Earth and Planetary Science Letters 304 (2011) 135–146 Contents lists available at ScienceDirect Earth and Planetary Science Letters 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 / e p s l Age and geochemistry of the oceanic Manihiki Plateau, SW Pacific: New evidence for a plume origin Christian Timm a,⁎,1, Kaj Hoernle a, Reinhard Werner a, Folkmar Hauff a, Paul van den Bogaard a, Peter Michael b, Millard F. Coffin c,2, Anthony Koppers d a IFM-GEOMAR Leibniz Institute of Marine Sciences, Wischhofstr. 1–3, 24148 Kiel, Germany Department of Geoscience, The University of Tulsa, Tulsa, OK 74104, USA Atmosphere and Ocean Research Institute, University of Tokyo, Kashiwa-shi, Chiba 277–8568, Japan d College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331–5503, USA b c a r t i c l e i n f o Article history: Received 9 August 2010 Received in revised form 24 January 2011 Accepted 26 January 2011 Available online 20 February 2011 Editor: R.W. Carlson Keywords: Manihiki Plateau oceanic large igneous province 40 Ar/39Ar age dates major and trace element and Sr–Nd–Hf–Pb isotope geochemistry volatiles Greater Ontong Java Event a b s t r a c t We present 40Ar/39Ar age and geochemical (major and trace element and Sr–Nd–Hf–Pb isotope) data from submarine samples recovered from the basement of the Manihiki Plateau during the R/V Sonne research expedition SO193. The samples, predominately tholeiites, with minor occurrences of basaltic andesites and hawaiites, give a mean age of 124.6 ± 1.6 Ma from four different localities on the plateau. Based on TiO2 content, we define two groups of volcanic rocks that differ in trace element and isotopic compositions. Partial melting modeling suggests that the low-Ti group lavas were derived through large degrees of melting (c. 30%) of a peridotitic source at mantle potential melting temperatures of c. Tp = 1510 °C, more than 100 °C above the ambient mantle potential melting temperature. Since the primary water contents of both groups of lavas are low (0.1–0.3g wt.%) and the source is peridotitic, excess temperature is most likely the reason for the large degrees of melting producing the large volume of plateau basalts, consistent with the involvement of a mantle plume. The incompatible element contents of the low-Ti group lavas show a multistage history with enrichment in the most incompatible elements of a previously highly depleted source. They have isotopic compositions similar to enriched mid-ocean-ridge basalt (EMORB) and similar to the common focal zone (FOZO) component. The high-Ti group lavas have more enriched incompatible element compositions overall. Their isotopic compositions tend towards an enriched mantle (EMI)-type endmember, similar, although less extreme, than lavas from the Pitcairn Islands. The geochemistry of the Manihiki Plateau can best be explained by a plume containing three components: 1) a dominant peridotitic FOZO-type component, 2) delaminated EMI-type subcontinental lithospheric mantle (SCLM), and 3) a HIMU (recycled oceanic crustal)-type component possibly in the form of eclogite/pyroxenite. The similarity in age and geochemical composition of Manihiki, Hikurangi and Ontong Java basement lavas, including volcanism in some adjacent basins, suggests that the Greater Ontong Java Volcanic Event covered c. 1% of the Earth's surface with volcanism. © 2011 Elsevier B.V. All rights reserved. 1. Introduction Large Igneous Provinces (LIPs) belong to the most extreme volcanic events on Earth, during which large volumes of volcanic rocks can be produced within a short time period (e.g. Self et al., 2008). Several models have been proposed for the formation of LIPs. Most authors attribute LIP formation to the arrival of a starting-plume head, resulting in extensive melting of upwelling lower mantle in the upper asthenosphere (e.g. Campbell et al., 1989; Campbell, 1998, 2003; Courtillot et al., 2003; Fitton and Godard, 2004; Hauff et al., ⁎ Corresponding author. Tel.: + 64 4 570 4391; fax: + 64 4 570 2600. E-mail address: c.timm@gns.cri.nz (C. Timm). 1 Now at GNS Science, 1 Fairway Dr, Avalon, Lower Hutt, New Zealand. 2 Now at Institute for Marine and Antarctic Studies, University of Tasmania, Hobart, Tasmania 7001, Australia. 0012-821X/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2011.01.025 2000; Hoernle et al., 2010; Ingle et al., 2007; Larson, 1991a,b, 1997; Mahoney, 1987; Mahoney and Spencer, 1991; Mahoney et al., 1993: Tejada et al., 1996, 2002, 2004). Such starting plume heads may broaden laterally to diameters of ~2500 km when they pond at the base of the lithosphere (Griffiths and Campbell, 1991; Richards et al., 1989). High magma production rates occur at the initial stage, which led to the formation of some LIPs within geologically short time scales of several millions of years (e.g. Duncan and Pyle, 1988; Peate, 1997; Renne et al., 1995). Other models include the formation of LIPs through 1) increased melt production by plume–ridge interaction (Mahoney, 1987; Mahoney et al., 1993), 2) upwelling and subsequent extensive melting induced by a meteoritic impact (Ingle and Coffin, 2004; Jones et al., 2002; Rogers, 1982), 3) extension and decompression melting (plate separation model; Anderson, 1996, 2000; Hames et al., 2000; King and Anderson, 1998), 4) enhanced partial melting due to the presence of eclogite (Cordery et al., 1997; Korenaga, 2005; Yasuda et al., 1997), 136 C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 5) accumulation of smaller volcanic terranes by subduction processes (Hoernle et al., 2004a), and 6) delamination of subcontinental lithospheric mantle followed by upwelling and decompression melting (e.g. Anderson, 2005; Hales et al., 2005). Cretaceous LIPs in the Pacific Ocean include the Ontong Java, Hikurangi, and Manihiki plateaus and the Hess, Shatsky and Magellan rises. Several models have been proposed for the formation of the Manihiki Plateau. Winterer et al. (1974) proposed its formation during active rifting in mid-Cretaceous time, near or at the triple junction between the Pacific, Antarctic and Farallon plates, whereas Mahoney and Spencer (1991) invoked the arrival of a plume head beneath the oceanic lithosphere. Beiersdorf et al. (1995) conducted a more detailed study on volcanic rocks from a seamount (“Mt. Eddie”) near Deep Sea Drilling Project (DSDP) Site 317, proposing the involvement of a plume head in the generation of the Manihiki Plateau basement. Larson (1997) instead favored the combination of plume activity and rifting. The presence of a paleo-spreading center, the Osbourn Trough, midway between the Manihiki and Hikurangi plateaus and evidence that the Rapuhia Scarp at the northern margin of the Hikurangi Plateau is a rifted margin, has led to the proposal that the Hikurangi and Manihiki plateaus might have once been connected (Billen and Stock, 2000; Hoernle et al., 2004b, 2010; Worthington et al., 2006). It has also been proposed that the Ontong Java, Hikurangi, and Manihiki plateaus possibly formed as a single mega-plateau (covering ~1% of the earth surface), which shortly after formation broke up forming the separate plateau fragments (Davy et al., 2008; Taylor, 2006). Until the marine expedition with R/V Sonne (SO193), few samples had been recovered from the Manihiki Plateau. Analyzed and/or dated igneous rock samples only exist from: 1) DSDP Site 317 (Hoernle et al., 2010; Jackson et al., 1976; Mahoney and Spencer, 1991), 2) a few dredge locations from the 1900 m high Mt. Eddie seamount and from the base of the Manihiki Atoll (Beiersdorf et al., 1995) and 3) from four locations along the flanks of the Danger Island Troughs (Clague et al., 1976; Ingle et al., 2007). Consequently the age and composition of the Manihiki Plateau are not well constrained. The uppermost 2 km of the Manihiki Plateau crust, however, is exposed along the Danger Island and Suvorov troughs (Fig. 1), and the northern margin of the High and West Plateaus, making the plateau basement accessible to sampling by dredging. During the SO193 expedition, igneous rocks were recovered from 70 basement and seamount sites, making the Manihiki Plateau the most extensively sampled submarine plateau to date. Here we present new 40Ar/39Ar age (from 5 samples) and geochemical (major and trace element and Sr–Nd– Hf–Pb isotope from 17 samples) data from volcanic rocks recovered from seven Manihiki basement sites during SO193 (Fig. 1), in order to improve our understanding of the temporal and geochemical evolution and ultimately the origin of the Manihiki LIP and test the existing models for the formation of LIPs. Data from the remaining locations, which contained younger, primarily alkalic, volcanism will be published separately. 2. Geological and tectonic setting The Manihiki Plateau is located in the SW Pacific extending from ~3°S to ~16°S and from 159°W to ~169°W. It covers an area of ~770,000 km2 (Coffin and Eldholm, 1994) and has an estimated overall volume of 8.8 million km3 (Eldholm and Coffin, 2000). The plateau is bordered by the Tokelau Basin to the west, the Samoan Basin to the south, and the Penrhyn and Central Pacific basins to the east and north, respectively. The plateau lies in water depths of ~3000 m, up to 2000 m above the surrounding Cretaceous seafloor, which is located in water depths of ~4000–5000 m (Fig. 1). The crustal thickness of the Manihiki Plateau is believed to range between 15 and 25 km (Hussong et al., 1979; Mahoney and Spencer, 1991; Viso et al., 2005). Numerous seamounts and some of the westernmost Cook Islands (e.g. the Manihiki and Rakahanga atolls and the Danger Islands) are scattered across the entire plateau. Based on geophysical surveys from the 1960s and early 1970s, Winterer et al. (1974) subdivided the Manihiki Plateau in three major morphological units: (1) the ‘High Plateau’ in the east, (2) the ‘North Plateau’, and the (3) the ‘Western Plateaus’ (Fig. 1). Deep fault systems, which are considered to be failed rift systems (e.g. Hoernle et al., 2010; Larson et al., 2002; Mahoney and Spencer, 1991; Viso et al., 2005), separate these units. The N–S trending Danger Island Troughs (named after the atolls at its southern end) separate the High from the Western Plateaus. They consist of three en echelon fault-bounded (up to ~6200 m deep) basins. The Danger Island Troughs bifurcate south of 10°S into the Suvorov Trough in the east and the southern Danger Island Troughs in the west. A broad basin