TECTONICS, VOL. 15, NO. 6, PAGES 1192-1212, DECEMBER 1996

Low-angle crustal ramp and basin geometry in the Gulf of Lion passive margin: Oligocene-Aquitanian Vistrenque graben, SE France

A. Benedicto, 1 P. Labaume, 2 M. S6guret, 1 and M. S6ranne 1

Abstract. With more than 4000 m of Oligocene-Aquitanian sediments, the Vistrenque graben (SE France) is the deepest synrift depocenter of the Gulf of Lion passive margin, NW Mediterranean. Detailed analysis of industrial seismic reflection profiles and borehole data show that the steep N•mes fault, which bounds the graben to the NW, forms at depth a low-angle (25 ø) crustal ramp. Along-strike changes of hangingwall geometry allow us to infer along-strike changes of fault shape: A rollover structure and divergent Oligocene- Aquitanian basin fill are associated with a listric geometry of the fault in the southern part of the graben, while a pseudo- rollover and compensation graben result from a two-segments planar geometry of the fault in the northern part. Preexisting structures inherited from Mesozoic extension and Late

Cretaceous-Eocene Pyrenean thickening controlled the location of the N•mes fault and the transfer zones which divide

the graben into different compartments. Since both hangingwall and fault profile are well constrained, restoration techniques can be used to estimate the prerift topography. The Vistrenque graben was formed close to sealevel, but at the front of a > 1 km-high elevated area resulting from the Pyrenean orogeny. In the studied transect, the N•mes fault formed the landward (NW) boundary of the basemenf faulted domain of the margin. Extensional deformation was restricted to this domain during most of the rifting interval. Small amounts of extension were transmitted landward to Mesozoic

cover d6collement rooted in the N•mes fault, only during short episodes, probably resulting from gravitational instability during margin collapse. The N•mes low-angle crustal ramp, as well as the other crustal ramps of the margin of similar orientation, are probably newly formed extensional structures rather than reactivated Pyrenean thrusts. Their activation at a low-angle may have been allowed by crustal weakening resulting from the previous Pyrenean thickening. Upper crustal extension corresponding to the graben formation was transmitted basinward through an intracrustal detachment, or/and distributed in the lower crust across the margin. In contrast to the more stretched areas of the margin which do not display thick synrift series due to their initial high surface

I G6ofiuides-Bassins-Eau, CNRS Universit6 Montpellier II, Montpellier, France.

2Laboratoire de G6ophysique Interne et Tectonophysique, CNRS Universit6 Joseph Fourier, Grenoble, France.

Copyright 1996 by the American Geophysical Union.

Paper number 96TC01097. 0278-7407/96/96TC-01097512.00

elevation, the Vistrenque basin fill records the whole rifting episode because of its location at the front of the Pyrenean orogen.

1. Introduction

In the Gulf of Lion passive margin (SE France, Figure 1), it has recently been suggested that upper crustal extension during the Oligocene-Aquitanian rifting was accommodated by low-angle (250-30 ø ) crustal ramps located in the shelf and slope of the margin, while the landward part was only affected by thin-skinned Mesozoic cover d6collement (Figure 2) [Sdranne et al., 1995]. The Nimes fault corresponds to the boundary between the two domains and controls the Vistrenque graben (Figure lb), the deepest synrift depocenter of the margin. The Vistrenque graben is thus a strategic feature for a better understanding of the margin evolution. Surprisingly, it remains practically unknown from a structural point of view, despite the existence of abundant seismic reflection and borehole data accumulated during 40 years of oil research and salt exploitation. This paper presents a new structural model for the Vistrenque graben and the Nimes fault based on the interpretation of the subsurface data. Basin model is analyzed in the regional context of the Gulf of Lion margin in order to better understand its kinematic evolution.

Geometry of normal faults in passive margins has been largely discussed. Normal faults flattening at depth associated to crustal detachments have classically been proposed as main structural feature in the evolution of extensional rifted margins [Wernicke and Burchfield, 1982; Gibbs, 1984; Lister et al., 1991]. Nevertheless, relationships between near-surface steeply dipping and deep low-angle detachments remains one of the most discussed point.

Linkage of surface high-angle fault and deep low-angle detachment can (1) be listric [Verral, 1981; Wernicke and Burchfield, 1982; Gibbs, 1983; Davison, 1986; White et al., 1986; McClay and Ellis, 1987; Ellis and McClay, 1988], or (2) by a set of planar faults with discrete change of dip [McClay and Ellis, 1987; Faure and Cherrnette, 1989], also known as kink-style fault [Groshong, 1989].

Analogue modelling [Closs, 1969; McClay and Ellis, 1987; Ellis and McClay, 1988] and geometric/numerical models [Jackson and McKenzie, 1983; Gibbs, 1983; White et al., 1986; Faure and Cherrnette, 1989; Groshong, 1989; Xiao and Suppe, 1992] of normal faults systems in the case of faults flattening at depth has pointed out the relationships between hangingwall deformation and fault section shape. These different approaches converge in the result that movement along a listric fault generates a half graben with a divergent basin fill towards the fault and a rollover structure of the

1192

BENEDICTO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY 1193

Lion

Prover•,al basin

, , Main normal faults +

--•-- Pyrenean thrusts (Eocene) • Alpine thrusts (Mio-Pliocene) • Pyrenean thrust reactived +

during Alpine compression + • Pyrenean compression

Oligocene-Miocene extension

43

ß •--• Main extensional basins !/i•.• Prerift cover detached during Pyrenean orogeny • .... Paleozoic involved in Pyrenean orogeny

Outcropping Paleozoic NW limit of oceanic crust

NW limit of intermediate continental crust

nque graben. (Fig.

44

Front

+ +

+

' 50km

'%

"";,' ,' ,, ,, q "'-.. .,_ ........

I 2 ø / 4 ø 5 ø 6 ø

Figure 1. (a)'Position of the Gulf of Lion margin in the Western Mediterranean basin. Stippled lines are prerift location of the Corsica-Sardinia and Balearic islands. (b) Simplified structural framework of the Gulf of Lion passive margin and location of the Vistrenque graben, AB, Albs basin; HB, H6rault basin; MB, Manosque basin; NPFZ, North Pyrenean fault zone; Sh, Saint Hippolyte well; Va, Vacquibres well; SM, Saintes Maries 101' Be, Beauduc well; Mo, Montpellier; Ni, N•mes; Ma, Marseille. Modified from S6ranne et al. [1995].

42

hangingwall. However, a planar tault generates a compensation graben with a horizontal basin fill and a pseudo-rollover structure of the hangingwall, these different models are not always well differentiated in the literature dealing with normal faults systems [e.g., Gibbs, 1984; Groshong, 1989; Dula, 1991; Xiao and Suppe, 1992; Withjack et al., 1995].

