Effusive history of the Grande Découverte Volcanic Complex, southern Basse-Terre (Guadeloupe, French West Indies) from new K–Ar Cassignol–Gillot ages

Journal of Volcanology and Geothermal Research 187 (2009) 117–130

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research

j ourna l homepage: www.e lsev ie r.com/ locate / jvo lgeores

Effusive history of the Grande Découverte Volcanic Complex, southern Basse-Terre(Guadeloupe, French West Indies) from new K–Ar Cassignol–Gillot ages

A. Samper a, X. Quidelleur a,⁎, J.-C. Komorowski b, P. Lahitte a, G. Boudon b

a IDES, Equipe Géochronologie et Dynamique des Systèmes Volcaniques, UMR 8148 CNRS-UPS, Université Paris-Sud11, Orsay, Franceb Institut de Physique du Globe de Paris (IPGP), CNRS UMR 7154, Equipe de Géologie des Systèmes Volcaniques, Tour 46, 2 place Jussieu, 75005 Paris, France

⁎ Corresponding author.E-mail address: [email protected] (X. Quid

0377-0273/$ – see front matter © 2009 Elsevier B.V. Aldoi:10.1016/j.jvolgeores.2009.08.016

a b s t r a c t

Article history:Received 2 December 2008Accepted 11 August 2009Available online 31 August 2009

Keywords:geochronologyK–Ar datingLesser Antilles arcGuadeloupeGrande DécouverteSoufrièreHolocene volcanism

The Grande Découverte Volcanic Complex (GDVC), active since at least 0.2 Ma, is the most recent volcaniccomplex of the Basse-Terre Island (Guadeloupe, Lesser Antilles Arc). A detailed geochronological study usingthe K–Ar Cassignol–Gillot technique has been undertaken in order to reconstruct the history of effusiveactivity of this long-lived volcanic system. Twenty new ages permit to suggest that the GDVC experienced atleast six main effusive stages, from 200 ka to present time. To the north of the GDVC, the GDS (GrandeDécouverte–Soufrière volcano) has been active since at least 200 ka, and to the south, the TRMF (Trois-Rivières–Madeleine Field), started to be emplaced 100 ka. Morphological investigations suggest that thewhole TRMF volcanism was emitted from vents distinct from the GDS, most probably a large E–W fissurenetwork linked to the Marie-Galante rift. The mean age of 62±5 ka, obtained for the E–W Madeleine–LePalmiste alignment suggests that a fissure-opening event occurred at that time. However, whole-rock majorand trace element signatures are similar for both systems, suggesting that a common complex magma-plumbing system has fed the overall GDVC. We report very young ages for lava flows from the TRMF, whichimplies that <10 ka volcanic activity is now identified for both massifs. Although hazards associated withsuch effusive volcanism are much lower than those associated with potential flank-collapse of the Soufrièrelava dome or a magmatic dome eruption with explosive phases within the GDS, the emplacement ofrelatively large Holocene age lava flows (3–1×108 m3) suggests that a revised integrated volcanic hazardassessment for Southern Basse-Terre should now consider the potential for renewed future activity from twoHolocene volcanic centers including the TRMF.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

The lifetime of an active composite volcano is punctuated byalternating constructive and destructive episodes, both of whichcontribute to shape the edifice. It is well documented that compositevolcanoes are built through pulses of activity (Davidson and De Silva,2000). For each of them, volumes erupted and time length of eacheffusive phase depend on the volcanic complex considered, as well ason those parameters that are linked to local geological and regionalgeodynamic settings (Wadge, 1984). A better understanding of avolcano's global past behavior can be achieved by recognizing theseveral effusive and explosive phases and destructive (collapse andflank-collapse) events as much as field conditions make possible anyexhaustive sampling of those. Combined with the use of geochrono-logical tools, such study can be relevant in establishing an accurateand comprehensive volcanic history (e.g., Ozawa et al., 2005; Calvertet al., 2006), which, by comparison with analogous volcanoes, leads

elleur).

l rights reserved.

towards an improved understanding of volcanic processes in general(White et al., 2006) and in particular of single volcanic complexes(e.g., Sherrod et al., 2007; Rouchon et al., 2008). Moreover, suchreconstructions are important for understanding the sequence oferuptive processes of a given volcano and to better estimate theassociated hazard assessment (e.g., Robertson, 1995; Sparks, 2003).Here we present the results of a K–Ar dating of the effusive phases ofthe Grande Découverte volcanic complex (GDVC), located in thesouthern part of the island of Basse-Terre, Guadeloupe, an activevolcanic complex of the Lesser Antilles Arc. The GDVC covers an areaof 129 km2. Today, 73,000 people (1999 census) live within 15 km ofthe Soufrière lava dome.

We have previously investigated the timing of the effusive activityin the northern part of this volcanic island by using the K–ArCassignol–Gillot technique (Samper et al., 2007). We confirmed thatvolcanism has been migrating southwards since 2.8 Ma, throughoutthewhole Basse-Terre Island, reaching its present-day focuswithin theGDVC. Following the same approach as Samper et al. (2007), this studyis dedicated at identifying the main successive effusive phases of theGDVC, at better constraining their chronology and geochemistry, inorder to reconstruct the multi-stage eruptive history of the GDVC.

118 A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

2. Geological setting

The volcanic island of Basse-Terre, part of the Guadeloupearchipelago, belongs to the inner arc of the Lesser Antilles (Fig. 1)and is located within the volcanic front active since the Plio-Pleistocene (Fig. 1). The present volcanic morphology of Basse-Terreresults from the competition from late Pliocene to late Pleistocene ofthe NW–SE Basse-Terre–Montserrat en-échelon normal fault systemwith the E–W rift systems of La Désirade and then of Marie-Galante(Feuillet et al., 2002, 2004). Thus, the volcanic activity of the island canbe divided into two lengthy eruptive episodes (Samper et al., 2007).From 2.8 to 1.15 Ma sub-aerial activity has been set in the northernhalf along a N–S trending lineament within two volcanic chains(Fig. 1). The Basal Complex, with ages between 2.79 and 2.69 Ma, isthe northernmost and smallest massif of the island. The NNW–SSESeptentrional Chain (SC), characterized by fissure volcanism, wasthen emplaced south of the Basal Complex (BC) between 1.8 and1.15 Ma. Volcanism of the last million years migrated southeastward.The Axial Chain (AC), made of several composite volcanoes of a globalvolume of 180 km3 stayed active during about 0.6 Myr, from 1 to0.435 Ma (Samper et al., 2007). It extends over most of the southernhalf of Basse-Terre Island and defines the northern limit of the GDVC.The GDVC, as themost recent volcanic complex on the island, occupiesthe southern one-third of Basse-Terre. Two distinct massifs have beendescribed within the GDVC on the basis of their different morphologyand location (e.g., Boudon et al., 1988; Komorowski et al., 2005): thecomposite volcano of Grande Découverte–Soufrière (GDS) and theTrois-Rivières–Madeleine volcanic field (TRMF).

The GDS is a composite volcano built upon older lava flowserupted from the southern end of the Axial Chain (Boudon et al.,1988; Komorowski et al., 2005). By the use of stratigraphic and

Fig. 1. (a) Localization of the Guadeloupe archipelago within the Lesser Antilles arc. Plate-mplate. (b) Guadeloupe Islands shaded digital elevation model (light from NW; data from Instisland are from Blanc (1983) and Samper et al. (2007), and localization of the normal faults frthe summit (1463 m) of Soufriere lava dome. (c) Close-up of the GDVC and location of the 2Blanc, 1983; six from Carlut et al., 2000). Note that the GDVC is located between the 600–4

petrologic arguments, GDS's activity was estimated to have started atabout 200 ka (Boudon et al., 1989, 1992). A K–Ar age of 205±28 kawas then obtained on the uppermost lava flow section from thenorthern rim of the Grande Découverte caldera (Carlut et al., 2000)and this now indicates that the GDS was already built by about200 ka. Three main volcanic stages have been defined for the GDS.The first stage is then constrained between at least 205±28 (Carlutet al., 2000) and 42±2 ka BP. It is known as the Grande Découvertephase. The latter age is a 14C age obtained for the Pintade andesiticPlinian pumice deposit (Boudon et al., 1988; Komorowski et al.,2005). The Pintade pumice consists of residual fallout tephra andthick valley-ponded pyroclastic flow deposits that are found inseveral valleys in southern Basse-Terre (over 120 km2) and representan estimated volume of 0.5 to 3 km3 (Komorowski et al., 2005).Residual outcrops of at least two additional quartz dacitic pumicedeposits, the Anse des Pères pumices and Montval pumices can befound scattered over large areas of the GDVC. Thermoluminescenceages of 140±14 ka and 108±10 ka were obtained for the Anse desPères and Montval deposits, respectively (Blanc, 1983). However,recent work on thermoluminescence dating on quartz from St. Luciadacite lavas suggests that thermoluminescence ages on LesserAntillean quartz should be treated with caution (Guérin and Samper,2007). However, it is reasonable to associate this explosive volcanismto the GDS's first stage, suggesting that otherwise chiefly effusiveactivity has been interrupted several times by major explosiveeruptions. Several flank-collapse episodes are also inferred from thepresence of older debris avalanche deposits (Komorowski et al.,2005; Boudon et al., 2007). The Pintade andesitic pumice deposit,which ended the first GDS stage, is thought to represent the sequenceassociated with the formation of the Grande Découverte caldera(Boudon et al., 1988).

otion vectors from Demets et al. (2000) and Dixon et al. (1998); NAM, North Americanitut Géographique National (IGN). Ages of the main volcanic massifs of the Basse-Terreom Feuillet et al. (2002). The thick square inset shows the location of panel c. The star is9 K–Ar ages from lava flows and lava domes (twenty ages from this study; three from00 ka southern Axial Chain and the Monts Caraïbes Massif (555–445 ka).