separates the High and the North plateaus. The High Plateau encompasses ~400,000 km2 above the 4000 m contour interval and represents the largest and shallowest morphological unit of the Manihiki Plateau. The center of the High Plateau lies in average water depths of ~2500 to 3000 m. According to Winterer et al. (1974), acoustic basement of the plateau has a flat relief and is covered by ≥1 km pelagic and/or volcaniclastic sediments. The Western Plateaus represent the second largest morphological unit of the Manihiki Plateau, encompassing ~250,000 km2 above the 5000 m contour interval, located in average water depths of ~3500 to 4000 m. The North Plateau, encompassing ~60,000 km2 above the 4500 m depth, forms the smallest portion of the Manihiki Plateau and is as shallow as ≤1500 mbsl. The central and western portions are characterized by rough topography, which becomes less pronounced as a result of being covered by ≤1 km of sediment (Winterer et al., 1974). Seamounts and atolls are concentrated in the marginal parts of the Manihiki Plateau. One K/Ar age of 106.0± 3.5 Ma was determined on an igneous basement sample drilled at DSDP Site 317 (Lanphere and Dalrymple, 1976), which is significantly younger than the directly overlying sediment ~116 Ma (Sliter, 1992). Two 40Ar/39Ar ages from tholeiitic basalts from DSDP Site 317 of 116.8±3.7 and 116.4 ±5.1 Ma (Hoernle et al., 2010) agree well with the age of the overlying sediment and are within error of an 40Ar/39Ar age of 117.9 ±3.5 for a tholeiitic sample dredged from the Danger Island Troughs (Ingle et al., 2007). The ages of these tholeiitic samples and an 40Ar/39Ar age of 99.5±0.7 Ma for an alkalic sample from the Danger Island Troughs also suggest that at least two episodes of volcanism are recorded on the Manihiki Plateau (Ingle et al., 2007). Three samples from Mt. Eddie seamount on the High Plateau yielded K/Ar total fusion ages of 81.6, 75.2 and 75.1 Ma, providing additional evidence for a late alkalic stage (Beiersdorf et al., 1995). During plateau formation in Early Cretaceous time, the seafloor drilled at DSDP Site 317 was located at shallow water depths of 200–300 m (possibly even subaerial), based on geochemical, sedimentological, and paleontological evidence in the deposited volcaniclastic material (Ai et al., 2008; Clift, 2005; Ito and Clift, 1998; Jenkyns, 1976). Shortly after the plateau was emplaced, it is believed that the Tongareva triple junction initiated, causing extension together with mantle upwelling, causing incipient rifting along the Danger Islands and Suvorov troughs at about 120–118 Ma (Larson et al., 2002; Viso et al., 2005). Shortly thereafter, renewed spreading formed the present eastern margin (Manihiki Scarp) and possibly the southern margin as a result of asymmetric spreading between the Manihiki and Hikurangi plateaus (Billen and Stock, 2000; Davy et al., 2008). A zircon age of 115 Ma for the seafloor exposed at the Wishbone Scarp just north of the Hikurangi Plateau indicates that the breakup of the Manihiki and Hikurangi plateaus at the Osbourn Trough spreading center happened at ~116 Ma, shortly after formation of these plateaus (Mortimer et al., 2006). After formation and ensuing breakup, the Manihiki Plateau cooled and subsided. Lithospheric subsidence, however, was less than that of normal seafloor (Clift, 2005). 3. Results 3.1. 40 Ar/39Ar dating A detailed list of the 40Ar/39Ar age data is presented in Table 1 (analytical method; age spectra and analytical data are in Supplementary file 1–4). All errors reported in this paper are stated as 2σ. C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 137 Fig. 1. Bathymetric map of the Manihiki Plateau. Black and white dots represent sample locations where the plateau basement has been recovered, including the location of DSDP Site 317. White diamonds are sample locations reported in Clague et al. (1976) and Ingle et al. (2007). Numbers beneath some sample numbers are the respective 40Ar/39Ar ages DSDP site 317 ages are taken from Hoernle et al. (2010). Lower left inset map shows the general location of the Manihiki Plateau. OT = Osbourn Trough. Our six new 40Ar/39Ar ages of plateau phase volcanic rocks from four different locations from the Manihiki Plateau range from 126.0 ± 1.5 to 122.9 ± 1.6 Ma and are within error of each other. Glass sample DR52, dredged along the northern margin of the High Plateau, produced 40Ar/39Ar step-heating ages of 126.0 ± 1.5 (IFM-GEOMAR geochronology laboratory) and 123.8 ± 0.8 Ma (Oregon State geochronology laboratory). Ages of 124.5 ± 1.5 Ma and 122.9 ± 1.6 Ma (both IFM-GEOMAR) were obtained on glass samples dredged along the eastern flank of the southern Danger Island Troughs (at location DR26). Feldspar step-heat ages of 125.2 ± 8.3 Ma (DR18) and 125.0 ± 2.1 Ma (DR46) were obtained from samples from the lower eastern flank of the Suvorov Trough and from a volcanic structure in the basin between the North and High plateaus, respectively. In summary, our new 40Ar/39Ar data set indicates that the main part of the Manihiki Plateau basement may have formed within a geologically short period of intense volcanism at 124.6 ± 1.6 Ma. 3.2. Geochemical characteristics of the plateau-phase lavas Major element compositions (normalized to 100% on a volatile-free basis) of volcanic rocks from the plateau basement range from tholeiitic basalts to basaltic andesites (after Le Maitre et al., 2002; except for two samples with MgO b 9.2 wt.% that extend to lower SiO2, suggesting clinopyroxene and plagioclase in addition to olivine fractionation in these samples). Relatively low loss of ignition (LOI b 2.5 wt.%) and P2O5 (b0.13 wt.%) suggests that the influence of seawater alteration on the chemistry of the selected whole rock samples was minor. Nevertheless, highly variable ratios of some large ion lithophile elements (LILEs) to rare earth elements (REEs) and high field strength elements (HFSEs) (e.g. (K, Rb)/Yb and U/(Nd, Nb)) in the whole rock samples suggest that some of the LILEs have been affected by seawater alteration. The glasses, on the other hand, do not show evidence of mobilization of LILEs. We therefore concentrate on immobile trace elements in this study, which 138 C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 Table 1 Results of step-heating 40 Ar/39Ar analyzes of the Manihiki basement. Sample Group Run/Lab No. Plateau age (Ma) ± 2s MSWD Probability % 39Ar in plateau No. of steps Dated material and type of analysis SO193 SO193 DR52–2 High-Ti 1.07 2.00 1.02 1.02 0.93 48.3 64.6 12 to 19 9 to 19 Glass step-heating Glass step-heating High-Ti 71.0 76.0 13 to 24 6 to 16 Glass step-heating Glass step-heating SO193 SO193 SO193 SO193 Low-Ti High-T Low-Ti Low-Ti 127.1 ± 2.6 125.5 ± 1.8 126.0 ± 1.5 124.2 ± 0.9 123.7 ± 0.9 123.8 ± 0.8 125.2 ± 8.3 125.0 ± 2.1 124.5 ± 1.5 122.9 ± 1.6 0.38 0.03 0.31 SO193 DR52-2a 1st /52-2gls 2nd /52-2gl2 Wtd. Mean 1st/3 G1-10 2nd/3 G2-10 Wtd. Mean DR18-4Bfs2 DR46-1fss DR26-1gls DR26-7gls 1.06 2.10 1.50 0.57 0.39 0.05 0.10 0.84 92.3 58.9 90.2 80.3 3 to 19 12 to 18 5 to 19 8 to 18 Feldspar step-heating Feldspar step-heating Glass step-heating Glass step-heating DR18-4B DR46-1 DR26-1 DR26-7 a University of Oregon ages. Bold ages are used in the text and figures. Number of Analyses 11 Manihiki (FOZO) Manihiki (EMI) Ontong Java (FOZO) Ontong Java (EMI) Hikurangi 9 7 tions (less radiogenic Nd, Hf, and Pb, but more radiogenic Sr isotope ratios than the low-Ti rocks; Figs. 4 and 5a–d, Table 2), confirming that their differences in TiO2 cannot solely be explained by fractional crystallization. Manihiki basement glasses have relatively low and variable H2O contents and H2O/Ce ratios similar to MORB (Michael, 1995) and OJP glasses (Michael, 1999; Roberge et al., 2004). The high-Ti glass (DR522) has low H2O/Ce like some EM1-type glasses (e.g., Dixon et al., 2004), while the low-Ti glasses have higher H2O/Ce (400) like some MORB and OJP glasses. Dissolved CO2− 3 contents in these same glasses show that they erupted deep enough below the sea surface to prevent H2O loss by degassing: thus the low H2O contents and ratios are believed to be primary. The primary water contents of the Manihiki 100 a) Ontong Java (Kwaimbaita) Ontong Java (Singgalo) Hikurangi 10 Rock/PRIMA (Hofmann, 1988) generally correlate well with TiO2 and each other, show similar compositions to related glass samples and show smooth patterns on multi-element diagrams (e.g. Figs. 3 and 4; Supplementary Table 1). Based on TiO2 content, the volcanic rocks can be grouped into a lowTi (TiO2 b 0.90 wt.%; MgO = 2.3–13.7) and a high-Ti group (TiO2 0.90 wt.%; MgO= 3.3–9.1). Fractional crystallization processes cannot explain the difference in TiO2 at the same MgO. At a given MgO content, SiO is generally higher and FeOt, Na2O, and P2O5 are mildly to moderately incompatible elements (e.g. heavy (H) and middle (M) rare earth elements (REE), Zr, Hf and Y) than in the high-Ti group and average N-MORB Th, Nb, Ta and La abundances than the high-Ti group (Fig. 3a). Whereas the U-shaped multi-patterns of the low-Ti group lavas are distinct from N-MORB, the high-Ti group lavas show relatively flat multi-element patterns between normal (N) and enriched (E) MORB lavas (Fig. 3a). In the glass samples, the large ion lithophile element (LILE; Rb, Ba, U, and K) within the low- and high-Ti group lavas is generally similar. In addition, moderately to less incompatible trace element ratios (e.g. (Nd, Sm)/Yb, (Tb/Yb)N and Zr/Y; Fig. 4a–e) are higher in the high-Ti group. The Manihiki basement samples form reasonably good correlations on radiogenic isotope correlation diagrams, suggesting that alteration has not destroyed the primary isotopic signatures of the samples. The low-Ti group samples overlap the isotopic range of the common focal zone (FOZO) component after Hauri et al. (1994) (FOZO A) and Stracke et al. (2005) (FOZO B). Their compositions are also similar to the dominant Kwaimbaita/Kroenke-type lavas from the Ontong Java Plateau (e.g. Mahoney et al., 1993; Tejada et al., 1996, 2002) and the Hikurangi Plateau A lavas (Hoernle et al., 2010), but extend to less radiogenic Sr and more radiogenic Pb isotopic compositions or towards more HIMU-like compositions (Fig. 5a–d). The high-Ti group, similar in composition to the less common Singgalo lavas, generally has more enriched mantle one (EM1)-type isotopic composi- Osbourn Seamounts 1 0.1 Ontong Java (Kroenke) Low-Ti group lavas (wr) High-Ti group lavas (wr) Low-Ti group (Ingle et al., 2007) High Ti group (Ingle et al., 2007) Th Nb La Ce Pr Nd Sm Hf Zr Eu Gd Tb Dy Ho Y Er Tm Yb Lu 100 b) EMORB 10 NMORB 1 5 Low-Ti group lavas (gl) High-Ti group lavas (gl) 3 0.1 1 80 90 100 40Ar/39Ar 110 120 130 Age (Ma) Fig. 2. 