If the relationships between fault and hangingwall/basin fill geometry in two dimension appears well established, the possible along-strike variation of the fault geometry has not been investigated and natural examples of such three- dimensional structures have not been described.

Here, we analyze the three-dimensional geometry of the Vistrenque graben and the N•mes fault, pointing out the along- strike variation of the fault geometry from listric to planar and the resulting contrasting basin geometries.

Relationships between hangingwall and fault geometries have lead to different techniques to reconstruct the fault shape

from the hangingwall shape as aids to field and seismic interpretations. These techniques consider an undeformable footwall and different modes of deformation for the

hangingwall (see review and discussion by Dula [ 1991]). Fault reconstruction techniques use a horizontal prerift surface (datum) and the geometry of the hangingwall (top of the prerifi) to reconstruct the fault shape at depth or vice versa. However, in the case of intracontinental basins, possible inherited prerifi elevation (irregular topography) and erosion of both footwall and hangingwall must be taken into account in analyzing the final geometry of the fault system. Besides, prerift elevation may be responsible for erosion instead of deposition during part of the activity of faults. In this case, calculated tectonic extension from the observed final

geometry may be minimized, !eading to differences in extension ratios calculated from tectonics or subsidence at the

scale of the margin. Our good control of hangingwall and fault geometry of the

1194 BENEDICTO ET AL.' FO('IENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY

o • o • • • •- c• c• • Vistrenque graben allow us to use the classical fault reconstruction techniques, in an inverse way, to reconstruct the prerift relief, inherited from the prerift Pyrenean orogeny which affected the area.

At the scale of the Gulf of Lion margin, constraint on Vistrenque graben and N•mes fault geometry gives valuable basis to discuss the role of a major crustal ramp in the structure of a passive margin. The Vistrenque graben location allows to discuss the kinematic relationships between basement fault and cover d6collement domains, and the genetic relationships between the low-angle crustal ramps of the margin and previous crustal thickening.

2. Geological Framework

The Gulf of Lion passive margin corresponds to the NW margin of the Proven9al basin (Figure l a) which opened due to Oligocene-Aquitanian rifting and Burdigalian oceanic accretion associated with the southeastward drift of the

Corsica-Sardinia block [Auzende et al., 1973; Biju-Duval et al., 1979; Rehault et al., 1984; Burrus, 1984]. This extensional episode corresponded to back-arc opening in the context of African and European plate convergence [Olivet et al., 1984; Rehault et al., 1984; Maillard and Mauffret, 1993].

The Gulf of Lion margin formed at the expense of a Paleozoic basement and Triassic to Eocene cover. It comprises Oligocene to Aquitanian basins overlain by Miocene (post- Aquitanian) to Plio-Quaternary sequences (Figure lb) [Lefevre, 1980; Arthaud et al., 1980; Gorini et al., 1993; Guennoc et

al., 1994]. The Oligocene-Aquitanian sediments are classically considered to represent the rifting stage and the post-Aquitanian to Plio-Quaternary deposits of the postrift stage [Burrus, 1984; Gorini et al., 1993]. The margin is bounded to the NW by the C6vennes fault, at the edge of the Massif Central. The NE boundary corresponds to the Ar16sienne transfer zone [Gorini et al., 1993' Guennoc et al., 1994] which separates the continental margin from the intracontinental rift of the Rh6ne valley between the C6vennes and Durance faults (Figure 1).

The N•mes fault, which bounds the Vistrenque graben to the NW, represents the northeastern boundary of a low-angle faulted basement domain (Figure 2) [Sdranne et al., 1995] which extends basinward (mainly offshore). The structure of this domain is well illustrated on the Etude Continental et

Oceanique par Reflexion et Refraction Sismique (ECORS) deep seismic reflection profiles [de Voogd et al., 1991] and on industrial seismic reflection lines [Gorini et al., 1993]. The landward (northwestern) part of the margin, between the N•mes and C6vennes faults, corresponds to a thin-skinned extensional domain, characterized by half grabens associated with listtic faults which detach in the Triassic shales and

evaporites [Stiranne et al., 1995]. The rift structures of the Gulf of Lion margin overprint older

structures which result from a complex evolution: 1. Mesozoic extension resulted in the formation of the SE

France basin (part of the Alpine Tethyan margin) [Beaudrirnont and Dubois, 1977; Curnelle and Dubois, 1986].The SE basin was controlled by NE-SW and E-W trending major faults inherited from the late Paleozoic strike- slip faulting [Arthaud and Matte, 1975]. More than 10 km of

BENEDICrO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY 1195

Triassic to Cretaceous sediments accumulated in the central

part of the basin, bounded to the NW by the C6vennes fault and to the south by an E-W trending uplifted domain of which the northern boundary lies close to the present-day coastline.

2. Late Cretaceous to Eocene N-S Pyrenean compression resulted in the formation of the E-W trending Pyrenean belt (Figure l a), which extended from the present-day range across the Gulf of Lion area. In the Corbinres region, the Pyrenean structural zones are transferred northwestward with respect to the present-day Pyrenean range (Figure lb). During Pyrenean compression, NE-SW trending Mesozoic faults were reactivated as sinistral strike-slip faults and E-W trending faults as N verging thrusts [Arthaud and $•guret, 1981; Ternpier, 1987]. Basement frontal thrust ramps were developed close to the present-day coastline (Figure lb). To the north, the Mesozoic cover was detached above the Triassic shales and

evaporites and affected by E-W trending folds and thrusts. Timing of the transition from Pyrenean compression to

Oligocene-Aquitanian extension is still poorly known. Northeast of the Ar16sienne transfer zone, in the Al•s and

Manosque basins (Figure lb), middle Ludian (Upper Eocene) continental sediments display extensional growth structures and are classically interpreted as the oldest synrift sediments [Cavelier et al., 1984]. Southwest of the Ar16sienne transfer zone, in the Montpellier region, the same sediments belong to the prerift succession and are involved in the latest Pyrenean folding. In this area, the oldest synrift sediments are continental deposits attributed to the middle Stampian on the base of vertebrate fauna correlated with the P19-P20

foraminifera zones of Blow (about 34 to 32 Ma, Crochet [1984]). This suggests a diachronous onset of the rifting, older in the north and younger southward, that is toward the front of the Pyrenean range.

3. Stratigraphy in the Vistrenque Graben

The general stratigraphy of the Vistrenque graben and adjacent areas is established from 48 petroleum research boreholes, 25 of which have reached the pre-Oligocene succession that forms the basin substratum (Figure 3). The most complete and best known basin fill sequence has been drilled in the Pierrefeu well (P in Figure 3), where the Oligocene to Quaternary succession is 4920 m thick (Caline, 1983 in Valette [ 1991 ]) and in the neighbouring Parrapon salt exploitation boreholes (PR in Figure 3, Valette [1991]; Valette and Benedicto [1995]).

