Magmatic peridotites and pyroxenites, Andong Ultramafic Complex, Korea: Geochemical evidence for supra-subduction zone formation and extensive melt–rock interaction

Magmatic peridotites and pyroxenites, Andong Ultramafic Complex, Korea: Geochemical evidence for supra-subduction zone formation and extensive melt–rock interaction

Lithos 127 (2011) 599–618 Contents lists available at ScienceDirect Lithos j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t ...

4MB Sizes 0 Downloads 10 Views

Lithos 127 (2011) 599–618

Contents lists available at ScienceDirect

Lithos j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / l i t h o s

Magmatic peridotites and pyroxenites, Andong Ultramafic Complex, Korea: Geochemical evidence for supra-subduction zone formation and extensive melt–rock interaction Scott A. Whattam a,⁎, Moonsup Cho b, Ian E.M. Smith c a b c

Smithsonian Tropical Research Institute, Apartado Postal 0843-03092, Balboa, Ancon, Panama School of Earth and Environmental Sciences, Seoul National University, Seoul, 151-747, Republic of Korea Department of Geology, University of Auckland, Private Bag 92019, Auckland, New Zealand

a r t i c l e

i n f o

Article history: Received 30 November 2010 Accepted 30 June 2011 Available online 8 July 2011 Keywords: South Korea Ultramafic magmas Peridotite Melt–rock reaction Supra-subduction zone (SSZ) Volcanic arc

a b s t r a c t The Andong Ultramafic Complex (AUC) mainly comprises peridotites (wehrlites ± plagioclase or spinel; or plagioclase + spinel) and related serpentinites with subordinate low-Al pyroxenites (clinopyroxenites, orthopyroxenites, and websterites). These rocks are compositionally similar to sub-continental lithospheric mantle peridotites and pyroxenites. Wehrlites formed predominantly by fractional crystallization processes within supra-subduction zone magmas and the pyroxenites are generally consistent with segregation and accumulation in similar magmas. Bulk rock ratios of Al2O3/SiO2 (0.01–0.03) and MgO/SiO2 (up to N 1) exhibited by the wehrlites and serpentinites indicate crystallization from a refractory source that underwent high degrees of melt extraction. Spinel chemistry confirms this and demonstrates that wehrlite and clinoproxenite protoliths underwent approximately 20–23% and 12–15% partial melting, respectively. Wehrlites and serpentinites also preserve evidence of extensive melt–peridotite interaction manifest as bulk rock SiO2-depletions and FeOt-enrichments relative to mantle residua as well as low Mg# (0.39–0.45) spinels with variable Ti contents but constant Cr# (0.42–0.47). These features are identical to those of ‘impregnated’ plagioclase-peridotites of abyssal and sub-continental environments and compositional trends in spinel space imply reaction between secondary, MORB-like melts saturated in olivine + clinopyroxene or olivine and a harzburgitic protolith. High olivine:pyroxene (~ 3:1) and clinopyroxene:orthopyroxene ratios of the wehrlites coupled with chemical data dictate that reactions entailed orthopyroxene dissolution and olivine recrystallization. All AUC rock types exhibit primitive mantle-normalized incompatible element signatures characterized by LILE-enrichments, high fluid-mobile/immobile element ratios (Sr/Nd, Ba/La and Pb/Ce ≫ 1) and prominent HFSE (Nb, Zr, and Ti) depletions indicative of generation in a sub-arc environment within a supra-subduction zone system. A candidate for the associated arc-system is the one responsible for nearby arc-related Jurassic granitoids. Southeast-directed thrusting along the Andong Fault System may account for subsequent emplacement of the AUC into the Gyeongsang Basin. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Partial melting of the Earth's mantle has resulted in mantle differentiation and the generation and extraction of basalt to form oceanic and continental crust. Among the Earth's ultramafic rocks the relatively fertile lherzolites (15% clinopyroxene) are generally regarded as fragments of pristine mantle that underwent relatively small degrees of partial melting; harzburgites (b5% clinopyroxene) are alternatively interpreted as mantle residues left behind following relatively higher degrees of partial melting (Bodinier, 1988; Frey et al., 1985). Recently however, this simple peridotite–basalt ‘residua⁎ Corresponding author at: Department of Earth and Environmental Sciences, Korea University, Seoul 136 701, Korea. Tel.: + 507 212 8000x8341. E-mail addresses: [email protected], [email protected] (S.A. Whattam). 0024-4937/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.lithos.2011.06.013

extraction’ relationship has been challenged, particularly in the case of western European sub-continental lithospheric mantle (SCLM) or ‘oceanic–continental transition’ (i.e., ‘Alpine’ or orogenic) peridotites (e.g., Leroux et al., 2007; Müntener et al., 2010; Piccardo et al., 2007; Piccardo and Guarnieri, 2010). In these complexes, the recognition of the role and importance of melt–rock interactions catalyzed by secondary partial melting processes and mantle ‘refertilization’ has come to the fore. Peridotites of SCLM and oceanic or intra-oceanic forearc (i.e., ophiolites) settings can also be of a magmatic (i.e., crystallization) origin and occur with associated cumulate ultramafic rocks and residual peridotite (e.g., the Cabo Ortegal SCLM peridotite complex, NW Spain, Santos et al., 2002; the Zambales ophiolite, Philippines, Yumul, 2004). In addition to sub-continental and oceanic environments, ultramafic complexes also comprise cratonic lithosphere and

600

S.A. Whattam et al. / Lithos 127 (2011) 599–618

are produced in a variety of different tectonic settings including passive margins, mid-ocean ridges, rifts and hotspots, oceanic forearcs and arc systems (Arai, 1992 and references therein). Exposure at the Earth's surface occurs as xenoliths entrapped not only in alkali but also in calc-alkaline (e.g., Ishimaru et al., 2007) basalts that erupt at rifts or oceanic hotspots; emplaced or underplated lowermost sections of ophiolites; and, as mentioned above, in Alpine-type peridotites (Ishiwatari, 1985). Ultramafic complexes of the Korean Peninsula (Fig. 1a and b) have not received much attention until quite recently (e.g., Arai et al., 2008; Hisada et al., 2008; Oh et al., 2010; Seo et al., 2005) due possibly in part to their relatively rare occurrence (e.g., Lee, 1987; Wee et al., 1994). In the Andong area i.e., the Andong Ultramafic Complex (AUC, Fig. 1c) has not yet been studied in detail but has the potential to yield important data on both the mantle composition beneath the Korean Peninsula and on interpretations of its tectonic evolution. The AUC is emplaced into sedimentary country rocks at the fault-bounded northern section of the Cretaceous Gyeongsang Basin (Fig. 1c; Choi et al., 2002) and comprises a bimodal suite of peridotites (wehrlites) and related serpentinites, and a fertile, near-basaltic suite of pyroxenites and gabbros; hornblende–biotite gabbros and biotite granites occur as meter-scale dykes or sills that intrude the ultramafic rocks. We report results of a petrological and geochemical study of the AUC that provide the first detailed treatment of the petrogenesis of

the AUC and demonstrate its complex evolution within a sub-arc (supra-subduction zone, SSZ) environment. By comparison of the AUC with other global and Korean peridotite complexes, we discuss its significance in an East Asian tectonic context. 2. Geological and tectonic setting The Korean Peninsula represents the easternmost extension of the Sino-Korean and the South China cratons and embodies a crucial link between the North and South China continental blocks and the island arc system of Japan (Fig. 1; see reviews of Chough et al., 2000; Chough and Sohn, 2010). The peninsula is dominated by three major Precambrian massifs and two Phanerozoic mobile belts which isolate the three massifs from each other. From north to south, the Nangrim and Gyeonggi massifs are subdivided by the Imjingang Belt and the southernmost Yeongnam Massif is separated from the Gyeonggi Massif by the intervening Okcheon (Fold) Belt (Fig. 1b). Basement rocks of the Gyeonggi and Yeongnam massifs encompass high-grade gneiss and schist. The southern and eastern segments of the peninsula comprise the expansive Cretaceous non-marine Gyeongsang Basin (Fig. 1b). The Korean peninsula was situated to the immediate rear of the Japanese Arc system in the Cretaceous. Production of the magmatic arc was the result of oblique northward subduction of the Izanagi Plate (a northern segment of the palaeo-Pacific plate) (Kim et al.,

Fig. 1. (a) Location of the Korean Peninsula in the North Pacific bounded by the North and South China Blocks to the west and SW Japan to the east. (b) Detail of the boxed region in (a) showing the locations of the Andong Ultramafic Complex (AUC) and five other ultramafic–mafic complexes exposed in Ulsan, Yugu, Baekdong, Bibong and Macheon. (c) Detail of the boxed region in (b) showing the location of the circa 2 × 1 km AUCe at the northern section of the Cretaceous non-marine Gyeongsan Basin which in turn is bounded by the Andong Fault System. Sketch in (a) is an amalgamation of maps modified from Seo et al. (2005) and Hisada et al. (2008); (b) and (c) are modified from Choi et al. (2002). LC beside Macheon in (b) stands for Layered Complex.

S.A. Whattam et al. / Lithos 127 (2011) 599–618

1997; Klimetz, 1983; Maruyama et al., 1997; Woods and Davies, 1982) beneath the eastern margin of the Asian continent circa ~ 130– 100 Ma. Subduction also resulted in a left-lateral wrench tectonics regime in a retro-arc setting and the ultimate formation of a number of relatively small transtensional basins in the southern and southwestern present-day Korean Peninsula and the expansive Gyeongsang Basin to the southeast (Chun and Chough, 1992). The evolution of the Gyeongsang Basin has been ascribed to pull-apart or transtensional tectonics (Lee, 1999) possibly accompanied by plumeassisted rifting (Okada, 1999, 2000); and alternatively upon a conjugate strike-slip faulting and compressional block rotation model (Hwang et al., 2008a, 2008b). Although volcanic deposits of the Yucheon Group (Chang, 1975) in the southeastern segment of the Gyeongsang Basin suggest proximity to a volcanic arc source, these deposits have traditionally been interpreted as an unconformable sedimentary unit which accumulated in the domain of the Gyeongsang Basin (Chang, 1975; Choi, 1986). In a recent innovative re-interpretation of the Early to Late Cretaceous tectonic evolution of the Korean Peninsula, Chough and Sohn (2010) suggest instead that volcanogenic deposits of the Yucheon Group represent an originally extensive volcanic arc platform which they coin the Gyeongsang Volcanic Arc (GVA). In such a scenario, the Gyeongsang Basin represents a back-arc basin that formed behind the GVA, which would have been a southern to south-western extension of the Japan Arc (see Fig. 6 of Chough and Sohn, 2010). Besides the AUC, the Gyeongsang Basin also hosts the (residual) Ulsan serpentinite complex (Fig. 1b; Choi et al., 1990; Hisada et al., 2008) which is exposed along the southeastern border of the basin. Three other peridotite bodies have also been identified in the southwestern section of the Gyeonggi Massif (Fig. 1b) in the Hongseong area. These include the Yugu peridotite complex (Arai et al., 2008) and the smaller Baekdong and Bibong peridotite/serpentinite complexes (Oh et al., 2010; Seo et al., 2005). While the Baekdong peridotites are interpreted as a section of passive margin lithospheric mantle (Seo et al., 2005), the Bibong peridotites are interpreted as a mantle slice of SSZ lithosphere (Oh et al., 2010). Peridotites of the Yugu complex are interpreted as sub-arc mantle wedge or abyssal peridotites (Arai et al., 2008). To the southwest of the Gyenongsang Basin the Macheon Layered Complex has been identified from the Yeongnam Massif (Fig. 1b) and dated at 223 ± 3 Ma by Kim and Turek (1996) on the basis of U–Pb zircon analyses. The complex is comprised mainly of olivine and hornblende gabbros (Song et al., 2007), but no mineral or whole rock chemical data are currently available. The tear-shaped, circa 2 × 1 km AUC lies within and at the northwestern margin of the Gyeongsang Basin to the immediate south of the Andong Fault System (Fig. 1c; Choi et al., 2002) which demarcates the boundary in this region between the Gyeongsang Basin and Precambrian to Jurassic basement rocks of the Yeongnam Massif to the north (Fig. 1b). Lithologies comprise dominantly ultramafic rocks (N90% by volume) with subordinate meter to tens-

601

of-meters scale gabbroic and granitic dykes and sills. Ultramafic exposures range from completely altered to very fresh. Secondary mm-cm scale hydrothermal veins associated with the serpentinites are ubiquitous and consist of epidote, altered plagioclase and clays, etc. Granites are manifest as dykes and sills intruding the ultramafic rocks, and the contact between the two is generally sharp and welldefined. Relations between the gabbroic intrusions and the ultramafic rocks they encroach, however, are much less clear although geochemical and mineral chemistry suggest they are related. Brittle deformation is manifest as faults on large (tens of meters) to small (cm) scales while ductile deformational features are abundant in the form of sinuous, felsic (rodingite?) cm-scale bands. Although we provide petrographic (Section 3), mineral (Section 4) and whole-rock chemistry (Section 5) evidence for a cumulate origin for the pyroxenites such evidence is not readily observed in the field. Some exceptions are seen in outcrops of gabbro and clinopyroxenite where samples AND-P:10 and 12 (gabbro) and AND-P:15 and 16 (clinopyroxenite) were collected. For example, gabbro sample ANDP:12 is from a relatively large (~20 × 5 m) stock that intrudes massive ultramafic rocks. Locally, compositional layering is observed in the gabbroic stock defined by circa 5–50 mm felsic (plagioclase-rich) bands (Fig. A1). Within the ultramafic outcrop where clinopyroxenite samples AND-P:15 and 16 were collected distinct subhorizontal compositional layering is defined by alternating ~5 mm felsic and ~2 mm mafic layers. 3. Nomenclature and petrography 3.1. Nomenclature and ultramafic classification Nomenclature used for the ultramafic rocks is from the classification scheme of Streckeisen (1976) which uses modal percentages of olivine, clinopyroxene and orthopyroxene (Fig. A2). Modal abundance of each mineral is visually estimated and listed in Table 1 and thin-section photomicrographs of representative samples are provided in Fig. 2. Due to inherent errors associated with classification via mode estimation and the moderate to severely serpentinized nature of the complex as a whole, classification was aided by whole-rock chemistry considerations (see Section 5). For example, sample AND-P:13a was difficult to classify as it comprises essentially 40% olivine and 60% clinopyroxene and thus plots at the boundary separating wehrlite from clinopyroxenite (Fig. A2). We term AND-P13a as an ‘anomalous’ clinopyroxenite but note that some major and trace element concentrations are more similar to wehrlites or possess concentrations intermediate to the clinopyroxenites and wehrlites. 3.2. Petrography Peridotites are comprised of olivine (70–85%) and lesser clinopyroxene ± spinel or plagioclase or spinel + plagioclase and

Table 1 Modalogy of Andong Ultramafic Complex (a) peridotites and pyroxenites and (b) gabbros. Lithology PERIDOTITE WEHR PL WEHR SP-PL WEHR PYROXENITE SP-PL-OL CPXN SP-PL-OL OPXN PL-OL WEB GABBRO HB-BT GABBRO

Representative samples

Olivine

Clinopyroxene

AND-P:18, 29, 30 AND-P:24 AND-P:01

85 70–75 70–75

15 20 20

AND-P:13a, 15, 16 AND-P:20, 21 AND-P:19 AND-P:10, 12 AND-P:09, 26, 27

30–35 25–30 25–30

60–65 Trace 10 50

Orthopyroxene

Plagioclase

Hornblende

Biotite

10

35

b5 Up to 12 Trace 60 55–65

10 10–15 40 55

Note: Spinel occurs in the spinel–plagioclase wehrlites as a minor phase (b10 modal%). BT = Biotite; CPXN = Clinopyroxenite; HB = Hornblende; OL = Olivine; OPXN = Orthopyroxenite; PL = Plagioclase feldspar; PXN = Pyroxenite; SP = Spinel; and WEHR = Wehrlite.

