Volcano-tectonic and geochemical evolution of an oceanic intra-plate volcano: Tahiti-Nui (French Polynesia)

Share Embed


Descripción

Available online at www.sciencedirect.com R

Earth and Planetary Science Letters 217 (2004) 349^365 www.elsevier.com/locate/epsl

Volcano-tectonic and geochemical evolution of an oceanic intra-plate volcano: Tahiti-Nui (French Polynesia)§ Anthony Hildenbrand  , Pierre-Yves Gillot, Isabelle Le Roy Laboratoire de Ge¤ochronologie Multi-Techniques, UPS-IPG Paris, Bat. 504, Sciences de la Terre, Universite¤ Paris Sud, 91405 Orsay, France Received 2 April 2003; received in revised form 10 October 2003; accepted 14 October 2003

Abstract Geological mapping, accurate K/Ar dating and geochemical analyses of lavas allow a detailed reconstruction of the geological history of Tahiti-Nui Island (French Polynesia). The exposed volcanic activity is first characterized by the construction of a main shield from 1.4 Ma to 870 ka, with a maximum aerial eruptive rate around 2 km3 /kyr. Lavas from this early building stage are alkaline, slightly silica-undersaturated, with 87 Sr/86 Sr and 143 Nd/144 Nd compositions rather constant and close to the enriched mantle II type. Vent locations were first concentrated along a main E^W rift zone, which was responsible for the lateral collapse of the northern and southern flanks of the main shield, around 0.87 Ma ago. The subsequent activity was first restricted to the northern depression, corresponding to an eruptive rate of about 5 km3 /kyr in the period 850^760 ka. Significant variations in La/Sm, 87 Sr/86 Sr and 143 Nd/144 Nd occur in lavas erupted immediately after the main northern landslide, indicating a sudden increase in the extent of partial melting likely caused by the decompression subsequent to collapse. However, the later activity declined, and the lavas exhibit a gradual change toward strongly silica-undersaturated basanites, likely indicating a decreasing extent of partial melting of the upper mantle. The evolution of radiogenic isotope ratios over time indicates a change in the source toward more depleted compositions, until around 500 ka. Post-erosion volcanic activity, following an apparent hiatus of 240 kyr, exhibits similar major and trace element and isotope compositions. The volcano-structural and geochemical evolution of Tahiti-Nui and the overall alkaline character of the lavas from other eruptive complexes of the Society alignment suggest a relative weakness (low temperature and low eruptive rate) of the Society plume compared to the Hawaiian hot-spot. @ 2003 Elsevier B.V. All rights reserved. Keywords: Tahiti-Nui; volcano-structural evolution; K^Ar dating; geochemical evolution

* Corresponding author. Present address: Laboratoire de Ge¤odynamique des Rifts et des Marges Passives, UFR Sciences et Techniques, Universite¤ du Maine, Avenue Olivier Messiaen 72085, Le Mans, France. Tel.: +33-2-43-83-32-35; Fax: +33-2-43-83-37-95. E-mail address: [email protected] (A. Hildenbrand). §

Supplementary data associated with this article can be found at doi:10.1016/S0012-821X(03)00599-5

0012-821X / 03 / $ ^ see front matter @ 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0012-821X(03)00599-5

EPSL 6897 24-12-03

350

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

1. Introduction The study of oceanic islands is of particular interest to understand the evolution of hot-spot magma sources, and to model the morpho-structural evolution of eruptive complexes in relation to volcano growth, huge landslides, further erosion, etc. In the southern Paci¢c Ocean, the tropical volcanic island of Tahiti is strongly eroded. Access to the deeper part of the system is thus possible, providing a unique insight into the eruptive history of a high island. It also allows analysis of the erosional evolution, which has important societal implications, e.g. in terms of risk assessment. On Tahiti, however, vertical cli¡s and dense vegetation strongly limit ¢eld investigations and observations, which explains the relatively poor geological knowledge and the lack of an accurate geological map of the whole island since the map of Deneufbourg [1]. The few studies have been carried on relatively restricted areas, generally con¢ned to the external part of the volcanic structure, or on the granulated rocks of the magma chamber (e.g. [2,3]) which is dissected by erosion in the central Maroto basin. A main problem lies in the poor quality of the rare outcrops in this tropical highly weathered complex. Under such hot and wet conditions, the alteration reaches ferralitic to lateritic facies [4,5], covering the whole external part of the shields. The only way to access well-preserved geological information then consists in a progression through the main valleys, which provide natural sections within the lava piles. However, the cli¡s are di⁄cult to reach because of scree and mud£ow deposits on their slopes. The present study results from geological observations and sampling during a 6 month survey. All the sectors of the main island were investigated, through more than 25 di¡erent valleys. Where su⁄ciently fresh, lava £ows were also sampled along the coastal road, i.e. at the periphery of the volcanic structure. The analytical part consists of 76 K/Ar determinations on fresh separated groundmass. These data provide insight into the eruptive history of Tahiti-Nui. Wholerock geochemical analyses of dated samples are presented. The volcano-tectonic and geochemical

evolutions of Tahiti-Nui are then reconstructed and compared to the case of other hot-spot oceanic islands.

2. Geologic setting Tahiti belongs to the Society Archipelago (Fig. 1), which is one of the ¢ve linear volcanic chains of the south-central Paci¢c Ocean. Located near the southeastern end of this alignment, it is composed of the two coalescent eruptive systems of Tahiti-Nui and Tahiti-Iti linked by an isthmus in Taravao which may correspond to the products of an independent edi¢ce [6]. Tahiti and its neighbor Moorea constitute the higher and younger inactive islands, in agreement with the northwestward displacement of the Paci¢c plate over a ¢xed hot-

Fig. 1. (Top) Location of Tahiti at the southeastern end of the Society volcanic chain. The circle indicates the current hot-spot zone in which crosses localize recent seamounts. The arrow shows the displacement of the Paci¢c plate. Nui: Tahiti-Nui. (Bottom) 3-D view of Tahiti-Nui to the north, obtained by draping the Spot image on a digital elevation model. Thin white lines delimit the preserved £anks of the main shield. Thick dotted lines indicate the inferred rims of the northern and southern landslide depressions. The solid white lines indicate the axis of the main rift zone. The white transparent areas highlight plateaus corresponding to depression in-¢ll. T: Tamanu; Or: Orofero; V: Viriviriterai; Te: Tefaiiti; F: Fau¢ru; Fa: Fataua.

EPSL 6897 24-12-03

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

spot during the last million years [7^10]. The current hot-spot zone is located around 50 km southeast, as indicated by volcano-seismic activity on the submarine £anks of Mehetia Island as well as the growth of several seamounts [11^13]. Although deeply eroded and subsided, Tahiti Island is among the younger and better preserved volcanic islands of the archipelago, with an aerial eruptive history comprised within the last two millions years (e.g. [9,14]). The ¢rst attempted geological mapping was at a scale of 1/40 000 [1]. Since then, several sectors have been mapped and studied in more detail [15^17]. They were complemented by geochemical, geochronological and paleomagnetic studies [14,18^20]. According to Brousse [21], three main volcanic stages are distinguished: (1) a shield-building phase, (2) a central circular calderic collapse episode, and (3) a late valley-¢lling phase. The existence and size of the so-called cal-

351

dera are deduced from the current circular shape of the central Maroto depression (Fig. 1). However, this model does not take into account important morphological and geological features. Indeed, Tahiti-Nui exhibits a strong asymmetry which is marked by the present location of the higher mountains on the northern half of the island (Fig. 1). In its central part, the higher crests are oriented with a preferential E^W direction and separated by the deep valleys of Punaruu, Papeihia, and the upper part of the Papenoo drainage system (Figs. 1 and 2). In those valleys, the presence of concentrations of intrusions shows the important controlling role of a main E^W axial rift zone during Tahiti-Nui’s eruptive history, which was responsible for a huge lateral collapse of the northern main shield slope [22]. A northern depression was created, in which a second shield volcano grew. To the south, a depression cut by erosion was later ¢lled up by the

Fig. 2. Location of the lava samples. The names of the main valleys are indicated. The cross indicates the location of Orohena massif. The circle and the triangle symbols respectively indicate the location of the main shield and second shield eruptive centers, as deduced from geomorphologic approach presented in Sections 6.1 and 6.3.

