspot tracks produced young andesitic continental crust beginning at 6.4.3. Role of mantle plumes in the generation of ~ 10 Ma as recorded in the Central American Land continental crust Bridge (CALB, Panama and Costa Rica) geochemical The role of enriched mantle sources in the generation of evolution. As shown in Figure 9, collective compositions continental crust has long been proposed (e.g. of Costa Rica plus Nicaragua are nearly identical to the Hawkesworth and Kemp 2006 and references therein). continental crust at 6–3 Ma, although when parsed into Primarily via isotopic evidence, we showed the contri- region, it is evident that whereas Costa Rica was clearly butions of the Galapagos Plume to CAVAS magmatism, continental-like, Nicaragua was not during this interval which has been noted by others (e.g. Wegner et al. (Figure 10). Thus, our data mostly agrees with those of 2011; for Panama, Whattam and Stern 2015; references Gazel et al.(2015) in the timing of when continental-like therein for various regions around the periphery and compositions appeared, but differs in (1) the rate at centre of the Caribbean Plate). A model of plume- which this continental crust composition was reached induced subduction initiation (PISI) around the CLIP, and (2) the cause of enrichment. We stress that conti- the plume head of the Galapagos Plume, to initiate nental-like compositions did not suddenly appear at the CAVAS (Whattam and Stern 2015; see also Gerya 10 Ma, but rather that the enrichment was gradual et al. 2015.) explains its isotopic and trace element over CAVAS history from PII (35–16 Ma) to the present similarity during PI with the Galapagos Plume and OPB in Panama and Costa Rica, as shown in Section 4 and in general; however, it is evident that plume-like mantle summarized in Figure 9. As Galapagos hotspot tracks contributions occurred all throughout the lifespan of appear to have been only subducted beneath Costa the CAVAS and continue to this day (see Whattam and Rica beginning about 10 Ma (Gazel et al. 2015 and Stern 2015 and references therein). Although emplace- Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 references therein), enrichment during PII cannot only ment of the CLIP oceanic plateau and continual feeding be the result of hotspot track subduction and the asso- of plume magmatism would have provided a head-start ciated mantle metasomatism. on enrichment via the production of an anomalously Moreover, the if the tenets of the model of Gazel thick arc substrate, enrichment cannot be solely attrib- et al.(2015) represent an overarching process necessary uted to plume contributions. The reason for this is that for enrichment in other enriched arcs (e.g. the whereas the isotopic affinities of CAVAS products kept Philippines), this then requires the serendipitous addi- pace with that of the Galapagos Plume, the trace ele- tion and interaction of hotspot-tracks (at other enriched ment chemistry subsequently changed to more arc systems). Alternatively, with our model, all that is enriched compositions (than the Galapagos Plume, i.e. required is the thickening of the arc edifice and a the CLIP) after 35 Ma. This also rules out enrichment as change in the associated sub-arc mantle wedge thermal the result of OIB subduction accompanying accretion structure. Enrichment at the CAVAS, and perhaps after 60 Ma. 28 S. A. WHATTAM AND R. J. STERN

6.4.4. Chemotemporal evolution of the CAVAS: (2) Elevated concentrations of some incompatible synthesis and model elements and element ratios, however, are higher Based on the results and interpretations of this study, than those of the IBM and similar in many cases we provide a graphical representation of the chemo- to mean OPB. This and the fact that early and temporal evolution of the CAVAS from arc establish- later magmas kept pace with the isotopic evolu- ment to the present day in Figure 17.PIat75–39 Ma tion of the Galapagos Plumes demonstrate sig- marked the establishment of the CAVAS magmatic arc nificant plume–subduction interactions that and the production of arc tholeiites. Subsequent occurred throughout the CAVAS lifespan. thickening of the sub-arc lithosphere during PII (35– (3) The most striking feature of CAVAS geochem- 16 Ma) and PIII (16–6 Ma) drove the isotherms deeper, ical evolution was the progression in incompa- with the consequent displacement of melting to tible-element enrichment with time after higherpressuresandlowertemperaturestoform 35 Ma. The composition of mean, linearly increasingly enriched magmas. This process continued regressed PII (35–16 Ma) lavas and intrusives in Costa Rica during PIV (6–3 Ma); however, extension of both Panama and Costa Rica and the collec- of the sub-arc lithosphere beneath Nicaragua begin- tive composition of PIV (6–3Ma,CostaRicaplus ning in the early Miocene resulted instead in the Nicaragua) magmas closely resemble that of upward displacement of hotter isotherms beneath BCC. The best explanation for CAVAS enrich- Nicaragua. This, in addition to the more steeply dip- ment is decreasing degrees of partial melting ping and serpentinized nature of the slab beneath with time as the result of crustal thickening and Nicaragua compared with Costa Rica, gave rise to changes to the sub-arc mantle thermal struc- more melting and the production of more depleted ture (progressive downwards displacement of magmas. Subsequent to the termination of subduc- isotherms) accompanied by cumulative metaso- tion beneath Panama and southeast Costa Rica during matic enrichment of the mantle wedge above Phase V (5.9–0.02 Ma), subsequent slab melting and the subducting Cocos Plate. partial melting of upwelling asthenosphere resulted in (4) A fundamental compositional divergence is the production of adakites and arc alkalic basalts, recorded in PVI Nicaraguan and Costa Rican mag- respectively. Enrichment after 6 Ma in central and mas manifest in elevated U/Th and Ba/La and a northeast Costa Rica was likely accentuated by sea- much more depleted source beneath Nicaragua, mount–subduction interaction as proposed by others. which is likely the result of higher degrees of However, we propose that continued extension, mantle melting. The elevated U/Th has been which first began beneath Nicaragua circa 20 Ma ascribed to the post-10 Ma carbonate crash, with the initial formation of the Nicaragua depression, which changed the subducting Cocos Plate sedi- propagated eastwards to beneath (both Nicaragua ments from U-poor carbonates to U-rich hemipe- and)CostaRicaduringPVI(2.6–0Ma(Figure 17). lagic muds, whereas the lower Ba/La in Costa Rica This resulted in attenuation of the sub-arc lithosphere, has been interpreted as the result of OIB sea- resulting in the driving of hotter isotherms upwards, mount–subduction interactions. The higher more melting, and the formation of depleted magmas degree of partial melting beneath Nicaragua similar in composition to those of PI–PIII. may be the result of steeper slab dip and upper plate extension, resulting in changes to the sub- Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 arc mantle thermal regime (displacement of the 7. Conclusions isotherms upwards) accompanying the formation Seven important conclusions about the geochemical of the Nicaragua depression, the combined evolution of the CAVAS are drawn from this study: effects of which resulted in higher degrees of melting. The higher degree of melting is also (1) Early CAVAS volcanism (PI, 75–39 Ma) was char- likely related to the more serpentinized nature acterized by low-K, tholeiitic to weakly calc-alka- of the Nicaraguan slab. line activity with compositions that are broadly (5) Chemotemporal enrichments of CAVAS and other similar to the depleted, sediment-poor lavas of arcs partly reflect whether or not an arc system is the IBM arc system. There was an apparent W–E associated with a backarc basin. Arcs associated and possibly older to younger transition from with backarc basins are fed by a more depleted MORB-like magmatism in the 73–39 Ma Sona– mantle and have taller melt columns; unrifted arc Azuero Arc to the 70–39 Ma Chagres–Bayano Arc. lithosphere in contrast thickens with time, INTERNATIONAL GEOLOGY REVIEW 29