Fig. 2. (a) Age-probability density diagram (Deino and Potts, 1992) outlining effusivephases maxima and maximum duration length for the GDVC over the last 350 ka(N=29; see Fig. 1 caption). (b) Close-up of the 150–0 ka interval, where five mainperiods of effusive activity can be defined: 150–100, 100–70, 70–48, 48–15 and <15 ka.

119A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

The Carmichaël composite volcano was then built within theGrande Découverte caldera, during the second stage of the GDS,beginning about 42 ka (Boudon et al., 1988; Komorowski et al., 2005).Its eruptive history, which consists of a succession of lava flows andlava domes associated with pyroclastic deposits (Komorowski et al.,2005), spanned until 11.5 ka. The Carmichaël composite volcano wasthen destroyed by at least two flank-collapse events at about 13.5 and11.5 ka (Boudon et al., 1987, 1988, 1989; Komorowski et al., 2005;Boudon et al., 2007). The GDS third and last eruptive stage is thepresent Soufrière phase. It is characterized by alternating lava domeeruptions and prolonged periods of phreatic explosive to non-explosive activity. Small volume edifice collapses have occurredfrequently, as revealed by the numerous debris avalanche depositsfound down to the Caribbean coast, in the SW part of the GDS(Komorowski et al., 2005; Boudon et al., 2007; Komorowski et al.,2008b). The three youngest edifices to have been emplaced on top ofthe GDS are the monogenetic edifices of L'Echelle, La Citerne and thelava dome of la Soufrière s.s. (Semet et al., 1981; Boudon et al., 1988).Recently, Boudon et al. (2008) revised the age of the last eruption – theemplacement of la Soufrière lava dome – to A.D. 1530, and theyproposed a scenario involving injection of basic magma into an upperdifferentiated magma-chamber. Such injection led to a summit flankcollapse of the highly altered edifice followed by an explosivesubplinian eruption, because of the sudden reduction of the edificeload (Komorowski et al., 2008a,b). These eventswere then followed bythe emplacement of the Soufrière lava dome, of less differentiatedmagma (Boudon et al., 2008).

The Trois-Rivières–Madeleine volcanic field (TRMF), located to thesouth of the GDS (Fig. 1), includes on its northern border the E–Walignment of Le Palmiste lava flow, Gros Fougas and La Madeleineedifices, and is bounded to the southwest by the Monts Caraïbes(Fig. 1). It consistsmainly of the prominent lava dome of LaMadeleine,of smaller domes and their associated lava flows, of thick lava flowsand some monogenetic centers. Unlike dome eruptions of the GDS,only minor pyroclastic deposits associated with lava domes arepresent in the TRMF (Komorowski et al., 2005). No K–Ar ages werepreviously available to constrain the eruptive effusive successions ofthis area.

3. Analyses and methodology

3.1. Sampling

Due to the tropical climate and because easterly winds are stoppedby the topographic highs of the GDVC, strong erosion, densevegetation and the development of thick soils are favored throughoutthe whole massif. Bedrock exposures are rather scarce and limited torivers, paths and road cuts, and quarries, even so sampling was alsoundertaken in some of the highest inner parts of the massif (between1000 and 1200 m elevation). Overall, during winter 2001 and March2004, 39 hand-sized lava blocks were taken on lava flows and domesfor both geochronological and geochemical studies. The SAM58sample was collected 58 m below ground level in a borehole locatedat the SW base of the La Soufrière lava dome on the Savane à Muletparking site.

3.2. Geochemical data

Whole-rock major and trace element analyses were performed atthe Service d'Analyse des Roches et des Minéraux (SARM)—Centre deRecherches Pétrographiques et Géochimiques (CRPG) (CNRS, Nancy,France) on all 29 dated samples. Samples were fused with LiBO2 thendissolved with HNO3. Major element contents were determined byICP-AES (Jobin-Yvon JY 70) and trace element concentrations by ICP-MS (Perkin Elmer 5000). Uncertainties for each element can be foundat: http://www.crpg.cnrs-nancy.fr/SARM/index.html.

3.3. K–Ar procedure

After careful examination of petrographic thin sections of samplesfrom the main lava flows, 20 samples (Fig. 1) were chosen for datingin the LGMT geochronology laboratory at Orsay (France) using the K–ArCassignol–Gillot technique (Cassignol and Gillot, 1982; Gillot andCornette, 1986). Sample procedure is described elsewhere (e.g., Chenetet al., 2007; Samper et al., 2007). Sampleswere crushed toa125–250 μmsize fraction and ultrasonically cleaned for 15 min in a 1 M nitricacid solution. In order to avoid weathered material (with itsattendant K loss), as well as excess argon carried by mafic mineralsand lowK-contentplagioclases (Harford et al., 2002; Samper et al., 2007,2008), analyseswere performed ona narrowdensity range groundmassphase which was carefully separated by the use of heavy liquids. Amagnetic separator was then used to remove remaining microlithicplagioclases. K was measured by flame emission spectrometry andcompared with reference values of MDO-G and ISH-G standards (Gillotet al., 1992). Ar was measured with a mass spectrometer similar to theone described by Gillot and Cornette (1986). The interlaboratorystandard GL-O, with the recommended value of 6.679×1014 atom/gof 40Ar* (Odin et al., 1982), was used for 40Ar signal calibration. Typicaluncertainties of 1% are achieved for the 40Ar signal calibration and for

120 A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

the K determination. The uncertainty on the 40Ar* determination is afunction of the radiogenic content of the sample. The detection limit ofthe system is 0.1% of 40Ar* (Quidelleur et al., 2001), which makes theCassignol–Gillot technique especially suitable for dating samples fromyoung volcanic complexes such as those investigated here. In order toavoid isotopic fractionation, no pre-degassing of the sample wasperformed (Gillot and Cornette, 1986). Decay constants and isotopicratios of Steiger and Jäger (1977) were used. K contents and Armeasurements were duplicated for each sample. Uncertainties hereinare given at the 1σ level.

4. Results

4.1. K–Ar results

We present in Table 1 twenty new K–Ar ages obtained on lavaflows from the GDVC. Our study provides seven new K–Ar ages on theGDS, which adds to the data set of nine K–Ar ages previously availablefor the composite edifice (Table 2; Blanc, 1983; Carlut et al., 2000).Wealso obtained thirteen new K–Ar ages within the TRMF where no datawere previously available. Radiogenic 40Ar concentrations range from0.0 to 14.5%, and all analyses have been duplicated at the 1σ level.

Previous K–Ar ages from the GDVC constrained its effusive activitybetween 205±28 and 47±21 ka (Table 2; Blanc, 1983; Carlut et al.,

Table 1New K–Ar ages from the GDVC. GDS: Grande Découverte–Soufrière volcano, TRMF: Trois Riargon. Mean: mean age obtained by weighting each age determination by its radiogenic argoreferable to WGS84.

Sample Site Massif Longitude Latitude

04GW04 Morne Laffite TRMF 61°37.60′ 16°00.38′

04GW14 Petite Montagne TRMF 61°36.88′ 16°00.08′

01GU55 Ravine Malanga GDS 61°40.74′ 16°02.60′

04GW22 Rivière Noire 3 GDS 61°40.73′ 16°02.30′

SAM58 Savane à Mulet GDS 61°39.91′ 16°02.33′

04GW30 Morne Joseph GDS 61°39.88′ 16°01.25′

00GU53 Roches Gravées TRMF 61°38.40′ 15°58.21′

04GW32b Petit Carbet TRMF 61°38.07′ 15°58.64′

01GU61b Fond-Graine W TRMF 61°38.91′ 15°00.33′01GU61a98GU32 Fond-Graine E TRMF 61°38.91′ 16°00.22′

01GU66 Le Palmiste TRMF 61°41.65′ 15°59.98′

04GW31 Bassin Bleu GDS 61°39.95′ 16°00.17′

04GW37 Gros Fougas TRMF 61°39.53′ 16°07.21′

01GU60 Piton Tarare TRMF 61°38.64′ 16°00.75′

01GU63 Nez Cassé GDS 61°40.28′ 16°02.93′

04GW26 Anse Duquery TRMF 61°37.91′ 15°58.33′

01GU65 La Coulisse TRMF 61°37.92′ 15°58.53′

04GW24 Rivière Noire 5 GDS 61°41.32′ 16°01.59′

04GW02 Racoon TRMF 61°15.09′ 16°00.60′

04GW05 Morne Liquin TRMF 61°36.48′ 16°00.83′

2000). Our new data extend this record into Holocene time (Table 1).The TRMF activity ranges from 87±2 ka (04GW30) to the Holocene(0±1 ka; 04GW02). Four very recent ages (<10 ka) have beenobtained within the GDS and TRMF.