On the base of different pre-Oligocene stratigraphy two areas can be distinguished from boreholes (Figure 3): in the north of the basin, the top of the pre-Oligocene consists of Neocomian (Valanginian-Berriasian) marine carbonates, whereas in the south it consists of Jurassic marine carbonates

locally overlain by Upper Cretaceous fluvial facies (Marette well, M in Figure 3) or/and continental Eocene carbonates (Baumelles well, B in Figure 3).

The Oligocene sediments form a succession of continental to lagoonal deposits with alternations of fluvial and lacustrine-palustrine facies. According to Valette [1991], this succession includes from the bottom to the top: (1) the "S6rie Grise" (2000 m thick), consisting of dark clays and sandstones with lignite intercalations in the lower part,

passing upward to lacustrine limestones alternating with marls, sandstones, and conglomerates; (2) the "S6rie Rouge" (200 m thick), comprising palustrine red clays and gypsiferous marls with several intercalations of marls and sandstones; (3) the "S6rie Calcar6o-salif'ere" (900 m thick), formed by rhythmic halite (exploited in Parrapon), anhydrite, and marly clay calcareous deposits.

The precise age of the Oligocene sediments remains poorly known due to the absence of fossils in the continental facies.

The "S6rie Calcar6o-salif'ere" is attributed to the Stampian on the base of correlation with the saliferous deposits in the Rh6ne Valley area. The age of the underlying "Oligocene" series remains unknown. The Chattian has not been

recognized in the Vistrenque graben, but is present in the outcropping basins of adjacent areas [Cavelief et al., 1984].

The Oligocene series are overlain by 800 to 1500 m of sediments attributed to the Aquitanian on the base of foraminifera fauna (Calina, 1983 in Valette [1991]), divided into three monotonous series. The Lower Aquitanian series consists of lagoonal to coastal sediments made of nonfossiliferous-variegated clays with rare intercalations of calcareous mudstones and sandstones. These deposits are overlain by marine clays with rare intercalations of limestones corresponding to the Middle Aquitanian. The Upper Aquitanian series consists of a coastal clay-sandstone coarsening-up succession, with few lagoonal dolomitic limestone and lignite beds. The Oligocene-Aquitanian succession indicates a transgressive trend from the Stampian to the Middle Aquitanian, whereas the Upper Aquitanian series corresponds to a regressive phase.

The Aquitanian is unconformably overlain by transgressive Burdigalian marine calcarenites passing upward to Langhian and Tortonian marine marls. In the Rh6ne Valley, this calcarenite-marl succession is classically considered as a transgressive trend. These deposits are truncated by a regional erosional surface of Messinian age [Beaufort et al., 1954; Dernarcq, 1970] related to the drying out of the Mediterranean sea [Ryan and Cita, 1978]. The erosional surface is buried under Pliocene marine clays and marls passing upward to Quaternary fluvial deposits of the Rh6ne delta.

By analogy with the geodynamical interpretation of Bessis and Burrus [1986] for the offshore area, the unconformity between the Aquitanian and Burdigalian observed in the Vistrenque graben wells and seismic lines was interpreted by Valette [1991] and Valette and Benedicto [1995] as the breakup unconformity separating Oligocene-Aquitanian synrift basin fill from Burdigalian to recent postrift sequences. This interpretation is discussed in later sections, on the basis of the new seismic data presented in this paper.

4. Structure of the Vistrenque Graben

The first structural interpretation of the Vistrenque graben, based only on borehole and gravimetric data, assumed a symmetric graben bounded by steeply dipping faults involving the Paleozoic basement [Arthaud et al., 1980]. In a detailed study of the Parrapon salt exploitation area, seismic interpretation led Valette [1991] to propose that the main structural control of this part of the graben was exerted by the N•mes fault, interpreted as a listric fault flattening at depth.

1196 BENEDICTO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY

i i Main normal taults Top of pre-Olig. in wells - - SG seismic profiles

VG seismic profiles • Prerift not reached PR seismic profiles •j• Lower Cretaceous

!lillil I Transfer zone 41• Jurassic, locally Outcropping pre-Oligoc. covered by Upper I • / • I Oligo-Aquitanian basin fill Cretaceous / Eocene

*"'"'• Outcropping post-Aquit. Miocene Outcropping Plio-Quaternary v35 Vaunage

L1

LJ •J•

Figure 3. Structural framework of the Vistrenque graben showing location of data base and cross sections. Boreholes cited in the text or reported on line drawings (Plate 1) are: V3, Vaunage 3; LJ, La Jassette; L1, Lunel 1; L2, Lunel 2; L3, Lunel 3; A, Aubord; SV, Saint Veran; P, Pierrefeu; A3, Albaron 3; A7, Albaron 7; A101, Albaron 101; M, Marette; M1, Montcalm 1; M2, Montcalm 2; B, Baumelles; V, Vaccar•s; SM,

Saintes Maries 101. PR, Parrapon salt exploitation; t.z., transfer zone; a, b and c, cross sections of Figure 6.

For the same area, Valette and Benedicto [1995] proposed a two-segments planar geometry for the N•mes fault. Here, we show that the N•mes fault corresponds to a low-angle basement ramp along the whole Vistrenque graben, and that along-strike variations of fault shape imply major changes in hangingwall geometry.

4.1. Seismic interpretation Data. The industrial multichannel seismic reflection

surveys (presented in Table 1) and boreholes used in this study are located in Figure 3. Comparison of unmigrated, migrated

and depth-converted seismic lines has been carried out to test the consistency of the seismic interoretation and validates the

use of the unmigrated sections for the interpretation of the structural style. Seismic facies (Figure 4) are best visible and correlatable in the Saint Gilles (SG) unmigrated survey.

