Tectonophysics 489 (2010) 210–226

Contents lists available at ScienceDirect

Tectonophysics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t e c t o

3D architecture of a complex transcurrent rift system: The example of the Bay of Biscay–Western Pyrenees Suzon Jammes a,⁎, Christel Tiberi b, Gianreto Manatschal c a b c

University of Bergen, Department of Earth Science, Allegaten 41, N-5007, Bergen, Norway Géosciences Montpellier UMR5243, Université Montpellier 2, Place E. Bataillon, 34095 Montpellier Cedex 5, France Institut de Physique du Globe de Strasbourg/EOST, Université de Strasbourg, 1 rue Blessig, F-67084 Strasbourg, France

a r t i c l e

i n f o

Article history: Received 20 July 2009 Received in revised form 30 March 2010 Accepted 19 April 2010 Available online 24 April 2010 Keywords: 3D structure of rifted basins Gravimetric inversion Seismic interpretation Bay of Biscay Western Pyrenees

a b s t r a c t The Parentis and Arzacq–Mauléon basins located in front of the V-shaped oceanic propagator in the Bay of Biscay present evidence for extreme crustal thinning. In this paper we investigate the 3D structure of these rift basins, based on field observations and the interpretation of seismic data. We compare these results with those obtained from two different and independent inversion methods: first a 3D gravity inversion and second the standard Euler deconvolution. For the Mauléon Basin our results show that the positive gravimetric anomaly identified above its southern margin is the consequence of two shallower high density bodies that are separated by the Pamplona fault and a deeper high density body. The high density bodies can be explained by the presence of mid-crustal and mantle rocks that were exhumed or uplifted at shallower depth during Early Cretaceous rifting before they were reworked and integrated to the Pyrenean chain during compression phase. Also, during this reactivation phase, some slices of the exhumed mid-crustal and mantle rocks were sheared off and were integrated in the present-day thrust belt in the Mauléon basin. For the Parentis Basin we can demonstrate, based on seismic data and gravimetric inversion methods, a decrease in extension from west to east, which is compatible with the V-shape geometry of the overall basin. Along strike, a change in the fault geometry from downward concave top-basement detachment faults to upward concave high-angle faults can be observed eastwards, i.e. towards the termination of the basin. A key structure, controlling the evolution of the Parentis Basin, is the east–west trending Ibis fault. We interpret this fault to have initially formed as a strike slip fault before it was reactivated during later crustal thinning. At present, it forms the limit between an upper plate sag basin to the north and a lower plate sag basin, floored at least locally by a top-basement detachment faults to the south. The strong asymmetry of the basin is supported by the shape of the basin and the results of standard Euler deconvolution. © 2010 Elsevier B.V. All rights reserved.

1. Introduction The structures observed in front of propagating oceans such as in the Woodlark basin (Taylor and Huchon, 2002) or the Gulf of California (Nagy and Stock, 2000) bear a lot of information of how rifts evolve into oceans. However, most of the present-day active systems, except the Red Sea and the Gulf of Aden, are related to subduction systems and hence do not represent, in a strict sense, Atlantic type rift systems. The Bay of Biscay located between France and Spain is one of the rare examples where rift structures are imaged and drilled (e.g. the Parentis Basin) and, due to its inversion along the southern margin, also exposed on land (e.g. southern Arzacq– Mauléon Basin) at the termination of V-shaped oceanic basins. Although single rift structures are exposed or drilled, the overall rift

system is hidden beneath an up to 10 km thick sedimentary succession. This is one of the reasons why the partitioning of the deformation on the scale of the entire system is difficult to image. In this study we use 3D inversion and standard Euler deconvolution of gravity data to investigate the crustal structure and their distribution in front of a propagating ocean. When combined with field observations, drill hole and seismic reflection data, this approach can reduce the non-uniqueness of the mathematical solutions, and thus provide interesting information on the density structure and distribution of major density contrasts. Consequently, it gives some insights into the crustal architecture within the study area. The aim of the paper is to describe the 3D structure of the overall rift system situated in front of the V-shaped oceanic domain in the Bay of Biscay. 2. Geological setting

⁎ Corresponding author. E-mail address: [email protected] (S. Jammes). 0040-1951/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.tecto.2010.04.023

The Bay of Biscay is a V-shaped propagating oceanic domain located between the Armorican and Cantabrian margins separating

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

France from Spain (Fig. 1a). Rifting affected a strongly pre-structured domain and was associated with left lateral movements between Iberia and Europe. In order to take into account all new observations made in the Iberia–Newfoundland rift system, i.e. the discovery of at least 300 km of exhumed mantle and hyper-extended crust and the reinterpretation of the M-series magnetic anomalies in the southern North Atlantic, Jammes et al. (2009) proposed a kinematic model for the Iberian plate which is similar to the one proposed by Wortmann et al. (2001). In contrast to the classical interpretations of Olivet (1996), Jammes et al. (2009) proposed that the left lateral displacement between Iberia and Europe had to initiate already before Aptian time along a diffuse plate boundary located in the Bay of Biscay–Pyrenean domain, which is compatible with the observed major pre-breakup extension observed in the Iberia–Newfoundland rift system. Although this stage is not well documented, due to its subsequent overprint, evidence for the occurrence of pre-Aptian basins within the Pyrenean domain have been reported (ex: Cameros basin: Platt, 1990; CasasSainz and Gil-Imaz, 1998; Mata et al., 2001). An example for such left lateral transcurrent segments would be the Ibis fault crossing the Parentis basin (Bois et al. 1997; Biteau et al., 2006). In Late Aptian time, a change from left lateral transcurrent to orthogonal extension resulted in the reactivation of these transcurrent fault segments as orthogonal normal faults leading to the formation of the Parentis and Arzacq–Mauléon basins that formed simultaneous to continental breakup in the Bay of Biscay. The change in the overall plate kinematics is well documented by the formation or reactivation of NE–SW trending transform faults (e.g. Pamplona fault, Claude, 1990; Larrasoaña et al., 2003) that are kinematically linked to exhumation in the Parentis and Arzacq–Mauléon basins (Fig. 1b) (Jammes et al., 2009). Both basins are characterized by a remarkable positive gravimetric anomaly (Daignières et al., 1981, 1982; Pinet et al., 1987; Casas et al., 1997, e.g. Fig. 5). The Parentis Basin is a 100 km wide weakly reactivated offshore basin bordered to the north by the Armorican shelf and to the south by the Landes high (Fig. 1b). The geometry of the Parentis basin is well imaged in the ECORS-Bay of Biscay profile (Bois et al., 1997) and East– West Expanding Spread profiles (ESPs) (Marillier et al., 1988). These data show that extreme crustal thinning to less than 10 km occurred over a zone of 60 km under the center of the basin, explaining the positive gravimetric anomaly (Pinet et al., 1987). The Arzacq–Mauléon basin is exposed in the northwestern foreland of the Pyrenees between Biarritz and Pau. A restoration of the ECORS-Arzacq deep seismic profile suggests that major crustal thinning occurred during Early Cretaceous time (Daignières et al., 1994). This basin was inverted as a pop-up structure during Pyrenean compression, with south verging thrust faults to the south (e.g. the Lakora thrust) (Teixell, 1996, 1998) (Fig. 1b) and the north vergent North Pyrenean thrust system to the north (e.g. Saint Palais and Sainte Suzanne thrusts faults; Figs. 1b and 2). Daignières et al. (1994) interpreted the Sainte Suzanne thrust as the present-day boundary between the southern Mauléon and the northern Arzacq basins. In this study we demonstrate, based on field and seismic observations and supported by gravimetric modelling that the Saint Palais thrust constitutes a key boundary that reactivated a major rift structure. Therefore we consider this structure as the boundary separating the Arzacq basin from the Mauléon basin (Fig. 2). This structure separates two very different basins: the Arzacq basin to the north formed by up to 10,000 m of lower and upper Cretaceous series overlying pre-rift sediments (Vergés and García Senz, 2001) and the Mauléon basin to the south, presenting a more complex structure that will be discussed in this paper. Several interpretations have been proposed to interpret the crustal structure within the ECORS-Arzacq profile. Daignières et al. (1994) suggested that the southern termination of the basin was reactivated as a crustal ramp that rooted in the mantle and coincided with the place where the Iberian crust was subducted underneath the

