Journal of Petrology Advance Access originally published online on May 20, 2005
Journal of Petrology 2005 46(9):1925-1962; doi:10.1093/petrology/egi043
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Origin of Exceptionally Abundant Phonolites on Ua Pou Island (Marquesas, French Polynesia): Partial Melting of Basanites Followed by Crustal Contamination
1 UMR 6538, IUEM, UNIVERSITÉ DE BRETAGNE OCCIDENTALE, 6 AVENUE LE GORGEU, CS 93837, 29238 BREST CEDEX 3, FRANCE
2 LSCE/CEA–CNRS UMR 1572, DOMAINE DU CNRS, 12 AVENUE DE LA TERRASSE, 91118 GIF-SUR-YVETTE, FRANCE
3 LGCA, 1381 RUE DE LA PISCINE, BP 53, 38000 GRENOBLE, FRANCE
4 UMR 6538, IUEM, UNIVERSITÉ DE BRETAGNE OCCIDENTALE, PLACE N. COPERNIC, 29280 PLOUZANÉ, FRANCE
5 CEA/LDG, BP 12, 91680 BRUYÈRES-LE-CHATEL, FRANCE
6 LABORATOIRE DE PÉTROLOGIE CRISTALLINE, GÉOSCIENCES RENNES, UMR 6118, UNIVERSITÉ DE RENNES 1, 35042 RENNES, FRANCE
7 BRGM, BP 6009, 45060 ORLÉANS CEDEX 2, FRANCE
RECEIVED JULY 30, 2004; ACCEPTED MARCH 21, 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGICAL SETTINGERUPTIVE HISTORY OF UA...PETROLOGICAL DATAGEOCHEMICAL RESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
On the basis of the first systematic mapping of Ua Pou, long known for its exceptionally abundant phonolites, we estimate that these rocks cover 65% of the surface of the island whereas mafic lavas cover 27% and intermediate ones 8%. The silica-undersaturated suite was erupted in a restricted time span (2·9–2·35 Myr), following the emplacement of tholeiites derived from a young HIMU-type source at c. 4 Ma. Primitive basanites, derived from a heterogeneous mantle source with a dominant EM II + HIMU signature, represent likely parental magmas. The series is characterized by a Daly gap defined by a lack of phonotephrites. We consider that the most likely model for the origin of evolved lavas is partial melting at depth of primitive basanites, leaving an amphibole-rich residuum and producing tephriphonolitic magmas. These tephriphonolitic magmas may have evolved by closed-system fractional crystallization towards Group A phonolites. Three other groups of phonolites could have been derived from tephriphonolitic magmas by open-system fractional crystallization processes, characterized respectively by seawater contamination (Group B), assimilation of nepheline syenite-type materials (Group C) and extreme fractionation coupled with assimilation of the underlying oceanic crust (Group D). The prominence of evolved lavas is a consequence of their origin from partial melting of mafic precursors followed by crustal contamination.
KEY WORDS: Marquesas; French Polynesia; phonolite; partial melting; contamination
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGERUPTIVE HISTORY OF UA...PETROLOGICAL DATAGEOCHEMICAL RESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Phonolites are common in oceanic islands, although they are usually present in small amounts (a few per cent of the total volume of the outcropping volcanics). They are generally associated in space and time with basanitic lavas, whereas trachytes are found together with mildly alkalic basalts, as already documented by Lacroix (1927)
for many South Pacific occurrences. Lavas of intermediate compositions (phonotephrites, trachyandesites) are often lacking or present in small amounts in these series, a feature commonly referred to as the ‘Daly gap’ (Daly, 1910, 1925). The existence of the Daly gap has been the subject of much controversy; nevertheless, since the work of Chayes (1963, 1977) it is generally accepted that the lavas of the majority of oceanic islands do indeed exhibit a bimodal distribution.
Another strongly debated issue concerns the origin(s) of these oceanic phonolites (and also trachytes and rhyolites). The most generally accepted explanation is that they are derived from the associated mafic lavas through fractional crystallization (Le Roex et al., 1990; Weaver, 1990; Kyle et al., 1992; Caroff et al., 1993; Nono et al., 1994; Mungall & Martin, 1995; Thompson et al., 2001; Cousens et al., 2003) coupled or not with assimilation of either the walls of the magma chamber in which they evolved or those of the conduits through which they ascend towards the surface (Davidson & Wilson, 1989; Freundt & Schmincke, 1995; Wolff et al., 2000). Such a process should result in lava volumes decreasing from mafic towards more evolved compositions, a feature observed in some oceanic islands; for example, Tristan da Cunha (basanites to phonolites: Le Roex et al., 1990), Fernando de Noronha (basanites to phonolites and trachytes: Weaver, 1990) or Mururoa atoll (alkali basalts to trachytes: Caroff et al., 1993). However, it does not easily account for the Daly gap commonly observed in other series; for example, most of those of the Canary and Polynesian Islands (see Bailey, 1987, for a review). Within the framework of the fractional crystallization hypothesis, this gap has been attributed to (1) sampling bias (Baker, 1968), (2) lack of emplacement at the surface of intermediate magmas present at depth (Mungall & Martin, 1995), (3) gently sloping liquidus surfaces (Wyllie, 1963), and (4) magmatic bifurcation in the reservoir (Bonnefoi et al., 1995), among other explanations.
An alternative to the fractional crystallization ± assimilation hypothesis involves low melting degrees of enriched mantle generating silica- and alkali-rich magmas (Bailey, 1987), more or less similar to the felsic interstitial glasses found in mantle xenoliths, some of which are trachytic or phonolitic in composition (Schiano et al., 1994; Neumann & Wulff-Pedersen, 1997; Wulff-Pedersen et al., 1999). Such a process could lead to the emplacement of small volumes of evolved magmas during the early stages of development of oceanic islands (Devey et al., 2001).
Finally, it has been proposed on the basis of experimental data and geochemical studies that the huge volumes of phonolites emplaced on the Kenya plateau derive from the melting at depth of plume-derived basaltic materials (Hay & Wendlandt, 1995; Hay et al., 1995; Kaszuba & Wendlandt, 2000). A similar mechanism has been proposed to account for the origin of the peralkaline trachytes and rhyolites that crop out on 80% of the surface of Socorro Island, Mexico (Bohrson & Reid, 1997), and has also been proposed for the peralkaline rhyolites of Pantelleria (Lowenstern & Mahood, 1991).
The aim of this study is to discuss the origin of the basanite–phonolite association (displaying a characteristic Daly gap) of Ua Pou Island (Marquesas archipelago, French Polynesia, South Pacific). Ua Pou has long been known for its abundant and petrologically diverse phonolites (Lacroix, 1928, 1931; Chubb, 1930; Bishop & Woolley, 1973; Brousse & Maury, 1978). Although no detailed geological map existed prior to the present study, Brousse (1978) proposed that phonolites and associated trachytes represent more than 50% of the surface of Ua Pou. We have revised this estimation to 73% intermediate and evolved lavas (including 65% phonolites) and 27% basanites, proportions that to our knowledge represent the most extreme abundance of phonolites recorded in a magmatic series from oceanic islands. We will show that crystal fractionation and wall-rock assimilation, although occurring within each of these groups, are not responsible for the genesis of the Ua Pou association and that it is more likely that the evolved lavas were derived by partial melting of basanites previously emplaced within the lithospheric mantle beneath the island.
|
GEOLOGICAL SETTING |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGERUPTIVE HISTORY OF UA...PETROLOGICAL DATAGEOCHEMICAL RESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
The Marquesas archipelago
Four Neogene–Quaternary volcanic island chains occur in French Polynesia. From south to north these are the Austral–Cook, Pitcairn–Gambier, Society and Marquesas archipelagos, respectively. Pitcairn–Gambier and Society represent (to a first approximation) typical Hawaiian-type linear chains trending N110–120°E, which correspond to the motion of the Pacific plate towards the NW. Each of them has a hotspot at its southeastern tip, marked by well-identified active submarine volcanoes, and the corresponding age vs distance relationships are consistent with the local Pacific plate motion of about 11 cm/year. The Austral–Cook chain is more complex and its origin has been attributed to coeval hotspots (Chauvel et al., 1997
; Guille et al., 1998; Maury et al., 2000).
The Marquesas archipelago (Fig. 1 inset) is atypical in many respects (Brousse et al., 1990; Guille et al., 2002). It includes eight main islands (Eiao, Nuku Hiva, Ua Huka, Ua Pou, Hiva Oa, Tahuata, Motane, Fatu Hiva) and a few islets and seamounts, all of which formed between 5·5 and 0·4 Ma on oceanic crust of 53–49 Ma age generated at the axis of the Pacific–Farallon ridge.
|
This 350 km long archipelago is only approximately linear (Fig. 1) and it has an unusual N140–150°E trend, which is consistent with the spreading motion of the Pacific–Farallon ridge prior to the onset of accretion along the East Pacific Rise. This feature suggests that the emplacement of the Plio-Quaternary Marquesas magmas was controlled by tectonic features and/or zones of weakness in the underlying Pacific plate (Crough & Jarrard, 1981
; McNutt et al., 1989; Brousse et al., 1990). Another characteristic feature of the Marquesas is the lack of any at present well-identified active volcanic zone at its southeastern edge; that is, close to the Marquesas Fracture Zone, which is supposed to have trapped the hotspot activity (McNutt et al., 1989). Although the ages of the islands tend to decrease towards the SE (Brousse et al., 1990; Desonie et al., 1993; Guille et al., 2002), the correlation between these ages and the corresponding distances to the presumed hotspot area is less clear than in most Hawaiian-type linear chains.
A deep crustal root has been identified by seismic studies below the central part of the archipelago, where the depth of the Moho reaches 15–20 km (Filmer et al., 1993; Caress et al., 1995) and the oceanic crust is covered by an archipelagic apron composed of volcanic debris removed from the islands (Filmer et al., 1994; Wolfe et al., 1994). Two alternative explanations are proposed to explain this thickening: (1) a magmatic body was underplated under the Moho during Plio-Quaternary hotspot activity (Caress et al., 1995; McNutt & Bonneville, 2000) or (2) an oceanic plateau, formed at the axis of a ridge at around 50–45 Ma, constitutes the Marquesas substratum (Gutscher et al., 1999; Guille et al., 2002).
The petrological and geochemical features of the Marquesas lavas are unusually complex. The chemical variability of basalts, from olivine tholeiites to basanites, is generally attributed to variable degrees of melting of heterogeneous sources containing DMM, HIMU and EM II mantle components (Duncan et al., 1986; Dupuy et al., 1989, 1993; Woodhead & McCulloch, 1989; Hart et al., 1992; Woodhead, 1992; Desonie et al., 1993; Hanan & Graham, 1996; Ielsch et al., 1998). However, several studies (e.g. Caroff et al., 1995, 1999; Le Dez et al., 1996) have suggested important chemical interactions between the ascending plume-related magmas (including the oldest ones, drilled in Eiao and dated back to 5·5 Ma) and lithosphere containing incompatible element enriched crustal materials.
Ua Pou island
Ua Pou, located 50 km south of Nuku Hiva in the Marquesas central group, is the fifth largest island of French Polynesia (105 km2). In contrast to most other Marquesas islands, it is devoid of any caldera. Ua Pou has a diamond-shaped outline (Fig. 1) and a north–south-trending central crest, c. 600 m high, from which diverge many secondary crests separated by deep valleys. The villages of Hakahau, Hakamoui, Paaumea, Hohoi, Hakatao, Hakamaii, Haakuti and Hakahetau are located in these valleys. The most spectacular features of the island are its phonolitic protrusions (Ua Pou means the Pillar Island), the summits of which tower several hundred metres above the radial crests (Oave 1203 m, Pouakei 1034 m, Matahenua 1028 m, Poumaka 979 m, Poutetainui 970 m). The Ua Pou volcanic edifice is made up of a thick lava pile, mostly phonolitic and basaltic, intruded by about 30 phonolitic and trachytic protrusions. Like the other Marquesas islands, Ua Pou is devoid of any coral reef or reefal limestone unit.
Geological features
A simplified geological map of the island of Ua Pou is presented in Fig. 1. Rock classification is based on major element compositions plotted in the total alkalis–silica (TAS) diagram of Le Bas et al. (1986). The oldest geological unit corresponds to tholeiitic basaltic flows, which occur in a very restricted area east of Hakahau, where they are exposed at present in a single outcrop.
A more than 200 m thick pile of laharic breccias is exposed along the western coast of the island between Hakahetau and Hikeu. These are composed of numerous debris flow and hyperconcentrated flow deposits, metres to tens of metres thick, in which traces of fossil leaves have been found (Plessis et al., 1978). In Haakuti and Hakamaii Bays, this sequence contains intercalated basanitic, tephritic and tephriphonolitic lava flows.
The lower phonolites are mostly exposed in the northern part of the island (Plateau des Anes) and occasionally in its eastern (Hakamoui) and southwestern bays. They consist of 5–8 m thick lava flows, often showing columnar jointing and a typical whitish or greenish altered surface. A pile of six such flows makes up the northern edge of the island.