The twenty-nine K–Ar ages now available (this study; Table 1, andBlanc, 1983; Carlut et al., 2000) have been plotted in age-probabilityspectra (Deino and Potts, 1992; Fig. 2). Such approach is biased by thedistribution of sampling sites; however, it allows us to identify themain time ranges of successive periods of construction of the wholemassif. These are 450–150, 150–00, 100–70, 70–48, 48–15, and<15 ka (Figs. 2 and 3), with most of our data falling into the 150–0 kainterval. A previous age of 205±28 ka (Carlut et al., 2000) wasobtained for a summit sample of the GDS (Table 2; Figs. 2 and 3). Thisage represents the lower bound of the 200–0 ka interval over whichGDS activity has been dated. The oldest ages (Fig. 2) obtained by us arefrom samples from the lava domes of Morne Laffite and PetiteMontagne (04GW04, 314±12 ka and 04GW14, 261±7 ka, respec-tively; Fig. 1), located in the south-east part of the GDVC. These twodomes belong to our oldest time interval. On the basis of the sixeffusive eruptive periods determined by the 29 twenty-nine K–Ar agesnow available, together with previous geological mapping (Boudonet al., 1988), we present in Fig. 3 an interpretative geochronologicalmap of the whole GDVC and in Fig. 4 a synthetic schematic chronologyof the whole volcanic history of Basse-Terre island.

vières–Madeleine volcanic Field. K (%): potassium content in percent. 40Ar*: radiogenicn content (40Ar*). Uncertainties are given at the one-sigma level. Latitude and longitude

K(%)

40Ar*(%)

40Ar* ×1010

(atom/g)Age(ka)

Mean(ka)

0.614 3.1 20.395 318±11 314±122.6 19.897 310±13

0.476 4.6 13.104 264±7 261±74.4 12.895 259±7

1.198 2.7 14.283 112±4 113±43.7 14.263 114±4

1.321 14.5 15.488 112±2 113±22.5 16.367 119±5

0.900 0.6 8.6229 92±16 94±160.6 8.9542 95±15

0.985 7.2 9.2776 90±2 89±21.9 8.9360 87±5

0.547 1.6 4.8099 84±5 87±51.8 4.9498 87±51.6 5.1296 90±6

0.800 2.5 7.0923 85±4 80±32.8 6.3636 76±3

0.249 0.7 2.2179 85±12 76±120.280 0.5 1.8373 63±120.335 0.5 2.5042 72±15 75±14

0.6 2.6999 77±141.073 2.4 7.6965 69±3 67±3

2.4 7.2329 65±30.991 2.3 6.6586 64±3 65±3

1.6 6.8426 66±41.011 1.5 6.9368 66±5 60±4

1.5 5.8277 55±40.560 2.1 3.3826 58±3 58±3

1.9 3.4539 59±30.712 0.8 2.4113 32±4 34±4

0.9 2.5974 35±40.834 0.4 2.4861 29±7 29±6

0.5 2.5015 29±60.593 0.1 0.44885 7±6 6±6

0.0 0.13568 2±70.847 0.3 0.2941 3±1 6±2

0.5 0.7254 8±21.151 –0.2 −0.395 –3±1 0±1

0.0 0.0679 1±10.724 0.0 0.0946 1±20 0±20

−0.1 −0.0185 −24±23

Fig. 3. Interpretative age-geologic map of the GDVC. (Basse-Terre Island digital elevation model from IGN). Age interpretation from data herein applied to pre-existing geologicalmap (Boudon et al., 1988). Ages from the southern Axial Chain and Monts Caraïbes volcanoes from Blanc (1983) and Samper et al. (2007). Grey areas are not discussed herein andinclude Axial Chain volcanism as well as more recent products up to the Holocene age.

121A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

4.2. Geochemistry results

Major and trace elements data are given in Table 3. For each of themain eruptive phases identified, a K2O vs SiO2 diagram (Gill, 1981) ispresented (Fig. 5). As already observed for the 2.8–0.435 Ma NorthernBasse-Terre volcanics (Samper et al., 2007), GDVC lavas younger than0.25 Ma also belong to the medium- and low-K types (Fig. 5). Mostsamples are andesites and basaltic andesites; the only basalt is the lavadome of Morne Liquin of the <15 ka phase (04GW05, Table 1 andFig. 1). When looking at the evolution of major elements as a functionof time (Fig. 6), an overall decrease of SiO2 and K2O can be observedthrough time, when an increase of MgO occurs, suggesting eruption ofincreasinglymore primitivemagmas. However, Fe2O3, Na2O and Al2O3

lack any clear trend.Similarly to the Northern Basse-Terre massifs (Samper et al.,

2007), GDVC lavas show systematic MgO concentrations lower than6 wt.% (from 1.91 to 5.88%, Fig. 7a), which is a recurrent feature ofthe Central and Northern islands of the Lesser Antilles arc(Macdonald et al., 2000). Although an overall increase can beinferred for MgO, it does not show any clear trend though time(Fig. 7a). GDVC lavas display narrow values for La/Yb chondrite-normalized ratios, which remain more or less constant through time(Fig. 7b). These values, from 1.96 to 2.5, indicate slight LREEenrichment typical of medium-K calc-alkaline rocks (Gill, 1981).Note that the two lava domes of the 450–250 ka period, PetiteMontagne (04GW14) and Morne Laffite (04GW04), display respec-tively the lowest and highest values of the data set for both MgO and

La/Yb ratios (Fig. 7a and b). When compared with the NorthernBasse-Terre complexes (Basal Complex, BC; Septentrional Chain, SC;Axial Chain, AC; Samper et al., 2007), Mg# values and La/Yb ratios forGDVC lavas are similar with the 1.8–1.15 Ma SC lavas. These values,which are much higher than for lavas from the 2.8 Ma BC and the1 Ma–0.435 ka AC (Fig. 7c and d), characterize more differentiatedlavas but do not show characteristic differentiation trends such asobserved for BC and AC lavas (Fig. 7d). The higher values of the La/Ybratio could be interpreted as a higher partial fusion of the source forboth SC and GDVC lavas.

Rare earth elements (REE) and multi-elements diagrams displaysimilar calc-alkaline patterns for both GDS and TRMF areas (Fig. 8a–d).Slight enrichments in LREE (Fig. 8a and b) typical of medium-K suitesand other typical subduction-related features can be observed, such asLILE, U, Pb and Sr enrichments, Nb and Ta depletions (Fig. 8c and d).Numerous lavas show negative Ce anomalies, which have beendescribed elsewhere as the effects of post-magmatic processes relatedto sub-aerial exposure under tropical conditions (Cotten et al., 1995).Multi-element patterns for each of the GDVC effusive phase (Fig. 8e–i)permit to identify lavas of very close chemistry. Our oldest GDS lava,GU08, is significantly different from the chronologically nearest lavas,the two 113 ka lavas, 01GU55 and 04GW22, which are identical exceptfor Ce depletion. Other lavas, such as 04GW37 and 04GW31, and,01GU66 and 01GU60, for the 70–48 ka phase (Fig. 8g), GU05 and04GW26 for the 48–15 ka phase (Fig. 8h) and 04GW24, 04GW02 and01GU65 for the< 15 kaphase (Fig. 8i) alsodisplay very similar patterns.Note that 01GU63 and 04GW05 from the 48–15 ka and the <15 ka

Fig. 4. Schematic chronology of the volcanism of Basse-Terre (see Section 2).

122 A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

phase, respectively, clearly differ from themain tendency of their phase(Fig. 8h and i).

5. Discussion

The 29 K–Ar ages now available for lava flows and domes (Fig. 3),together with previous geological mapping and geochemical correla-tions (Blanc, 1983; Boudon et al., 1988; Komorowski et al., 2005;Samper et al., 2007), allow us to propose a scenario for the successionof the six characteristic effusive stages identified here for the past450 ka (Fig. 9).