40Ar/39Ar ages of Manihiki (medium gray), Ontong Java (light gray) and Hikurangi (black) basement lavas, including an age from the Danger Island Troughs (Ingle et al., 2007) and two ages from DSDP Site 317 (Hoernle et al., 2010). Ages from the Ontong Java Plateau are from Mahoney et al. (1993), Tejada et al. (1996) and Tejada et al. (2002) and those from the Hikurangi Plateau are from Hoernle et al. (2010). Samples from Manihiki (high-Ti group) and Ontong Java (Singgalo group), which have enriched (EMI-type) compositions, are marked with a cross. Rb Th Nb K Ce Pr Sr Sm Zr Eu Tb Ho Er Yb Ba U Ta La Pb Nd P Hf Ti Gd Dy Y Tm Lu Fig. 3. a) Immobile incompatible element patterns on a multi-element diagram of the Manihiki basement lavas normalized to primitive mantle (Hofmann, 1988). Additional Manihiki data (black and white diamonds) are from Ingle et al. (2007). The dark gray field shows the range in primitive-mantle-normalized whole rock incompatible element abundances of the Hikurangi Plateau, whereas the medium gray field shows the field for the incompatible element contents of the Ontong Java Plateau. (Fitton and Godard, 2004; Mahoney et al., 1993; Tejada et al., 2002, 2004 and from Hoernle et al., 2010). E- and N-MORB (after Sun and McDonough, 1989) are added as reference. b) Primitive mantle normalized multi-element diagram of the glass samples. 139 C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 1.4 a) 8 c) r2 1.2 = 0.85 4 0.8 r2 = 0.73 0.6 2 0.4 b) εNd(t) (Tb/Yb)N 6 1.0 15 d) r2 = 0.61 13 3 9 2 εHf(t) Zr/Y 11 7 1 0 0.0 2 r = 0.79 (ZrXRF) 5 3 0.5 1.0 1.5 2.0 2 r = 0.78 20 Low-Ti group lavas (FOZO) High-Ti group lavas (EMI) 19 Low-Ti group lavas (gl) (t) High-Ti group lavas (gl) 206Pb/204Pb TiO2 (wt %) e) 18 LT group literature data HT group literature data 0.0 0.5 1.0 1.5 2.0 2.5 TiO2 (wt %) Fig. 4. TiO2 (wt%) correlates well with a) (Tb/Yb)N, b) Zr/Y, c) εNd(t), d) εHf(t) and e) 206Pb/204Pb (t) in the Manihiki basement rocks, where N denotes normalized to primitive mantle and t the initial isotopic composition at the age of sample formation. The low-Ti group (LT) lavas, compared to the high-Ti rocks (HT), have lower (Tb/Yb)N and Zr/Y but higher εNd(t), εHf(t) and 206Pb/204Pb (t). Additional Manihiki data are from Ingle et al. (2007), Mahoney and Spencer (1991) and Hoernle et al. (2010). Light gray rectangles are data from the Ontong Java Plateau (from the GEOROC database and Fitton and Godard, 2004; Mahoney et al., 1993; Tejada et al., 1996, 2002, 2004). glasses are low and fall into the range for the Kwaimbaita/Kroenke glasses from Ontong Java (0.1–0.3 wt.%). This is surprising for DR52, which has an isotopic composition more similar to the Singgalo lavas from Ontong Java. The Singgalo glasses have higher primary water contents of 0.4–0.5 wt.%. In addition, the low and high-Ti group Manihiki glasses have lower S contents (c. 0.05 wt.%) than glasses from the Ontong Java Plateau (0.09–0.11 wt.%; Roberge et al., 2004) and MORB (e.g. Wallace and Carmichael, 1992). 4. Discussion 4.1. Similar ages and geochemistry of the Manihiki, Ontong Java and Hikurangi Plateau basements The Manihiki Plateau basement volcanic rocks have similar ages and geochemical compositions to the Ontong Java and Hikurangi Plateau fragments. Our new 40Ar/39Ar age data from four new Manihiki basement locations yield an age of 124.6 ± 1.5 Ma (Fig. 1). The age of 117.9 ± 3.5 Ma (2σ) determined from a sample from the Danger Island Troughs (Ingle et al., 2007) and the DSDP Site 317 sample with an age of 116.4 ± 5.1 Ma (Hoernle et al., 2010) overlap the youngest age in this study within error. The age of 116.8 ± 3.7 Ma from DSDP Site 317 lies slightly outside of the 2σ error for the youngest sample dated in this study at 122.9 ± 1.6 Ma, suggesting that the uppermost part of the plateau drilled at Site 317 may be slightly younger than basement lavas dredged at from the Danger Islands and Suvorov troughs and the northern margins of the plateau. Taken together the ages for the Manihiki basement localities fall in the range of 117–126 Ma. This age range overlaps with the main pulse of activity at Ontong Java ranging from 119 to 129 Ma (Mahoney et al., 1993; Tejada et al., 1996, 2002) and includes the oldest age from the Hikurangi Plateau of 118 ± 4.0 Ma (Hoernle et al., 2010), consistent with all three plateaus having formed contemporaneously with a peak in activity between 121 and 125 Ma (Fig. 2). Despite the large extent of the Ontong Java Plateau, samples from the basement analyzed thus far form three very restricted geochemical groups: 1) the dominant Kwaimbaita-type lavas, found throughout the plateau, which are characterized by relatively flat incompatible element patterns on multi-element diagrams and isotopic compositions similar to E-MORB, 2) the Kroenke group lavas that are characterized by higher MgO contents and lower incompatible element abundances than the Kwaimbaita group lavas but have similar incompatible element and isotopic ratios and have been interpreted as parental to the Kwaimbaita group lavas (and thus these two groups will be referred to collectively henceforth), and 3) the Singgalo group lavas that have more enriched incompatible element characteristics and more EMI-type isotopic compositions (e.g. Fitton and Godard, 2004; Tejada et al., 1996, 2002, 2004). The Hikurangi Plateau lavas (group A) have very similar geochemical characteristics to the Kwaimbaita/Kroenke group lavas from Ontong Java, whereas one sample (group B) has a composition similar to the Singgalo group lavas (Hoernle et al., 2010). As noted above, the Manihiki basement rocks also form two distinct geochemical groups: 1) a low-Ti group with isotopic compositions ranging from the Kwaimbaita/Kroenke type lavas from Ontong Java (and Hikurangi) to HIMU or FOZO (B)-like compositions, and 2) a high-Ti group with isotopic compositions similar to the Singgalo lavas from Ontong Java with more EM1-type compositions than the low-Ti group lavas. The similarity in age and geochemical composition of the Manihiki, Ontong Java and Hikurangi Plateaus is consistent with these plateaus having been formed by the same event and as a single mega-plateau (e.g. Taylor, 2006; Davy et al., 2008), but does not require formation as a single plateau. 140 C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 12 FOZO A (t) 0.703 b) 15.5 Ptc (t) Ptc (t) a) Pacific MORB (t) 0.705 0.704 87Sr/86Sr 20 EMI (t) Hikurangi Seamounts (t) d) FOZO B (t) εHf(t) 5 r eA tl an OS (t ) Pacific MORB (t ) Hikurangi Seamounts (t) ray M 0 Ptc (t) Osbourn Seamounts (t) Pacific MORB (t) EMI -1 38 HIMU t) c( Pt -5 -3 EMI 1 3 5 7 9 11 17 18 19 Low-Ti group lavas (FOZO) HT group literature data High-Ti group lavas (EMI) OJP-Kwaimbaita/Kroentke High-Ti group lavas (gl) 37 21 206Pb/204Pb(t) εNd(t) Low-Ti group lavas (gl) 20 208Pb/204Pb(t) 39 FOZO A (t) 10 40 HIMU FOZO B (t) FOZO A (t) 15 15.7 15.6 Hikurangi OS (t) Seamounts (t) 0.702 FOZO B (t) 207Pb/204Pb(t) 0 HIMU Osbourn Seamounts (t) Seawater alteration 4 2 Marine Sediments (t) FOZO A (t) 6 15.8 Hikurangi Seamounts (t) FOZO B (t) 8 εNd(t) c) Pacific MORB (t) 10 OJP-Singallo Hikurangi A Hikurangi B LT group literature data Fig. 5. a) Initial 87Sr/86Sr vs. initial εNd, b) initial εNd vs. initial εHf, c) initial 206Pb/204Pb vs. 207Pb/204Pb and d) 206Pb/204Pb vs. 208Pb/204Pb. To minimize the seawater alteration effect the 87Sr/86Sr age correction of the samples DR 26–1 (wh) and DR 47–1 a Rb/Sr ratio of 0.27 was assumed. Additional Manihiki data are from Hoernle et al. (2010), Ingle et al. (2007), Mahoney and Spencer (1991). Pacific MORB is based on data from Meyzen et al. (2007), the Pitcairn (Ptc) field is based on data from Eisele et al. (2002) and the Hikurangi and Osbourn Seamount (OS) fields are based on data from Hoernle et al. (2010). Mantle array in Fig. 5b is as defined in Geldmacher et al. (2003). Also shown are the FOZO-type Kwaimbaita/Kroenke (white rectangles), the EMI-type Singgalo (black rectangles) formations of the Ontong Java Plateau (from Mahoney et al., 1993; Tejada et al., 1996, 2002, 2004) and the FOZO-type Hikurangi A (white triangles) and EMI-type Hikurangi B (black triangles) after Hoernle et al. (2010). MORB data have been corrected for radiogenic ingrowth over 125 Ma assuming 87Rb/86Sr = 0.005, 147Sm/144Nd = 0.25, 176Lu/177Hf = 0.04, μ = 10 and κ = 40. FOZO fields (FOZO A after Hauri et al., 1994 and FOZO B after Stracke et al., 2005) have been age corrected assuming 87Rb/86Sr = 0.015, 147Sm/144Nd = 0.1, 176Lu/177Hf = 0.005, μ = 12 and κ = 80. The Pitcairn data have been age corrected assuming 87Rb/86Sr = 0.01, 147 Sm/144Nd = 0.2, 176Lu/177Hf = 0.03, μ = 8 and κ = 80. Although there is no discernable difference in the ages of the highand low-Ti volcanic rocks on Manihiki, all basement samples drilled at DSDP Site 317 belong to the high-Ti group. Based on the stratigraphy, they formed at the end of the plateau stage of volcanism, which is consistent with the Singgalo Formation lavas in Central Malaita and at DSDP Site 807 on the Ontong Java Plateau being stratigraphically younger than the main Kwaimbaita/Kroenke phase (Tejada et al., 2002). It is, however, noteworthy that the lavas from the Ontong Java and Hikurangi Plateau basements with ages younger than 116 Ma have Kwaimbaita/Kroenke (low-Ti) type compositions, rather than Singgalo (high-Ti) type compositions (Fig. 2). Below we will evaluate the geochemical data from the low- and high-Ti group volcanic rocks separately to elucidate the origin of both geochemical groups and to assess which processes were responsible for the excess volcanism that formed the Manihiki Plateau. 4.2. Low-Ti-group volcanic rocks — evidence for an upwelling thermal anomaly (mantle plume) The CaO concentration of very mafic volcanic rocks (MgO N 9 wt.% to minimize the possible effects of fractionation of clinopyroxene and plagioclase) can be used to determine if the melts were derived from peridotitic and/or pyroxenitic sources (Herzberg and Asimow, 2008). The CaO content of the most mafic low-Ti group volcanic rocks from the Manihiki Plateau ranges from 10.3 to 12.6 wt.