The basanitic to intermediate lava flow pile, up to 300 m thick, is mainly located in the eastern, southern and northern parts of the island, where it directly overlies the lower phonolites (Hakamoui, Hakatao), and fills pre-existing depressions (Plateau des Anes). It is composed of a sequence of metre thick lava flows, mostly basanitic with subsidiary tephrites, tephriphonolites and benmoreites, which recalls the shield sequences of Hawaiian-type volcanoes. Laharic breccias are sometimes intercalated between these lava flows. Blocks of syenite occur within these breccias between Paaumea and Hohoi. In addition, a nepheline syenite xenolith has been found within a tephriphonolitic dyke, which crosses the basanitic to intermediate lava flow pile near Hakamoui bay.
The upper phonolitic flows represent the prominent geological unit of Ua Pou (Fig. 1). These flows, sometimes more than ten metres thick, constitute a pile that makes up most of the central part of the island. They radiate from the centre of the island towards the coasts, where they overlie either the basanitic edifice in the eastern part of the island or the laharic breccias in its western part. The total thickness of this phonolitic sequence reaches up to 300 m.
The phonolitic protrusions, which crosscut the former units, are distributed all over the island. However, the major (and most spectacular) peaks are mostly located in its central part, along north–south and NW–SE trends that compose the central ridge and may define fracture zones along which the upper phonolitic flows were erupted. Their erosion feeds well-developed scree breccias, which often mask their contacts. A few trachytic domes have also been mapped, but in terms of volume they are very minor in comparison with the phonolites.
Our estimates of the relative surfaces covered by these units on the new geological map provided in Fig. 1 yield 65% phonolites (12% lower flows, 53% upper flows and protrusions), 29% for the basanitic to intermediate flow edifice and 6% for the laharic breccias and intercalated flows. Converted into surface frequencies per petrographic type, they yield 27% mafic lavas (basanites and less than 1% tephrites plus tholeiites) and 73% intermediate and evolved lavas (including 65% phonolites).
Volumetrically dominant phonolites occur in other basanite–phonolite volcanoes in both oceanic and continental settings. For example, the percentages of phonolitic clasts in some diatremes from the Miocene Dunedin volcano (New Zealand) range from 56% to almost 100% (Price et al., 2003). In Tenerife (Canary Islands), Martí et al. (1994) have subdivided the Las Cañadas edifice into a predominantly mafic Lower Group and an Upper Group corresponding to thick piles of phonolitic lavas and proximal pyroclastic deposits. Three types of phonolites have been recognized in central Kenya, including Late Miocene plateau-type phonolitic lava flows that cover nearly 50000 km3 (Williams, 1972; Lippard, 1973). Mount Erebus (Antarctica) is composed of voluminous anorthoclase-phyric tephriphonolite and phonolite lavas (modern Erebus) overlying unknown volumes of poorly exposed less differentiated lavas (basanites, tephrites and phonotephrites from proto-Erebus) and its summit crater contains a convecting phonolitic lava lake (Esser et al., 2004; Harpel et al., 2004).
|
ERUPTIVE HISTORY OF UA POU |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGERUPTIVE HISTORY OF UA...PETROLOGICAL DATAGEOCHEMICAL RESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Radiometric K–Ar ages: analytical method
K contents and Ar isotopic compositions were measured at Brest and at the ‘Laboratoire des Sciences du Climat et de l’Environnement' of Gif-sur-Yvette, respectively. Only rocks with minimal traces of alteration were selected [samples with loss on ignition (LOI) <2%]. The samples were crushed, sieved to 0·25–0·125 mm size fraction and ultrasonically washed in acetic acid. Potassium and argon were measured on the microcrystalline groundmass, after removal of phenocrysts and xenocrysts using heavy liquids of appropriate densities and magnetic separations. This process improves the K yield as well as the percentage of radiogenic argon, and removes at least some potential sources of systematic error related to the presence of excess 40Ar known to occur in olivine, feldspars and pyroxene, which are ubiquitous phases in the Ua Pou lava flows. Ar analyses were performed using the Cassignol technique (Cassignol & Gillot, 1982
; Gillot & Cornette, 1986). The volumetric calibration of the spike-free introduction line was refined by the cross-calibration of GL-O (Odin et al., 1982; Charbit et al., 1998), Mmhb-I (Samson & Alexander, 1987), LP-6 (Odin et al., 1982), and HD-B1 (Fuhrmann et al., 1987). This calibration (Charbit et al., 1998) allows Ar content to be determined with a precision of 0·2% (±2
).
Geological history
Previous isotopic ages on samples from Ua Pou, given by Duncan et al. (1986), Brousse et al. (1990) and Diraison (1991), range from 5·61 to 1·78 Ma. They allowed those workers to define three stages of edifice construction: (1) an olivine tholeiite episode from 5·61 to 4·46 Ma; (2) an alkali basaltic event from 2·88 to 2·70 Ma; (3) an episode of emplacement of intermediate and evolved lavas from 2·49 to 1·78 Ma.
Detailed geological mapping of the island leads us to recognize more stages of edifice construction than previously proposed, based on stratigraphic distinctions and correlations between cross-sections. In a further step, representative lava samples from these different stratigraphic units were dated. Fifteen new age dates are presented in Table 1.
|
The tholeiitic sample UP33 is dated to 4·00 ± 0·06 Ma, an age slightly younger than previous determinations for that unit. Although this tholeiitic unit located at the base of the volcano pile (at its northeastern edge) is the oldest exposed on the island, it was subaerially emplaced. One might expect, however, that the previous submarine activity history of Ua Pou also involved the emplacement of such tholeiitic lavas, as temporal patterns of magmatic activity in the Marquesas commonly range from tholeiitic towards alkalic magmas (Guille et al., 2002
).
The other ages range from 2·93 ± 0·04 to 2·35 ± 0·03 Ma and define a relatively short subaerial period of activity of 0·58 ± 0·07 Myr, which is considerably shorter than that previously published (from 2·88 to 1·78 Ma, i.e. 1·10 Myr; Duncan et al., 1986; Brousse et al., 1990; Diraison, 1991). These new ages are in agreement with the volcanic stratigraphy described above, although the periods of emplacement of the units overlap considerably (Fig. 2).
|
A significant period of quiescence, 1·07±0·10 Myr long, followed the emplacement of the tholeiitic basaltic flows. This could correspond to a period of destabilization of the initial subaerial volcanic edifice. The basanitic and intermediate lava flows intercalated within the laharic breccia pile were erupted from 2·93 ± 0·04 to 2·91 ± 0·04 Ma (i.e. their mean age is 2·92 ± 0·05 Ma) and the lower phonolites were emplaced from 2·90 ± 0·04 to 2·86 ± 0·04 Ma (mean age 2·88 ± 0·06 Ma). These overlapping ages suggest that the emplacement of these two units was more or less simultaneous, although in some cross-sections along the western coast the lower phonolites overlie the laharic breccias that mainly crop out in that part of the island (Hikeu bay, Fig. 1). Elsewhere, the basal phonolites usually represent the lowermost exposed stratigraphic unit and they are overlain by the basanitic and intermediate edifice (Fig. 1).
During a relatively long period, from 2·76 ± 0·04 to 2·38 ± 0·03 Ma (mean age 2·57 ± 0·23 Ma), the basanitic and intermediate edifice was emplaced over the previous units. The upper phonolites, including both lava flows and protrusions, were emplaced between 2·60 ± 0·04 and 2·35 ± 0·03 Ma (mean age 2·48 ± 0·16 Ma). Once again, these overlapping ages indicate that mafic–intermediate and phonolite lavas erupted more or less simultaneously, although the upper phonolites usually overlie the basanitic and intermediate edifice (Fig. 1).
It is difficult to locate the eruptive vents of the oldest units (e.g. the basanitic to intermediate lava flows intercalated within the laharic breccias and the lower phonolites). The basanitic to intermediate and the upper phonolitic flows were clearly erupted along the central ridge of the island, the trend of which might indicate a fracture zone. The major phonolitic peaks are also located along this ridge (Fig. 1).
According to these data, the average construction rate of the roughly 900 m high lava flow sequence is c. 6 mm/year. This value is only indicative because of the complex geological relationships between the units, including many palaeolandscapes and channellings that provide evidence of periods of erosion and/or destabilization during the building of the emerged part of Ua Pou.
Some similarities, or contrasts, of Ua Pou with other basanite–phonolite volcanoes in oceanic settings can be highlighted. A significant contrast with Dunedin (New Zealand) and Tenerife (Canary Islands) is that the phonolitic pyroclastic activity typical of these volcanoes (Martí et al., 1994; Price et al., 2003) is absent from Ua Pou. In addition, Ua Pou and Dunedin volcano display particularly striking similarities: these two volcanoes contain phonolitic protrusions together with extensive phonolitic lava flows (Price & Chappell, 1975), and in both cases the feeding centres are located along a central fracture zone. The Ua Pou phonolites were emplaced during two distinct phases, whereas three periods of volcanic activity, each involving phonolitic materials, have been recognized in Dunedin volcano (Price & Chappell, 1975). In Tenerife, the Upper Group of the Las Cañadas edifice comprises three phonolitic formations, one of which, the Diego Herñandez Formation, corresponds to seven major phonolitic events (Wolff et al., 2000). Moreover, in this island, the younger Pico Teide and Pico Viejo stratovolcanoes also include significant volumes of phonolitic magmas (Ablay & Martí, 2000).
|
PETROLOGICAL DATA |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGERUPTIVE HISTORY OF UA...PETROLOGICAL DATAGEOCHEMICAL RESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Petrographic types
Two groups of samples can be clearly distinguished in the TAS diagram (Le Bas et al., 1986
) shown in Fig. 3. The mafic lavas plot in the field of basanites and tephrites, with the exception of tholeiitic sample UP33 and one trachybasalt. Most of the mafic lavas are basanites, with normative nepheline contents ranging generally from 5 to 11% (Table 2). The studied basanitic lavas are porphyritic and contain <10% modal phenocrysts, among which olivine usually dominates over clinopyroxene and Fe–Ti oxides. Some basanites contain up to 20% modal olivine. Plagioclase microlites are present in the groundmass together with clinopyroxene, olivine, titanomagnetite, very uncommon hemoilmenite and apatite. Some basanites contain resorbed amphibole phenocrysts (Table 3). Two samples (UP52 and UP86) can be classified as tephrites because their normative olivine content is <10%. The tephrites are porphyritic and contain 10–15% phenocrysts. They contain plagioclase phenocrysts in addition to the usual basanitic mineral assemblage. The highly porphyritic olivine-rich tholeiite sample UP33 has an intergranular texture. It contains augite phenocrysts, and olivine and augite are present in its groundmass, together with ilmenite and plagioclase.
|
|
|
A typical Daly gap, indicated by the lack of phonotephrites and mugearites in the Ua Pou series, separates the mafic group from the intermediate–evolved group, which is more diversified. The latter includes dominant phonolites (non-peralkaline and peralkaline), non-peralkaline trachytes, tephriphonolites and trachyandesites (potassic benmoreites). All these rocks are silica-undersaturated (Table 2), with normative nepheline contents up to 32% in peralkaline phonolites. Four groups (labelled A, B, C and D) of intermediate and evolved silica-undersaturated lavas have been identified based on chemical (trace element and isotopic) criteria, which will be detailed in a later section.
Group A tephriphonolites are usually porphyritic and contain 10–15 modal % phenocrysts. Amphibole is the most frequent phenocryst and the first to appear in the crystallization sequence, with the exception of apatite, which it often contains as inclusions. Amphibole is usually rimmed or even totally replaced by titanomagnetite. Other phenocrysts and groundmass minerals are mainly represented by clinopyroxene, titanomagnetite, plagioclase, apatite, K-feldspar (only as a groundmass mineral) and uncommon titanite microphenocrysts. Group C tephriphonolites are rather similar to Group A tephriphonolites. In the Group C tephriphonolite UP56, isolated crystals or veinlets of high-temperature carbonate are observed. Their La2O3 and Ce2O3 contents reach 2000 ppm and 3500 ppm, respectively (Table 4).
|
K-benmoreites are porphyritic and contain 10–15% phenocrysts. They contain clinopyroxene, amphibole, plagioclase, titanomagnetite and apatite as phenocrysts and groundmass crystals. K-feldspar is present exclusively in the groundmass. Trachytes are also porphyritic and contain c. 10% phenocrysts (clinopyroxene, amphibole, plagioclase and K-feldspar).
Non-peralkaline Group A, B and C phonolites are usually porphyritic with a fluidal texture. They contain 5–15% phenocrysts. Amphibole (almost always irregularly shaped and resorbed) and K-feldspar are the most abundant. They are accompanied by minor amounts of clinopyroxene, titanite, plagioclase and titanomagnetite. Groundmass minerals are mainly K-feldspar, plagioclase and titanomagnetite. In addition, idiomorphic nepheline and sometimes rare aegirine–augite occur in the groundmass.