The seismic interpretation is summarized in the line drawings presented in Figure 5. Isobath/isopach maps (Figure 5) and depth cross sections (Figure 6) have been constructed from seismic and bore hole data. Depth conversion of the seismic data has been done using the seismic stack velocities, but only at points where the stack velocities were in

BENEDICTO ET AL.: EXTF./qSIONAL CRUSTAL RAMP AND BASIN GEOMETRY 1197

Table 1. Seismic Reflection Data Used in this Work (Location in Figure 3)

Survey Company Seismic Lines Processing

Saint Gilles

(SG)

Vauvert-

Gallician (VG)

ESSO-REP, 1969 SG. 1 to 18 SG.2, 3, 6, 8, 9, 10& 11

SG.I, 2 &7

SG.4 & 9

Eurafrep, 1986 VG. 1 to 8 VG.I, 3,5 &7

Parrapon Elf-Atochem, 1986-87 (PR) [Valette, 1991 ]

PR. 1 to 7

Satan

Migration by Eurafrep, 1986 Migration by Elf-Atochem, 1975 Migration and depth convertion by IFP, 1994 Satan

Migration by Eurafrep, 1986

Migrated

Figure 4. Typical pattern of seismic stratigcaphic units and discontinuities used for seismic interpretation; 2, discontinuity 2 (Aquitanian/Oligocene); 3, discontinuity 3 (post-Aquitanian unconformity, Burdigalian/Aquitanian); 4, discontinuity 4 (Messinian erosion, Pliocene/Miocene); b, unit b (Oligocene); c, unit c (Aquitanian); d, unit d (post- Aquitanian Miocene); e, unit e (Pliocene); b I, group b I ("S6rie Grise"); bii, group bii ("S6rie Rouge"); bii I, group bill ("S6rie Calcareo-salif'ere"). Discontinuity 1 (top of the pre-Oligocene) and unit a (pre-Oligocene) are not visible in this section of the unmigrated line SG2.

agreement with velocity analysis in the Pierrefeu, Parrapon- 18, and La Galine- 1 wells [Valette, 1991].

Seismic stratigraphy. The seismic stratigraphy used in this work is derived from that proposed by Valette [1991] for the Parrapon survey (PR in Figure 3). Validity of Valette's seismic stratigraphy is ensured by the goccl quality and dense grid of PR lines, and correlation using geophysical logs and vertical seismic profiles in closely spaced wells. Although Valette's seismic sequences were defined in a small area in the northern part of the Vistrenque graben, they have proved to have excellent lateral continuity across the whole graben.

In this study, we consider five seismic stratigraphic units, a to e (units b to e are showed in Figure 4), characterized by specific seismic facies patterns and bounded by 4 major discontinuities, 1 to 4 (discontinuities 2 to 4 are shown in

Figure 4). Although some of these units may correspond to seismic sequences in the sense of Mitchurn [1977], they are referred to as units and not sequences due to the difficulty of defining the precise nature of some of the boundaries on the seismic lines. The typical seismic facies are found in the central part of the graben. Their characteristics become progressively less distinct southeastward due (1) to lateral thinning and facies change of the basin fill and (2) to seismic noise resulting from the Messinian erosional surface. The five seismic units correlate well with the litho-biostratigraphic units drilled in boreholes.

Unit a (sequence 1 by Valette [1991]) is truncated upward by discontinuity 1 and exhibits a variable opaque to chaotic seismic facies. Unit a corresponds to the pre-Oligocene succession, and discontinuity 1 to the top of the basin substratum (isobath map b in Figure 5).

Unit b (sequences 2 to 5 by Valette [1991]) is concordant with or onlaps discontinuity 1 and is topped by concordant discontinuity 2. In the southern part of the Vistrenque graben, unit b is made up of an alternance of high- and low-amplitude NW divergent reflectors. In the central part of the graben the reflectors are subparallel and form three groups: b I, bii and blii, from the bottom to the top. Here bii shows low-amplitude reflectors or transparent facies between high amplitude reflectors of b I, and bill,. A bright continuous reflector forms the top of bii I. In the northern part of the Vistrenque graben,

• ,•. ;•; ===================================================================== ;.,::::::::•.::-..::: •..::::..:..:.::•:::. ,. '-..- •.' •' • '•".-':-:':•.::•{•{•:.-:...::;.".-'•ig'.-:-:.-"• :::i•i;i: :':':":'"-•'.:;:::

• E"'"'""""•'•'.---...•:½ i:i :::"-':--'4--- --:%: '"'" • '• • ,,c: . 2' ........ :;:,•;½:'*:';':'-:-'"•ii':""':-'":::•-:---'-"•i • o • -,-, • • ',-,

= c: 0,,, o-___. .L,00 • • "= 2 E • •

-.-- n-, _•, •-',• ,• I::: 0 ->,, •'- ."•4 c•.'• • "-' •'

"'-= " • ol

m o o • • -w-.. o

•'½; / •2_-_• •----- • • •,.• _,/ o 0 o o

• 0 '- •_. • (p 0 -'" r- c

I•• '• I o o ,,, =o . • • • • •

"E:- .../ *.., • • -,•- .--:.::-: ,•.e ,•. '."4

"•""--. ß '. , '% '":":•----."-'- -"--- -'-':•.•,•"::--....':, .•....'.":.4.'.:..'.-'..*..--':•;•

,// ' "'x. "'.' •'•• o'. , ...... :;,:,:...-';:.,,::•-;, .......................... I • k •'-_,/.• '.:. %%,\', ,•,.2..• I :•':•'::' "•'••' •" •::";":;'" fr'"'""'•'"':"":•:•••• .................... _ ..

BENEDICTO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY 1199

NW

Vauvert

compartment •'-,. •3• ;--I extension = 8750-9000m !--;.-• .,"• La Jassette (p. 3km) St. Veran (p. 1kin)

Vaunage 3 (p.6km) Aubord 1 (p. 1kin) Pierrefeu

• d6collement •

Albaron 101 (p. Skm) Albaron 3 I Vaccar•s

Si= NE

•;•d/; Jurasic , ..,, basin inverted during the -o '. ..... ' .'. -, . • •o'•.• .:;~, • -• - , •mes •au•r '•', .... 2, ' .... '. • - '..'• '. ' '- 'Mesozoic E-W fa, Pyrenean compression .•; • • . •..: -,' • . '.•, • •* •..- .... ; • • •: 5 ,, • . ,f, • ß ß. ............ • • .... ß .... • ..... •'•.• which controled erosion of

,-, - -. ß • •. , •....... -, .- • ,-- •, • ,.'.•. .' .......,.. ,z •, • •.' . ' ' 'Cretaceous to the south, rotated and "• • reactivated as a N-verging thrust during:

the Pyrenean compression

Marsillargues e•=; ;• compartment •'•= --l extension = 9150-9500m }-- .. • Petit Rh6ne graben J• J• II Vistre7 uegraben j• • •

break-up unconformity Messinian erosion ........

Mesozoic NTmes fault -,

E-WPyrenean the Pyrenean compression

Okra

1

2

3

4

5 •-- Marly 6 Lias 7

8

9

lO

•,,•,•:•t'• extension = 3600.4100m !

Vistreique graben Marette break-up unconformity

•| extension=4600m l--;._-•,",• Marette Petit Rh6ne graben C O rn p art rn e n t

Aigues Mortes I Stes. Marie,102

Mesozoic Nimes fault

2•.. -ø ;, '•

Messinian erosion

POSTRIFT

PRERIFT

• Plio-Quaternary ........ ß '"':• Miocene pot-Aquitanian

!:-',"."1 Eocene

• Triassic • Lower Cretaceous

•-:• Paleozoic

SYNRIFT / F" '"'" "'"'• Aquitanian i!:i...' Oligocene r-T'-i upper Jurassic | ! Middle and

Lower Jurassic

SW

Figure 6. Cross sections through the Vistrenque graben. Note the compensation graben geometry in cross section (Figure 6a) and the rollover geometry in cross sections (Figure 6b) and (Figure 6c) (location in Figure 3).