Phase I (75-39 Ma) Farallon Plate development of magmatic arc; contributions from Galapagos Plume result in construction of anomalously thick arc substrate; eruption of arc tholeiites

Phase II (35-16 Ma) thickening of lithosphere beneath magmatic arc results in shorter melt column - displacement of isotherms downwards - less melting - more enriched magmas; extension beneath Nicaragua (not shown) begins around 23 Ma and results in extension of melt column - displacement of hot isotherms upwards - more melting beneath Nicaragua relative to Costa Rica beginning in PIII

Seamount Province Phase III (16-6 Ma) continued thickening of lithosphere beneath Costa Rica results in even shorter melt column - continual displacement of isotherms downwards - even less melting and even more enriched magmas; continued extension beneath Nicaragua however, results in incipient divergence in source compositions beneath

ions ions Nicaragua vs. Costa Rica to more depleted and enriched beneath the former and latter, respectively

Phase IV (6-3 Ma): Costa Rica continued thickening of lithosphere beneath magmatic arc results in even shorter melt column - continual displacement of isotherms even further downwards -

overt lifespan of CAVAS of lifespan overt even less melting - even more enriched magmas; subduction of Seamount Province seamounts accentuates enrichment subducting plate getting younger; getting plate subducting contribut Plume Galapagos continue Phase IV (6-3 Ma): Nicaragua continued extension thins lithosphere, lengthens melt column - displacement of isotherms upwards - more melting - more depleted magmas

Phase V (5.9-0.02 Ma) termination of subduction beneath Panama and SE Costa Rica leads to slab melting (adakite formation) alkalic basalts and alkalic basalt formation adakites asthenosphere slab tear flow

Cocos Plate Phase VI (2.6-0 Ma) continued extension beneath Nicaragua and NW Costa Rica lengthens melt column - displacement of isotherms upwards - more melting - more depleted magmas Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015

Figure 17. Graphical representation of the chemotemporal evolution of the CAVAS between magmatic arc establishment and the present day based on Turner and Langmuir (2015) and Plank and Langmuir (1988) (see text for details). Note that the green lines beneath the CAVAS arc edifice in Costa Rica and Panama between PII (35–16 Ma) and PIV (6–3 Ma) represent gradual lithospheric thickening, which resulted in the progressive displacement of isotherms downwards, lower degrees of partial melting, and the generation of increasing enriched magmas. An opposite mechanism of lithospheric attenuation/extension occurred beneath Nicaragua beginning ~23 Ma, concomitant with the commencement of Nicaragua depression formation resulting in the displacement of isotherms upwards, higher degrees of melting, and the production of more depleted magmas. By 3 Ma, extension beneath Nicaragua had propagated to the SW beneath Costa Rica, resulting in the resetting to the production of more depleted magma compositions similar to those produced between PI and PIII.

resulting in progressively lower degrees of melt- arc magmatic evolution. Early Galapagos Plume ing and more enriched lava compositions. contributions in forming the CAVAS mantle Opening of a backarc basin essentially reverses wedge kick-started CAVAS enrichment. 30 S. A. WHATTAM AND R. J. STERN