%, falling into the field of peridotite partial melts defined by Herzberg and Asimow (2008) (Fig. 6a). The sample with low MgO (6.4 wt.%) also has low CaO (8.6 wt.%) but high Al2O3 (16.7), reflecting primarily clinopyroxene fractionation. The low Zr/Hf (b39), Sr/Y (b7), and CaO/Al2O3 (b0.9) further support derivation from a peridotitic source. In accordance with melting pressures N30 kb (see below), the low FeOt contents (8.9– 9.6 wt.%) in the mafic low-Ti volcanic rocks (MgO N 9 wt.%) point to derivation from a depleted peridotitic source similar to mantle peridotite KLB-1 with a forsterite content in the olivine of 89.1 mol% (Hirose and Kushiro, 1993). The fresh glass samples can be used to evaluate the volatile contents of the melts forming the Manihiki basement rocks. The melts for the Manihiki glass samples (high- and low-Ti groups) had low H2O (0.18– 0.26 wt.%), similar to MORB (Michael, 1995) and the Kwaimbaita/ Kroenke lavas from the Ontong Java Plateau (Michael, 1999; Roberge et al., 2004). This is supported by H2O/Ce of 140–430, which overlaps with those from Ontong Java (200–400) and MORB (120–410). The low H2O of the melts indicates a fairly dry mantle source for the Manihiki Plateau rocks, which precludes flux melting to produce the observed large degrees of melting and large volumes of lavas. Since the volatile content of the melts appears to have been low, we use the method of Herzberg and Asimow (2008) to evaluate the temperatures and pressures of melting and the degrees of melting. The primary composition of the three low-Ti group Manihiki tholeiite samples (two whole rock and one glass; which only had olivine on the liquidus) can be calculated by using the Primelts2 program. The calculated primary-melt composition of these lavas can be generated 141 0.283014 0.282947 0.283019 0.282999 – 0.282992 0.282827 0.283095 0.282861 0.282873 a) 0.283488 (7) 0.283238 (53) 0.283132 (5) – 0.283141 (6) – 0.283036 (4) 0.282923 (3) 0.283214 (5) 0.282925 (5) 0.282932 (4) CaO (wt%) L - Ol L - Ol - Cpx - Plag Peridotite Partial Melts +Cpx 3 10 0.47b 0.37 0.58b 1.00b 0.68 0.50 0.78 0.65 0.82 0.87b 0.83 1.03 =1 3.8 1- 0.2 74 Mg O 5 12 % Melt fraction 0 10 20 30 40 b) FeO (wt%) 10 50 4 Gpa 11 L+Ol -Ol +Ol 9 3 GPa 8 -Ol 7 6 0 Measured Calculated 2 GPa 10 20 30 MgO (wt%) Fig. 6. a) Diagram showing MgO vs. CaO. The dashed line marks at MgO= 9 wt.% the onset of clinopyroxene and plagioclase fractionation. High concentrations of CaO in the mafic basement lavas (MgON 9 wt.%) from the Manihiki Plateau are consistent with their derivation from a peridotitic source. Light gray rectangles are data from the Ontong Java Plateau (from the GEOROC database and Fitton and Godard, 2004; Mahoney et al., 1993; Tejada et al., 1996, 2002). Additional data for the high-Ti rocks comes from Hoernle et al. (2010). b) Diagram showing MgO vs. FeOt modified after Herzberg and Asimow (2008). Primary melt compositions of three Low-Ti group lavas (two whole rock and one glass sample) were calculated by using the Primelts2 program. The primary melt composition of these lavas corresponds to c. 30% partial melting, which is broadly consistent with the extent of partial melting for volcanic basement lavas from the Ontong Java Plateau (e.g. Tejada et al., 2002). Furthermore the FeOt concentration relates to melting pressures of c. 3.2 GPa, which corresponds to the melting depth of c. 100 km. Modified after Herzberg and Asimow, 2008. 2.97 1.48 5.25 5.34 5.26 5.25 5.47 4.83 7.38 3.92 7.52 10.5 11.2 11.4 12.3 0.512931 (6) 0.512954 (3) 0.512895 (2) 0.512897 (3) 0.512899 (4) 0.512895 (3) 0.512901 (2) 0.512907 (3) 0.512735 (2) 0.513015 (5) 0.512773 (2) 0.512749 (2) 0.512738 (2) 0.512745 (2) 0.512741 (3) 0.512809 0.512785 0.512776 0.512785 0.512782 0.512782 0.512789 0.512745 0.512564 0.512859 0.512616 0.512587 0.512590 0.512584 0.512590 0.86 0.88 0.84 0.85 0.21 0.09 0.41 0.36 0.46 0.60 0.48 0.52 0.19 0.14 0.19 0.20 0.14 0.05 0.32 0.30d 0.12 0.18 0.26 0.38 5 6 7 SOLIDUS Pi (GPa) -Cpx Ca O 0.23 20.000 (16) 19.498 15.632 (12) 15.608 39.719 (32) 39.080 0.23 0.21 20.311 (1) 19.778 15.659 (1) 15.653 39.920 (2) 39.141 – 20.223 (1) 19.725 15.677 (1) 15.653 40.111 (4) 39.103 0.25 20.310 (1) 19.892 15.679 (1) 15.659 39.921 (2) 39.305 – 20.291 (1) 20.035 15.675 (1) 15.662 39.894 (2) 39.534 18.475 (1) 18.213 15.517 (1) 15.505 38.272 (2) 38.149 0.17 18.802 (1) 17.904 15.501 (1) 15.495 38.224 (4) 38.154 0.32 19.848 (1) 19.325 15.723 (1) 15.697 39.360 (3) 39.136 0.23 18.494 (1) 17.918 15.504 (1) 15.476 38.372 (3) 38.148 18.124 (1) 17.939 15.508 (1) 15.499 38.363 (2) 38.136 0.45 18.120 (1) 17.862 15.509 (1) 15.496 38.355 (2) 38.077 18.273 (1) 17.889 15.524 (1) 15.505 38.424 (2) 38.193 18.270 (1) 17.820 15.529 (1) 15.507 38.424 (2) 38.222 0.45 0.21 0.14 0.05 4 Pyroxenite Partial Melts d c b a Glass samples. Pb measured by laser ablation ICPMS. For age correction an 87Rb/86Sr of 0.27 has been assumed. Assumed. 0.73 0.51 1.27 1.21 1.25 1.20 1.24 1.59 2.56 1.25 2.41 3.47 3.38 3.72 3.76 0.703648 0.703639 0.702559 – 0.702864 0.702769 0.702781 0.703542 0.704611 0.703747c 0.705484 0.704628 0.705657 0.704338 0.704391 0.705574 (6) 0.704100 (5) 0.703348 (5) – 75.6 0.703331 (5) 82.7 0.703277 (2) 83.8 0.703274 (3) 7.44 0.707422 (6) 191 0.704722 (5) 67.5 0.704227 (5) 194 0.705585 (6) 125 0.704723 (6) 185 0.705885 (3) 130 0.704650 (6) 143 0.704526 (5) 125 8.33 125 3.15 125 11.3 125 125 6.87 125 8.18 125 8.05 125 5.61 125 4.16 125 28.1 125 3.82 125 2.30 125 8.19 125 7.88 125 3.76 SO193DR18-1 SO193DR18-4B SO193DR26-1 SO193DR26-1a SO193DR26-2 SO193DR26-3a SO193DR26-10a SO193DR38-2 SO193DR46-1 SO193DR47-2 SO193DR49-1 SO193DR52-1A SO193DR52-2a SO193DR52-3A SO193DR52-3B 22.2 32.8 73.8 176 143 15 0.17 0.27 0.65 – 0.60 – 1.32 1.14 0.67 2.42 2.63 177 204 204 204 204 204 204 144 Age Rb Sample Table 2 Sr–Nd–Hf–Pb isotope ratios. Sr 87 Sr/86Srm 87 Sr/86Sri Sm Nd 143 Nd/ Ndm 144 Nd/ Ndi U Th Pb 206 Pb/ Pbm 206 Pb/ Pbi 207 Pb/ Pbm 207 Pb/ Pbi 208 Pb/ Pbm 208 Pb/ Pbi Lu Hf 176 Hf/ Hfm 177 Hf/ Hfi C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 by c. 30% partial melting of a (garnet–) peridotite at a pressure of c. 3.2 GPa, corresponding to a depth of partial melting of c. 105 km (Fig. 6b). Mantle potential temperatures of 1503 °C (SO193DR18-4B), 1509 °C (SO193DR26-10) and 1516 °C (SO193DR26-13) were calculated for the three samples. These temperatures are about 160 °C above the average ambient mantle potential temperature of 1350 °C, slightly above the average for ocean island basalts (~1480 °C), and similar to average temperatures for LIPs, but about 40 °C less than those calculated for Ontong Java melts (Herzberg and Gazel, 2009). The lower calculated melting temperatures for the Manihiki basement rocks, combined with lower crustal thickness for the plateau, are consistent with the lower overall magma production during the formation of the Manihiki compared to the Ontong Java Plateau or portions of a mega-plateau. In addition to the low total FeO, the moderately to mildly incompatible element abundances and ratios also indicate derivation of the low-Ti group lavas from a depleted source. The low TiO2, Na2O, P2O5 and moderately incompatible trace element contents (e.g. MREE, Zr, and Hf) and the low moderately to less incompatible element ratios (e.g. (Nd, Sm, Tb)/Yb, Zr/Y) of the low-Ti group support a combination of greater dilution through higher degrees of partial melting and/or derivation from a more depleted source compared to the high-Ti lavas and N-MORB (Fig. 4a–b). Despite the evidence for source depletion, the enrichment of the most incompatible elements (e.g. Th, Nb and La), compared to less incompatible elements (e.g. Zr, Hf and the M- and HREE), can be explained by enrichment through small degree melts. Ratios of highly to moderately incompatible elements correlate well with Pb isotope ratios in the low-Ti volcanic rocks: for example, 206Pb/204Pb isotope ratios correlate positively with Th/Zr (r2 = 0.80), Th/Hf (r2 = 0.77), Th/Sm (r 2 = 0.78), Nb/Zr 142 C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 (r2 = 0.82), Nb/Hf (r2 = 0.82), Nb/Sm (r2 = 0.84), La/Zr (r2 = 0.89), La/Hf (r2 = 0.82) and La/Sm (r2 = 0.82) (not shown). The good correlation with these highly immobile incompatible element ratios suggests that the Pb isotope composition of the low-Ti samples, which include four fresh glass samples, cannot be explained by posteruption alteration. In conclusion, the aforementioned correlations indicate that the component causing the source enrichment is not only highly enriched in the most incompatible elements but also has radiogenic Pb isotope ratios and therefore could be a HIMU (high time-integrated μ (U/Pb))-type component. As pointed out by Ingle et al. (2007), the unusual incompatible element characteristics (depletion of moderately relative to mildly incompatible elements and enrichment of highly relative to moderately incompatible elements) of the low-Ti group lavas from the Danger Island and Suvorov troughs (our data) require a multistage source history. Ingle and colleagues attributed the incompatible element characteristics to initial depletion of a mantle wedge of a subduction zone through melt extraction and subsequent introduction of melts from subducted volcaniclastic sediments, which implies that the Manihiki low-Ti volcanic rocks were derived from a distinct source compared to the Ontong Java Kwaimbaita/Kroenke (and Hikurangi A) type rocks. We will discuss an alternative model that can also relate the Manihiki to Ontong Java and Hikurangi sources. On Sr–Nd–Pb–Hf isotope correlation diagrams, the low-Ti group volcanic rocks largely overlap the FOZO-type mantle source fields of Hauri et al. (1994) (FOZO A) and Stracke et al. (2005) (FOZO B), projected back to 125 Ma (Fig. 5a–d), suggesting that FOZO-type lower mantle may be involved in the generation of the low-Ti group lavas. We note, however, that the incompatible element characteristics of the lowTi basalts (very low moderately incompatible element abundances and extreme depletion of the moderately compared to mildly and highly incompatible elements) are distinct from any oceanic basalts for which data has been published thus far (Ingle et al., 2007), including OIBs with isotopic compositions falling within the range of the FOZO A and FOZO B components. Therefore we feel that a FOZO B type source for the Manihiki rocks is unlikely. The only rocks to date that have been found related to the Greater Ontong Java Event that have similarly depleted incompatible