602

S.A. Whattam et al. / Lithos 127 (2011) 599–618

comprise wehrlite (samples AND-P:18, 29, 30), plagioclase wehrlite (AND-P:24) and spinel plagioclase wehrlite (AND-P:01) (Fig. 2). Plagioclase-bearing wehrlites contain lesser olivine and more clinopyroxene than the wehrlites (20% vs. 15%) (Table 1). Wehrlite olivines are euhedral to subhedral, ~1.5 mm crystals (Fig. 2a) or N2 mm, elongated ‘stringers’ (Fig. 2b and c) and are ubiquitously ‘meshed’ due to alteration to serpentine along fractures. Clinopyroxenes occur rarely as 1–5 mm stubby, subhedral to euhedral prisms that pokilitically enclose small (b0.5 mm) olivine chadacrysts but predominantly as anhedral masses interstitial to olivine (Fig. 2a and b) A conspicuous feature of some wehrlites is the presence of plagioclase (up to 10– 15 modal%) which we describe as troctolitic/gabbroic ‘pockets’ interstitial to olivine (Fig. 2b and c). Replacement of primary pyroxene by carbonates in the wehrlite is evident by the presence of ample pseudomorphs. Other secondary phases include fine-grained amphiboles after clinopyroxene and sericite after plagioclase. Carbonates are also present as anastomosing ~0.1 mm veinlets. Spinels (b5% by volume) primarily occur as rounded 0.2–0.3 mm blebs embedded in olivine. The pyroxenites (Fig. A2, Fig. 2d–f) contain either clinopyroxene or orthopyroxene and olivine (~ 25–35% of the mode) (Table 1)

AND-P:01 sp-pl websterite

irregularly with trace (≪ 5%) amounts of the other pyroxene and are termed olivine clinopyroxenites and olivine orthopyroxenites. Exceptions include the orthopyroxene- and plagioclase-rich (~10– 15 modal%) pyroxenite sample (AND-P:19) which has ~ 10% clinopyroxene and is best termed an olivine websterite (Fig. 2b and c); and samples AND-P:04 and 6b which contain appreciable hornblende (~20–30%) and are best classified as hornblende pyroxenites (Fig. A2). In contrast to the wehrlites, the pyroxenites are generally very fresh although plagioclase is commonly saussuritized and/or altered to sericite. The clinopyroxenites (AND-P:13a, 15, 16) comprise ~60–65% clinopyroxene and 30–35% olivine (Table 1; Fig. 2d). Clinopyroxenite olivines are meshed similar to those of the wehrlite but are virtually free of secondary alteration products. Olivines are generally subhedral and range from 0.5 to 5.0 mm whereby anhedral to nearly euhedral clinopyroxenes are on average of the order of ~ 2 mm and commonly display simple twins. Plagioclase–olivine websterites comprise ~55– 60% orthopyroxene, 25–30% olivine, rare interstitial clinopyroxene (b5%) as in the wehrlites and appreciable plagioclase (~10–15%) (Table 1). Cumulate texture is manifest as large (2–6 mm), anhedral orthopyroxenes poikilitically enclosing much smaller olivine chadacrysts (Fig. 2e and f). Unlike the wehrlites many pyroxenites display

AND-P:15 sp-pl-ol clinopyroxenite

ol

ol

interstitial cpx

cpx

carb

sp

ol

ol cpx

cpx ol ol

ol

cpx (replaced)

ol

a

1 mm

primary cpx

DSCN 8089 pl wehrlite

ol

pl

ol

ol

AND-P:31B pl-ol websterite

ol opx

opx ol

ol

opx opx

ol

pl

d

1 mm

pl

ol

opx

b

1 mm

primary cpx

DSCN 8089 pl wehrlite

AND-P:31B pl-ol websterite

ol

ol

opx

cpx ol ol

ol

ol

opx

ol opx

pl

e

1 mm

ol

pl 1 mm

c

1 mm

f

Fig. 2. Representative photomicrographs of AUC wehrlites (a–c) illustrating the typical coarse-grained, proto-granular to porphyroclastic nature defined by subhedral meshed olivine porphyroclasts (a) or olivine ‘stringers’ (b, c) surrounded by interstitial clinopyroxene and mm-sized troctolitic or gabbroic ‘pockets’ comprised of commonly highly sericitized plagioclase laths; and (d–f) pyroxenites. The clinopyroxenites (d) and orthopyroxenites exhibit weak equilibrium textures defined by triple junctions (white arrowheads) indicative of crystal boundary migration and recrystallization. Pokilitic textures predominate in the pyroxenites and are best developed in the websterites (e, f) which exhibit large anhedral orthopyroxenes that enclose much smaller olivine chadacrysts. (a, b) and (d, e) are under crossed nicols and (c, f) under plain light.

S.A. Whattam et al. / Lithos 127 (2011) 599–618

equilibrium textures such as sutured grain boundaries and triple junctions (Fig. 2d) indicative of equilibrium recrystallization. Serpentinites retain rare remnant olivines; whole-rock data (Section 5) suggests the serpentinites and wehrlites share common protoliths but underwent different post-formation petrogeneses (see Section 6). There are two distinct types of gabbros: low-Ti (0.14 wt.%) gabbros (Fig. A2) that generally lie along compositional trends from the wehrlites and a silica- and alkali-rich hornblende–biotite-bearing gabbro with whole-rock compositions (see Sections 4 and 5) different from those of the gabbros. The gabbros comprise approximately 50% clinopyroxene and 40% plagioclase whereby the hornblende–biotite gabbros have ~ 55% plagioclase and 35% biotite with hornblende concentrations typically less than ~ 5–10 modal% (Table 1). Both gabbro types are relatively fresh. Minor secondary phases include fine-grained amphiboles after clinopyroxene in the gabbro and minor chlorite after biotite in the hornblende–biotite gabbros. Gabbros texturally range from coarse to fine whereby the hornblende–biotite gabbros are coarse-grained. Ophitic and subophitic textures dominate in both gabbro types with clinopyroxenes enclosed or partially enclosed in plagioclase in the gabbros and amphibole and biotite enclosed in plagioclase in the hornblende–biotite gabbros. Mineralogy of the granites in descending order is potassium feldspar, quartz, biotite and subordinate plagioclase and amphibole. The granites are porphyritic defined by phenocrysts of potassium feldspar; minor alteration is manifest as sericite and/or clays after potassium feldspar. 4. Mineral chemistry Mineral chemistry data were obtained on polished, carbon-coated thin sections using a JEOL 8900 electron microprobe at Seoul National University. Operating voltage was 15 kV with a beam current of 10 nA and a beam diameter of 5 μm. Natural and synthetic oxides as well as silicate minerals were used as standards. Data acquisition and reduction were done using an automated ZAF correction program. A summary of compositional characteristics of olivine, clinopyroxene, orthopyroxene and plagioclase is provided in the online appendix as Table A.1 and complete mineral analyses of these phases are provided in Table A.2. Representative spinel analyses are provided in Table A.3. 4.1. Olivine Olivine Fo contents are relatively low and composition is mostly homogenous within a given sample and rock type (Tables A.1a, A.2a). An exception is the disparity between olivines of the (plagioclasefree) wehrlites (Fo84.5) (which have a similar composition to those of the olivine websterites and orthopyroxenites with means of Fo86) and the spinel–plagioclase wehrlites (mean Fo87.2). Clinopyroxenite olivines have Fo contents slightly higher (Fo88) than those of the plagioclase wehrlites. All peridotite and pyroxenite olivines have very low Ti and Cr concentrations (typically less than 0.1 wt.%) and correspondingly low NiO contents. Olivine core NiO concentrations are 0.23–0.25 wt.% in the wehrlite and 0.25–0.29 in the anomalous spinel–olivine clinopyroxenite AND-P13a. Olivines of the spinel– plagioclase–olivine clinopyroxenites (samples AND-P:15, 16) alternatively, display a large range in NiO (0.18–0 wt.%) as do orthopyroxenite and websterite olivines which range to higher minimum values than the clinopyroxenites (NiO = 0.28–0.15 wt.%). NiO concentrations of all ultramafic olivines are significantly lower than those of olivines encompassing the mantle olivine array (Takahashi et al., 1987). Collectively, these observations are consistent with a magmatic origin for the AUC ultramafic rocks (see Section 6).

603

from Morimoto, 1989). Diopside rims are commonly enriched in iron and depleted in Ca relative to cores indicative of normal zonation, while augite commonly records a reverse pattern. In contrast to other phases, many clinopyroxenes show broad compositional variations and large compositional discrepancies exist between cores and rims. An exception is the spinel–plagioclase wehrlite (sample AND-P:01) which solely comprises diopside (Mg# = 100 Mg/[Mg + Fe 2+] = 89– 90) with homogenous cores (Wo47–48En46–47Fs6). Diopsides also predominate in the clinopyroxenite dyke (AND-P:06a) and the spinel–olivine clinopyroxenite (AND-P:13a). In contrast, clinopyroxenes of the clinopyroxenites are compositionally heterogeneous. Diopside cores (Mg# = 85–90) in the clinopyroxenite dyke vary moderately from Wo45–49En45–49Fs5–9 and exhibit homogeneous diopside rims (Mg# = 89) that fall within the compositional range of the cores. In all cases, diopside is Cr-rich with upper limits of Cr2O3 N 1.0 wt.% (Table A.2b). Clinopyroxenes in the spinel– plagioclase–olivine clinopyroxenites (samples AND-P:15, 16) are predominantly augite (cores: Mg# = 88–90, Wo38–45En50–55Fs6–7). A rare clinopyroxene from the websterite (AND-P:19) has an augite core (Mg# = 86, Wo32En58Fs10) and a diopside rim. Cores of gabbro clinopyroxenes (Mg# = 88–93) are (relatively) high-Ca diopside (Wo48En47Fs5–6) or low-Ca augite and rims are homogenous low-Ca augite that record iron enrichment and Ca-depletion trends (Tables A.1b and A.2b). Orthopyroxenes in the plagioclase–olivine websterites and plagioclase–olivine orthopyroxenites are enstatite (Tables A.1c and A.2c) and share compositional similarities. Enstatite cores and rims are exceptionally homogenous and record uniformity between cores and rims. Orthopyroxene core compositions of websterite and orthopyroxenites are Wo1–3En83–85Fs13–14 and Wo2–3En84–85Fs13–14, respectively. Similarly, Mg# is nearly equivalent for orthopyroxenes of the websterites and orthopyroxenites (86 and 87, respectively), very similar to the Fo content (i.e., Mg#) of pyroxenite olivines (~86–88) but lower than the Mg# (~ 89–93) recorded by clinopyroxenes in the peridotites, clinopyroxenites and gabbros. 4.3. Plagioclase Apart from plagioclase in the hornblende–biotite gabbros, composition is homogenous within a given rock type (Tables A.1d and A.2d). Wehrlite plagioclase (labradorite) is markedly more Ca-poor (core average of An65) than the pyroxenites (An77–79) (Table A.1d). A single plagioclase analysis of olivine clinopyroxenite yields a composition of An85 identical to that of plagioclase cores (An83–86) in gabbro. Plagioclase in the orthopyroxenites and websterites is identical with cores recording compositions of An78–79 and An76–78, respectively. In contrast, plagioclase compositions in the hornblende– biotite gabbros are Na-rich and suggestive of secondary alteration. 4.4. Spinel Type I spinel is Cr-rich and occurs exclusively in plagioclase–spinel wehrlite (AND-P:01) whereby Type II spinel is Al-rich and comprises anomalous spinel–olivine clinopyroxenite (AND-P:13a) and two spinel–plagioclase–olivine clinopyroxenites (AND-P:15,16) (Fig. A3, Table A.3). Type I spinel has higher Cr, Fe and Ti but lower Al than the Type II spinels. Type I spinels are also distinguishable from Type II on the basis of lower Mg# (36–45 vs. 42–56) and higher Cr# (100 Cr/[Cr + Al]= 42–47 vs. 25–31). Paired Al2O3 vs. Cr2O3 concentrations of both spinel types fall within the spinel mantle array of Franz and Wirth (2000) (see Section 6).

4.2. Pyroxene

5. Bulk rock chemistry

Peridotites and pyroxenites comprise roughly equivalent amounts of diopside and augite (Tables A.1b, A.2b) (pyroxene classification is

Samples were crushed and ground in an agate mill for whole rock major, trace and rare earth element (REE) determinations. Major

604

Table 2 Whole-rock XRF and ICP-MS analyses of Andong Ultramafic Complex rocks. Mg# = [100 Mg/Mg + Fe2+]. Major oxides (wt.%)