EPSL 6897 24-12-03

352

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

products of the second volcano. The deposits associated with a large southern landslide have been recently identi¢ed o¡-shore [23].

3. Field investigations 3.1. Geological description of the main volcano-structural units Geological mapping led to the de¢nition of the

major volcano-structural units. The main shield is preserved in the northwestern, northeastern, southwestern and southeastern sectors of TahitiNui (Figs. 1^3). Its external slopes de¢ne morphological surfaces with a moderate 8‡ dip. The lava pile comprises basaltic aa £ows. The uppermost lava accumulations are quite thin ( 6 50 cm thick) and extend seaward over several kilometers. Because the lateral rims of the northern depression are partly hidden by the subsequent lavas,

Fig. 3. Synthetic geological map showing the main volcano-structural units. The dashed captions symbolize the valley-¢lling character of the £ows at di¡erent epochs, as well as the in-¢ll of the southern depression. The K/Ar determinations are indicated in black character (ages in ka). Numbered lines refer to pro¢les exposed in Fig. 4.

EPSL 6897 24-12-03

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

the precise contact between the two shields is not easily identi¢able in the ¢eld. However, it can be inferred from morphological and geological evidence. The northern and southern in-¢lls present higher slopes (12^15‡) than the external £anks of the main shield (Figs. 1 and 2). Moreover, the lateral contact between the two nested volcanic systems is highlighted by the geometry of the northern coral barrier, which displays two in£ection points near the Tipaerui and Onoheha river outlets (Figs. 1 and 3). The northwestern rim of the depression is characterized by the occurrence of parasitic Strombolian cones (Pic Rouge, Pic Vert) located far from the main central magma supply. Related feeding dykes outcrop in the adjacent valleys of Tipaerui and Fataua (Figs. 2 and 3). These intrusions de¢ne a N150 ( T 10‡) rift zone, which gradually bifurcates up to N110 in the deeper parts of the two valleys. The northeastern rim is harder to de¢ne because later £ows hide its location. However, the presence of parasitic cones and dyke concentrations indicates that it coincides with the Onoheha valley. It is bordered by a complementary rift zone, trending roughly N050 in the Tahaute valley (Figs. 2 and 3). Particular attention has been paid to the postcollapse northern shield. The base is represented by a thick, vesicular to completely auto-brecciated unit, likely indicating that rapid gas decompressions followed the landslide. It outcrops in the center of Tahiti-Nui, constituting the base of the Orohena massif (Fig. 2). There, it is a¡ected by secondary hydraulic fracturing observed along the latest intrusions of the main E^W rift zone. To the north, the brecciated unit outcrops in the deepest parts of most of the northern valleys, forming plateaus. One can observe that the vesicularity gradually decreases from the bottom to the top of the sequence. This thick unit is overlain by porphyric pahoehoe accumulations largely represented at the periphery of Tahiti-Nui. In the northeastern sector, they are known as the ‘lava-tube’ unit [24]. There, they often present a steep channeled geometry and associated thickness up to 10 m. Apart from the zones of the inferred U-shaped depression, they cover the primitive shield whereas some of them are found in a valley-¢lling position (e.g. in the Tahaute and Orofero valleys, Figs. 2 and 3).

353

The upper stage of the second shield is represented by a thick columnar, pyroclastic formation, observed in the upper part of the Orohena massif [25]. The related products, sampled as blocks in the river Vaitamanu under Orohena, constitute an ignimbrite facies. Finally, recent activity is represented by thick valley-¢lling, peridotitic nodule-rich £ows. They are hung to the main walls of the northern valleys of Tuauru and Papenoo, which incise the upper part of the second shield, including the ignimbritic formation. These £ows were thus erupted during a post-erosion phase. They stand in the lower part of the basins forming the plateaus of Fau¢ru and Tefaiiti (Figs. 1^3). 3.2. Sampling strategy All the sectors of Tahiti-Nui were investigated (Fig. 2). Taking advantage of the di¡erent valleys, lava samples were thus collected at various levels of the volcanic succession. The main shield was mostly sampled along the main E^W rift zone, in the valleys of Punaruu, Papenoo and Papeihia. Shield lavas were also reached at the periphery of the volcanic structure. The base of the second shield was approached in the deeper part of the northern in-¢ll, within the valleys of Fataua, Tuauru, Papenoo and Onohea (Fig. 2). Moreover, a £ow intercalated in the breccia in the upper part of the Vaitamanu valley was sampled on a plateau located 1000 m above the Orohena massif (sample 56Q, Fig. 2). In the same valley, dykes intruding the brecciated unit were also collected. They are related to the intermediary to upper part of the second shield activity. Intermediary £ows were reached along the ‘Sentier des Mille Sources’, which runs along the eastern rim of the Tuauru drainage basin. Others were also collected at altitudes of 1600 m and 1700 m on the pathway to Mount Aorai (Fig. 2). The later £ows of the second shield were sampled mainly at the northern and northeastern periphery of the volcanic structure (Fig. 3), whereas some in a valley-¢lling position were also collected. The southern in-¢ll was extensively sampled through its whole width, at various levels

EPSL 6897 24-12-03

354

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

of the succession in the valleys of Papeiti, Taharuu, Moaroa, Tahiria, and Vaite (Fig. 2). Finally, two £ows from the late post-erosion activity were sampled along the Papenoo and Tuauru rivers. The K/Ar dating was then mainly achieved on £ows from the basal and summit parts of the di¡erent piles, in order to estimate the mean duration of activity within a given volcanic unit, and thus estimate mean rates of volcano growth.

4. K/Ar dating of the samples 4.1. Sample preparation and analytical procedure Thin sections were made from the freshest part of the lava £ows. The microcrystalline volcanic groundmass was selected for the K^Ar analyses. Taking into account the size of the phenocrysts (mostly olivine, pyroxene and K-poor plagioclase feldspars), crushing and sieving at 125^250 Wm was performed. Heavy liquids were used to remove all phenocrysts, which may bring possible inherited argon. They are in any case K-poor and may create heterogeneity in the sample. Tightened density spans, generally between 2.95 and 3.05, were realized in order to select the freshest fraction of the groundmass and thus avoid alteration and secondary zeolitization. Potassium and argon were measured on two aliquots of the selected groundmass grains, by two separate analyses: the former by £ame emission spectroscopy, the latter by mass spectrometry according to the Cassignol^Gillot unspiked technique, which has been shown to be suitable for dating young basalts [26,27]. Potassium was determined with 1% relative uncertainty. Argon was analyzed at least twice in order to obtain a reproducible age within range of error determined from replicated measurements of standards. The technique and principles have been described elsewhere [28]. The results are presented in Table 11 . Errors are quoted at 1c.

