- Email: [email protected]
A new 3D fault model of the Bouillante geothermal province combining onshore and offshore structural knowledge (French West Indies)
Tectonophysics 526–529 (2012) 185–195
Contents lists available at SciVerse ScienceDirect
Tectonophysics journal homepage: www.elsevier.com/locate/tecto
A new 3D fault model of the Bouillante geothermal province combining onshore and offshore structural knowledge (French West Indies) P. Calcagno ⁎, V. Bouchot, I. Thinon, B. Bourgine BRGM, 3 avenue C. Guillemin, 45060 Orléans Cedex 2, France
a r t i c l e
i n f o
Article history: Received 28 February 2011 Received in revised form 12 August 2011 Accepted 16 August 2011 Available online 30 August 2011 Keywords: 3D modelling Fault modelling Onshore and offshore combination Geothermal energy Geothermal exploration Bouillante
a b s t r a c t The Bouillante area hosts geothermal resources located in a complex structural area (Guadeloupe Island, French West Indies). On one hand, faults observed on the field mainly elongate along the E–W direction. On the other hand, offshore structures interpreted from marine seismic lines shows a larger range of directions. A coherent 3D interpretation is proposed through a fault model combining onshore and offshore structural knowledge in a zone crossing the island coastline. The fault network constructed reveals a hierarchy in the family of structures and highlights the prevalence of the NNW–SSE direction, associated with secondary NE–SW-trending structures, and the E–W direction. On a geographical point of view, the modelled faults are gathered in 3 clusters. Data available to build the 3D fault model are sometimes sparse, especially inland because of intense vegetation cover. Consequently, not only the results and impacts of the 3D fault model are discussed but also its limitations as well as its possible evolution. © 2011 Elsevier B.V. All rights reserved.
1. Introduction
2. Regional structural setting
The Bouillante geothermal province area is located in the French West Indies, on the West of the Guadeloupe Basse-Terre Island (Fig. 1). That zone is known for its high-temperature geothermal field (e.g. Bouchot et al., 2010; Fabriol et al., 2005; Sanjuan and Traineau, 2008) which produce electricity at the 15 MWe installed power plant in Bouillante (Figs. 1 and 2). Our study is dedicated to better understand the structural pattern of the Bouillante geothermal province. This is a prerequisite for a better knowledge of the Bouillante geothermal field understanding. At the scale of the geothermal province, onshore and offshore data are required to interpret in a coherent way the faults in the area because of its location both inland and at sea. Previous works aimed at describing the structural geology on land (e.g. Bouchot et al., 2010; Boudon et al., 2007; Feuillet et al., 2002; Mathieu, 2010; Samper et al., 2007; Traineau et al., 1997) and at sea (e.g. Bouysse et al., 1983, 1988; Feuillet, 2000; Thinon et al., 2010). We propose to merge data and knowledge from field work and from marine exploration to produce a model in 3 dimensions representing the faults down to 2 km depth at the scale of the Bouillante geothermal province that is about 250 km 2.
The Bouillante geothermal province is a regional key geodynamic area where the major tectonic and volcanic structures of the inner arc of the Lesser Antilles join (e.g. Bouysse et al., 1988; Feuillet, 2000; Feuillet et al., 2001, 2002; Thinon et al., 2010). As far as can be determined from our present state of knowledge, the Bouillante geothermal field is located at the junction of two regional fault systems (Fig. 1a):
⁎ Corresponding author. Tel.: + 33 238643054. E-mail address: [email protected] (P. Calcagno). 0040-1951/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.tecto.2011.08.012
(i) The NW–SE Montserrat–Bouillante (MB) fault system was observed from the northern of the Directeur volcano seamount to the foot of the NW Basse-Terre slope (Bouysse et al., 1988). It is composed of several ‘en echelon’ oblique normal faults that accommodate a component of sinistral motion (Duperret, 1991; Feuillet et al., 2002; Kenedi et al., 2010). (ii) The ESE–WNW to E–W Bouillante–Capesterre (BC) fault system which could belongs to the E–W Marie–Galante graben system crosscuts the volcanic island of Basse-Terre from Bouillante town (western side) to Capesterre town (eastern side) with a normal component (Feuillet et al., 2002). This hypothetic south dipping fault zone that separates the Axial Chain from the Bouillante–Sans–Toucher complex (Figs. 1b and 2) is mainly interpreted from topographic evidences (e.g. continuous step offsetting the topography), from few outcrops on the west coast (Feuillet et al., 2002), and is partly suggested by gas anomalies close to Capesterre town (Baubron, 1990). However, according to Boudon et al. (2007), these topographic steps represent the scare of two collapse structures (northern part of the
186
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
Fig. 1. (a) Geodynamic context of the West Indies. NAM: North American plate. SAM: South American plate. The grey area covers the fore arc developed on ~ 250 km wide in E–W between the subduction zone and the western edge of the volcanic arc of the Lesser Antilles. From North to South, the motion ranges from sinistral extensional shear to dextral oblique thrusting (Feuillet et al., 2002). (b) Regional structural setting of the Guadeloupe Island. Thick square: location of the 3D fault model including the Bouillante field. BCF: Bouillante–Capesterre fault system. The trajectory of this hypothetical fault is extracted from the geological, morphological and structural map of southern Basse-Terre realised by Feuillet et al. (2002). It mainly based on alignments of steps offsetting the topography. a, modified from Feuillet et al., 2002; b, modified from Feuillet et al., 2001 and Thinon et al., 2010.
Beaugendre and Vieux-Habitants horseshoe-shaped depression) rather than a fault plan. Even if occurrence of debris avalanche on the mouth of the structure argues in favour of flank collapses, the origin of these collapses — volcano-hydrothermal, or structural (normal or strike-slip faulting) — are highly debated (Boudon et al., 2007; Mathieu et al., 2011; Samper et al., 2007).
From Feuillet et al. (2001), the MB fault system would join the BC fault system, thus drawing at the scale of the northern Caribbean arc a sinistral extensional horsetail. From these authors, the western end of the BC fault is correlated to the Machette–Pointe à Sel fault observed on the coast at north of the Bouillante Bay. Thinon et al. (2010) propose that the BC fault system extends into the Bouillante Bay and cuts the whole shelf. A third fault system, described by Thinon et al. (2010), is a N160°E-trending escarpment which limits the shelf of the Bouillante sector. This escarpment is interpreted as a major sinistral strike–slip system, playing the role of transfer zone between the MB and Les Saintes (LS) fault systems. This discontinuity belongs to the western limit of a zone of sinistral extensional shear, developed on 250 km wide in E–W between the subduction zone and the western edge of the volcanic arc of the Lesser Antilles (Feuillet et al., 2002; Fig. 1a). 3. Onshore structural knowledge
Fig. 2. Onshore structural knowledge. A selection of structural data and interpretations available onshore. The geothermal boreholes BO5 and BO6, belonging to the power plant in Bouillante, give structural information at depth.
The Bouillante geothermal field is contained within an andesitic volcanic substratum largely attributed to sub-products of the axial Pitons de Bouillante chain cropping out along the axis of Basse-Terre Island. Recent small volcanic centres (0.8 and 0.2 Ma) of the ‘Bouillante Chain complex’ defined by Gadalia et al. (1988) lie on this volcanic substratum. Immediately north of Bouillante City, several volcanic centres and related deposit, i.e. lava flow or pyroclastic deposits, have been dated between 840,000 y, e.g. strombolian cone of Pointe à Sel, and 490,000 y, e.g. Desmarais centre, from K/Ar dating (Gadalia et al., 1988 and recent BRGM unpublished data). Aligned along a NNW–SSE trending band of some 20 km by 4 km between Anse Colas (North) and the Basse-Terre City (South), the ‘Bouillante Chain complex’ volcanism, and its offshore and onshore volcanic centres seems controlled by the offshore NNW–SSE MB fault zone. These volcanic rocks often emplaced under hydro-magmatic condition (e.g. hyaloclastite) and have the characteristics of a weakly potassic tholeiitic series that locally evolved through fractional crystallisation (Gadalia et al., 1988). The studied area is located on the West side of the volcanic BasseTerre Island where the prevailing climate is wet tropical. This specific geographic location constrains the acquisition of structural data.
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
Exposures suitable for structural measurements are mainly found only on the seaside where coastal erosion has stripped away the soil (Fig. 2). Since the 1970s when geothermal exploration began around Bouillante, several field works have succeeded in acquiring sets of structural data on outcrops (e.g. Bouchot et al., 2010; Sanjuan et al., 2004; Traineau et al., 1997). From a recent synthesis of existing structural interpretation (Bouchot et al., 2008) and acquisition of new structural measurements by the authors (Fig. 2), it appears that most of the measured faults have a trend E–W to ESE–WNW, with a dominant normal throw showed by the evidence of normal offset of volcanic layer on each side of the fault surface with striae and drag fold in few stations. Moreover, a very limited second family, consisting of only two faults visible onshore, trends NE–SW with a strong dextral strike–slip component, according to tectoglyphs such as striae and tension gashes. In detail, onshore structural analysis provides essential information for our understanding, from south to north: - On the Bouillante geothermal field, fracturing is a 2.5 km wide network of seven faults. These faults strike E–W to ESE–WNW and dip steeply (~ 70–80°) to the north or the south. They accommodated a dominant normal slip according to some drag folds (Fig. 3), rare fault striate on fault plan, and the shift of the volcanic levels on both sides of faults. Spaced from 200 to 700 m, these normal faults delimit horst and graben compartments and in particular the Bouillante graben located on the southern part of the Bouillante Bay (Bouchot et al., 2010). - In the northern part of the Bouillante Bay, an E–W-striking fault zone is cut by a 15 m-large, steeply dipping, NE–SW-striking fault zone of Pointe Marsolle, which is characterised by an atypical anastomosing pattern. This shear zone shows a dextral kinematics according to en echelon networks of vertical tension gashes, partly filled by quartz and/or calcite. The structure continues offshore along a NE–SW ridge marked in the bathymetry, and onshore, marks a sharp NE–SW trending contrast in terms of resistivity anomaly showed by a MELOS campaign (Fabriol, 2001; Fig. 2). - Further north around Malendure village, although less intense than in the Bouillante area, deformation is also characterised by several E–W to ESE–WNW-striking normal faults and one NE– SW-striking fault showing a dextral–normal throw. Along this
187
fault, a N65°E-striking, 85°S dipping fault plan shows brittle striaes with a pitch of 50°W, indicating a dextral–normal oblique motion. Some limited fractures trending NW–SE, associated with reddish hydrothermal alteration, have also been identified to the north of Malendure village, close to Anse Colas. As soon as one moves away from the coast, the volcanic island is covered by dense tropical forest on very steep relief. Given a rainfall that is 10 times higher than at the coast, this forest zone is the site of intense supergene alteration that impedes ground observation, except where ravines generated by torrents have been cut into the weathering profile. Nevertheless, the hilly areas where topographic regular structures are sharp can provide indirect information on fault traces (Feuillet et al., 2002; Fig. 2) and on flank collapse (Boudon et al., 2007; Fig. 2). But supplementary data are needed to confirm the origin of the topographic anomalies and collapses. Similarly abnormal geophysical signatures (Fabriol et al., 2005) provide indirect data likely to indicate the presence of a fault at depth, e.g. NE–SW-striking Pointe Marsolle fault mark by a sharp contact between different ground resistivities (Fig. 2). Overall, the combination of direct and indirect structural data has enabled an interpretative onshore structural pattern to be considered, albeit poorly constrained at depth given the small amount of information available (borehole logs, electrical surveys, extrapolation of observed dips). Finally, it is clear that the patchy distribution of direct data is a real challenge to build an interpretative structural pattern of the studied area. 4. Offshore structural knowledge Shallow water multibeam echosounder (MBES) bathymetric, very high resolution (VHR) reflection seismic (Sparker source, 6-channel) and high resolution (HR) magnetic surveys were conducted in 1998 and 2003 by the BRGM (Bathybou98 and Geoberyx03 cruises) on the west Basse-Terre Island shelf, centred on the Bouillante Bay sector (Guennoc et al., 2001; Thinon et al., 2010). The bathymetry (Fig. 4a) shows the present-day morphology of the shelf, its extension, the dimensions and forms of the bays, the occurrence of escarpments and the prolongation of promontories. Furthermore, the morphology underlines some main directions, as
Fig. 3. Picture of the N95°E-striking, dipping 70°S normal Machette–Pointe à Sel fault located in the Anse Machette (see Fig. 2 for location). This normal fault put in tectonic contact a white layered deposit characterised by alternating deposits of coeruptive pumices fall and pyroclastic flows (north block) and a brown slightly reworked coarse pyroclastic deposit (south block). The dip-slip of the fault is estimated around 20 m from the evidence of brown pyroclastic deposit above the white layered in the north block. The layering (S0) is affected by drag fold along the main normal fault. The red arrow locates the S0-N130°E-17°S measurement.