Peralkaline phonolites [with (Na + K)/Al > 1] are always porphyritic, with 5–15% phenocrysts, and belong to three petrographic types. The first type, which corresponds to most of the Group B phonolites, contains abundant K-feldspar phenocrysts together with some irregularly shaped and resorbed amphibole, very scarce clinopyroxene and titanite phenocrysts. Groundmass minerals are K-feldspar, aegirine–augite, idiomorphic nepheline and sodalite. The second type, which corresponds to most of the Group C phonolites, contains only K-feldspar phenocrysts, together with clusters of aegirine–augite and titanomagnetite microphenocrysts, and K-feldspar, nepheline, sodalite, aegirine–augite and rare titanomagnetite as groundmass minerals. The third type, which corresponds to Group D phonolites UP26 and UP61, is almost identical to the second type but for the occurrence of numerous idiomorphic nepheline microphenocrysts, clusters of aegirine–augite and sodalite microphenocrysts, and lack of titanomagnetite. Aegirine–augite crystals are generally aligned with the flow planes and tend to mould around the phenocrysts.
Potassic feldspar is the most abundant phenocryst in the syenites. It is accompanied by biotite, amphibole, minor amounts of clinopyroxene, titanite and titanomagnetite. Biotite and amphibole contain apatite and zircon inclusions. Nepheline syenite contains abundant nepheline and potassic feldspar phenocrysts, together with biotite (containing apatite and zircon inclusions) and minor amounts of titanite.
Mineralogical features
Mineral compositions: analytical method
Mineral compositions were determined using a CAMECA SX50 electron microprobe (Microsonde Ouest, Brest) using the following analytical conditions: acceleration voltage 15 kV, beam diameter 1 µm, beam current 15 nA, 6 s counting time and correction by the PAP method (Pouchou & Pichoir, 1984). Detection limits are c. 0·2% for major elements and concentrations of <0·3 wt % are considered semi-quantitative. A detailed account of the analytical procedure has been given by Defant et al. (1991).
Olivines
Olivine phenocrysts from olivine tholeiite and basanite display relatively large compositional variations from Fo80 to Fo69 for the olivine tholeiite and from Fo87 to Fo72 for basanites (Tables 3 and 5). Olivine phenocrysts are commonly rimmed by iddingsite. Sometimes, iddingsite is in turn rimmed by fresh olivine, a feature that indicates the magmatic character of the oxidation (HTI, high temperature iddingsite; Goff, 1996; Caroff et al., 1997).
|
Clinopyroxenes
Figure 4 and Tables 3 and 6 show that clinopyroxene phenocrysts are augitic in olivine tholeiite and either fassaitic or diopsidic in basanites, tephrites and tephriphonolites. They display chemical zoning with increasing Ti from their cores to their rims. Fassaites (Table 6) are typical of highly silica-undersaturated, generally potassic, alkaline lavas (Vicat & Pouclet, 1995
). Clinopyroxene phenocrysts are diopsidic in K-benmoreites and trachytes. Fassaite, diopside, hedenbergite and aegirine–augite occur in non-peralkaline phonolites, and hedenbergite and aegirine–augite (Table 4) in peralkaline phonolites. Hedenbergite generally occurs in evolved lavas (characterized by low Mg-numbers) and has not been found together with aegirine–augite.
|
|
Feldspars and feldspathoids
Plagioclase phenocrysts occur in all lavas with the exception of tholeiite, basanites (where they are present only as groundmass minerals) and peralkaline phonolites. They are usually unzoned, and within the basanites (Tables 3 and 7; Fig. 5), tephriphonolites and K-benmoreites their composition is highly variable (An70–35). They are more sodic in non-peralkaline phonolites (An40–28) and trachytes (An15) than in other rock types. In some intermediate lavas, plagioclase phenocrysts are partly resorbed and show textural evidence for disequilibrium with their host magma. In some non-peralkaline phonolites, cores of plagioclase (An30) are rimmed by anorthoclase. Groundmass anorthoclase is observed within tephriphonolites and K-benmoreites, and anorthoclase phenocrysts occur in both types of phonolites and trachytes (Fig. 5). Anorthoclase and Na-sanidine are the prominent phenocrysts of peralkaline phonolites.
|
Groundmass nepheline occurs in all lavas with the exception of K-benmoreites and trachytes. Idiomorphic nepheline (Table 7) microphenocrysts are present in both types of phonolites and can be very abundant in some peralkaline occurrences together with microphenocrysts of sodalite (Table 4).
|
Fe–Ti oxides
With the exception of one basanite where both titanomagnetite and hemoilmenite occur in the groundmass, the lavas studied contain exclusively titanomagnetite. The olivine tholeiite sample, UP33, contains exclusively hemoilmenite microcrysts (ilm88–86; Table 8). Titanomagnetite phenocrysts often display exsolution features. Their ferric iron contents increase progressively from basanites and tephrites (usp70–45), towards phonolites (usp45–32) where they are uncommon, through K-benmoreites and trachytes (usp62–48) and tephriphonolites (usp55–35; Tables 3 and 8). Groundmass titanomagnetites are richer in ferric iron than the corresponding phenocrysts.
|
Amphiboles
Except in a few basanites, amphibole is ubiquitous as phenocrysts and microphenocrysts in Ua Pou lavas. The mineral is scarce and shows partly resorbed rims in basanites and tephrites, but it is very abundant in tephriphonolites, where it contains ubiquitous apatite inclusions and is, in all cases, partially or entirely replaced by titanomagnetite. In phonolites, amphibole is entirely converted to titanomagnetite aggregates preserving its initial shape.
Representative amphibole compositions are listed in Table 9. Amphiboles are kaersutitic in basanites, tephrites, tephriphonolites, K-benmoreites and trachytes. A few edenitic hornblende crystals also occur in some basanites (Tables 3 and 9). Kaersutites from basanites and tephrites display a rather wide range in Mg numbers [Mg/(Mg + Fe2+), from 0·71 to 0·61] and TiO2 contents (4·23 to 7·94 wt %). Edenitic hornblendes from basanites are more magnesian and less titaniferous than the kaersutites. Besides their lower Mg numbers, kaersutites from tephriphonolites and K-benmoreites are rather similar to those from basanites and tephrites, which display wider ranges in TiO2, Na2O and K2O.
|
|
GEOCHEMICAL RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGERUPTIVE HISTORY OF UA...PETROLOGICAL DATAGEOCHEMICAL RESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Whole-rock major and trace elements: analytical methods
Whole-rock major and trace elements, except Rb, were measured by inductively coupled plasma–atomic emission spectrometry (ICP-AES) at UMR 6538 Domaines Océaniques, Brest, using a Jobin-Yvon system. Rb was measured by flame atomic emission using a Perkin–Elmer 5000 spectrometer. The 68 samples were handcrushed into
1·5 mm grains with an Abisch mortar and then finely powdered using a planetary agate grinder at moderate speed. Contamination during these steps is considered negligible. Powdered samples (250 mg) were then digested with 0·5 ml of HNO3 (70%) and 2·5 ml of HF (40%) in a screw-top Teflon PFA vessel. Then, 97 ml of H3BO3 aqueous solution (20 g/l H3BO3 and 0·5 g/l CsCl) were added to neutralize the excess HF and dissolve the precipitated fluoride, CsCl acting as a buffer of the ionization phenomena in the flame and the argon plasma. The final solution (100 ml) was obtained after a 2 day complexation process of fluoride as fluoboric acid HBF4. All the elements were determined from the final solution without selective extraction, boron being used as an internal standard for ICP-AES analyses. Corresponding detection limits (in ppm) are shown in Table 2. For major elements, relative standard deviation is c. 1% for SiO2 and 2% for the other major elements, except for low values (0·50 wt %), for which the absolute standard deviation is ±0·01 wt %. For trace elements, relative standard deviation is c. 5% for all the trace elements except for Sr and Rb (c. 2%). Results for BE-N standard are shown in Table 2. A more detailed description of the analytical procedures has been given by Cotten et al. (1995).
Additional trace element data were obtained by inductively coupled plasma–mass spectrometry (ICP-MS) at UMR 5025 Grenoble using a Plasma QUAD-Fisons spectrometer. Powdered samples (100 mg) were dissolved for 2 days in a screw-top Teflon bomb using HF and HNO3. After evaporation to dryness of the HF–HNO3 mixture, the samples were taken up in 40 g of HCl and transferred to new, clean bottles. An aliquot of 5 g of these solutions was spiked with a solution of pure Tm (typically 0·4 ml of Tm) and then evaporated to dryness. Then, the samples were taken up in diluted HNO3. All the trace elements were determined from this final solution. Results for BHVO standard are shown in Table 2. More detailed analytical procedures have been described by Barrat et al. (1996).
Major elements
Most of the samples analysed (Table 2) are relatively fresh, as shown by LOI values ranging usually from 0·7 to 3%. However, some samples display evidence of post-magmatic alteration as indicated by the occurrence of secondary calcite, zeolites and iddingsite. In these samples, LOI ranges from 3 to 5%. Samples with LOI >5% are not considered in this study.
In most major element diagrams, e.g. the TAS (Fig. 3) and the K2O–Na2O and K2O–SiO2 plots (Fig. 6a and b), our data set shows an obvious gap between the mafic and intermediate–evolved lava groups, which is evidenced by the lack of phonotephrites and basaltic trachyandesites (mugearites). This corresponds to a SiO2 gap of 2·5% and to a rather remarkable shift in K2O from 2·4% to 4·7%. In addition, the two groups have contrasted K2O/Na2O ratios (Fig. 6a), significantly lower than 1:2 (commonly 1:3 to 1:4) for mafic lavas and higher than 1:2 (usually close to unity) for intermediate–evolved lavas. Our data are generally consistent with those of Bishop & Woolley (1973). Although they used a smaller sample set, those workers described the Ua Pou volcanic association as a continuous alkali basalt–trachyte–phonolite series.
|
The Harker diagrams for major elements (not shown) display, besides the Daly gap discussed above, systematic decreases of TiO2, total iron (FeO*), MgO, CaO and P2O5 abundances with increasing SiO2 contents. Within the phonolitic group, Al2O3 abundance decreases as SiO2 content increases. This group displays a wide range of Na2O contents from c. 5% to more than 11% (Fig. 6c). Some phonolite analyses display sums of alkalis higher than 16% (Table 2 and Fig. 3), values unmatched in the South Pacific as previously pointed out by Bishop & Woolley (1973)
.
Trace elements
Compatible trace element contents of most Mg-rich basanites (MgO >8%) fit the range predicted for basaltic magmas in equilibrium with a lherzolitic source, e.g. 50–70 ppm Co and 200–500 ppm Ni (Allègre et al., 1977; Villemant et al., 1981). Abundances of these elements rapidly decrease within the mafic lava group with increasing SiO2 contents.
Incompatible multielement patterns normalized to the primitive mantle of Sun & McDonough (1989) are shown in Fig. 7 for selected lavas. The tholeiitic sample UP33 differs from all others. It has a relatively flat pattern, similar to those of previously analysed samples from the same unit (Duncan et al., 1986; Dupuy et al., 1987; Fig. 7a). Basanites display patterns with fractionated rare earth element (REE) spectra and strong negative anomalies in Pb and K (Fig. 7b). Tephrites show quite similar patterns, with less pronounced negative K anomalies. Middle REE (MREE) are more fractionated in tephriphonolites than in other mafic rock types. Tephriphonolites also display higher concentrations of the most incompatible trace elements without noticeable negative anomalies in K and Pb, clear negative P and Ti anomalies and heavy REE (HREE) and Y contents lower than those of basanites and tephrites (Fig. 7b). Patterns of K-benmoreites and trachytes shown in Fig. 7c are rather similar, except that Ba, Sr, P and Ti display strong negative anomalies in the trachytes. Selected patterns of phonolites are shown in Fig. 7d and e. They present very spiky patterns with variable but pronounced negative anomalies in Ba, K, Sr, P and Ti. Their REE are heavily depleted relative to Zr and Hf.
|
The trace element patterns of syenites and nepheline syenite (Fig. 7f) are rather different from those of phonolites. They are characterized by negative anomalies in Th and Zr but not in Sr.