1

See online version of this article.

4.2. Results The ages measured in this study (Table 11 , Fig. 3) generally con¢rm the previous radiometric data from Tahiti-Nui (e.g. [7,9,14,29]). The K/Ar determinations obtained on £ows from the main shield volcano are in agreement with the available stratigraphic control (Fig. 4). The older ages were obtained on lavas from the major valleys (Punarru, Papeihia, Tahaute, Orofero) at a constant distance from the coast of about 5 km (Fig. 4, sections 2^4), due to rather constant dipping of the £ows (about 8‡). Note that other ages obtained in these valleys decrease toward the periphery, but also toward the center of the volcanic structure, which is a consequence of the logarithmic pro¢le of the river (Fig. 4, sections 2^4). A maximum mean age of 1.37 T 0.02 Ma was determined (sample 62CH). This is 0.3 Ma younger than the older age of Duncan et al. [14], obtained on £ow T85-41, which was sampled in the center of the island at a higher level of the stratigraphic pile. De facto, two £ows from the southern wall of the Maroto basin were dated in the present study: one at the outlet of the basin (sample 62BO),

Fig. 4. Stratigraphic sections showing the relation between the di¡erent volcano-structural units. Ages and uncertainties are indicated in ka.

EPSL 6897 24-12-03

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

dated at 1.16 T 0.02 Ma, and the other at the top of the sequence at Urufau pass, dated at 882 T 12 ka (Figs. 3 and 4). The latter age is indicative of the late activity of the main shield, which is better constrained by the age determinations on £ows 81A and 62AO (Table 11 , Fig. 2), sampled at the periphery of the structure, and dated at 866 T 17 ka and 867 T 18 ka respectively. The exposed main shield activity was thus dated here between 1.37 T 0.02 Ma and 0.87 T 0.02 Ma, whereas intermediary £ows in the volcanic pile (samples 81W1 and 65J) yielded coherent intermediate values (Figs. 3 and 4). Flows from the lower part of the second shield (56AL, 56Q, 81N) are dated between 872 T 10 ka and 851 T 11 ka (sample 56Q, Fig. 4, section 1), yielding a mean age of 0.85 T 0.01 Ma. Within uncertainties, this is very close to the mean age from the later activity of the main shield, although the two were obtained on £ows sampled at di¡erent geologic levels of the volcanic structure (Fig. 4, section 1). The timing of the northern collapse is thus constrained very accurately between 0.87 T 0.02 Ma and 0.85 T 0.01 Ma (Fig. 4). The K/Ar determinations on £ows from the base of the second shield (samples 56N, 56K1, 56K2, 56M, 62AV2, 81AAH, 62AD, 56X, 81U1, 56I, 56W, 56AK) range from 850 ka to 700 ka, but are mainly between 850 ka and 760 ka, indicating that the northern U-shaped depression was rapidly ¢lled following the collapse. Note that the ages generally decrease to the periphery of the structure (Figs. 3 and 4, section 1), which is in agreement with the main dip of the £ows. However, £ows located at the periphery of Tahiti-Nui are dated around 760 ka (samples 62CA2, 65BE) and even sometimes as old as 790 ka (samples 56AA and 62F). This indicates that some of the post-collapse eruptions occurred at this time in a zone extending far out of the northern depression area (Fig. 3). The samples from the late e¡usive activity of the second shield are mainly dated between 750 ka and 450 ka. However, £ows caping the preserved slopes of the main shield (samples 56AD, 81M1, 81W), as well as valley-¢lling £ows (81G, 81T, 56AB, 56AK, Fig. 4), are dated between 650 ka and 550 ka. Thick £ows from the base of

355

Mount Tetufera (sample 56H3) yield a similar age. Thus, lavas from the second shield likely began to reach the southern part of the island around 650 ka. Lava £ows dated younger than 550 ka are mainly distributed within the zones of the northern and southern in-¢lls (Figs. 3 and 4). In the southern sector, the ages cluster between about 580 ka and 500 ka (Table 11 ). Comparable determinations were obtained on several £ows from the external part of the northern in-¢ll (81U, 62AC1, 62AD, 81AAF; Table 11 ). The £ows from the late post-erosion valley-¢lling stage were dated at 187 T 3 ka (sample 56L) and 227 T 3 ka (sample 62Y) indicating an apparent gap of around 250 ka between the end of the second shield building and the post-erosion recovery.

5. Geochemistry of the lavas Tahiti-Nui lavas have been exposed to heavy rainfall for a long time. Under such conditions, alteration may be a crucial issue, which has to be taken into account before considering the data for geochemical analysis. Loss on ignition (LOI) for most samples from the present study is lower than 1%, indicating that alteration is overall negligible. However, a few samples have LOI s 1%, i.e. 62Q (LOI = 1.4), 62BZ2 (LOI = 1.42), 62AL (LOI = 1.84), 81N (LOI = 2.14), 56AD (LOI = 2.6) and 81I (LOI = 3.47). Among them, samples 62Q, 62BZ2, 56AD and 81N are characterized by a di¡erentiation index (DI) greater than or equal to 35 (Table 21 ). Their high LOI thus likely results from di¡erentiation processes and subsequent increase in the content of volatiles dissolved in the magma. Samples 62AL and 81N, however, are not di¡erentiated lavas. Alteration could thus be suspected for these samples. Weathering a¡ects the concentration of more mobile elements such as the alkalis, and Ba/Rb in weathered samples may be signi¢cantly di¡erent from the value of 11 according to White and Duncan [10]. Nevertheless, Ba/Rb for samples 62AL and 81N is about 11 and 13 respectively, indicating that alteration may have been minor for these two samples,

EPSL 6897 24-12-03

356

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

which were thus retained for the geochemical analysis. 5.1. Major elements All of the Tahiti-Nui samples belong to the alkaline series (Fig. 5). The main shield lavas plot along or just above the alkaline and sub-alkaline boundary [30]. The compositions are rather uniform, ranging from picro-basalts to slightly evolved basalts. Most of the post-collapse samples plot in the basanitic ¢eld, up to tephritic (sample 81I) and even nephelinitic (sample 62BZ2). The most di¡erentiated rock is from the late activity, and re£ects the e¡ect of di¡erentiation processes within the shallow reservoir, as suggested by the presence of plutonic equivalents in the center of the Maroto basin [2,3]. The two main volcano-structural units distinguished in this study thus correspond to two successive magma series with distinct a⁄nities. They are similar to the two trends previously described for lavas [14,20,31] and plutonic rocks from

Fig. 6. Variations in SiO2 , Al2 O3 , MgO, Fe2 O3 *, P2 O5 major oxides (wt%) and nepheline contents in Tahiti-Nui lavas. The symbols are the same as in Fig. 5. Error bars are included in the size of the symbols. Di¡erentiated samples (81I, 62 BZ2) and samples with high MgO (62BW, 56A3) are labelled on plots.