(6) Models calling for the chemical evolution of mag- Brown, G.C., 1982, Calc-alkaline intrusive rocks: Their diversity, matic arcs from depleted to enriched with time evolution and relation to volcanic arcs, in Thorpe, R.S., ed., (i.e. early arc evolution studies) are generally sup- Orogenic andesites and related rocks: London, Wiley, p. 437–461. ported by the study of the CAVAS system. Bryant, C.J., Arculus, R.J., and Eggins, S.M., 2003, The geochem- Tectonic events can complicate simple models ical evolution of the Izu-Bonin Arc System: A perspective for arc evolution, and even the reversal of long- from tephras recovered by deep-sea drilling: Geochemistry, term chemotemporal enrichment trends. Geophysics, Geosystems, v. 4, doi:10.1029/2002GC000427 (7) Future studies are needed to improve coarse Buchs, D.M., Arculus, R.J., Baumgartner, P.O., Baumgartner- Mora, C., and Ulianov, A., 2010, Late Cretaceous arc devel- temporal resolution in especially PI (75–39 Ma), opment on the SW margin of the Caribbean Plate: Insights – – PII (35 16 Ma), and PIII (16 6 Ma). More inte- from the Golfito, Costa Rica, and Azuero, Panama, com- grated geochronologic studies of especially plexes: Geochemistry, Geophysics, Geosystems, v. 11, CAVAS PI, II, and III are required. doi:10.1029/2009GC002901 Buchs, D.M., Arculus, R.J., Baumgartner, P.O., and Ulianov, A., Acknowledgements 2011, Oceanic intraplate volcanoes exposed: Example from seamounts accreted in Panama: Geology, v. 39, p. 335–338. We greatly appreciate the detailed and constructive com- doi:10.1130/G31703.1 ments of four anonymous reviewers, which led to a greatly Buchs, D.M., Baumgartner, P.O., Baumgartner-Mora, C., revised – and we believe improved – manuscript. Bandini, A.N., Jackett, S.-J., Diserens, M.-O., and Stucki, J., 2009, Late Cretaceous to Miocene seamount accretion and mélange formation in the Osa and Burica Peninsulas (south- Disclosure statement ern Costa Rica): Episodic growth of a convergent margin, in James, K., et al., eds., The origin and evolution of the No potential conflict of interest was reported by the authors. Caribbean Plate: Geological Society of London Special Publication 328, p. 411–456. Carr, M.J., Feigenson, M.D., and Bennett, E.A., 1990, Funding Incompatible element and isotopic evidence for tectonic control of source mixing and melt extraction along the SAW acknowledges financial support from the Smithsonian Central American arc: Contributions to Mineralogy and (Washington) and the Smithsonian Tropical Research Petrology, v. 105, p. 369–380. doi:10.1007/BF00286825 Institute (Panama) during his 1 year tenure at the latter during Carr, M.J., Feigenson, M.D., Patino, L.C., and Walker, J.A., 2003, the period 2009–2010. Volcanism and geochemistry in Central America: Progress and problems, in Eiler, J., ed., Inside the Subduction Factory: Geophysical Monograph Series 138, p. 153–174. References Defant, M.J., Clark, L.F., Stewart, R.H., Drummond, M.S., De Boer, J.Z., Maury, R.C., Bellon, H., Jackson, T.E., and Abratis, M., and Wörner, G., 2001, Ridge collision, slab-window Restrepo, J.F., 1991a, Andesite and dacite genesis via con- formation, and the flux of Pacific asthenosphere into the trasting processes: The geology and geochemistry of El – Caribbean realm: Geology, v. 29, p. 127 130. doi:10.1130/ Valle Volcano, Panama: Contributions to Mineralogy and 0091-7613(2001)029<0127:RCSWFA>2.0.CO;2 Petrology, v. 106, p. 309–324. doi:10.1007/BF00324560 Annen, C., Blundy, J.D., and Sparks, R.S.J., 2006, Genesis of Defant, M.J., Jackson, T.E., Drummond, M.S., De Boer, J.Z., intermediate and Silicic Magmas in deep crustal hot Bellon, H., Feigenson, M.D., Maury, R.C., and Stewart, R.H., – zones: Journal of Petrology, v. 47, p. 505 539. doi:10.1093/ 1992, The geochemistry of young volcanism throughout petrology/egi084 western Panama and southeastern Costa Rica: An overview: Arculus, R.J., 1994, Aspects of magma genesis in arcs: Lithos, v. Journal of the Geological Society of London, v. 149, p. 569– Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 – 33, p. 189 208. doi:10.1016/0024-4937(94)90060-4 579. doi:10.1144/gsjgs.149.4.0569 Arculus, R.J., 2003, Use and abuse of the terms calcalkaline and Defant, M.J., Richerson, P.M., De Boer, J.Z., Stewart, R.H., Maury, – calcalkalic: Journal of Petrology, v. 44, p. 929 935. R.C., Bellon, H., Drummond, M.S., Feigenson, M.D., and doi:10.1093/petrology/44.5.929 Jackson, T.E., 1991b, Dacite Genesis via both Slab Melting Arculus, R.J., and Johnson, R.W., 1978, Criticism of generalised and Differentiation: Petrogenesis of La Yeguada Volcanic models for the magmatic evolution of arc-trench systems: Complex, Panama: Journal of Petrology, v. 32, p. 1101– – Earth and Planetary Science Letters, v. 39, p. 118 126. 1142. doi:10.1093/petrology/32.6.1101 doi:10.1016/0012-821X(78)90148-6 Dickinson, W.R., and Hatherton, T., 1967, Andesitic volcanism Barth, M.G., McDonough, W.F., and Rudnick, R.L., 2000, and seismicity around the Pacific: Science, v. 157, p. 801– Tracking the budget of Nb and Ta in the continental crust: 803. doi:10.1126/science.157.3790.801 – Chemical Geology, v. 165, p. 197 213. doi:10.1016/S0009- Drummond, M.S., Bordelon, M., De Boer, J.Z., Defant, M.J., 2541(99)00173-4 Bellon, H., and Feigenson, M.D., 1995, Igneous petrogenesis Baumgartner, P.O., Flores, K., Bandini, A.N., Girault, F., and Cruz, and tectonic setting of plutonic and volcanic rocks of the D., 2008, Upper Triassic to Cretaceous radiolaria from Cordillera de Talamanca, Costa Rica-Panama, Central – Nicaragua and northern Costa Rica The Mesquito compo- American arc: American Journal of Science, v. 295, p. 875– – site oceanic terrane: Ophioliti, v. 33, p. 1 19. 919. doi:10.2475/ajs.295.7.875 INTERNATIONAL GEOLOGY REVIEW 31