188
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
Fig. 4. Bathymetry and acoustic basement maps. Isobaths in metres. (a) Bathymetry of the NW shelf of Basse-Terre. Compilation of MBES data: BRGM Bathybou98 survey, 1998; DTM, mesh 5 m and 25 m (Guennoc et al., 2001); IPG Aguadomar survey, 1998–1999, DTM 250 m (Deplus et al., 2001). Stars: Location of faults shifting the seabed. (b) Isobaths map of top of acoustic basement of the NW shelf of Basse-Terre and main faults derived from the seismic interpretations of the Geoberyx03 survey.
for example the E–W-trending Marsolle escarpment in the northern part of the Bouillante Bay (see Fig. 4b for location). The depth of the seabed is a reference data for the 3D model, as much as the topography. The Geoberyx03 cruise has provided a dense network — about 180 km — of the VHR reflection seismic images of the sedimentary cover (geometry, thickness, seismic facies, etc.) and the volcanic basement in second Two Way Travel Time (TWTT, Fig. 5). The acoustic basement is interpreted as being the volcanic basement of BasseTerre Island. These data show the tectonic structures and the relationships between the acoustic basement and the sedimentary cover. According to the shifts of the seismic reflectors along faults, vertical throws of faults are a few metres to tens of metres at least. The hori-
zontal displacement is not deductible and the faults within the acoustic basement are not visible. The processing of the HR magnetic survey of the Geoberyx03 cruise (e.g. Truffert et al., 2004) has highlighted a strong circular magnetic anomaly off the Bouillante Chain on the shelf between the South Bouillante Bay and the Vieux Habitants town (Fig. 4b). This anomaly has been interpreted by Thinon et al. (2010) as the marker of a part of a volcanic seamount located over the shelf in the axis of Upper Miocene to actual submarine volcanoes (Feuillet et al., 2001). The association of the geophysical data provides a good image of the geometry of the structure. However, no geological data are available to constrain the age and the kind of formations, nor structural measurements to quantify true dips.
Fig. 5. One of the Geoberyx03 VHR seismic profiles through the NW shelf offshore Bouillante (Thinon et al., 2010). It shows the thickness of the sedimentary cover (UP, UC, RB, UB), the morphology of the acoustic basement (US), the apparent dip and the vertical throw of the major faults, the N160° basement high and the geometry of basins (BSP) on the internal shelf. RB: surficial formation. Various seismic units are defined and noted by the symbol U. Box: location of the marine seismic section (West Guadeloupe offshore).
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
The interpretation of these surveys has helped to highlight several major structural features (Thinon et al., 2010): i. The sedimentary cover is generally thin, except on the outer shelf where thick sedimentary unit deposition induced an important widening of the shelf. The thickness of the sedimentary cover is between 0.02 and 0.04 s twtt. It reaches 0.085 s twtt thick (~80 m) in the small basins, i.e. the BSP basin (Fig. 5). The sedimentary cover thickens to 0.18 s twtt on the external shelf to the west of the basement high (Fig. 5). The bathymetry map doesn't reflect the morphology of the volcanic basement (Fig. 4a). ii. The main directions of structures, observed on the shelf, are by increasing importance NNW–SSE, E–W, NW–SE and NE–SW. iii. The western edge of the shelf is a NNW–SSE-trending escarpment, may be a major sinistral strike–slip system, playing the role of transfer zone between the NW–SE MB and LS fault systems. The N160°E-trending escarpment is composed by segments shifted southward from West to East. This structure limits also an asymmetric basement high, faulted and shifted. iv. The E–W-trending Marsolle escarpment, located in the Bouillante area (Fig. 4b), could be the offshore prolongation of the E–W-trending BC fault system. Indeed, an E–W major structure extends into the Bouillante Bay from the coast to the shelfbreak crossing the NNW–SSE- and NW–SE-striking faults. It consists of two segments shifted southward from east to west which coincides with the northern bound of the Bouillante Bay. v. The NW–SE-striking faults are mainly observed on the insular slope and on the external shelf off the Bouillante Bay, with a “staircase”-type distribution. vi. A major basement high (Figs. 4b and 5) forms a 500 m wide ridge extending for 1600 m in a N160°E direction. This high reaches a sea-bottom depth of about 70 m to the north and 75 m to the south on both sides of the E–W-trending Marsolle escarpment. It is bounded to the west by the NNW–SSE-trending escarpment which underlies the shelf-break of the acoustic basement. vii. A volcano edifice extends over the shelf on south of Bouillante Bay, according to the observed large magnetic anomaly. Based from only indirect observations, the relationships between the structures of the shelf are not very well defined. 5. Data and methodology A 15 km × 16 km zone (see Fig. 1b for location) is modelled down to 2 km below sea level in the local “Guadeloupe, St. Anne–UTM20– IGN”, WGS84 projection system. In that zone, faults are interpreted by taking into account both offshore and onshore geological knowledge. Onshore, data consist in strike and dip measurements of fault measured during field work. Due to outcrops accessibility, these data are located on the coast. Inland, some structures are interpreted from previous works (e.g. Feuillet et al., 2002). Offshore, location and dip of the faults are derived from the interpretation of 52 Geoberyx03 seismic profiles (Thinon et al., 2010). The magnetic lineaments, interpreted as markers such as faults or dykes are clues for the existing orientations of structures on the shelf (Finn and Morgan, 2002; Henkel and Guzman, 1977). The information on the age and nature of geological formations are derived from geological surveys at land. Preliminary to the 3D model, a Digital Elevation Model (DEM) was created to combine onshore and offshore data in a coherent 50 m resolution 2D grid covering the whole modelled area. The GDM1 software was used for this work (Bourgine et al., 2008). Onshore data are provided by the IGN (Institut Géographique National, France) on a 50 m grid in the local UTM coordinate system. Offshore, two bathymetric surveys are
1 GDM is a commercial software developed by BRGM. For further information visit: http://gdm.brgm.fr/?lang=en.
189
used: Deep water multibeam echosounder of the AGUADOMAR cruise (1998–1999, IPG, Deplus et al., 2001) and the shallow water multibeam echosounder of the Bathybou98 cruise (1998, BRGM, Guennoc et al., 2001). The AGUADOMAR data are available on a 100 m grid in latitude–longitude coordinates. The Bathybou98 data are available on a 50 m grid in the local UTM coordinate system. Data are kriged with a linear variogram to fill the gaps of few hundreds of metres between onshore and offshore. The full grid is interpolated using a bilinear interpolation to get a grid in the WGS84 coordinate system used in this study. The resulting DEM (Fig. 6) limits the top of the structural 3D model. The way for interpreting the faults is driven by the size of the zone. In particular, only relevant faults at the scale of the model were taken into account. Some assumptions regarding their geometry were also made. For instance, the Montserrat–Bouillante–Les Saintes (MBLS) fault system was interpreted as a continuous fault crossing the whole modelled area and extending to the bottom of the box. To note, the geometry of the MBLS fault system is extrapolated on the Southern and the Northern parts of the zone due to the lack of data: From North Malendure to the MB fault system observed in the abyssal plain at the North of Directeur volcano seamount, the direction and localisation of the N160°E-striking fault system is hypothetical and based on the present-day shelf-break (bathymetric data). The southern part of the MBLS fault system was also extrapolated to the south of Coreil. It doesn't reflect the morphology of the basement, the shelf is much closer. The horizontal extension of the other faults is limited by the location of their data. At depth, no data is available to quantify the extension of the faults. They were limited by taking into account their horizontal extension, i.e. the geometry of finite faults is assumed to be limited by half a sphere centred on the middle of the horizontal trace of the fault. A fault network was built where information was available to determine the relation between contiguous faults. Where such information is not available, modelled faults cross each other. In addition, some of the structures are not directly observed but deduced from the observed faults. As an example, an offset or a termination of an observed fault drives the interpretation of a nonobserved fault. To construct a coherent interpretation, the structures observed onshore — mainly E–W oriented — were as much as possible used as a guide to interpret and to connect the offshore structures oriented in the same direction. However, our methodology does not simply consist in trying to connect faults, separately interpreted or observed, at sea and inland. Offshore and onshore information are used together to build a coherent structural 3D interpretation covering the whole area. To achieve this goal, the geometry of a fault is computed using a set of data merging marine and inland locations of this fault (e.g. the Bouillante fault on Fig. 7).Data are interpolated using the potential field method developed in BRGM (Calcagno et al., 2008; Lajaunie et al., 1997) and implemented in the 3D GeoModeller2 software. 6. 3D fault model 6.1. Faults analysis Statistical analyses of the faults coming from the 3D regional fault model bring relevant information on the distribution of onshore and offshore deformation. The pole density contour plotted for the offshore faults (Fig. 8a) shows that the main set of fault is striking ESE–WNW, and mainly dipping SSW. Two others sets are striking NNW–SSE and NE–SW with opposite dipping. In contrast, the pole density contour plot of the onshore faults (Fig. 8b) indicates that the main set of faults is striking from E–W to ESE–WNW and dipping
2 3D GeoModeller is a commercial software developed by BRGM and Intrepid Geophysics. For further information visit: http://www.geomodeller.com.