211

European crust. On the contrary, Muñoz (2002) suggested that the European crust formed an indenter that separated the Iberian upper crust from its lower crust and upper mantle. While the upper crust was thrust south as well as northwards, forming an orogenic lid, the remainder of the Iberian crust was subducted underneath the European margin. A major question is whether the positive gravimetric anomaly recorded along the southern Mauléon margin is, as proposed by Pedreira et al. (2007) related to the emplacement during compression at shallower depth or at the seafloor of lower crustal rocks already uplifted during crustal extension, or, as suggested in this study, related to the exhumation at the seafloor of crustal material during Early Cretaceous extension and kept in place during inversion. 3. Field observations in the southern Mauléon basin During Pyrenean compression the southern margin of the Arzacq– Mauléon basin was inverted and most of the rift-related structures were reactivated and either buried below Tertiary sediments or exposed and eroded in the Pyrenees. Rift-related structures are well preserved in the Basque massifs and the Béarnais ranges (from north to south: Mail Arrouy, Sarrance and Layens) and the Arbailles (Fig. 2). A description of these structures can be found in Jammes et al. (2009) and the key observations are summarized below. In the Labourd massif, situated at the western termination of the Mauléon basin (Figs. 1b and 2), Paleozoic upper crustal rocks and granulites are sealed by Albian to Santonian sediments. Our field observations made in this area show that granulites belonging to the pre-rift middle crust were exhumed along an extensional detachment system at the seafloor and were partly eroded and redeposited in Albo-Cenomanian sediments (e.g. Bonloc Breccia; Claude, 1990). Evidence for mantle exhumation at the seafloor within the Mauléon Basin comes from the Urdach area (Fig. 2). The occurrence of sedimentary breccias containing clasts of crustal and mantle rocks over exhumed mantle (e.g. Urdach, Fig. 2) implies that mantle rocks had to be exhumed at the seafloor and may have locally floored the Mauléon basin. Due to the close neighborhood of the Upper Triassic to Lower Cretaceous fault blocks in the Béarnais ranges (Fig. 2) and mantle rocks at their base, as well as based on structural arguments, Jammes et al. (2009) interpreted these blocks as extensional allochthons overlying exhumed mantle. Another important structural elements are N40° to N60° directed transfer faults that segment the basin (Peybernès and Souquet, 1984). Claude (1990) and Larrasoaña et al. (2003) showed that at least one of these transform faults associated with the Pamplona fault was active during the deposition of Cretaceous sediments. During Pyrenean reactivation, the extensional detachment system responsible for the exhumation of deeper crustal and mantle rocks was reactivated in many places as a thrust fault. Where the thrust faults incised into the footwall, pieces of the previous footwall, consisting of exhumed crustal or mantle rocks, were thrust together with their cover sequence over Upper Cretaceous post-rift sediments. During the final stage of convergence in Oligocene to earliest Miocene times, the major shortening occurred outside the former rift basin and affected the domain located south of the Mauléon basin (Jammes et al., 2009). 4. Interpretation of seismic reflection data from the Parentis basin In the ECORS-Bay of Biscay section, the Parentis Basin was originally interpreted as a synclinal basin affected by little normal faults and floored by continuous Triassic and Jurassic sediments (Bois et al., 1997). Using new data issued from the MARCONI campaign, Ferrer et al. (2008) proposed that the basin is formed by a major halfgraben bounded by a major north-dipping high-angle fault. None of the two interpretations can, however, explain the extreme observed crustal thinning. Based on the available data we propose a new

212 S. Jammes et al. / Tectonophysics 489 (2010) 210–226 Fig. 1. a: Map of the Bay of Biscay and Pyrenees displaying the major structures identified in the area and the different domains situated between the continental and the oceanic domains in the Bay of Biscay and western Iberia margin. Magnetic anomalies proposed by Sibuet et al. (2004) are also reported for the oceanic part (HF: Hendaya fault; PF: Pamplona fault; TF: Toulouse fault). b: Location map of the Parentis basin and the Labourd–Mauléon area showing the major tectonic structures referred to in the text. Red lines show the location of the ECORS-Bay of Biscay reflection seismic profile, the Aquitaine coast profile and the ECORS-Arzacq profile. Wells are indicated by violet triangles.

S. Jammes et al. / Tectonophysics 489 (2010) 210–226 Fig. 2. Simplified geological map of the Labourd–Mauléon area modified after the BRGM 1/50,000 geological maps of France (maps of Bayonne (n° 1001), Hasparren (n° 1002), Orthez (n° 1003), Espelette (n° 1026), Iholdy (n° 1027), MauléonLicharre (n° 1028), Pau (n° 1029), Saint Jean Pied-de-Port (n° 1049), Tardets-Sorholus (n° 1050), Oloron-Ste-Marie (n° 1051), Lourdes (n° 1052), Larrau (n° 1068), Laruns-Somport (n° 1069), Argelès-Gazost (n° 1070), Gavarnie (n° 1082)). The cross section shown in the inset is a restored section of the ECORS-Arzacq reflection seismic interpretation (Jammes et al., 2009) showing the situation in Albian–Cenomanian time. Major field relations described in the text are shown in the section.

213

214

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

interpretation which can take into account the observed extreme crustal thinning. The ECORS-Bay of Biscay profile shows an important asymmetry of the two conjugate margins, which are separated by a deeply rooted vertical fault referred to as the “Ibis fault” (Figs. 3a and 4c). This structure is associated to an east–west alignment of salt ridges: the Ibis, Eridan and Antares ridge (Biteau et al., 2006). We interpret this structure as a remnant of a former strike slip system along which the left lateral displacement of Iberia relative to Europe was accommodated. The Ibis fault separates the basin into two very different parts. To the north (CDP 4300-2200), drill hole data in combination with seismic reflection data show that Jurassic to Albian sequences, characterized by a set of strong continuous reflections, are tilted southwards and thicken towards the Ibis fault (Fig. 3a). We described this geometry as a sag geometry. In this part of the ECORS-Bay of Biscay profile, normal faults are rare, but seem to become more abundant eastwards, on profiles situated along the Aquitaine coast (Fig. 3b). These high-angle faults affect Lower Aptian sediments and are generally sealed by Upper Aptian to Albian sediments (Biteau et al., 2006; Serrano et al., 2006). To the south the margin is strongly affected by diapirism and is consequently less well imaged (CDP 5300-4300). However, where structures are imaged, they appear more complex. Based on the resemblance of the northern Parentis and Arzacq basins, both showing large and deep depressions filled by lower and upper Cretaceous series (Fig. 3a and section in Fig. 2), we assume that these basins are the result of a similar tectonic evolution. As a consequence, we assume that the two southern extensions of these basins, i.e. southern Parentis and Mauléon may be similar as well. This is supported by the occurrence of a positive gravity anomaly over these basins. Assuming that these basins may have evolved in a similar way, we explored for the possibility that the southern Parentis basin could be floored, like the Mauléon basin, by a top-basement detachment system. However, since top-basement detachment faults cannot be observed directly in seismic sections, their interpretation needs to be based on other arguments. In a NW–SE profile crossing the southern border of the Parentis basin to the east of the ECORS-Bay of Biscay profile (73BY13, Fig. 4a), we observe strong reflectors between 2 and 3 s TWT in the basement that are tilted towards the southeast (continentwards), truncated by the top-basement reflection and onlapped by Cretaceous sediments. This relationship is comparable to those described from the Labourd massif by Jammes et al. (2009), where upper crustal meta-sediments (likely to be very reflective) are juxtaposed against granulitic middle crust, capped by a top-basement detachment fault and sealed by Cretaceous sediments. Thus, in analogy to this interpretation, we interpret the top-basement reflection as an exhumation fault along which Paleozoic metasediments were exhumed and onlapped by Cretaceous sediments (Fig. 4a). Another evidence for the occurrence of a top-basement detachment faults can be found in the Aquitaine coast profile (Fig. 3b). Along this line, the SGM and CTS wells penetrated beneath the salt into a strongly deformed basement formed by cataclasites and gouges. This observation is compatible with the existence of a top-basement fault, which is not resolved in the seismic section because of the Triassic salt structures (for further details see Jammes, S. et al., submitted for publication). Also, based on these indirect evidence we suggest that the southern margin of the Parentis basin is locally floored by tectonically exhumed basement. In the central part (CDP 4400-4700), the Albian anticlinal structure described by Masse (1997) is