5.1. The GDVC area prior to 200 ka

The three most recent volcanic edifices of the Axial Chain (AC),Icaques (IcV), Capesterre (CV) and Sans Toucher (STV) volcanoesdefine the northern limit of the GDVC; their topographic reliefprevented its extension to the north. It has been proposed that theIcV and CVwere emplaced after major flank collapses that affected thesouthern AC, whereas the STV was emplaced later within an erosionaldepression (Samper et al., 2007). Located beneath the western part ofthe GDS, the lava flow dated at 445±6 ka (GU14, Table 2; Figs. 1 and9a; Carlut et al., 2000) can be attributed either to the 435–447 ka oldSTV (Samper et al., 2007) or to a lava flow from the earliest stage ofeffusive activity of the proto-GDS volcano of unknown extent. Ourpresent study cannot resolve this issue. Except for the two domes ofMorne Laffite (04GW14, 314±12 ka, Fig. 1) and Petite Montagne(04GW04, 261±7 ka, Fig. 1),wehave no age data between445 ka and200 ka. TheMorne Laffite and PetiteMontagne ages are in conflict withthe hypothesis that these two lava domes are coeval with the 0.6–

0.4 Ma Mont Caraïbes (Blanc, 1983; Fig. 1), as previously suggestedfrom petrographic similarities (Boudon et al., 1988). The apparenthiatus of 200 kyrmost likely reflects an artifact of insufficient exposureof the basal GDS lavaflows, so that no constraining age can be obtainedfor initial GDS activity. Alternatively this hiatus might reflect a longperiod of explosive activity and lava dome eruptions that producedpyroclastic deposits that are ephemeral in the geologic record intropical setting and have not been considered in the present study.Morne Laffite and PetiteMontagne are evidences that lava domeswereextruded in the Grande Découverte area during the period 445–200 ka, as already observed on its western part with the emplacementof a rhyolite obsidian lavas at 325±8 ka (Blanc, 1983).

5.2. GDS effusive activity from 200 to 34 ka

The diameter of the composite cone of the GDS is 7 to 9 km (Figs. 1and 8). The destruction of the volcano summit during the Pintadeandesitic pumice at 42±2 ka led to the formation of the GrandeDécouverte caldera (Boudon et al., 1988; Komorowski et al., 2005) withamean altitude of about 1000 mabove sea level. Subsequently, eruptiveproducts of the GDS (the Carmichaël and Soufrière phases) wereextruded within or on the borders of the caldera from 42 ka until thepresent time. Products from the four effusive phases older than 42 ka(>150 ka, 150–100 ka, 100–70 ka and 70–48 ka; Fig. 3) are distributedat various altitudes from the bottom of the edifice to the top of thepreserved rims of the caldera. Remnants of the oldest lava flows arerestricted to the northern slopes of the GDS (Fig. 9a). Most lava flowswithin the 150–100 ka interval (01GU55, 04GW22) are found on thenorthwestern slopes although an age of 129±5 ka is obtained for a lavaflow located to the southeast (Figs. 1 and 9b). Lava flows from the 100–70 ka interval (SAM58, 04GW30, 00GU53, GU29, 04GW32b, GU04,01GU61, 98GU32; Figs. 1 and 9c) are found mostly on eastern slopes,and appear beneath the <15 ka volcano of La Citerne, which isassociated with the Soufrière eruptive sequence (Boudon et al., 1988;Komorowski et al., 2005). Whereas most lava flows emplaced between70 and 48 ka are located within the TRMF, one age within this timeinterval was obtained on the southwestern slope of the GDS (04GW31,65±3 ka; Figs. 1 and 9d). The age of 34±4 ka obtained for the NezCassé (01GU63; Figs. 1 and 9e), a remnant part of the rim of theCarmichaël Volcano, fully supports the hypothesis that this volcanowasemplaced after the 42 ka Pintade plinian eruption (Komorowski et al.,2005).

5.3. The TRMF activity: 100 ka–present

The base of the TRMF complex is composed of the 100–70 ka Trois-Rivières thick lava flows that reached the southern coast. We obtainedfor these basal lavaflows two undistinguishable ages of 87±5 and 80±3 ka(00GU53and04GW32b;Table1 andFig. 9c).On topof thenorthernpart of the Trois-Rivières lava flow pile, three discrete edifices and theirrelated lava flows were emplaced during the 100–70 and 70–48 kaphases. From east to west, La Madeleine lava dome, the Gros Fougasscoria cone and lava flow and the Palmiste flows (Fig. 1). La Madeleinelava dome appears to be a complex of several lava domes (Boudon et al.,1988). For basal lavas we obtained two indistinguishable ages of 76±12 ka and 75±14 ka (01GU61 and 98GU32, respectively) and one ageof 58±3 ka (01GU60) on top of the lava dome, which suggest thatbuildingof theMadeleine lava domestartedduring the100–70 kaphaseand continued through the 70–48 ka phase (Fig. 9d). The Gros Fougascomplex, overlaid by the strombolianedificeofGros Fougas, consists of apile of thick lava flows, of which the uppermost has an age of 60±4 ka(04GW37). The Palmiste is a thick andesitic lava flow (67±3 ka,01GU66) thatmay have originated from an eruptive vent located to theeast (Fig. 1). Coeval emplacement of these vents and lava flows duringthe 70–48 ka phase (Fig. 9e) occurred along this E–W-trendinglineament. Three samples located close to the Madeleine lava dome

Fig. 5. K2O versus SiO2 diagram for lavas of each of the seven effusive phases of the Grande Découverte–Soufrière Massif. Fields for rock names and relative K content from (Gill,1981); analyses from this study, Carlut et al. (2000) and Blanc (1983). (a) 2.8–0.25 Ma: grey areas for Northern Basse-Terre volcanic massifs: BC for Basal Complex, SC forSeptentrional Chain, AC for Axial Chain, data from (Samper et al., 2007). Black star for the 445±6 ka Axial Chain lava flow of Cascade Vauchelet (GU14; Carlut et al., 2000; Samperet al., 2007); white stars for edifices of Morne Laffite (04GW04) and Petite Montagne (04GW14) of the Trois-Rivières area (Fig. 1). For the next phases, grey symbols are for GDVClavas, white symbols for TRM lavas. (b) 250–100 ka: black cross for GU08 (205±28 ka; Carlut et al., 2000), hexagons for 150–100 ka lavas. (c) 100–70 ka: triangles. (d) 70–48 ka:squares. (e) 48–15 ka: diamonds. (f) <15 ka: circles. Note that the only basalt sampled (04GW05) belongs to the TRMF area.

123A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

display ages younger than 15 ka (Fig. 9f): two lava flows, La Coulisse(01GU65; 6±6 ka) and Racoon (04GW02; 0±1 ka); and the andesiticlava dome of Morne Liquin 04GW05 (0±20 ka). As further discussedbelow, this is the first recognition of such recent volcanic activity withinthe TRMF. This zone of recent activity extends to areas located east of LaMadeleine lava dome. Only one age from the 48–15 ka phase (04GW26,29±6 ka) has been identified in the TRMF (Fig. 9e) for a lava flowlocated below the 6±6 ka La Coulisse lava flow (01GU65; Fig. 9f).

5.4. Coeval activity from 100 ka to present

The 4-kilometer-long Trois-Rivières lava flows (00GU53 and04GW32b, Fig. 1) were emplaced within the 100–70 kyr interval,when effusive activity was also taking place at the GDS cone itself(Fig. 9c). They both display similar basaltic-andesite compositions,but are slightly more silicic than the coeval lavas of the GDS (Fig. 7c).Later, a fissure vent gave birth to the E–W Madeleine–Le Palmistealignment that is now covering the Trois-Rivières lava flows.

As noted above, multi-elements diagrams (Fig. 8f–i) showrelatively homogeneous compositions with significant differencesbetween samples over the most mobile elements. However, withineach time interval, lava groups from TRMF and GDS cannot bedistinguished. For instance, within the 100–70 ka interval (Fig. 8f),SAM98, GU29, 04GW30, and 04GW32b display strikingly identicalpatterns: the former three are from the GDS and the latter from theTRMF (Fig. 1). Similarly, within the 70–48 ka time interval (Fig. 8g),04GW31 and 04GW37, from the GDS and TRMF, respectively, andwithin the 48–15 ka interval (Fig. 8h), GU05 and 04GW26, from theGDS and TRMF, respectively, show similar patterns. Such undifferen-tiated behavior for coeval lavas originating from distinct centers canbe interpreted as reflecting a single magma source for both GDS and

TRMF. This hypothesis is supported by the homogeneity of major andtrace elements compositions (Figs. 7 and 8) and by similar Sr, Nd andPb isotopic ratios (Samper et al., 2005) displayed by lavas from thesedifferent complexes. Slight changes through time for both volcaniccomplex lavas in the more mobile elements can then be explained interms of crustal contamination within superficial reservoirs, prior toeruption.