40.14 0.13 1.19 12.99 0.18 33.86 2.58 0.06 0.02 0.02 0.99 6.60 98.76 84

ANDP:29 WEHR

ANDP:30 WEHR

ANDP:24 PL WEHR

ANDP:01 SPPL WEHR

ANDP:13a SP-OL CPXN

AND-P: 15 SP-PLOL CPXN

AND-P:16 SP-PL-OL CPXN

ANDP:19 PL-OL WEB

ANDP:20 OL OPXN

ANDP:21 PL-OL OPXN

ANDP:13b SERP

ANDP:17 SERP

ANDP:23 SERP

ANDP:25 SERP

ANDP:10 GB

AND-P: 12 GB

ANDP:09 HB-BT GB

ANDP:26 HB-BT GB

ANDP:27 HB-BT GB

ANDP:07 GR

ANDP:08 GR

36.56 0.09 1.02 11.39 0.15 37.35 2.60 0.05 0.01 0.02 1.25 8.39 98.88 87

38.68 0.11 1.24 11.61 0.15 36.74 3.38 0.08 0.02 0.02 0.94 6.29 99.28 86

37.92 0.09 2.55 11.72 0.15 35.66 2.91 0.06 0.02 0.02 1.03 6.88 99.01 86

40.05 0.12 4.96 10.84 0.15 34.68 2.50 0.36 0.04 0.02 0.38 5.70 99.80 86

41.72 0.20 1.68 9.45 0.12 30.32 8.33 0.15 0.01 0.01 0.37 5.83 98.19 86

42.83 0.13 6.93 7.93 0.11 26.37 10.17 0.35 0.10 0.01 0.25 3.79 98.97 87

43.28 0.13 7.14 7.75 0.11 26.21 10.50 0.34 0.13 0.01 0.49 3.27 99.36 87

43.14 0.08 9.46 10.21 0.13 27.07 5.36 0.41 0.44 0.02 0.41 2.74 99.47 84

49.44 0.34 2.64 7.24 0.13 23.32 11.29 0.30 0.05 0.01 0.52 3.49 98.77 86

48.79 0.29 3.49 9.70 0.15 26.12 7.29 0.25 0.10 0.02 0.34 2.30 98.85 84

36.22 0.04 0.63 12.05 0.09 35.83 0.03 0.00 0.00 0.01 1.09 11.88 97.87 85

36.94 0.06 0.77 13.53 0.10 36.37 0.05 0.01 0.01 0.02 1.42 9.52 98.80 84

37.55 0.04 0.53 12.78 0.09 36.43 0.10 0.00 0.01 0.02 1.49 9.95 98.99 85

37.32 0.11 1.13 10.07 0.15 35.28 3.47 0.02 0.01 0.02 1.49 9.99 99.06 87

47.50 0.14 19.62 3.77 0.07 10.33 14.00 1.36 0.36 0.01 0.21 2.23 99.59 84

47.61 0.13 19.58 2.86 0.05 10.44 12.93 1.43 1.11 0.01 0.29 2.93 99.38 88

51.93 0.80 17.96 5.84 0.09 8.16 7.48 3.50 0.25 0.29 0.35 2.66 99.30 73

48.74 1.19 14.75 9.05 0.14 10.53 7.44 3.33 1.43 0.30 0.29 1.95 99.15 70

49.22 1.69 16.49 9.98 0.15 6.76 8.25 3.77 1.46 0.42 0.15 1.00 99.34 57

72.45 0.16 14.54 1.29 0.02 0.57 0.44 4.94 3.81 0.05 0.31 0.80 99.38 47

73.45 0.17 14.05 0.99 0.01 0.36 0.84 4.17 4.14 0.03 0.19 0.76 99.17 42

Trace elements (ppm) Sc 15 12 V 61 38 Cr 3167 2612 Ni 1591 1815 Cu 54 504 Zn 111 90 Rb 2.42 0.25 Sr 20.89 23.82 Y 3 1.92 Zr 9 7 Nb 0.07 0.05 Cs 0.32 bdl Ba 5.06 5.89 La 1.22 1.18 Ce 1.76 1.39 Pr 0.36 0.28 Nd 1.61 1.28 Sm 0.47 0.35 Eu 0.15 0.12 Gd 0.54 0.44 Tb 0.12 0.12 Dy 0.58 0.44 Ho 0.11 0.08 Er 0.35 0.25 Tm 0.05 0.03 Yb 0.33 0.21 Lu 0.05 0.03 Hf 0.19 0.16 Ta bdl bdl Pb 0.55 1.12 Th 0.07 0.03 U 0.02 bdl

16 44 1587 1054 61 79 0.31 23.39 2.84 9 0.06 0.03 6.5 0.91 1.83 0.36 1.78 0.52 0.16 0.66 0.13 0.56 0.12 0.33 0.05 0.30 0.05 0.22 0.02 0.88 0.04 bdl

12 44 3263 1754 63 100 0.41 24.59 2.01 8 0.08 0.06 6.48 1.01 1.62 0.31 1.39 0.39 0.12 0.43 0.11 0.44 0.08 0.24 0.03 0.24 0.03 0.16 bdl 0.67 0.05 0.02

8 44 1987 1444 6 85 0.9 250.59 1.3 7 0.03 0.35 36.07 1.05 1.75 0.22 1.06 0.26 0.24 0.26 0.05 0.26 0.06 0.17 0.03 0.14 0.03 0.07 bdl 0.75 bdl bdl

27 74 2868 3221 576 50 0.64 25.87 5.21 11 0.07 0.39 4.52 0.76 2.43 0.47 2.96 0.99 0.30 1.15 0.17 1.11 0.22 0.61 0.07 0.51 0.07 0.35 bdl 4 0.03 0.03

23 59 1945 606 65 48 2.72 236.17 3.07 10 0.04 0.24 8.79 0.58 1.53 0.28 1.67 0.56 0.23 0.64 0.09 0.66 0.13 0.35 0.04 0.29 0.04 0.18 bdl 0.82 bdl bdl

26 65 2080 596 57 47 4.09 266.53 3.34 12 0.04 0.31 12.71 1.58 1.74 0.38 1.76 0.62 0.26 0.77 0.14 0.69 0.15 0.38 0.05 0.29 0.05 0.23 bdl 0.77 bdl bdl

13 30 1118 1661 204 66 9.93 414.97 1.18 10 0.03 0.09 32.29 4.15 1.56 0.26 0.84 0.18 0.18 0.21 0.08 0.23 0.05 0.15 0.02 0.16 0.02 0.08 bdl 0.6 bdl bdl

41 148 2695 960 255 48 2.9 71.63 8.73 19 0.15 0.27 13.06 2.06 5.58 1.03 5.51 1.51 0.44 1.75 0.31 1.77 0.35 0.94 0.14 0.91 0.13 0.49 0.02 0.72 0.11 0.03

33 104 2913 1753 229 69 4.93 169.33 6.56 17 0.11 0.3 42.51 1.74 4.07 0.76 3.96 1.15 0.35 1.34 0.25 1.27 0.26 0.75 0.11 0.71 0.12 0.44 0.02 0.75 0.08 0.02

3 26 3321 5570 582 74 0.07 1.61 0.83 4 0.06 0.09 8.32 0.30 0.59 0.09 0.48 0.11 0.04 0.18 0.02 0.18 0.04 0.10 0.01 0.10 0.02 0.05 bdl 1.99 0.02 0.04

8 38 3815 1804 79 98 0.41 2.12 0.88 6 0.08 0.09 8.73 0.77 0.89 0.18 0.60 0.14 0.05 0.17 0.07 0.17 0.03 0.10 0.02 0.13 0.02 0.09 bdl 0.49 0.05 0.03

5 27 2730 2240 228 76 0.2 2.68 0.54 5 0.04 bdl 3.53 1.34 0.36 0.10 0.26 bdl 0.03 0.11 0.05 0.11 0.03 0.07 bdl 0.10 0.02 0.05 bdl 0.84 0.02 0.02

20 58 3872 1257 94 76 3.97 872.93 12.47 9 3.74 0.15 149.7 19.18 37.68 4.46 18.04 3.22 1.11 2.92 0.39 2.46 0.46 1.34 0.19 1.29 0.19 2.86 0.26 5.73 2.98 0.68

23 62 201 153 42 18 8.31 847.74 2.97 16 0.08 0.28 129.62 1.24 2.22 0.33 1.81 0.57 0.35 0.67 0.09 0.67 0.12 0.34 0.05 0.31 0.04 0.16 0.03 1.54 0.1 0.02

22 56 819 95 28 17 34.54 1331.39 2.71 20 0.03 1.56 500.16 0.91 2.18 0.36 1.97 0.61 0.31 0.58 0.08 0.58 0.11 0.30 0.04 0.26 0.04 0.18 bdl 2.15 bdl 0.03

17 94 86 80 12 48 4.71 1036.07 15.04 170 4.63 0.19 196.69 25.63 50.15 5.86 23.21 4.22 1.40 3.63 0.48 3.00 0.56 1.63 0.24 1.60 0.25 3.62 0.33 7.44 4.18 0.93

24 167 539 186 38 88 34.55 750.45 24.8 152 6.38 4.47 457.49 25.26 55.10 7.19 29.59 5.74 1.65 5.49 0.83 4.87 0.95 2.74 0.39 2.63 0.36 3.49 0.39 5.7 1.85 0.4

18 131 119 76 25 98 33.54 1465.77 29.72 251 8.08 1.45 513.91 27.05 57.57 7.37 31.96 6.63 2.10 6.28 0.94 5.82 1.14 3.29 0.46 2.98 0.45 5.39 0.47 8.3 1.89 0.55

3 15 18 22 7 18 110.2 318.1 7.87 95 5.48 1.06 516.58 8.62 15.18 1.57 5.95 1.25 0.37 1.19 0.17 1.32 0.25 0.84 0.13 0.89 0.13 3.11 1.1 25.31 11.56 9.2

3 16 3 0 5 9 95.27 402.87 9.15 90 5.5 0.87 366.16 10.15 19.76 2.01 7.50 1.71 0.34 1.50 0.25 1.51 0.30 0.96 0.14 1.05 0.16 3.37 0.91 20.45 14.05 3.06

Abbreviations are as in Table 1.

S.A. Whattam et al. / Lithos 127 (2011) 599–618

SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 H2O LOI total Mg#

ANDP:18 WEHR

S.A. Whattam et al. / Lithos 127 (2011) 599–618

element concentrations were determined using X-ray fluorescence (XRF) on fused glass beads by standard XRF techniques at the Department of Geology, University of Auckland. Abundances of selected trace elements (Zr, V, Cr, Ni, Cu and Zn) were also obtained via XRF but on pressed pellets. Concentrations of all other trace and REE were obtained by inductively coupled plasma mass spectrometry (ICP-MS) at Australian National University. Accuracies of the XRF analyses are ±2% for major elements present in concentrations greater than 5% and ±5% for trace elements. The ICP-MS analyses yield accuracies greater than ±5%. Whole-rock XRF and ICP-MS data are provided in Table 2.

5.1. Major and minor elements Loss on ignition (LOI) values of wehrlites and the anomalous spinel–olivine clinopyroxenite are relatively high (5.70–8.39) (Table 2) reflecting their moderate to severe serpentinization (for comparison, LOI of serpenitinites is in the range of 9.52–11.88). The generally fresher pyroxenites yield much lower LOI values (2.30– 3.79) similar to those of the gabbros (2.23–2.93). Major element compositional variations between AUC peridotites (wehrlites and serpentinites) and pyroxenites primarily reflect differing modal abundances of olivine, pyroxene and plagioclase feldspar. Wehrlites and serpentinites plot within or on the cusp of the (global) peridotite field (Fig. 3) whereby all pyroxenites plot within the low-Al pyroxenite field (Bodinier and Godard, 2003) (which is analogous to the Group I and Cr-diopside pyroxenite classification schemes of Frey and Prinz (1978) and Shervais (1979) respectively, and agrees with the high Cr-content of pyroxenite diopside, e.g. 0.67– 1.08 wt.% Cr2O3). Gabbros have higher Al2O3 (N20 wt.%) and lower SiO2/MgO ratios than MORB and plot within the field of ophiolitic gabbro. SiO2/MgO ratios of serpentinites (1.01–1.06) overlap those of plagioclase-free wehrlites (0.98–1.15) but possess lesser Al2O3 than the latter (0.61– 1.29 wt.% vs. 1.14–1.35). Compared with plagioclase-free peridotites, the plagioclase wehrlites have significantly higher Al2O3 (2.80– 5.30 wt.%) similar to the orthopyroxenites (2.79–3.64 wt.%) reflecting relatively high plagioclase content (up to ~ 10–15 modal%) (Table 1). The relatively high SiO2/MgO ratios of the pyroxenites in general reflects higher pyroxene/olivine ratios, while higher SiO2/MgO ratios in the orthopyroxenites relative to peridotites and other pyroxenites

Fig. 3. Distribution of AUC peridotites (wehrlites), serpentinites, pyroxenites and gabbros in Al2O3–SiO2/MgO space (Bodinier and Godard, 2003). All wehrlites apart from the spinel–plagioclase wehrlite plot within the peridotite field while all pyroxenites fall into the low-Al pyroxenite field (also the Group I and Cr-diopside pyroxenite fields of Frey and Prinz, 1978 and Shervais, 1979, respectively) with b 10 wt.% Al2O3. Wehrlites and serpentinites are distinguishable from pyroxenites on the basis of major element chemistry by virtue of lower SiO2/MgO ratios (b 1.20 vs. N1.35).

605

is a reflection of higher modal orthopyroxene and relatively lesser olivine (Table 1). Variations in MgO vs. Al2O3, TiO2, CaO and Na2O (Fig. 4) demonstrate that wehrlites have TiO2 (0.10–0.14 wt.%) and CaO (2.67–3.68 wt.%) abundances similar to but Al2O3 (1.19–4.96 wt.%) and Na2O (0.06–0.38 wt.%) concentrations generally lower than primitive mantle (PM) (Hart and Zindler, 1986; McDonough and Sun, 1995). The wehrlites also display major element abundances and variations at equivalent MgO wt.% comparable to shallow European sub-continental lithospheric mantle (SCLM) Alpine peridotites (Downes, 2001) and Al2O3 and Na2O concentrations that are typically less than abyssal peridotites (Fig. 4) (Niu, 2004). Apart from the wehrlites having significantly higher MgO (33.86–37.35 wt.%) and lesser CaO than the pyroxenites (23.63–30.89 wt.% MgO and 5.42– 11.44 wt.% CaO) the wehrlites also generally display lesser Al2O3, TiO2 and Na2O than the pyroxenites (Table 2). The serpentinites are interpreted as altered peridotites and generally plot with the wehrlites but have lower CaO and Al2O3 at a given MgO wt.% (Fig. 4). Three serpentinites have anomalously very low CaO (0.03– 0.11 wt.%) which is possibly the result of Ca-loss during serpenitization. Alternatively, the Ca-enrichment recorded by the wehrlites (see Section 6) may indicate Ca-exchange between the serpentinite protoliths and the wehrlites which was facilitated by melt-induced metasomatism. In addition to having higher MgO and lesser CaO, Al2O3, TiO2 and Na2O (Fig. 4), the peridotites have higher total FeO (FeO t) and lower SiO2 than the pyroxenites (Table 2; see also Section 6) and range to lower ratios of CaO/Al2O3 (0.50–2.72) than those of the pyroxenites (0.57–4.97) (Fig. 4). Apart from the ‘anomalous’ clinopyroxenite (AND-P:13a, see below) which exhibits very low Al2O3 (1.82 wt.%) and a correspondingly anomalously high CaO/Al2O3 ratio (4.97) and an olivine websterite which has unusually high Al2O3 (9.46 wt.%), the pyroxenites plot within fields defined by European SCLM mantle pyroxenites. The wehrlites have Ni and Cr concentrations (1880–2100 ppm Ni, 1780–3590 ppm Cr) similar to PM (2090 ppm Ni, 3240 ppm Cr; Hart and Zindler, 1986). Three highly depleted and refractory serpentinites which display U-shaped REE patterns (see Section 5.2) have higher Cr (3060–4330 ppm) and much higher Ni (2400–6540 ppm) but relatively lesser Sc (8–9 ppm) than the wehrlites (9–18 ppm) which probably reflects higher olivine/pyroxene ratios and relatively more Cr-spinel. Apart from anomalous clinopyroxenite sample AND-P:13a (3610 ppm Ni, 3420 ppm Cr) the peridotites generally have higher concentrations of Ni and Cr than the remaining pyroxenites (Table 2). Overall, gabbros have similar MgO (10.63–10.86 wt.%) as MORB but higher Al2O3 and CaO, lesser Na2O and much lesser TiO2 (0.14 wt.%) similar to the wehrlites (0.10–0.14 wt.%) and the plagioclase-bearing clinopyroxenites (0.14 wt.%). On a MgO/SiO2 vs. Al2O3/SiO2 plot (Fig. 5) the terrestrial peridotite array (Jagoutz et al., 1979) delineates the progressive magmatic depletion trend (gray line and arrow, Fig. 5) of primitive and increasingly more residual mantle subjected to increasingly higher degrees of melting (and hence melt extraction, as recorded by Al2O3/SiO2). The most depleted (refractory) compositions are defined by Al2O3/SiO2 and MgO/SiO2 ratios of b0.01 and N1, respectively (Hart and Zindler, 1986; Jagoutz et al., 1979). Thus, peridotites of residual mantle origin subjected to increasing degrees of melting should follow a progressive trend towards lower Al2O3/SiO2 (higher degrees of melt extraction) and higher MgO/SiO2 along a trend that parallels the terrestrial array. Fundamentally, serpentinization does not obscure primary magmatic signatures as documented by MgO/SiO2 and Al2O3/SiO2 ratios (Niu, 2004; see also Godard et al., 2008) and thus, signatures recorded by Andong serpentinites can be considered as reliable as those of the less altered wehrlites. Peridotites of the Baekdong and Bibong ultramafic complexes are interpreted as residual (Oh et al., 2010; Seo et al., 2005) and clearly follow a progressive residual trend from the more fertile Baekdong

606

S.A. Whattam et al. / Lithos 127 (2011) 599–618

lherzolites (Al2O3/SiO2 = 0.07–0.08 and MgO/SiO2 = 0.89–0.91) to the more refractory Bibong lherzolites and harzburgites and serpentinites (Al2O3/SiO2 = 0.01–0.03 and MgO/SiO2 = 0.98–1.13) which all plot either along or above the terrestrial array. The higher MgO/SiO2 of the

Fig. 5. Plot of AUC peridotites and serpentinites and (inset) all ultramafic and mafic rocks in Al2O3/SiO2–MgO/SiO2 space (symbols for AUC rocks as in Fig. 3). Shown for comparison are the compositions residual peridotites of the Baekdong and Bibong ultramafic complexes (serpentinites are indicated by an ‘x’) and abyssal melt ‘impregnated’ plagioclase lherzolites (ODP Leg 209, Sites 1270, 1271, Godard et al., 2008). Superimposed on the plots is the terrestrial (peridotite) array (Jagoutz, 1979) which defines the compositional trend followed by increasingly refractory (i.e., progressively Al2O3/SiO2-depleted and MgO/SiO2-enriched) residua subjected to progressive degrees of melt extraction (thick gray arrow). The hatched box represents a hypothetical composital range that residual, refractory, parental AUC peridotite might be expected to occupy.