interpreted as a pre-rift sequence overlying the detachment structure and affected by gravitational deformation after extension ceased (see Jammes, S. et al., submitted for publication). Evidence for other extensional structures exists over the Landes high to the south and southwest. This is well documented on Line 75BY13 (Fig. 4b) in which some sets of layered, southward tilted reflections bounded by normal faults can be observed. Moreover, also over the Landes high, Lefort et al. (1997) highlighted the existence of strong reflectors in the seismic basement in the ECORSBay of Biscay presenting evidence for abrupt changes in the tilting (CDP 7900-5300). They interpreted them as the result of major tilting related to pre-Mesozoic deformation events. To explain these geometries, we suggest the existence of two more extensional detachment systems in this area (dashed violet lines in Fig. 3a). In mapping these structures eastwards on profile 75BY13 (Fig. 4b), it can be shown that these detachment faults change along strike and become high-angle normal faults. A similar observation showing that high-angle faults may develop laterally into low-angle detachment faults was shown by Péron-Pinvidic et al. (2007) in the IberiaAbyssal Plain. In conclusion, the Parentis and Arzacq–Mauléon basins show many similarities in the overall, large-scale architecture of the basin. Strong similarities exist between the northern Parentis and Arzacq basins while correlations between the southern Parentis and Mauléon basins are more difficult to establish, because of the inversion of the latter. However, the observed top-basement detachment faults in the Mauléon basin may correspond and be analogue to top-basement detachment faults interpreted in the southern Parentis basin. To the south of this basin over the Landes high, detachment faults can also be identified along ECORS-Bay of Biscay profile. These structures become high-angle faults going eastwards, which is compatible with the observed decrease of extension and termination of the basin in the same direction. 5. Gravity modelling A compilation of onshore and offshore gravity data issued from the BGI (Bureau Gravimetric International) associated with values obtained in the Bay of Biscay from the satellite-derived database of Sandwell and Smith (2009) are used for this study. The complete Bouguer terrain corrections (Fig. 5a) are calculated from simple Bouguer anomalies and digital elevation and bathymetric data using the Cogbill (1990) method. In this calculation, the density used for the land and water masses are respectively 2.67 g cm− 3 and 1.03 g cm− 3. In the study area the complete Bouguer gravity anomaly mainly comes from two different sources: the Moho topographic variations and the crustal heterogeneities. We will make the assumption that short wavelengths are mainly the expression of small-scale crustal heterogeneities, whereas the longest wavelengths mainly result from the regional component (mainly variations in the Moho topography). To independently study both causes, we separate wavelengths greater than 185 km (Fig. 6a) from the short wavelengths (Fig. 5b). First, we use the Oldenburg inversion scheme (Oldenburg, 1974) on the largest wavelength anomalies to extract information on the deep crustal structure. Then, the shortest wavelength anomalies are inverted using two different and independent methods: a 3D gravity inversion and the standard Euler deconvolution. The first one allows us to map the distribution of density contrasts in a 3D volume. The second one allows us to map and localize the depths of source edges that can be at

Fig. 3. Interpretation of seismic reflection profiles from the Parentis basin (for location of the sections see Fig. 4c. a) ECORS-Bay of Biscay seismic profile without interpretation (top) and with a new interpretation (bottom); the line-drawing of deep structures was made by Lefort et al. (1997) b) Aquitaine coast profile seismic section (from Serrano et al. 2006) without interpretation (top) and with a new interpretation (bottom): we observe an upper plate sag basin to the north of the Parentis ridge and a more complex structure affected by salt diapirs to the south. In this southern part we can observe tilted blocks underlying a strong reflector which was drilled and identified to coincide with a decollement in Triassic salt (Jammes S. et al., submitted for publication). Our interpretation suggests that a detachment fault floored both the southern part of the Parentis basin and the northern part of the Landes high. This structure was responsible for the exhumation of deeper crustal levels beneath the salt layer while the upper sedimentary cover was deformed by tilted blocks and salt diapir.

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

215

216

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

Fig. 4. Interpretation of seismic reflection profiles from the Landes High. a) Seismic line 73BY13 (from Total) without interpretation (top) and with proposed interpretation (bottom) showing basement reflectors that are tiled and truncated by the top-basement reflector. This structure is interpreted as a top-basement detachment fault; b) Seismic line 75BY13 (from Total) without interpretation (top) and with proposed interpretation (bottom) showing normal faults over the Landes High; c) location map of all seismic sections used to draw the 3D block of the Parentis basin (see Fig. 11). The seismic sections in red are those presented and interpreted in this paper.

the origin of the gravimetric anomaly. These two methods were applied on 61101 data points, situated between 2.65°W and 0.057°W in longitude and 42.71°N and 45.22°N in latitude. 5.1. Oldenburg inversion To image crustal thickness variations in the studied area we use the approach of Oldenburg (1974). This method is based on the direct formula of Parker (1972) that calculates the gravity signal of a 2-D uneven layer with a constant density:

−jkjz0

F ðΔg ðxÞÞ = −2πGΔρe

∞

∑

n=1

jkjn−1 ! n " F h ðxÞ : n!

Here F indicates the Fourier Transformation, g(x) is the gravity signal, k is the wave number, x is the horizontal coordinate, G is Newton's gravitational constant, z0 is a reference depth, n is an integer

and ∆ρ is the density contrast between the two media. The function h (x) represents the topography of the interface. Oldenburg (1974) solved the inverse problem by an iterative approach, and thus he retrieved the layer's thickness from the gravity signal. In our case, we choose n = 4 which is far enough to find an adequate solution. The problem is typically non-unique and requires one to fix first a reference depth z0 from which the topographic variation h(x) is calculated and second a constant density contrast ∆ρ between the two layers. The additional use of a cosine taper during the inversion helps the inversion to stabilize (Oldenburg, 1974). We use here a reference depth of 30 km and a density contrast of 530 kg m− 3 that corresponds to the difference between a crust of 2670 kg m− 3 (value used for the Bouguer anomaly calculation) and a mantle of 3200 kg m− 3. The mean Moho depth (30 km) is taken from previous studies (e.g. Lefort and Agarwal, 2002). Of course, the obtained results are only relative results, because the amount of crustal thinning and thickening is strongly dependent on the choice of

S. Jammes et al. / Tectonophysics 489 (2010) 210–226 Fig. 5. a) Unfiltered complete Bouguer anomaly map of the studied area based on BGI (Bureau Gravimetrique international) database; b) filtered complete Bouguer anomaly map of the studied area; only wavelengths lower than 185 km are displayed. Note on the southern border of the Parentis basin the remaining oval shaped positive anomaly; c) calculated complete Bouguer anomaly map of the studied area from the 3D inversion model presented in this paper. The comparison between calculated and filtered data reveals few differences, mainly coming from very short wavelengths unable to be taken into account by our parametrization (blocks of 10 km). Note that the principal features are very well retrieved. Major tectonic features and seismic lines are displayed to ease the comparison with the other figures.