190
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
Fig. 6. DEM reconstructed from onshore data (IGN) and offshore data (Bathybou98 and AGUADOMAR surveys). The coastline (dark green line) separates onshore and offshore domains. Size of the area: 15 km × 16 km, vertical exaggeration × 2, elevations ranging from − 1120 m to + 1150 m from sea level. Corner coordinates (W–E; S–N) in metres and in the local “Guadeloupe, St. Anne–UTM20–IGN”, WGS84 projection system. View from the S–SW.
north or south as a horst and graben pattern. Taking into account the whole set of onshore and offshore faults, it appears that at the model scale, faults are mainly striking from ESE–WNW to NNW–SSW and mainly dipping SSW to WSW (Fig. 8c and d). The final 3D interpretation highlights coherent structures onshore and offshore (Fig. 9). The modelled faults can be classified in two main families: The MBLS fault system and the E–W- to ESE–WNWstriking fault system. Both are detailed below. 1. The MBLS fault system: the case of the NNW–SSE-trending section around Bouillante The MBLS fault system is considered as a sinistral strike slip fault zone recognised over 300 km long between St Kitts and Les Saintes north to south (Fig. 1b). We combine the NW–SE MB fault system between Montserrat and Directeur Volcano seamount (Duperret, 1991; Feuillet, 2000), the NNW–SSE MB fault system between Malendure and Coreil and
Fig. 7. Example of onshore and offshore knowledge combination: The Bouillante fault (yellow surface) is constructed from onshore field observation (dark blue point), onshore interpretation (red points) and offshore seismic profiles interpretation (E–Wtrending Marsolle escarpment, light blue points). View from WSW. The coastline (white crosses) separates onshore and offshore domains. See Fig. 1b for modelled area location.
the NW–SE LS fault system (Deplus et al., 2001) to model the MBLS fault system. This fault system is generally oriented NW– SE, but draws/forms a NNW–SSE-trending virgation along the Northern part of Basse-Terre volcanic island (Fig. 9). The latter would be accompanied by a reversal of dip. The virgation could appear as the result of shearing deformation moulded on the coastal shelf of the Basse-Terre Island. It could be induced by the association of the occurrence of the old volcanic complex of the North Basse-Terre, the collapse of the southern part of Basse-Terre linked to the Marie–Galante graben system. The NNW–SSW-trending virgation is characterised by a sub-continuous 30 km long fault zone which strikes N160°E and dips west. In contrast, on both end of the virgation, the NW–SE-striking corridors consist of east-dipping ESE–WNW-striking oblique fault. This “en echelon” oblique normal faults network is interpreted as an indicator of a sinistral transtension (Fig. 1) and the NNW–SSE section of the MBLS fault zone as a N160°E sinistral strike slip fault (Thinon et al., 2010). According to our fault model (Fig. 9), the NNW–SSE-trending section consists of two main parallel N160°E-striking ‘en relais’ faults,
Fig. 8. Statistical analysis of the faults. (a) Pole density contour plot of the offshore faults. (b) Pole density contour plot of the onshore faults. (c) Stereonet plot of poles to onshore (+) and offshore (x) fault planes. (d) Pole density contour plot of the onshore and offshore faults.
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
191
Fig. 9. The whole 3D fault interpretation combines onshore and offshore knowledge. The coastline (white crosses) separates onshore and offshore domains. Two main families of faults are modelled: The MBLS fault system (green and blue faults); The E–W to ESE–WNW-striking fault system (yellow faults). See Fig. 1b for modelled area location. View from SE.
developed along the shelf break. They show an opposite dip (Fig. 10). Separated by 150 m, these two faults limit the development of a 5 km long, NNE–SSW-trending basement high, indicating a normal component (Figs. 4b and 5). We interpret this basement high as a horst structure related to the sinistral strike– slip movement along the N160°E-striking faults system. In addition, a network of several parallel faults, well developed offshore, striking NE–SW, and 3 km long is linked to the offshore N160°E-striking faults. The onshore expression of these faults is very limited (Figs. 9 and 10). Identified only twice on the coast (Bouillante and Malendure outcrops), the NE–SW-striking faults show a dextral movement with a weak normal component. We interpret these short faults as a dextral conjugate network, grafted onto the main N160°E-striking sinistral fault, as a pinnate pattern. In the 3D fault model, some NW–SE-striking faults were modelled mainly on the insular slope (2 dark blue faults) and on the external shelf (2 blue faults). They are located offshore the Bouillante Bay (Figs. 9 and 10). They are limited and sometimes shifted by the E– W and NE–SW faults. The NW–SE faults are also present at North of Malendure, near Pointe Noire town (see Malendure cluster below). 2. The E–W- to ESE–WNW-striking fault system According to our fault model (Fig. 9), the main fracturing strikes E–
W to ESE–WNW, onshore and on the shelf. However, this deformation is very heterogeneous and is divided into 3 clusters of faults: (i) the Malendure cluster on the Northern part of the zone, (ii) the Bouillante cluster on the geothermal field, and (iii) the Coreil cluster on the Southern part of the zone.
Fig. 10. W–E vertical section in the 3D fault model. No vertical exaggeration. See Fig. 9 for location.
192
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
i. Malendure cluster As stated by the regional structural pattern, this fracturing cluster is located at the junction between the NW–SE section and the NNW–SSE section of the MB fault system, in the Malendure area (Figs. 1b and 2). On the shelf, the vertical offsets of two NW–SE-striking faults are low. Onshore, these faults are correlated to discrete fracture zones, controlling reddish hydrothermal alteration. Concerning the E–W- to ESE–WNW-striking normal faults, they do not seem connected to a regional E–W fault but they are well developed between the NW–SE and the NNW–SSE sections of the MBLS fault system. At north of the Malendure cluster, the NW–SE faults observed on coast have been correlated to the faults observed on the seismic profiles. The low density of marine data prevents from specifying the tectonic network and the relationships between the NW–SE Pointe–Noire faults and the EW faults of Malendure. ii. Bouillante cluster The ~EW-striking, steeply dipping, normal faults network forms a horst and graben pattern, and consists in main conduits for geothermal fluids in depth. However, on the shelf, i.e. approaching the main N160°E-striking fault system, the E–W fault trajectories are discontinuous because these faults are stopped on continuous transverse NE–SW-striking faults (Fig. 11). This geometry defines structural blocks composed of E–W normal faults bounded by dextral NE–SW-striking faults. This logic of block suggests that the continuous strike slip plays the role of transfer fault during the development of the E–W-striking normal faults under a ~N–S regional extension. On our fault model (Fig. 11), evidence of intense deformation, focused on the Bouillante cluster, is a strong argument to defend the existence of the major fault zone of BC fault system (Feuillet et al., 2002). The bursting of the E–W fracturing in this area (over 2.5 km in N–S) suggest the evidence of a horsetail end of this regional fault, near the function with the regional NNW–SSE-striking MBLS fault system (Bouchot et al., 2010). Moreover, unlike previous works defending the idea of Machette–Pointe à Sel fault as the major member of the regional E–W fault zone (Feuillet et al., 2002), our fault model shows that this Machette–Pointe à Sel fault (Fig. 11), measured on outcrop (Figs. 2 and 3), has a very limited continuity offshore. By contrast, the steeply south dipping E–W-striking Bouillante fault (Figs. 9, 11, and 12) is the only fault which i) cuts across the shelf, ii) shows a strong offset of the basement at depth (more than 25 m) and the E–W-trending Marsolle escarpment at the surface in the Northern part of the Bouillante Bay, as a result of the normal movement along the fault, and iii) cuts and
Fig. 12. S–N vertical section in the 3D fault model. No vertical exaggeration. See Fig. 9 for location. The interpretation is extrapolated at depth to give a predominant behaviour to the E–W-striking Bouillante fault.
offsets the major N160°E-striking faults at the level of a Bouillante Valley canyon (Fig. 4a, Thinon et al., 2010). This latter characteristic is not represented at the scale of the regional 3D fault model. Consequently, the Bouillante fault appears as a main member of the regional E–W fault, on which are associated the other E–W-striking normal faults whose dips are opposed (Figs. 11 and 12). On the Bouillante cluster, seismic images show two faults which shift the seabed off Bouillante (Fig. 4a), suggesting that the tectonic activity is recent. Finally, we interpret the fracturing cluster of Bouillante as the junction zone between the major E–W fault zone through Bouillante town, and the major sinistral–(normal) N160°E
Fig. 11. Zoom of the 3D fault model on the Bouillante cluster area. View from SE.
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
strike–slip faults on which were grafted dextral–(normal) NE– SW strike slip fault (Fig. 11). This structural pattern is consistent with roughly N–S extension, nearly parallel to the arc, as suggested by Feuillet et al. (2002). The abnormal high density of faults in the Bouillante cluster and its geometric distribution compatible with a horsetail end of major E–W fault argue in favour of a regional E–W striking normal fault going through the Bouillante Bay. However, out of Bouillante cluster, it needs further geological and geophysical investigations to define the trajectory of this fault zone and to understand its possible relationship with the Beaugendre and Vieux–Habitants horseshoe-shaped depression, interpreted as flank collapses by Boudon et al. (2007). iii. Coreil cluster Two sets of faults are identified offshore: a NE–SW-striking faults set and an ESE–WNW-striking faults set (Fig. 9) and few indicators of deformation exist onshore. Concerning the ESE–WNW-striking faults set, the reduction of the shelf south of this cluster could be related to the passage of a hypothetical
193
fault oriented ESE–WNW (called Vieux Habitants, Fig. 13), deduced from a lineament observed on aerial photographs by Grellet et al. (1988). Note that this hypothetical fault is not presented in the 3D model (Fig. 9) because it is only a lineament with no indication of dip. The set of 4 short NE–SW-striking faults (Fig. 9) are interpreted as a conjugate strike slip network, grafted onto the main N160°E-striking sinistral fault. A recent tectonic activity near Coreil is suggested by a fault shifting the seabed (Fig. 4a). The relationships between the different sets of faults are not defined because of the low density of seismic data and the lack of inland outcrops. Between the Coreil and Bouillante clusters, the 3D model highlights an area without observed faults, which coincides with a strong sub-circular magnetic anomaly and a convex morphology of the shelf (Fig. 4b). These observations could mark the occurrence of young volcanic magnetic product — hiding the faults — as suggested by youngest K/Ar age obtained around Bouillante, on the Muscade volcanic rock (479 ± 16 ky; Sanjuan et al., 2008).
Fig. 13. Interpretative structural pattern of the Bouillante geothermal province built from the 3D fault model and kinematics indications deduced from marine seismic profiles and onshore micro-structural measurements. White arrow: N–S regional extension direction.