Bibong vs. the Baekdong peridotites presumably reflects a higher olivine/pyroxene ratio of the former vs. the latter. The AUC serpentinites and plagioclase-free wehrlites also exhibit relatively low Al2O3/SiO2 ratios (0.01–0.03 and 0.03, respectively) which overlap those of the Bibong harzburgites and lherzolites illustrating their depleted nature. Similarly, the MgO/SiO2 ratios exhibited by the plagioclase-free wehrlites and serpentinites range from 0.84 to 1.02 and 0.95 to 0.99; the higher end of these values are close to those typical of refractory mantle rocks (i.e., MgO/SiO2 N 1). However, the Andong peridotites and serpentinites also demonstrate considerable diversity in MgO/SiO2 vs. Al2O3/SiO2 space (Fig. 5) and unlike the residual Bibong and Baekdong peridotites, do not follow a trend that parallels the terrestrial peridotite array. Instead, the wehrlites follow a near-vertical trend towards lower MgO/SiO2 (1.02–0.84) at constant Al2O3/SiO2 (0.03) (Fig. 5, inset) suggesting secondary SiO2 addition (i.e., melt- or fluid-derived SiO2 metasomatism). Secondary compositional modifications aside, in addition to low primitive-mantle normalized high-field strength (HFSE) and HREE concentrations which are below primitive-mantle and low HFSE/LREE ratios (Section 5.2) the relatively high MgO/SiO2 and low Al2O3/SiO2 ratios of the plagioclase-free wehrlites and serpentinites indicate that they are derived from a relatively refractory, residual peridotite parent.

5.2. Trace and rare earth elements PM-normalized incompatible trace element signatures of ultramafic rocks and gabbros are grossly similar and are characterized by significant but highly variable enrichments in fluid-mobile elements

Fig. 4. Selected major elements vs. MgO plots of AUC peridotites and pyroxenites. Wholerock concentrations here and in subsequent figures are normalized to 100% on a volatilefree, anhydrous basis. The field of sub-continental lithospheric mantle (SCLM) European peridotites is from Downes (2001). Data for the pyroxenites are from the Alps (Bodinier, 1988), Pyrenees (Bodinier et al., 1987), Ronda (Garrido and Bodinier, 1999; Schubert, 1977; Suen and Frey, 1987), Beni Bousera (Pearson et al., 1993) and Cabo Ortegal (Gravestock, 1992). The field for abyssal peridotite is from Niu (2004) and MORB data is from Melson et al. (1976); Schilling et al. (1983); and Korenaga and Kelemen (2000). Primitive mantle (PM) compositions (stars) are from Hart and Zindler (1986) (black) and McDonough and Sun (1995) (white).

S.A. Whattam et al. / Lithos 127 (2011) 599–618

(e.g., Cs, Rb, Ba, Th, U, K, LREE, Pb, Sr), high mobile/immobile elements ratios (e.g., Sr/Nd, Ba/La, and Pb/Ce ≫ 1) and prominent HFSE depletions (Fig. 6). Variable enrichments in highly incompatible elements (e.g., Cs, Th) relative to LREE are manifest for example by [Th/Ce]PM-N values that range from 28 to 49 in the LREE-enriched wehrlites, gabbros and websterites and MREE-enriched clinopyroxenites, to 82–112 in the MREE-enriched orthopyroxenites. The serpentinites with U-shaped REE patterns (see below) exhibit modest incompatible element enrichments (e.g., [Th/Ce]PM-N = 7–18) whereby the extreme LREE-enriched serpentinite displays correspondingly extreme Th-enrichment ([Th/Ce]PM-N = 754). High mobile/immobile elements ratios are manifest by prominent PM-normalized Sr and Pb spikes and selective enrichments in U relative to Th and LREE (Fig. 6). Sr-enrichments and depletions appear to be related to plagioclase accumulation (or lack thereof) as relative degrees of Sr-enrichment are correlated with chondrite-normalized Eu-anomalies (Fig. 7) and because Sr exhibits high compatibility in plagioclase (distribution coefficients ≫ 1 for plagioclase (Blundy and Wood, 1991). While the wehrlites, clinopyroxenites and orthopyroxenites exhibit modest Sr-enrichments ([Sr/Ce]PM-N = 1–13,) the websterite and gabbros exhibit significantly greater enrichments ([Sr/Ce]PM-N = 22 and 32–51, respectively). Alternatively, the three LREE- and extreme-HFSE-depleted serpentinites exhibit Sr-troughs ([Sr/Ce]PM-N = 0.2–0.6) while the LREE-enriched serpentinite exhibits a modest enrichment ([Sr/Ce]PM-N = 2). The plagioclase–olivine websterite exhibits the highest Sr-enrichment of ultramafic rocks and also the most exaggerated Eu-anomaly (Eu/Eu* = 2.95); in contrast the plagioclase-poor wehrlites which exhibit negligible Sr-enrichments ([Sr/Ce]PM-N = 1.0–1.33) also demonstrate correspondingly slight negative Eu-anomalies (Eu/Eu* = 0.60–0.92). Pbenrichment trends are striking but also highly variable. [Pb/Ce]PM-N averages 33 in wehrlites and clinopyroxenites (excluding anomalous clinopyroxenite AND-P:13a with [Pb/Ce]PM-N = 44 which is identical to the average for the gabbros) and 31 in the websterite but jumps to 96 in the orthopyroxenites. HFSE (e.g., Nb, Zr and Ti) depletions are striking and typically much lower than PM (Fig. 6). Negative-Nb depletions for example range from [Nb/Ce]PM-N = 0.05–0.10 in the wehrlites and all pyroxenites while gabbros exhibit [Nb/Ce]PM-N = 0.03–0.09 and the serpentinites display [Nb/Ce]PM-N (0.22–0.28). Similarly, apart from the unique LREE and MREE-enriched serpentinite, Zr concentrations are depleted with respect to HREE and Hf (not plotted) (Fig. 7). Wehrlites exhibit chondrite-normalized LREE-enrichments and MREE-enrichments manifest as sub-horizontal signatures of negative slope (Fig. 7a; Table 2). In contrast, pyroxenites generally display convex-up REE patterns (i.e., MREE enrichment relative to both LREE and HREE) (Fig. 7b and c). Compared to clinopyroxenites, the orthopyroxenites exhibit slighter LREE-depletion and flatter LREE to MREE segments (Table 2, Fig. 7c). The plagioclase–olivine websterite is anomalously LREE-enriched ([La/Sm]C-N = 14.62, [La/Yb]C-N = 16.82) with a flat HREE segment (i.e., [Gd/Yb]C-N = 1) and very low HREE concentrations. Two serpentinite samples (AND-P:17, 23) have the lowest absolute REE concentrations of all rocks and exhibit prominent LREE-enrichment ([La/Sm]C-N, [La/YbC-N] ratios = 3.39, 3.79–10.22) (Table 2, Fig. 7d). Of the other two serpentinites, one (AND-P:13b) exhibits modest LREE-enrichment and a gently negative-sloping REE pattern similar to the wehrlites; the other serpentinite (AND-P:25, Fig. 7f) exhibits a REE pattern analogous to that of the extreme LREE- and MREE- and enriched hornblende–biotite gabbros. Gabbros display very similar patterns to the wehrlites (compare Fig. 7a and e) and have comparable LREE and MREE enrichments (Table 2). The hornblende–biotite gabbros and biotite granites are significantly enriched in LREE (Table 2, Fig. 7f). Prominent positive Eu-anomalies are exhibited by the plagioclase– olivine websterite (Eu/Eu* = 2.95), spinel–plagioclase wehrlite (Eu/Eu* = 2.84) and the gabbros (Eu/Eu* = 1.60–1.74) (Table 2,

607

Fig. 7). In addition to exhibition of positive Eu-anomalies, these samples also possess excess Sr (see below) and Al2O3 (Table 2) and these features probably reflect plagioclase accumulation. Slight negative Euanomalies exhibited by the orthopyroxenites (Eu/Eu* = 0.83–0.87) and spinel–olivine clinopyroxenite (Eu/Eu* = 0.87) on the other hand likely indicates plagioclase removal. A binary plot of Al2O3/SiO2 vs. [La/Sm]C-N (not shown) illustrates decreasing Al2O3/SiO2 with increasing LREE-enrichment for the wehrlites and serpentinites (but not the plagioclase-bearing wehrlites or pyroxenites) and suggests that LREE-enrichment was concomitant with source depletion and probably not related to secondary metasomatic events (see Section 6). Moreover, positive correlations also exist between HFSE-depletions and other highly mobile, subduction–environment elements (e.g., Th, Sr). Some features are attributable to cumulate processes (e.g., Sr spikes and Eu-anomalies in the pyroxenites and gabbros), but the prominent HFSE depletions coupled with LILE enrichments are indicative of concomitant subduction-derived fluid enhancement in the melting source (Ellam and Hawkesworth, 1988; Gamble et al., 1996) and are typical of arclike magmas (e.g., Pearce and Peate, 1995). 6. Discussion 6.1. Wehrlite and pyroxenite: residual, referitilized or magmatic products? It has become increasingly recognized that compositional variations of many peridotite complexes (or at least segments therein), are the result of melt–rock reactions catalyzed by secondary partial melting events and melt infiltration. For example, in many large, orogenic Alpine lherzolite massifs (e.g., Lanzo, Italy; Ronda, Spain; Horoman, Japan), meter–decimeter scale bands of harzburgite and dunite originally interpreted as slices of highly refractory and residual mantle have been re-interpreted as lherzolite that underwent extensive melt–rock interactions associated with secondary melt percolation (e.g., Bodinier et al., 1991; Takazawa et al., 1992). Melt– rock reactions have also dominated the petrogeneses of Alpine plagioclase-peridotites which are interpreted as remnant refractory residues subsequently ‘impregnated’ by secondary, percolating MORB-like melts (e.g., Rampone et al., 1994, 1997; Müntener et al., 2004). These ‘mantle re-fertilization’ (e.g., Lenoir et al., 2001; Van der Wal and Bodinier; 1996) reactions entail two key processes: (i) a secondary partial melting event that produces the metasomatic agent (melt/fluid); and (ii) melt/fluid–peridotite interaction/reaction which causes modal and compositional modifications in the residual peridotite. As we illustrate below, AUC peridotites and pyroxenites are consistent with a magmatic crystallization and cumulate origin; however, the wehrlites also record evidence of extensive secondary peridotite–melt interactions similar to those described above that resulted in modal, whole-rock and mineral chemistry modifications typical of such interactions. 6.1.1. Establishment of a ‘non-residual mantle’ and primary magmatic origin for the AUC On the spinel Al2O3 vs. Cr2O3 plot (Franz and Wirth, 2000) all Andong wehrlite and clinopyroxenite spinels plot within the spinel mantle array along with spinels from ultramafic rocks of the Baekdong, Bibong (Oh et al., 2010; Seo et al., 2005) and Yugu (Arai et al., 2008) ultramafic complexes (Fig. 8). However, coexisting olivines of the Andong wehrlites and clinopyroxenites exhibit lower Fo contents (Fo85–88) than mantle olivines (typically higher than ~Fo89) and as a result plot outside of the olivine–spinel mantle array (OSMA) (Arai, 1992) (Fig. 9). Arai (1994) considers the OSMA to be a field which encompasses the spinel Cr# vs. olivine Fo content spectrum of residual mantle peridotite compositions and that

608

S.A. Whattam et al. / Lithos 127 (2011) 599–618

S.A. Whattam et al. / Lithos 127 (2011) 599–618

(magmatic) cumulates plot to the right of this array. Below we show that a cumulate origin is probable for the clinopyroxenites at least and note also that on a plot of spinel Cr# vs. Fe# plot (not shown) all wehrlite and clinopyroxenite spinels lie to the right of the mantle array and are compositionally similar to Archaean layered gabbroperidotites, Greenland (Rollinson, 2007; Rollinson et al., 2002). Similarly, along with olivines of the websterite and plagioclase– olivine orthopyroxenite, Andong wehrlite and clinopyroxenite olivines have much lower NiO concentrations (b0.3 wt.%) than (residual) mantle olivines (typically with N0.3 wt.% NiO, Fig. 9b) and plot well below the mantle olivine array of Takahashi et al. (1987). Thus, a residual origin for the AUC is categorically ruled out. Trends for olivines of the plagioclase-free wehrlite and ‘anomalous clinopyroxenite’ AND-P:13a are defined by significant decreases in NiO with decreasing Fo# and plot along olivine fractional crystallization trends of Sato (1977), indicating that these rocks are indeed magmatic. Furthermore, olivines of the remaining clinopyroxenites dramatically fall off the crystallization trend at Fo88–87 and plot along a near vertical trend below 0.18 wt.% NiO. Thus, the majority of clinopyroxenites comprise cumulate olivine and are probably magmatic cumulates.