Fig. 5. Total alkali vs. silica diagram for lavas from the main volcano-structural units of Tahiti-Nui. Black circles indicate the main shield lavas, triangles the second shield lavas, squares lavas from the late, post-erosion activity. The bold solid line separates the alkaline and sub-alkaline ¢elds [30]. Error bars are included in the size of the symbols.

the center of the island [2,3]. Additionally, basanites from the late post-erosion activity are the most silica-undersaturated samples, and could thus de¢ne a third volcanic series. The silica content in the Tahitian lavas clearly decreases over time (Fig. 6) with two exceptions corresponding to more di¡erentiated terms (samples 81I and 62BZ2, with DIs of 58 and 59 respectively). The earlier lavas from the main shield are close to SiO2 saturation (SiO2 ranging between 48 and 50 wt%), whereas the later activity, basanitic, is more undersaturated, with silica contents as low as 41%. Al2 O3 variations are not clear but one can observe a general decrease over time, whereas MgO and Fe2 O3 present an overall increase. Outliers from the general trend (Fig. 6) have DI v 35 (Table 21 ). Most of them are lavas erupted between 700 ka and 500 ka (samples 56K1, 56W, 81I, 56X, 81AAF, 56AD, 62BZ2), with a few from the main shield phase (samples 62CK, 62Q, 62AM). Note that samples 56A3 and 62BW exhibit surprisingly low values for Al2 O3 , associated with a high MgO content (Fig. 6). This may be due to cumulative processes

EPSL 6897 24-12-03

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

357

linked to mineral fractionation, since these two samples are olivine-rich. The other major element oxides do not show a signi¢cant evolution. Alkalis and P2 O5 are rather uniform. However, quite important variations occur at some stages of Tahiti-Nui’s geochemical history (Fig. 6), corresponding again to samples with DI v 35, which are also characterized by a high content of normative nepheline (Table 21 ). These variations may thus be linked to episodic di¡erentiation processes possibly associated with crystallization of feldspars and apatite. However, the overall increase in normative nepheline over time, especially for samples with DI 6 35, likely results from a decreasing extent of partial melting of the mantle. 5.2. Trace elements Lavas from the main shield have concentrations of Ba mostly comprised between 250 ppm and 450 ppm, except sample 62BW for which the curiously low value (Table 21 ) may be due to olivine cumulation. Lavas from the post-collapse second shield have more dispersed values (Fig. 7), ranging between 250 ppm and 1000 ppm, with more than 50% of the analyzed concentrations exceeding 500 ppm (Table 21 ). Among these lavas, samples with DI s 35 have the highest concentrations of Ba, which is a highly lithophile element. This pattern likely emphasizes the increase of di¡erentiation processes for lavas erupted between 700 ka and 500 ka. The two analyzed samples from the post-erosion late activity are about 500 ppm which is not signi¢cantly di¡erent from lavas of the post-collapse second shield. To better constrain the chemical evolution of Tahiti-Nui parental magmas, ratios involving rare earth elements (REE, La/Sm) and other elements (Th/Ta) are plotted versus time (Fig. 7). Samples with DI v 35 are not taken into account in the diagram, since di¡erentiation processes may disturb the general pattern. Although the values are not strongly contrasted, La/Sm ratios gradually increase as a function of time (Fig. 7), especially during the main shield phase. Tahiti-Nui lavas thus evidence ¢rst a relative enrichment in light REE with respect to

Fig. 7. Evolution of Ba abundances, La/Sm, and Ta/Th in Tahiti-Nui lavas over time. The symbols are the same as in Fig. 5. Error bars are included in the size of the symbols.

heavier REE, which is most likely a consequence of a decreasing degree of partial melting of the mantle (e.g. [32^34]). Note, however, that the lavas erupted immediately after the collapse (samples 56Q, 81N) are characterized by low La/Sm ratios, which could indicate a rapid jump in the degree of partial melting following the collapse. Nevertheless, La/Sm values for younger lavas from the post-collapse second shield are rather dispersed and do not show an overall trend. This would indicate a relative instability in the conditions of mantle partial melting during the late evolution of Tahiti-Nui. Th/Ta ratios do not clearly change over time (Fig. 7) except at some moments, e.g. during the short period following the landslide. This may be associated with episodic variations in the composition of the magma source (e.g. [35]).

EPSL 6897 24-12-03

358

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

5.3. Radiogenic isotopes Strontium and neodymium isotopic ratios were measured by thermo-ionization mass spectrometry on 12 samples from the three volcano-structural stages, i.e. main shield, post-collapse second shield and post-erosion activity (Table 21 ). Samples 62BF1 and 62BG3 (Table 21 ) were not retained for the geochemical major and trace study, since they have LOIs of 3% and 6% respectively despite their basic character. However, because alteration processes do not di¡erentially a¡ect Sr or Nd isotopes, those two samples were used for isotopic measurements. The points concerned have been distinguished (Figs. 8 and 9). Tahiti-Nui lavas show a wide range of compositions (Fig. 8), with extreme values comparable with previous data [10]. The data de¢ne a linear trend. Note that the samples show a fairly good repartition according to the volcano-structural units, except for sample 56AA, which is characterized by higher 87 Sr/86 Sr and 143 Nd/144 Nd ratios. The lavas from the main shield have rather constant 87 Sr/86 Sr and lower 143 Nd/144 Nd ratios (Table 21 , Fig. 9), indicating an enriched component that would be closer to the enriched mantle (EM) II type [36]. Immediately after the landslide, 87 Sr/ 86 Sr and 143 Nd/144 Nd are higher and lower respectively than values for the main shield, likely indi-

Fig. 9. Variation over time of the 87 Sr/86 Sr and 143 Nd/144 Nd isotopic ratios in Tahiti-Nui lavas. The symbols are the same as in Fig. 8. Error bars are included in the size of the symbols.

cating a sudden evolution in the source possibly associated with a change in the geometry of the zone of magma generation. Then, 87 Sr/86 Sr rapidly decreased and 143 Nd/144 Nd correlatively increased until around 500 ka. This might re£ect a decrease in the EMII component over time possibly indicating a decreasing in£uence of the Society plume under Tahiti-Nui. However, the late valley¢lling £ows, although erupted after an apparent gap of around 250 kyr, show isotopic values comparable to those from 500 ka, possibly indicating a stabilization of the source.

6. Discussion 6.1. The main shield Fig. 8. 143 Nd/144 Nd vs. 87 Sr/86 Sr for Tahiti-Nui lavas. The symbols are the same as in Fig. 5. Symbols with no shading correspond to samples 62BF1 and 62BG3 (Table 21 ), and sample 56AA is labelled on plot. Error bars are included in the size of the symbols. Crosses indicate data from White and Duncan [10].