1200 BENEDICTO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY

unit b passes to subparallel disrupted reflectors, locally to chaotic facies, making difficult the differentiation of the three groups. Unit b corresponds to the Oligocene deposits. The bi group correlates with the "S6rie Grise", the bxi group with the "S6rie Rouge" and the bxi I group with the "S6rie Calcar6o- salif•re". Discontinuity 2 corresponds to the boundary between the Oligocene and Aquitanian series (Figure 5d).

Unit c (sequence 6 by Valette[ 1991 ]) onlaps discontinuity 2 toward the $E and is truncated upward by the erosional discontinuity 3. Unit c is characterized by a transparent seismic facies with a few bright reflections of high-amplitude at the top, and a few easily correlatable high amplitude reflectors in the middle part. It shows rapid lateral seismic facies changes. In the southern part of the Vistrenque graben, unit c shows a NW divergent pattern, while in the southern part shows a subparallel one. This unit c corresponds to the Aquitanian deposits, and the erosional discontinuity 3 corresponds to the breakup unconformity defined by Valette [ 1991 ] and Valette and Benedicto [ 1995] (Figure 5e).

Unit d (sequences 7 and 8 by Valette [1991 ]) is concordant with or onlaps discontinuity 3 and is truncated upward by the erosional discontinuity 4. The lower part comprises a group of bright subparallel and continuous reflectors of moderate amplitude. The upper part shows hummocky reflections and frequently disrupted configuration. The lower part corresponds to the transgressive Burdigalian postrift deposits [Gorini et al., 1993] and the upper part to the post-Burdigalian Miocene succession (Langhian to Tortonian). Discontinuity 4 corresponds to the regional Messinian erosional surface.

Unit e (sequence 9 by Valette [1991]) onlaps discontinuity 4. It is mainly transparent in $G lines and shows subparallel discontinuous reflectors in VG lines, with local hummocky clinoforms. Unit e corresponds to the Plio-Quaternary succession.

Geometry of the N•mes fault. Truncated reflectors of the hangingwall and footwall, as well as reflectors corresponding to the fault itself allow accurate seismic identification of the Ntmes fault in the whole surveyed area (Figures 5a and 6).

The upper part of the N•mes fault, down to about 2 s (twt), is a SE dipping steep surface (about 70 ø in depth-converted sections; Figure 6) defined by truncated reflectors of the hangingwall (Oligocene, Aquitanian, and post-Aquitanian Miocene) and footwall (pre-Oligocene unit) (SG2, SG3, SG4, and SG 5 in Plate 1).

The deep portion of the N•mes fault is a $E dipping low- angle surface (about 25 ø in depth-converted sections; Figure 6) defined by (1) reflectors related to the fault itself, visible in all lines between about 2 and 3 s (twt), and (2) truncated reflectors of the hangingwall and footwall. In the footwall, most of the lines show bright continuous subhorizontal reflectors between 2 and 3 s (twt) beneath the central part of the Vistrenque graben. To the SE, these reflectors are either truncated by the SE dipping reflectors of the N•mes fault or interrupted against the NW-tilted hangingwall reflectors (e.g., $G3 and 4 in Plate 1). In the latter case, the limit between both groups of reflectors is aligned with the N•mes fault reflectors. The truncation of the deep horizontal reflectors thus defines a footwall ramp in the low-angle part of the fault. The deep horizontal reflectors are interpreted as being within or at the

top of the Paleozoic series, based on lateral correlations between seismic lines and the Castries and La Jassette wells (C and LJ in Figure 3) in the footwall. Northwest of the N•mes fault (footwall), these reflectors are segmented and their upper envelope gently dips southeastwards, suggesting that the Paleozoic is lowered southeastward by one or several steeply dipping normal faults (SG 3, 4 and in Plate 1).

The N•mes fault is thus characterized in seismic profiles by a steep SE dipping ramp in the upper part and by a low-angle (25 ø) SE dipping ramp which affects the Paleozoic basement at depth. Depth-converted sections show that the low-angle ramp is defined down to about 8 km (Figure 6). Isobath mapping (Figure 5a) shows that the low-angle ramp is affected by along-strike depth changes which define NW-SE trending lateral ramps. The regional direction of extension indicates that the N•mes fault ramp was a frontal ramp in the rift fault system (Figure lb).

Contrasting hangingwall geometries. T w o distinct geometries of the Oligocene-Aquitanian basin fill and its pre-Oligocene substratum are seen on the seismic lines and isopach/isobath maps (Figures 5 and 6).

(1) In the southern part, the Vistrenque graben shows an asymmetric geometry corresponding to a hangingwall rollover structure of the basin substratum and Oligocene- Aquitanian infill units against the N•mes fault. On the southeastern side of the graben, the Oligocene-Aquitanian basin fill dips northwestward and displaying a pattern divergent toward the NW (VG4-VG7, SG4-SG9 and SG5-SG10 in Plate 1). The NW dipping reflectors of the Oligocene unit abut against the low-angle part of the N•mes fault, while the Aquitanian unit dips slightly to the SE against the steep part of the fault, defining a slight syncline truncated upward by the post-Aquitanian unconformity (SG4 in Plate 1). At the crest of the rollover, reflectors are offset by a SE dipping fault forming the northwestern boundary of another asymmetric half graben, the Petit Rh6ne graben (SG9 and 10 in Plate 1). Details of its geometry are masked by seismic artefacts related to velocity contrasts associated with the Messinian erosional surface, but

seismic correlations define the Petit Rh6ne graben infill as Oligocene in age (group bi). The Oligocene-Aquitanian basin fill units extend southeastward beyond the rollover crest with a horizontal pattern and reduced thickness.

(2) In the northern part of the Vistrenque graben, the basin fill units also form a slight syncline, but with a different, more symmetric geometry than in the southern part (SG2 and SG3 in Plate 1). In the axis of the graben, the Oligocene-Aquitanian units shows a subparallel and subhorizontal pattern. In the SE, both the pre-Oligocene and Oligocene-Aquitanian units show a confused image with disrupted reflectors resulting from faulting and dip slightly to the NW. Dipmeter logs from the Pierrefeu and Parrapon wells show that faults are parallel, closely spaced and dip 70 ø to the NW [Valette, 1991; Valette and Benedicto, 1995]. In the NW, the Oligocene-Aquitanian units dip slightly to the SE and are affected by steep (70 ø) SE dipping normal faults synthetic with the N•mes fault.