6.1.2. Pyroxenites as crystalline segregates On the CaO–Al2O3–MgO plot (Fig. 10) of Coleman (1977) all pyroxenites fall within the field of ultramafic cumulates while four of five wehrlites and one serpentinite overlap the fields of ultramafic cumulates and metamorphic peridotite. The three remaining serpentinites plot distinctly near the MgO apex similar to all Bibong residual lherzolites and harzburgites; the other (spinel–plagioclase) wehrlite plots uniquely in a region intermediate to the peridotite and cumulate fields. Apart from the spinel–plagioclase wehrlite which displays some evidence of plagioclase accumulation manifest as a relatively large positive chondrite-normalized Eu-anomaly (Eu/Eu* = 2.84) (Table 2) and anomalously high Sr (251 ppm which is an order of magnitude higher than the remaining wehrlites) there is no evidence to support a cumulate origin for the wehrlites. As we discuss in Section 6.1.4, bulk rock and mineral chemistry of the wehrlites indicate that they have undergone extensive melt–rock interaction which has modified their bulk composition resulting in enrichments of Fe and Ca and depletions in Mg and Si. Thus, the original composition of the wehrlites probably plotted nearer to the MgO apex and completely within the peridotite field. A cumulate origin for the pyroxenenites on the other hand is in general agreement with petrographic and field evidence (e.g., poikilitic texture, pyroxenite layers within massive wehrlite), whole-rock geochemistry and mineral chemistry considerations. As explained by Downes (2007), compositions of mantle pyroxenites by and large do not lie between compositions of peridotite and ‘common’ basalts (e.g., MORB) and hence must represent crystalline cumulates in a broad sense. As shown in Fig. 4, AUC pyroxenite compositions also generally do not lie between the compositional fields of peridotite and MORB and excluding anomalous clinopyroxenite P:13a, the pyroxenites have relatively high and variable concentrations of Al2O3 (2.79– 9.87 wt.%), CaO (5.60–11.72) and Sc (11–47 ppm). The Al2O3 and CaO abundances lie along a near vertical trend at relatively constant MgO (24.63–28.24 wt.%) suggesting plagioclase and pyroxene accumulation. Similarly, the pyroxenites with the highest Al2O3 concentrations (clinopyroxenites and the olivine websterite with N7 wt.% Al2O3) also have anomalously high concentrations of Sr (236–415 ppm) and exhibit pronounced positive Eu-anomalies (Eu/Eu* = 1.14–2.95) (Fig. 7) suggestive of plagioclase accumulation. Moreover, as illustrated in Fig. 9b, clinopyroxenite olivines fall off the fractional crystallization trend defined by the plagioclase

609

wehrlites and the anomalous clinopyroxenite which suggests that the remaining clinopyroxenites comprise cumulate olivine. We provide further evidence of a magmatic origin for the clinopyroxenites in Section 6.1.6 on the basis of spinel and olivine chemistry. The nearly identical mantle-normalized ‘arc-like’ signatures and Mg# of the wehrlites and clinopyroxenites (84–87 and 86–87 respectively) suggest that the pyroxenites formed within the same SSZ magmas crystallizing the wehrlites. In this scenario, the pyroxenites for whatever reason apparently did not interact with secondary MORB-like melts to the degree with which did the wehrlite protoliths. Since first proposed by Obata (1980) for pyroxenites of the Ronda peridotite, a cumulate origin has been further invoked for many dykes and layers of pyroxenite in ultramafic massifs particularly of SLCM European peridotite complexes, e.g., the Balmuccia peridotite, Italy (Mukasa and Shervais, 1999; Shervais and Mukasa, 1991; Sinigoi et al., 1983) and the Ronda peridotite (Suen and Frey, 1987). 6.1.3. Significance of plagioclase-bearing peridotites Troctolitic or gabbroic ‘pockets’ occur interstitially to olivine in some AUC peridotites (Fig. 2). Various petrogenetic models have been proposed for plagioclase-peridotites including the formation of ‘fertile’ upper mantle diapirs which have undergone incomplete melt extraction and subsequent rapid cooling and partial crystallization of melt pockets (Boudier and Nicolas, 1986; Nicolas, 1986); the production of fertile rocks via melt ‘impregnation’ of basaltic liquids and subsequent melt– peridotite reaction and equilibration (Boudier and Nicolas, 1986; Dick, 1989), i.e. via melt or fluid metasomatism in the plagioclase-stability field (Müntener et al., 2004); and generation of plagioclase feldspar by subsolidus metamorphic reactions during uplift from spinel to plagioclase facies conditions (Rampone et al., 1993). As we illustrate below, bulk rock and mineral major element chemistry is consistent with melt– peridotite interaction and melt impregnation. 6.1.4. Bulk rock major element evidence of melt–rock interaction Calculated compositions of refractory mantle residua subjected to progressively higher degrees of melt extraction follow a characteristic melt extraction trend towards decreasing SiO2 and increasing MgO at relatively constant FeO t regardless of the mode of melting (Fig. 11) (Niu, 1997). Thus, as in Fig. 5 where peridotites of residual mantle origin are expected to plot along the terrestrial peridotite mantle array of Jagoutz et al. (1979), residual peridotites should also plot similarly to the calculated extraction trends of Niu (1997) in MgO/SiO2 and MgO/FeOt space. As in Fig. 5, the Bibong and Baekdong peridotites unambiguously follow residual mantle trends toward lesser SiO2 and higher MgO at stable FeO t concentrations. In contrast, the AUC peridotites, i.e. the wehrlites and serpentinites, do not follow such a trend but instead exhibit much lower SiO2 and higher FeO t compared with refractory residua at comparable MgO abundances (Fig. 11). This feature is identical to characteristics exhibited by ‘meltimpregnated’ plagioclase–lherzolites of Alpine peridotite complexes (Piccardo et al., 2007; Piccardo and Guarnieri, 2010) and cratonic lherzolite–wehrlite xenoliths metasomatically altered by ascending, evolved, silica-undersaturated melts (Ionov et al., 2005) (Fig. 11). Similarly, the AUC wehrlites and peridotites also record Ca-enrichment with respect to mantle residua of comparable MgO concentrations (not shown) which is also characteristic of the plagioclase– lherzolites and wehrlites outlined above. Furthermore, AUC wehrlites and serpentinites clearly define a metsomatic trend towards higher SiO2 and lesser FeO t and MgO which is not seen in refractory peridotites. Consensus is that the compositional variations recorded by the Alpine plagioclase–lherzolites are the result of metasomatic effects caused by MORB–melt/peridotite interaction and consequent sublithospheric

Fig. 6. Primitive mantle-normalized incompatible trace element plots of AUC ultramafic rocks and of gabbroic and granitic dykes that cut the complex. Primitive mantle concentrations are from Sun and McDonough (1989).

610

S.A. Whattam et al. / Lithos 127 (2011) 599–618

Fig. 7. Chondrite-normalized REE plots of AUC ultramafic rocks and of gabbroic and granitic dykes that cut the complex. Chondritic concentrations are from Nakamura (1974).

S.A. Whattam et al. / Lithos 127 (2011) 599–618

611

Fig. 8. Distribution of spinels from ultramafic rocks of the Andong, Baekdong and Bibong (Oh et al., 2010; Seo et al., 2005), and Yugu (Arai et al., 2008) ultramafic complexes on the mantle array plot of Franz and Wirth (2000).

mantle re-fertilization (Piccardo and Guarnieri, 2010 and references therein). A similar interpretation appears reasonable for the AUC wehrlites and serpentinites. Moreover, the significant iron enrichment recorded by AUC peridotites and serpentinites (FeO t = 10.41–13.85 wt.%) with respect to refractory melting residues at equivalent MgO (Fig. 11) cannot be explained by melting but can be explained rather as a result of refertilization or modal metasomatism as is widely accepted (e.g., Bodinier et al.,, 2003 and references therein). An enrichment in MnO (with respect to concentrations typical of melting residues) in the peridotites and serpentinites (MnO = 0.10–0.19 wt.%) is also indicative of melt metasomatism (e.g., Ionov et al., 2005). These bulk chemical trends exhibited by the AUC peridotites and serpentinites are entirely consistent with re-fertilization and impregnation of plagioclase–wehrlite protoliths with secondary, percolating basaltic melts.

6.1.5. Mineral major element evidence of melt–rock interaction Ti contents of spinel preserve a record of the nature of the impregnating melt as opposed to the primary composition of the host peridotite (Cannat et al., 1990); because impregnating melts are generally richer in Ti relative to the latter, spinel equilibrating with the latter should have higher Ti concentrations than non-reacted peridotites (Pearce et al., 2000). Plots of spinel Cr# vs. spinel Mg# (Fig. 12a), Cr# vs. spinel TiO2 (Fig. 12b) and spinel TiO2 vs. co-existing olivine NiO content (Fig. 12c) demonstrate that plagioclase–wehrlite spinels (and hence plagioclase–wehrlite protoliths) underwent extensive melt–rock interactions with a Ti-rich melt subsequent to having undergone ~20–23% partial melting. The high TiO2 and wide range in concentration of TiO2 (0.09–0.66) at relatively constant Cr# (43–48) and olivine NiO (0.18–0.25 wt.%) exhibited by wehrlite spinels indicate that the metasomatising melts were similar in composition to secondary MORB-like melts that interacted with Alpine and abyssal peridotites (Fig. 12b and c). In contrast to wehrlite spinels, those of the clinopyroxenites have relatively low Ti and a limited range in concentrations (0.01–0.18 wt.% TiO2) and thus display little evidence of melt–rock reaction (Fig. 12b and c). Clinopyroxenite spinels also display a trend towards decreasing NiO (0.29–0.06) at relatively constant TiO2 (Fig. 12c) which is consistent with fractional crystallization based on the modeling of a dunitic assemblage of olivine (95%) + spinel (5%) (fractional crystallization expression of Gast (1968) applied to an initial liquid with 15 wt.% MgO, 0.6 wt.% TiO2 and 400 ppm Ni; see Fig. 12 caption and Marchesi et al. (2009) for further details).

6.1.6. Tectonomagmatic affinity of the reacting melt and inference of reacting mantle composition Although Ti/Fe 3# (Fe 3/Fe 3 + Cr + Al) ratios (Dare et al., 2009) of peridotite spinels are mostly independent of melt–rock reactions, the Ti/Fe 3# can increase during interactions between mantle and an evolving MOR melt as exemplified by melt–rock interactions recorded by peridotite at the Hess Deep (Allan and Dick, 1996; Arai and Matsukage; 1996; Edwards and Malpas, 1996). Spinel Fe 3+# will remain constant but Ti/Fe 3+# will necessarily increase because (e.g., MORB-like) melts normally contain a higher concentration of Ti than the host peridotite and hence host peridotite spinels should record a post-equilibration increase in Ti-content (Pearce et al., 2000) as explained above. Thus, the Ti/Fe 3# ratio of peridotite spinel serves as an effective gauge with which to measure melt–rock reactions at MOR and SSZ settings (Dare et al., 2009). AUC wehrlite spinels have an identical mean Fe3+ concentration (0.06) as SSZ spinels (e.g., harzburgite spinels of the Torishima forearc seamount, Izu–Bonin forearc, Parkinson and Pearce, 1998; Dare et al., 2009) which are significantly more refractory than those of abyssal harzburgite spinels (Fe3+ = 0.03, e.g., Hess Deep, Dare et al., 2009). We therefore expect AUC wehrlite spinel to record a steeper trend (i.e., higher Ti/Fe3+# ratios) in Ti/Fe3+# space than MORB harzburgite spinel if reaction occurred between SSZlike mantle and a MORB-like melt. Indeed, Fig. 13a illustrates that the trend of AUC wehrlite spinels follow a steeper path in Ti/Fe3# space with relatively higher Ti/Fe3+# ratios. Thus, in addition to spinel Cr#/Mg# relations (Fig. 12a), which require that the protolith of the plagioclase– spinel wehrlite underwent approximately 20–23% partial melting, we interpret that the wehrlite protolith was probably a refractory and residual harzburgite. That the reacting melts were ‘MORB-like’ as opposed to ‘SSZ-like’ (i.e., island arc tholeiitic or boninitic) is confirmed on Fig. 13b which also corroborates the estimations of relative degrees of partial melting undertaken by the precursor plagioclase–wehrlites and clinopyroxenites (~20–23% and 12–15%, respectively) on the basis of spinel Mg#/Cr# relations (Fig. 12a). Via extrapolation of these trends from the fields which define the particular tectonomagmatic affinity of the reacting melt back to the melting curve it is possible to ascertain the approximate composition of the precursor mantle (Pearce et al., 2000). In addition to the high degree of partial melting undertaken by the AUC plagioclase–wehrlite protoliths and the intersection of the spinel compositional trend with the region on the melting curve defined by harzburgite (Fig. 13b), we again interpret that the wehrlite protolith was a harzburgite. Similarly, the relatively lower degree of partial melting recorded by clinopyroxenite spinels (i.e., ~ 12–15%) indicates a more fertile, i.e., lherzolite protolith. These observations

612

S.A. Whattam et al. / Lithos 127 (2011) 599–618

Fig. 9. Plots of olivine Fo content vs. (a) spinel Cr# and (b) olivine NiO wt.% for AUC peridotites and pyroxenites. Olivine spinel mantle array (OSMA) and fields for abyssal, passive margin and supra-subduction zone (SSZ) peridotites in (a) are from Arai (1994) and Arai et al. (1998; FC stands for fractional crystallization). The mantle olivine array in (b) is from Takahashi et al. (1987). The fine dotted curves in (b) are olivine crystallization trends of Sato (1977) (A = fractional, B = equilibrium) and the replacement trend represents the compositional trend defined by replacive olivine formed via open-system melting between harzburgite and infiltrating melts of evolved composition (Niida, 1997). Plotted for comparison on (a) are the compositions of peridotite spinels and coexisting olivines from various Korean ultramafic complexes. Fields for these complexes are as follows: (1a) Bibong lherzolites and (1b) harzburgites (Oh et al., 2010); (2a) Yugu lherzolites and websterites and (2b) harzburgites and dunites (Arai et al., 2008); (3) Ulsan peridotites (Hisada et al., 2008); and (4) Andong serpentinites (Hisada et al., 2008). Also shown for comparison on (a) and (b) are the compositions of olivines and spinels of pyroxenites of the Cabo Ortegal ultramafic complex (light gray triangles) which formed via crystallization and cumulate processes (Santos et al., 2002). Symbols for Andong peridotites and pyroxenites from this study in (a) are the same as in Fig. 8 and those for (b) are as in Fig. 3 and symbols for Baekdong and Bibong peridotites are as in Fig. 5. The larger and smaller symbols in (b) represent core and rim compositions respectively. The data is consistent with a magmatic origin for the Andong ultramafic rocks. Furthermore, the decrease in olivine NiO at nearconstant olivine Fo# in (b) (dashed line) demonstrates that the clinopyroxenites are cumulates. The plagioclase-free wehrlites (black squares) have much a much lower olivine Fo# (Fo84.5) vs. the spinel–plagioclase wehrlites (gray squares, Fo87) at equivalent NiO content which suggests that olivines of the plagioclase-free wehrlites and possibly the websterites (average Fo85.5) and orthopyroxenites (average Fo85.9) may be replacive (solid black line and arrow in (b)), i.e.,the wehrlites represent ‘reactive’ peridotites which reacted with Ti-rich, MORB-like melts that impregnated the plagioclase–wehrlite protoliths (see Fig. 12 and text).