Main shield £ows have been dated between 1.37 T 0.02 Ma and 0.87 T 0.02 Ma (Fig. 10a), but as the core of the volcanic structure is not entirely exposed, the duration of subaerial activity

EPSL 6897 24-12-03

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

Fig. 10. Volcano-structural and geochemical model of the evolution of Tahiti-Nui. The colors refer to the volcanostructural units described in Fig. 3. The late post-erosion phase is not drawn. Ages are indicated in million years (Ma).

is signi¢cantly longer. Although the shape of Tahiti-Nui is elongated in the direction of the main E^W rift zone (Fig. 1), the geometry of the northwestern, northeastern, southwestern and southeastern sectors indicates a mean conic morphology for the main shield. Extrapolating inward the slope of these preserved morphological elements (8‡) de¢nes a main eruptive center located at the current center of Tahiti-Nui (Fig. 2), with a minimum altitude of 3250 m. For a conical shape, this corresponds to a mean aerial volume of 1000 km3 . A maximum rate of construction of 2 km3 /kyr is then calculated from 1.37 Ma to 0.87 Ma. This is very low compared to eruptive rates of 0.1 km3 /yr determined for present shield-building activity in Hawaii [37], but in the same order of magnitude as the rates determined for La Palma Island

359

(Canaries) over the past 125 kyr [38]. Lavas from the Tahiti-Nui main shield phase, although hypersthene-normative (Table 21 ), overall plot in the weakly alkaline ¢eld, in agreement with previous geochemical data (e.g. [3]). Nevertheless, some authors [14] have suggested the existence of an earlier transitional to tholeiitic phase, on the basis of their sample T85-41, which has been dated at 1.67 Ma. However, the sum of major elements for this sample is equal to 96.04%, corresponding to a LOI of about 4%, which, for a hawaiite, is too high and only explained by weathering. This £ow is stratigraphically located within the sequence of the South Maroto wall that we date from the base to the top between 1.16 T 0.02 Ma and 0.88 T 0.02 Ma. K loss by meteoric water leaching is a possible cause to explain the singular ‘old’ age found for this sample, as well as its apparent low alkali content. The present study shows a light evolution in the composition of lavas erupted during the main shield-building phase. This consists in a gradual decrease in SiO2 over time (Fig. 5), which is associated with an increase in La/Sm, suggesting a decreasing degree of partial melting of the mantle. However, £uctuations in SiO2 , MgO, P2 O5 and Ba for some of the samples are likely associated with olivine cumulation and episodic di¡erentiation processes. 87 Sr/86 Sr and 143 Nd/144 Nd isotopic ratios seem rather constant, indicating a homogeneous character of the lava source during the main shield phase. Although Pb isotopic ratios would be necessary to better constrain the nature of the source, our data indicate an enriched mantle component, which would be close to the EMII type [36], in agreement with previous geochemical studies (e.g. [10,14,20]). 6.2. The northern and southern collapses The timing of the northern collapse is accurately constrained between 0.87 T 0.02 Ma and 0.85 T 0.01 Ma, given the ages obtained for the late activity of the main shield on the one hand and the base of the post-collapse second shield (Figs. 3 and 4) on the other hand. Note that the two ages obtained here are indistinguishable at the 1c level and hence support evidence for the

EPSL 6897 24-12-03

360

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

absence of any interruption of volcanic activity between the two stages. The timing of the southern collapse is not so well constrained. Lava £ows as old as 870 ka are cut by the collapse in the upper Taharuu valley, whereas the late in-¢ll provides a minimum age of 550 ka for the southern landslide. It is thus possible that the southern collapse occurred synchronously with the main northern landslide (Fig. 10b). Both could have been triggered as a consequence of magma concentration along the main central E^W rift zone. However, several collapse episodes may have a¡ected the southern £ank of Tahiti-Nui, as suggested by an important discrepancy between the volume of submarine debris avalanches and the slide volume estimated from on-land reconstruction [23]. 6.3. Rapid in-¢ll of the northern depression Geological and geochronological data (Figs. 3 and 4) support evidence for a rapid construction of the second shield following the northern landslide. The presence of the thick brecciated unit in the deeper part of the main northern valleys of Tahiti-Nui indicates that an intense volcanic activity was ¢rst restricted into the northern landslide depression. It rapidly ¢lled the depression, from 850 ka to 760 ka (Figs. 4 and 10c). Some of the £ows might have over£owed the depression, as suggested by the presence of £ows dated as old as 760 ka in the southwestern and southeastern external parts of the island (Fig. 4). However, such £ows could also have been erupted from parasitic cones distributed along the rims of the northern depression as supported by the presence of subsidiary rift zones, oriented at N150 and N050 in the Fataua and Tahaute valleys. The subaerial volume of lava erupted subsequent to the northern landslide is broadly estimated from a geomorphological reconstruction. The total geometry of the second shield is ¢rst inferred from the prolongation of the northern and southern in-¢lls, as well as the extrapolation of the slope of the northeast lava-tube unit. The eruptive center thus deduced is located slightly to the north of the pre-collapse main shield, for a

mean elevation estimated about 2900 m. In the hypothesis of a conical shape, a grid is computed to reconstruct the morphology of the second shield. A total volume of 450 km3 is estimated for the part of the second shield located above sea level, inside the inferred area of the northern and southern landslide depressions. The volume of lava outside the depressions is in any case negligible since it corresponds to: (1) £ows caping the preserved slopes of the main shield and (2) restricted valley-¢lling £ows. The thick base of the second shield (Figs. 3 and 4) represents about two thirds of its total aerial volume, since it forms most of the northern in-¢ll, up to altitudes of 1500 m (Fig. 4). The subaerial volume of lava erupted soon after the collapse is thus estimated to be about 300 km3 . This rate is probably underestimated since the submarine part of the northern in-¢ll was not considered here in the volume computation. Bathymetric imaging of TahitiNui’s northern slope, indeed, provides evidence for a continuation of post-collapse lavas up to 1500 m below sea level [39]. Assimilating the submarine part to a triangular edge (height about 1.5 km, length about 15 km, Fig. 4) on the whole width of the northern depression (20 km) yields an additional volume of at least 200 km3 . On the period 870 ka^760 ka, the total eruptive rate for the base of the second shield then likely reaches up to 5 km3 /kyr, which is signi¢cantly greater than in the case of the main shield. This is probably a consequence of the huge northern landslide, which favored magma extrusion preferentially to the north. Signi¢cant variations in La/ Sm, Ta/Th, 87 Sr/86 Sr and 143 Nd/144 Nd for lavas from the base of the second shield also suggest a brutal change in the conditions of magma generation subsequent to collapse. Decompression associated with collapse deloading might thus have provoked a sudden increase in the extent of partial melting, as observed on other ocean volcanic islands (e.g. [40]). 6.4. Late activity of the second shield This was dated mainly between 750 ka and 450 ka. Widespread distribution of the £ows con¢rms that the headwalls of the northern depression

EPSL 6897 24-12-03

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

were rapidly reached and even over£owed (Fig. 10d). Lavas then capped the preserved slopes of the main shield, or ¢lled existing valleys, gradually migrating toward the southern part of the island, which was reached around 650 ka (Fig. 10d). Lava trapping by the southern depression then extensively occurred, until around 450 ka (Fig. 10e). The volume of lava erupted between 750 ka and 450 ka is estimated to be about one third of the total subaerial volume of the second shield, which corresponds to approximately 150 km3 and gives an eruptive rate of 0.5 km3 /kyr. Such a rate indicates a diminishing magma supply during the late history of Tahiti-Nui, which might reveal the di⁄culty for magma to reach the surface due to the load exerted by the second shield [41]. This is associated with the geochemical evolution of the lavas over time. Indeed, the overall increasing degree of silica undersaturation (Fig. 6) and the apparent enrichment in incompatible elements for samples with DI 6 35 (Fig. 7) suggest a decrease in the degree of partial melting. This may be associated with a chemical evolution of the source, revealed by the evolution of radiogenic isotope ratios over time (Fig. 8). The geochemical change of lavas during the post-collapse activity of Tahiti-Nui might then be associated with the decreasing contribution of the enriched EMII plume component. Although post-erosion lavas erupted after an apparent volcanic hiatus of about 250 kyr, and are even more basanitic in composition, their trace element and radiogenic isotope compositions suggest a general stabilization at a low degree of partial melting of the upper mantle. 6.5. Comparison with other oceanic islands 6.5.1. Volcano-tectonic evolution This study on Tahiti, evidencing two huge lateral collapses linked to a central rift zone, demonstrates that such mass wasting events are common during the eruptive history of oceanic islands. Catastrophic lateral sector collapses have been described in the case of other oceanic eruptive complexes: e.g. the Hawaiian Ridge [42^ 44], Re¤union Island [27,45], the western Canary