Du Bray, E.A., and John, D.A., 2011, Petrologic, tectonic and Goss, A.R., and Kay, S.M., 2006, Steep REE patterns and metallogenic evolution of the Ancestral Cascades magmatic enriched Pb isotopes in Central American arc magmas: arc, Washington, Oregon and northern California: Evidence for forearc erosion?: Geochemistry, Geophysics, Geosphere, v. 7, p. 1102–1133. doi:10.1130/GES00669.1 Geosystems, v. 7, no. 5, Q05016. doi:10.1029/2005GC001163 Ehrenborg, J., 1996, A new stratigraphy for the Tertiary volca- Hauff, F., Hoernle, K., Tilton, G., Graham, D.W., and Kerr, A.C., nic rocks of the Nicaraguan highland: Geological Society of 2000a, Large volume recycling of oceanic lithosphere over America Bulletin, v. 108, p. 830–842. doi:10.1130/0016-7606 short time scales: Geochemical constraints from the (1996)108<0830:ANSFTT>2.3.CO;2 Caribbean large Igneous Province: Earth and Planetary Ellam, R.M., and Hawkesworth, C.J., 1988, Is average continental Science Letters, v. 174, p. 247–263. doi:10.1016/S0012- crust generated at subduction zones?: Geology, v. 16, p. 314– 821X(99)00272-1 317. doi:10.1130/0091-7613(1988)016<0314:IACCGA>2.3.CO;2 Hauff, F., Hoernle, K., Van Den Bogaard, P., Alvarado, G., and Elliot, T., Plank, T., Zindler, A., White, W., and Bourdon, B., 1997, Garbe-Schönberg, D., 2000b, Age and geochemistry of Element transport from slab to volcanic front at the Mariana basaltic complexes in western Costa Rica: Contributions to arc: Journal of Geophysical Research, v. 102, p. 14991– the geotectonic evolution of Central America: 15019. doi:10.1029/97JB00788 Geochemistry, Geophysics, Geosystems, v. 1, Paper no. Elming, S.-A., Layer, P., and Ubieta, K., 2001, A palaeomagnetic 1999GC000020. doi:10.1029/1999GC000020 study and age determinations of Tertiary rocks in Hawkesworth, C.J., Gallagher, K., Hergt, J.M., and McDermott, Nicaragua, Central America: Geophysical Journal F., 1993a, Mantle and slab contributions in arc-magmas: International, v. 147, p. 294–309. doi:10.1046/j.0956- Annual Review of Earth and Planetary Sciences, v. 21, p. 540x.2001.01526.x 175–204. doi:10.1146/annurev.ea.21.050193.001135 Farris, D.W., Jaramillo, C., Bayona, G., Restrepo-Moreno, S.A., Hawkesworth, C.J., Gallagher, K., Hergt, J.M., and McDermott, Montes, C., Cardona, A., Mora, A., Speakman, R.J., Glascock, F., 1993b, Trace element fractionation processes in the M.D., and Valencia, V., 2011, Fracturing of the Panamanian generation of island arc basalts: Philosophical Transactions isthmus during initial collision with South America: of the Royal Society A: Mathematical, Physical and Geology, v. 39, p. 1007–1010. doi:10.1130/G32237.1 Engineering Sciences, v. 342, p. 179–191. doi:10.1098/ Feigenson, M.D., Carr, M.J., Maharaj, S.V., Julianao, S., and rsta.1993.0013 Bolge, L.L., 2004, Lead isotope composition of central Hawkesworth, C.J., and Kemp, A.I.S., 2006, Evolution of the American volcanoes: Influence of the Galapagos plume: continental crust: Nature, v. 443, p. 811–817, doi:10.1038/ Geochemistry, Geophysics, Geosystems, v. 5, no. 6, nature05191 Q06001. doi:10.1029/2003GC000621 Heydolph, K., Hoernle, K., Hauff, F., Bogaard, P.V.D., Portnyagin, Funk, J., Mann, P., McIntosh, K., and Stephens, J., 2009, M., Bindeman, I., and Garbe-Schönberg, D., 2012, Along and Cenozoic tectonics of the Nicaraguan depression, across arc geochemical variations in NW Central America: Nicaragua, and Median trough, El Salvador, based on seis- Evidence for involvement of lithospheric pyroxenite: mic-reflection profiling and remote-sensing data: Geochimica Et Cosmochimica Acta, v. 84, p. 459–491. Geological Society of America Bulletin, v. 121, p. 1491– doi:10.1016/j.gca.2012.01.035 1521. doi:10.1130/B26428.1 Hildreth, W.E., and Moorbath, S., 1988, Crustal contributions to Gazel, E., Alvarado, G.E., Obando, J., and Alfaro, A., 2005, arc magmatism in the Andes of central Chile: Contributions Evolución magmática del arco de Sarapiquí, Costa Rica: to Mineralogy and Petrology, v. 98, p. 455–489. doi:10.1007/ Reviews of the Geology of Central America, v. 32, p. 13–31. BF00372365 Gazel, E., Carr, M.J., Hoernle, K., Feigenson, M.D., Hauff, F., Hoernle, K., Abt, D.L., Fischer, K.M., Nichols, H., Hauff, F., Abers, Szymanski, D., and Van Den Bogaard, P., 2009, The G.A., Van Den Bogaard, P., Heydolph, K., Alvarado, G., Protti, Galapagos-OIB signature in southern Central America: M., and Strauch, W., 2008, Arc-parallel flow in the mantle Mantle re-fertilization by arc-hotspot interaction: wedge beneath Costa Rica and Nicaragua: Nature, v. 451, p. Geochemistry, Geophysics, Geosystems, Q02S11, 1094–1097. doi:10.1038/nature06550 doi:10.1029/2008GC002246 Hoernle, K., and Hauff, F., 2007, Oceanic igneous complexes, in Gazel, E., Hayes, J.L., Hoernle, K., Kelemen, P., Everson, E., Bundschuh, J., and Alvarado, G., eds., Central America, geol- Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 Holbrook, W.S., Hauff, F., Van Den Bogaard, P., Vance, E.A., ogy, resources, hazards: Volume 1: London, Taylor & Francis, Chu, S., Calvert, A.J., Carr, M.J., and Yogodzinski, G.M., 2015, p. 523–548. Continental Crust generated in oceanic arcs: Nature Hoernle, K., Van Den Bogaard, P., Werner, R., Lissinna, B., Hauff, Geosciences, v. 8, p. 321–327. doi:10.1038/ngeo2392 F., Alvarado, G., and Garbe-Schönberg, D., 2002, Missing Gazel, E., Hoernle, K., Carr, M.J., Herzberg, C., Saginor, I., Van history (16–71 Ma) of the Galápagos hotspot: Implications Den Bogaard, P., Hauff, F., Feigenson, M., and Swisher, C., III, for the tectonic and biological evolution of the Americas: 2011, Plume-subduction interaction in southern Central Geology, v. 30, p. 795–798. doi:10.1130/0091-7613(2002) America: Mantle upwelling and slab melting: Lithos, v. 030<0795:MHMOTG>2.0.CO;2 121, p. 117–134. doi:10.1016/j.lithos.2010.10.008 Hoernle, K.A., Werner, R., Phipps Morgan, J., Garbe-Schönberg, Gerya, T., Stern, R.J., Baes, M., Sobolev, S., and Whattam, S.A., D., Bryce, J., and Mrazek, J., 2000, Existence of 2015, Plume-induced subduction initiation triggered plate complex spatial zonation in the Galápagos plume: tectonics on Earth: Nature (in press). Geology, v. 28, p. 435–438. doi:10.1130/0091-7613(2000) Gill, J.B., 1970, Geochemistry of Viti Levu, Fiji, and its evolution 28<435:EOCSZI>2.0.CO;2 as an island arc: Contributions to Mineralogy and Petrology, Irvine, T.N., and Baragar, W.R.A., 1971, A Guide to the Chemical v. 27, p. 179–203. doi:10.1007/BF00385777 Classification of the Common Volcanic Rocks: Canadian 32 S. A. WHATTAM AND R. J. STERN