In the whole Vistrenque graben, the NW dipping basin fill units, on the southeastern side, are also affected by several antithetic d6collements which correspond to the deel• part of NW dipping listric faults (VG7, SG9, and SG10 in Plate 1). In the southern part of the graben, the associated minor rollovers

BENEDICTO ET AL.: EXTF3qSIONAL CRUSTAL RAMP AND BASIN GEOMETRY 1201

in the Oligocene-Aquitanian units are truncated by the post- Aquitanian unconformity, indicating that most of the movement occurred during the Oligocene-Aquitanian extension. However, the base of the post-Aquitanian succession is locally affected by late movements of the listric faults (VG7 in Plate 1). In the northern area, dtcollements are not visible on the seismic lines, but have been revealed by the detailed study of the Parrapon salt exploitation area [Valette, 1991; Valette and Benedicto, 1995].

The contrasting geometries of the Oligocene-Aquitanian basin fill observed along the strike of the Nimes fault in seismic lines allow to differentiate three compartments in the Vistrenque graben, namely, from the SW to the NE, the Marette, Marsillargues, and Vauvert compartments (Figure 3). The Marette and Marsillargues compartments are characterized by the divergent Oligocene-Aquitanian basin fill above a rollover. They differ by the thickness of the Oligocene unit, thinner in the Marette compartment. The Vauvert compartment is characterized by a horizontal pattern of the Oligocene- Aquitanian reflectors in the axis of the graben, and steep basinward dipping faults on the SE border of the graben.

The post-Aquitanian succession is not affected by hangingwall tilting and graben compartimentalization. This subhorizontal succession rests unconformably over previously tilted and eroded underlying units. Although it is thicker above the Vistrenque graben, it extends out of it, rapidly thinning up to disappear northwestward above the footwall of the Nimes fault (Figure 5e). A recent unpublished seismic survey in the northern part of the Vistrenque graben suggests that the upper part of the fault truncates the post- Aquitanian Miocene unit and is sealed by the Messinian erosional surface (M. Perrissol, personal communication, 1995).

Transfer zones between the Vistrenque graben compartments. The three compartments differentiated in the graben are separated by zones of accommodation subperpendicular to the Nimes fault (Figures 3 and 5b). The Aigues-Mortes transfer zone separates the Marette and Marsillargues compartments. On the SG6 strike line (Plate 1), it corresponds to a NE-dipping low-angle fault which branches on the Nimes fault. Its activity was responsible for the thicker Oligocene unit in the Marsillargues compartment. It remained inactive during Aquitanian times.

The Gallician transfer zone separates the Marsillargues and Vauvert compartments. On the VG2 strike line (Plate 1), it corresponds to a NW trending anticlinal warping of both the basin substratum and Oligocene-Aquitanian units. Oligocene- Aquitanian reflectors show a slight divergent geometry toward the NE, and the anticline is truncated by the post-Aquitanian unconformity. The anticline was developed above a SE trending lateral ramp of the Nimes fault which gently deepens northward beneath the Vauvert compartment (Figure 5a).

To the NE, the Vauvert compartment boundary corresponds to the Arltsienne transfer zone (Figure lb). The SG1 strike line (Plate 1) shows that this boundary corresponds to a NE upward flexure of both the basin substratum and Oligocene- Aquitanian units. On this flexure, the Oligocene unit presents a divergent geometry toward the SW and onlaps northeastward onto an erosional surface at the top of the pre-Oligocene. Seismic data suggests. that the flexure results from forced

folding of the hangingwall units of a SW dipping lateral ramp of the Nimes fault. Thus, the Vauvert compartment overlies the deepest level of detachment of the Nimes fault.

The strike lines VG2 and SG1 also show that lateral ramps of the antithetic dtcollements which affect the Oligocene- Aquitanian units occur at the level of the Gallician and Arltsienne transfer zones.

4.2. Relationships Between the Nimes Fault Profile and Hanging Wall Geometry

On seismic lines (Plate 1), the N•mes fault is characterized in the whole Vistrenque graben by a low-angle profile at depth. However, hangingwall geometry varies (1) along-strike for the basin substratum and Oligocene-Aquitanian units, and (2) vertically, between the Oligocene-Aquitanian and post- Aquitanian Miocene units. In this section, we discuss the relationships between fault profile and variable hangingwall geometries.

For the Oligocene-Aquitanian period, the Marette and Marsillargues compartments, on the one hand, and the Vauvert compartment, on the other hand, show a fundamental difference in hangingwall geometry, which can be explained by along-strike change of the shape of the curvature between the steep and low-angle parts of the N•mes fault.

Flattening at depth of a normal fault may be achieved by either a progressive curvature, that is, a listtic fault sensu stricto, or an abrupt change of dip, that is, a two-segments planar fault (Figure 7). Each case implies a specific geometry of the hangingwall. Movement along a listric fault generates a classical hangingwall rollover geometry and a divergent basin fill toward the fault [e.g., Verrall, 1981; Gibbs, 1983; Davison, 1986; White et al., 1986]. Movement along a two- segment planar fault generates a compensation graben, induced by synthetic and antithetic faults [Faure and Chermette, 1989]. For small offset along the major fault, the basin substratum of the hangingwall forms a symmetric, fiat

TWO-SEGMENTS

PLANAR FAULT LISTRIC FAULT

compensation graben

pseudo-rollover

cl ...... rollover

c2

rollover

Figure 7. Two types of fault shape with their associated hangingwall and basin fill geometries, modified from Faure and Cherrnette [1989]. (a) initial geometry, (b) potential void resulting from hangingwall displacement, and (c) evolution of hangingwall and basin fill geometries.

1202 BENEDICTO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY

0km

NW SE - 0km

I .__ Pliocene _ 1 _

4- 4

6

7 7

8 8

9 9

10 10

Aquitanian Oligocene

el ext = 5 km ext = 4,5 km I • total extension = 9,5 km

e2 =•.total extension = 8,8 km. l,,,, [• 1• • •-•,••,• '-'-' ____ ,__-••__,,,'

2 -

5

Present elevation of the post-Pyrenean erosional surface (suposed topfootwall cutoff)

Depth of the present topfootwall cutoff

/

...... ,,/z""1•=6! H=V

Figure 8. (a) Simplified line drawing of depth-converted lines SG4 and SG9 (by IFP, 1994). (b) Sketch showing calculation of extension in the Vistrenque graben through the cross section b of Figure 6, using the inclined-shear ((• = 60 ø) technique. Calculation is made using the simplified line drawing of a, considering no postrift tectonic movement and after restoration of the Petit Rh6ne graben. el, value of extension using the present elevation of the preserved post-Pyrenean erosional surface to locate the original footwall cutoff (1); e2, minimized value of extension if the present top footwall cutoff visible in seismic data is used (2). Taking a possible maximum postrift vertical tectonic movement for this section of about 500m, extension results between 350m (using 1) and 300m (using 2) smaller. Taking compaction into account for calculating the distribution of extension between the Oligocene and Aquitanian would give a higher value for the Oligocene and a lower value for the Aquitanian extension.