361

Islands, [38,46,47], some of the Galapagos Islands [48]. Those wasting stages seem mostly favored by dyke intrusions along rift zones (e.g. [49,46]), although gravitational spreading [50] as well as reactivation of faults in the basement [51] may also in£uence the development of such instabilities. Two main types of rift zones are classically distinguished [46]: (1) triple-armed rift zones, observed for example in Hawaii [52] and Tenerife Island [46], are likely developed by the fracturing of the oceanic crust in response to magma upward loading; (2) simple rift zones (e.g. La Palma, Canary Islands) may result from the concentration of magmas along a structural weakness in the basement [46,38]. On Tahiti-Nui, the main linear E^W rift zone would belong to the second type, although it is not aligned with the structural directions of the crust (N160 for the orientation of the ancient ridge and N070 for transform faults, e.g. [53]), nor with the direction of plate motion (N120). However, E^W submarine lineaments have recently been recognized in a wide zone to the south of Tahiti [54], indicating that this direction is an important feature of the plate. The Tahiti-Nui E^ W rift zone thus appears to be part of a major regional system. It may have been active through the whole history of the eruptive complex, since it in£uenced the E^W elongated shape of the island. It also clearly in£uenced the development of the northern and southern landslides, which occurred in opposite directions, at right angles to the main direction of dyke concentrations. In addition, subsidiary rift zones developed along the lateral rims of the northern U-shaped depression. They intrude the main shield but also the second shield, e.g. in the Fautaua valley (Fig. 3), showing that rift zone recon¢guration likely occurred subsequent to the main northern collapse. Those subsidiary rift zones, respectively oriented at N150 and N050 along the northwestern and northeastern rims of the northern collapse, are rather parallel to the structural directions of the underlying crust. This might indicate a reactivation of preexisting fractures in the basement subsequent to the load by the landslide products. Tahiti-Nui’s northern collapse thus resulted in a major structural change, which may have a¡ected both the

EPSL 6897 24-12-03

362

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

architecture of the island and the geometry of its oceanic basement. This may explain why the post-collapse second shield activity was ¢rst captured to the north which is in contradiction with the expected volcanic migration given the displacement of the plate to the northwest (Fig. 1). This evolution is signi¢cantly di¡erent from Hawaiian Islands, where landslides mainly occurred on both sides of the ridge [42], except the Hilina slump (e.g. [44]), while the activity mainly migrated along the ridge, in agreement with the plate displacement over the Hawaii hot-spot. Tahiti-Nui’s post-collapse evolution also di¡ers from the late history of Piton de La Fournaise at La Re¤union, where each major £ank collapse was followed by subsequent construction of a volcanic cone in the landslide depression [27,45], following an apparent migration of volcanic activity at almost right angles to the displacement of the plate. On Tahiti-Nui, the second shield was active for at least 400 kyr, but the magma productivity visibly declined rapidly after the ¢rst 100 kyr. This may be linked to a transfer of the main volcanic construction toward Tahiti-Iti and seamounts from the current hot-spot zone [55], since the growth of adjacent volcanoes may partly overlap in time in a hot-spot context [56]. The late posterosion phase here enhanced on Tahiti-Nui indicates an apparent volcanic hiatus of about 250 kyr, which is rather short compared to the case of Tahaa Island [10]. 6.5.2. Geochemical evolution The chemical evolution of Tahiti-Nui parental magmas is marked by a general increase in silica undersaturation together with gradually more depleted isotopic compositions, which is overall similar to trends previously described for Tahiti [10,14,20] and other oceanic islands, such as Hawaiian Islands (e.g. [57,58]). However, lavas from the Tahiti-Nui main shield-building phase are weakly alkaline, whereas typical tholeiites seem totally absent during the subaerial construction, in contrast with Hawaiian Islands, which are primarily composed of tholeiitic basalts (e.g. [30]). Lavas from older Society Islands and seamounts from the currently active hot-spot zone are also

weakly to strongly alkaline (e.g. [55,59,60]). This might indicate that the Society magmas have been persistently generated by low degrees of partial melting of the mantle, suggesting that the Society plume has been colder than the Hawaii hot-spot. Signi¢cant variations in La/Sm, 87 Sr/86 Sr and 143 Nd/144 Nd occur in lavas erupted immediately after the main northern landslide, indicating a sudden change in the conditions of magma generation. This may result from the decompression subsequent to landslide, which favored the brutal extrusion of magmas to the north and possibly provoked an increase in the extent of partial melting, as observed in the case of other ocean volcanic islands (e.g. [40]). However, the change here enhanced appears rather episodic since later lavas from the second shield follow the general pattern of increasing silica undersaturation. From 760 ka to about 480 ka, the source of Tahiti-Nui lavas rapidly changed toward more depleted signatures, whereas the global eruptive rate declined. In the meantime, episodic di¡erentiation processes occurred, likely due to crystal fractionation in the shallow reservoir, which partly intrudes the base of the second shield (Figs. 3 and 4). Tahiti-Nui’s late geochemical evolution thus shows similarities with the Hawaiian Islands, where extensive fractional crystallization mainly occurred during the later stages of volcanic history (e.g. [30]). However, lavas from Tahiti-Nui late post-erosion activity do not show di¡erences in isotopic composition compared to lavas from the later activity of the second shield, which is signi¢cantly di¡erent from the case of Tahaa (Society Islands) where post-erosion activity is characterized by strongly depleted compositions [10].

7. Conclusions The present study, based on combined geological, geochronological and geochemical data, has led to a detailed reconstruction of the volcanostructural evolution of Tahiti-Nui. The exposed volcanic history is ¢rst characterized by the construction of a main shield from 1.37 T 0.02 Ma to 0.87 T 0.02 Ma at a maximum eruptive rate of about 2 km3 /kyr. The lavas from this early build-

EPSL 6897 24-12-03

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365

ing stage are weakly alkaline, and slightly silicaundersaturated. La/Sm gradually increases over time, indicating a decreasing degree of partial melting of the mantle, whereas isotopic compositions are rather constant and close to the EMII. The northern collapse episode is dated between 0.87 T 0.02 Ma and 0.85 T 0.01 Ma, ruling out any possible period of volcanic quiescence. The subsequent activity was ¢rstly restricted to the northern depression, although some lavas were likely erupted from parasitic cones distributed along the rim of the northern depression. In around 100 kyr, most of the depression was ¢lled, corresponding to a high eruptive rate of about 5 km3 / kyr. Signi¢cant variations in La/Sm, 87 Sr/86 Sr and 143 Nd/144 Nd occur in lavas erupted immediately after the main northern landslide, indicating a sudden increase in the extent of partial melting likely caused by the decompression subsequent to collapse. However, the later activity was less productive, as evidenced by the lower eruptive rate in the period 760 ka^450 ka (around 0.5 km3 /kyr). In the meantime, lavas evolved toward more silicaundersaturated compositions indicating a decrease in the extent of partial melting of the mantle. The associated evolution of 87 Sr/86 Sr and 143 Nd/144 Nd ratios suggests a gradual change in the source linked to a decreasing involvement of the plume. Post-erosion lavas, although erupted after an apparent volcanic hiatus of about 250 kyr, are even more basanitic in composition, but the trace elements and radiogenic isotopes suggest a general stabilization of the source. The overall alkaline character of Tahiti-Nui lavas, as well as lavas from other Society eruptive complexes, suggests a relative weakness (low temperature and low eruptive rate) of the Society plume compared to the Hawaii hot-spot.