5.5. Recent (<10 ka) effusive activity

Our new age determinations provide evidence for a very recentperiod of effusive activity in the Trois-Rivières–La Madeleine area. Ofspecial consideration is the La Coulisse lava flow (01GU65; Figs. 1 and9f), once thought to originate from the LaMadeleine lavadome(Boudonet al., 1988) which has an age of 58±3 ka (01GU60; Fig. 1). Despite theage of 6±6 ka we obtained for La Coulisse lava flow (Table 1) conflictswith the hypothesis that it is a basal flow coming from La Madeleinecomplex, several other arguments support our radiometric age deter-mination. First, another lava flow (04GW26; Fig. 1) located underneathLa Coulisse lava flow, has been dated at 29±6 ka (Fig. 3). Secondly,although it could not be precisely dated by K–Ar (Blanc, 1983), an agerange of 0–15 ka was also obtained previously on La Coulisse lava flowbut from a different sampling location than ours (Blanc, 1983). Thirdly,based on a well-defined elevated paleomagnetic intensity determina-tion, an age within the 0–5 or 8–10 ka intervals was proposed for thislava flow (Carlut and Quidelleur, 2000), in full agreement with ourradiometric determination. Moreover, the presence of well-preservedlevées, under tropical conditions, on both sides of La Coulisse lava flow(01GU65) supports its young age (Fig. 1).We propose the source to be avent located at the base of La Madeleine lava dome or a fissure locatedsouth of the dome. Finally, a verywell-defined age of 0±1 ka (04GW02;

Fig. 6. Major elements evolution of GDVC lava flows as a function of time (in ka).

124 A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

Table 1) obtained on lavas on the lower east side of La Madeleinecomplex (Fig. 1), reinforces our interpretation that magmatic activitytook place in this area evenmore recently during theHolocene (<10 ka),than previously considered (Boudon et al., 1988; Komorowski et al.,2005). Ar radiometric measurements close to the detection limit (Blanc,1983) on the l'Habituée lava flow provide an older bound of 10 ka. Note

Fig. 7. (a) MgO versus time (in ka), (b) chondrite-normalized (Hofmann, 1988) La/Yb ratio vYb ratio versus Mg#.

that the La Coulisse and l'Habituée lava flows are the biggest of this area,with volumes calculated here at 2.5 and 1.5×108 m3, respectively.

Adjacent to the l'Habituée lava flow, the Morne Liquin edifice(04GW05; Fig. 1) yielded a late Pleistocene–Holocene age of 0±20 ka(Table 1 and Fig. 3). Unfortunately, no radiometric constraints couldbe obtained in this study for other lava flows located between the

ersus time (in ka), (c) SiO2 versus Mg#, (d) chondrite-normalized (Hofmann, 1988) La/

Fig. 8. (a) GDS rare earth elements (REE) patterns, chondrite-normalized (Hofmann, 1988). (b) Same as (a) for TRMF. (c) GDS multi-element patterns, normalized to the primitivemantle (Hofmann, 1988). (d) Same as (c) for TRMF. (e) to (i) Trace element patterns for each of the GDVC effusive phases, respectively (e) 250–100 ka, (f) 100–70 ka, (g) 70–48 ka(h) 48–15 ka, (i) <15 ka.

125A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

,

Table 2Previously published Cassignol–Gillot K–Ar ages from the GDVC. GDS: Grande Découverte–Soufrière volcano, STV: Sans–Toucher volcano (see Samper et al., 2007). References:1. Blanc (1983) ; 2. Carlut et al. (2000).

Sample Site Massif Longitude Latitude Age±1σ(ka)

Reference

GU14 Cascade Vauchelet STV/GDS 61°41.18′ 16°02.11′ 445±6 2GU08 Grande Découverte GDS 61°39.93′ 16°03.77′ 205±28 2GU19 Fond Bernard GDS 61°41.47′ 16°03.66′ 143±6 1GU03 Carbet (3e chute) GDS 61°37.27′ 16°02.84′ 129±5 1N1105 Morne Goyavier GDS 61°40.27′ 16°02.14′ 100±10 1GU29 Ravine Longuetau GDS 61°38.61′ 16°02.81′ 83±2 2GU02 Chute du Galion GDS 61°39.51′ 16°02.03′ 79±3 2GU04 Grosse Corde GDS 61°38.04′ 16°02.41′ 77±4 2GU05 Rivière Noire GDS 61°40.32′ 16°02.41′ 47±21 2

126 A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

southern slopes of the Grande Découverte cone and the Trois-Rivières–La Madeleine alignment (Fig. 1), because of a very highatmospheric argon contamination and consequently a low radiogenicargon yield, that prevented precise age determinations. However aswe consider them to be recent, they are attributed to the <15 karange (Fig. 3).

Table 3Major and trace element data.

Sample 98GU14 04GW04 04GW14 98GU08 01GU55 04GW

Massif STV/GDS TRMF TRMF GDS GDS GDS

Age (ka) 445±6 314±12 261±7 205±28 113±2 113±SiO2 (wt.%) 58.89 52.13 57.32 60.49 58.02 56.99Al2O3 17.47 17.77 20.34 17.16 17.43 17.62Fe2O3 8.14 9.60 7.07 7.62 8.61 9.10MnO 0.16 0.18 0.16 0.15 0.15 0.18MgO 2.83 5.86 1.93 2.75 3.22 3.62CaO 6.92 10.14 8.74 6.48 7.32 7.70Na2O 3.46 2.58 3.69 3.19 3.08 2.96K2O 0.81 0.66 0.57 1.04 0.97 0.81TiO2 0.74 0.71 0.66 0.62 0.75 0.78P2O5 0.13 0.11 0.19 0.11 0.12 0.15LOI 0.26 0.21 0.25 0.20 0.15 0.37SUM 99.81 99.96 100.91 99.81 99.82 100.2Ni (ppm) – 18.1 6.482 – – 8.881Cu 61.20 91.98 33.06 54.54 156.1 91.33Zn 63.90 76.11 93.00 59.59 76.93 75.58Cr 5.547 55.25 – – – 6.552V 169.6 300.8 85.88 121.1 202.6 189.1Co 17.44 29.04 12.78 14.94 18.66 21.17Cs 0.276 – – 0.546 0.356 0.316Rb 15.77 7.157 9.976 21.86 21.83 18.45Ba 126 83.1 99.7 159 150 129Th 1.549 1.29 1.112 2.44 2.366 1.665U 0.467 0.448 0.339 0.676 0.68 0.628Nb 2.379 1.178 2.462 2.463 2.336 2.528Ta 0.191 0.094 0.175 0.222 0.211 0.208La 7.553 6.432 7.354 8.26 8.288 7.64Ce 18.49 12.63 13.04 18.94 19.35 14.05Pr 2.673 2.055 2.486 2.527 2.733 2.423Pb 2.1628 3.5831 8.9427 3.3642 3.5654 2.369Nd 12.72 9.24 11.73 11.29 11.99 11.25Sr 210.6 468.0 226.3 205.7 198.4 224.2Sm 3.546 2.253 3.351 2.941 3.271 3.075Zr 83.67 52.76 75.51 87.66 86.16 85.72Hf 2.523 1.581 2.129 2.609 2.511 2.401Eu 1.07 0.80 1.15 0.87 0.95 0.97Gd 4.057 2.369 3.861 3.268 3.647 3.536Tb 0.688 0.393 0.661 0.537 0.611 0.59Dy 4.56 2.50 4.39 3.53 3.97 3.89Ho 0.945 0.531 0.951 0.765 0.842 0.812Er 2.79 1.56 2.80 2.24 2.46 2.42Y 29.02 15.05 29.13 23.69 24.42 23.76Tm 0.425 0.239 0.437 0.355 0.386 0.381Yb 3.004 1.641 2.996 2.466 2.516 2.626Lu 0.468 0.269 0.475 0.394 0.415 0.417

On the GDS western flank, an age of 6±2 ka was obtained from alava flow in the thalweg of Rivière Noire (04GW24; Figs. 1 and 3).Such a young age, together with the morphology of this area (Fig. 1),and the age of 113±2 ka (04GW22; Table 1) obtained for a flowlocated higher in the massif (Fig. 3) suggest that 04GW24 is a valley-filling flow that originated from the Soufrière location, prior to the

22 SAM58 04GW30 00GU53 S04GW32b 98GU29 98GU02

GDS GDS TRMF TRMF GDS GDS

2 94±16 89±2 87±5 80±3 83±2 79±356.88 56.58 54.21 54.97 58.72 53.1017.76 17.78 18.21 17.76 17.49 18.589.12 8.99 9.47 9.70 8.32 9.580.18 0.17 0.17 0.18 0.15 0.173.50 3.39 4.34 4.27 2.98 4.677.48 7.70 9.18 8.95 6.63 9.452.90 3.05 2.91 2.96 3.06 2.930.67 0.84 0.73 0.75 0.94 0.640.68 0.80 0.84 0.90 0.69 0.860.15 0.16 0.15 0.16 0.10 0.130.48 0.43 −0.37 −0.09 0.74 −0.27