element contents and characteristics to the mildly and moderately incompatible elements of the low-Ti group lavas from Manihiki are samples from the Osbourn Seamounts (Tuatara and Moa Seamounts) adjacent to the northwest margin of the Hikurangi Plateau, which are also highly depleted in the highly incompatible elements (Fig. 3a). These rocks have the appropriate isotopic compositions and depleted incompatible element abundances to serve as the depleted endmember for the basement array formed by the Ontong Java and Hikurangi Plateaus and the previously published Manihiki basement data (Hoernle et al., 2010). Despite the extremely depleted incompatible element compositions of the Osbourn Seamount samples, the isotopic compositions are similar to E-MORB or the FOZO A component of Hauri et al. (1994) with more radiogenic Nd, Hf and Pb but less radiogenic Sr isotopic compositions than the Kwaimbaita/Kroenke and Hikurangi A lavas. The less radiogenic Sr and more radiogenic Pb isotopic composition of the lowTi group rocks, which include fresh glass samples and therefore the unradiogenic Sr similar to the radiogenic Pb cannot simply be explained by post-eruption alteration, and could reflect minor addition of a HIMUtype component, similar in composition to late-stage volcanism on the Hikurangi, Ontong Java and Manihiki Plateaus (Hoernle et al., 2008, 2009, 2010; Ingle et al., 2007; Tejada et al., 1996). This HIMU-type component could be present in the source as pyroxenite/eclogite, possibly ultimately derived from recycled oceanic crust. Addition of small degree melts from ancient recycled oceanic crust to a source similar to the Osbourn Seamounts could explain the isotopic and trace element compositions of the low-Ti melts. One possible scenario is that recycled ocean crust (in the form of eclogite/pyroxenite) is contained in a depleted peridotitic matrix with a composition similar to the Osbourn Seamount source. Upon upwelling of such a source, the eclogite (especially if carbonated) would form melts that metasomatize the surrounding depleted peridotitic mantle (e.g. Dasgupta et al., 2007; Timm et al., 2009). Melting of such a source upon further upwelling could generate the low-Ti melts. 4.3. High-Ti-group lavas — contribution of subcontinental lithospheric mantle Now we will discuss the origin of the high-Ti group lavas and their role in the generation of the Manihiki Plateau. Whereas all but one of the low-Ti samples had MgO N 9.3 wt.%, all of the high-Ti samples have MgO b 9.1 wt.%. Fractionation of clinopyroxene and plagioclase is likely to have lowered the CaO and Al2O3 contents of the high-Ti Manihiki samples (Fig. 6a). Therefore, their parental melts are likely to have had higher CaO contents and thus also to have been derived from a peridotitic source. Alternatively the low MgO and CaO could reflect derivation from a pyroxenitic source. The low CaO/Al2O3 (b0.9), Sr/Y (b9.5) and Zr/Hf (b39.5) ratios of all Manihiki basement samples, however, also favor derivation from a peridotitic rather than pyroxenitic/eclogitic source. Therefore both the low- and high-Ti group rocks appear to be derived from melting of peridotitic sources. The differences in trace element and isotopic compositions between the groups, however, require that at least two distinct peridotitic source components were involved in the formation of the plateau basement. The high-Ti group has minor and trace element characteristics similar to DSDP Site 317 samples (e.g. Hoernle et al., 2010; Ingle et al., 2007; Mahoney and Spencer, 1991). The generally flat immobile, primitive-mantle-normalized incompatible element distribution on multi-element diagrams of the high-Ti group volcanic rocks suggests a common mantle source for the widely-distributed dredge samples that is very similar to the source for the Ontong Java, Singgalo and the Hikurangi B basement lavas (Fig. 5). Following Hirose and Kushiro (1993) and Walter (1998), the TiO2 concentration in mafic volcanic rocks mainly depends on the degree of partial melting and/or pressure conditions, behaving like a moderately incompatible element. Thus the high TiO2, combined with high Na2O, P2O5 and moderately incompatible trace element contents (e.g. MREE, Zr, and Hf) and the low moderately to less incompatible element ratios (e.g. (Nd, Sm, Tb)/Yb, Zr/Y), of the high-Ti group lavas could represent lower degrees of partial melting at greater depths and/or derivation from a more enriched source. Generally higher FeOt and lower SiO2 at a given MgO are also consistent with greater depths of melting (Hirose and Kushiro, 1993) in the formation of the high-Ti group lavas. Low ratios of (Th, U, Ba, Rb, Pb, and K)/Zr in the glass sample argue against the addition of a marine sedimentary or fluid-bearing component into the high-Ti group rocks, which is consistent with the low H2O content and H2O/Ce ratio of 140 in the high-Ti glass sample (DR52-2; H2O= 0.19 wt.%). Lack of evidence for a significant role of fluids in the generation of the high-, as well as low-, Ti group lavas excludes flux-melting from being a major melting mechanism for all Manihiki basement lavas. The general origin of the EMI-type source component, however, is controversial (see Lustrino and Dallai, 2003 and Geldmacher et al., 2008 for summaries). In the SW Pacific, the Pitcairn archipelago exemplifies one EMI-type locality, where the EMI-type component has been attributed to incorporation of continentally- derived material containing various proportions of pelagic sediments (e.g. Eisele et al., 2002). The absence of typical depletions in Th, Nb, and Ta in the Manihiki basement lavas argues against the involvement of significant amounts of continental crust (Fig. 3), which is also consistent with the eruption of the Manihiki Plateau through oceanic crust within the Pacific Ocean (which was larger in Cretaceous time; e.g. Mueller et al., 2008) far away from active subduction zones and continents. There is a striking similarity in radiogenic isotopic composition between the high-Ti group Manihiki basement lavas and lavas from Hawaiian volcanoes Mauna Loa and Koolau, which gives valuable insights into the origin of enriched mantle (Fig. 7). C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 12 Pacific MORB (t) 10 εNd(t) 8 6 4 FOZO B (t) Osbourn Seamounts (t) Kilauea Loihi Mauna Loa Koolau FOZO A (t) Hikurangi Seamounts (t) 2 0 17.0 Ptc (t) 18.0 19.0 20.0 206Pb/204Pb(t) Fig. 7. 206Pb/204Pb (t) vs. εNd (t). The low-Ti group lavas from Manihiki overlap the Kwaimbaita/Kreonke group rocks from Ontong Java but extend to more radiogenic Pb isotopic compositions, suggesting involvement of an additional component, not yet observed in the plateau basement lavas from the Ontong Java and Hikurangi Plateaus. Hawaiian lavas (data from the GEOROC database) have been age corrected for radiogenic ingrowths assuming 147Sm/144Nd= 0.202 and μ = 12. Since the Hawaiian lavas are the classical example for lower mantle source plume-related melts, the similarity of Hawaiian lavas and those from the Manihiki/Ontong Java/Hikurangi Plateau suggests a lower mantle source for the lavas of the “Greater Ontong Java Event”. The sub-continental lithospheric mantle (SCLM) composed of residual lherzolite and harzburgite may be the best source for the EMI-type component. Throughout the complex geological history of the SCLM, several geological processes may have produced trace element enrichment and depletion (see Geldmacher et al., 2008 for a summary). Although spinel peridotite xenoliths from continental lithosphere commonly show strongly fractionated incompatible-element patterns (e.g. McDonough, 1990), multiple melting episodes and refertilization of Achaean SCLM (e.g. Griffin et al., 2009) may ultimately have smoothed the trace element distribution. Drilled DSDP Site 527 lavas from the EMI-type Walvis Ridge in the southern Atlantic for which SCLM has been proposed as the possible origin (e.g. Gibson et al., 2005; Salters and Sachi-Kocher, 2010) show similar trace element patterns to the high-Ti Manihiki lavas. Even if the trace element composition is more enriched, large degrees of partial melting during plateau formation may produce relatively flat trace element patterns. Finally it has been demonstrated that depleted subcontinental lithosphere also plays an important role in the EMI-type Koolau lavas (Salters et al., 2006). Delamination could transfer EMI-type SCLM to a zone of neutral buoyancy, such as the 660 km upper–lower mantle or the core–mantle boundary (Elkins-Tanton, 2007; O'Reilly et al., 2009; Timm et al., 2010). 4.4. Geodynamic model of the Manihiki Plateau New age data from four additional widely distributed locations on the Manihiki Plateau provide further support that the Manihiki Plateau was formed during the “Greater Ontong Java Event”, which includes the Ontong Java, Manihiki and Hikurangi Plateaus and possibly the related volcanism in the East Marianas, Lyra and Nauru Basins around Ontong Java (Fitton and Godard, 2004). This event was the most extreme volcanic episode on Earth during the Phanerozoic (covering c. 1% of the Earth's surface), requiring an extreme melting event. Increased partial melting through plume–ridge interaction as proposed for the formation of the Ontong Java Plateau (Mahoney, 1987; Mahoney et al., 1993) cannot sufficiently explain the eruption of such large volumes of lavas across the entire plateau with similar ages (124.6 Ma). Furthermore, the relatively high CaO at a given MgO and the low Sr/Y and Zr/Hf are MgO inconsistent with plateau formation through partial melting of an upwelling primarily eclogitic source as proposed by Korenaga (2005). High partial melting temperatures and the presumed proximity to spreading centers argue against the plate separation model. No evidence for a meteorite impact (e.g. the presence of shocked quartz, iridium 143 anomaly or associated tsunami deposits) has been found at c. 125 Ma. The low H2O content in the Manihiki submarine glasses (184– 258 ppm), similar to water content from the Ontong Java Plateau lavas and MORB (170 ± 30 and 140 ± 40 ppm H2O, respectively; Roberge et al., 2004) excludes flux melting as the major mechanism for generating the high degrees of melting and large volumes of magmatism. Together with evidence for significant uplift in the early phase of volcanic activity at Manihiki (i.e. evidence for shallow water or subaerial formation), the high temperatures and