Acknowledgements We are grateful to M. Garcia and W.M. White for their constructive reviews, which helped to signi¢cantly improve the manuscript. We wish to thank Drs. Reymond and Talandier, from the Laboratory of Geophysics (CEA), and A. Bonne-

363

ville (University of French Polynesia) for their logistical help. Advice from A. Peccerillo was also very useful. Funding was obtained by the Re¤gion Ile de France Grant SESAME No. 947. This is LGMT contribution number 47.[BARD]

References [1] G. Deneufbourg, Carte ge¤ologique de Tahiti, notice explicative, BRGM, Orle¤ans, 1965. [2] J.M. Bardintze¡, H. Bellon, B. Bonin, R. Brousse, A.R. McBirney, Plutonic rocks from Tahiti-Nui caldera (Society Archipelago, French Polynesia): a petrological, geochemical and mineralogical study, J. Volcanol. Geotherm. Res. 35 (1988) 255^262. [3] B. Bonin, J.M. Bardintze¡, Plutonic rocks from TahitiNui caldera (French Polynesia). II. Evolution of thermodynamical parameters during magma di¡erentiation and emplacement. Applications to other alkaline magmatic provinces, Bull. Soc. Ge¤ol. Fr. 8 (1989) 1091^1099. [4] P. Petit, Etudes hydroge¤ologiques en Polyne¤sie Franc[aise, Service des Travaux Publics et des Mines - Centre Expe¤rimental de Recherches et d’Etudes du Ba“timent et des Travaux Publics, Papeete, 1969, 378 pp. [5] R. Jamet, Carte pe¤dologique de la Polyne¤sie Franc[aise ^ notice explicative, feuille de Tahiti, 1983. [6] C. Le¤otot, P.Y. Gillot, F. Guichard, R. Brousse, Le volcan de Taravao (Tahiti): un exemple de volcanisme polyphasique associe¤ a' une structure d’e¡ondrement, Bull. Soc. Ge¤ol. Fr. 8 (1990) 951^961. [7] R.A. Duncan, I. McDougall, Linear volcanism in French Polynesia, J. Volcanol. Geotherm. Res. 1 (1976) 197^227. [8] R. Brousse, C. Le¤otot, Mode'le d’e¤di¢cation de l’archipel de la Socie¤te¤ (Polyne¤sie Franc[aise), C.R. Acad. Sci. Paris 307 (1990) 533^536. [9] C. Diraison, H. Bellon, C. Le¤otot, R. Brousse, H.G. Barsczus, L’alignement de la Socie¤te¤ (Polyne¤sie franc[aise): aise): volcanologie, ge¤ochronologie, proposition d’un mode'le de point chaud, Bull. Soc. Ge¤ol. Fr. 162 (1991) 479^496. [10] W.M. White, R.A. Duncan, Geochemistry and geochronology of the Society Islands: New evidence for deep mantle recycling, in: A. Basu, S. Hart (Eds.), Earth processes: Reading the Isotopic Code, AGU Geophys. Monogr. 95 (1996) 183^206. [11] J. Talandier, E.A. Okal, The volcanoseismic swarms of 1981^1983 in the Tahiti-Mehetia area, French Polynesia, J. Geophys. Res. 89 (1984) 11,216^11,234. [12] J.L. Chemine¤e, R. He¤kinian, J. Talandier, F. Albare'de, C.W. Devey, J. Francheteau, Y. Lancelot, Geology of an active hot-spot Teahiti-Mehetia region in the south central Paci¢c, Mar. Geophys. Res. 11 (1989) 27^50. [13] N. Binard, R.C. Maury, G. Guille, J. Talandier, P.Y. Gillot, J. Cotton, Mehetia Island, South Paci¢c: geology

EPSL 6897 24-12-03

364

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21] [22]

[23]

[24]

[25]

[26]

[27]

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365 and petrology of the emergent part of the Society hot spot, J. Volcanol. Geotherm. Res. 55 (1993) 239^260. R.A. Duncan, M.R. Fisk, W.M. White, R.L. Nielsen, Tahiti: geochemical evolution of a French Polynesian volcano, J. Geophys. Res. B Solid Earth Planets 99 (1994) 24,341^24,357. R. Brousse, G. Boutault, A. Eisenstein, P. Gelugne, Tahiti-Nui, carte ge¤ologique (1/25000), carte des formations de surface (1/25000), carte des instabilite¤s et risques naturels (1/50000), feuille Papeete, Ministe're de l’e¤quipement, de l’ame¤nagement, de l’energie et des mines, 1985. R. Brousse, P. Gelugne, Tahiti-Nui, carte ge¤ologique (1/25000), carte des formations de surface (1/25000), carte des instabilite¤s et risques naturels (1/50000), feuille Punaauia-Paea, Ministe're de l’e¤quipement, de l’ame¤nagement, de l’energie et des mines, 1987. R. Brousse, P. Lavest, C. Le¤otot, Tahiti-Nui, carte ge¤ologique (1/25000), carte des formations de surface (1/25000), carte des instabilite¤s et risques naturels (1/50000), feuille Papara^Teva I Uta^Taravao, Ministe're de l’e¤quipement, de l’ame¤nagement, de l’energie et des mines, 1990. J. Dostal, C. Dupuy, J.M. Liotard, Geochemistry and origin of basaltic lavas from Society Islands, Bull. Volcanol. 45 (1982) 51^62. P. Roperch, A. Chauvin, Transitional geomagnetic ¢eld behavior: Volcanic records form French Polynesia, Geophys. Res. Lett. 14 (1987) 151^154. Q.C. Cheng, J.D. MacDougall, G.W. Lugmair, Geochemical studies of Tahiti, Teahetia and Mehetia, Society Island Chain, J. Volcanol. Geotherm. Res. 55 (1993) 155^ 184. R. Brousse, La ge¤ologie des “|les hautes, Atlas de la Polyne¤sie Franc[aise, O.R.S.T.O.M., planche 28, 1993. P.Y. Gillot, H. Talandier, H. Guillou, I. Leroy, Temporal evolution of the volcanism at Tahiti Island: Geology and structural evolution, in: International Workshop on Intraplate Volcanism: The Polynesian Plume Province, Papeete, Tahiti, 1993. V. Clouard, A. Bonneville, P.Y. Gillot, A giant landslide on the southern £ank of Tahiti Island, French Polynesia, Geophys. Res. Lett. 28 (2001) 2253^2256. G. Boutault, Tahiti (Polyne¤sie Franc[aise, Archipel de la Socie¤te¤): volcanologie, pe¤trographie et ge¤otechnie du secteur co“tier nord-est, The'se 3‡ cycle, Universite¤ Paris-Sud, Orsay, 1985, 262 pp. A. Hildenbrand, P.Y. Gillot, Evidence for a di¡erentiated ignimbritic activity ending the building-stage of Tahiti (French Polynesia), in: EUG XI TERRA Nova, Strasbourg, 2001, 85 pp. C. Cassignol, P.Y. Gillot, Range and E¡ectiveness of Unspiked Potassium-Argon Dating: Experimental Groundwork and Applications, Wiley, New York, 1982, pp. 159^ 179. P.Y. Gillot, J.C. Lefe'vre, P.E. Nativel, Model for the structural evolution of the volcanoes of Re¤union Island, Earth Planet. Sci. Lett. 122 (1994) 291^302.