217

218

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

thickness. However, it is interesting to use this method on large wavelength anomalies (wavelengths greater than 185 km) in order to estimate what the deep crustal information removed in our filtered data (wavelengths smaller than 185 km) are, and to get a first approximation on the Moho variations. The results are shown in Fig. 6. Convergence was obtained after five iterations only, with a final Root Mean Square (RMS) of 1.15 mGal, compared to the initial 4.76 mGal. The gravity signal calculated (Fig. 6b) from the resulting topography (Fig. 6c) is extremely close to the initial one. On the Moho topographic map we can see that in the Parentis basin the large wavelength anomaly is associated to the Moho uplift under the central part of the basin as suggested by Pinet et al. (1987). Concerning the Labourd–Mauléon area, these results show that the large wavelength anomaly characterizes a steep variation of the Moho depth that becomes deeper southward. This topographic variation of the Moho is also well observed in ECORS-Arzacq profile (Grandjean, 1992; Daignières et al., 1994). Westward this topographic variation is NNE–SSW oriented which corresponds to the orientation of the Pamplona fault, a structure that will be discussed below. 5.2. Filtered complete Bouguer anomaly map The previous study demonstrates that in filtering the largest wavelengths, the regional component of the topographic Moho variation is suppressed from the gravimetric signal. This would suggest that the remaining anomalies on the filtered complete Bouguer anomaly map mainly result from small-scale crustal heterogeneities (Fig. 5b). In the Parentis basin these crustal heterogeneities create a small-scale positive anomaly centered above the southern border of the Parentis basin (Fig. 5b). This anomaly seems to be relayed westwards to the north by a similar anomaly observable on the western border of the map at about 44.5°N. To the south of this positive anomaly, above the Cantabrian shelf, a strong negative anomaly is present. Concerning the anomaly located above the Mauléon–Labourd region, the filtering proceeded on the data does not fundamentally modify the shape and the location of the anomaly (compare Fig. 5a and b). The resulting anomaly is centered on the southern Mauléon basin, whereas no positive anomaly is observed above the northern Arzacq basin. 5.3. 3D density model

Fig. 6. a) Initial large wavelength (N 185 km) gravity signal; b) retrieved signal from Oldenburg inversion; c) Moho topography variation obtained from the Oldenburg inversion for a reference depth equal to 30 km.

the reference depth and the density contrast (Oldenburg, 1974). Moreover this method does not take into account the lateral changes in the density contrast that will be mapped into variations of crustal

To account for the lateral changes in density contrast within the study area, we perform a 3-D inversion of the filtered gravity data (wavelength smaller than 185 km). The aim of the inversion is to obtain a distribution of density variations in a 3-D model space that can fit the observed gravity data in a least-squares sense. The causative sources of the local complete Bouguer gravity anomaly are assumed to be distributed in N horizontal layers of thickness Hi, i = 1, 2…, N, in the crust. Every layer is subdivided into a number of rectangular blocks in which an initially homogeneous density contrast is assigned. In this method, the average density contrast and gravity anomaly are both assumed to be zero (see Widiwijayanti et al. (2004) for a detailed presentation of the method). The final density contrasts retrieved by the inversion are then relative to each layer, and cannot be directly compared between different depths. In order to exclude possible boundary effects, the area from which data are inverted exceeds the study area by about 30 km in each direction (Fig. 7). For the same reason, the used model is even larger. Depending on the tested parameterization, four, five or six layers constitute the model down to 45 km depth. Each layer is equally divided into 10 × 10 km blocks in the central study area, and with larger blocks on the edges (Fig. 7). Each block is also associated with a standard deviation that accounts for an a priori information: by increasing or decreasing this parameter for a given block we allow the inversion to put higher or lower density contrasts in this block, respectively. We start the inversion with a homogeneous

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

density contrast model. We are fully aware of the difficulty for a gravity inversion to properly retrieve the depth location of an anomaly without any more information. Our seismic constraints cannot be properly input as a priori information because of the lower resolution we have in the gravity inversion (10 × 10 km blocks). In order to investigate this question, we thus carried out several inversions with various layers with different depths and widths. Within all our models, the Root Mean Square (RMS) decreases by about 86–87%, and the retrieved Bouguer anomaly is very close to the observed one (Fig. 5b and c), with only differences for the very short wavelengths. These very short wavelengths are principally expressed in the most superficial layer (0–4 km) and come from discrepancy between the signal shortest wavelength and our block size. This layer will be consequently not discussed in this paper. We will concentrate our discussion on the deeper levels, where major structures remain coherent for all the models. Our preferred model is presented in Fig. 8. It is a six layer model (layer 1: 0–4 km, layer 2: 4–16 km; layer 3: 16–19 km; layer 4: 19–26 km: layer 5: 26–30 km and layer 6: 30–45 km) where standard deviation is set to 0.1 g cm− 3 for layers 1 to 5 and set to 0.01 for layer 6. We decrease this parameter in the deepest layer to prevent the apparition of artifacts at deeper levels. In all our models, the positive gravimetric anomaly observed on the southern border of the Parentis basin is interpreted by the presence of a well marked but small-scale positive density contrast between 4 and 16 km. To the south, the negative gravimetric anomaly of the Cantabrian shelf is interpreted by a high amplitude negative density contrast also well marked in all models between 4 and 16 km. Concerning the Labourd–Mauléon anomaly, all models suggest that this positive anomaly is created at shallower depth (between 4 and 16 km) by a high density body and a deeper large-scale density contrast located between 19 and 26 km. Between 4 and 16 km the high density body can be separated into two different parts: an oval-shape body to the east and westward and southward a smaller scale body bounded to the east by a NNW–SSE oriented structure. The easternmost body was hereafter referred to as the Labourd–Mauléon body and the westernmost body as

Fig. 7. Model used for the 3-D gravity inversion. The six layers are gridded into rectangular blocks shown by the light black lines. The thin black lines represent the coast lines. The dashed rectangle represent the area of data inverted and the inside bold rectangle represent the studied area.