194
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
6.2. Dynamic interpretation The integration of kinematic criteria obtained onshore (from micro-tectonic measurements) and offshore (from offsets of seismic reflector) in the regional geometric fault model provides a dynamic interpretation compatible with a regional ~N–S extension, as discussed by Feuillet et al. (2002). One distinguishes the major NNW– SSE-striking sinistral strike slip fault with its NE–SW-striking dextral secondary structures and the E–W-striking faulting network (major fault or not), forming horst and graben perpendicular to the direction of extension (Fig. 13). 7. Discussion and perspectives Merging onshore and offshore data improves the geological interpretation of the MBLS fault system. The resulting 3D fault model helps to better understand the regional structural geology of the Bouillante geothermal province. Data from the sea and data from the land are used simultaneously to build a coherent interpretation across the coast (e.g. Fig. 7 for the Bouillante fault). In addition, the geological knowledge from the land benefits to the comprehension of the structures at sea and reciprocally. This methodology is significantly more powerful than trying to connect two separate interpretations, one onshore and the other offshore. The interpretation presented in previous works (e.g. Thinon et al., 2010) is refined. In particular, our methodology highlights NE–SW structures linked to the MBLS fault system. Such structures define, limit and shift compartments where the E–W to ESE–WNW faults elongate. The marine seismic interpretation combined to the field observation is a guide to define the faults hierarchy by showing that the NE–SW-striking faults cut the E–W to ESE–WNW ones. This example demonstrates how the methodology allows to build a fault network giving clues for geometry and kinematics pattern. The 3D fault model provides indications for future field work. It was not possible to connect some E–W structures interpreted offshore to observation onshore and vice versa. However places were these modelled E–W faults cross the coast will be interesting spots to investigate. Some NE–SW structures modelled offshore would affect inland terrains if prolonged. In that way this 3D fault model is a tool to point out probable areas where to explore for such structures onshore. The regional scale of the model needs to be considered because some assumptions were made. Consequently, various issues need to be tackled to enhance the fault model. - For instance, the NNW–SSE-trending section of the MBLS fault system is modelled as a single infinite fault crossing the zone, along with some correlated finite faults offshore the Bouillante bay. However, the interpretations from seismic data show that this section is an “en relais” fault system, and locally cut and offset by some E–W faults such as the Bouillante one. - The geometry of the MBLS fault system is extrapolated on the Southern and Northern borders. Some extra marine seismic profiles would be sufficient to define its location in these areas. - Offshore, the top of the volcanic basement is not represented in this 3D-fault model, because its depth is not enough constraint. The depth of the basement and the offsets of faults have been estimated from seismic interpretations in time converted in depth by a simple seismic velocity law. A two layers velocity model was applied by assuming a water layer (1500 m/s) and a sedimentary layer (2000 m/s). That does not affect significantly the geometry at regional scale but a more accurate estimation of the basement depth would give better quantification of the dips and the vertical offsets of the faults and consequently refine the fault model. A more realistic estimation for time–depth law is needed to improve the geometry of the fault model. This could be partly
achieved by using the velocities measured in the onshore boreholes. - The fault network is uncompleted; some faults are crossing each other because no clue is available to define their hierarchy. Thus, relations between faults are to be determined to solve the current ambiguity. The 3D fault model of the Bouillante province presented in our study is the final state of numerous iterations. Preliminary models were constructed to test various hypotheses for interpreting faults. These models were modified, rejected or refined but they are not shown here. Such an iterative process lead to the current model described upper. Even if the methodology is deterministic — i.e. a single model is presented — the evolution of the interpretation is intrinsic to the model. In addition, the current model might be modified and turned into an up-to-date structural interpretation if new data are available. Another step is to fill the fault model with relevant geological formations at the regional scale. Outcrops correlated to the seismic units would help to set the geometry of the geological formations. Furthermore, integrating deep refraction seismic data would help to extend the 3D model at depth in order to describe deeper geological structures. Then, geophysical data, e.g. gravity and magnetics, would be additional information to validate and refine the model by combining its geometry to geophysical data in a forward or an inversion process. On a geothermal point of view, the clusters of faults identified from the fault model could be a guide for exploration. They emphasise areas where permeability is a priori favourable, making them potential geothermal fields. In addition, the geometry of the fault model will be used to constrain hydrothermal simulations. Such simulations will help to understand how fluids circulate within these promising permeable areas. Acknowledgement This work is supported by ADEME (French Agency for Energy and Environment): contract 0805C0044 (GHEMOD) and contract 0805C0039 (GEO3BOU). The AGUADOMAR bathymetric data are provided by the Institut de Physique du Globe de Paris. Our manuscript was significantly enhanced by the input of two anonymous reviewers. References Baubron, J., 1990. Prospection géochimique par analyse des gaz des sols en vue de la localisation d'une fracture majeure sous recouvrement. Faille Montserrat-Marie Galante: secteur de Marie-Galante et Capesterre-Belle-Eau, Basse-Terre (Guadeloupe) = Geochemical prospection by gas analysis of a major burried fracture: the Montserrat–Marie Gallante fault. BRGM Report R-31069. Bouchot, V., Traineau, H., Sanjuan, B., Gadalia, A., Guillou-Frottier, L., Thinon, I., Fabriol, H., Bourgeois, B., Baltassat, J.M., Pajot, G., Jousset, Ph., Lasne, E., Genter, A., 2008. Modèle conceptuel du champ géothermique haute température de Bouillante, Guadeloupe, Antilles françaises. Rapport final. BRGM/RP-57252-FR. 60 pp. Bouchot, V., Sanjuan, B., Traineau, H., Guillou-Frottier, L., Thinon, I., Baltassat, J.M., Fabriol, H., Bourgeois, B., Lasne, E., 2010. Assessment of the Bouillante geothermal field (Guadeloupe, French West Indies): toward a conceptual model of the high temperature geothermal system. WGC-2010 - World Geothermal Congress - Bali - Indonesia - 25-30/04/2010. Boudon, G., Le Friant, A., Komorowski, J.-C., Deplus, C., Semet, M.P., 2007. Volcano flank instability in the Lesser Antilles arc: diversity of scale, processes, and temporal recurrence. Journal of Geophysical Research, Solid Earth 112, B08205. Bourgine, B., Prunier-Leparmentier, A.M., Lembezat, C., Thierry, P., Luquet, C., Robelin, C., 2008. Tools and methods for constructing 3D geological models in the urban environment. In: Ortiz, J.M., Emery, X. (Eds.), The Paris Case: Proceeding of the Eighth international Geostatistics congress, Vol. 2, pp. 951–960http://www.geostats2008. com/2008/cierre2008/?pag=download2&leng=in&IDSESION=155. Bouysse, P., Robert, S., Guennoc, P., Monti, S., 1983. Bathymétrie détaillée (seabeam), anomalie magnétique des Antilles françaises: Interprétation morphostructurale de la vallée et de l'escarpement de la Désirade et des côtes occidentales de basse-terre de Guadeloupe et de Martinique (campagne ARCANTE 2-thermosite N/O Jean-Charcot, decembre 1980). Document du BRGM, no. 63, pp. 261–287.
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195 Bouysse, P., Mascle, A., Mauffret, A., Mercier De Lepinay, B., Jany, I., Leclere-Vanhoeve, A., Montjaret, M.C., 1988. Reconnaissance des structures tectoniques et volcaniques sous-marines de l'arc récent des Petites Antilles. Marine Geology 81, 261–287. Calcagno, P., Chilès, J.P., Courrioux, G., Guillen, A., 2008. Geological modelling from field data and geological knowledge, part I — modelling method coupling 3D potentialfield interpolation and geological rules. Physics of the Earth and Planetary Interiors 171, 147–157. Deplus, C., Le Friant, A., Boudon, G., Komorowski, J.C., Villemant, B., Harford, C., Ségoufin, J., Cheminée, J.L., 2001. Submarine evidence for large-scale debris avalanches in the Lesser Antilles Arc. Earth and Planetary Science Letters 192, 145–157. Duperret, A., 1991. Étude structurale de la zone Monserrat–Marie-Galante (Petites Antilles) : Analyse des levés sismiques de la campagne FAGUAD. Mémoire DEA, Univ. Bretagne occidentale, Brest — France, pp. 1–74. Fabriol, H., 2001. Champ géothermique de Bouillante : synthèse des études géophysiques. BRGM/RP-50259-FR, 43 pp., 17 Fig. Fabriol, H., Bitri, A., Bourgeois, B., Debeglia, N., Genter, A., Guennoc, P., Jousset, P., Miehe, J.M., Roig, J.Y., Thinon, I., Traineau, H., Sanjuan, B., Truffert, C., 2005. Geophysical methods applied to the assessment of the Bouillante geothermal field (Guadeloupe, French West Indies). Proceedings World Geothermal Congress 2005, Antalya, Turkey, pp. 24–29. April 2005, 6 pp. Feuillet, N., 2000, Sismotectonique des Petites Antilles. Liaison entre activité sismique et volcanique, PhD Thesis, Univ. Paris VII — Denis Diderot Univ., France, 1–283. Feuillet, N., Manighetti, I., Tapponier, P., 2001. Extension active perpendiculaire à la subduction dans l'arc des Petites Antilles (Guadeloupe, Antilles Françaises). Comptes Rendus de l' Academie des Sciences Paris 333, 583–590. Feuillet, N., Manighetti, I., Tapponier, P., Jacques, E., 2002. Arc parallel extension and localization of volcanic complexes in Guadeloupe, Lesser Antilles. Journal of Geophysical Research 107 (B12), 2331. Finn, C.A., Morgan, L.A., 2002. High-resolution aeromagnetic mapping of volcanic terrain Yellowstone National Park. Journal of Volcanology and Geothermal Research 115, 207–231. Gadalia, A., Gstalter, N., Westercamp, D., 1988. La chaîne volcanique de Bouillante, BasseTerre de Guadeloupe, (Petites Antilles) : Identité pétrographique, volcanologique et géodynamique. Géologie de la France 2–3, 101–130. Grellet, B., Sauret, B., Bonneton, J.-R., 1988. Cadre général de la tectonique récente de la Guadeloupe. Rapport BRGM, 88SGN 627 GEG. Guennoc, P., Traineau, H., Castaing, C., 2001. Levé bathymétrique le long de la côte ouest de l'île de Basse-Terre, Guadeloupe. Apports à la compréhension du contexte structural du champ géothermique de Bouillante. BRGM R 40944. 30 pp. Henkel, H., Guzman, M., 1977. Magnetic features of fracture zones. Geoexploration (ISSN: 0016-7142) 15 (3), 173–181. doi:10.1016/0016-7142(77)90024-2 July 1977.
195
Kenedi, C.L., Sparks, R.S.J., Malin, P., Voight, B., Dean, S., Minshull, T., Paulatto, M., Peirce, C., Shalev, E., 2010. Contrasts in morphology and deformation offshore Montserrat: new insights from the SEA-CALIPSO marine cruise data. Geophysical Research Letters 37. Lajaunie, C., Courrioux, G., Manuel, L., 1997. Foliation fields and 3D cartography in geology: principles of a method based on potential interpolation. Mathematical Geology 29, 571–584. Mathieu, L., 2010, L'impact des mouvements décrochants sur la structure des volcans : une étude de cas des volcans de la Guadeloupe, de Maderas et du Mont Cameroun. Thèse des Univ. de Dublin et de Clermont-Ferrand. Mathieu, L., de Vries, W., Pilato, M., Troll, V.R., 2011. The interaction between volcanoes and strike–slip, transtensional and transpressional fault zones: analogue models and natural examples. Journal of Structural Geology. doi:10.1016/j.jsg.2011.03.003. Samper, A., Quidelleur, X., Lahitte, P., Mollex, D., 2007. Timing of effusive volcanism and collapse events within an oceanic arc island: Basse-Terre, Guadeloupe archipelago (Lesser Antilles arc). Earth and Planetary Science Letters 258, 175–191. Sanjuan, B., Traineau, H., 2008. French West Indies: development of the Bouillante geothermal field in Guadeloupe. IGA News, Newsletter of the International Geothermal Association, pp. 5–9. July–September 2008, Quarterly no. 73. Sanjuan, B., Genter, A., Roig, J.-Y., Patrier, P., Beaufort, D., Bes de Berc, S., Brach, M., Foucher, J.C., 2004. Travaux de reconnaissance géologique et de prospection de gaz dans les sols appliqués à la région de Bouillante, en Guadeloupe. Rapport BRGM/RP-53445-FR. 61 pp., 2 tabl., 15 fig., 1 ann. Sanjuan, B., Lopez, S., Guillou-Frottier, L., Le Nindre, Y.-M., Menjoz, A., 2008. Travaux de recherche liés au projet GHEDOM-ADEME. Rapport final BRGM/RP-56432-FR. 214 pp., 75 fig., 9 ann. Thinon, I., Guennoc, P., Bitri, A., Truffert, C., 2010. Study of the Bouillante Bay (West Basse-Terre Island shelf): contribution of geophysical surveys to understanding of the structural context of Guadeloupe (French West Indies–Lesser Antilles). Bulletin de la Societe Geologique de France 181 (1), 51–65. Traineau, H., Sanjuan, B., Beaufort, D., Castaing, B.M.C., Correia, H., Genter, A., Herbrich, B., 1997. The Bouillante geothermal field revisited: new data on the fractured geothermal reservoir in light of a future stimulation experiment in a low productive well. 22nd Workshop Geothermal Reservoir Engineering, Stanford, CA. 27–29 January 1997. Truffert, C., Thinon, I., Bitri, A., Lalanne, X., 2004. Using MAGIS for geothermal application — Guadeloupe archipelago in French West Indies. Hydro International 8 (6), 55–57.