Plots of selected incompatible trace elements against Th contents and MgO/(MgO + FeO*) weight ratios are shown in Fig. 8. Th abundances increase regularly in the Ua Pou series without any significant gaps (except within Group D phonolites). However, the patterns of trace element abundances against Th contents do not match the succession of petrographic types defined using the TAS classification (e.g. the Th contents of some tephriphonolites or phonolites are equivalent to those of some basanites). The weight ratio MgO/(MgO + FeO*) seems to be a better indicator of differentiation as, with the exception of the most evolved phonolites, which are characterized by very low MgO, it fits with the petrographic succession. These plots show features usually considered as consistent with fractionation-related processes. For instance, Nb and La contents show rough positive correlations with Th abundances and negative correlations with MgO/(MgO + FeO*) whereas Sr, Ba and Eu contents increase then decrease, as in most series where feldspar fractionation plays an important role. Plots of Th and Zr abundances against MgO/(MgO + FeO*) show generally negative correlations but some samples, together with syenites and nepheline syenite, plot well below the main trend.
|
Sr–Nd–Pb isotopes
Analytical methods
Sr, Nd and Pb isotopic separations were performed in Grenoble on rock chips that were strongly leached with 6N HCl and dissolved using HF and HNO3. Chemical separation of Pb was performed using the classical anionic resin technique of Manhès et al. (1984)
, whereas light REE (LREE) and Sr were isolated using cationic resin. Nd was further separated from Sm using LnSpec Eichrom resin and the Sr fraction was cleaned up using SrSpec Eichrom resin. Total blanks were lower than 90 pg for Pb, 45 pg for Sr and 50 pg for Nd, and were therefore negligible relative to the amounts of elements present in the samples. All samples were loaded and run on a FINNIGAN MAT 261 mass spectrometer (in Brest) for Sr, on a Triton spectrometer (in Brest) for Nd and on a P54 VG multi-collector ICPMS system (in Lyon) for Pb measurements. Complete duplicate analyses, including dissolution and chemistry, were performed for three samples.
Results
Sr, Nd and Pb isotopic data are given in Table 10, together with standard values, and data are plotted in Fig. 9. Ua Pou lavas display a large degree of isotopic heterogeneity. They plot within or near the previously determined Ua Pou field in the 143Nd/144Nd vs 87Sr/86Sr and 208Pb/204Pb vs 206Pb/204Pb diagrams (Fig. 9). Olivine tholeiite UP33 has the least radiogenic 87Sr/86Sr ratio (0·702871), the most radiogenic 206Pb/204Pb ratio (20·142) and plots close to the HIMU end-member, confirming previously published results for samples from the same unit (Dupuy et al., 1987). The alkaline lavas have rather homogeneous 206Pb/204Pb ratios whereas their 87Sr/86Sr, 143Nd/144Nd and 208Pb/204Pb compositions vary between DMM and EM II end-members. The Sr–Nd–Pb compositions of most of the intermediate and evolved lavas (e.g. phonolites UP8, UP64 and UP112; Fig. 9) are nearly identical to that of the mafic lavas. However, the isotopic compositions of some intermediate–evolved lavas deviate significantly from the field of the mafic lavas. For instance, tephriphonolite UP56 and phonolites UP14 and UP95 have less radiogenic 87Sr/86Sr ratios than the basanites and are displaced towards more radiogenic 206Pb/204Pb values. Other phonolites have higher (UP59 and UP76) or lower (UP31) 87Sr/86Sr ratios and deviate from the field of mafic lavas towards the EM II and DMM end-members, respectively.
|
|
Intermediate and evolved lava types
Four groups of intermediate and evolved silica-undersaturated lavas have been recognized acccording to their Sr–Nd–Pb isotope signatures and major and trace element features.
The first group (Group A) corresponds to tephriphonolites and to non-peralkaline lower and upper phonolites. It corresponds to lavas plotting within the upper trend in the Zr vs Th diagram (Fig. 8), which are characterized by regular variations of major and trace elements together with isotopic signatures identical to those of the basanites (Fig. 9).
The second group (Group B) is exclusively phonolitic, with both non-peralkaline and peralkaline phonolites (lower and upper ones). The trace element features of this group are similar to those of Group A, but MREE contents are lower (Fig. 7d). Samples from this group are also rich in Na2O, especially the peralkaline phonolites (6·5% < Na2O < 8%; Fig. 6c). They plot within the upper trend in the Zr vs Th diagram (Fig. 8). The analysed lavas from this group display the highest 87Sr/86Sr ratio of our set together with a 143Nd/144Nd ratio equivalent to that of Group A non-peralkaline phonolites (Fig. 9).
Group C is composed of tephriphonolites, non-peralkaline and peralkaline phonolites (only upper ones) that plot within the alkali-rich and SiO2-poor trend in the TAS diagram (e.g. SiO2 54–55·5%, total alkali 14–16·8% for phonolites; Fig. 3). They are relatively enriched in Al2O3 and in Na2O (5% < Na2O < 10%; Fig. 6c) with respect to Group A and Group B phonolites. They belong to a group of lavas that define a trend depleted in Th and Zr (Fig. 8). Their Th/Zr ratios (0·04–0·05; Fig. 10) as well as other ratios of incompatible elements vs MgO/(MgO + FeO*) are remarkably constant, with the exception of sample UP69, which might have experienced larger amphibole fractionation (Fig. 11). Tephriphonolite UP56 and phonolites UP14 and UP95 from this group have the most radiogenic 206Pb/204Pb and the least radiogenic 87Sr/86Sr ratios observed in our sample set (Fig. 9).
|
|
Group D comprises the three most evolved peralkaline phonolites, which are exposed as upper phonolitic protrusions (Th 51–85 ppm; Fig. 8). They are highly enriched in Na2O (>8·5%), Th and Zr, and display the highest Th/Zr (0·05–0·06) ratio of our sample set (Fig. 10). Sample UP31 from this group has the most unradiogenic 87Sr/86Sr and 206Pb/204Pb ratios observed in our sample set (Fig. 9).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGERUPTIVE HISTORY OF UA...PETROLOGICAL DATAGEOCHEMICAL RESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
General
As discussed in the Introduction, the petrogenetic processes controlling the evolution of the basanite–phonolite series (or their mildly silica-undersaturated equivalents, alkali basalt–trachyte series) are widely debated, especially in connection with the still problematic origin(s) of the Daly gap. Most researchers would agree that phonolites can be derived from the fractional crystallization of basanitic magmas (Le Roex et al., 1990
; Weaver, 1990; Thompson et al., 2001). In a somewhat extreme case, both types can occur in single hand-size samples that have experienced vapour-differentiation processes (Caroff et al., 1999). However, fractional crystallization models do not easily explain bimodal series, nor associations dominated by evolved lavas. In the framework of this debate, the Ua Pou series displays several specific features that might allow better constraints on its petrogenesis.
- Given the short geological history of the emerged part of the island and the overlap between the ages of the various units, one can consider that mafic and intermediate–evolved magmas erupted simultaneously over a time span of c. 0·6 Myr. Interestingly, intermediate lava flows (tephriphonolites and K-benmoreites) that are geochemically close to phonolites and trachytes are not associated with them in the field. Indeed, they are interbedded with mafic lavas, within either the laharic breccia pile or the mafic–intermediate flow sequence.
- The intermediate–evolved lavas are clearly much more abundant (especially the phonolites, which make up 65% of the island) than the associated mafic lavas, a rather uncommon feature in oceanic islands.
- The majority of the mafic and evolved lavas share common Sr, Nd and Pb isotopic signatures, a feature that suggests that they are derived (directly or not) from a similar enriched mantle source.
- The series displays a typical Daly gap (lack of phonotephrites). Some major (K2O) and trace (Pb) elements exhibit abrupt shifts from mafic to evolved lavas, which seem inconsistent with simple fractionation processes. On the other hand, most major and trace element patterns of Ua Pou lavas are similar to those of the basanite–phonolite series thought to result from fractional crystallization, as already pointed out by earlier workers (Bishop & Woolley, 1973; Brousse, 1978).
- Finally, the fact that some samples (especially phonolitic ones) display distinctive isotopic and trace element signatures suggests that complex open-system processes might have controlled their genesis.
Sources of primitive basalts
Major and trace element data allow us to identify two kinds of primitive or near-primitive mafic magmas in Ua Pou. Both are rich in MgO (6–10·25%), Co (40–60 ppm) and Ni (100–425 ppm). They are olivine tholeiites (sample UP33) and basanites. Two sub-types can be distinguished in the latter group, according to the presence or absence of amphibole phenocrysts.
The Marquesas ocean island basalts (OIB) are isotopically heterogeneous (Vidal et al., 1984) and the variations in their Sr, Nd and Pb signatures are adequately explained by multicomponent mixing processes involving HIMU, EM II and DMM mantle end-members, as noted by Duncan et al. (1986), Dupuy et al. (1987) and Vidal et al. (1987). According to these earlier workers the emplacement of the tholeiites and alkali basalts of Ua Pou was separated by a 1·25 Myr hiatus. They showed that the younger alkali basalts derive from a mantle source enriched in the EM II end-member relative to the source of the tholeiites. Similar relationships are found in Hiva Oa and Nuku Hiva, where the shield phase is characterized by an important contribution from a DMM end-member and the post-shield phase is characterized by a more significant involvement of the EM II end-member (Le Dez et al., 1996; Legendre et al., 2005). On the island of Ua Pou, no shield or post-shield phases can be observed, and the tholeiites (4·00 Ma, sample UP33) display a rather typical young HIMU signature characterized by elevated 206Pb/204Pb (>20·5) and relatively low 207Pb/204Pb ratios (Chauvel et al., 1992; Thirlwall, 1997).
Olivine tholeiites of Ua Pou are likely to have been derived from a source different from that which produced the alkali basalts of the island. The young HIMU signature of the tholeiites is interpreted to indicate the presence of a subducted oceanic crust component recycled into the Marquesan hotspot. Ua Pou is the only island of the Marquesas where the HIMU signature is so clearly expressed. However, although more diluted, this signature has also been recognized in lavas from Eiao islet (Caroff et al., 1995). Ua Pou olivine tholeiites could thus be considered to represent an end-member source component of the Marquesas magmas, the mixing of which with DMM and EM II end-members contributes to the large geochemical heterogeneity observed in the Marquesas archipelago (Desonie et al., 1993).
Closed-system fractional crystallization
Fractionation within the mafic magma group
The oldest basanites (intercalated within the laharic breccias) and the more recent ones (from the basanitic to intermediate lava flow pile) do not display significant mineralogical or geochemical differences and their Sr–Nd–Pb isotopic ratios are rather homogeneous. Major and trace element variation is regular from primary to evolved basanites for the two basanitic types (with or without amphibole), but some incompatible trace elements, particularly MREE, are less enriched in amphibole-bearing basanites. As amphiboles have relatively high partition coefficients for MREE (Wörner et al., 1983; Caroff et al., 1993), such a feature is consistent with fractionation of amphibole within this group.
A simple model of closed-system fractional crystallization (CSFC) following Rayleigh's law has been tested to account for the derivation of evolved basanites from primitive basanites:
 | (1) |
The compositions of the most primitive [i.e. displaying the highest MgO/(MgO + FeO*) ratio] amphibole-free basanite UP20 and amphibole-bearing basanite UP42 were taken as parental liquids and those of the most evolved [lowest MgO/(MgO + FeO*) ratio] amphibole-free basanite UP49 and amphibole-bearing basanite UP81 as daughter liquids. Microprobe analyses of phenocrysts from the two daughter liquids were used for least-squares mass balance calculations based on major elements following the method of Bryan et al. (1969) and Wright & Doherty (1970). The results are shown in Table 11. Sums of the squares of residuals are lower than unity for the two models. In the case of amphibole-free basanites, the cumulate contains olivine, clinopyroxene and magnetite, and for amphibole-bearing basanites it contains the same minerals plus plagioclase, apatite and significant percentages of amphibole (20%). Bulk distribution coefficients (D) can be estimated by using the proportions of minerals in the cumulates in combination with mineral–liquid distribution coefficients from the literature [Henderson (1984) for olivine, clinopyroxene, plagioclase and apatite and Caroff et al. (1993) for magnetite and amphibole]. Parameters F and D have been put into equation (1) to calculate the trace element concentrations of the liquids derived from the parental liquids. The results are shown in Fig. 12a and b as incompatible multi-element patterns. The good agreement between observed and calculated trace element patterns for the two models suggests that the derivation of evolved basanites from primitive basanites is consistent with a closed-system fractional crystallization process involving variable amounts of amphibole. However, these models cannot be regarded as providing strong constraints because of the wide range of literature-derived distribution coefficients and the number of fractionating phases involved (six for the amphibole-bearing basanites).
|
|
Lack of evidence for the derivation of intermediate–evolved lavas through CSFC of mafic magmas
Most major and trace element data available on Ua Pou lavas might seem consistent with an evolution from basanite to phonolite controlled by CSFC, a hypothesis formerly proposed by Bishop & Woolley (1973)
. Basanites, the majority of intermediate lavas and trachytes share a common Sr, Nd and Pb isotopic signature, which suggests that they come from similar enriched mantle sources and that none of them have suffered contamination. However, CSFC (or also open-system fractional crystallization) would normally lead to the emplacement of small volumes of evolved lavas whereas Ua Pou displays an unusually high proportion of phonolites. The fact that K2O and Pb contents exhibit abrupt shifts from mafic to evolved lavas is also hardly consistent with CSFC. For instance, when plotted on log(K2O) vs log(Th) and log(Pb) vs log(Th) diagrams, data for basaltic lavas and tephriphonolites display regression lines with slopes clearly higher than unity (1·75 and 1·21, respectively). In the framework of CSFC, this would indicate that K2O and Pb are more incompatible than Th, which would be unusual. An alternative explanation might be that K2O and Pb concentrations have been increased by an additional process. Nevertheless, we have tested numerous CSFC models in an effort to account for the transition from basanites or tephrites to tephriphonolites. All these tests gave sums of squares of residuals ranging from 4·55 to 15·1, and can thus be considered inconsistent with any CSFC model.