219

Cinco-Villas body. The NNE–SSW oriented structure bounding the Cinco-Villas body is well identified on the southern border of the model at about 1.5°W. This structure splits a high density body situated to the west, beneath Cinco-Villas massif (Figs. 1 and 6) from a small-scale low density body to the east (Fig. 8). We link this structure to the Pamplona fault (Fig. 5) and will discuss it later. The second body responsible for the positive gravimetric anomaly of the Labourd–Mauléon area is identified at shallower depth between 5 and 15 km and is located NE from the Pamplona pattern (Fig. 8). At a deeper level, the density contrast corresponding to the Cinco-Villas anomaly disappears meanwhile under the Mauléon basin, density contrast is first attenuated but increases again between 19 and 26 km to create a large-scale density contrast. 5.4. Standard Euler deconvolution method The standard Euler deconvolution method is based on Euler's homogeneity equation that relates the potential field (gravimetric or magnetic) and its gradient components to the location of the source, with the degree of homogeneity n which may be interpreted as a structural index (SI) (Thompson, 1982; Blakely, 1996). For a three dimensional function f(x, y, z) the Euler's equation is (where n corresponds to the degree of homogeneity): x

∂f ∂f ∂f +y +z = −nf : ∂x ∂y ∂z

Considering potential field data, Euler's equation can be re-stated as follow: ðx−x0 Þ

∂T ∂T ∂T + ðy−y0 Þ + ðz−z0 Þ = −NðB−T Þ ∂x ∂y ∂z

where (x0, y0, z0) is the position of the source whose total field T is measured at (x, y, z), B is the regional value of the total field and N is the structural index, which is a measure of the rate of change with distance of a potential field. Given a set of observed total field data, this method determines an optimum source location (x0, y0, z0) by solving Euler's equations for a given index N by least-squares inversion of the data. This method operates on gridded data and solves Euler's equation simultaneously for each grid position within a square window that moves along each grid row. The inversion process will also yield an uncertainty (standard deviation) for each of the fitted parameters. If the calculated depth is less than a specified tolerance (typically 15%) and the solution is within a limiting distance of the center of the data window, the solution is accepted. We inverted the gravity data with the Oasis Montaj Geosoft software. We tested the parameters by processing several inversions with different structural indexes (N = 0, N = 0.5 and N = 1) and window sizes (6 km, 14 km, 20 km, 24 km, 30 km, 36 km and 40 km). A structural index equal to 0 represents a source with a planar geometry (fault, dyke, sill…), a structural index equal to 0.5 can be associated to a fault with a certain width or finite dimension in one axe and a structural index equal to 1 corresponds to a cylinder. Also, the solution obtained will localize the edges of these planar or cylindrical sources. Some results obtained with this method are shown in Fig. 9, with different values of structural index (N) and different window sizes. For N = 0, all solutions are displayed meanwhile for N = 0.5 or N = 1 only solutions with error in depth location less than 10% are displayed. By superimposing these results on the simple Bouguer gravimetric map, we check that the Euler sources coincide well with the most important gravimetric gradients. Moreover, to identify the structures marked by the alignment of our Euler sources, we superimpose to these results the main geological features. In all inversions, calculated sources are situated between 0 and 13 km depth but as usually observed with this method, those

220

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

Fig. 8. 3-D density distribution within each of the 6 layers constituting the model issued from our 3D inversion. The standard deviation (σ for each layer is indicated in the lower right corner). Note the two high density bodies creating a part of the Labourd–Mauléon anomaly and the Parentis anomaly between 4 and 16 km depth and the high density body between 19 and 26 km depth under the Mauléon basin.

calculated depth have to be considered as estimations rather than absolute values. For N = 0 (Fig. 9a and b), we observe in the western Pyrenean domain that for both window sizes sources line up with major structural features. Indeed, north of the Mauléon basin, sources coincide with the Saint Palais thrust in the western part of the basin and the North Pyrenean thrust in the eastern part of the basin. Moreover, to the south these sources coincide with south verging thrust system located along the southern margin of the Mauléon basin (e.g. Arbailles thrust). We can notice that with a window size equal to 36 km the North Pyrenean thrust seems to be less well localized. In this case the window size is too big and significant effects from multiple sources are included and interfere with the localization of the sources. In the Parentis basin, solutions are only observable with a window size equal to 14 km. In this result, superficial Euler sources underline a SW–NE oriented structure that does not correspond to the southern border of the basin as defined by Bois et al. (1997) (Fig. 9f) and a N–S

oriented structure located at about 1.6°W between 43.9°N and 44.2°N. In the north-eastern part of the basin, Euler sources align with the salt ridges defined as the eastern termination of the Ibis fault. For N = 0.5 (Fig. 9c and d) and N = 1 (Fig. 9e), the results are very poor in the Mauléon area justifying that displayed structures are planar and similar to faults. In the Parentis area, the previously described structures are more or less also outlined with N = 0.5 and N = 1. This observation suggests that in the Parentis basin sources are not exactly planar structure such as faults with two infinite dimensions, and can be associated to either large faults, or blocks. 6. Discussion and conclusion 6.1. Labourd–Mauléon gravity anomaly The 3D inversion method suggests that the anomaly of the Labourd–Mauléon area is created by the presence of a high density body located between 4 and 16 km depth and a deeper high density

Fig. 9. Superposition of Euler sources obtained in this paper for different structural index (N) and several window sizes (W) on the complete filtered Bouguer gravimetric map to verify that the Euler deconvolution calculated sources coincide well with the most important gravimetric gradients; Major tectonic structures and seismic profiles are displayed to ease the interpretation and comparison of figures. a) N = 0, W = 14 km; b) N = 0, W = 36 km; c) N = 0.5, W = 20 km, d) N = 0.5, W = 30 km; e) N = 1, W = 20 km; f) superposition of Euler sources obtained for N = 0 and W = 14 km on a geological and structural map: the Euler sources line up with major structural features, such as Saint Palais thrust.

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

221

222

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

Fig. 10. Comparison of our results (density and Euler sources) with geological interpretations. a) observed (plain) and calculated (dashed line) Bouguer anomaly along the ECORS-Arzacq profile. The calculated data correspond to the gravimetric anomaly obtained with the density model shown in Fig. 5. b) Vertical slice through our density model along the ECORS-Arzacq profile (Fig. 5). The main structures issued from the above geological cross sections (Moho, top of the lower crust, top of the basement) are superimposed. c) Geological sections made along ECORS-Arzacq profile. For the external parts of the orogen this interpretation used cross sections made by Teixell (1998) and Schellart (2002) for the Spanish part, and Ducasse and Velasque (1988) and Serrano et al. (2006) for the French part. For the internal parts, we show a new interpretation, which is based on our field observations. The lower crustal structure is based on the interpretation of the ECORS-Arzacq profile proposed by Daignières et al. (1994).

body located between 19 and 26 km (Fig. 8). The existence of this shallower body was previously proposed by Grandjean (1992), Casas et al. (1997) and Pedreira et al. (2007). The standard Euler deconvolution method suggests moreover that this body is situated in a domain delimited at the surface by the south verging thrust system to the south and the north verging Saint Palais thrust to the north (Fig. 9). The high density body responsible for the positive gravimetric anomaly is consequently located inside or below the popup structure resulting from the inversion of the Mauléon basin or under this inverted basin. To better visualize this result, we show a vertical slice in our density model along the ECORS-Arzacq profile (Fig. 10). We are totally aware that a direct comparison of the vertical density contrasts obtained by our 3D inversion is not perfectly correct (see e.g. Widiwijayanti et al., 2004), but such a vertical slice helps to visualize the localization of density contrast and allows comparison with geological sections. We compare our results with a geological section made along the ECORS-Arzacq profile (Fig. 10b and c). For the external parts of the sections we used interpretations proposed by Teixell (1998) and Schellart (2002) for the Spanish part, and Ducasse

