Le volcanisme de l’île de Pâques (Chili)

Abstract

Easter Island forms part of the Easter Line (Fig. 1, inset), a continuous latitudinal chain of volcanic seamounts and islands (Bonatti et al., 1977; Pilger and Handschumacher, 1981). The island’s roughly triangular shape is determined by its three main volcanoes -Rano Kau, Terevaka, Poike- which form its main mass (Fig. 1). Vélain (1879) first described the sideromelane tuff from which the statues were carved. Lacroix (1928; 1936a and b) later presented petrological data and Chubb (1933) and Bandy (1937), after their visits, provided more extensive geological descriptions of the island. Gonzalez et al. (1968) established the first comprehensive geological map and Baker et al. (1974) realized an extensive petrological and geochemical study of the island. More recently, new geochemical and geochronological data have been presented by Boven et al. (1997), De Paepe and Vergauwen (1997), Haase et al. (1997), McCormick (1997) and Haase (2002). Radiogenic K-Ar and Ar-Ar data have given ages between 3.0 and 0.13 Ma (Gonzalez et al., 1976; Clark and Dymond, 1977; Kaneoka and Katsui, 1985; Boven et al., 1997) in accordance with paleomagnetic data (Isaacson and Heinrichs, 1976; Miki et al., 1988). The Rano Kau volcano is made up of numerous basaltic lava flows and has been reduced in size by faulting and marine erosion. Its crater (1.4 km wide) is a small caldera that collapsed after a late, large explosive phase, as attested by the presence of breccia deposits around the eastern rim of the crater. This breccia comprises gabbroic blocks dispersed in a completely altered white matrix. A late aa-type lava flow of benmoreite composition extends to the east of the crater. The three small islets (“motu”), to the south of Rano Kau are composed of whitish and obsidian rhyolite. Te Mamavai is a parasite pyroclastic cone adjacent to a small dome of white rhyolite and spherulitic obsidian. Another exposure of obsidian occurs on the upper slope of Rano Kau, north of the crater. The Orito dome is probably related to Rano Kau; it essentially consists of whitish banded rhyolitic lava interlayered with (? decimetre- to metre-thick) obsidian layers. Terevaka is the largest and highest (507 m a.s.l.) volcano of the island. It consists essentially of basaltic lava that flowed mainly southwards and of more than 50 pyroclastic cones. Rano Raraku is composed of layered deposits of coherent tuff probably erupted from a shallow submarine vent. The statues have been carved on its flanks. Poike, which forms the eastern part of the island, is a basaltic volcano made up of numerous basaltic lava flows (up to 30 flows have been identified in the eastern cliff) with a small central crater. The volcano has been cut by faulting and is limited by vertical cliffs. Three small parasite trachyte domes occur on its eastern flank. Petrography and mineralogy Basalt s.s. (D.I. < 35, olivine phenocrysts) Is rare on Easter Island. Most of the basaltic lavas are hawaiite (D.I. = 35 - 50) that contains abundant plagioclase phenocrysts, but very rare olivine and clinopyroxene phenocrysts. Few mugearite and benmoreite (D.I.= 50 - 65 and 65 - 80, respectively) lava flows have been described. The trachyte (D.I. > 80) from Poike contains anorthoclase and fayalite phenocrysts in an essentially feldspathic groundmass of anorthoclase microlites. The whitish banded rhyolite (D.I. > 80) displays a trachytoid texture marked by the alignment of anorthoclase and fayalite phenocrysts, whereas the obsidian contains numerous tiny microlites of clinopyroxene and anorthoclase. The plagioclase ranges from An76-50 (phenocrysts in the hawaiite) to An22 (microlites in the benmoreite). The anorthoclase phenocrysts have a composition that ranges from Or8 to Or25 in the trachyte and reaches Or38 in the rhyolite; the microlites are more potassic (Table 1; Fig. 2a). Olivine phenocrysts in the basaltic lavas are in the range Fo78-65; the microphenocrysts are less magnesian. Olivine phenocrysts in the trachyte and rhyolite are almost pure fayalite (Fa99). Clinopyroxene phenocrysts in the basaltic lavas are diopside or augite (Wo47-42) whereas the microlites are less calcic. Green clinopyroxene microlites in the trachyte plot near the triple junction between the diopside, hedenbergite and augite fields in the quadrilateral of Morimoto et al. (1988), and ferroan hedenbergite occurs in the obsidian (Fig. 2b). Fe-Ti oxide phenocrysts are TiO2-rich magnetite in the basaltic lavas and ilmenite in the