5 99.80 99.88 99.84 100.49 99.82 99.84– 9.679 7.636 9.581 7.168 10.4245.31 47.52 41.69 52.04 51.80 78.9070.71 72.31 67.33 73.94 60.05 73.555.179 12.53 14.19 17.82 8.883 41.2160.7 184.5 241 266.7 160.2 241.418.76 20.61 24.17 24.57 17.93 24.680.554 0.226 0.29 0.535 0.381 –

18.34 17.46 13.66 14.37 20.68 11.11118 137 106 120 145 1031.993 1.737 1.61 1.718 2.217 1.6930.563 0.657 0.455 0.561 0.659 0.4362.001 2.645 1.996 2.28 2.181 2.080.175 0.203 0.156 0.181 0.203 0.167.448 8.555 6.883 8.127 8.217 7.80917.49 15.6 16.49 14.53 18.39 18.232.412 2.689 2.35 2.605 2.627 2.554

2 2.5197 2.865 2.2191 2.7943 2.4274 1.92910.71 12.17 10.35 11.92 11.9 11.77218.9 312.8 346.5 352.8 203.6 353.52.914 3.188 2.746 3.122 3.04 3.05776.31 87.86 64.72 76.22 81.58 61.492.291 2.484 1.946 2.185 2.456 1.8320.90 0.98 0.95 1.00 0.97 1.023.217 3.442 3.001 3.447 3.635 3.3470.56 0.574 0.515 0.573 0.600 0.5383.61 3.80 3.42 3.77 3.81 3.420.756 0.808 0.691 0.794 0.803 0.6862.26 2.35 2.04 2.31 2.24 2.0222.03 23.32 20.08 22.40 24.20 20.670.347 0.36 0.307 0.350 0.367 0.3092.348 2.519 2.136 2.380 2.452 2.1010.377 0.402 0.328 0.382 0.414 0.330

127A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

emplacement of the Soufrière lava dome. Additional sampling andanalyses are required to further investigate this hypothesis.

5.6. Geodynamic control

As described elsewhere, the significant differences between La/Smratios measured on lava flows throughout Basse-Terre and the highervalues observed for SC and GDVC massifs have been associated withincreased partial fusion of their magmatic source (Samper et al.,2007). Those authors proposed that such increase was related to theE–Wpropagation of the La Désirade graben during the 1.8–1.15 Ma SCactivity, and of the Marie-Galante active rift during the GDVC activity,superimposed on the en-échelon Dominica–Basse-Terre–Montserratnormal faults system. The latter system controls the southwardvolcanic propagation of Basse-Terre, by accommodating a left-lateralcomponent of slip along the arc (Feuillet et al., 2002).

The three overlapping ages of 67±3, 60±4 and 58±3 ka,obtained here for Le Palmiste, Gros Fougas, and La Madeleine,respectively, (Table 1 and Figs. 1 and 3), constrained the emplacementof this E–Wvolcanic lineament and, hence, the on-land propagation oftheMarie-Galante rift at amean age of about 62±5 ka. In this area, the

98GU04 01GU61 98GU32 01GU66 04GW31 01GU60 S04GW37 9

GDS TRMF TRMF TRMF GDS TRMF TRMF G

77±4 76±12 75±14 67±3 65±3 58±3 60±4 451.97 52.40 52.60 55.69 56.29 55.80 55.76 518.46 18.84 18.68 19.06 17.86 18.75 17.72 19.90 9.79 9.87 8.54 9.06 8.85 9.10 80.18 0.16 0.17 0.16 0.18 0.18 0.18 05.10 4.45 4.51 3.54 3.49 3.82 3.78 39.73 9.28 9.56 8.62 7.71 8.01 7.85 72.54 2.86 2.88 2.77 3.08 2.72 2.97 20.60 0.63 0.62 0.79 0.85 0.77 0.82 00.73 0.91 0.92 0.63 0.81 0.65 0.82 00.09 0.12 0.10 0.11 0.16 0.12 0.16 00.53 0.40 −0.09 −0.10 0.20 0.23 0.79 099.83 99.84 99.82 99.81 99.69 99.90 99.93 99.135 7.744 8.620 – 10.91 – 10.20 −80.28 80.51 53.69 63.84 45.56 65.74 32.57 583.95 72.61 71.81 60.30 75.96 65.81 72.96 636.51 16.55 15.86 – 17.05 – 17.53 5271.9 274.3 264.8 165.0 178.8 174.2 201.6 126.78 26.4 24.67 18.63 21.50 19.99 22.10 10.322 – – 0.286 0.238 0.201 0.225 08.890 9.679 9.581 16.26 17.26 16.13 14.31 187.9 95.4 95.9 114 141 121 139 11.235 1.416 1.424 1.811 1.917 1.872 1.752 10.416 0.39 0.391 0.539 0.660 0.520 0.645 01.372 1.952 1.923 1.695 2.625 1.801 2.446 20.108 0.148 0.147 0.148 0.200 0.161 0.199 05.738 7.148 7.375 6.682 8.639 6.985 8.114 713.66 16.54 16.81 16.16 15.19 16.31 14.36 11.97 2.376 2.443 2.198 2.683 2.25 2.606 22.8222 2.7027 2.7818 2.5238 2.6703 2.6441 2.7143 29.034 11.19 11.11 9.893 12.04 10.01 11.78 1404.8 388.0 389.9 271.6 280.9 256.6 312.1 22.504 2.803 3.072 2.565 3.132 2.661 3.124 349.11 63.86 58.41 67.79 86.89 71.04 66.52 71.523 1.902 1.827 2.061 2.443 2.048 2.398 20.89 0.96 0.95 0.79 0.98 0.83 0.99 02.667 2.980 3.063 2.826 3.446 2.857 3.394 30.450 0.533 0.522 0.469 0.582 0.488 0.568 02.91 3.30 3.34 3.00 3.77 3.06 3.77 30.618 0.693 0.700 0.631 0.786 0.647 0.781 01.76 2.00 1.96 1.88 2.31 1.96 2.31 217.49 19.92 20.43 18.97 23.34 19.35 22.46 20.270 0.301 0.311 0.300 0.354 0.309 0.358 01.791 2.037 2.057 1.922 2.463 2.131 2.446 20.290 0.322 0.339 0.331 0.400 0.338 0.393 0

occurrence of ages younger than 10 ka (Fig. 9f) suggests that this rifthas been active recently and that a renewed propagation might be atriggering factor for future effusive volcanism in this area. In a go-stop-go rifting propagation context, these two clusters of ages suggest arecurrence time of about 50 ka, similar to the ones observed in otherrift-propagation contexts (Manighetti et al., 1998; Lahitte et al., 2003;Audin et al., 2004).

6. Conclusions

The long-lived system of GDVC has been active over a period of atleast 0.2 Myr. Since 2.8 Ma, volcanism has been ongoing on Basse-Terre island, with a north-to-south overall migration (Samper et al.,2007). GDVC is thus the most recent complex of Basse-Terre island. Itexperienced at least six main effusive stages, which have contributedto shape the present-day GDS stratovolcano and the TRMF, in thenorthern and southern parts of the GDVC, respectively. Already builtby at least 205±28 ka (Carlut et al., 2000), the cone of the GDS hascontinued to grow, with a substantial constructional phase betweenabout 130 and 80 ka, so that by 70 ka it had reached most of itspresent-day volume of 16 km3. About 42 ka, the caldera-forming

8GU05 01GU63 04GW26 01GU65 04GW24 04GW02 04GW05

DS GDS TRMF TRMF GDS TRMF TRMF

7±21 34±4 29±6 6±6 6±2 0±1 0±207.24 55.52 54.74 58.69 57.33 58.43 51.327.75 18.14 17.90 17.70 18.00 18.03 19.23.68 9.28 9.63 8.16 9.03 8.40 11.48.16 0.18 0.18 0.16 0.18 0.18 0.21.66 4.03 4.14 3.06 3.46 3.30 4.52.49 7.89 8.95 7.02 7.85 7.75 9.56.94 2.90 2.91 3.05 3.09 3.01 2.71.71 0.65 0.84 0.90 0.76 0.86 0.52.67 0.72 0.91 0.66 0.75 0.69 0.89.12 0.11 0.15 0.10 0.16 0.12 0.14.39 0.39 −0.11 0.31 −0.22 0.09 −0.049.81 99.81 100.23 99.81 100.38 100.84 100.54

13.56 10.35 – 10.07 5.905 5.433.53 63.5 49.39 35.19 35.74 41.03 38.257.86 84.83 73.24 64.89 73.77 62.65 90.86.407 25.54 19.32 – 10.17 4.186 –