also generally correlate with the SSZ affinity recorded by wehrlite spinels and the MORB affinity recorded by clinopyroxenite spinels (see next section), i.e., SSZ mantle is predominantly harzburgite as opposed to lherzolite. Whole rock bivariate plots of Ni vs. elements that partition into pyroxene, e.g., Sc and Al (Fig. A4) illustrate that serpentinites show increases in Ni with corresponding slight to moderate decreases in Sc and Al, but that the wehrlites on the other hand show very slight to moderate increases in Al with decreasing Ni and variable (i.e., both increasing and decreasing Sc) with decreasing Ni. This suggests that

the reacting melt was saturated in olivine (vector r1, Fig. A4) or olivine + clinopyroxene (vector r2, Fig. A4). As wehrlite samples have and presumably the related serpentinites had very high olivine: clinopyroxene ratios (wehrlites, ~ 75:25, Table 1) and clinopyroxene: orthpyroxene ratios (wehrlites, 20:0), this suggests that reactions entailed orthopyroxene dissolution and olivine ± clinopyroxene recrystallization (i.e., reaction of harzburgitic mantle with melts saturated in olivine + clinopyroxene or olivine). These trends define reactions between clinopyroxene–harzburgite (or lherzolite) (Pearce, 2008) and confirm the intersection of the wehrlite spinel trend with the melting curve in the harzburgite region in spinel TiO2 vs. spinel Cr# space (Fig. 13b). Thus, melt–peridotite interactions as recorded by AUC wehrlites appear consistent with dissolution of orthopyroxenes by infiltrating olivine + clinopyroxene- or olivine-saturated melts that drove subsequent olivine re-crystallization (Arai and Matsukage, 1996; Kelemen, 1990; Kelemen et al., 1995, 1997; Quick, 1981; Suhr et al., 2003). One plagioclase-free wehrlite and the plagioclase wehrlites on the other hand, appear to have recorded net ‘impregnation’ (i1, i2, ‘trends, Fig. A4) in contrast to the serpentinites which record net reaction. This suggests that while the serpentinites underwent reaction entailing orthopyroxene dissolution and olivine re-crystallization that the wehrlites alternatively underwent net impregnation defined by plagioclase ± clinopyroxene crystallization. Another apparent trend on the olivine Fo# vs. olivine NiO plot (Fig. 9b) is the displacement of olivine compositions of the plagioclase-free wehrlites (e.g., AND-P:18 as plotted on Fig. 9b) to lower olivine Fo# (mean Fo84.5) at relatively constant NiO concentrations from the most NiO-rich spinel–plagioclase wehrlites (mean Fo87.2) and the anomalous clinopyroxenite (AND-P:13a, mean Fo87.9). This trend parallels the ‘replacive’ trend (dark gray line and arrow; Niida, 1997) of olivine that forms due to interaction of secondary infiltrating melts with refractory harzburgitic residue. Thus, olivine of the plagioclase-free wehrlites could be of a replacive origin and this, albeit limited data is consistent with the petrogenetic model presented above. A replacive origin may also be supported by the fact that the plagioclase-free wehrlite olivine plotted on Fig. 9b (ANDP:18) is the one which records the highest ‘degree’ of SiO2metasomatism on the basis of bulk rock major element chemistry (Fig. 5) and exhibits the highest degree of Fe-enrichment and MgO depletion (Fig. 11). In such a scenario, the wehrlites and possibly the most refractory serpentinites represent ‘reactive’ peridotite that interacted with secondary melts that also impregnated the plagioclase-bearing wehrlite protoliths. We note also that olivine compositions of the orthopyroxenites and websterites have Fo totals (Fo86–87) intermediate to the spinel–plagioclase wehrlites and clinopyroxenites (Fo87.3–87.9) and plagioclase-free wehrlites (mean Fo84.5). Thus, some of the websterite and orthopyroxenite olivine are possibly replacive, but most websterite olivines appear to follow a fractional crystallization trend from the wehrlites similar to bulk chemistry trends (Figs. 9b and 11).

6.1.7. Petrogenesis summary The most reasonable explanation for the compositional diversity exhibited by the serpentinites and wehrlites is that they represent harzburgites that crystallized from SSZ magmas prior to being metasomatized and possibly impregnated by olivine+ clinopyroxene saturated MORB-like melts in the plagioclase-peridotite stability field. Clinopyroxenites on the other hand, probably represent crystalline segregates deposited by the SSZ melts that crystallized the wehrlites, but subsequently underwent negligible metasomatism relative to the wehrlites. As spinels of the plagioclase wehrlite appear to record a record of intense metasomatism it is possible that many olivines are replacive. Although this is not obvious in thin section or in the field, limited olivine data from plagioclase-free wehrlites also suggests a replacive origin.

S.A. Whattam et al. / Lithos 127 (2011) 599–618

613

Fig.10. Plot of AUC ultramafic and mafic rocks in CaO–Al2O3–MgO space. Shown for comparison are peridotites of the Baekdong and Bibong ultramafic complexes. The AUC wehrlites are significantly more. ‘fertile’ than the AUC serpentinites and are compositionally similar to two of three Baekong peridotites (lherzolites). Note that three of the four AUC serpenitinites on the other hand, possess very low Al2O3 and CaO and plot similar to the Bibong peridotites (harzburgites) near the MgO apex. Fields and plot of peridotites are from Coleman (1977). Symbols are as in Figs. 3 and 5.

6.2. Tectonomagmatic affinities and tectonic implications 6.2.1. Tectonomagmatic affinities of the AUC and relation with other Korean ultramafic complexes As outlined above, various lines of evidence suggest that the AUC ultramafic rocks and gabbros formed within a SSZ below an arc system. On the basis of spinel chemistry, Al2O3 vs. Fe 2+/Fe 3+

variations (Fig. 14a) demonstrate that wehrlite spinels plot within the field defined by Bibong harzburgites and fall completely within the field of SSZ peridotite spinels (Kamenetsky et al., 2001). The pyroxenite spinels plot similarly to those of the Bibong lherzolites and overlap the fields of MORB and SSZ peridotite spinels. Similarly on a plot of Al2O3 vs. TiO2 (Fig. 14b) (Kamenetsky et al., 2001), the clinopyroxenite spinels again overlap the fields of SSZ and MORB

Fig. 11. Plot of bulk rock MgO vs. (a) SiO and (b) FeOt (c) and (d) are the same plots but with addition of AUC pyroxenites and gabbros (symbols are as in Fig. 3). Superimposed are the compositional fields for Alpine–Apennine sub-continental lithospheric melt-impregnated plagioclase–lherzolites (from the south Lanzo peridotite complex) as compiled by Piccardo and Guarnieri (2010) and the field of Siberian Fe-rich lherzolite–wehrlite series xenotiths (LW series xenos) which are interpreted as metasomatized, partially melted residues formed via melt percolation in refractory mantle peridotites (Ionov et al., 2005). Also plotted for comparison are peridotites of the Baekdong and Bibong complexes (Oh et al., 2010; Seo et al., 2005) (symbols are as in Fig. 5). The boxed numbers on the melting trends correspond to polybaric melting at P of (1) 2.5–0.8 GPa and (2) 1.5–0.8 Gpa (Niu, 1997). The thick black arrows in (a) and (b) indicate the trend of metasomatic effects (based on Fig. 5) which are manifest as progressive enrichments in SiO2 and FeOt as metasomatism progressed. However, there must firstly have been a ‘reactive’ trend (thick white arrow) emanating from (hitherto unrecognized) precursor and parental residual, refractory peridotite plotting roughly along the residual melt extraction (or melting) trend. The arrows in (c) and (d) are simple trend lines from the peridotites and serpentinites through the pyroxenites and ultimately the gabbros.

614

S.A. Whattam et al. / Lithos 127 (2011) 599–618

Fig. 12. (a) Spinel Cr# (cationic Cr/Cr + Al) vs. Mg# (cationic Mg/Mg + Fe2+), (b) spinel TiO vs. spinel Cr# and (c) spinel TiO vs. co-existing olivine NiO content variations in AUC wehrlite and clinopyroxenite (symbols are as in Fig. 8). The curve in (a) is from Hirose and Kawamoto (1995) and represents the experimental melting trend and estimation of the degree of partial melting of hydrous lherzolite based on spinel Mg# vs. Cr#. Superimposed on (b) and (c) are the compositional fields of spinels and olivines of (basaltic) meltimpregnated abyssal peridotites and Alpine–Apennine peridotites (b only) both of which are interpreted as the result of peridotite–metasomatic melt interaction. Also superimposed on (c) and represented by the near-vertical, solid black line with small black squares is a fractional crystallization curve modeling the crystallization of olivine + spinel (dunite, with 0.95 olivine and 0.05 spinel from an initial liquid with 400 ppm Ni, 0.6 wt.% TiO and 15 wt.% MgO fractional crystallization expression is from Gast, 1968). Each olivine/melt olivine/melt spinel/melt black square along the curve corresponds to a 2% increment. Constant partition coefficients: DNi = 8.4; DTi = 0.15; DNi = 10 are from Beattie et al. (1991); Kelemen et al. (1993); and Righter et al. (2006), respectively. While the data for wehrlite spinel clearly demonstrate the metasomatic addition of Ti, the data for the clinopyroxenite spinel suggests limited melt–rock interaction and a magmatic origin, i.e., a trend that roughly parallels the fractional crystallization curve in (c).

fields with the majority lying along the margin of the two fields. Wehrlite spinels with the lowest TiO2 and hence the least evidence of metasomatic alteration plot within the SSZ field whereby the highly metasomatically altered spinels (i.e., ones with higher TiO2) plot above the field (Fig. 14b). This also demonstrates that while the least altered samples provide an accurate record of the original tectonomagmatic affinity prior to melt–rock interaction the more altered samples provide instead a record of the secondary metasomatic event. As well, no clinopyroxenite spinels exhibit anomalously high Ti which demonstrates along with other plots presented in Section 6.1.6 that in contrast to wehrlite spinels, those of the clinopyroxenites show little evidence of metasomatic alteration. Overlapping MORB and SSZ affinities is not unusual for SSZ settings and is actually diagnostic of the best-documented ophiolite complexes (e.g., the classical Tethyantype ophiolites of the Eastern Mediterranean–Persian Gulf region) and intra-oceanic forearc systems (e.g. Izu–Bonin Marianas) which characteristically exhibit a chemotemporal evolution from MORB to SSZ (i.e., island arc tholeiite and sometimes boninite) (e.g., see Reagan et al., 2010; Whattam and Stern, 2011). In Section 5 we illustrated that AUC wehrlites and serpentinites exhibit bulk rock major element composition similar to European SCLM peridotites and primitive mantle but generally dissimilar to abyssal peridotites. On the basis of REE systematics, Fig. 15 illustrates a similar theme and also shows that the AUC peridotites are similar in composition to continental peridotites in general and specifically to continental arc peridotites. 6.2.2. Implications for the tectonic evolution of the Korean Peninsula According to a recent tectonic model (Chough and Sohn, 2010), the present-day Gyeongsang Basin comprised an extensive volcanic arc platform in the Early to Late Cretaceous that was coined the

Gyeongsang Volcanic Arc (GVA). The GVA would have been a southern extension of the Japanese Arc system and the Gyenongsang Basin would have been a back arc basin (or back arc rift) behind the volcanic front. If this is accurate, similarly aged arc-related rocks of the Gyeongsang Basin should record appropriate chemistry defined by subduction-modified, arc-like affinities. On the basis of petrological, petrogenetic and chemical data we have documented that the ultramafic and mafic rocks of the AUC are probably of an arc-related origin. However, age constraints for the AUC are lacking and the GVA is but one arc-system candidate for AUC formation. Another candidate is the arc-related system responsible for Early Jurassic granitic batholiths that crop out to the north of the Andong Fault System (AFS) (Fig. 1c; Sagong et al., 2005; Kee et al., 2010). Proximity of the AUC with the Jurassic granitoids and the Cretaceous (or younger?) timing of southward thrusting along the AFS (Choi et al., 2002) may favor formation of the AUC within the Jurassic arc system (Kee et al., 2010). Yet another possibility is the Triassic arc system which is poorly constrained in its distribution but probably responsible for the formation of the Macheon Layered Complex as well as other granitoids in the Yeongnam Massif (Fig. 1). We have demonstrated some similarities between the AUC wehrlite spinels with spinels of the Bibong harzburgites and the compositional similarities of the clinopyroxenite spinels with those of the Baekdong peridotites (Seo et al., 2005) and Bibong lherzolites (Oh et al., 2010) (Figs. 14). Caution needs to be heeded when considering possible relationships of these complexes with the AUC, however, as age constraints are lacking and the Baekdong and Bibong ultramafic complexes are situated some 150 km to the NW of the AUC where they intrude basement rocks of the Gyeonggi Massif. Data from other Korean mafic–ultramafic complexes (the Ulsan serpentinite complex to the SW of the AUC within the Gyeongsang Basin, the Yugu

S.A. Whattam et al. / Lithos 127 (2011) 599–618

615

Fig. 14. Tectonomagmatic classification of Type I and II spinels of Andong peridotites (wehrlites) and clinopyroxenites on the basis of spinel Al2O3 vs. (a) Fe2+/Fe3+ and (b) TiO2 concentrations (Kamenetsky et al., 2001). Shown for comparison on (a) are compositions of spinels from ultramafic rocks of the Baekdong and Bibong ultramafic complexes (Oh et al., 2010). Symbols are as in Fig. 8.