trachyte. The trachyte also contains microphenocrysts of katophorite (Fig. 2b), zircon and apatite in a matrix of anorthoclase and scarce hedenbergite microlites. Tiny crystals (<20 µm) of rhabdophane-(Nd) and churchite occur in REE-rich basalt. Variations in major- and trace-element concentrations (Table 2) are presented in binary SiO2- and Th-element diagrams respectively (Figs. 3 and 4). Thorium content increases with SiO2 content from the basalt to the rhyolite, except for the trachyte, which has a higher Th content than the rhyolite. Transition-element and Sr contents decrease strongly from the basaltic to the felsic lavas. The distribution of REE (Fig. 5) reveals two groups of basalt with parallel patterns: a group with very high REE contents and a strong negative Ce anomaly, and a group with “normal” REE contents similar to alkaline basalt in general. It also reveals three groups among the felsic lavas: obsidian, whitish rhyolite and trachyte. The obsidian pattern parallels that of the basalts. The trachyte and whitish rhyolite show slightly concave HREE patterns. REE-rich basalt has low K and Sr contents (Fig. 6). The trachyte and obsidian have higher contents for most trace elements, apart from Sr, P and Ti. Zirconium and hafnium contents are high in these lavas. Initial 87Sr/86Sr isotopic ratios vary from 0.70303 to 0.70336 for eight basalts, in agreement with the data of White and Hofmann (1982) and with the recent compilation of Haase et al. (1997). Ten felsic lavas have initial (recalculated at 1 Ma) Sr isotopic compositions (0.70300 - 0.70331, with one sample at 0.70360) largely overlapping those of the basalts. Discussion and conclusions Do the basaltic and felsic (evolved) lavas belong to the same magmatic series? Although the typical bimodality of the alkaline series is not as marked at Easter Island as in other series world-wide (cf. Ngounouno et al., 2000, and references herein) due to the presence of lavas of intermediate composition (mugearite and benmoreite), the similarity of the Sr-isotopic initial ratios strongly suggests a co-magmatic origin for the two groups of lava. With this hypothesis, the felsic lavas may simply have been derived from the basalt through fractionation of olivine, clinopyroxene, Fe-Ti oxides, plagioclase, anorthoclase, and apatite, such as modelled for major elements by Haase et al. (1997). Other hypotheses concerning the genesis of felsic lavas in oceanic islands, such as partial melting of oceanic crust or of granophyres, fractional crystallization combined with assimilation of oceanic crust, have already been discarded for the Galapagos rhyolite by Geist et al. (1995). How to explain the very high REE contents in some basalt? Some of the basalts have classical REE contents similar to those of other Polynesian alkali basalts (cf. Le Dez et al., 1998), whereas others have exceptionally high REE contents -such high contents have already been measured in basalts from Hawaii (Fodor et al., 1987) and French Polynesia (Joron et al., 1991). Three hypotheses have been proposed to explain the high REE contents: (1) enrichment from a mantle zone that contains halogen-CO3-S-K-bearing metasomatic phases (Fodor et al., 1987); (2) enrichment by late crystallization of phosphates, hydroxides and carbonates from H2O + CO2 + Cl rich-fluids of magmatic origin (Joron et al., 1991; Schiano et al., 1992); (3) the presence of rhabdophane-(La), a supergene phosphate that occurs in veinlets in the basalts (Fodor et al., 1992; Cotten et al., 1995) and that has high REE (except Ce) and Y contents. Tiny (20 µm) REE-rich phosphate crystals (rhabdophane-[Nd] and churchite) have been identified in the matrix of the REE-rich basalt of Easter Island. As the REE-rich basalt has the same Si isotopic composition as the other basalts, it is unlikely that it derives from heterogeneous mantle sources. For the same reason, contamination of the REE-rich basalt by exotic materials (seawater, guano) is also precluded. Thus, we suggest, as proposed by Cotten et al. (1995) for the Polynesian basalts, that the REE + Y enrichments and the strong Ce depletion probably result from a supergene process. These elements would have been extracted from the overlying basalt by meteoric waters during weathering and soil development, with REE + Y being redeposited as phosphates in the underlying basalts, and Ce being oxidized as Ce4+ and remaining in solution.

Key words: Volcanic rocks, Major-element analyses, Trace-element analyses, Rare earths, Bimodal magmatism, Easter Island.