Major elements show regular variations from K-benmoreites to trachytes and corresponding Sr–Nd–Pb isotopic compositions are similar (Fig. 9). These features would appear to be consistent with derivation of trachytes through CSFC from K-benmoreites. For instance, Ba and Sr display strong negative anomalies in trachytes (Fig. 7), suggesting an important fractionation of plagioclase and K-feldspar. However, the trace element patterns of K-benmoreites and trachytes are roughly similar and some incompatible elements such as Th, Nb and Zr are lower in trachytes than in K-benmoreites. These features are difficult to reconcile with an origin of trachytes through fractional crystallization of K-benmoreites but nevertheless their closely related trace element patterns suggest that both types have a similar origin.
Fractionation within the intermediate–evolved magmas: origin of Group A phonolites
Group A phonolites have major and trace element features similar to those of tephriphonolites. Both groups plot within the upper trend in the Zr vs Th diagram (Fig. 8). Their Sr–Nd–Pb isotopic ratios are similar (Fig. 9). All these features are consistent with the derivation of Group A non-peralkaline phonolites from tephriphonolites by CSFC processes.
Mantle-normalized trace element patterns of Group A tephriphonolites and phonolites are shown in Fig. 7b and d. Phonolites are depleted in MREE (from Nd to Er) and in LREE with respect to tephriphonolites. These features are consistent with separation of minerals that could fractionate MREE and LREE, i.e. amphibole and/or apatite. Group A phonolites are also characterized by relatively low Nb concentrations, suggesting fractionation of amphibole and/or titanomagnetite, and by negative Eu anomalies possibly linked to plagioclase and/or K-feldspar fractionation.
Equation (1) can be used to test CSFC models for the derivation of these non-peralkaline phonolites from tephriphonolites parents. The less evolved [i.e. displaying the highest MgO/(MgO + FeO*) ratio] tephriphonolite UP71 and the most evolved non-peralkaline phonolite UP112 [with the lowest MgO/(MgO + FeO*) ratio] have been selected as parent and daughter liquids, respectively. Least-squares calculations based on major elements are shown in Table 11. The sum of the squares of residuals is lower than unity (
r2 = 0·345) and minerals extracted from the parental liquid to obtain the derived liquid are amphibole, anorthoclase, plagioclase, clinopyroxene, magnetite and apatite. Mineral–liquid distribution coefficients taken from the literature [Wörner et al. (1983) for plagioclase and amphibole, Henderson (1984) for K-feldspar and clinopyroxene, Caroff et al. (1993) for magnetite, and Chazot et al. (1996) for apatite] have been put into equation (1) to calculate the trace element concentrations of the daughter liquid derived from the parental liquid. The results are shown in Fig. 12c as incompatible multielement patterns. The composition of the calculated liquid is almost identical to that of non-peralkaline phonolite UP112 except for P and Ba. These discrepancies might be due to underestimation of the proportion of apatite in the cumulate or of the Ba distribution coefficients used [3·02 for K-feldspar (Henderson, 1984); 1·08 for plagioclase and 0·86 for amphibole (Wörner et al., 1983)].
This model is consistent with the genesis of Group A non-peralkaline phonolites from tephriphonolites by CSFC mainly governed by fractionation of amphibole and feldspar. However, in several plots of trace element ratios vs MgO/(MgO + FeO*) (Fig. 11) a scatter is observed for Group A phonolites. This scatter could be explained by variations of mineral proportions in the cumulate involved in the CSFC process. The arrows in the small inset boxes in Fig. 11 show the effects of fractionation of the main minerals that could be involved (amphibole, apatite, titanite, K-feldspar) on La/Th, Th/Yb and Nb/Zr ratios. These trends suggest that the fractional crystallization model tested (from tephriphonolite UP71 to non-peralkaline phonolite UP112) corresponds to the maximum proportion (c. 40%) of amphibole within the cumulate and that the scatter observed for the other phonolites indicates varying but lesser fractionation of this mineral for some Group A phonolites.
Partial melting: origin of intermediate lavas and trachytes
Numerous workers have demonstrated, on the basis of experimental data (Hay & Wendlandt, 1995; Hay et al., 1995; Kaszuba & Wendlandt, 2000) or the study of natural field associations (Spulber & Rutherford, 1983; Lowenstern & Mahood, 1991; Bohrson & Reid, 1997), that abundant evolved silica- and alkali-rich magmas can be generated by partial melting at relatively shallow depths of mafic materials trapped in a deep intrusive network of dykes or sills, and equivalent in composition to the mafic lavas erupted at the surface. The genesis of intermediate lavas and trachytes by the melting at depth of plume-derived basaltic materials could be consistent with several characteristics discussed above (i.e. shifts in K2O and Pb; common isotopic Sr, Nd and Pb signatures). Beneath Ua Pou, intermediate lavas could be generated by the partial melting of gabbroic or doleritic rocks, similar in composition to the basanites erupted in the subaerial part of the volcano, and derived from magmas trapped in the upper part of the oceanic crust or within the volcano itself. The adequacy of such a model can be quantitatively evaluated by mass balance calculations and trace element modelling (Hay et al., 1995).
Partial melting: major element models
The derivation of intermediate lavas by partial melting of basaltic material has been evaluated using the program MONA (Mode Near Analysis; Metzner & Grimmeisen, 1990), which reconstructs the initial modal proportions of mafic solids by combining the composition of mineral phases with that of selected intermediate lavas. The mineral phases are assumed to have the compositions of residual minerals in the source that was in equilibrium with a given intermediate melt. As two types of intermediate lavas (K-benmoreites and tephriphonolites) occur in Ua Pou, two models were tested. We assume that the initial mafic solid has the composition of a representative primitive basanite UP54. The two intermediate lavas correspond to the least evolved [highest MgO/(MgO + FeO*) ratio] K-benmoreite UP66 and the least K2O-enriched tephriphonolite UP48. The compositions of residual minerals (olivine, clinopyroxene, plagioclase, magnetite, amphibole) are taken from microprobe analyses from the basanite assumed to represent the source composition, except for plagioclase and olivine, the compositions of which were obtained by entering their respective mineral end-members anorthite and albite, and forsterite and fayalite into the program. The composition obtained for olivine in equilibrium with the intermediate melt is relatively iron-rich (Fo 73%) whereas that of the corresponding plagioclase is rather sodic (andesine) compared with the Ca-rich basaltic plagioclase, in agreement with albite–anorthite phase relationships. We are aware that these calculations are rather rough and need to be refined with a better knowledge of melting conditions (especially total pressure and water content, which are critical for plagioclase–liquid equilibria).
For each of the two models, sums of the squares of residuals are lower than unity (Table 12) and the proportions of intermediate liquid obtained correspond to low melting degrees (c. 9% for tephriphonolites and 11·5% for K-benmoreites). The most significant residual is obtained for P2O5 because the residue is assumed not to contain apatite (phosphorus, which behaves as an incompatible element, is assumed to be extracted early during the melting process). Removing P2O5 from consideration in the calculations lowers the sum of squared residuals without changing significantly the residue composition or the melting degrees. It is noteworthy that the partial melting model accounts for the observed abrupt shift in K2O contents (between 2·4 and 4·7%) from mafic to intermediate compositions.
|
Thus, our calculations indicate that tephriphonolites and K-benmoreites (respectively silica-undersaturated and silica-saturated) can be derived by partial melting of 9 and 11·5% respectively of a mafic material the composition of which is similar to that of basanite UP54. Compared with the residue left by the extraction of K-benmoreites, that of tephriphonolites contains more amphibole and less olivine (Table 12). The tephriphonolite residue is more amphibole-rich than the K-benmoreite one and the calculated partial melting degree is lower for the former.
Inverse and dehydration melting experiments carried out by Hay & Wendlandt (1995) and Kaszuba & Wendlandt (2000) respectively were performed from 0·5 to 1·0 GPa with H2O and/or CO2 added. The latter researchers observed that, for the same basaltic source, the presence of H2O + CO2 leads to silica-undersaturated evolved lavas whereas the presence of only H2O leads to silica-saturated evolved lavas. High-temperature carbonates have been found within tephriphonolites (Table 4), a feature that indicates the presence of high levels of magmatic CO2. According to these results, similar degrees of melting of a single mafic source might generate both tephriphonolites and K-benmoreites depending on the presence of CO2 + H2O for the former and nearly pure H2O for the latter. The c. 10% melting degrees are consistent with those generally obtained in such modelling (e.g. 10–30%, Bohrson & Reid, 1997).
Partial melting: trace element models
Trace element modelling calculations were performed to assess the validity of the major element model. We used the model of batch melting proposed by Schilling & Winchester (1967):
 | (2) |
Bulk distribution coefficients (D) can be estimated by combining the proportions of minerals within the residue (obtained by the major element models and recalculated to 100%) with mineral–liquid distribution coefficients from the literature. Mineral–liquid distribution coefficients are from Henderson (1984) for olivine, clinopyroxene, plagioclase and amphibole, from Caroff et al. (1993) for magnetite, and from White (1997) and Blundy et al. (1998) for Pb and U. Parameters F and D have been put into equation (2) to calculate the theoretical trace element concentrations of the liquids generated by partial melting of the assumed mafic source. The results are shown in Fig. 13 as trace element patterns. The patterns of calculated liquids are very similar to those of the tephriphonolites and K-benmoreites. The calculated trace element contents are a little higher than the observed values but the shapes of the patterns are nearly identical.
|
K-benmoreites and trachytes display very similar patterns, except for Ba, Sr, Eu and Gd (Fig. 13), a feature that suggests that the trachytes might have originated from lower partial melting degrees of mafic material compared with K-benmoreites, and that partial melts were then modified by fractionation of plagioclase and/or K-feldspar. The derivation of trachyte UP94 from partial melting of basanite UP54 has been evaluated using major and trace elements. The sum of the squares of residuals is lower than unity (
r2 = 0·53), and the composition of the residue is 37·9% plagioclase, 27·6% clinopyroxene, 10·2% amphibole, 8·5% titanomagnetite and 7·8% olivine. The corresponding melting degree is lower (c. 8%) than that obtained for K-benmoreites. Nevertheless, the composition of the calculated liquid obtained by trace element modelling does not account for the strong Sr negative anomaly observed in the trachytes. This feature could indicate that partial melts derived from mafic materials were subsequently modified by fractionation of plagioclase to generate trachytes.
The negative Pb anomaly observed in the basanites is preserved in the calculated intermediate lavas, whereas it does not exist in the natural lavas. Thus, even if the partial melting model accounts for the genesis of the intermediate lavas, it does not account for Pb contents of tephriphonolites and K-benmoreites. Additional processes have to be evoked to explain the Pb enrichment within the intermediate lavas. In particular, the occurrence of water–magma interactions can be envisioned because Pb is highly soluble in high-temperature hydrous fluids (Kogiso et al., 1997).
Wall-rock assimilation
Open-system differentiation effects
Group C phonolites are isotopically different from Group A phonolites. They also display distinctive elemental characteristics (e.g. lower contents of Th and Zr, Fig. 8; Na2O enrichment, Fig. 6c; and high Th/Zr, Fig. 10). These features are not consistent with closed-system fractionation and suggest that their genesis might have been controlled by open-system fractionation processes.
Compared with the other basanitic lavas, basanite UP19 is Na2O-enriched (6 wt %; Fig. 6c) and plots within the lower trend in the Zr vs Th diagram (Fig. 8). Its Th/Zr ratio (Th/Zr = 0·05; Fig. 10) and other ratios of incompatible elements vs MgO/(MgO + FeO*) are remarkably identical with those of Group C tephriphonolites and phonolites (Fig. 11). These features suggest that this sample could also have undergone some degree of contamination.
AFC models
DePaolo's (1981) equation (6a) can be rearranged to calculate the composition of the contaminant involved in an AFC process:
 | (3) |
- F values are given by mass balance calculations: F
0·5 for basanite UP19 (mineral assemblage: 19% olivine, 38% clinopyroxene, 12% plagioclase, 9% titanomagnetite, 1·6% apatite, 20% amphibole) and F 0·6 for Group C non-peralkaline phonolite UP13 (mineral assemblage: 13% clinopyroxene, 35% plagioclase, 4% titanomagnetite, 30% K-feldspar, 0·7% apatite, 18% amphibole). Relatively high sums of squared residuals are obtained (1·43 and 4·42, respectively) because simple fractionation models are not consistent with the Na2O enrichment of the contaminated lavas, which could instead result from an open-system differentiation effect.