low volatile contents support the presence of a thermal and probably also compositional anomaly, thus favoring the plume hypothesis. Two major components have been previously identified in volcanism associated with the Greater Ontong Java Event: 1) the regionally dominant Kwaimbaita/Kroenke component (with Kroenke-type melts being parental to the Kwaimbaita lavas) with flat incompatible element patterns and isotopic compositions similar to the FOZO of Hauri et al. (1994), and 2) the stratigraphically younger Singgalo component with more enriched incompatible element abundances and EMI-type isotopic signatures. The Manihiki low-Ti volcanic rocks have similar Nd and Hf isotopic compositions, but they have distinct incompatible element abundances and more radiogenic Pb but less radiogenic Sr isotope ratios than the Kwaimbaita/Kroenke and Hikurangi A lavas. As discussed above, a multistage history involving at least two distinct source components is required: 1) one with more depleted incompatible element abundances than the Kwaimbaita/Kroenke lavas (possibly similar to those of the Osbourn seamounts located just off the conjugate rifted margins of the Hikurangi and Manihiki Plateaus), and 2) one with enriched highly incompatible elements and radiogenic Pb and less radiogenic Sr, similar to the late-stage HIMU-type alkalic volcanism on the plateaus. In summary, at least three distinct components are required to explain the geochemistry of the Manihiki basement lavas: 1) Osbourn seamount, possibly FOZO-type, 2) HIMU-type and 3) EM1-type components. The Kwaimbaita/Kroenke component at Ontong Java can either be a mixture of the Osbourn and Singgalo (or more extreme EMI-type component such as sampled at Pitcairn) components. Alternatively the compositions of the Osbourn and Kwaimbaita/Kroenke lavas could reflect heterogeneity in the FOZO-type source. It has been proposed that the FOZO component represents a common and ubiquitous component present in the lower mantle (Hart et al., 1992; Hauri et al., 1994; Stracke et al., 2005; Zindler and Hart, 1986), although the proposed compositions for FOZO show considerable variation (see Stracke et al., 2005, for a summary). It is, however, to be expected that this component, if it indeed represents the composition of the lower mantle, will also vary regionally as is the case with MORB, which shows that the upper mantle is regionally heterogeneous. Excluding the Manhiki low-Ti lavas, the remaining plateau basement lavas (Kwaimbaita/Kroenke, Hikurangi A, Singgalo, Hikurangi B and the Manihiki high-Ti groups) and Osbourn Seamounts form a very similar isotopic array to the lavas from the Hawaiian volcanoes of Kilauea, Loihi and Mauna Loa with the Osbourn samples overlapping the field for Kilauea (e.g. Fig. 7). High 3He/4He ratios in lavas from Kilauea, Loihi, and Mauna Loa (e.g. Farley and Neroda, 1998), together with the presence of low velocity zones beneath the Hawaiian volcanoes (e.g. Montelli et al., 2004, 2006; Nolet et al., 2006), strongly argue for a lower mantle origin of the Hawaiian plume. The similarity of the isotopic composition of the Hawaiian volcanoes and the Manihiki basement lavas provides further support that the Manihiki, Ontong Java and Hikurangi Plateaus were derived from similar, but heterogeneous source in the lower mantle (e.g Hoernle et al., 2010; Tejada et al., 2002). During Late Jurassic and Early Cretaceous times, the Panthalassa (proto-Pacific) Ocean was more extensive than today (e.g. Mueller et al., 2008). The emplacement of the Manihiki Plateau nearly in the center of the Pacific Ocean argues against the direct involvement of continental crust and/or sediments. Thermal erosion and delamination of the SCLM during the breakup of Gondwana and subsequent recycling into the upper mantle and subsequent storage at a thermal boundary layer (such 144 C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 as the 660 km boundary) have been proposed for the origin of the EMItype signature in the southern Atlantic and Indian Oceans (Geldmacher et al., 2008; Hawkesworth et al., 1986; O'Reilly et al., 2009). Gondwana SCLM could also have been delaminated during subduction around the proto Pacific Ocean. If delamination of SCLM has been an ongoing process since ancient times (O'Reilly et al., 2009), the SCLM could have developed into a widespread enriched mantle reservoir along the transition zone or enriched pockets could be distributed throughout the Earth's mantle down to the D″ layer at the core–mantle boundary. A plume, initiated at the core mantle boundary, could entrain significant amounts of such enriched material while rising through the mantle, as well as HIMU-type recycled oceanic crust in the form of eclogite/pryoxenite. It, however, cannot be excluded that all three of the components (FOZO-type, EM1-type and HIMU-type) in the basement lavas were derived from the plume source, e.g. the D″ boundary layer. Delaminated subcontinental lithospheric mantle could contain all three of these components. Depleted harzburgitic portions of the subcontinental lithosphere can have highly depleted incompatible element and overall depleted isotopic compositions resulting from depletion related to crust formation and later magmatism. Various metasomatic processes can cause enrichment of the subcontinental lithosphere, leading to an EM1type composition for example through carbonatite metasomatism and a HIMU-type composition through metasomatism with very low-degree silica-undersaturated melts (e.g. see Geldmacher et al., 2008). Based on experimental results, pyroxenite/eclogite and enriched mantle peridotite will begin melting at deeper depth compared to depleted upper mantle (e.g. Dasgupta et al., 2006, 2007; Hirschmann, 2000; Hirschmann and Stolper, 1996). Therefore the first melts would be derived from more fertile parts of the plume, for example pyroxenite/ eclogite with a HIMU-type composition and the enriched peridotite with an (EMI)-type composition. Entrained parts of the plume, which would then become progressively more diluted with increasing degree of partial through decompression as the plume ascended. Recent studies of seismic tomography in the south Pacific show a large low-velocity zone located beneath the transition zone extending down to the lower mantle (Li et al., 2008; Suetsugu et al., 2009; Tanaka et al., 2009). Small plumes rising from a stalled super-plume are believed to be responsible for the formation of the ocean islands of French Polynesia (Suetsugu et al., 2009). Assuming its existence since mid-Cretaceous time, this super-plume could explain the formation of not only the Manihiki Plateau but also the Greater Ontong Java Event (Hoernle et al., 2010; Larson, 1991a,b) (Fig. 8). Larson (1991a,b) attributed the intense magmatism/volcanism in the Pacific Basin during mid-Cretaceous time to an even larger superplume (c. 6000 to 10,000 km in diameter), which he proposed formed at c. 125 Ma at the core–mantle boundary. Significant uplift should have accompanied the arrival of such a large-scale upwelling event, which could account for the shallow water to subaerial formation of parts of the Manihiki Plateau. A super-plume, stalled at the transition zone (as shown by Li et al., 2008; Suetsugu et al., 2009; Tanaka et al., 2009) and feeding a widespread swarm of secondary plumes, would also result in less significant uplift than impingement of the super-plume at the base of the lithosphere. Such a model could also explain why the temperatures calculated for Ontong Java and Manihiki lavas, despite the magnitude of this event, are lower than strong plumes such as Hawaii. In conclusion, we favor the formation of the Manihiki Plateau and possibly the Greater Ontong Java Event through numerous secondary plumes coming off a newly arrived heterogeneous superplume head stalled at the transition zone with at least three distinct compositional endmembers. Fig. 8. Schematic model for the origin of the Manihiki Plateau and possibly also for the Ontong Java and Hikurangi Plateaus. A deep derived (FOZO-type) superplume, most likely from a thermal boundary layer such as the core–mantle boundary (CMB), ascends until the transition zone at c. 660 km, either already containing or entraining domains of EMI-type SCLM and eclogite/pyroxenite with HIMU-type compositions (recycled ocean crust (ROC)) during its ascent. The EMI-and HIMU-type components are sampled at lower degrees of melting and the FOZO type at higher degrees of melting. plateau formation continuing until c. 117 Ma. Therefore its formation was contemporaneous with the Ontong Java and Hikurangi Plateaus. 2) The geochemical data suggest the presence of two groups of basement lavas with different minor and trace element and isotopic compositions: a) High-Ti group, compositionally similar to the Singgalo lavas on the Ontong Java Plateau, with an EMI-type isotopic composition, and b) Low-Ti group of lavas ranging in composition from the Kwaimbaita/Kroenke lavas at Ontong Java to more HIMU-like compositions. Following the methods of Herzberg and Asimow (2008), we propose that both groups formed through partial melting of peridotite. Calculated primary partial melts for the low-Ti group lavas were formed through c. 30% partial melting at mantle potential temperatures of c. 1510 °C, falling into the average mantle potential temperature for the formation of LIPs (Herzberg and Gazel, 2009). The incompatible element and isotopic composition of the low-Ti group lavas can best be explained by the enrichment of a highly depleted source (similar in composition to the Osbourn Seamounts located adjacent to the rifted northeast margin of the Hikurangi Plateau) with small degree melts from a HIMU-type source, similar to the source of late-stage alkalic volcanism on each of the plateaus. This is the first time that involvement of a HIMU-type component has been identified in the Greater Ontong Java volcanism. 