Journal of Earth Sciences, v. 8, p. 523–548. doi:10.1139/e71- Geochemistry, Geophysics, Geosystems, v. 15, p. 691–739. 055 doi:10.1002/2013GC005132 Ishikawa, T., and Tera, F., 1997, Source, composition and dis- Kuno, H., 1966, Lateral variation of basalt magma type across tribution of the fluid in the Kurile mantle wedge: continental margins and island arcs: Bulletin Constraints from across- arc variations of B/Nb and B iso- Volcanologique, v. 29, p. 195–222. doi:10.1007/BF02597153 topes: Earth and Planetary Science Letters, v. 152, p. 123– Le Bas, M.J., Le Maitre, R.W., Streckeisen, A., and Zanettin, B., 138. doi:10.1016/S0012-821X(97)00144-1 1986, A chemical classification of volcanic rocks based on Jakeš, P., and White, A.J.R., 1969, Structure of the Melanesian the total alkali-Silica diagram: Journal of Petrology, v. 27, p. arcs and correlation with distribution of magma types: 745–750. doi:10.1093/petrology/27.3.745 Tectonophysics, v. 8, p. 223–236. doi:10.1016/0040-1951 Lee, J.M., Stern, R.J., and Bloomer, S.H., 1995, Forty million (69)90099-7 years of magmatic evolution in the Mariana arc: The tephra Jakeš, P., and White, A.J.R., 1972, Major and trace element glass record: Journal of Geophysical Research, v. 100, no. B9, abundances in volcanic rocks of orogenic areas: Geological p. 17671–17687. doi:10.1029/95JB01685 Society of America Bulletin, v. 83, p. 29–40. doi:10.1130/ Lissinna, B.A., 2005, Profile though the Central American 0016-7606(1972)83[29:MATEAI]2.0.CO;2 Landbridge in western Panama: 115 Ma interplay between Jakeš, P., and Gill, J.B., 1970, Rare earth elements and the the Galápagos Hotspot and the Central American island arc tholeiitic series: Earth and Planetary Science Subduction Zone [Ph.D. thesis]: Kiel, Germany, Christian- Letters, v. 9, p. 17–28. doi:10.1016/0012-821X(70)90018-X Albrechts University, 102 pp. Jolly, W.T., Lidiak, E.G., Dickin, A.P., and Wu, T.-W., 1998b, MacMillan, I., Gans, P.B., and Alvarado, G., 2004, Middle Geochemical diversity of Mesozoic island arc tectonic Miocene to present plate tectonic history of the southern blocks, eastern Puerto Rico, in Lidiak, E.G., and Larue, D.K., Central American volcanic arc: Tectonophysics, v. 392, p. eds., Tectonics and geochemistry of the Northeast 325–348. doi:10.1016/j.tecto.2004.04.014 Caribbean: Volume 322: Geological Society of America, Mann, P., Rogers, R., and Gahagan, L., 2007, Overview of plate Special Paper p. 67–98. tectonic history and its unresolved tectonic problems, in Jolly, W.T., Lidiak, E.G., Dickin, A.P., and Wu, T.-W., 2001, Secular Bundschuh, J., and Alvarado, G., eds., Central America: Geochemistry of Central Puerto Rican Island Arc Lavas: Geology, Resources, and Hazards Volume 1: Leiden, The Constraints on Mesozoic Tectonism in the Eastern Greater Netherlands, Taylor and Francis/Balkema, p. 201–237. Antilles: Journal of Petrology, v. 42, p. 2197–2214. Martin, H., Smithies, R.H., Rapp, R., Moyen, J.