- We have postulated that r can vary within a ‘reasonable’ range, i.e. from 0·2 to 0·5.
- The bulk distribution coefficients D have been approximated by combining the mineral weight fractions determined from least-squares calculations and the individual mineral–liquid distribution coefficients taken from the literature and used for previous CSFC models. The bulk distribution coefficients D used in the modellings for basanite UP19 are 0·2 for Ba, 0·05 for Th, 0·38 for Nb, 0·25 for La, 0·28 for Ce, 0·4 for Sr, 0·4 for Nd, 0·15 for Zr, 0·61 for Sm, 0·55 for Eu, 0·62 for Gd, 0·62 for Dy, 0·64 for Y, 0·57 for Er, and 0·52 for Yb. In the case of non-peralkaline phonolite UP13 we have used D values of 1·42 for Ba, 0·28 for Th, 0·77 for Nb, 0·72 for La, 0·86 for Ce, 0·61 for Sr, 1·06 for Nd, 0·25 for Zr, 1·42 for Sm, 1·33 for Eu, 1·36 for Gd, 1·33 for Dy, 0·91 for Y, 1·1 for Er, and 0·99 for Yb.
The incompatible trace element composition of the contaminant has been calculated by putting these values into equation (3). The method has been applied assuming the derivation of basanite UP19 from basanite UP20 and the derivation of Group C non-peralkaline phonolite UP13 from tephriphonolite UP71. The latter samples were selected for modelling because they are the least differentiated of their category in our set according to their MgO/(MgO + FeO*) ratio. The results are shown as primitive mantle-normalized elements patterns in Fig. 14a and b for the basanitic and intermediate–evolved pairs, respectively.
|
For the basanitic pair and for low values of r (r = 0·2–0·3, not shown), the modelling yields contaminants with very low Dy contents together with negative values for HREE (Er and Yb). For higher values of r (r = 0·4–0·5, Fig. 14a), all the calculated contaminants display a strong LREE enrichment, negative anomalies in Th and Yb, and near-zero values for Zr and Er.
For the intermediate–evolved pair and for low values of r (r = 0·2–0·3, not shown), the modelling yields contaminants with very low Yb contents together with negative Ce and positive Sr anomalies. For higher values of r (r = 0·4–0·5, Fig. 14b), the patterns of all the calculated contaminants are strongly enriched in the highly incompatible elements and near-zero values are obtained for Th and Zr. They are relatively similar to the contaminants calculated for the basanitic pair except for their more enriched patterns, their more pronounced negative Th anomaly and the lack of negative Er anomaly.
Two contaminant compositions corresponding to different values of r (r = 0·4 and r = 0·5) are shown for each pair in Fig. 14. However, the patterns of these two contaminants are very similar and they will not be considered separately in the following discussion.
Possible nature of the contaminant
The trace element compositions of the calculated contaminants are shown as primitive mantle normalized element patterns in Fig. 15. The calculated patterns (for basanite UP19 and Group C non-peralkaline phonolite UP13) are strongly fractionated with respect to HREE. The phonolite contaminant is more enriched than that for the basanite, but both display similar slopes and Th and Zr negative anomalies. The basanite contaminant also differs from the phonolitic one by its less pronounced Th negative anomaly and its negative Er anomaly.
|
Compositions of syenitic blocks (UP302, 303, 305) and nepheline syenite xenolith UP57c are shown in Fig. 15 for comparison. They are characterized by high LREE/HREE ratios and slight Th and strong but variable Zr negative anomalies. Their patterns are relatively similar, or at least parallel to, those of calculated contaminants, except for the more pronounced Th and Zr negative anomalies and the very low HREE contents of the basanite contaminant. Thus, we propose that the calculated contaminants could be equivalent to the syenites and nepheline syenite that were sampled on Ua Pou island. More precisely, nepheline syenite UP57c seems to be more appropriate as a contaminant than the syenites because of its highly sodic and silica-undersaturated character. The assimilation of such a rock would be consistent with the Na2O enrichment together with the increase in silica-undersaturated character observed in the contaminated lavas. The Th and Zr negative anomalies, together with the low HREE contents of the calculated contaminants, may indicate that it has experienced separation of zircon, a mineral observed in the syenitic blocks and the nepheline syenite xenolith, and that is known to behave as a refractory phase during partial melting. If the lava contamination process operated through wall-rock melting and subsequent melt assimilation, one may expect that zircon behaved as a refractory phase and did not contribute to the contamination of Ua Pou lavas.
Other specific cases of contamination
Role of seawater
At equivalent SiO2 contents, Group B non-peralkaline and peralkaline phonolites are clearly enriched in Na2O with respect to Group A phonolites (Fig. 6c). Lavas of the former group are devoid of any petrographic features suggesting alteration processes that could affect Na2O contents (e.g. Bohrson & Reid, 1997). In addition, they do not display petrographic evidence for alkali feldspar accumulation. Phonolites from Group B are clearly depleted in MREE compared with those of Group A (Fig. 7d), suggesting extensive separation of amphibole, and they display negative Ba anomalies consistent with K-feldspar fractionation. Non-peralkaline Group B phonolite samples UP59 and UP76 have higher 87Sr/86Sr isotopic ratios than samples of Group A despite their almost equivalent 143Nd/144Nd ratios (Fig. 9).
This isotopic signature could reflect the incorporation of seawater within Group B phonolites. In particular, it could result either from the interaction between seawater and the mafic rock source of the phonolites, or alternatively from the assimilation by ascending evolved melts of rocks previously altered by hydrothermal circulation involving seawater. As our samples were heavily leached before chemical separation, this signature is not likely to result from sample alteration. We suggest that seawater, or a material with a seawater isotopic imprint, might have been incorporated within the magmatic plumbing system, a hypothesis consistent with the Na2O enrichment of Group B lavas. Such incorporation of water could trigger the fractionation of amphibole, a feature suggested by the MREE depletion of Group B phonolites.
Strongly evolved Group D phonolites
Group D is composed of the three most evolved peralkaline phonolites UP26, UP31 and UP61 [(Na + K)/Al = 1–1·2]. They are clearly enriched in Na2O (Fig. 6c) and Al2O3, and their contents of the most incompatible elements are very high (Th 51–84 ppm, Zr 940–1650 ppm, U 20 ppm). Their Th/Zr ratios are the highest observed in our sample set (Th/Zr = 0·05–0·06; Fig. 10) even though they plot within the lower trend in the Zr vs Th diagram (Fig. 8). The 206Pb/204Pb isotopic ratio of sample UP31 is shifted towards the DMM end-member (Fig. 9). Their major and trace element features are very similar to those of Group C phonolites but their isotopic signatures are rather different.
Several lines of evidence suggest genetic links between Group C and D peralkaline phonolites. Despite their contrasting Pb isotopic signatures, and even if Group D phonolites are far more evolved than their Group C equivalents, their trace element patterns are roughly parallel except for more pronounced Ba and Sr negative anomalies in Group D phonolites (Fig. 7d and e). We suggest that Group D peralkaline phonolites could be derived from Group C phonolites trapped within the crust or within the volcano itself, where they could have experienced extensive fractionation of feldspar (Ba and Sr negative anomalies). This open-system differentiation could have been coupled with assimilation of surrounding wall-rock with a DMM isotopic signature (e.g. the Pacific oceanic crust underlying Ua Pou island).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGERUPTIVE HISTORY OF UA...PETROLOGICAL DATAGEOCHEMICAL RESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
We propose that the exceptional abundance of evolved lavas (especially phonolites) in Ua Pou reflects their derivation from basanites through partial melting rather than fractional crystallization. Fractional crystallization is, however, the most likely process explaining the derivation of basanites and tephrites from primitive basanitic magmas, as well as the evolution from tephriphonolites towards Group A non-peralkaline phonolites. Most of the Ua Pou phonolites probably are derived from open-system fractional crystallization of tephriphonolitic magmas, possibly associated with wall-rock assimilation of nepheline syenites and syenites trapped within the underlying oceanic lithosphere or within the volcano basement.
The combination of petrogenetic processes responsible for the genesis of Ua Pou lavas is schematically depicted in Fig. 16.
|
Stage 1. At c. 4 Ma, the partial melting of a mantle plume displaying a young HIMU signature produced Ua Pou olivine tholeiites.
Stage 2. Magmatic activity resumed at c. 2·9 Ma when melting of a heterogeneous mantle with a prominent EM II + HIMU signature produced primitive basaltic melts. We propose that these melts either were emplaced at depth or at the surface, or evolved by CSFC towards basanites and tephrites (and possibly towards nepheline syenites and syenites). Minor contamination linked to open-system fractional crystallization could have led to the generation of basanite sample UP19. We suggest that the remelting of the basanites at depth during stage 2 produced tephriphonolitic magmas, leaving an amphibole-rich residuum. These magmas could evolve by CSFC towards Group A non-peralkaline phonolites, and then towards Group B non-peralkaline and peralkaline phonolites by open-system fractional crystallization involving minor contamination by seawater.
Stage 3. The genesis and emplacement of Group A and B phonolites was continuing, but possible additional open-system fractional crystallization processes were involved between c. 2·6 and 2·4 Ma. These could have produced either contaminated Group C tephriphonolites and phonolites by assimilation of nepheline syenites and syenites, or Group D peralkaline phonolites where extreme fractionation was coupled with assimilation of oceanic crustal materials.
The study of Ua Pou allows us to document a rather atypical petrogenetic history of an oceanic basanite–phonolite series. Several of its features may appear at first to be consistent with intermediate and evolved lavas originating through fractional crystallization of mafic magmas, as proposed by many earlier workers. Indeed, mafic, intermediate and evolved magmas erupted simultaneously during the short geological history of the island (c. 0·6 Myr); the majority of the mafic and evolved lavas (Group A phonolites) share common isotopic signatures and ultimately derive from a similar mantle source. However, the exceptional abundance of phonolites (65% of the surface of the island), the paucity of intermediate lavas (lack of phonotephrites) and the abrupt shift in K2O from 2·4% to 4·7% observed at the level of the Daly gap in the Ua Pou series are not consistent with the fractional crystallization hypothesis. The extensive remelting at depth of basanitic materials, which produced tephriphonolitic magmas and left amphibole-rich residues, appears to be the most plausible petrogenetic process accounting for both the Daly gap and the unusual amount of evolved lavas.
|
ACKNOWLEDGEMENTS |
|---|
This study is part of the Ph.D. thesis of the senior author at the Université de Bretagne Occidentale in Brest. Sampling and mapping were carried out in 2001, 2002 and 2003 with the financial and logistic help of BRGM, CEA/DASE, UMR 6538 Domaines océaniques and Géosciences Rennes. We thank M. Bohn (Brest) for his contribution to microprobe analyses. The detailed reviews and pertinent suggestions of Drs R. C. Price and R. Wendlandt are gratefully acknowledged.
* Corresponding author. Present address: Laboratoire de Planétologie et de Géodynamique UMR-6112, Université de Nantes, 2 rue de la Houssinière, BP 92208, 44322 Nantes Cedex 3, France. Telephone: (33) 2 51 12 52 82. Fax: (33) 2 51 12 52 68. E-mail: christelle.legendre{at}univ-nantes.fr
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGERUPTIVE HISTORY OF UA...PETROLOGICAL DATAGEOCHEMICAL RESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Ablay, G. J. & Martí, J. (2000). Stratigraphy, structure and volcanic evolution of the Pico Teide–Pico Viejo formation, Tenerife, Canary Islands. Journal of Volcanology and Geothermal Research 103, 175–208.[CrossRef][Web of Science]
Allègre, C. J., Treuil, M., Minster, J. F., Minster, B. & Albarède, F. (1977). Systematic use of trace element in igneous processes. Part 1: fractional crystallisation processes in volcanic suites. Contributions to Mineralogy and Petrology 60, 57–75.[CrossRef][Web of Science]
Bailey, D. K. (1987). Mantle metasomatism—perspective and prospect. In: Fitton, J. G. & Upton, B. G. J. (eds) Alkaline Igneous Rocks. Geological Society, London, Special Publications 30, 1–14.