[28] P.Y. Gillot, Y. Cornette, The Cassignol technique for potassium-argon dating, precision and accuracy examples from late Pleistocene to recent volcanics from southern Italy, Chem. Geol. 59 (1986) 205^222. [29] M. Becker, R. Brousse, G. Guille, H. Bellon, Phases d’e¤rosion-comblement de la valle¤e de la Papenoo et volcanisme subre¤cent a' Tahiti, en relation avec l’e¤volution des “|les de la Socie¤te¤ (Paci¢que Sud), Mar. Geol. 16 (1974) 71^77. [30] G.A. Mcdonald, T. Katsura, Chemical composition of Hawaiian lavas, J. Petrol. 5 (1964) 82^133. [31] A.R. McBirney, K.J. Aoki, Petrology of the island of Tahiti, GSA Mem. 116 (1968) 523^556. [32] M. Wilson, Igneous Petrogenesis. A Global Tectonic Approach, Chapman and Hall, London, 1989. [33] W.M. White, Geochemical tracers of mantle processes, Rev. Geophys. 33 (Suppl. S) (1995) 19^24. [34] W.M. White, Lecture Notes in Isotope Geochemistry, http://www.geo.cornell.edu/-geology/classes/Geo656/ 656Notes00.html, 2000. [35] H.R. Rollinson, Using Geochemical Data: Evaluation, Presentation, Interpretation, Longman, London, 1993, 352 pp. [36] C. Chauvel, A.W. Hofmann, P. Vidal, HIMU-EM: The French Polynesian connection, Earth Planet. Sci. Lett. 110 (1992) 99^119. [37] J.J. Dvorak, D. Dzurisin, Variations in magma supply rate at Kilauea Volcano, Hawaii, J. Geophys. Res. 98 (1993) 22,255^22,268. [38] A. Hildenbrand, P.Y. Gillot, V. Soler, P. Lahitte, Evidence for a persistent uplifting of La Palma (Canary Islands), inferred from morphological and radiometric data, Earth Planet. Sci. Lett. 210 (2003) 277^289. [39] A. Hildenbrand, Etude ge¤ologique de l’|“le volcanique de Tahiti-Nui (Polyne¤sie franc[aise): e¤volution morpho-structurale, ge¤ochimique et hydrologique, Paris-Sud, 2002, 250 pp. [40] T.K. Presley, J.M. Sinton, M. Pringle, Postshield volcanism and catastrophic mass wasting of the Waianae Volcano, Oahu, Hawaii, Bull. Volcanol. 58 (1997) 597^616. [41] V. Pinel, C. Jaupart, The e¡ect of edi¢ce load on magma ascent beneath a volcano, Phil. Trans. R. Soc. London 358 (2000) 1515^1532. [42] J. Moore, D. Clague, R. Holcomb, P. Lipman, W. Normark, M. Torresan, Prodigious submarine landslides on the Hawaiian ridge, J. Geophys. Res. 94 (1989) 17645^ 17664. [43] R.T. Holcomb, R.C. Searle, Large landslides from oceanic volcanoes, Mar. Geotechnol. 10 (1991) 19^32. [44] J.R. Smith, A. Malaho¡, A.N. Shor, Submarine geology of the Hilina slump and morpho-structural evolution of Kilauea volcano, Hawaii, J. Volcanol. Geotherm. Res. 94 (1999) 59^88. [45] W.A. Du⁄eld, L. Stieljes, J. Varet, Huge landslide blocks in the growth of Piton de la Fournaise, La Re¤union and Kilauea volcano, Hawaii, J. Volcanol. Geotherm. Res. 12 (1982) 147^160.

EPSL 6897 24-12-03

A. Hildenbrand et al. / Earth and Planetary Science Letters 217 (2004) 349^365 [46] J.C. Carracedo, The Canary Islands: an example of structural control on the growth of large oceanic-island volcanoes, J. Volcanol. Geotherm. Res. 60 (1994) 225^241. [47} N.C. Mitchell, D.G. Masson, A.B. Watts, M.J.R. Martin, J.R. Gee, R. Urgeles, The morphology of the submarine £anks of volcanic ocean islands: A comparative study of the Canary and Hawaiian hotspot islands, J. Volcanol. Geotherm. Res. 115 (2002) 83^107. [48] D. Geist, W.M. White, F. Albare'de, K. Harpp, R. Reynolds, R. Blichert-Toft, M. Kurz, Volcanic evolution in the Galapagos The dissected shield of Volcan Ecuador, Geochem. Geophys. Geosyst. 3 (2003) 1^32. [49] G.P.L. Walker, Three Hawaiian calderas: An origin through loading by shallow intrusions?, J. Geophys. Res. 93 (1988) 14773^14784. [50] J.P. Callot, L. Geo¡roy, O. Merle, Approche expe¤rimentale de la tectonique de l’Etna, Bull. Soc. Ge¤ol. Fr. 171 (2000) 59^70. [51] N. Vidal, O. Merle, Reactivation of basement faults beneath volcanoes: a new model of £ank collapse, J. Volcanol. Geotherm. Res 99 (2000) 9^26. [52] G.A. Mcdonald, Volcanoes, Englewood Cli¡s, NJ, 1972. [53] J. Mammerickx, E. Herron, L. Dorman, Evidence for two fossil spreading ridges in the southeast Paci¢c, Geol. Soc. Am. Bull. 91 (1980) 263^271.

365

[54] V. Clouard, A. Bonneville, P.Y. Gillot, The Tarava Seamounts: a newly characterized hotspot chain on the South Paci¢c Superswell, Earth Planet. Sci. Lett. 207 (2003) 117^130. [55] I. Leroy, Evolution des volcans en syste'me de point chaud: Ile de Tahiti, Archipel de la Socie¤te¤ (Polyne¤sie Franc[aise), Universite¤ Paris-Sud, 1994, 271 pp. [56] D.J. DePaolo, E.M. Stolper, Models of Hawaiian volcano growth and plume structure: Implications of results from the Hawaii Scienti¢c Drilling Project, J. Geophys. Res. 101 (1996) 11,643^11,654. [57] H.T. Stearns, Geology of the Hawaiian Islands, Hawaiian Div. Hydrogr. 8 (1946) 112 pp. [58] C.Y. Chen, F.A. Frey, Origin of Hawaiian tholeiite and alkalic basalt, Nature 302 (1983) 785^789. [59] A. Le Dez, R. Maury, H. Guillou, J. Cotten, S. Blais, G. Guille, Moorea (Society Islands): rapid build-up of a Polynesian volcano, Geol. Fr. Surrounding Areas 3 (1998) 51^64. [60] C.W. Devey, F. Albare'de, J.L. Chemine¤e, A. Michard, R. Mu«he, P. Sto¡ers, Active submarine volcanism on the Society hot-spot swell (West Paci¢c): A geochemical study, J. Geophys. Res. 95 (1990) 5049^5067.

EPSL 6897 24-12-03

Lihat lebih banyak...

Comentarios

Copyright © 2017 DATOSPDF Inc.