3) Based on shallow emplacement (e.g. Ai et al., 2008) and evidence for unusually high degree of melting, and strongly elevated melting temperatures, the plume head theory can best explain the origin of the Manihiki Plateau. The plume head may have stalled at the transition zone, consistent with the present occurrence of the superplume beneath the SW Pacific. Supplementary materials related to this article can be found online at doi:10.1016/j.epsl.2011.01.025. Acknowledgements 5. Conclusions 1) The Manihiki Plateau formed in Early Cretaceous time primarily at 124.6 ± 1.6 Ma but with volcanism associated with the late stages of We thank Dagmar Rau, Jan Fietzke and Silke Hauff for their technical assistance with major element and isotopic analyses. We are grateful to SO193 Captain Mallon, his crew, and the shipboard scientists for their expert support. A. Ehmer and J. Leppin helped with C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 processing of the SIMRAD data. Constructive reviews from two anonymous reviewers helped to improve the manuscript. Richard Carlson is thanked for editorial handling and additional comments. The German Ministry of Education and Research (BMBF; Grant SO193 Manihiki) and German Research Foundation (DFG; Grant HO1833/191 to cover costs for a substitute to cover KH's teaching load during preparation of the manuscript) are thanked for providing funds to carry out and publish this study. Fig. 1 was prepared with GMT public domain software (Wessel and Smith, 1995). References Ai, H.A., Stock, J.M., Clayton, R., Luyendyk, B., 2008. Vertical tectonics of the High Plateau region, Manihiki Plateau, Western Pacific, from seismic stratigraphy. Mar. Geophys. Res. 29, 13–26, doi:10.1007/s11001-008-9042-0. Anderson, D., 1996. Enriched asthenosphere and depleted plumes. Int. Geol. Rev. 38 (1), 1–21, doi:10.1080/00206819709465320. Anderson, D., 2000. The thermal state of the upper mantle; no role for mantle plumes. Geophys. Res. Lett. 27 (22), 3623–3626. Anderson, D.L., 2005. Large igneous provinces, delamination, and fertile mantle. Elements 1, 271–275. Beiersdorf, H., Bickert, T., Cepek, P., Fenner, J., Petersen, N., Schönfeld, J., Weiss, W., Won, M.Z., 1995. High-resolution stratigraphy and the response of biota to Late Cenozoic environmental changes in the central equatorial Pacific Ocean (Manihiki Plateau). Mar. Geol. 125, 29–59. Billen, M.I., Stock, J., 2000. Morphology and origin of the Osbourn Trough. J. Geophys. Res. 105 (B6), 13,481–13,489. Campbell, I.H., 1998. The Mantle's chemical structure: insights from melting products of mantle plumes. In: Jackson, I. (Ed.), The Earth's Mantle; Composition, Structure and Evolution. Cambridge University Press, pp. 259–310. Campbell, I.H., 2003. Large igneous provinces and the mantle plume hypothesis. Elements 1 (5), 265–269, doi:10.2113/gselements.1.5.265. Campbell, I.H., Griffith, R.W., Hill, R.I., 1989. Melting in an Archaean mantle plume: heads it's basalts, tails it's komatiites. Nature 339, 697–699, doi:10.1038/339697a0. Clague, D.A., et al., 1976. Petrology of basaltic and gabbroic rocks dredged from the Danger Island Troughs, Manihiki Plateau. In: Schlanger, S.O., Jackson, E.D. (Eds.), Initial Reports of the Deep Sea Drilling Project, Volume 33. U.S. Government Printing Office, Washington, D.C., pp. 891–907. Clift, P., 2005. Sedimentary evidence for moderate mantle temperature anomalies associated with hotspot volcanism. Geol. Soc. Am. Spec. Pap. 388, 279–287. Coffin, M., Eldholm, O., 1994. Large ingneous provinces: crustal structure, dimensions and external consequences. Rev. Geophys. 32, 1–36. Cordery, M.J., Davies, G.F., Campbell, I.H., 1997. Genesis of flood basalts from eclogitebearing mantle plumes. J. Geophys. Res. 102, 20179–20197. Courtillot, V., Davaille, A., Besse, J., Stock, J., 2003. Three distinct types of hotspots in the Earth's mantle. Earth Planet. Sci. Lett. 205, 295–308, doi:10.1016/S0012-821X(02)01048-8. Dasgupta, R., Hirschmann, M.M., Stalker, K., 2006. Immiscible transition from carbonaterich to silicate-rich melts in the 3 GPa melting interval of eclogite +CO3 and genesis of silica-undersaturated ocean island lavas. J. Petrol. 47 (4), 647–671. Dasgupta, R., Hirschmann, M.M., Smith, N.D., 2007. Partial melting experiments of peridotite + CO2 at 3 GPa and genesis of alkalic ocean island basalts. J. Petrol. 48 (11), 2093–2124. Davy, B., Hoernle, K., Werner, R., 2008. Hikurangi Plateau: crustal structure, rifted formation and Gondwana subduction history. Geochem. Geophys. Geosyst. 9, Q07004, doi:10.1029/2007/GC0011855. Dixon, J.E., Dixon, T.H., Bell, D.R., Malservisi, R., 2004. Lateral variation in the upper mantle viscosity: the role of water. Earth Planet. Sci. Lett. 222, 451–467, doi:10.1016/j.epsl.2004.03.022. Duncan, R.A., Pyle, D.G., 1988. Rapid eruption of the Deccan flood basalts at the Cretaceous/Tertiary boundary. Nature 333, 841–843. Eisele, J., Sharma, M., Galer, S.J.G., Blichert-Toft, J., Devey, C.W., Hofmann, A.W., 2002. The role of sediment rcycling in EM-1 inferred from Os, Pb, Hf, Nd, Sr isotope and trace element systematics of the Pitcairn hotspot. Earth Planet. Sci. Lett. 196, 197–212. Eldholm, O., Coffin, M.F., 2000. Large igneous provinces and plate tectonics. The History and Dynamics of Global Plate Motions. : Geophysical Monograph, 121, pp. 309–326. Elkins-Tanton, L.T., 2007. Continental magmatism, volatile recycling, and a heterogeneous mantle caused by lithospheric gravitational instabilities. J. Geophys. Res. 112, B03405, doi:10.1029/2005JB004072. Farley, K.A., Neroda, E., 1998. Noble gases in the Earth's mantle. Annu. Rev. Earth Planet. Sci. 26, 189–218. Fitton, G., Godard, M., 2004. Origin and evolution of magmas on the Ontong Java Plateau. In: Fitton, J.G., Mahoney, J.J., Wallace, P.J., Saunders, A.D. (Eds.), Origin and evolution of the Ontong-Java Plateau: Geological Society London, Special Publications, 229, pp. 151–178. Geldmacher, J., Hanan, B.B., Blichert-Toft, J., Harpp, K., Hoernle, K., Hauff, F., Werner, R., Kerr, A.C., 2003. Hafnium isotopic variations in volcanic rocks from the Carribean Large Igneous Province and Galapagos hot spot tracks. Geochemistry Geophysics Geosystems 4 (7), 1062, doi:10.1029/2002GC000477. Geldmacher, J., Hoernle, K., Kluegel, A., Bogaard, Pvd, Bindeman, I., 2008. Geochemistry of a new enriched mantle type locality in the northern hemisphere: implications for the origin of the EM-I source. Earth Planet. Sci. Lett. 265, 167–182. 145 Gibson, S.A., Thompson, R.N., Day, J.A., Humphris, S.E., Dickin, A.P., 2005. Melt generating processes associated with the Tristan mantle plume: constraints on the origin of EM-1. Earth Planet. Sci. Lett. 237, 744–767. Griffin, W.L., O'Reilly, S.Y., Afonso, J.C., Begg, G.C., 2009. The composition and evolution of lithospheric mantle: a re-evaluation and its tectonic implications. J. Petrol. 50 (7), 1185–1204, doi:10.1093/petrology/egn033. Griffiths, R.W., Campbell, I.H., 1991. Interaction of mantle plume heads with the Earth's surface and onset of small-scale convection. J. Geophys. Res. 96, 18,295–18,310. Hales, T.C., Abt, D.L., Humphreys, E.D., Roering, J.J., 2005. A lithospheric instability origin for the Columbia River flood basalts and Wallowa Mountains uplift in northeast Oregon. Nature 438, 842–845. Hames, W.E., Renne, P.R., Ruppel, C., 2000. New evidence for geologically instantaneous emplacement of earliest Jurassic Central Atlantic magmatic province basalts on the North American margin. Geology 28 (9), 859–862, doi:10.1130/0091-7613(2000)28. Hart, S.R., Hauri, E.H., Oschmann, L.A., Whitehead, J.A., 1992. Mantle plumes and entrainment: isotopic evidence. Science 256, 517–520, doi:10.1126/ science.256.5056.517. Hauff, F., Hoernle, K., Tilton, G., Graham, D.W., Kerr, A.C., 2000. Large volume recycling of oceanic lithosphere over short time scales: geochemical constraints from the Caribbean Large Igneous Province. Earth Planet. Sci. Lett. 174, 247–263. Hauri, E.H., Whitehead, J.A., Hart, S.R., 1994. Fluid dynamic and geochemical aspects of entrainment in mantle plumes. J. Geophys. Res. 99 (B12), 24,275–24,300. Hawkesworth, C.J., Mantovani, M.S.M., Taylor, P.N., Palacz, Z., 1986. Evidence from the Parana of South Brasil for a continental contribution to Dupal basalts. Nature 322. Herzberg, C., Asimow, P., 2008. Petrology of some Oceanic Island Basalts: PRIMELT2.XLS software for Primary Magma Calculation. Geochem. Geophys. Geosyst. 9 (9), Q09001, doi:10.1029/2008GC002057. Herzberg, C., Gazel, E., 2009. Petrological evidence for secular cooling in mantle plumes. Nature 458, 619–622, doi:10.1038/nature07857. Hirose, K., Kushiro, I., 1993. Partial melting of dry peridotites at high pressures: determination of compositions of melts segregated from peridotite using aggregates of diamond. Earth Planet. Sci. Lett. 114, 477–489 Geosystems 9 (9) Q09001, doi:10.1029/2008GC002057. Hirschmann, M.M., 2000. Mantle solidus: experimental constraints and the effects of peridotite composition. Geochem. Geophys. Geosyst. 1 2000GC000070. Hirschmann, M.M., Stolper, E.M., 1996. A possible role for garnet pyroxenite in the origin of the "garnet signature" in MORB. Contributions to Mineralogy and Petrology 124, 185–208. Hoernle, K., Bogaard, Pvd, Hauff, F., 2004a. A 70 Myr history (69–139 Ma) for the Caribbean Large Igneous Province. Geology 32, 697–700. Hoernle, K., Hauff, F., Werner, R., Mortimer, N., 2004b. New insights into the origin and evolution of the Hikurangi Oceanic Plateau (Southwest Pacific) from multi-beam mapping and sampling. EOS Trans. AGU Feature 85 (41), 401–408. Hoernle, K., Hauff, F., Werner, R., van den Bogaard, P., Timm, C., Coffin, M., Mortimer, N.N., Davy, B.W., 2008. A Similar Multi-Stage Geochemical Evolution for the Manihiki, Hikurangi and Ontong Java Plateau? AGU Fall Meeting, San Francisco, USA, December 14–19, Eos Transactions AGU 89 (53), Fall Meeting Supplement Abstract V23H-05. Hoernle, K., Timm, C., Hauff, F., Rupke, L., Werner, R., Bogaard, P.v.d., Michaet, P.J., Coffin, M., Mortimer, N.N., Davy, B., 2009. New Results for the Multi-Stage Geochemical Evolution of the Manihiki and Hikurangi Plateaus. AGU Fall Meeting, San Francisco, USA, December 14–18, Eos Transactions AGU 90 (53), Fall Meeting Supplement Abstract V51H-03. Hoernle, K., Hauff, F., Werner, R., van den Bogaard, P., Mortimer, N., Geldmacher, J., Garbe-Schoenberg, D., Davy, B., 2010. Age and geochemistry of volcanic rocks from the Hikurangi and Manihiki oceanic plateaus. Geochim. Cosmochim. Acta 74, 7196–7219, doi:10.1016/j.gca.2010.09.030. Hofmann, A.W., 1988. Chemical differentiation of the Earth: the relationship between mantle, continental and oceanic crust. Earth Planet. Sci. Lett. 90, 297–314. Hussong, D.M., Wipperman, L.K., Kroenke, L.W., 1979. The crustal structure of the Ontong Java and Manihiki oceanic plateaus. J. Geophys. Res. 84 (B11), 6003–6010. Ingle, S., Coffin, M.F., 2004. Impact origin for the greater Ontong Java Plateau? Earth Planet. Sci. Lett. 218, 123–134. Ingle, S., Mahoney, J.J., Sato, H., Coffin, M.F., Kimura, J.