and Velasque (1988) and Serrano et al. (2006) for the French part. For the internal parts, we show a new interpretation, based on new field observations described in Jammes et al. (2009). For the deeper crustal structure, we show the interpretation proposed by Daignières et al. (1994) of the ECORS-Arzacq profile even if the interpretation of Muñoz (2002) is also possible. At this stage, our gravimetric inversion lacks constraints at depth to fully advocate for one of these interpretations, and further geophysical studies have to be realized to better constrain the deep crustal structure of the Western Pyrenees. The comparison between the cross sections and the corresponding slice in our 3D density model allows us to discuss the origin of these high density bodies. The recent study realized in Mauléon basin shows that before Pyrenean compression, mid-crustal and mantle rocks were exhumed at the seafloor at the base of this basin (Jammes et al., 2009). During inversion of the basin, thrust faults reactivated on a first order the former rift-related detachment faults responsible for mantle exhumation. Wherever on a more local scale the thrust faults were unable to use the pre-existing detachment, they incised into the footwall. As a consequence, slivers of the underlying exhumed material (mantle or crustal rocks) were accreted together with the former hanging wall rocks (allochthons and post-rift sedimentary cover) in the orogenic wedge. This can explain the occurrence of mantle rocks at the base of Jurassic to Lower Aptian calcareous blocks in the southern Mauléon basin (Mail Arrouy, Sarrance and Layens). However such slices of mantle rocks are too thin to correspond to the high density body located between 4 and 16 km depth identified in this study. They only prove that mantle and lower crustal rocks were at the seafloor before convergence. We therefore think that this high density body located between 4 and 16 km depth is created by remnants of these previously exhumed crust or mantle rocks that were kept at shallower position during the development of main thrust faults resulting in the formation of the antiformal stack corresponding to the Axial zone. The presence of such bodies at shallow crustal level could moreover explain the attenuation of guided waves observed in this area (Chazalon et al., 1993; SensSchönfelder et al., 2009). Concerning the deeper high density body (located between 19 and 26 km), it can also be explained by a piece of lower crust or mantle rocks emplaced during rifting in a shallower position and subsequently reactivated during convergence or by a pronounced Moho topography resulting from rifting. In the latter case, this would suggest a short wavelength of the Moho topography that was not filtered.

6.2. Cinco-Villas anomaly and Pamplona fault The Cinco-Villas anomaly is characterized in all our 3D density models by a high density contrast bounded to the east by a NNW–SSE oriented structure which is associated at depth to a steep variation of the Moho topography (Fig. 6c). The Pamplona fault interpreted by Larrasoaña et al. (2003) as a transfer fault separating two segments of the Pyrenean rift, is a good candidate to explain such NNW–SSE structure. In Spain, this structure separates the western deep Biscay basin from the eastern shallower Ebro basin (Schoeffler, 1982; Vergés and García Senz, 2001; Vergés, 2003). To the north in France, a flip in basement depth can be observed across the fault. The deeper basin (the Mauléon basin) is located on the eastern side of the fault and the shallower basin (the Hendaya basin) on the western side. It is true that in our density model the position and the orientation of this structure does not correspond perfectly to the geological trace of the Pamplona fault but we have to keep in mind that these two parameters are not absolute value and strongly depend on the position of chosen blocks in the model. Consequently, due to the lack of outcrops, more local studies including local tomography or seismic surveys are needed to better constrain the location of the Pamplona fault.

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

6.3. Parentis anomaly The interpretation of several seismic profiles crossing the Parentis Basin including the deep reflection ECORS-Bay of Biscay profile, suggests that the Parentis basin is strongly asymmetric and separated by the Ibis fault (Fig. 3a) into a northern and southern part. Well identified in the ECORS-Bay of Biscay profile, this structure can be traced eastwards and correlated with the Ibis, Eridan and Antares salt ridges. The basin to the north is floored by a complete stratigraphic section, whereas to the south, top-basement detachment faults may floor the basin (e.g. 73BY13 profile and Aquitaine coast profile; Figs. 3b and 4a). Detachment faults are also observed on the western Landes High, however, eastwards the displacement along these faults decreases and a lateral transition from low-angle to high-angle

223

normal faults can be observed (profile 75BY13 in Fig. 4b). To better visualize the results of this work concerning the Parentis basin we made two 3D blocks. The first one (Fig. 11a) summarizes the results of gravimetric inversion. The cross section bounding the block on the upper right side corresponds to a vertical slice through our density model along the ECORS-Bay of Biscay profile on which the interpretation of the seismic section is superimposed. We use a reinterpretation of a depth section published in Bois et al. (1997) to properly compare the geological interpretation with the density section, and to correctly locate major structures such as the Moho, the top of the lower crust and basement in a depth section. The second block (Fig. 11b) represents the geometry of the Parentis basin between the ECORS-Bay of Biscay profile (redrawn on the western border of the 3D block) and the Aquitaine coast profile deduced from

Fig. 11. a) 3D block summarizing gravimetric results obtain from the superposition of Euler sources on the filtered simple Bouguer gravimetric map of the Parentis basin. On the cross section bounding the block on its left side a vertical slice through our density model (Fig. 8) is shown that coincide with the ECORS-Bay of Biscay profile (Fig. 3a). Black triangles correspond to neighboring projected Euler sources. We report on the density profile major structures issued from a reinterpretation of the depth section published in Bois et al. (1997); b) the 3D structural below summarizes the major 3D structures observed in the Parentis basin between the ECORS-Bay of Biscay profile (redrawn on the western border of the 3D block) and the Aquitaine coast profile. This section is made by gathering seismic data and results obtained with gravimetric inversion methods. All Late Aptian and younger sediments have been removed to highlight the basement structure of the basin.

224

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

seismic observations and Euler deconvolution results. In order to highlight the crustal structure of the basin all sediments younger than Late Aptian have been removed. In Fig. 11a we can note that projected Euler sources issued from the SW–NE alignment observed on the southern part of the basin correspond to the break-away of the northern detachment fault identified in the seismic section (Fig. 3a). Likewise, on the eastern part of the basin, this alignment of Euler sources that becomes N–S oriented superimposes the break-away of the northern detachment

system identified on the 73BY13 seismic section (Figs. 4a and 11). In the same way, we can observe in Fig. 11 that Euler sources line up on an E–W structure that correspond to the southward boundary of the block of pre-rift sediment overlying the detachment fault and forming the Albian anticlinal structure (CDP 4400-4700, Fig. 3a). It seems therefore that Euler sources line up along the boundary of the exhumed material. Consequently, the northern alignment of the Euler sources defines the southern boundary of the block of pre-rift sediment meanwhile the southern and eastern alignments localized

Fig. 12. 3D reconstruction of the entire studied area before Pyrenean compression based on field observation, seismic interpretations and results obtained by gravimetric inversion methods. Syn- and post-rift sediments are removed to better highlight the rift structures. A map showing the filtered simple Bouguer anomaly is underlined in the 3D block in order to show the correlation between gravity data and the geological interpretation.