Fig. 13. Plots of (a) spinel Fe3+# [(Fe3+)/Fe3++ Cr + Al] vs. spinel TiO2 (Dare et al., 2009); and (b) spinel TiO2 vs. spinel Cr#. Symbols are as in Fig.8.

peridotite of the Gyeonggi Massif to the NW and the Macheon Layered Complex to the west within the Yeongnam Massif) at present is too scarce for detailed comparison with the AUC; the relation, if any, of these complexes with the AUC will need to await further studies. Nonetheless, we note also that the Yugu peridotite is considered as probable SSZ residual mantle (Arai et al., 2008) and that spinels of the Ulsan peridotites are similar to AUC wehrlite and Bibong harzburgite spinels. Collectively, the data for the Bibong, Yugu and Ulsan peridotite complexes suggest that they represent residual, subductionmodified forearc mantle. The major difference between the AUC and those of the aforementioned Korean ultramafic complexes is that ultramafic rocks of the AUC display petrographic, geochemistry and mineral chemistry consistent with crystallization and accumulation within SSZ magmas and also record evidence of extensive secondary melt–rock interactions. 7. Conclusions 1. The Andong Ultramafic Complex consists predominantly of peridotites (wehrlites ± plagioclase or spinel; or plagioclase + spinel) and related serpentinites with subordinate low-Al pyroxenites (clinopyroxenites, orthopyroxenite, websterites) and gabbros. The wehrlites exhibit bulk rock major element concentrations similar to PM and the wehrlites and pyroxenites are compositionally similar to SCLM peridotites and pyroxenites. 2. Wehrlites formed predominantly by fractional crystallization processes in SSZ magmas. Pyroxenites are consistent with fractionation plus accumulation and the clinopyroxenenites spe-

cifically are interpreted as crystalline segregates deposited in similar magmas that crystallized the wehrlites. 3. The wehrlites and serpentinites are moderately refractory e.g., exhibit Al2O3/SiO2 = 0.01–0.03 and MgO/SiO2 = 0.84–1.02 which indicate derivation from a depleted source. The range to lower MgO/SiO2 values significantly below the terrestrial peridotite reflects secondary metasomatism which caused SiO2 addition. 4. Wehrlites and serpentinites preserve a record of extensive melt– rock interaction. Bulk rock major element compositional trends manifest depletions in SiO2 and enrichments in FeO t and CaO relative to mantle residua of equivalent MgO content and clearly define a progressive metasomatic trend towards higher SiO2 and FeO t and lesser MgO which is opposite to those exhibited by (metasomatically unmodified) refractory peridotites. These trends are identical to those demonstrated by ‘impregnated’ plagioclase peridotites (lherzolites and wehrlites) from both abyssal and subcontinental environments and of metasomatically altered cratonic lherzolites and wehrlites. The clearest support for melt–rock reactions is preserved in wehrlite spinels which show a large range in TiO2 content (0.09–0.66 wt.%) at constant spinel Cr# and olivine NiO content indicative of reaction with a secondary Ti-rich melt. The melts may represent asthenosphere derived MORB-melts that reacted with refractory, serpentinite and wehrlite protoliths and subsequently impregnated the protoliths of spinel–plagioclase wehrlites. 5. Spinel Cr#/Mg# and Cr#/TiO2 relations suggest that protoliths of the wehrlites and clinopyroxenites underwent approximately 20– 23% and 12–15% melting respectively, and that secondary reacting melts were olivine + clinopyroxene (or olivine) saturated and MORB-like. Extrapolation of spinel compositional trends in Cr#/TiO2 space from the MORB field back to the melting curve suggests, in concert with the required high degree of partial

616

S.A. Whattam et al. / Lithos 127 (2011) 599–618

ultramafic–mafic complexes described from South Korea. Age constraints are needed to determine the probable arc system (e.g., the Jurassic one responsible for formation of arc-related granitoids) in which the AUC formed. Supplementary materials related to this article can be found online at doi:10.1016/j.lithos.2011.06.013. Acknowledgements Editor A. Kerr and reviewers S. Arai and O. Parlak are thanked for detailed and constructive comments which greatly aided in revision of the original version of the manuscript. M. Lee, SNU probe technician, is thanked for her assistance in the operation of and acquisition of data from the JEOL 8900 electron microprobe. Y. Lee is thanked for assistance in the field and J.K. Kim for assistance in data manipulation. S.A. Whattam thanks R.J. Stern for comments and ideas related to this study. This work was supported in part by a NRF-MEST grant 20100000296 to M. Cho and a postdoctoral fellowship to S.A. Whattam via the Brain Korea (BK) 21 grant, Seoul National University. References

Fig. 15. REE systematics of AUC peridotites and pyroxenites that demonstrate probable tectonic affinities (symbols are as in Fig. 3). (a) Both La/Sm and Sm/Yb ratios of AUC peridotites are greater than or the same as primitive mantle (white star, McDonough and Sun, 1995) and peridotite compositions suggest an affinity for continental arcrelated peridotites. Similarly, in (b) the primitive mantle-normalized concentrations of Yb vs. Ce/Sm suggest a (continental) sub-arc affinity. In (a) the field of abyssal peridotites is from Niu (2004) and the field of continental peridotites is constructed from Knippa peridotite xenoliths, Ouachita Belt, southern Laurentia (Young and Lee, 2009) and Canadian Cordillera lherzolites (Peslier et al., 2002). In (b), the peridotite fields are from Niu (2004) (absyssal); Bodinier and Godard (2003) (orogenic); Takazawa et al. (2000) (sub-arc, Horoman peridotite); Parkinson and Pearce (1998) (Mariana Arc). Primitive mantle-normalized abundances of Yb and Ce/Sm are from Sun and McDonough (1989).

melting, that wehrlite protoliths were harzburgites. Coupled with the high olivine:pyroxene and clinopyroxene:orthopyroxene ratios of the wehrlites these observations suggest that reactions entailed orthopyroxene dissolution and olivine recrystallization. Although the serpentinites follow trends that suggest ‘net reaction’, i.e. interaction with a melt saturated in olivine or olivine + clinopyroxene, ‘impregnation’ trends are not as convincing for the plagioclase peridotites. However, we note that the plagioclase wehrlites display textures consistent with melt impregnation and that wehrlite plagioclase and clinopyroxene compositions are distinct from those of the pyroxenites. Nevertheless, discrimination of ‘cumulate’ from ‘reaction’ and ‘impregnation’ textures is not obvious in most cases. 6. A feature common to all rock types is primitive mantle-normalized trace element patterns defined by fluid-mobile LILE enrichments, high mobile/immobile element ratios (Sr/Nd, Ba/La and Pb/Ce ≫ 1) and prominent HFSE (Nb, Zr, Ti; e.g., [Nb/Ce]PM-N b 0.1) depletions indicative of formation in a SSZ environment. Spinel data is also consistent with a primary SSZ origin. 7. At present, a lack of geochronological data precludes establishing the genetic relationship, if any, of the AUC with five other

Allan, J.F., Dick, H.J.B., 1996. Cr-rich spinel as a tracer formeltmigration and melt–wall rock interaction in the mantle. Hess Deep, Leg 147. In: Mével, C., Gillis, K.M., Allan, J.F., Meyer, P.S. (Eds.), Proceedings of the Ocean Drilling Program: Scientific Results, 147, pp. 157–172. Arai, S., 1992. Chemistry of chromian spinel in volcanic rocks as a potential guide to magma chemistry. Mineralogical Magazine 56, 173–184. Arai, S., 1994. Characterization of spinel–peridotites by olivine–spinel compositional relationships: review and interpretation. Chemical Geology 113, 191–204. Arai, S., Matsukage, K., 1996. Petrology of gabbro-troctolite–peridotite complex from Hess Deep, equatorial Pacific: implications for mantle–melt interaction within the oceanic lithosphere. In: Mével, C., Gillis, K.M., Allan, J.F., Meyer, P.S. (Eds.), Proceedings of the Ocean Drilling Program: Scientific Results, 147, pp. 135–155. Arai, S., Abe, N., Hirai, H., 1998. Petrological characteristics of the sub-arc mantle: an overview on petrology of peridotite xenoliths from the Japan Arcs. Trends in Mineralogy 2, 39–55. Arai, S., Tamura, A., Ishimaru, S., Kadoshima, K., Lee, Y.-I., Hisada, K.-I., 2008. Petrology of the Yugu peridotites in the Gyeonggi Massif, South Korea: implications for its origin and hydration process. Island Arc 17, 485–501. Beattie, P., Ford, C., Russell, D., 1991. Partition coefficients for olivine-melt and orthopyroxene-melt systems. Contributions to Mineralogy and Petrology 109, 212–224. Blundy, Wood, 1991. Crystal-chemical controls on the partitioning of Sr and Ba between plagioclase feldspar, silicate melts and hydrothermal solution. Geochimica et Cosmochimica Acta 55, 193–209. Bodinier, J.L., 1988. Geochemistry and petrogenesis of the Lanzo peridotite body, western Alps. Tectonophysics 149, 67–88. Bodinier, J.L., Godard, M., 2003. Orogenic, ophiolitic, and abyssal peridotites. Treatise on Geochemistry, vol. 2. Elsevier, Amsterdam, pp. 103–170. Bodinier, J.L., Guiraud, M., Fabriès, J., Dostal, J., Dupuy, C., 1987. Petrogenesis of layered pyroxenites from the Lherz, Freychinede and Prades ultramafic bodies (Ariège, French Pyrenees). Geochimica et Cosmochimica Acta 51, 279–290. Bodinier, J.L., Menzies, M.A., Thirlwall, M.F., 1991. Continental to oceanic mantle transition: REE and Sr–Nd isotopic geochemistry of the Lanzo Lherzolite Massif. Journal of Petrology 191–210. Boudier, F., Nicolas, A., 1986. Harzburgite and lherzolite subtypes in ophiolitic and oceanic environments. Earth and Planetary Science Letters 76, 84–92. Cannat, M., Bideau, D., Hebert, R., 1990. Plastic deformation and magmatic impregnation in serpentinized ultramafic rocks from the Garrett transform fault (East Pacific Rise). Earth and Planetary Science Letters 101, 216–232. Chang, K.H., 1975. Cretaceous stratigraphy of southeast Korea. Journal of the Geological Society of Korea 11, 1–23. Choi, H.I., 1986. Sedimentation and evolution of the Cretaceous Gyeongsang Basin, southeastern Korea. Journal of the Geological Society of London 143, 29–40. Choi, S.Y., Hwang, J.Y., Hwang, J.Y., Kim, J.J., Yoon, J.L., 1990. Studies on mineralogy and geochemistry of Ulsan serpentinites in the Ulsan mine area, Korea. Journal of the Geological Society of Korea 26, 105–118. Choi, P.Y., Lee, S.R., Choi, H.-I., Hwang, J.-A., Kwon, S.-K., Ko, I.-S., An, G.-O., 2002. Movement history of the Andong Fault System: geometric and tectonic approaches. Geosciences Journal 6, 91–102. Chough, S.K., Sohn, Y.K., 2010. Tectonic and sedimentary evolution of a Cretaceous continental arc–backarc system in the Korean Peninsula: new view. Earth-Science Reviews 52, 175–235. Chough, S.K., Kwon, S.T., Ree, J.H., Choi, D.K., 2000. Tectonic and sedimentary evolution of the Korean peninsula: a review and new view. Earth-Science Reviews 52, 175–235. Chun, S.S., Chough, S.K., 1992. Tectonic history of Cretaceous sedimentary basins in the southwestern Korean Peninsula and Yellow Sea. In: Chough, S.K. (Ed.), Sedimentary

S.A. Whattam et al. / Lithos 127 (2011) 599–618 Basins in the Korean Peninsula and Adjacent Seas. : Korean Sedimentology Research Group, Special Publication. Hanlimwon Publishers, Seoul, pp. 60–76. Coleman, R.G., 1977. Ophiolites. Springer-Verlag, New York. 229 pp. Dare, S.A.S., Pearce, J.A., McDonald, I., Styles, M.T., 2009. Tectonic discrimination of peridotites using fO2–Cr# and Ga–Ti–FeIII systematics. Chemical Geology 261, 199–216. Dick, H.J.B., 1989. Abyssal peridotites, very slow spreading ridges and ocean ridge magmatism. In: Saunders, A.D., Norry, M.J. (Eds.), Magmatism in the Ocean Basins: Geological Society of London Special Publications, 42, pp. 71–105. Downes, H., 2001. Formation and modification of the shallow subcontinental lithospheric mantle: a review of geochemical evidence from ultramafic xenolith suites and tectonically emplaced ultramafic massifs of western and central Europe. Journal of Petrology 42, 233–250. Downes, H., 2007. Origin and significance of spinel and garnet pyroxenites in the shallow lithospheric mantle: ultramafic massifs in orogenic belts in Western Europe and NW Africa. Lithos 99, 1–24. Edwards, S.J., Malpas, J., 1996. Melt–peridotite interactions in shallow mantle at the East Pacific Rise: evidence from ODP Site 895 (Hess Deep). Mineralogical Magazine 60, 191–206. Ellam, R.M., Hawkesworth, C.J., 1988. Elemental and isotopic variations in subductionrelated basalts: evidence for a three component model. Contributions to Mineralogy and Petrology 98, 72–80. Franz, L., Wirth, R., 2000. Spinel inclusions in olivine of peridotite xenoliths from TUBAF seamount (Bismark Archipelago/Papua New Guinea): evidence for the thermal and tectonic evolution of the oceanic lithosphere. Contributions to Mineralogy and Petrology 140, 283–295. Frey, F.A., Prinz, M., 1978. Ultramafic inclusions from San Carlos, Arizona: petrological and geochemical data bearing on their genesis. Earth and Planetary Science Letters 129–176. Frey, F.A., Suen, C.J., Stockman, H.W., 1985. The Ronda high temperature peridotite: geochemistry and petrogenesis. Geochimica et Cosmochimica Acta 49, 2469–2491. Gamble, J., Woodhead, J., Wright, I.C., Smith, I.E.M., 1996. Basalt and sediment geochemistry and magma petrogenesis in a transect from oceanic island arc to continental margin arc: the Kermadec–Hikurangi margin, S.W. Pacific. Journal of Petrology 37, 1523–1546. Garrido, C.J., Bodinier, J.L., 1999. Diversity of mafic rocks in the Ronda peridotite: evidence for pervasive melt–rock reaction during heating of subcontinental lithosphere by upwelling asthenosphere. Journal of Petrology 40, 729–754. Gast, P.W., 1968. Trace element fractionation and the origin of tholeiitic and alkaline magma types. Geochimica et Cosmochimica Acta 32, 1057–1086. Godard, M., Lagabrielle, Y., Alard, O., Harvey, J., 2008. Geochemistry of the highly depleted peridotites drilled at ODP Sites 1272 and 1274 (Fifteen–Twenty Fracture Zone, Mid-Atlantic Ridge): implications for mantle dynamics beneath a slow spreading ridge. Earth and Planetary Science Letters 267, 410–425. Gravestock, P.J., 1992. The chemical causes of uppermost mantle heterogeneities. PhD thesis, Open University. Hirose, K., Kawamoto, T., 1995. Hydrous partial melting of lherzolite at 1 GPa: the effect of H2O on the genesis of basaltic magmas. Earth and Planetary Science Letters 133, 463–473. Hart, S.R., Zindler, A., 1986. In search of a bulk-earth composition. Chemical Geology 57, 247–267. Hisada, K.-I., Takashima, S., Arai, S., Lee, Y.I., 2008. Early Cretaceous paleogeography of Korea and Southwest Japan inferred from occurrence of detrital chromian spinels. Island Arc 17, 471–484. Hwang, B.-H., Ernst, W.G., McWilliams, M., Yang, K., 2008a. Geometric model of conjugate faulting in the Gyeongsang Basin, southeast Korea. Tectonics 27. doi:10.1029/2008TC002343. Hwang, B.-H., Son, M., Yang, K., Yoon, J., Ernst, W.G., 2008b. Tectonic evolution of the Gyeongsang Basin, southeastern Korea from 140 Ma to the present, based on astrike-slip and block rotation tectonic model. International Geology Review 50, 343–363. Ionov, D.A., Chanefo, I., Bodinier, J.L., 2005. Origin of Fe-rich lherzolites and wehrlites from tok, SE Serbia by reactive melt percolation in refractory mantle peridotites. Contributions to Mineralogy and Petrology 150, 335–353. Ishimaru, S., Arai, S., Ishida, Y., Shirasaka, M., Okrugin, V.M., 2007. Melting and multistage metasomatism in the mantle wedge beneath a frontal arc inferred from highly depleted peridotite xenoliths from the Avacha volcano, southern Kamchatka. Journal of Petrology 48, 395–433. Ishiwatari, A., 1985. Apline ophiolites: product of low-degree mantle melting in a transcurrent zone rift zone. Earth and Planetary Science Letters 76, 93–108. Jagoutz, E., Palme, H., Baddenhausen, H., Blum, H., Cendales, M., Dreibus, G., Spettel, B., Lorenz, V., Wanke, H., 1979. The abundances of major, minor and trace elements in the Earth's mantle as derived from primitive ultramafic nodules. Geochimica et Cosmochimica Acta 11, 2031–2050. Kamenetsky, V.S., Crawford, A.J., Meffre, S., 2001. Factors controlling chemistry of magmatic spinel: an empirical study of associated olivine, Cr-spinel and melt inclusions from primitive rocks. Journal of Petrology 42, 655–671. Kee, W.-S., Kim, S.W., Jeeong, Y.J., Kwon, S., 2010. Characteristics of Jurassic continental arc magmatism in South Korea: Tectonic implications. Journal of Geology 118, 305–323. Kelemen, P.B., 1990. Reaction between ultramafic rock and fractionating basaltic magma I. Phase relations, the origin of calc-alkaline magma series, and the formation of discordant dunite. Journal of Petrology 31, 51–98. Kelemen, P.B., Shimizu, N., Dunn, T., 1993. Relative depletion of niobium in some arc magmas and the continental crust: partitioning of K, Nb, La and Ce during melt/ rock reaction in the upper mantle. Earth Planet. Sci. Lett. 120, 111–134.