-F., and Champion, doi:10.1093/petrology/42.12.2197 D., 2005, An overview of adakite, tonalite–trondhjemite– Jolly, W.T., Lidiak, E.G., Schelleckens, H.S., and Santos, S., 1998a, granodiorite (TTG), and sanukitoid: Relationships and Volcanism, tectonics and stratigraphic relations in Puerto some implications for crustal evolution: Lithos, v. 79, p. 1– Rico, in Lidiak, E.G., and Larue, D.K., eds., Tectonics and 24. doi:10.1016/j.lithos.2004.04.048 geochemistry of the Northeast Caribbean: Volume 322: McCulloch, M.T., and Gamble, J.A., 1991, Geochemical and Geological Society of America, Special Paper p. 1–34. geodynamical constraints on subduction zone magmatism: Jordan, E.K., Lieu, W., Stern, R.J., Carr, M.J., Gill., J.B., and Earth and Planetary Science Letters, v. 102, p. 358–374. Lehnert, K., 2012, Geochemical data library for Quaternary doi:10.1016/0012-821X(91)90029-H lavas from the magmatic front of the Central American and Meschede, M., and Frisch, W., 1998, A plate-tectonic model for Izu-Bonin-Mariana arcs: Explanation to accompany excel the Mesozoic and early Cenozoic history of the Caribbean workbook “CentAm and IBM geochem database”, version Plate: Tectonophysics, v. 296, p. 269–291. doi:10.1016/ 1.02, located at: http://www.earthchem.org/grl. S0040-1951(98)00157-7 Karlstrom, L., Lee, C.-T.A., and Manga, M., 2014, The role of Miyashiro, A., 1974, Volcanic rock series in island arcs and magmatically driven lithospheric thickening on arc front active continental margins: American Journal of Science, v. migration: Geochemistry, Geophysics, Geosystems, v. 15, p. 274, p. 321–355. doi:10.2475/ajs.274.4.321 2655–2675. doi:10.1002/2014GC005355 Montes, C., Bayona, G.A., Cardona, A., Buchs, D.M., Silva, C.A., Kay, R.W., 1985, Island arc processes relevant to crustal and Morón, S.E., Hoyos, N., Ramírez, D.A., Jaramillo, C.A., and mantle evolution: Tectonophysics, v. 112, p. 1–15. Valencia, V., 2012a, Arc-continent collision and orocline for- Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 doi:10.1016/0040-1951(85)90169-6 mation: Closing of the Central American seaway: Journal of Kerr, A.C., Marriner, G.F., Tarney, J., Nivia, A., Saunders, A.D., Geophysical Research, v. 117, p. B04105. doi:10.1029/ Thirlwall, M.F., and Sinton, C.W., 1997, Cretaceous basaltic 2011JB008959 terranes in Western Columbia: Elemental, chronological and Montes, C., Cardona, A., McFadden, R., Moron, S.E., Silva, C.A., Sr-Nd isotopic constraints on Petrogenesis: Journal of Restrepo-Moreno, S., Ramirez, D.A., Hoyos, N., Wilson, J., Petrology, v. 38, p. 677–702. doi:10.1093/petroj/38.6.677 Farris, D., Bayona, G.A., Jaramillo, C.A., Valencia, V., Bryan, Kerr, A.C., Tarney, J., Kempton, P.D., Spadea, P., Nivia, A., J., and Flores, J.A., 2012b, Evidence for middle Eocene and Marriner, G.F., and Duncan, R.A., 2002, Pervasive mantle younger land emergence in central Panama: Implications plume head heterogeneity: Evidence from the late for Isthmus closure: Geological Society of America Bulletin, Cretaceous Caribbean-Colombian oceanic plateau: Journal v. 124, p. 780–799. doi:10.1130/B30528.1 of Geophysical Research, v. 107, doi:10.1029/2001JB000790 Nakamura, N., 1974, Determination of REE, Ba, Fe, Mg, Na and Kimura, J.I., Gill, J.B., Kunikiyo, T., Osaka, I., Shimoshioiri, Y., K in carbonaceous and ordinary chondrites: Geochimica Et Katakuse, M., Kakubuchi, S., Nagao, T., Furuyama, K., Cosmochimica Acta, v. 38, p. 757–775. doi:10.1016/0016- Kamei, A., Kawabata, H., Nakajima, J., Van Keken, P.E., and 7037(74)90149-5 Stern, R.J., 2014, Diverse magmatic effects of subducting a Patino, L.C., Carr, M.J., and Feigenenson, M.D., 2000, Local hot slab in SW Japan: Results from forward modeling: and regional variations in Central American arc lavas INTERNATIONAL GEOLOGY REVIEW 33