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

the northern limit of the major detachment fault and its SW–NE and then N–S orientation. The resulting geometry reflects the widening of the Parentis basin and the increasing importance of exhumation processes westward (Fig. 11b). Concerning the Ibis fault, it appears that this fault is neither well marked by the Euler sources (except its eastern termination; Fig. 9a) nor well identified in density models. Either this fault is not associated with strong density contrasts, or the strong positive anomaly observed on the western border of the map at about 44.5°N (Fig. 5b), may hide the effect of this fault. 6.4. 3D reconstruction of the eastern Bay of Biscay–Western Pyrenees and conclusions Based on field and seismic observations and gravimetric modelling results we make a 3D block for the entire study area illustrating the 3D architecture of the eastern termination of the Bay of Biscay–NW Pyrenees before the onset of Pyrenean compression (Fig. 12). In order to make the crustal structure more visible, we do not show the Late Aptian and younger sediments. The cross section bounding the block on the upper left side corresponds to the ECORS-Bay of Biscay profile, the one on the lower right side to a reconstruction across the Arzacq– Mauléon basin. The major structures that can be observed on the 3D block are the Parentis and Arzacq–Mauléon basins that correspond to hyperextended rift basins. The comparison with the filtered Bouguer gravimetric map shows that there is a strong correlation between areas of major thinning and the positive gravimetric anomalies. This correlation is despite of the compressional overprint of the Mauléon basin and the thick sedimentary infill. The reconstruction shows that the basins are made of complex geometries. This is particularly shown for the Parentis basin, for which the 3D architecture is better resolved due to the weak inversion and denser seismic data set (Fig. 11b). This basin shows windows of exhumed crustal rocks, which wedge out eastwards, suggesting a decrease in accommodated extension from west to east. This change is also associated with a change from downwards-concave faults in the west (e.g. top-basement detachment faults) to upwards-concave faults in the east (e.g. high-angle faults) (Figs. 11b and 12). Another interesting observation is that there is evidence in the western Parentis basin that there is not one single detachment fault, but a detachment system formed by several detachments that extended and thinned the crust (Fig. 3a). How these structures evolved in time and space is not yet understood. The observation that high-angle faults can be traced laterally into lowangle detachment faults suggests that the faults may have initiated as high-angle faults and developed, with increasing displacement, into downward concave faults (Fig. 11b). Since the transition from highangle to low-angle faulting can also be observed parallel to strike, a likely scenario that needs to be further investigated, is that the change from high-angle to low-angle faulting is probably also a function of increasing thinning of the crust. This interpretation does, however, not answer the question of how the extensional structures were linked and interacted with major structures such as the Ibis, Saint Palais and Pamplona faults, to mention just the most prominent structures identified in the study area. This work shows that the Ibis and Saint Palais faults are key structures aligned parallel to the axes of the Parentis and Arzacq– Mauléon basins and separating within these basins two parts that show contrasting gravimetric anomaly, different structures and tectonic evolutions (Fig. 12). Based on seismic interpretations and considering the paleogeographic framework (see Jammes et al., 2009), we interpret these structures as the result of an early stage strike slip movement that may have happened before early Aptian time. Although there are no direct constrains for such an interpretation, we propose that these structures: 1) post-date initial Triassic to Jurassic rifting and predated the final Late Aptian to Albian exhumation

225

event, and 2) represent a major crustal or even lithospheric weakness that guided and localized the subsequent extension. The Pamplona fault is well imaged in our density model between 4 and 16 km depth (Fig. 8) and seems to be associated to a steep variation of the Moho topography (Fig. 6c) which is coherent with the interpretation of this fault proposed by Larrasoaña et al. (2003). Indeed, they suggest that during the Mesozoic, the Pamplona fault behaved as a transfer fault that separated two segments of the Pyrenean rift where differences in the amount and/or style of extension occurred. Based on the results presented in this study and observations published in Jammes et al. (2009) we suggest that this fault acted during final thinning and exhumation as a transfer fault linking the eastern termination of the Cantabrian basin with the Arzacq–Mauléon. During Pyrenean compression Larrasoaña et al. (2003) suggested that the Pamplona fault behaved as a transfer fault that separated segments of the Pyrenean orogen with different tectono-stratigraphic evolution that could explain the steep variation of Moho topography. On the scale of the entire studied area, we suggest that Parentis and Arzacq–Mauléon are two V-shaped basins widening westward. We propose that the formation of the hyper-extended Parentis and Arzacq–Mauléon basins were localized along pre-existing strike slip lineaments (Ibis and Saint Palais faults). During subsequent rotation of the Iberian plate away from Europe, the propagation of the deformation was accommodated along extensional faults that show decreasing amounts of extension along strike, resulting in wedges of exhumation surfaces and a change from high-angle to low-angle structures along strike. This kind of structures results in frayed terminations of the basins (e.g. the eastern Parentis basin) (Fig. 11). Elsewhere, transfer structures such as the Pamplona fault segmented basins during final rifting and can explain sharp basin terminations as is the case for the western Mauléon Basin. Acknowledgment We thank M. Schaming and the ECORS program partners (CNRSINSU, IFP, IFREMER, SNEA (P) now TOTAL) for providing us the ECORSBay of Biscay profile and G. Nesen, P. Unternehr and B. Ros from TOTAL for providing us seismic and drill hole data. We thank also A. Bouzeghaia for his help to improve the quality of the figures and the two reviewers D. Pedreira and an anonymous reviewer for their valuable comments that helped to improve the quality of the manuscript. The research presented in this paper was supported by the GDR marges/Action marges and the Ministry of Education and Research through a PhD Grant to S. Jammes. References Biteau, J.J., Le Marrec, A., Le Vot, M., Masset, J.M., 2006. The Aquitaine Basin. Petrol. Geosci. 12, 247–273. Blakely, R., 1996. Potential Theory in Gravity and Magnetic Applications, p. 441. Cambridge University Press. Bois, C., Gariel, O., Lefort, J.P., Rolet, J., Brunet, M.F., Masse, P., Olivet, J.L., 1997. Geologic contribution of the Bay of Biscay deep seimic survey: a summary of the main scientific results, a discussion of the open questions and suggestions for further investigations. Soc. Géol. Fr. Mém. 171, 193–209. Casas, A., Kearey, P., Rivero, L., Adam, C.R., 1997. Gravity anomaly map of the Pyrenean region and a comparison of the deep geological structure of the western and eastern Pyrenees. Earth Planet. Sci. Lett. 150, 65–78. Casas-Sainz, A.M., Gil-Imaz, A., 1998. Extensional subsidence, contractional folding and thrust inversion of the eastern Cameros basin, northern Spain. Geol. Rundsch. 86, 802–818. Chazalon, A., Campillo, M., Gibson, R., Carreno, E., 1993. Crustal wave propagation anomaly across the Pyrenean Range. Comparison between observations and numerical simulations. Geophys. J. Int. 115, 829–838. Claude, D., 1990. Etude stratigraphique, sédimentologique et structurale des dépôts mésozoïques au nord du Labourd massif. Rôle de la faille de Pamplona (Pays Basque). thèse de Doctorat, 436pp, Université de Bordeaux II, Bordeaux, France. Cogbill, A.H., 1990. Gravity terrain corrections calculated using digital elevation models. Geophysics 55, 102–106.