Baker, I. (1968). Intermediate oceanic volcanic rocks and the ‘Daly gap’. Earth and Planetary Science Letters 4, 103–106.[CrossRef][Web of Science]
Barrat, J. L., Keller, F., Amossé, J., Taylor, R. N., Nesbitt, R. W. & Hirata, T. (1996). Determination of rare earth elements in sixteen silicate reference samples by ICP-MS after Tm addition and ion exchange separation. Geostandards Newsletter 20, 133–139.[Web of Science]
Bishop, A. C. & Woolley, A. R. (1973). A basalt–trachyte–phonolite series from Ua Pou, Marquesas islands, Pacific Ocean. Contributions to Mineralogy and Petrology 39, 309–326.[CrossRef][Web of Science]
Blundy, J. D., Robinson, J. A. C. & Wood, B. J. (1998). Heavy REE are compatible in clinopyroxene on the spinel lherzolite solidus. Earth and Planetary Science Letters 160, 493–504.[CrossRef][Web of Science]
Bohrson, W. A. & Reid, M. R. (1997). Genesis of silicic peralkaline volcanic rocks in an oceanic island setting by crustal melting and open-system processes: Socorro Island, Mexico. Journal of Petrology 38, 1137–1166.[CrossRef][Web of Science]
Bonnefoi, C. C., Provost, A. & Albarède, F. (1995). The ‘Daly gap’ as a magmatic catastrophe. Nature 378, 270–272.[CrossRef]
Brousse, R. (1978). Eléments d'analyses de quelques îles des Marquises: Fatu Hiva, Tahuata, Motane, Ua Pou. Cahiers du Pacifique, Volume Spécial ‘Marquises’. Paris: Fondation Singer-Polignac, pp. 107–144.
Brousse, R. & Maury, R. C. (1978). Minéralogie d'une phonolite à grenat de l'île de Ua Pou (Marquises, Pacifique central). Bulletin de Minéralogie 101, 3–10.
Brousse, R., Barsczus, H. G., Bellon, H., Cantagrel, J. M., Diraison, C., Guillou, H. & Leotot, C. (1990). Les Marquises (Polynésie française): volcanologie, géochronologie, discussion d'un modèle de point chaud. Bulletin de la Société Géologique de France 6, 933–949.[Web of Science]
Bryan, W. B., Finger, L. A. & Chayes, F. (1969). Estimating proportions in petrologic mixing equations by least square approximation. Science 163, 926–927.
Caress, D. W., McNutt, M. K., Detrick, R. S. & Mutter, J. C. (1995). Seismic imaging of hotspot related underplating beneath the Marquesas Islands. Nature 373, 600–603.[CrossRef]
Caroff, M., Maury, R. C., Leterrier, J., Joron, J. L., Cotten, J. & Guille, G. (1993). Trace element behaviour in the alkali basalt–comenditic trachyte series from Mururoa atoll, French Polynesia. Lithos 30, 1–22.[CrossRef][Web of Science]
Caroff, M., Maury, R. C., Vidal, Ph., Guille, G., Dupuy, C., Cotten, J., Guillou, H. & Gillot, P. Y. (1995). Short temporal changes in oceanic island basalt series: evidence from an 800 m-deep drill hole in Eiao shield (Marquesas). Journal of Petrology 36, 1333–1365.
Caroff, M., Ambrics, C., Maury, R. C., & Cotten, J. (1997). From alkali basalt to phonolite in hand-size samples: vapor-differentiation effects in the Bouzentes lava flow (Cantal, France). Journal of Volcanology and Geothermal Research 79, 47–61.
Caroff, M., Guillou, H., Lamiaux, M., Maury, R. C., Guille, G. & Cotten, J. (1999). Assimilation of ocean crust by hawaiitic and mugearitic magmas: an example from Eiao (Marquesas). Lithos 46, 235–258.[CrossRef][Web of Science]
Cassignol, C. & Gillot, P. Y. (1982). Range and effectiveness of unspiked potassium–argon dating: experimental groundwork and applications. In: Odin, G. S. (ed.) Numerical Dating in Stratigraphy. Chichester: John Wiley, pp. 159–179.
Charbit, S., Guillou, H. & Turpin, L. (1998). Cross calibration of K–Ar standard minerals using an unspiked Ar measurement technique. Chemical Geology 150, 147–159.[CrossRef][Web of Science]
Chauvel, C., Hofmann, A. W. & Vidal, Ph. (1992). HIMU-EM: the French Polynesian connection. Earth and Planetary Science Letters 110, 99–119.[CrossRef][Web of Science]
Chauvel, C., McDonough, W., Guille, G., Maury, R. C. & Duncan, R. (1997). Contrasting old and young volcanism in Rurutu Island, Austral chain. Chemical Geology 139, 125–143.[CrossRef][Web of Science]
Chayes, F. (1963). Relative abundance of intermediate members of the oceanic basalt–trachyte association. Journal of Geophysical Research 68, 1519–1534.
Chayes, F. (1977). The oceanic basalt–trachyte relation in general and in the Canary Islands. American Mineralogist 62, 666–671.[Abstract]
Chazot, G., Menzies, M. A. & Harte, B. (1996). Determination of partition coefficients between apatite, clinopyroxene, amphibole and melt in natural spinel lherzolite from Yemen: implications for wet melting of the lithospheric mantle. Geochimica et Cosmochimica Acta 60, 423–437.[CrossRef][Web of Science]
Chubb, L. J. (1930). The geology of the Marquesas islands. Bernice P. Bishop Museum 68, 1–71.
Cotten, J., Le Dez, A., Bau, M., Caroff, M., Maury, R. C., Dulski, P., Fourcade, S., Bohn, M. & Brousse, R. (1995). Origin of anomalous rare-earth element and yttrium enrichments in subaerially exposed basalts: evidence from French Polynesia. Chemical Geology 119, 115–138.[CrossRef][Web of Science]
Cousens, B. L., Clague, D. A. & Sharp, W. D. (2003). Chronology, chemistry and origin of trachytes from Hualalai volcano, Hawaii. Geochemistry, Geophysics, Geosystems 4(9), 1078, doi: 10.1029/2003GC000560, 1–27.
Crough, S. T. & Jarrard, R. D. (1981). The Marquesas-line swell. Journal of Geophysical Research 86, 11763–11771.[CrossRef]
Daly, R. A. (1910). Origin of alkaline rocks. Geological Society of America Bulletin 21, 87–118.
Daly, R. A. (1925). The geology of Ascension Island. Proceedings of the American Academy of Arts and Sciences 60, 3–80.
Davidson, J. P. & Wilson, I. R. (1989). Evolution of an alkali basalt–trachyte suite from Jebel Marra volcano, Sudan, through assimilation and fractional crystallization. Earth and Planetary Science Letters 95, 141–160.[CrossRef][Web of Science]
Defant, M. J., Richerson, M., de Boer, J. Z., Stewart, R. H., Maury, R. C., Bellon, H., Drummond, M. S., Feigenson, M. D. & Jackson, T. E. (1991). Dacite genesis via both slab melting and differentiation: petrogenesis of La Yeguada volcanic complex, Panama. Journal of Petrology 32, 1101–1142.
DePaolo, D. J. (1981). Trace element and isotopic effects of combined wall rock assimilation and fractional crystallization. Earth and Planetary Science Letters 53, 189–202.[CrossRef][Web of Science]
Desonie, D. L., Duncan, R. A., Nielsen, R. N. & Natland, J. H. (1993). Temporal and geochemical variability of volcanic products of the Marquesas hotspot. Journal of Geophysical Research 98(B10), 17649–17665.[CrossRef]
Devey, C. W., Lackschewitz, K. S., Mertz, D. F., Bourdon, B., Cheminée, J. L., Dubois, J., Guivel, C., Hekinian, R. & Stoffers, P. (2001). Evidence for preferential melting of the enriched components in Polynesian plumes. EOS Transactions, American Geophysical Union 82(47), Fall Meeting, Supplement, Abstract V51C-1025.
Diraison, C. (1991). Le volcanisme aérien des archipels polynésiens de la Société, des Marquises et des Australes–Cook. Téphrostratigraphie, datation isotopique et géochimie comparées. Contribution à l'étude des origines du volcanisme intraplaque du Pacifique Central. Ph.D. thesis, Université de Bretagne Occidentale, Brest, 413 pp.
Duncan, R. A., McCulloch, M. T., Barsczus, H. G. & Nelson, D. R. (1986). Plume vs lithospheric melts at Ua Pou, Marquesas Islands. Nature 322, 534–538.[CrossRef]
Dupuy, C., Vidal, Ph., Barsczus, H. G. & Chauvel, C. (1987). Origin of basalts from the Marquesas archipelago (South Central Pacific Ocean): isotope and trace element constraints. Earth and Planetary Science Letters 117, 89–100.[CrossRef]
Dupuy, C., Barsczus, H. G., Dostal, J., Vidal, P. & Liotard, J. M. (1989). Subducted and recycled lithosphere as the mantle source of ocean island basalts from southern Polynesia, central Pacific. Chemical Geology 77, 1–18.[CrossRef][Web of Science]
Dupuy, C., Vidal, P., Maury, R. C. & Guille, G. (1993). Basalts from Mururoa, Fangataufa and Gambier islands (French Polynesia): geochemical dependence on the age of the lithosphere. Earth and Planetary Science Letters 117, 89–100.[CrossRef][Web of Science]
Esser, R. P., Kyle, P. R. & McIntosh, W. C. (2004). 40Ar/39Ar dating of the eruptive history of Mount Erebus, Antarctica: volcano evolution. Bulletin of Volcanology 66, 671–686.[CrossRef][Web of Science]
Filmer, P. M., McNutt, M. K. & Wolfe, C. J. (1993). Elastic thickness of the lithosphere in the Marquesas and Society islands. Journal of Geophysical Research 98, 19565–19577.[CrossRef]
Filmer, P. M., McNutt, M. K., Webb, H. F. & Dixon, D. J. (1994). Volcanism and archipelagic aprons: a comparison of the Marquesas and Hawaiian islands. Marine Geophysical Research 16, 385–406.[CrossRef]
Freundt, A. & Schmincke, H. U. (1995). Petrogenesis of rhyolite–trachyte–basalt composite ignimbrite P1, Gran Canaria, Canary Islands. Journal of Geophysical Research 100(B1), 455–474.
Fuhrmann, U., Lippolt, H. & Hess, J. C. (1987). HD-B1 Biotite reference material for K–Ar chronometry. Chemical Geology 66, 41–51.[Web of Science]
Gillot, P. Y. & Cornette, Y. (1986). The Cassignol technique for potassium–argon dating, precision and accuracy: examples from the late Pleistocene to Recent volcanics from southern Italy. Chemical Geology 59, 205–222.[Web of Science]
Goff, F. (1996). Vesicle cylinders in vapor-differentiated basalt flows. Journal of Volcanology and Geothermal Research 71, 167–185.[CrossRef][Web of Science]
Govindaraju, K. & Roelandts, I. (1993). Second report (1993) on the first three GIT-IWG rock reference samples: anorthosite from Greenland, An-G; basalte d'Essey-la-Côte, BE-N; granite de Beauvoir, MA-N. Geostandards Newsletter 17(2), 227–294.[Web of Science]
Guille, G., Guillou, H., Chauvel, C., Maury, R. C., Blais, S. & Brousse, R. (1998). L'île de Rurutu (Archipel des Australes, Polynésie française): une édification complexe liée au fonctionnement de deux points chauds. Géologie de la France 3, 65–85.
Guille, G., Legendre, C., Maury, R. C., Caroff, M., Munschy, M., Blais, S., Chauvel, C., Cotten, J. & Guillou, H. (2002). Les Marquises (Polynésie française): un archipel intraocéanique atypique. Géologie de la France 2, 5–36.
Gutscher, M. A., Olivet, J. L., Aslanian, D., Eissen, J. P. & Maury, R. C. (1999). The ‘lost Inca Plateau’: cause of flat subduction beneath Peru? Earth and Planetary Science Letters 171, 335–341.[CrossRef][Web of Science]
Hanan, B. B. & Graham, D. W. (1996). Lead and helium isotope evidence from oceanic basalts for a common deep source of mantle plumes. Science 272, 991–995.[Web of Science][Medline]
Harpel, C. J., Kyle, P. R., Esser, R. P., McIntosh, W. C. & Caldwell, D. A. (2004). 40Ar/39Ar dating of the eruptive history of Mount Erebus, Antarctica: summit flows, tephra and caldera collapse. Bulletin of Volcanology 66, 687–702.[CrossRef][Web of Science]
Hart, S. R. (1984). A large scale isotope anomaly in the southern hemisphere mantle. Nature 309, 753–757.[CrossRef]
Hart, S. R. & Zindler, A. (1989). Constraints on the nature and development of chemical heterogeneities in the mantle. In: Peltier, W. R. (ed.) Mantle Convection. New York: Gordon and Breach, pp. 261–387.
Hart, S. R., Hauri, E. H., Oschmann, L. A. & Whitehead, J. A. (1992). Mantle plumes and entrainment: isotopic evidence. Science 256, 517–520.
Hay, D. E. & Wendlandt, R. F. (1995). The origin of Kenya rift flood phonolites: results of high-pressure/high temperature experiments in the systems phonolite–H2O and phonolite–H2O–CO2. Journal of Geophysical Research 100(B1), 401–410.[CrossRef]
Hay, D. E., Wendlandt, R. F. & Wendlandt, E. D. (1995). The origin of Kenya rift plateau-type flood phonolites: evidence from geochemical studies for fusion of lower crust modified by alkali basaltic magmatism. Journal of Geophysical Research 100(B1), 411–422.[CrossRef]
Henderson, P. (1984). General geochemical properties and abundances of the rare earth elements. In: Henderson, P. (ed.) Rare Earth Element Geochemistry. Development in Geochemistry 2, 1–32.