51.5 202.4 315.8 150.6 167.8 144.4 364.58.49 24.67 24.27 16.18 19.72 16.73 28.94.214 0.224 0.491 0.525 0.267 0.307 0.3194.02 14.93 15.83 19.08 16.73 17.91 9.9020 119 118 140 127 134 75.7.984 1.690 1.498 2.170 1.601 1.936 1.030.562 0.489 0.549 0.623 0.613 0.611 0.316.120 2.015 2.303 2.122 2.497 2.251 1.800.189 0.184 0.179 0.196 0.199 0.196 0.133.493 7.992 7.889 6.66 7.679 8.354 5.8387.51 19.77 14.42 16.33 14.34 16.41 11.13.456 2.791 2.543 2.279 2.506 2.625 1.86.4685 2.0143 2.599 2.9164 2.3779 3.3814 1.48721.29 13.22 11.64 10.43 11.39 11.89 8.56512.0 236.3 355.4 206.9 232.3 248.0 290.8.035 3.394 3.033 2.807 3.052 3.068 2.3607.03 72.90 76.51 85.64 85.09 86.82 50.41.327 2.245 2.158 2.583 2.365 2.491 1.446.95 1.07 0.97 0.92 0.9 0.93 0.89.264 3.996 3.339 3.235 3.344 3.325 2.686.546 0.692 0.561 0.553 0.573 0.554 0.443.62 4.36 3.60 3.63 3.70 3.64 2.86.745 0.886 0.753 0.746 0.783 0.769 0.605.17 2.62 2.22 2.20 2.32 2.25 1.742.24 26.83 21.66 21.91 22.37 21.84 16.90.346 0.412 0.335 0.352 0.362 0.360 0.265.369 2.706 2.325 2.337 2.476 2.439 1.774.387 0.451 0.364 0.381 0.401 0.397 0.288

Fig. 9. Evolution of the GDVC from 450 ka to present day: main effusive phases. (a) 450–150 ka; (b) 150–100 ka; (c) 100–70 ka, onset of activity south of the GDS with theemplacement of the Trois-Rivières lava flows; (d) 70–48 ka emplacement of the E–W La Madeleine–Le Palmiste alignment; (e) 48–15 ka Carmichaël Volcano phase; (f) <15 kaedifices and lava flows associated with the La Soufrière phase of GDS and the TRMF activity.

128 A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

Pintade plinian eruption occurred (Boudon et al., 1988; Komorowskiet al., 2005). Subsequent effusive construction resumed within thecaldera with the emplacement of the Carmichaël volcano dated hereat 34±4 ka, and was followed by repeated dome growth and partialedifice collapse episodes within the summit area (Komorowski et al.,2005; Boudon et al., 2007; Komorowski et al., 2008b).

The lava flows of the GDVC all display very similar compositions,between andesite and basaltic-andesite. Only one basalt flow, from the

volcanic cone ofMorne Liquin, has been sampled.Major elements showa weak trend toward decreasing level of differentiation with time andLa/Yb ratios suggest that, relative to the earlier Axial Chain magmatism,an increase in the partial fusion of the source could have occurred.

We have provided evidence for coeval effusive activity of the GDSand the TRMF since the onset of the Trois-Rivières lava flows eruptionwithin the 100–70 ka interval, at about 87±2 ka. Trace elementsdisplay strikingly similar signatures, which we interpret as due to a

129A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

commonmagmatic source. The location of eruptive vents of the TRMFhas been controlled by tectonic activity. Morphological investigationsallowed us to suggest that the whole TRMF volcanism was emittedfrom vents distinct from the GDS, most probably a large E–W fissurenetwork linked to the Marie-Galante rift (Feuillet et al., 2002).Tectonic activity probably contributed to extend the entire feedingmagma-plumbing system beneath the GDVC massif. The mean age of62±5 ka, obtained for Le Palmiste–Gros Fougas–La Madeleine E–Walignment suggests that a volcanic phase of the on-land propagationof the Marie-Galante active rift occurred at that time.

The GDS and the TRMF have been active during the Holocene(younger than 10 ka), with ages as young as only a few thousand yearsreported for lavaflows anddomes located east of the LaMadeleine lavadome.

Our new data provide strong temporal constrains to the reconstruc-tion of volcanic activity of the TRMF. Thus these data support theconcept of the Southern Basse-Terre (Guadeloupe) Volcanic-triggeredEventDecision TreeproposedbyKomorowski et al. (2008a) as a startingpoint for the elaboration of an integrated volcanic hazard assessment forSouthern Basse-Terre thatmust include potential activity from both theGDS and the TRMF. Even so, the hazard associated with potentialrenewed effusive volcanism in the TRMF is far lower than the oneassociated with renewed eruptive activity from the GDS, particularlywhen considering the most credible and probable scenarios of partialedifice collapse of the Soufrière domeor a domeeruptionwith explosivephases producing tephra fallout and pyroclastic flows (Komorowskiet al., 2005; Le Friant et al., 2006; Komorowski et al., 2008a; Boudonet al., 2008).

Finally, the occurrence of volcanism in both GDS and TRMF withinthe GDVC is tectonically controlled by the combined influence of theE–W Marie-Galante on-land propagating rift and the NW–SE en-échelon Dominica–Basse-Terre–Montserrat normal faults system.

Acknowledgments

We would like to thank all the staff from the Volcanological andSeismological Observatory of Guadeloupe OVSG of Houëlmont fortheir assistance and discussion. Gilbert Hammouya is also thanked forproviding us with sample SAM58. Precious help from Damien Mollexduring field sampling and mineralogical preparation was greatlyappreciated. Throughout reviews, comments and suggestions by D.Sherrod and P. Layer greatly improved this paper. This study wasfunded by the European Commission EXPLORIS project (EVCR1-CT-2002-40026) and the Centre National de la Recherche Scientifique'sInstitut National des Sciences de l'Univers DyETI programs. This isLaboratoire de Géochronologie Multi-Techniques (LGMT) contribu-tion 82 and IPGP contribution 2552.

References

Audin, L., Quidelleur, X., Coulié, E., Courtillot, V., Gilder, S., Manighetti, I., Gillot, P.Y.,Tapponnier, P., Kidane, T., 2004. Palaeomagnetism and K–Ar and 40Ar/39Ar ages in theAli Sabieh area (Republic of Djibouti and Ethiopia): constraints on the mechanism ofAden ridge propagation into southeastern Afar during the last 10 Myr. Geophys. J. Int.158, 327–345.

Blanc, F., 1983. Corrélations chronologiques et géochimiques des formations volcani-ques du sud de la Basse-Terre de Guadeloupe (Petites Antilles). Début du cyclerécent. 3e cycle Thesis, Univ. Sci. Medic. Grenoble, 171 pp.

Boudon, G., Semet, M.P., Vincent, P.M., 1987. Magma and hydrothermally driven sectorcollapses: the 3100 and 11, 500 y.b.p. eruptions of La Grande Découverte (LaSoufrière) volcano, Guadeloupe, FrenchWest Indies. J. Volcanol. Geotherm. Res. 33,317–323.

Boudon, G., Dagain, J., Semet, M.P. and Westercamp, D., 1988. Carte géologique1/20000e du massif volcanique de la Soufrière (Département de la Guadeloupe,Petites Antilles). BRGM, Orléans.

Boudon, G., Semet, M.P., Vincent, P.M., 1989. The evolution of la Grande Découverte (laSoufrière) volcano, Guadeloupe (F.W. I.). In: Latter, J. (Ed.), Volcanic hazards:assessment and monitoring. IAVCEI Proceedings in Volcanology, Berlin, pp. 86–109.

Boudon, G., Semet, M.P., Vincent, P.M., 1992. Les éruptions à écroulement de flanc sur levolcan de la Grande-Découverte (La Soufrière) de Guadeloupe: implications sur lerisque volcanique. Bull. Soc. Géol. France 163, 159–167.

Boudon, G., Le Friant, A., Komorowski, J.C., Deplus, C., Semet, M.P., 2007. Volcano flankinstability in the Lesser Antilles Arc: diversity of scale, processes, and temporalrecurrence. J. Geophys. Res. 112, B08205.

Boudon, G., Komorowski, J.C., Villemant, B., Semet, M.P., 2008. A new scenario for thelast magmatic eruption of La Soufrière of Guadeloupe (Lesser Antilles) in 1530 A.D.Evidence from stratigraphy radiocarbon dating and magmatic evolution of eruptedproducts. J. Volcanol. Geotherm. Res. 178, 474–490.

Calvert, A.T., Moore, R.B., McGeehin, J.P., da Silva, A.M.R., 2006. Volcanic history and40Ar/39Ar and 14C geochronology of Terceira Island, Azores Portugal. J. Volcanol.Geotherm. Res. 156, 103–115.

Carlut, J., Quidelleur, X., 2000. Absolute paleointensities recorded during the Brunheschron at La Guadeloupe Island. Phys. Earth Planet. Int. 120, 255–269.

Carlut, J., Quidelleur, X., Courtillot, V., Boudon, G., 2000. Paleomagnetic directionsand K/Ar dating of 0–1 Ma lava flows from La Guadeloupe Island (French WestIndies): implications for time averaged field models. J. Geophys. Res. 105,835–849.