controlled by variations in subducted sediment input: Rudnick, R.L., and Gao, S., 2003, Composition of the continen- Contributions to Mineralogy and Petrology, v. 138, p. tal crust, in Rudnick, R.L., Holland, H.D., and Turekian, K.K., 265–283. eds., Treatise on Geochemistry, Volume 3: Oxford, Elsevier, Pearce, J.A., and Peate, D.W., 1995, Tectonic implications of the p. 1–64. composition of volcanic arc magmas: Annual Review of Saginor, I., Gazel, E., Carr, M.J., Swisher, C., III, and Turrin, B., Earth and Planetary Sciences, v. 23, p. 251–285. 2011, New Pliocene–Pleistocene 40Ar/39Ar ages fill in tem- doi:10.1146/annurev.ea.23.050195.001343 poral gaps in the Nicaraguan volcanic record: Journal of Pearce, J.A., Stern, R.J., Bloomer, S.H., and Fryer, P., 2005, Volcanology and Geothermal Research, v. 202, p. 143–152. Geochemical mapping of the Mariana arc-basin system: doi:10.1016/j.jvolgeores.2011.02.002 Implications for the nature and distribution of subduction Saginor, I., Gazel, E., Condie, C., and Carr, M.J., 2013, Evolution components: Geochemistry, Geophysics, Geosystems, v. 67, of geochemical variations along the Central American vol- doi:10.1029/2004GC000895 canic front: Geochemistry, Geophysics, Geosystems, v. 14, p. Peccerillo, A., and Taylor, S.R., 1976, Geochemistry of eocene 4504–4522. doi:10.1002/ggge.20259 calc-alkaline volcanic rocks from the Kastamonu area, Stamatelopoulou-Seymour, K., Vlassopoulos, D., Pearce, T.H., Northern Turkey: Contributions to Mineralogy and and Rice, C., 1990, The record of magma chamber processes Petrology, v. 58, p. 63–81. doi:10.1007/BF00384745 in plagioclase phenocrysts at Thera Volcano, Aegean Pichler, H., and Weyl, R., 1975, Magmatism and crustal evolu- Volcanic Arc, Greece: Contributions to Mineralogy and tion in Costa Rica (Central America): Geol. Rundsch., v. 64, p. Petrology, v. 104, no. 1, p. 73–84. doi:10.1007/BF00310647 457–475. doi:10.1007/BF01820678 Stern, R.J., 2010, The anatomy and ontogeny of modern intra- Plank, T., Balzer, V., and Carr, M.J., 2002, Nicaraguan oceanic arc systems, in Kusky, T.M., Zhai, M.-G., and Xiao, W., volcanoes record paleoceanographic changes accompanying eds., The evolving continents: Understanding processes of closure of the Panama gateway: Geology, v. 30, p. 1087–1090. continental growth: Geological Society of London, Special doi:10.1130/0091-7613(2002)030<1087:NVRPCA>2.0.CO;2 Publication 338, p. 7–34. Plank, T., and Langmuir, C.H., 1988, An evaluation of the global Straub, S.M., 2003, The evolution of the Izu-Bonin-Mariana variations in the major element chemistry of arc basalts: volcanic arcs (NW Pacific) in terms of major elements: Earth and Planetary Science Letters, v. 90, p. 349–370. Geochemistry, Geophysics, Geosystems, v. 4, p. 1018. doi:10.1016/0012-821X(88)90135-5 doi:10.1029/2002GC000357 Plank, T., and Langmuir, C.H., 1993, Tracing trace elements Straub, S.M., Woodhead, J.D., and Arculus, R.J., 2015, Temporal from sediment input to volcanic output at subduction evolution of the Mariana Arc: Mantle wedge and subducted zones: Nature, v. 362, p. 739–743. doi:10.1038/362739a0 slab controls revealed with a tephra perspective: Journal of Reagan, M.K., and Gill, J.B., 1989, Coexisting calcalkaline and Petrology, v. 56, p. 409–439. doi:10.1093/petrology/egv005 high-niobium basalts from Turrialba Volcano, Costa Rica: Sugimura, A., 1968, Spatial relations of basaltic magmas in Implications for residual titanates in arc magma sources: island arcs, in Hess, H.H., and Poldervaart, A., eds., Basalts: Journal of Geophysical Research, v. 94, p. 4619–4633. The Poldervaart treatise on rocks of basaltic composition: doi:10.1029/JB094iB04p04619 New York, Wiley, p. 537–572. Ringwood, A.E., 1974, The petrological evolution of island arc Sun, S.S., and McDonough, W.F., 1989, Chemical and isotopic systems: Twenty-seventh William Smith Lecture: Journal of systematics of oceanic basalts: Implications for mantle com- the Geological Society of London, v. 130, p. 183–204. position and processes, in Saunders, A.D., and Norry, M.J., doi:10.1144/gsjgs.130.3.0183 eds., Magmatism in the Ocean Basins: Geological Society of Rogers, R.D., Mann, P., and Emmet, P.A., 2007b, Tectonic ter- London, Special Publication 42, p. 313–345. ranes of the Chortis Block based on integration of regional Syracuse, E.M., Abers, G.A., Fischer, K., MacKenzie, L., Rychert, aeromagnetic and geologic data, in Mann, P., ed., Geologic C., Protti, M., González, V., and Strauch, W., 2008, Seismic and tectonic development of the Caribbean plate in north- tomography and earthquake locations in the Nicaraguan ern Central America: The Geological Society of America and Costa Rican upper mantle: Geochemistry, Geophysics, Special Paper 428, doi:10.1130/2007.2428(04) Geosystems, v. 9, no. 7, p. Q07S08. doi:10.1029/ Rogers, R.D., Mann, P., Emmet, P.A., and Venable, M.E., 2007a, 2008GC001963 Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015 Colon fold belt of Honduras: Evidence for Late Cretaceous Taylor, S.R., and McLennan, S.M., 1985, The continental crust: collision between the continental Chortis block and intra- Its composition and evolution: Oxford, Blackwell Scientific. oceanic Caribbean arc, in Mann, P., ed., Geologic and tec- Turner, S.J., and Langmuir, C.H., 2015, What processes control tonic development of the Caribbean plate in northern the chemical compositions of arc front stratovolcanoes?: Central America: Volume 428: The Geological Society of Geochemistry, Geophysics, Geosystems, v. 16, p. 1865– America Special Paper, doi:10.1130/2007.2428(06) 1893. doi:10.1002/2014GC005633 Rooney, T.O., Franceschi, P., and Hall, C.M., 2010, Water-satu- Van Avendonk, H.J., Holbrook, W.S., Lizarralde, D., and Denyer, rated magmas in the Panama Canal region: A precursor to P., 2011, Structure and serpentinization of the subducting adakite-like magma generation?: Contributions to Cocos plate offshore Nicaragua and Costa Rica: Mineralogy and Petrology, doi:10.1007/s00410-010-0537-8 Geochemistry, Geophysics, Geosystems, v. 12, doi:10.1029/ Rudnick, R.L., 1995, Making continental crust: Nature, v. 378, p. 2011GC003592 571–578. doi:10.1038/378571a0 Venable, M., 1994, A geological, tectonic, and metallogenic Rudnick, R.L., and Fountain, D.M., 1995, Nature and composi- evaluation of the Siuna terrrane (Nicaragua) [Ph.D. disserta- tion of the continental crust: A lower crustal perspective: tion]: Tucson, AZ, University of Arizona, 154 p. Reviews of Geophysics, v. 33, p. 267–309. doi:10.1029/ Vogel, T.A., Patino, L.C., Alvarado, G.E., and Gans, P.B., 2004, 95RG01302 Silicic ignimbrites within the Costa Rican volcanic front: 34 S. A. WHATTAM AND R. J. STERN