226

S. Jammes et al. / Tectonophysics 489 (2010) 210–226

Daignières, M., Gallart, J., Banda, E., 1981. Lateral variation of the crust in the North Pyrenean zone. Ann. Geophysicae 37, 435–456. Daignières, M., Gallart, J., Banda, E., Hirn, A., 1982. Implications of the seismic structure for the orogenic evolution of the Pyrenees range. Earth Planet. Sci. Lett. 57, 88–110. doi:10.1016/0012-821X(82)90175-3. Daignières, M., Séguret, M., Specht, M., ECORS team, 1994. The Arzacq-Western Pyrenees ECORS deep seismic profile. Hydrocarbon and Petroleum Geology of France: Spec. Publ. Eur. Assoc. Pet. Geosci., 4, pp. 199–208. Ducasse, L., and Velasque, P. C., 1988. Geotraverse dans la partie occidentale des Pyrénées de l'avant-pays Aquitain au bassin de l'Ebre, effet d'une inversion structurale sur l'édification d'une chaîne intracontinentale. Thèse de Doctorat, 248pp., Université d'Aix-Marseille III, Marseille, France. Ferrer, O., Roca, E., Benjumea, B., Munoz, J.A., Ellouz, N., MARCONI, team, 2008. The deep seismic reflection MARCONI-3 Profile: Role of extensional mesozoic structure during the Pyrenean contractional deformation at the eastern part of the Bay of Biscay. Mar. Petrol. Geol 25, 714–730. Grandjean, G., 1992. Etudes des structures crustales dans une portion de chaîne et de leur relation avec les bassins sédimentaires. Application aux Pyrénées Occidentales au travers du projet ECORS-Arzacq-Pyrénées. Thèse de Doctorat, 300pp., Université de Montpellier III, Montpellier, France. Jammes, S., Manatschal, G., and Lavier, L., submitted for publication What is the interaction between pre-rift salt and detachment faulting in hyper-extended rift systems: the example of the Parentis and Mauléon basins (Bay of Biscay-Western Pyrenees). AAPG Bulletin. Jammes, S., Manatschal, G., Lavier, L., Masini, E., 2009. Tectono-sedimentary evolution related to extreme crustal thinning ahead of a propagating ocean: the example of the western Pyrenees. Tectonics 28, TC4012. doi:10.1029/2008TC002406. Larrasoaña, J.C., Parés, J.M., Millàn, H., del Valle, J., Pueyo, E.L., 2003. Paleomagnetic, structural, and stratigraphic constraints on transverse fault kinematics during basin inversion: the Pamplona Fault (Pyrenees, north Spain). Tectonics 22, 1071–1092. doi:10.1029/2002TC001446. Lefort, J.P., Bois, C., Liewig, N., Peucat, J.J., Agarwal, B., Gariel, O., 1997. Contribution of the ECORS Bay of Biscay deep seismic profile to the location of the southern Variscan front beneath the Aquitaine basin (France). Soc. Géol. Fr. Mém. 171, 79–96. Lefort, J.-P., Agarwal, B.N.P., 2002. Topography of the Moho undulations in France from gravity data: their age and origin. Tectonophysics 350, 193–213. Marillier, F., Tomassino, A., Patriat, P., Pinet, B., 1988. Deep structure of the Aquitaine shelf: constraints from expanding spread profiles on the ECORS Bay of Biscay transect. Mar. Petrol. Geol. 5, 65–74. Masse, P., 1997. The early Crataceous Parentis basin (France). A basin associated with a wrench fault. Soc. Géol. Fr. Mém. 171, 177–185. Mata, M.P., Casas, A.M., Canals, A., Gil, A., Pocovì, A., 2001. Thermal history during Mesozoic extension and Tertiary uplift in the Cameros Basin, northern Spain. Basin Res. 13, 91–111. Muñoz, J.A., 2002. Alpine tectonics I: the Alpine system north of the Beltic Cordillera: the Pyrenees. In: Moreno, A.T. (Ed.), The Geology of Spain: In: G., W. (Ed.), Geol. Soc., London, pp. 370–385. Nagy, E.A., Stock, J.M., 2000. Structural controls on the continent–ocean transition in the northern Gulf of California. J. Geophys. Res. 105, 16251–16269. doi:10.1029/ 1999JB900402. Oldenburg, D.W., 1974. The inversion and interpretation of gravity anomalies. Geophysics 39, 526–536. Olivet, J.L., 1996. La cinématique de la plaque Ibérique. Bull. Centres Rech. Explor.-Prod. Elf-Aquitaine 20 (1), 131–195.

Parker, R.L., 1972. The rapid calculation of potential anomalies. Geophys. J. R. astr. Soc. 31, 447–455. Pedreira, D., Pulgar, J.A., Gallart, J., Torné, M., 2007. Three-dimensional gravity and magnetic modeling of crustal indentation and wedging in the western Pyrenees– Cantabrian Mountains. J. Geophys. Res. 112. doi:10.1029/2007JB005021. Péron-Pinvidic, G., Manatschal, G., Minshull, T.A., Sawyer, D.S., 2007. Tectonosedimentary evolution of the deep Iberia–Newfoundland margins: evidence for a complex breakup history. Tectonics 26. doi:10.1029/2006TC001970. Peybernès, B., Souquet, P., 1984. Basement blocks and tecto-sedimentary evolution in the Pyrenees during Mesozoic times. Geol. Mag. 121, 397–405. doi:10.1017/ S0016756800029927. Pinet, B., Montader, L., Curnelle, R., Cazes, M., Marillier, F., Rolet, J., Tomassino, A., Galdeano, A., Patriat, P., Brunet, M.F., Olivet, J.L., Schaming, M., Lefort, J.P., Arrieta, A., Riaza, C., 1987. Crustal thinning on the Aquitaine shelf, Bay of Biscay, from deep seismic data. Nat., Lett. Nat. 325. Platt, N.H., 1990. Basin evolution and fault reactivation in the western Cameros Basin, Northern Spain. J. Geol. Soc. Lond. 147, 165–175. Sandwell, D.T., Smith, W.H.F., 2009. Global marine gravity from retracked Geosat and ERS-1 altimetry: ridge segmentation versus spreading rate. J. Geophys. Res. 114, B01411. doi:10.1029/2008JB006008. Schellart, W.P., 2002. Alpine deformation at the western termination of the Axial Zone, Southern Pyrenees. J. Virtual Explorer 8, 35–55. doi:10.3809/jvirtex.2002.00055. Schoeffler, J., 1982. Les transversales Basco-Landaises. Bull. Cent. Rech. Explor. Prod Elf Aquitaine 6, 257–263. Sens-Schönfelder, C., Margerin, L., Campillo, M., 2009. The influence of laterally variable heterogeneity on the propagation of Lg-waves in the Pyrenees. Eos Trans. AGU 90 (52) Fall Meet. Suppl., Abstract S32A-08. Serrano, O., Delmas, J., Hanot, F., Vially, R., Herbin, J.P., Houel, P., Tourlière, B., 2006. Le bassin d'Aquitaine: Valorisation des données sismiques, cartographie structurale et potentiel pétrolier, Rapport Régional d'Evaluation Pétrolière. Bureau de la Recherche Géologique et Minière, Orléans, France, p. 245. Sibuet, J.C., Srivastava, S.P., Spakman, W., 2004. Pyrenean orogeny and plate kinematics. J. Geophys. Res. 109. doi:10.1029/2003JB002514. Taylor, B., Huchon, P., 2002. Active continental extension in the western Woodlark Basin: a synthesis of Leg 180 results. In: Taylor, B., Huchon, P., Klaus, A. (Eds.), Proc. ODP, Sci. Results 180, pp. 1–36. Teixell, A., 1996. The Ansó transect of the southern Pyrenees: basement and cover thrust geometry. J. Geol. Soc. London 153 (301–310), 301–310. doi:10.1144/gsjgs.153.2.0301. Teixell, A., 1998. Crustal structure and orogenic material budget in the west central Pyrenees. Tectonics 17, 395–406. doi:10.1029/98TC00561. Thompson, D.T., 1982. EULDPH: a new technique for making computer-assisted depth estimates from magnetic data. Geophysics 47, 31–37. Vergés, J., 2003. South Pyrenean oblique thrust system evolution: Segre and Pamplona faults (Evolución de los sistemas de rampas oblicuas de los Pirineos meridionales: fallas del Segre y Pamplona). Boletin Geologico y Minero Esp. 114, 87–101. Vergés, J., García Senz, J.M., 2001. Mesozoic evolution and Cenozoic inversion of the Pyrenean Rift, in Peri-Tethys memoir 6: Pery-Tethyan rift/wrench basins and passive margins. In: Ziegler, P.A., et al. (Ed.), Memoires Muséum National d'Histoire Naturelle, 186, pp. 187–212. Widiwijayanti, C., Tiberi, C., Deplus, C., Diament, M., Mikhailov, V., Louat, R., 2004. Geodynamic evolution of the northern Molucca Sea area (Eastern Indonesia) constrained by 3-D gravity field inversion. Tectonophysics 386, 203–222. doi:10.1016/j.tecto.2004.05.003. Wortmann, U.G., Weissert, H., Funk, H., Hauck, J., 2001. Alpine plate kinematics revisited: the Adria problem. Tectonics 20 (1), 134–147.