Contents lists available at SciVerse ScienceDirect
Tectonophysics journal homepage: www.elsevier.com/locate/tecto
A new 3D fault model of the Bouillante geothermal province combining onshore and offshore structural knowledge (French West Indies) P. Calcagno ⁎, V. Bouchot, I. Thinon, B. Bourgine BRGM, 3 avenue C. Guillemin, 45060 Orléans Cedex 2, France
a r t i c l e
i n f o
Article history: Received 28 February 2011 Received in revised form 12 August 2011 Accepted 16 August 2011 Available online 30 August 2011 Keywords: 3D modelling Fault modelling Onshore and offshore combination Geothermal energy Geothermal exploration Bouillante
a b s t r a c t The Bouillante area hosts geothermal resources located in a complex structural area (Guadeloupe Island, French West Indies). On one hand, faults observed on the field mainly elongate along the E–W direction. On the other hand, offshore structures interpreted from marine seismic lines shows a larger range of directions. A coherent 3D interpretation is proposed through a fault model combining onshore and offshore structural knowledge in a zone crossing the island coastline. The fault network constructed reveals a hierarchy in the family of structures and highlights the prevalence of the NNW–SSE direction, associated with secondary NE–SW-trending structures, and the E–W direction. On a geographical point of view, the modelled faults are gathered in 3 clusters. Data available to build the 3D fault model are sometimes sparse, especially inland because of intense vegetation cover. Consequently, not only the results and impacts of the 3D fault model are discussed but also its limitations as well as its possible evolution. © 2011 Elsevier B.V. All rights reserved.
1. Introduction
2. Regional structural setting
The Bouillante geothermal province area is located in the French West Indies, on the West of the Guadeloupe Basse-Terre Island (Fig. 1). That zone is known for its high-temperature geothermal field (e.g. Bouchot et al., 2010; Fabriol et al., 2005; Sanjuan and Traineau, 2008) which produce electricity at the 15 MWe installed power plant in Bouillante (Figs. 1 and 2). Our study is dedicated to better understand the structural pattern of the Bouillante geothermal province. This is a prerequisite for a better knowledge of the Bouillante geothermal field understanding. At the scale of the geothermal province, onshore and offshore data are required to interpret in a coherent way the faults in the area because of its location both inland and at sea. Previous works aimed at describing the structural geology on land (e.g. Bouchot et al., 2010; Boudon et al., 2007; Feuillet et al., 2002; Mathieu, 2010; Samper et al., 2007; Traineau et al., 1997) and at sea (e.g. Bouysse et al., 1983, 1988; Feuillet, 2000; Thinon et al., 2010). We propose to merge data and knowledge from field work and from marine exploration to produce a model in 3 dimensions representing the faults down to 2 km depth at the scale of the Bouillante geothermal province that is about 250 km 2.
The Bouillante geothermal province is a regional key geodynamic area where the major tectonic and volcanic structures of the inner arc of the Lesser Antilles join (e.g. Bouysse et al., 1988; Feuillet, 2000; Feuillet et al., 2001, 2002; Thinon et al., 2010). As far as can be determined from our present state of knowledge, the Bouillante geothermal field is located at the junction of two regional fault systems (Fig. 1a):
⁎ Corresponding author. Tel.: + 33 238643054. E-mail address: [email protected] (P. Calcagno). 0040-1951/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.tecto.2011.08.012
(i) The NW–SE Montserrat–Bouillante (MB) fault system was observed from the northern of the Directeur volcano seamount to the foot of the NW Basse-Terre slope (Bouysse et al., 1988). It is composed of several ‘en echelon’ oblique normal faults that accommodate a component of sinistral motion (Duperret, 1991; Feuillet et al., 2002; Kenedi et al., 2010). (ii) The ESE–WNW to E–W Bouillante–Capesterre (BC) fault system which could belongs to the E–W Marie–Galante graben system crosscuts the volcanic island of Basse-Terre from Bouillante town (western side) to Capesterre town (eastern side) with a normal component (Feuillet et al., 2002). This hypothetic south dipping fault zone that separates the Axial Chain from the Bouillante–Sans–Toucher complex (Figs. 1b and 2) is mainly interpreted from topographic evidences (e.g. continuous step offsetting the topography), from few outcrops on the west coast (Feuillet et al., 2002), and is partly suggested by gas anomalies close to Capesterre town (Baubron, 1990). However, according to Boudon et al. (2007), these topographic steps represent the scare of two collapse structures (northern part of the
186
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
Fig. 1. (a) Geodynamic context of the West Indies. NAM: North American plate. SAM: South American plate. The grey area covers the fore arc developed on ~ 250 km wide in E–W between the subduction zone and the western edge of the volcanic arc of the Lesser Antilles. From North to South, the motion ranges from sinistral extensional shear to dextral oblique thrusting (Feuillet et al., 2002). (b) Regional structural setting of the Guadeloupe Island. Thick square: location of the 3D fault model including the Bouillante field. BCF: Bouillante–Capesterre fault system. The trajectory of this hypothetical fault is extracted from the geological, morphological and structural map of southern Basse-Terre realised by Feuillet et al. (2002). It mainly based on alignments of steps offsetting the topography. a, modified from Feuillet et al., 2002; b, modified from Feuillet et al., 2001 and Thinon et al., 2010.
Beaugendre and Vieux-Habitants horseshoe-shaped depression) rather than a fault plan. Even if occurrence of debris avalanche on the mouth of the structure argues in favour of flank collapses, the origin of these collapses — volcano-hydrothermal, or structural (normal or strike-slip faulting) — are highly debated (Boudon et al., 2007; Mathieu et al., 2011; Samper et al., 2007).
From Feuillet et al. (2001), the MB fault system would join the BC fault system, thus drawing at the scale of the northern Caribbean arc a sinistral extensional horsetail. From these authors, the western end of the BC fault is correlated to the Machette–Pointe à Sel fault observed on the coast at north of the Bouillante Bay. Thinon et al. (2010) propose that the BC fault system extends into the Bouillante Bay and cuts the whole shelf. A third fault system, described by Thinon et al. (2010), is a N160°E-trending escarpment which limits the shelf of the Bouillante sector. This escarpment is interpreted as a major sinistral strike–slip system, playing the role of transfer zone between the MB and Les Saintes (LS) fault systems. This discontinuity belongs to the western limit of a zone of sinistral extensional shear, developed on 250 km wide in E–W between the subduction zone and the western edge of the volcanic arc of the Lesser Antilles (Feuillet et al., 2002; Fig. 1a). 3. Onshore structural knowledge
Fig. 2. Onshore structural knowledge. A selection of structural data and interpretations available onshore. The geothermal boreholes BO5 and BO6, belonging to the power plant in Bouillante, give structural information at depth.
The Bouillante geothermal field is contained within an andesitic volcanic substratum largely attributed to sub-products of the axial Pitons de Bouillante chain cropping out along the axis of Basse-Terre Island. Recent small volcanic centres (0.8 and 0.2 Ma) of the ‘Bouillante Chain complex’ defined by Gadalia et al. (1988) lie on this volcanic substratum. Immediately north of Bouillante City, several volcanic centres and related deposit, i.e. lava flow or pyroclastic deposits, have been dated between 840,000 y, e.g. strombolian cone of Pointe à Sel, and 490,000 y, e.g. Desmarais centre, from K/Ar dating (Gadalia et al., 1988 and recent BRGM unpublished data). Aligned along a NNW–SSE trending band of some 20 km by 4 km between Anse Colas (North) and the Basse-Terre City (South), the ‘Bouillante Chain complex’ volcanism, and its offshore and onshore volcanic centres seems controlled by the offshore NNW–SSE MB fault zone. These volcanic rocks often emplaced under hydro-magmatic condition (e.g. hyaloclastite) and have the characteristics of a weakly potassic tholeiitic series that locally evolved through fractional crystallisation (Gadalia et al., 1988). The studied area is located on the West side of the volcanic BasseTerre Island where the prevailing climate is wet tropical. This specific geographic location constrains the acquisition of structural data.
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
Exposures suitable for structural measurements are mainly found only on the seaside where coastal erosion has stripped away the soil (Fig. 2). Since the 1970s when geothermal exploration began around Bouillante, several field works have succeeded in acquiring sets of structural data on outcrops (e.g. Bouchot et al., 2010; Sanjuan et al., 2004; Traineau et al., 1997). From a recent synthesis of existing structural interpretation (Bouchot et al., 2008) and acquisition of new structural measurements by the authors (Fig. 2), it appears that most of the measured faults have a trend E–W to ESE–WNW, with a dominant normal throw showed by the evidence of normal offset of volcanic layer on each side of the fault surface with striae and drag fold in few stations. Moreover, a very limited second family, consisting of only two faults visible onshore, trends NE–SW with a strong dextral strike–slip component, according to tectoglyphs such as striae and tension gashes. In detail, onshore structural analysis provides essential information for our understanding, from south to north: - On the Bouillante geothermal field, fracturing is a 2.5 km wide network of seven faults. These faults strike E–W to ESE–WNW and dip steeply (~ 70–80°) to the north or the south. They accommodated a dominant normal slip according to some drag folds (Fig. 3), rare fault striate on fault plan, and the shift of the volcanic levels on both sides of faults. Spaced from 200 to 700 m, these normal faults delimit horst and graben compartments and in particular the Bouillante graben located on the southern part of the Bouillante Bay (Bouchot et al., 2010). - In the northern part of the Bouillante Bay, an E–W-striking fault zone is cut by a 15 m-large, steeply dipping, NE–SW-striking fault zone of Pointe Marsolle, which is characterised by an atypical anastomosing pattern. This shear zone shows a dextral kinematics according to en echelon networks of vertical tension gashes, partly filled by quartz and/or calcite. The structure continues offshore along a NE–SW ridge marked in the bathymetry, and onshore, marks a sharp NE–SW trending contrast in terms of resistivity anomaly showed by a MELOS campaign (Fabriol, 2001; Fig. 2). - Further north around Malendure village, although less intense than in the Bouillante area, deformation is also characterised by several E–W to ESE–WNW-striking normal faults and one NE– SW-striking fault showing a dextral–normal throw. Along this
187
fault, a N65°E-striking, 85°S dipping fault plan shows brittle striaes with a pitch of 50°W, indicating a dextral–normal oblique motion. Some limited fractures trending NW–SE, associated with reddish hydrothermal alteration, have also been identified to the north of Malendure village, close to Anse Colas. As soon as one moves away from the coast, the volcanic island is covered by dense tropical forest on very steep relief. Given a rainfall that is 10 times higher than at the coast, this forest zone is the site of intense supergene alteration that impedes ground observation, except where ravines generated by torrents have been cut into the weathering profile. Nevertheless, the hilly areas where topographic regular structures are sharp can provide indirect information on fault traces (Feuillet et al., 2002; Fig. 2) and on flank collapse (Boudon et al., 2007; Fig. 2). But supplementary data are needed to confirm the origin of the topographic anomalies and collapses. Similarly abnormal geophysical signatures (Fabriol et al., 2005) provide indirect data likely to indicate the presence of a fault at depth, e.g. NE–SW-striking Pointe Marsolle fault mark by a sharp contact between different ground resistivities (Fig. 2). Overall, the combination of direct and indirect structural data has enabled an interpretative onshore structural pattern to be considered, albeit poorly constrained at depth given the small amount of information available (borehole logs, electrical surveys, extrapolation of observed dips). Finally, it is clear that the patchy distribution of direct data is a real challenge to build an interpretative structural pattern of the studied area. 4. Offshore structural knowledge Shallow water multibeam echosounder (MBES) bathymetric, very high resolution (VHR) reflection seismic (Sparker source, 6-channel) and high resolution (HR) magnetic surveys were conducted in 1998 and 2003 by the BRGM (Bathybou98 and Geoberyx03 cruises) on the west Basse-Terre Island shelf, centred on the Bouillante Bay sector (Guennoc et al., 2001; Thinon et al., 2010). The bathymetry (Fig. 4a) shows the present-day morphology of the shelf, its extension, the dimensions and forms of the bays, the occurrence of escarpments and the prolongation of promontories. Furthermore, the morphology underlines some main directions, as
Fig. 3. Picture of the N95°E-striking, dipping 70°S normal Machette–Pointe à Sel fault located in the Anse Machette (see Fig. 2 for location). This normal fault put in tectonic contact a white layered deposit characterised by alternating deposits of coeruptive pumices fall and pyroclastic flows (north block) and a brown slightly reworked coarse pyroclastic deposit (south block). The dip-slip of the fault is estimated around 20 m from the evidence of brown pyroclastic deposit above the white layered in the north block. The layering (S0) is affected by drag fold along the main normal fault. The red arrow locates the S0-N130°E-17°S measurement.