-I., Hirano, N., Nakanishi, M., 2007. Depleted mantle wedge and sediment signature in unusual basalts from the Manihiki Plateau, Central Pacific. Geology 35, 595–598. Ito, G., Clift, P., 1998. Subsidence and growth of Pacific Cretaceous plateaus. Earth Planet. Sci. Lett. 161, 85–100. Jackson, E.D., Bargar, K.E., Fabbi, B.P., Heropoulos, C., 1976. Petrology of the basaltic rocks drilled on Leg 33. Initial Reports DSDP, 33, pp. 571–630. Jenkyns, H., 1976. Sediments and sedimentary history of the Manihiki Plateau, South Pacific Ocean. Init Rep DSDP, 33, pp. 873–890. Jones, A.P., Price, D.G., Price, N.J., DeCarli, P.S., Clegg, R.A., 2002. Impact induced melting and the development of large igneous provinces. Earth Planet. Sci. Lett. 202, 551–561. King, S., Anderson, D., 1998. Edge-driven convection. Earth Planet. Sci. Lett. 160, 289–296. Korenaga, J., 2005. Why did not the Ontong Java Plateau form subaerially? Earth Planet. Sci. Lett. 234, 385–399. Lanphere, M.A., Dalrymple, G.B., 1976. K–Ar ages of basalts from DSDP Leg 33: Sites 315 (Line Islands) and 317 (Manihiki Plateau). Initial Report DSDP, 33, pp. 649–653. Larson, R.L., 1991a. Geological consequences of superplumes. Geology 19 (10), 963–966. Larson, R.L., 1991b. Latest pulse of Earth: evidence for a mid-Cretaceous superplume. Geology 19 (6), 547–555. Larson, R.L., 1997. Superplumes and ridge interactions between Ontong Java and Manihiki Plateaus and the Nova-Canton Trough. Geology 25, 779–782. Larson, R.J., Pockalny, R.A., Viso, R.F., Erba, E., Abrams, L.J., Luyendyk, B.P., Stock, J.M., Clayton, R.W., 2002. Mid-Cretaceous tectonic evolution of the Tongareva triple junction in the southwest Pacific Basin. Geology 30, 663–678. 146 C. Timm et al. / Earth and Planetary Science Letters 304 (2011) 135–146 Le Maitre, R.W., Streckeisen, A., Zanettin, B., Le Bas, M.J., Bonin, B., Bateman, P., Bellieni, G., Dudek, A., Efremova, S., Keller, J., Lamere, J., Sabine, P.A., Schmid, R., Sörensen, H., Woolley, A.R., 2002. Igneous Rocks: A Classification and Glossary of Terms, Recommendations of the International Union of Geological Sciences, Subcommission of the Systematics of Igneous Rocks. Cambridge University Press. Li, C., van der Hilst, R.D., Engdahl, E.R., Burdick, S., 2008. A new global model for P wave speed variations in Earth's mantle. Geochem. Geophys. Geosyst. 9 (5), Q05018, doi:10.1029/2007GC001806. Lustrino, M., Dallai, L., 2003. On the origin of the EM-I end-member. Neues Jahrb. Mineral. Abh. 179/1, 85–100. Mahoney, J.J., 1987. An isotopic survey of Pacific oceanic plateaus: implication for their nature and origin. In: Keating, B.H., Fryer, P., Batiza, R., Boehlert, G.W. (Eds.), Seamounts, Islands, and Atolls: American Geophysical Union Monograph, 43, pp. 207–220. Washington, D.C. McDonough, W.F., 1990. Constraints on the composition of the continental lithospheric mantle. Earth and Planetary Science Letters 101, 1–18. Mahoney, J.J., Spencer, K., 1991. Isotopic evidence for the origin of the Manihiki and Ontong Java oceanic plateaus. Earth Planet. Sci. Lett. 104, 196–210. Mahoney, J.J., Storey, M., Duncan, R.A., Spencer, K.J., Pringle, M., 1993. Geochemistry and Geochronology of Leg 130 Basement Lavas: Nature and Origin of the Ontong Java Plateau. In: Pringle, M. (Ed.), AGU, Washington, D.C. Meyzen, C.M., Blichert-Toft, J., Ludden, J.N., Humler, E., Mevel, C., Albarede, F., 2007. Isotopic portrayal of the Earth's upper mantle flow field. Nature 447, doi:10.1038/ nature05920. Michael, P., 1995. Regionally distinctive sources of depleted MORB: evidence from trace elements and H2O. Earth Planet. Sci. Lett. 131, 301–320. Michael, P.J., 1999. Implications for magmatic processes at Ontong Java Plateau from volatile and major element contents of Cretaceous basalt glasses. Geochem. Geophys. Geosyst. 1, 1008, doi:10.1029/1999GC000025. Montelli, R., Nolet, G., Dahlen, R.A., Masters, G., Engdahl, E.R., Hung, S.-H., 2004. Finitefrequency tomography reveals a variety of plumes in the mantle. Science 303, 338–343. Montelli, R., Nolet, G., Dahlen, R.A., Masters, G., 2006. A catalogue of deep mantle plumes: new results from finite frequency tomography. Geochem. Geophys. Geosyst. 7 (11), Q11007, doi:10.1029/2006GC001248. Mortimer, N., Hoernle, K., Hauff, F., Palin, J.M., Dunlap, W.J., Werner, R., Faure, K., 2006. New constraints on the age and evolution of the Wishbone Ridge, southwest Pacific Cretaceous microplates, and Zealandia–West Antarctica breakup. Geology 34 (3), 185–188. Mueller, D.R., Sdrolias, M., Gaina, C., Steinberger, B., Heine, C., 2008. Long-term sea-level fluctuations driven by ocean basin dynamics. Science 319, 1357–1362, doi:10.1126/science.1151540. Nolet, G., Karato, S.-I., Montelli, R., 2006. Plume fluxes from seismic tomography. Earth Planet. Sci. Lett. 248, 685–699. O'Reilly, S.Y., Zhang, M., Griffin, W.L., Begg, G., Hronsky, J., 2009. Ultradeep continental roots and their oceanic remnants: a solution to the geochemical “mantle reservoir” problem? Lithos 211, 1043–1054, doi:10.1016/j.lithos2009.04.028. Peate, D.W., 1997. The Parana-Etendeka Province. In: Mahoney, J.J., Coffin, M.F. (Eds.), Large Igneous Provinces, pp. 217–245. Renne, R.R., Zichao, Z., Richards, M.A., Black, M.T., Basu, A.R., 1995. Synchrony and causal relations between Permian–Triassic boundary crises and Siberian flood volcanism. Science 269, 1413–1416. Richards, M.A., Duncan, R.A., Courtillot, V.E., 1989. Flood basalts and hot spot tracks: plume heads and tails. Science 246, 103–107. Roberge, J., White, R.V., Wallace, P., 2004. Volatiles in submarine basaltic glasses from the Ontong Java Plateau (ODP Leg 192): implications for magmatic processes and source region compositions. Geol. Soc. Lond. Spec. Publ. 229, 239–257, doi:10.1144/ GSL.SP.2004.229.01.14. Rogers, G.C., 1982. Oceanic Plateaus as meteorite impact signatures. Nature 299, 341–342. Salters, V.J.M., Sachi-Kocher, A., 2010. An ancient matesomatic source for the Walvis Ridge basalts. Chem. Geol. 273, 151–167. Salters, V.J.M., Blichert-Toft, J., Fekiacova, Z., Sachi-Kocher, A., Bizimis, M., 2006. Isotope and trace element evidence for depleted lithosphere in the source of enriched Ko'olau basalts. Contrib. Mineralog. Petrol. 151, 297–312, doi:10.1007/s00410-005-0059-y. Self, S., Jay, A.E., Widdowson, M., Keszthelyi, L.P., 2008. Correlation of the Deccan and Rajahmundry Trap lavas: are these the longest and largest lava flows on Earth? J. Volcanol. Geoth. Res. 172, 3–19, doi:10.1016/j.jvolgeores.2006.11.012. Sliter, W.V., 1992. Cretaceous planktonic foraminiferal biostratigraphy and plaeoceanographic events in the Pacific Ocean with emphasis on indurated sediments. In: Ishizaki, K., Saito, S. (Eds.), Centenary of Japanese Micropaleontology. Terra Science, Tokyo, pp. 261–299. Stracke, A., Hofmann, A.W., Hart, S.R., 2005. HIMU, FOZO and the rest of the mantle zoo. Geochem. Geophys. Geosyst. 6 (5), Q05007, doi:10.1029/2004GC000824. Suetsugu, D., Isse, T., Tanaka, S., Obayashi, M., Shiobara, H., Sugioka, H., Kanazawa, T., Fukao, Y., Barroul, G., Reymond, D., 2009. South Pacific mantle plume imaged by seismic observation on islands and seafloor. Geochem. Geophys. Geosyst. 10 (11), Q11014, doi:10.1029/2009GC002533. Sun, S.-s., McDonough, W.F., 1989. Chemical and isotopic systematics 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 Publications, 42, pp. 313–345. Tanaka, S., Suetsugu, D., Shiobara, H., Sugioka, H., Kanazawa, T., Fukao, Y., Barroul, G., Reymond, D., 2009. On the vertical extent of the large low shear velocity province beneath the South Pacific Superswell. Geophys. Res. Lett. 36, L07305, doi:10.1029/ 2009GL037568. Taylor, B., 2006. The single largest oceanic plateau: Ontong Java‚ Manihiki‚ Hikurangi. Earth Planet. Sci. Lett. 241, 372–380. Tejada, M.L.G., Mahoney, J.J., Duncan, R.A., Hawkins, M.P., 1996. Age and geochemistry of basement and alkalic rocks of Malaita and Santa Isabel, Solomon Islands, southern margin of the Ontong Java Plateau. J. Petrol. 37, 361–394. Tejada, M.L.G., Mahoney, J.J., Neal, C.R., Duncan, R.A., Petterson, M.G., 2002. Basement geochemistry and geochronology of Central Malaita, Solomon Islands, with implications for the origin and evolution of the Ontong Java Plateau. J. Petrol. 43 (3), 449–484. Tejada, M.L.G., Mahoney, J.J., Castillo, P.R., Ingle, S.P., Sheth, H.C., Weis, D., 2004. Pinpricking the elephant: evidence on the origin of the Ontong-Java Plateau from Pb– Sr–Hf–Nd isotopic characteristics of ODP Leg 182 basalts. In: Fitton, J.G., Mahoney, J. J., Wallace, P.J., Saunders, A.D. (Eds.), Origin and Evolution of the Ontong-Java Plateau: Geological Society London, Special Publications, 229, pp. 133–150. Timm, C., Hoernle, K., van den Bogaard, P., Bindemann, l., Weaver, S., 2009. Geochemical evolution of intraplate volcanism at Banks Peninsula, New Zealand: interaction between lithospheric and asthenospheric melts. J. Petrol. 50, 989–1023. Timm, C., Hoernle, K., Werner, R., Hauff, F., van den Bogaard, P., White, J., Mortimer, N., Garbe-Schoenberg, D., 2010. Temporal and geochemical evolution of the Cenozoic intraplate volcanism of Zealandia. Earth Sci. Rev. 98, 38–64, doi:10.1016/j. earscirev.2009.10.002. Viso, R.F., Larson, R.L., Pockalny, R.A., 2005. Tectonic evolution of the Pacific–Phoenix– Farallon triple junction in the South Pacific. Earth Planet. Sci. Lett. 233, 179–194. Wallace, P., Carmichael, I.S.E., 1992. Sulfur in basaltic magmas. Geochim. Cosmochim. Acta 56, 1863–1874. Walter, M., 1998. Melting of garnet peridotite and the origin of komatiite and depleted lithosphere. J. Petrol. 39 (1), 29–60. Wessel, P., Smith, W.H.F., 1995. New version of the generic mapping tool released. EOS Trans. AGU 328. Winterer, E.L., Lonsdale, P.F., Matthews, J.L., Rosendahl, B.R., 1974. Structure and acoustic stratigraphy of the Manihiki Plateau. Deep Sea Res. 21, 793–814. Worthington, T.J., Hekinian, R., Stoffers, P., Kuhn, T., Hauff, F., 2006. Osbourn Trough: structure, geochemistry and implications of a mid-Cretaceous paleospreading ridge in the South Pacific. Earth Planet. Sci. Lett. 245, 685–701, doi:10.1016/j. epsl.2006.03.018. Yasuda, A., Fuji, T., Kurita, K.A., 1997. A composite diapir model for extensive basaltic volcanism: magmas from subducted oceanic crust entrained with mantle plumes. Proc. Jpn Acad. 73, 201–204. Zindler, A., Hart, S., 1986. Chemical geodynamics. Annu. Rev. Earth Planet. Sci. 14, 493–571.