Evidence for the formation of continental crust: Earth and White, W.M., McBirney, A.R., and Duncan, R.A., 1998, Petrology Planetary Science Letters, v. 226, p. 149–159. doi:10.1016/j. and geochemistry of the Galapagos Islands: Portrait of a epsl.2004.07.013 pathological mantle plume: Journal of Geophysical Wegner, W., Worner, G., Harmon, R.S., and Jicha, B.R., 2011, Research, v. 98, no. B11, p. 19533–19563. Magmatic history and evolution of the Central American land White, W.M., and Patchett, J., 1984, HfNdSr isotopes and bridge in Panama since Cretaceous times: Geological Society of incompatible element abundances in island arcs: America Bulletin, v. 123, p. 703–724. doi:10.1130/B30109.1 Implications for magma origins and crust-mantle evolution: Werner, R., Hoernle, K., Barckhausen, U., and Hauff, F., 2003, Earth and Planetary Science Letters, v. 67, p. 167–185. Geodynamic evolution of the Galapagos hot spot system doi:10.1016/0012-821X(84)90112-2 (Central East Pacific) over the past 20 m.y.: Constraints from Wörner, G., Harmon, R.S., and Wegner, W., 2009, Backbone of morphology, geochemistry, and magnetic anomalies: the Americas, in Kay, S.M., Ramos, V.A., and Dickinson, W., Geochemistry, Geophysics, Geosysttems, v. 4, doi:10.1029/ eds., Shallow subduction, plateau uplift and ridge and ter- 2003GC000576 rane collision, Geological Society of America, Memoir 204, p. Whattam, S.A., Montes, C., McFadden, R.R., Cardona, A., 183–196. Ramirez, D., and Valencia, V., 2012, Age and origin of ear- Zernack, A.V., Price, R.C., Smith, I.E.M., Cronin, S.J., and Stewart, liest adakitic-like magmatism in Panama: Implications for R.B., 2012, Temporal evolution of a high-K andesitic mag- the tectonic evolution of the Panamanian magmatic arc matic system: Taranaki Volcano, New Zealand: Journal system: Lithos, v. 142-143, p. 226–244. doi:10.1016/j. of Petrology, v. 53, p. 325–363. doi:10.1093/petrology/ lithos.2012.02.017 egr064 Whattam, S.A., and Stern, R.J., 2015, Late Cretaceous plume- Zimmer, M.M., Plank, T., Hauri, E.H., Yogodzinski, G.M., Stelling, induced subduction initiation along the southern margin of P., Larsen, J., Singer, B., Jicha, B., Mandeville, C., and Nye, C. the Caribbean and NW South America: The first documen- J., 2010, The role of water in generating the calc-alkaline ted example with implications for the onset of plate tec- trend: New volatile data for Aleutian magmas and a new tonics: Gondwana Research, v. 27, p. 38–63. http://dx.doi. tholeiitic index: Journal of Petrology, v. 51, p. 2411–2444. org/10.1016/j.gr.2014.07.011 doi:10.1093/petrology/egq062 Downloaded by [The University of Texas at Dallas] 18:30 15 December 2015