188
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
Fig. 4. Bathymetry and acoustic basement maps. Isobaths in metres. (a) Bathymetry of the NW shelf of Basse-Terre. Compilation of MBES data: BRGM Bathybou98 survey, 1998; DTM, mesh 5 m and 25 m (Guennoc et al., 2001); IPG Aguadomar survey, 1998–1999, DTM 250 m (Deplus et al., 2001). Stars: Location of faults shifting the seabed. (b) Isobaths map of top of acoustic basement of the NW shelf of Basse-Terre and main faults derived from the seismic interpretations of the Geoberyx03 survey.
for example the E–W-trending Marsolle escarpment in the northern part of the Bouillante Bay (see Fig. 4b for location). The depth of the seabed is a reference data for the 3D model, as much as the topography. The Geoberyx03 cruise has provided a dense network — about 180 km — of the VHR reflection seismic images of the sedimentary cover (geometry, thickness, seismic facies, etc.) and the volcanic basement in second Two Way Travel Time (TWTT, Fig. 5). The acoustic basement is interpreted as being the volcanic basement of BasseTerre Island. These data show the tectonic structures and the relationships between the acoustic basement and the sedimentary cover. According to the shifts of the seismic reflectors along faults, vertical throws of faults are a few metres to tens of metres at least. The hori-
zontal displacement is not deductible and the faults within the acoustic basement are not visible. The processing of the HR magnetic survey of the Geoberyx03 cruise (e.g. Truffert et al., 2004) has highlighted a strong circular magnetic anomaly off the Bouillante Chain on the shelf between the South Bouillante Bay and the Vieux Habitants town (Fig. 4b). This anomaly has been interpreted by Thinon et al. (2010) as the marker of a part of a volcanic seamount located over the shelf in the axis of Upper Miocene to actual submarine volcanoes (Feuillet et al., 2001). The association of the geophysical data provides a good image of the geometry of the structure. However, no geological data are available to constrain the age and the kind of formations, nor structural measurements to quantify true dips.
Fig. 5. One of the Geoberyx03 VHR seismic profiles through the NW shelf offshore Bouillante (Thinon et al., 2010). It shows the thickness of the sedimentary cover (UP, UC, RB, UB), the morphology of the acoustic basement (US), the apparent dip and the vertical throw of the major faults, the N160° basement high and the geometry of basins (BSP) on the internal shelf. RB: surficial formation. Various seismic units are defined and noted by the symbol U. Box: location of the marine seismic section (West Guadeloupe offshore).
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
The interpretation of these surveys has helped to highlight several major structural features (Thinon et al., 2010): i. The sedimentary cover is generally thin, except on the outer shelf where thick sedimentary unit deposition induced an important widening of the shelf. The thickness of the sedimentary cover is between 0.02 and 0.04 s twtt. It reaches 0.085 s twtt thick (~80 m) in the small basins, i.e. the BSP basin (Fig. 5). The sedimentary cover thickens to 0.18 s twtt on the external shelf to the west of the basement high (Fig. 5). The bathymetry map doesn't reflect the morphology of the volcanic basement (Fig. 4a). ii. The main directions of structures, observed on the shelf, are by increasing importance NNW–SSE, E–W, NW–SE and NE–SW. iii. The western edge of the shelf is a NNW–SSE-trending escarpment, may be a major sinistral strike–slip system, playing the role of transfer zone between the NW–SE MB and LS fault systems. The N160°E-trending escarpment is composed by segments shifted southward from West to East. This structure limits also an asymmetric basement high, faulted and shifted. iv. The E–W-trending Marsolle escarpment, located in the Bouillante area (Fig. 4b), could be the offshore prolongation of the E–W-trending BC fault system. Indeed, an E–W major structure extends into the Bouillante Bay from the coast to the shelfbreak crossing the NNW–SSE- and NW–SE-striking faults. It consists of two segments shifted southward from east to west which coincides with the northern bound of the Bouillante Bay. v. The NW–SE-striking faults are mainly observed on the insular slope and on the external shelf off the Bouillante Bay, with a “staircase”-type distribution. vi. A major basement high (Figs. 4b and 5) forms a 500 m wide ridge extending for 1600 m in a N160°E direction. This high reaches a sea-bottom depth of about 70 m to the north and 75 m to the south on both sides of the E–W-trending Marsolle escarpment. It is bounded to the west by the NNW–SSE-trending escarpment which underlies the shelf-break of the acoustic basement. vii. A volcano edifice extends over the shelf on south of Bouillante Bay, according to the observed large magnetic anomaly. Based from only indirect observations, the relationships between the structures of the shelf are not very well defined. 5. Data and methodology A 15 km × 16 km zone (see Fig. 1b for location) is modelled down to 2 km below sea level in the local “Guadeloupe, St. Anne–UTM20– IGN”, WGS84 projection system. In that zone, faults are interpreted by taking into account both offshore and onshore geological knowledge. Onshore, data consist in strike and dip measurements of fault measured during field work. Due to outcrops accessibility, these data are located on the coast. Inland, some structures are interpreted from previous works (e.g. Feuillet et al., 2002). Offshore, location and dip of the faults are derived from the interpretation of 52 Geoberyx03 seismic profiles (Thinon et al., 2010). The magnetic lineaments, interpreted as markers such as faults or dykes are clues for the existing orientations of structures on the shelf (Finn and Morgan, 2002; Henkel and Guzman, 1977). The information on the age and nature of geological formations are derived from geological surveys at land. Preliminary to the 3D model, a Digital Elevation Model (DEM) was created to combine onshore and offshore data in a coherent 50 m resolution 2D grid covering the whole modelled area. The GDM1 software was used for this work (Bourgine et al., 2008). Onshore data are provided by the IGN (Institut Géographique National, France) on a 50 m grid in the local UTM coordinate system. Offshore, two bathymetric surveys are
1 GDM is a commercial software developed by BRGM. For further information visit: http://gdm.brgm.fr/?lang=en.
189
used: Deep water multibeam echosounder of the AGUADOMAR cruise (1998–1999, IPG, Deplus et al., 2001) and the shallow water multibeam echosounder of the Bathybou98 cruise (1998, BRGM, Guennoc et al., 2001). The AGUADOMAR data are available on a 100 m grid in latitude–longitude coordinates. The Bathybou98 data are available on a 50 m grid in the local UTM coordinate system. Data are kriged with a linear variogram to fill the gaps of few hundreds of metres between onshore and offshore. The full grid is interpolated using a bilinear interpolation to get a grid in the WGS84 coordinate system used in this study. The resulting DEM (Fig. 6) limits the top of the structural 3D model. The way for interpreting the faults is driven by the size of the zone. In particular, only relevant faults at the scale of the model were taken into account. Some assumptions regarding their geometry were also made. For instance, the Montserrat–Bouillante–Les Saintes (MBLS) fault system was interpreted as a continuous fault crossing the whole modelled area and extending to the bottom of the box. To note, the geometry of the MBLS fault system is extrapolated on the Southern and the Northern parts of the zone due to the lack of data: From North Malendure to the MB fault system observed in the abyssal plain at the North of Directeur volcano seamount, the direction and localisation of the N160°E-striking fault system is hypothetical and based on the present-day shelf-break (bathymetric data). The southern part of the MBLS fault system was also extrapolated to the south of Coreil. It doesn't reflect the morphology of the basement, the shelf is much closer. The horizontal extension of the other faults is limited by the location of their data. At depth, no data is available to quantify the extension of the faults. They were limited by taking into account their horizontal extension, i.e. the geometry of finite faults is assumed to be limited by half a sphere centred on the middle of the horizontal trace of the fault. A fault network was built where information was available to determine the relation between contiguous faults. Where such information is not available, modelled faults cross each other. In addition, some of the structures are not directly observed but deduced from the observed faults. As an example, an offset or a termination of an observed fault drives the interpretation of a nonobserved fault. To construct a coherent interpretation, the structures observed onshore — mainly E–W oriented — were as much as possible used as a guide to interpret and to connect the offshore structures oriented in the same direction. However, our methodology does not simply consist in trying to connect faults, separately interpreted or observed, at sea and inland. Offshore and onshore information are used together to build a coherent structural 3D interpretation covering the whole area. To achieve this goal, the geometry of a fault is computed using a set of data merging marine and inland locations of this fault (e.g. the Bouillante fault on Fig. 7).Data are interpolated using the potential field method developed in BRGM (Calcagno et al., 2008; Lajaunie et al., 1997) and implemented in the 3D GeoModeller2 software. 6. 3D fault model 6.1. Faults analysis Statistical analyses of the faults coming from the 3D regional fault model bring relevant information on the distribution of onshore and offshore deformation. The pole density contour plotted for the offshore faults (Fig. 8a) shows that the main set of fault is striking ESE–WNW, and mainly dipping SSW. Two others sets are striking NNW–SSE and NE–SW with opposite dipping. In contrast, the pole density contour plot of the onshore faults (Fig. 8b) indicates that the main set of faults is striking from E–W to ESE–WNW and dipping
2 3D GeoModeller is a commercial software developed by BRGM and Intrepid Geophysics. For further information visit: http://www.geomodeller.com.