Ielsch, G., Caroff, M., Barsczus, H. G., Maury, R. C., Guillou, H., Guille, G. & Cotten, J. (1998). Géochimie des basaltes de l'île de Ua Huka (archipel des Marquises): variation du taux de fusion partielle et hétérogénéité de la source mantellique. Comptes Rendus de l'Académie des Sciences 326, 413–420.
Kaszuba, J. P. & Wendlandt, R. F. (2000). Effect of carbon dioxide on dehydration melting reactions and melt compositions in the lower crust and the origin of alkaline rocks. Journal of Petrology 41, 363–386.
Kogiso, T., Tatsumi, Y., Shimoda, G. & Barsczus, H. G. (1997). High µ (HIMU) ocean island basalts in southern Polynesia: new evidence for whole mantle scale recycling of subducted oceanic crust. Journal of Geophysical Research 102(B4), 8085–8103.[CrossRef]
Kyle, P. R., Moore, J. A. & Thirlwall, M. F. (1992). Petrologic evolution of anorthoclase phonolite lavas at Mount Erebus, Ross Island, Antarctica. Journal of Petrology 33, 849–875.
Lacroix, A. (1927). La constitution lithologique des îles volcaniques de la Polynésie Australe. Mémoires de l'Académie des Sciences 5, 1–80.
Lacroix, A. (1928). Nouvelles observations sur les laves des îles Marquises et de l'île Tubuai (Polynésie australe). Comptes Rendus de l'Académie des Sciences 187, 397.
Lacroix, A. (1931). Les phonolites néphéliniques et leucitiques de l'île Ua Pou (Archipel des Marquises). Comptes Rendus de l'Académie des Sciences 20, 1161–1166.
Le Bas, M. J., Le Maitre, R. W., Streckeisen, A. & Zanettin, B. (1986). A chemical classification of volcanic rocks based on the total-alkali–silica diagram. Journal of Petrology 27, 745–750.
Le Dez, A., Maury, R. C., Vidal, Ph., Bellon, H., Cotten, J. & Brousse, R. (1996). Geology and geochemistry of Nuku Hiva, Marquesas: temporal trends in a large Polynesian shield volcano. Bulletin de la Société Géologique de France 167, 197–209.[Abstract]
Legendre, C., Maury, R. C., Savanier, D., Cotten, J., Chauvel, C., Hémond, C., Bollinger, C., Guille, G., Blais, S. & Rossi, Ph. (2005). The origin of intermediate and evolved lavas in the Marquesas archipelago: an example from Nuku Hiva island (French Polynesia). Journal of Volcanology and Geothermal Research (in press).
Le Roex, A. P., Cliff, R. A. & Adair, B. J. I. (1990). Tristan da Cunha, South Atlantic, geochemistry and petrogenesis of a basanite–phonolite lava series. Journal of Petrology 31, 779–812.
Lippard, S. J. (1973). The petrology of phonolites from the Kenya rift. Lithos 6, 217–234.[CrossRef]
Lowenstern, J. B. & Mahood, G. A. (1991). New data on magmatic H2O contents of pantellerites with implications for petrogenesis and eruptive dynamics at Pantelleria. Bulletin of Volcanology 54, 78–83.[CrossRef][Web of Science]
Manhès, G., Allègre, C. J. & Provost, A. (1984). U–Th–Pb systematics of the eucrite ‘Juvinas’: precise age determination and evidence for exotic lead. Geochimica et Cosmochimica Acta 48, 2247–2264.[CrossRef][Web of Science]
Martí, J., Mitjavila, J. & Araña, V. (1994). Stratigraphy, structure and geochronology of the Las Cañadas caldera (Tenerife, Canary Islands). Geological Magazine 131, 715–727.[Web of Science]
Maury, R. C., Guille, G., Guillou, H., Blais, S., Chauvel, C. & Brousse, R. (2000). Notice explicative de la feuille Rurutu et Tubuai à 1/25 000 (Polynésie française, archipel des Australes). Orléans: BRGM, 81 pp.
McNutt, M., Fischer, K., Kruse, S. & Natland, J. (1989). The origin of the Marquesas fracture zone ridge and its implications for the nature of hot spots. Earth and Planetary Science Letters 91, 381–393.[CrossRef][Web of Science]
McNutt, M. K. & Bonneville, A. (2000). A shallow, chemical origin for the Marquesas swell. Geochemistry, Geophysics, Geosystems 1(6), doi: 10.1029/1999GC000028.
Metzner, C. & Grimmeisen, W. (1990). MONA: a user-friendly computer program for calculating the modal mineralogy of rocks from chemical analyses. European Journal of Mineralogy 2, 735–738.
Morimoto, N., Fabriès, J., Ferguson, A. K., Ginzburg, I. V., Malcom, R., Seiffert, F. A. & Zussman, J. (1988). Nomenclature of pyroxenes. Mineralogical Magazine 52, 535–550.[CrossRef][Web of Science]
Mungall, J. E. & Martin, R. F. (1995). Petrogenesis of basalt–comendite and basalt–pantellerite suites, Terceira, Azores, and some implications for the origin of ocean-island rhyolites. Contributions to Mineralogy and Petrology 119, 43–55.[CrossRef][Web of Science]
Neumann, E. R. & Wulff-Pedersen, E. (1997). The origin of highly silicic glass in mantle xenoliths from the Canary Islands. Journal of Petrology 38, 1513–1539.[CrossRef][Web of Science]
Nono, A., Déruelle, B., Demaiffe, D. & Kambou, R. (1994). Tchabal Nganha volcano in Adamawa (Cameroon): petrology of a continental alkaline lava series. Journal of Volcanology and Geothermal Research 60, 147–178.[CrossRef][Web of Science]
Odin, G. S., Adams, C. J., Armstrong, L. R., et al. (1982). Interlaboratory standards for dating purposes. In: Odin, G. S. (ed.) Numerical Dating in Stratigraphy. Chichester: John Wiley, pp. 123–149.
Plessis, Y., Brousse, R. & Mouton, J. A. (1978). Découverte de fossiles dans les cendres volcaniques de l'île de Ua Pou (groupe des îles Marquises). Cahiers du Pacifique, Volume spécial «Marquises», 237–241.
Pouchou, J. L. & Pichoir, F. (1984). A new model for quantitative X-ray-microanalysis. 1—Application to the analysis of homogeneous samples. La Recherche Aérospatiale 3, 167–192.
Price, R. C. & Chappell, B. W. (1975). Fractional crystallisation and the petrology of Dunedin volcano. Contributions to Mineralogy and Petrology 53, 157–182.[CrossRef][Web of Science]
Price, R. C., Cooper, A. F., Woodhead, J. D. & Cartwright, I. (2003). Phonolitic diatremes within the Dunedin volcano, South Island, New Zealand. Journal of Petrology 44, 2053–2080.
Samson, S. D. & Alexander, E. C., Jr (1987). Calibration of the intralaboratory 40Ar/39Ar dating standard MMhb-1. Chemical Geology 6, 27–34.
Schiano, P., Clocchiatti, R., Shimizu, N., Weis, D. & Mattielli, N. (1994). Cogenetic silica-rich and carbonate-rich melts trapped in mantle minerals in Kerguelen ultramafic xenoliths: implications for metasomatism in the oceanic upper mantle. Earth and Planetary Science Letters 23, 167–178.
Schilling, J. G. & Winchester, J. W. (1967). Rare-earth fractionation and magmatic processes. In: Runcorn, S. K. (ed.) Mantles of Earth and Terrestrial Planets. New York: Interscience, pp. 267–283.
Spulber, S. D. & Rutherford, M. J. (1983). The origin of rhyolite and plagiogranite in oceanic crust: an experimental study. Journal of Petrology 24, 1–25.
Steiger, R. H. & Jäger, E. (1977). Subcommission on geochronology: convention on the use of decay constants in geo- and cosmochronology. Earth and Planetary Science Letters 36, 359–362.[CrossRef][Web of Science]
Sun, S. S. & McDonough, W. F. (1989). Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and process. In: Saunders, A. D. & Norry, M. J. (eds) Magmatism in the Ocean Basins. Geological Society, London, Special Publications 42, 313–345.
Thirlwall, M. F. (1997). Pb isotopic and elemental evidence for OIB derivation from young HIMU mantle. Chemical Geology 139, 51–74.[CrossRef][Web of Science]
Thompson, G. M., Smith, I. E. M. & Malpas, J. G. (2001). Origin of oceanic phonolites by crystal fractionation and the problem of the Daly gap: an example from Rarotonga. Contributions to Mineralogy and Petrology 142, 336–346.[Web of Science]
Vicat, J. & Pouclet, A. (1995). Nature du volcanisme lié à une extension pré-panafricaine: les dolérites des bassins de Comba et de Sembe-Ouesso (Congo). Bulletin de la Société Géologique de France 166, 355–364.[Web of Science]
Vidal, Ph., Chauvel, C. & Brousse, R. (1984). Large mantle heterogeneity beneath French Polynesia. Nature 307, 536–538.[CrossRef]
Vidal, Ph., Dupuy, C., Barsczus, H. G. & Chauvel, C. (1987). Hétérogénéités du manteau et origine des basaltes des Marquises (Polynésie). Bulletin de la Société Géologique de France 4, 633–642.[Web of Science]
Villemant, B., Jaffrezic, H. & Treuil, M. (1981). Distribution coefficients of major and trace elements; fractional crystallisation in alkali basalt series of Chaîne des Puys (Massif Central, France). II. Géochimie. Bulletin de la Société Française de Minéralogie et de Cristallographie 103, 267–286.
Weaver, B. L. (1990). Geochemical of the highly-undersaturated ocean island basalt suites from the South Atlantic Ocean Fernando de Noronha and Trindade. Contributions to Geology and Petrology 105, 502–515.
White, W. M. (1997). Geochemistry, on-line textbook. Baltimore, MD: Johns Hopkins University Press. Available at www.geo.cornell.edu/geology/classes/Chapters/.
White, W. M., Albarède, F. & Télouk, P. (2000). High-precision analysis of Pb isotope ratios by multi-collector ICP-MS. Chemical Geology 167, 257–270.[CrossRef][Web of Science]
Williams, L. A. J. (1972). The Kenya rift volcanics: a note on volumes and chemical compositions. Tectonophysics 15, 83–96.[CrossRef][Web of Science]
Wolfe, C. J., McNutt, M. & Detrick, R. S. (1994). The Marquesas archipelagic apron: seismic stratigraphy and implications for volcano growth, mass wasting, and crustal underplating. Journal of Geophysical Research 99(B7), 13591–13608.[CrossRef]
Wolff, J. A., Grandy, J. S. & Larson, P. B. (2000). Interaction of mantle-derived magma with island crust? Trace element and oxygen isotope data from the Diego Hernandez Formation, Las Cañadas, Tenerife. Journal of Volcanology and Geothermal Research 103, 343–366.[CrossRef][Web of Science]
Woodhead, J. D. (1992). Temporal geochemical evolution in oceanic intra-plate volcanics: a case study from the Marquesas (French Polynesia) and comparison with other hotspots. Contributions to Mineralogy and Petrology 111, 458–467.[CrossRef][Web of Science]
Woodhead, J. D. & McCulloch, M. T. (1989). Ancient seafloor signals in Pitcairn island lavas and evidence for large amplitude, small-scale mantle heterogeneities. Earth and Planetary Science Letters 94, 257–273.[CrossRef][Web of Science]
Wörner, G., Beusen, J. M., Duchateau, N., Gijbels, R. & Schmincke, H. U. (1983). Trace element abundances and mineral melt distribution coefficients in phonolites from the Laacher Seevolcano (Germany). Contributions to Mineralogy and Petrology 84, 152–173.[CrossRef][Web of Science]
Wright, T. L. & Doherty, P. C. (1970). A linear programming and least squares computer method for solving mixing problems. Geological Society of America Bulletin 81, 1995–2008.
Wulff-Pedersen, E., Neumann, E. R., Vannucci, R., Bottazzi, P. & Ottolini, L. (1999). Silicic melts produced by relation between peridotite and infiltrating basaltic melts: ion probe data on glasses and minerals in veined xenoliths from La Palma, Canary Islands. Contributions to Mineralogy and Petrology 137, 59–82.[CrossRef][Web of Science]
Wyllie, P. J. (1963). Effects of the changes in slope occurring on liquidus and solidus paths in the system diopside–anorthite–albite. Mineralogical Society of America, Special Publications 1, 204–212.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. R. Vidal-Solano, A. Demant, F. A. Paz Moreno, H. Lapierre, M. A. Ortega-Rivera, and J. K.W. Lee Insights into the tectonomagmatic evolution of NW Mexico: Geochronology and geochemistry of the Miocene volcanic rocks from the Pinacate area, Sonora Geological Society of America Bulletin, May 1, 2008; 120(5-6): 691 - 708. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||