Cassignol, C., Gillot, P.-Y., 1982. Range and effectiveness of unspiked potassium–argondating: experimental groundwork and applications. Numerical Dating in Stratig-raphy. InJohn Wiley, New York. 159–179 pp.

Chenet, A.L., Quidelleur, X., Fluteau, F., Courtillot, V., Bajpai, S., 2007. 4K–40Ar dating ofthe Main Deccan large igneous province: further evidence of KTB age and shortduration. Earth Planet. Sci. Lett. 263, 1–15.

Cotten, J., Le Dez, A., Bau, A., Caroff, M., Maury, R.C., Dulski, P., Fourcade, S., Bohn, M.,Brousse, R., 1995. Origin of anomalous rare-earth element and yttrium enrichmentsin subaerially exposed basalts: evidence from French Polynesia. Chem. Geol. 119,115–138.

Davidson, J., De Silva, S., 2000. Composite volcanoes. In: Sigurdsson, H. (Ed.), Encyclopediaof Volcanoes. Academic Press, San Diego, London, pp. 663–681.

Deino, A., Potts, R., 1992. Age-probability spectra for examination of single-crystal 40Ar/39Ardating results: examples from Olorgesailie, Southern Kenya Rift. Quat. Int. 13/14,47–53.

DeMets, C., Jansma, P.E., Mattioli, G.S., Dixon, T.H., Farina, F., Bilham, R., Calais, E., Mann,P., 2000. GPS geodetic constraints on Caribbean–North America plate motion.Geophys. Res. Lett. 27, 437–440.

Dixon, T.H., Farina, F., DeMets, C., Jansma, P., Mann, P., Calais, E., 1998. Relative motionbetween the Caribbean and North American plates and related boundary zonedeformation from a decade of GPS observations. J. Geophys. Res. 103, 15157–15182.

Feuillet, N., Manighetti, I., Tapponnier, P., 2002. Arc parallel extension and loca-lization of volcanic complexes in Guadeloupe, Lesser Antilles. J. Geophys. Res.107, 1–29.

Feuillet, N., Tapponnier, P., Manighetti, I., Villemant, B., King, G., 2004. Differential upliftand tilt of Pleistocene reef platforms and Quaternary slip rate on the Morne-Pitonnormal fault (Guadeloupe, French West Indies). J. Geophys. Res. 109, B02404.

Gill, J.B., 1981. Orogenic Andesites and Plate Tectonics. Springer-Verlag, NewYork. 390 pp.Gillot, P.Y., Cornette, Y., 1986. The Cassignol technique for potassium–argon dating,

precision and accuracy: examples from late Pleistocene to recent volcanics fromsouthern Italy. Chem. Geol. 59, 205–222.

Gillot, P.Y., Cornette, Y., Max, N., Floris, B., 1992. Two reference materials, trachytesMDO-G and ISH-G, for argon dating (K–Ar and 40Ar/39Ar) of Pleistocene andHolocene rocks. Geostand. Newsl. 16, 55–60.

Guérin, G., Samper, A., 2007. Aberrant thermoluminescence dates obtained fromprimary volcanic quartz. Radiat. Meas. 42, 1453-1459.

Harford, C.L., Pringle, M.S., Sparks, R.S.J., Young, S.R., 2002. The volcanic evolution ofMontserrat using 40Ar/39Ar geochronology. Geol. Soc. Lond. Mem. 21, 93–113.

Hofmann, A., 1988. Chemical differentiation of the Earth: the relationship betweenmantle, continental crust, and oceanic crust. Earth Planet. Sci. Lett. 90, 297–314.

Komorowski, J.C., Boudon, G., Semet, M., Beauducel, F., Anténor-Balzac, C., Bazin, S.,Hammouya, G., 2005. Guadeloupe. In: Unit, S.R. (Ed.), Volcanic Hazard Atlas of theLesser Antilles. University of theWest Indies, St Augustine, Trinidad, W.I., pp. 67–104.

Komorowski, J.C., Legendre, Y., Caron, B., Boudon, G., 2008a. Reconstruction and analysisof sub-plinian tephra dispersal during the 1530 A.D. Soufrière (Guadeloupe)eruption: implications for scenario definition and hazards assessment. J. Volcanol.Geotherm. Res. 178, 491–515.

Komorowski, J-C., Boudon, G., Le Friant, A., Legendre, Y., 2008b. A remarquable Holocenerecordofflank-collapses at La Soufrière volcano (Guadeloupe): implications for futurehazards and scenarios. International Association of Volcanology and Chemistry of theEarth's Interior (IAVCEI), General Assembly, Reijkavik, Iceland, Aug. 17–23.

Lahitte, P., Gillot, P.-Y., Kidane, T., Courtillot, V., Bekele, A., 2003. Newage constraints on thetimingof volcanism incentralAfar, in thepresenceof propagating rifts. J. Geophys. Res.108, 2123.

Le Friant, A., Boudon, G., Komorowski, J.C., Heinrich, P., Semet, M.P., 2006. Potential flank-collapse of Soufrière Volcano, Guadeloupe, Lesser Antilles? Numerical simulation andhazards. Nat. Hazards 39, 381–393.

Macdonald, R., Hawkesworth, C.J., Heath, E., 2000. The Lesser Antilles volcanic chain: astudy in arc magmatism. Earth-Sci. Rev. 49, 1–76.

Manighetti, I., Tapponnier, P., Gillot, P.-Y., Jacques, E., Courtillot, V., Armijo, R., Ruegg, J.C.,King, G., 1998. Propagation of rifting along the Arabia Somalia plate boundary intoAfar. J. Geophys. Res. 103, 4347–4974.

Odin, G.S., et al., 1982. Interlaboratory standards for dating purposes. In: Odin, G.S. (Ed.),Numerical Dating in Stratigraphy. John Wiley and Sons, Chichester, pp. 123–150.

Ozawa, A., Tagami, T., Garcia, M.O., 2005. Unspiked K–Ar dating of the Honolulu re-juvenated and Koʻolau shield volcanism on Oʻahu Hawaiʻi. Earth Planet. Sci. Lett.232, 1–11.

Quidelleur, X., Gillot, P.Y., Soler, V., Lefèvre, J.C., 2001. K/Ar dating extended into the lastmillennium: application to the youngest effusive episode of the Teide volcano(Spain). Geophys. Res. Lett. 28, 3067–3070.

130 A. Samper et al. / Journal of Volcanology and Geothermal Research 187 (2009) 117–130

Robertson, R.E.A., 1995. An assessment of the risk from future eruptions of the SoufriereVolcano of St-Vincent West-Indies. Nat. Hazards 11, 163–191.

Rouchon, V., Gillot, P.Y., Quidelleur, X., Chiesa, S., Floris, B., 2008. Temporal evolution of theRoccamonfina volcanic complex (Pleistocene), Central Italy. J. Volcanol. Geotherm.Res. 177, 500–514.

Samper, A., Chauvel, C. and Quidelleur, X., 2005. Geochemical and isotopic study of K–Ardated volcanics from Basse-Terre, Guadeloupe, Lesser Antilles Arc. In: Geophys.Res. Abs. (Editor), EGU 2005 Meeting, Vienna.

Samper, A., Quidelleur, X., Lahitte, P., Mollex, D., 2007. Timing of effusive volcanism andcollapse events within an oceanic arc island: Basse-Terre, Guadeloupe archipelago(Lesser Antilles Arc). Earth Planet. Sci. Lett. 258, 175–191.

Samper, A., Quidelleur, X., Boudon, G., Le Friant, A., Komorowski, J.C., 2008. Radiometricdating of three large volume flank collapses in the Lesser Antilles Arc. J. Volcanol.Geotherm. Res. 176, 485–492.

Semet, M., Vatin-Pérignon, N., Vincent, P.-M., Joron, J.-L., 1981. L'éruption volcanique duXVIème siècle de la Soufrière de Guadeloupe, mélanges de magmas et dynamismeéruptif. Bull. PIRPSEV, CNRS-INAG, Paris, France, vol. 60, pp. 1–63.

Sherrod, D.R., Murai, T., Tagami, T., 2007. New K–Ar ages for calculating end-of-shieldextrusion ratesatWestMauivolcano,Hawaiian islandchain. Bull. Volcanol. 69, 627–642.

Sparks, R.S.J., 2003. Forecasting volcanic eruptions. Earth Planet. Sci. Lett. 201, 1–15.Steiger, R.H., Jäger, E., 1977. Subcommission on geochronology: convention on the use

of decay constants in geo and cosmochronology. Earth Planet. Sci. Lett. 36, 359–362.Wadge, G., 1984. Comparison of volcanic production rates and subduction rates in the

Lesser Antilles and Central America. Geology 12, 555–558.White, S.M., Crisp, J.A., Spera, F.J., 2006. Long-term volumetric eruption rates and

magma budgets. Geochem. Geophys. Geosyst. 7, Q03010.