190
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
Fig. 6. DEM reconstructed from onshore data (IGN) and offshore data (Bathybou98 and AGUADOMAR surveys). The coastline (dark green line) separates onshore and offshore domains. Size of the area: 15 km × 16 km, vertical exaggeration × 2, elevations ranging from − 1120 m to + 1150 m from sea level. Corner coordinates (W–E; S–N) in metres and in the local “Guadeloupe, St. Anne–UTM20–IGN”, WGS84 projection system. View from the S–SW.
north or south as a horst and graben pattern. Taking into account the whole set of onshore and offshore faults, it appears that at the model scale, faults are mainly striking from ESE–WNW to NNW–SSW and mainly dipping SSW to WSW (Fig. 8c and d). The final 3D interpretation highlights coherent structures onshore and offshore (Fig. 9). The modelled faults can be classified in two main families: The MBLS fault system and the E–W- to ESE–WNWstriking fault system. Both are detailed below. 1. The MBLS fault system: the case of the NNW–SSE-trending section around Bouillante The MBLS fault system is considered as a sinistral strike slip fault zone recognised over 300 km long between St Kitts and Les Saintes north to south (Fig. 1b). We combine the NW–SE MB fault system between Montserrat and Directeur Volcano seamount (Duperret, 1991; Feuillet, 2000), the NNW–SSE MB fault system between Malendure and Coreil and
Fig. 7. Example of onshore and offshore knowledge combination: The Bouillante fault (yellow surface) is constructed from onshore field observation (dark blue point), onshore interpretation (red points) and offshore seismic profiles interpretation (E–Wtrending Marsolle escarpment, light blue points). View from WSW. The coastline (white crosses) separates onshore and offshore domains. See Fig. 1b for modelled area location.
the NW–SE LS fault system (Deplus et al., 2001) to model the MBLS fault system. This fault system is generally oriented NW– SE, but draws/forms a NNW–SSE-trending virgation along the Northern part of Basse-Terre volcanic island (Fig. 9). The latter would be accompanied by a reversal of dip. The virgation could appear as the result of shearing deformation moulded on the coastal shelf of the Basse-Terre Island. It could be induced by the association of the occurrence of the old volcanic complex of the North Basse-Terre, the collapse of the southern part of Basse-Terre linked to the Marie–Galante graben system. The NNW–SSW-trending virgation is characterised by a sub-continuous 30 km long fault zone which strikes N160°E and dips west. In contrast, on both end of the virgation, the NW–SE-striking corridors consist of east-dipping ESE–WNW-striking oblique fault. This “en echelon” oblique normal faults network is interpreted as an indicator of a sinistral transtension (Fig. 1) and the NNW–SSE section of the MBLS fault zone as a N160°E sinistral strike slip fault (Thinon et al., 2010). According to our fault model (Fig. 9), the NNW–SSE-trending section consists of two main parallel N160°E-striking ‘en relais’ faults,
Fig. 8. Statistical analysis of the faults. (a) Pole density contour plot of the offshore faults. (b) Pole density contour plot of the onshore faults. (c) Stereonet plot of poles to onshore (+) and offshore (x) fault planes. (d) Pole density contour plot of the onshore and offshore faults.
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
191
Fig. 9. The whole 3D fault interpretation combines onshore and offshore knowledge. The coastline (white crosses) separates onshore and offshore domains. Two main families of faults are modelled: The MBLS fault system (green and blue faults); The E–W to ESE–WNW-striking fault system (yellow faults). See Fig. 1b for modelled area location. View from SE.
developed along the shelf break. They show an opposite dip (Fig. 10). Separated by 150 m, these two faults limit the development of a 5 km long, NNE–SSW-trending basement high, indicating a normal component (Figs. 4b and 5). We interpret this basement high as a horst structure related to the sinistral strike– slip movement along the N160°E-striking faults system. In addition, a network of several parallel faults, well developed offshore, striking NE–SW, and 3 km long is linked to the offshore N160°E-striking faults. The onshore expression of these faults is very limited (Figs. 9 and 10). Identified only twice on the coast (Bouillante and Malendure outcrops), the NE–SW-striking faults show a dextral movement with a weak normal component. We interpret these short faults as a dextral conjugate network, grafted onto the main N160°E-striking sinistral fault, as a pinnate pattern. In the 3D fault model, some NW–SE-striking faults were modelled mainly on the insular slope (2 dark blue faults) and on the external shelf (2 blue faults). They are located offshore the Bouillante Bay (Figs. 9 and 10). They are limited and sometimes shifted by the E– W and NE–SW faults. The NW–SE faults are also present at North of Malendure, near Pointe Noire town (see Malendure cluster below). 2. The E–W- to ESE–WNW-striking fault system According to our fault model (Fig. 9), the main fracturing strikes E–
W to ESE–WNW, onshore and on the shelf. However, this deformation is very heterogeneous and is divided into 3 clusters of faults: (i) the Malendure cluster on the Northern part of the zone, (ii) the Bouillante cluster on the geothermal field, and (iii) the Coreil cluster on the Southern part of the zone.
Fig. 10. W–E vertical section in the 3D fault model. No vertical exaggeration. See Fig. 9 for location.
192
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
i. Malendure cluster As stated by the regional structural pattern, this fracturing cluster is located at the junction between the NW–SE section and the NNW–SSE section of the MB fault system, in the Malendure area (Figs. 1b and 2). On the shelf, the vertical offsets of two NW–SE-striking faults are low. Onshore, these faults are correlated to discrete fracture zones, controlling reddish hydrothermal alteration. Concerning the E–W- to ESE–WNW-striking normal faults, they do not seem connected to a regional E–W fault but they are well developed between the NW–SE and the NNW–SSE sections of the MBLS fault system. At north of the Malendure cluster, the NW–SE faults observed on coast have been correlated to the faults observed on the seismic profiles. The low density of marine data prevents from specifying the tectonic network and the relationships between the NW–SE Pointe–Noire faults and the EW faults of Malendure. ii. Bouillante cluster The ~EW-striking, steeply dipping, normal faults network forms a horst and graben pattern, and consists in main conduits for geothermal fluids in depth. However, on the shelf, i.e. approaching the main N160°E-striking fault system, the E–W fault trajectories are discontinuous because these faults are stopped on continuous transverse NE–SW-striking faults (Fig. 11). This geometry defines structural blocks composed of E–W normal faults bounded by dextral NE–SW-striking faults. This logic of block suggests that the continuous strike slip plays the role of transfer fault during the development of the E–W-striking normal faults under a ~N–S regional extension. On our fault model (Fig. 11), evidence of intense deformation, focused on the Bouillante cluster, is a strong argument to defend the existence of the major fault zone of BC fault system (Feuillet et al., 2002). The bursting of the E–W fracturing in this area (over 2.5 km in N–S) suggest the evidence of a horsetail end of this regional fault, near the function with the regional NNW–SSE-striking MBLS fault system (Bouchot et al., 2010). Moreover, unlike previous works defending the idea of Machette–Pointe à Sel fault as the major member of the regional E–W fault zone (Feuillet et al., 2002), our fault model shows that this Machette–Pointe à Sel fault (Fig. 11), measured on outcrop (Figs. 2 and 3), has a very limited continuity offshore. By contrast, the steeply south dipping E–W-striking Bouillante fault (Figs. 9, 11, and 12) is the only fault which i) cuts across the shelf, ii) shows a strong offset of the basement at depth (more than 25 m) and the E–W-trending Marsolle escarpment at the surface in the Northern part of the Bouillante Bay, as a result of the normal movement along the fault, and iii) cuts and
Fig. 12. S–N vertical section in the 3D fault model. No vertical exaggeration. See Fig. 9 for location. The interpretation is extrapolated at depth to give a predominant behaviour to the E–W-striking Bouillante fault.
offsets the major N160°E-striking faults at the level of a Bouillante Valley canyon (Fig. 4a, Thinon et al., 2010). This latter characteristic is not represented at the scale of the regional 3D fault model. Consequently, the Bouillante fault appears as a main member of the regional E–W fault, on which are associated the other E–W-striking normal faults whose dips are opposed (Figs. 11 and 12). On the Bouillante cluster, seismic images show two faults which shift the seabed off Bouillante (Fig. 4a), suggesting that the tectonic activity is recent. Finally, we interpret the fracturing cluster of Bouillante as the junction zone between the major E–W fault zone through Bouillante town, and the major sinistral–(normal) N160°E
Fig. 11. Zoom of the 3D fault model on the Bouillante cluster area. View from SE.
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
strike–slip faults on which were grafted dextral–(normal) NE– SW strike slip fault (Fig. 11). This structural pattern is consistent with roughly N–S extension, nearly parallel to the arc, as suggested by Feuillet et al. (2002). The abnormal high density of faults in the Bouillante cluster and its geometric distribution compatible with a horsetail end of major E–W fault argue in favour of a regional E–W striking normal fault going through the Bouillante Bay. However, out of Bouillante cluster, it needs further geological and geophysical investigations to define the trajectory of this fault zone and to understand its possible relationship with the Beaugendre and Vieux–Habitants horseshoe-shaped depression, interpreted as flank collapses by Boudon et al. (2007). iii. Coreil cluster Two sets of faults are identified offshore: a NE–SW-striking faults set and an ESE–WNW-striking faults set (Fig. 9) and few indicators of deformation exist onshore. Concerning the ESE–WNW-striking faults set, the reduction of the shelf south of this cluster could be related to the passage of a hypothetical
193
fault oriented ESE–WNW (called Vieux Habitants, Fig. 13), deduced from a lineament observed on aerial photographs by Grellet et al. (1988). Note that this hypothetical fault is not presented in the 3D model (Fig. 9) because it is only a lineament with no indication of dip. The set of 4 short NE–SW-striking faults (Fig. 9) are interpreted as a conjugate strike slip network, grafted onto the main N160°E-striking sinistral fault. A recent tectonic activity near Coreil is suggested by a fault shifting the seabed (Fig. 4a). The relationships between the different sets of faults are not defined because of the low density of seismic data and the lack of inland outcrops. Between the Coreil and Bouillante clusters, the 3D model highlights an area without observed faults, which coincides with a strong sub-circular magnetic anomaly and a convex morphology of the shelf (Fig. 4b). These observations could mark the occurrence of young volcanic magnetic product — hiding the faults — as suggested by youngest K/Ar age obtained around Bouillante, on the Muscade volcanic rock (479 ± 16 ky; Sanjuan et al., 2008).
Fig. 13. Interpretative structural pattern of the Bouillante geothermal province built from the 3D fault model and kinematics indications deduced from marine seismic profiles and onshore micro-structural measurements. White arrow: N–S regional extension direction.