617

Kelemen, P.B., Shimizu, N., Salters, V.J.M., 1995. Extraction of mid-ocean-ridge basalt from the upwelling mantle by focused flow of melt in dunite channels. Nature 375, 747–753. Kelemen, P.B., Hirth, G., Shimizu, N., Spiegelman, M., Dick, H., 1997. A review of melt migration processes in the adiabatically upwelling mantle beneath oceanic spreading ridges. Philosophical Transactions of the Royal Society of London. Series A 355, 283–318. Kim, C.-B., Turek, A., 1996. Advances in U–Pb zircon geochronology of Mesozoic plutonism in the southwestern part of Ryeongnam massif, Korea. Geochemical Journal 30, 323–338. Kim, S.B., Chun, S.S., Chough, S.K., 1997. Discussion on structural development and stratigraphy of the Kyokpo pull-apart basin, south Korea and tectonic implications for inverted extensional basins. Journal of the Geological Society of London 154, 369–372. Klimetz, M.P., 1983. Speculations on the Mesozoic plate tectonic evolution of eastern China. Tectonics 2, 139–166. Korenaga, J., Kelemen, P.B., 2000. Major element heterogeneity in the mantle source of the North Atlantic igneous province. Earth and Planetary Science Letters 184, 251–268. Lee D.-S., 1987. Geology of Korea. Kyohak-Sa Publishing Co., Seoul. Lee, D.W., 1999. Strike-slip fault tectonics and basin formation during the Cretaceous in the Korean Peninsula. Island Arc 8, 218–231. Lenoir, X., Garrido, C.J., Bodinier, J.-L., Dautria, J.-M., Gervilla, F., 2001. The recrystallization front of the Ronda peridotite: evidence for melting and thermal erosion of subcontinental lithospheric mantle beneath the Alboran Basin. Journal of Petrology 42, 141–158. Leroux, V., Bodinier, J.L., Tommasi, A., Alard, O., Dautria, J.M., Vauchez, A., Riches, A.J.V., 2007. The Lherz lherzolite: refertilized rather than pristine mantle. Earth and Planetary Science Letters 259, 599–612. Marchesi, C., Garrido, C.J., Godard, M., Belley, F., Ferré, E., 2009. Migration and accumulation of ultra-depleted, subduction-related melts in the Massif du Sud ophiolite (New Caledonia). Chemical Geology 266, 180–195. Maruyama, S., Isozaki, Y., Kimura, G., Terabayashi, M., 1997. Paleogeographic maps of the Japanese Islands: plate tectonic synthesis from 750 Ma to the present. Island Arc 6, 121–142. McDonough, W.F., Sun, S.S., 1995. The composition of the Earth. Chemical Geology 120, 223–253. Melson, W.G., Vallier, T.L., Wright, T.L., Byerly, G., Nelen, J., 1976. Chemical diversity of abyssal volcanic glass erupted along the Pacific. Atlantic and Indian Ocean sea-floor spreading centres. The geophysics of the Pacific Ocean basin and its margin. American Geophysical Union, Washington DC. pp. 351–368. Morimoto, N., 1989. Nomenclature of pyroxenes. The Canadian Mineralogist 27, 143–156. Mukasa, S.B., Shervais, J.W., 1999. Growth of sub-continental lithosphere: evidence from repeated injections in the Balmuccia lherzolite massif, Italian Alps. Lithos 48, 287–316. Müntener, O., Pettke, T., Desmurs, L., Meier, M., Schaltegger, U., 2004. Refertilization of mantle peridotite in embryonic ocean basins: trace element and Nd isotopic evidence and implications for crust–mantle relationships. Earth and Planetary Science Letters 221, 293–308. doi:10.1016/S0012-821X(04)00073-1. Müntener, O., Manatschal, G., Desmurs, L., Petkke, T., 2010. Plagioclase peridotites in ocean–continent transitions: refertilized mantle domains generated by melt stagnation in the shallow mantle lithosphere. Journal of Petrology 51, 255–294. Nakamura, N., 1974. Determination of REE, Ba, Fe, Mg, Na and K in carbonaceous and ordinary meteorites. Geochimica et Cosmochimica Acta 38, 757–775. Nicolas, A., 1986. A melt extraction model based on structural studies in mantle peridotites. Journal of Petrology 27, 999–1022. Niida, K., 1997. Mineralogy of MARK peridotites: replacement through magma channeling examined from 920D, MARK area. In: Karson, J.A., Cannat, M., Miller, D.J., Elthon, D. (Eds.), Proceedings of the Ocean Drilling Program. : Scientific Results, 153. Ocean Drilling Program, College Station, TX, pp. 265–275. Niu, Y., 1997. Mantle melting and melt extraction processes beneath ocean ridges: evidence from abyssal peridotites. Journal of Petrology 38, 1047–1074. Niu, Y., 2004. Bulk rock major and trace element compositions of abyssal peridotites: implications for mantle melting, melt extraction and post-melting processes beneath mid-ocean ridges. Journal of Petrology 45, 2423–2458. Obata, M., 1980. The Ronda peridotite: garnet–spinel and plagioclase lherzolite facies and the P–T trajectories of a high temperature mantle intrusion. Journal of Petrology 21, 533–572. Oh, C.W., Rajesh, V.J., Seo, J., Choi, S.-G., Lee, J.-H., 2010. Spinel compositions and tectonic relevance of the Bibong ultramafic bodies in the Hongseong collision belt, South Korea. Lithos 117, 198–208. Okada, H., 1999. Plume-related sedimentary basins in East Asia during the Cretaceous. Palaeogeography, Palaeoclimatology, Palaeoecology 150, 1–11. Okada, H., 2000. Nature and development of Cretaceous sedimentary basins in East Asia: a review. Geosciences Journal 4, 271–282. Parkinson, I.J., Pearce, J., 1998. Peridotites from the Izu–Bonin–Mariana forearc (ODP Leg 125): evidence for mantle melting and melt–mantle interaction in a suprasubduction zone setting. Journal of Petrology 39, 1577–1618. Pearce, J.A., 2008. Geochemical fingerprinting of oceanic basalts with applications to ophiolite classification and the search for Archean oceanic crust. Lithos 100, 14–48. Pearce, J.A., Peate, D.W., 1995. Tectonic implications of the composition of volcanic arc lavas. Annual Review of Earth and Planetary Science 23, 251–285. Pearce, J.A., Barker, P.F., Edwards, S.J., Parkinson, I.J., Leat, P.T., 2000. Geochemistry and tectonic significance of peridotites from the South Sandwich arc-basin system, South Atlantic. Contributions to Mineralogy and Petrology 139, 36–53. Pearson, D.G., Davies, G.R., Nixon, P.H., 1993. Geochemical constraints on the petrogenesis of diamond facies pyroxenites from the Beni Bousera peridotite massif, North Morocco. Journal of Petrology 34, 125–172.

618

S.A. Whattam et al. / Lithos 127 (2011) 599–618

Peslier, A.H., Francis, D., Ludden, J., 2002. The lithospheric mantle beneath continental margins; melting and melt-rock reaction in Canadian Cordillera xenoliths. Journal of Petrology 43, 2013–2047. doi:10.1093/petrology/43.11.2013. Piccardo, G.B., Guarnieri, L., 2010. Alpine peridotites from the Ligurian Tethys: an updated critical review. International Geology Review 52, 1138–1159. Piccardo, G.B., Zanetti, A., Müntener, O., 2007. Melt/peridotite interaction in the Lanzo South peridotite: Field, textural and geochemical evidence. Lithos 94, 181–209. Quick, J.E., 1981. The origin and significance of large, tabular dunite bodies in the Trinity peridotite, northern California. Contributions to Mineralogy and Petrology 78, 413–422. Rampone, E., Piccardo, G.B., Vannucci, R., Bottazzi, P., Ottolini, L., 1993. Subsolidus reactions monitored by trace element partitioning: the spinel- to plagioclase-facies transition in mantle peridotites. Contributions to Mineralogy and Petrology 115, 1–17. Rampone, E., Piccardo, G.B., Vannucci, R., Bottazzi, P., Zanetti, A., 1994. Melt impregnation in ophiolitic peridotite: an ion microprobe study of clinopyroxene and plagioclase. Mineralogical Magazine 58A, 756–757. Rampone, E., Piccardo, G.B., Vannucci, R., Bottazzi, P., 1997. Chemistry and origin of trapped melts in ophiolitic peridotites. Geochimica et Cosmochimica Acta 61, 4557–4569. Reagan, M.K., Ishizuka, O., Stern, R.J., Kelley, K.A., Ohara, Y., Blichert-Toft, J., Bloomer, S.H., Cash, J., Fryer, P., Hanan, B.B., Hickey-Vargas, R., Ishii, T., Kimura, J.I., Peate, D.W., Rowe, M.C., Woods, M., 2010. Forearc basalts and subduction initiation in the Izu–Bonin–Mariana system. Geochemistry, Geophysics, Geosystems 11, Q03X12. doi:10.1029/2009GC002871. Righter, K., Leeman, W.P., Hervig, R.L., 2006. Partitioning of Ni, Co and V between spinel structured oxides and silicate melts: importance of spinel composition. Chemical Geology 227, 1–25. Rollinson, J., 2007. Recognising Early Archaen mantle: a reappraisal. Contributions to Mineralogy and Petrology 154, 241–252. Rollinson, H.R., Appel, P.W.U., Frei, R., 2002. A metamorphosed, early Archaean chromitite from west Greenland: implications for the genesis of Archaean anorthositic chromitites. Journal of Petrology 43, 2143–2170. Sagong, H., Kwon, S.-T., Ree, J.-H., 2005. Mesozoic episodic magmatism in South Korea and its tectonic implication. Tectonics 24, TC5002. doi:10.1029/2004TC001720. Santos, J.F., Schärer, U., Gil Ibarguchi, J.I., Girardeau, J., 2002. Genesis of pyroxenite-rich peridotite at Cabo Ortegal (NW Spain): geochemical and Pb–Sr–Nd isotope data. Journal of Petrology 43, 17–43. Sato, H., 1977. Nickel content of basaltic magmas: identification of primary magmas and a measure of the degree of olivine fractionation. Lithos 10, 113–120. Schilling, J.G., Zajac, M., Evans, R., Johnston, T., White, W., Devine, J.D., Kingsley, R., 1983. Petrologic and geochemical variations along the Mid-Atlantic Ridge from 27°N to 73°N. American Journal of Science 283, 510–586. Schubert, W., 1977. Reaktionen im alpinotypen Peridotitmassiv von Ronda (Spanien) und seinen partiellen Schmeltzproduckten. Contributions to Mineralogy and Petrology 62, 205–220. Seo, J., Choi, S.G., Oh, C.W., Kim, S.W., Song, S.H., 2005. Genetic implications of two different ultramafic rocks from Hongseong area in the southwestern Gyeonggi Massif, South Korea. Gondwana Research 8, 539–552.

Shervais, J.W., 1979. Ultramafic layers in the Alpine-type lherzolite massif at Balmuccia, NW Italy. Memorie di Scienze Geologiche, Universita' di Padova 33, 135–145. Shervais, J.W., Mukasa, S.B., 1991. The Balmuccia orogenic lherzolite massif, Italy. Journal of Petrology Special Volume 155–174. Sinigoi, S., Comin-Chiaramonti, P., Demarchi, G., Siena, F., 1983. Differentiation of partial melts in the mantle: evidence from the Balmuccia peridotite, Italy. Contributions to Mineralogy and Petrology 82, 351–359. Song, Y.-S., Kim, D.-Y., Park, K.-H., 2007. The overview of layered structures in mafic– ultramafic Macheon intrusion. Journal of the Petrological Society of Korea 16, 162–179 (in Korean with English abstract). Streckeisen, A., 1976. To each plutonic rock its proper name. Earth-Science Reviews 12, 1–33. Suen, C.J., Frey, F.A., 1987. Origins of the mafic and ultramafic rocks in the Ronda peridotite. Earth and Planetary Science Letters 85, 183–202. Suhr, G., Hellebrand, E., Snow, J.E., Seck, H.A., Hofmann, A.W., 2003. Significance of large, refractory dunite bodies in the upper mantle of the Bay of Islands Ophiolite. Geochemistry, Geophysics, Geosystems 4, 8605. doi:10.1029/2001GC000277. Sun, S.S., McDonough, W.F., 1989. Chemical and isotopic systematic of oceanic basalts: implications for mantle composition and processes. In: Saunders, A.D., Norry, M.J. (Eds.), Magmatism in the Ocean Basins: Geological Society London Special Publication, 42, pp. 313–345. Takahashi, E., Uto, K., Schilling, J.G., 1987. Primary magma compositions and mg/Fe ratios of the mantle residues along Mid Atlantic Ridge 29°N to 73°N. Technical reports of ISEI, Okayama University, A9, pp. 1–14. Takazawa, E., Frey, F.A., Shimuzu, N., Obata, M., Bodinier, J.L., 1992. Geochemical evidence for melt migration and reaction in the upper mantle. Nature 359, 55–58. Takazawa, E., Frey, F.A., Shimuzu, N., Obata, M., 2000. Whole-rock compositional variations in an upper mantle peridotite (Horoman, Hokkaido, Japan): implications for melt segregation, migration and reaction. Geochimica et Cosmochimica Acta 64, 695–716. Van der Wal, D., Bodinier, J.-L., 1996. Origin of the recrystallisation front in the Ronda peridotite by km-scale pervasive porous melt flow. Contributions to Mineralogy and Petrology 122, 387–405. Wee, S.-M., Choi, S.-G., So, C.-S., 1994. Preliminary study on the ultramafic rocks from the Chungnam Province, Korea. Economic and Environmental Geology 27, 171–180 (in Korean with English abstract). Whattam, S.A., Stern, R.J., 2011. The ‘subduction initiation rule’: a key for linking ophiolites, intra-oceanic forearcs and subduction. Contributions to Mineralogy and Petrology. doi:10.1007/s00410-011-0638-z. Woods, M.T., Davies, G.F., 1982. Late Cretaceous genesis of the Kula Plate. Earth and Planetary Science Letters 58, 161–166. Young, H.P., Lee, C.-T.A., 2009. Fluid-metasomatized mantle beneath the Ouachita belt of southern Laurentia: fate of lithospheric mantle in a continental orogenic belt. Lithosphere 1, 370–383. Yumul, G., 2004. Zambales ophiolite complex (Philippines) Transition-zone dunites: restite, cumulate or replacive products? International Geology Review 46, 259–272.