0km

1

2

3

4

5

6

7

8

9

10

bottom syncline, and the basin fill displays an horizontal parallel pattern. When the offset along the major fault is large enough to allow. the basin substratum hangingwall cut off to reach the low-angle segment of the fault, shearing of the hangingwall faulted blocks induces a basinward tilting of the basin substratum resulting in a pseudo-rollover geometry.

In the Marette and Marsillargues compartments, the transition between the steep and low-angle parts of the N•mes

fault is not imaged by seismic lines (SG4 and SG5 in Plate 1), but the hangingwall rollover and divergent basin fill indicates a listric geometry for the fault which passes from 70 ø SE upsection to 250-30 ø SE at depth (Figure 6b and c).

In contrast, the horizontal basin fill and the faulted SE

margin in the Vauvert compartment suggest a two-segments planar geometry of the N•mes fault. This geometry is rather well constrained by hangingwall and footwall cutoffs (SG2 and

BENEDICTO ET AL.: F3('IENSIONAL CRUSTAL RAMP AND BAS1N GEOMETRY 1203

SG3 in Plate 1). The fault dip changes from about 70 ø SE upsection to about 15 ø at depth (Figure 6a). The Oligocene reflectors rest directly on the low-angle segment of the fault, indicating that the hangingwall pseudo-rollover stage has been reached.

The along-strike variation of the Nimes fault shape is well imaged by the isobath map of the Nimes fault (Figure 5a), where the southern listric and the northern two-parts planar fault sectors are well differentiated. The related change in hangingwall geometry is well imaged by the isobath map of the Aquitanian base (Figure 5d) where contours draw a faultward dip surface in the listric fault sector and a syncline pattern in the two-parts planar fault sector.

The Nimes fault differs from the theoretical models of Figure 8 by the fact that the deep part is not horizontal but shows a low-angle inclination. This difference did not influence the structural style of the hangingwall but introduced a vertical component of hangingwall displacement, responsible for the deposition of the Oligocene-Aquitanian sediments SE of the rollover crest.

The SE dip of the Oligocene-Aquitanian deposits along the NW border of the graben is interpreted as being related not to the particular geometry of the fault, but to differential compaction of the underlying sediments [White et al., 1986].

In the three compartments, the antithetic d6collements in the Oligocene-Aquitanian succession are interpreted as the basal surfaces of gravity slides toward the deepest part of the basin, resulting from the hangingwall tilting induced by the Nimes fault.

The distribution and subhorizontal geometry of the post- Aquitanian Miocene cannot be explained by the activity of a low-angle fault at depth. This implies that the low-angle basement ramp of the N•mes fault was no longer active after the Aquitanian. However, the base of the post-Aquitanian deposits is located at 1000 to 1500 m deep in the Vistrenque graben while it is outcropping northwestward on the footwall of the N•mes fault (Figures 5e and 6). This difference must be explained. Three possibilities can be considered: (1) tectonic activity of the steep basement faults observed on seismic lines in the footwail of the N•mes fault ß these deep faults may have propagated upward and reutilized the steep upper part of the Nimes fault, (2) compaction of the Oligocene-Aquitanian sediments, and (3) burial of a topography inherited from the post-Aquitanian erosion (paleo-fault).

Differential compaction of the Oligocene-Aquitanian sediments may explain the gentle syncline geometry of the post-Aquitanian Miocene unit in the center of the graben, but cannot explain its important thickness in the whole area located SE of the Nimes fault including sectors where thickness of the Oligocene-Aquitanian deposits is reduced (compare SG4 and SG5 in Plate 1).

Since the Nimes fault is sealed by the Messinian erosion, tectonic activity of steep deep faults would have only occurred during the post-Aquitanian Miocene, the Pliocene-Quaternary units occupying the space created later by the Messinian erosion. However, respective importance of offset by post- Aquitanian tectonic movement and topographic offset inherited from the post-Aquitanian erosion cannot be differentiated with available data.

Our structural analysis emphasizes the different tectonics of the basin between the Oligocene-Aquitanian and post- Aquitanian periods. This analysis supports the interpretation of the post-Aquitanian discontinuity in terms of "breakup unconformity", separating (1) a synrift Oligocene-Aquitanian episode related to pluri-kilometric extension along a low- angle crustal ramp, from (2) a postrift post-Aquitanian episode characterized by a combination of large-scale erosions (breakup unconformity and Messinian erosion) and vertical accommodation compensated by steeply dipping basement faults. Although precise postrift tectonic movement is difficult to appreciate with available data, its maximum possible vertical amplitude would be about 1000 m (in the south of the graben, Figure 5e), corresponding to a maximum horizontal extension of about 500 m.

4.3. Amount of Synrift (O!igocene-Aquitanian) Extension

The amount of extension calculated on a fault flattening at depth depends on the assumed modes of hangingwall deformation. For instance, assuming hangingwall deformation by simple shear [Verrall, 1981; Gibbs, 1983; White et al., 1986; Faure and Cherrnette, 1989; Dula, 1991], varying the angle of shear (c0 from 45 ø to 90 ø induces differences in the estimated horizontal extension ranging from 100 to 300% [White et al., 1986].

In the case of the Vistrenque graben, we have calculated the value of synrift extension for each compartment by restoring the top of the pre-Oligocene (prerift) succession. We assume a 60 ø inclined shear deformation, compatible with the classical angle of fracturing of rocks [Faure and Cherrnette, 1989] and which provides a first order approximation in calculating net extension (Figure 8).

Whatever mode of deformation of the hangingwall and restoration technique is used, calculating extension in the Vistrenque graben poses two problems: (1) the present top footwall cutoff visible on seismic lines may not correspond to the original cutoff which may have been eroded during extension and the formation of the post-Aquitanian breakup unconformity. Extension calculated using the present top footwall cutoff will be a minimum as is shown in Figure 8b. In the case of the Vistrenque graben, we assume that the original top footwall cutoff was located at 225m above the present sea level (1 in Figure 8b). This level corresponds to the average elevation of the preserved subhorizontal post-Pyrenean erosion surface in the Nimes fault footwall (Vaunage area, Figure 3); and (2) possible post-Aquitanian (postrift) tectonic movement must be restored before calculation of the synrift extension. As this movement is not precisely known, we have calculated the possible extreme values of synrift extension for both minimum (0 m) and maximum (1000 m) possible postrift vertical movements in each compartment.