194
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195
6.2. Dynamic interpretation The integration of kinematic criteria obtained onshore (from micro-tectonic measurements) and offshore (from offsets of seismic reflector) in the regional geometric fault model provides a dynamic interpretation compatible with a regional ~N–S extension, as discussed by Feuillet et al. (2002). One distinguishes the major NNW– SSE-striking sinistral strike slip fault with its NE–SW-striking dextral secondary structures and the E–W-striking faulting network (major fault or not), forming horst and graben perpendicular to the direction of extension (Fig. 13). 7. Discussion and perspectives Merging onshore and offshore data improves the geological interpretation of the MBLS fault system. The resulting 3D fault model helps to better understand the regional structural geology of the Bouillante geothermal province. Data from the sea and data from the land are used simultaneously to build a coherent interpretation across the coast (e.g. Fig. 7 for the Bouillante fault). In addition, the geological knowledge from the land benefits to the comprehension of the structures at sea and reciprocally. This methodology is significantly more powerful than trying to connect two separate interpretations, one onshore and the other offshore. The interpretation presented in previous works (e.g. Thinon et al., 2010) is refined. In particular, our methodology highlights NE–SW structures linked to the MBLS fault system. Such structures define, limit and shift compartments where the E–W to ESE–WNW faults elongate. The marine seismic interpretation combined to the field observation is a guide to define the faults hierarchy by showing that the NE–SW-striking faults cut the E–W to ESE–WNW ones. This example demonstrates how the methodology allows to build a fault network giving clues for geometry and kinematics pattern. The 3D fault model provides indications for future field work. It was not possible to connect some E–W structures interpreted offshore to observation onshore and vice versa. However places were these modelled E–W faults cross the coast will be interesting spots to investigate. Some NE–SW structures modelled offshore would affect inland terrains if prolonged. In that way this 3D fault model is a tool to point out probable areas where to explore for such structures onshore. The regional scale of the model needs to be considered because some assumptions were made. Consequently, various issues need to be tackled to enhance the fault model. - For instance, the NNW–SSE-trending section of the MBLS fault system is modelled as a single infinite fault crossing the zone, along with some correlated finite faults offshore the Bouillante bay. However, the interpretations from seismic data show that this section is an “en relais” fault system, and locally cut and offset by some E–W faults such as the Bouillante one. - The geometry of the MBLS fault system is extrapolated on the Southern and Northern borders. Some extra marine seismic profiles would be sufficient to define its location in these areas. - Offshore, the top of the volcanic basement is not represented in this 3D-fault model, because its depth is not enough constraint. The depth of the basement and the offsets of faults have been estimated from seismic interpretations in time converted in depth by a simple seismic velocity law. A two layers velocity model was applied by assuming a water layer (1500 m/s) and a sedimentary layer (2000 m/s). That does not affect significantly the geometry at regional scale but a more accurate estimation of the basement depth would give better quantification of the dips and the vertical offsets of the faults and consequently refine the fault model. A more realistic estimation for time–depth law is needed to improve the geometry of the fault model. This could be partly
achieved by using the velocities measured in the onshore boreholes. - The fault network is uncompleted; some faults are crossing each other because no clue is available to define their hierarchy. Thus, relations between faults are to be determined to solve the current ambiguity. The 3D fault model of the Bouillante province presented in our study is the final state of numerous iterations. Preliminary models were constructed to test various hypotheses for interpreting faults. These models were modified, rejected or refined but they are not shown here. Such an iterative process lead to the current model described upper. Even if the methodology is deterministic — i.e. a single model is presented — the evolution of the interpretation is intrinsic to the model. In addition, the current model might be modified and turned into an up-to-date structural interpretation if new data are available. Another step is to fill the fault model with relevant geological formations at the regional scale. Outcrops correlated to the seismic units would help to set the geometry of the geological formations. Furthermore, integrating deep refraction seismic data would help to extend the 3D model at depth in order to describe deeper geological structures. Then, geophysical data, e.g. gravity and magnetics, would be additional information to validate and refine the model by combining its geometry to geophysical data in a forward or an inversion process. On a geothermal point of view, the clusters of faults identified from the fault model could be a guide for exploration. They emphasise areas where permeability is a priori favourable, making them potential geothermal fields. In addition, the geometry of the fault model will be used to constrain hydrothermal simulations. Such simulations will help to understand how fluids circulate within these promising permeable areas. Acknowledgement This work is supported by ADEME (French Agency for Energy and Environment): contract 0805C0044 (GHEMOD) and contract 0805C0039 (GEO3BOU). The AGUADOMAR bathymetric data are provided by the Institut de Physique du Globe de Paris. Our manuscript was significantly enhanced by the input of two anonymous reviewers. References Baubron, J., 1990. Prospection géochimique par analyse des gaz des sols en vue de la localisation d'une fracture majeure sous recouvrement. Faille Montserrat-Marie Galante: secteur de Marie-Galante et Capesterre-Belle-Eau, Basse-Terre (Guadeloupe) = Geochemical prospection by gas analysis of a major burried fracture: the Montserrat–Marie Gallante fault. BRGM Report R-31069. Bouchot, V., Traineau, H., Sanjuan, B., Gadalia, A., Guillou-Frottier, L., Thinon, I., Fabriol, H., Bourgeois, B., Baltassat, J.M., Pajot, G., Jousset, Ph., Lasne, E., Genter, A., 2008. Modèle conceptuel du champ géothermique haute température de Bouillante, Guadeloupe, Antilles françaises. Rapport final. BRGM/RP-57252-FR. 60 pp. Bouchot, V., Sanjuan, B., Traineau, H., Guillou-Frottier, L., Thinon, I., Baltassat, J.M., Fabriol, H., Bourgeois, B., Lasne, E., 2010. Assessment of the Bouillante geothermal field (Guadeloupe, French West Indies): toward a conceptual model of the high temperature geothermal system. WGC-2010 - World Geothermal Congress - Bali - Indonesia - 25-30/04/2010. Boudon, G., Le Friant, A., Komorowski, J.-C., Deplus, C., Semet, M.P., 2007. Volcano flank instability in the Lesser Antilles arc: diversity of scale, processes, and temporal recurrence. Journal of Geophysical Research, Solid Earth 112, B08205. Bourgine, B., Prunier-Leparmentier, A.M., Lembezat, C., Thierry, P., Luquet, C., Robelin, C., 2008. Tools and methods for constructing 3D geological models in the urban environment. In: Ortiz, J.M., Emery, X. (Eds.), The Paris Case: Proceeding of the Eighth international Geostatistics congress, Vol. 2, pp. 951–960http://www.geostats2008. com/2008/cierre2008/?pag=download2&leng=in&IDSESION=155. Bouysse, P., Robert, S., Guennoc, P., Monti, S., 1983. Bathymétrie détaillée (seabeam), anomalie magnétique des Antilles françaises: Interprétation morphostructurale de la vallée et de l'escarpement de la Désirade et des côtes occidentales de basse-terre de Guadeloupe et de Martinique (campagne ARCANTE 2-thermosite N/O Jean-Charcot, decembre 1980). Document du BRGM, no. 63, pp. 261–287.
P. Calcagno et al. / Tectonophysics 526–529 (2012) 185–195 Bouysse, P., Mascle, A., Mauffret, A., Mercier De Lepinay, B., Jany, I., Leclere-Vanhoeve, A., Montjaret, M.C., 1988. Reconnaissance des structures tectoniques et volcaniques sous-marines de l'arc récent des Petites Antilles. Marine Geology 81, 261–287. Calcagno, P., Chilès, J.P., Courrioux, G., Guillen, A., 2008. Geological modelling from field data and geological knowledge, part I — modelling method coupling 3D potentialfield interpolation and geological rules. Physics of the Earth and Planetary Interiors 171, 147–157. Deplus, C., Le Friant, A., Boudon, G., Komorowski, J.C., Villemant, B., Harford, C., Ségoufin, J., Cheminée, J.L., 2001. Submarine evidence for large-scale debris avalanches in the Lesser Antilles Arc. Earth and Planetary Science Letters 192, 145–157. Duperret, A., 1991. Étude structurale de la zone Monserrat–Marie-Galante (Petites Antilles) : Analyse des levés sismiques de la campagne FAGUAD. Mémoire DEA, Univ. Bretagne occidentale, Brest — France, pp. 1–74. Fabriol, H., 2001. Champ géothermique de Bouillante : synthèse des études géophysiques. BRGM/RP-50259-FR, 43 pp., 17 Fig. Fabriol, H., Bitri, A., Bourgeois, B., Debeglia, N., Genter, A., Guennoc, P., Jousset, P., Miehe, J.M., Roig, J.Y., Thinon, I., Traineau, H., Sanjuan, B., Truffert, C., 2005. Geophysical methods applied to the assessment of the Bouillante geothermal field (Guadeloupe, French West Indies). Proceedings World Geothermal Congress 2005, Antalya, Turkey, pp. 24–29. April 2005, 6 pp. Feuillet, N., 2000, Sismotectonique des Petites Antilles. Liaison entre activité sismique et volcanique, PhD Thesis, Univ. Paris VII — Denis Diderot Univ., France, 1–283. Feuillet, N., Manighetti, I., Tapponier, P., 2001. Extension active perpendiculaire à la subduction dans l'arc des Petites Antilles (Guadeloupe, Antilles Françaises). Comptes Rendus de l' Academie des Sciences Paris 333, 583–590. Feuillet, N., Manighetti, I., Tapponier, P., Jacques, E., 2002. Arc parallel extension and localization of volcanic complexes in Guadeloupe, Lesser Antilles. Journal of Geophysical Research 107 (B12), 2331. Finn, C.A., Morgan, L.A., 2002. High-resolution aeromagnetic mapping of volcanic terrain Yellowstone National Park. Journal of Volcanology and Geothermal Research 115, 207–231. Gadalia, A., Gstalter, N., Westercamp, D., 1988. La chaîne volcanique de Bouillante, BasseTerre de Guadeloupe, (Petites Antilles) : Identité pétrographique, volcanologique et géodynamique. Géologie de la France 2–3, 101–130. Grellet, B., Sauret, B., Bonneton, J.-R., 1988. Cadre général de la tectonique récente de la Guadeloupe. Rapport BRGM, 88SGN 627 GEG. Guennoc, P., Traineau, H., Castaing, C., 2001. Levé bathymétrique le long de la côte ouest de l'île de Basse-Terre, Guadeloupe. Apports à la compréhension du contexte structural du champ géothermique de Bouillante. BRGM R 40944. 30 pp. Henkel, H., Guzman, M., 1977. Magnetic features of fracture zones. Geoexploration (ISSN: 0016-7142) 15 (3), 173–181. doi:10.1016/0016-7142(77)90024-2 July 1977.
195
Kenedi, C.L., Sparks, R.S.J., Malin, P., Voight, B., Dean, S., Minshull, T., Paulatto, M., Peirce, C., Shalev, E., 2010. Contrasts in morphology and deformation offshore Montserrat: new insights from the SEA-CALIPSO marine cruise data. Geophysical Research Letters 37. Lajaunie, C., Courrioux, G., Manuel, L., 1997. Foliation fields and 3D cartography in geology: principles of a method based on potential interpolation. Mathematical Geology 29, 571–584. Mathieu, L., 2010, L'impact des mouvements décrochants sur la structure des volcans : une étude de cas des volcans de la Guadeloupe, de Maderas et du Mont Cameroun. Thèse des Univ. de Dublin et de Clermont-Ferrand. Mathieu, L., de Vries, W., Pilato, M., Troll, V.R., 2011. The interaction between volcanoes and strike–slip, transtensional and transpressional fault zones: analogue models and natural examples. Journal of Structural Geology. doi:10.1016/j.jsg.2011.03.003. Samper, A., Quidelleur, X., Lahitte, P., Mollex, D., 2007. Timing of effusive volcanism and collapse events within an oceanic arc island: Basse-Terre, Guadeloupe archipelago (Lesser Antilles arc). Earth and Planetary Science Letters 258, 175–191. Sanjuan, B., Traineau, H., 2008. French West Indies: development of the Bouillante geothermal field in Guadeloupe. IGA News, Newsletter of the International Geothermal Association, pp. 5–9. July–September 2008, Quarterly no. 73. Sanjuan, B., Genter, A., Roig, J.-Y., Patrier, P., Beaufort, D., Bes de Berc, S., Brach, M., Foucher, J.C., 2004. Travaux de reconnaissance géologique et de prospection de gaz dans les sols appliqués à la région de Bouillante, en Guadeloupe. Rapport BRGM/RP-53445-FR. 61 pp., 2 tabl., 15 fig., 1 ann. Sanjuan, B., Lopez, S., Guillou-Frottier, L., Le Nindre, Y.-M., Menjoz, A., 2008. Travaux de recherche liés au projet GHEDOM-ADEME. Rapport final BRGM/RP-56432-FR. 214 pp., 75 fig., 9 ann. Thinon, I., Guennoc, P., Bitri, A., Truffert, C., 2010. Study of the Bouillante Bay (West Basse-Terre Island shelf): contribution of geophysical surveys to understanding of the structural context of Guadeloupe (French West Indies–Lesser Antilles). Bulletin de la Societe Geologique de France 181 (1), 51–65. Traineau, H., Sanjuan, B., Beaufort, D., Castaing, B.M.C., Correia, H., Genter, A., Herbrich, B., 1997. The Bouillante geothermal field revisited: new data on the fractured geothermal reservoir in light of a future stimulation experiment in a low productive well. 22nd Workshop Geothermal Reservoir Engineering, Stanford, CA. 27–29 January 1997. Truffert, C., Thinon, I., Bitri, A., Lalanne, X., 2004. Using MAGIS for geothermal application — Guadeloupe archipelago in French West Indies. Hydro International 8 (6), 55–57.