Calculated extension (Figure 6) is comprised between 8750 m (400 m postrift movement) and 9000 m (no postrift movement) across the Vauvert compartment, 9200 m (500 m postrift movement) and 9500 m (no postrift movement) across the Marsillargues compartment and 3600 m (1000 m postrift movement) and 4400 m (no postrift movement) across the

1204 BENEDICTO ET AL.: EX'IENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY

Marette compartment. Extension has also been calculated in the Petit Rh6ne graben, resulting in values of about 2200 m across the Marsillargues compartment and 4600 m in the Marette compartment.

These values of extension point out the role of the Aigues- Mortes transfer zone which separates two compartments with notably different amounts of extension, explaining the thin Oligocene series in the Marette compartment. On the other hand, the similar values in the Marsillargues and Vauvert compartments indicate that the different structural style between the southern and northern parts of the Vistrenque graben was not related to a change in the amount of extension across the Gallician transfer zone.

5. Structural Inheritance and Origin of the Nimes Fault Low-Angle Crustal Ramp

The Vistrenque graben formed in an area previously affected by Mesozoic extension (SE France basin) and by Upper Cretaceous-Eocene compression (Pyrenean orogeny). The N•mes fault is thought to have acted as a normal fault during the Mesozoic extension and a sinistral strike-slip fault during the Pyrenean compression [Arthaud and Sdguret, 1981]. Evaluation of the role of this structural heritage is important in the understanding of the origin and compartmentalization of the graben. Specifically, we have investigated the possible existence of an inherited topography and the correlation of Tertiary extensional structures with preexisting structures.

5.1. Reconstruction of the Prerift (Pre-Oligocene) Topography

Classical fault reconstruction techniques are commonly used to calculate the fault geometry from the hangingwall geometry. This implies the definition of a prerift level datum, usually assumed to be a horizontal line at the elevation of the top footwall cutoff. In the case of intramontane basins, it is obvious that the prerift surface was not horizontal because of important initial morphology. In this case, reconstruction techniques can be used in an inverse way for the reconstruction of the prerift topography if both the fault and hangingwall geometries are well constrained.

We have applied this approach to the Vistrenque graben in order to evaluate a possible prerift topography inherited from the Pyrenean orogeny (Figure 9). The recent reprocessing and depth conversion of lines SG4 and SG9 (by the IFP, 1994) defined precisely the N•mes fault and hangingwall geometries (Figure 8a). The extension restored is 9500 m corresponding to the maximum horizontal extension as explained before. In a first step, the Petit Rh6ne graben was restored using the simple shear ((• = 60 ø ) technique. According to this reconstruction, the Petit Rh6ne fault is listtic and flattens at

about 4800 m (marly Lias). In a second step, we have reconstructed the prerift topography using three different techniques, the constant bed length [Davison, 1986], constant heave (• = 90 ø) [Verrall, 1981; Gibbs, 1983] and simple shear (• = 60 ø ) [White et al., 1986; Faure, 1990; Dula, 1991] techniques (Figure 9).

topfootwall cutoff

225m Okm

1

2

3

4

5

6

7

8

9

10

- - 2

3

4

5

6

7

8

-9 _

-10

Figure 9. Prerift relief reconstruction through the cross section b in Figure 6 using three different reconstruction techniques: inclined shear ((• = 60ø), constant heave, and constant bed length. (a) Theoretical geometry and depth of the top hangingwall assuming an horizontal prerift level datum at 225 m high and using the inclined shear technique ((• = 60ø). Note the vertical amplitude of the relief in the three cases.

BENEDICTO ET AL.: EXTE2qSIONAL CRUSTAL RAMP AND BASIN GEOMETRY 1205

Present structuration

'• ;-*:-:.:-::•.*.:-:•*':•;•.•.•.:::...'.'- ..:• ...... ..:.'.::x*.;x•:::.-./'•x::;= :•- -- --:3;;':': .J• .:.:.;•;,:;::::::P-:::k:'?:::::•::.:.:•...;'• • '.::.• -

..... • '*%•.*:.L;"X;'::;':.

1206 BENEDICTO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY

major late Hercynian sinistral strike-slip fault outcropping in the Massif Central (Figure lb).

The Gallician transfer zone separates the Lower Cretaceous prerift series in the north from the Jurassic prerift rocks directly overlain by Upper Cretaceous or younger deposits in the south. This transfer zone is aligned with E-W trending, N verging Pyrenean thrusts, the Montpellier thrust to the west, and south-Alpilles thrust to the east, which display the same stratigraphic difference between their footwall and their hangingwall (Figures lb and 10). This suggests that (1) the Gallician transfer zone was controlled by this thrust system, and (2) the latter was formed at the level of a N facing Cretaceous normal fault responsible for the stratigraphic difference between the southern and northern blocks (Saint

Gilles fault in Figure 10). Thickening of the Mesozoic succession and deepening of the top of the Paleozoic basement north of this fault were probably responsible for the formation of the transfer zone, and the difference in extensional structural style between the Vauvert compartment to the north and the Marsillargues compartment to the south.

A similar origin is proposed for the N verging thrusts inferred at the northern edge of the Saintes Maries high (Figure 6c), explaining the northward thickening of the Jurassic succession between the Saintes Maries well (2485 m) and the Albaron 101 well (> 3583 m) (Saintes Maries fault in Figure 10). The westward continuation of this fault may have controlled the southern boundary of the Marette compartment. Although this boundary occurs outside the seismic surveys studied in this paper, its location is inferred from the different structural style between the Marette compartment and the offshore continuation of the Vistrenque graben imaged on the NW-ECORS seismic profile [de Voogd et al., 1991]. On another hand, there is no evidence of a prerift control on the Aigues-Mortes transfer zone.

"•:..."-• Mesozoic cover d6collement domain '• Low-angle basement ramp domain • Zone undeformed during extension

, , Extensional cover fault

• Extensional basement fault

IIII Transfer zone

T Front of basement Pyrenean thrust

==•m Pyrenean basement ramp reactivated as oblique ramp during extension • Direction of extension

Figure 11. Location of the Vistrenque graben with respect to (1) the domains of different extensional structural style of the Gulf of Lion margin, and (2) the Pyrenean thrust front. Modified from Sdranne et al. [1995]. Discussion in the text.

5.3. Origin of the Nimes Fault Low-Angle Crustal Ramp

The low-angle geometry of the Nimes fault at depth is similar to that of the others synrift basement faults observed in seismic lines in the offshore part of the margin [de Voogd et al., 1991, Gorini et al., 1993]. Thus the origin of the NE trending extensional low-angle crustal ramps must be discussed on a regional scale.

At this scale, the basement-faulted domain during rifting matches the area of crustal thickening during the previous Pyrenean orogeny (Figures lb and 11), suggesting that the latter played a major role in controlling the formation of the extensional low-angle crustal ramps [Gorini et al., 1993; Sdranne et al., 1995]. The orogenic lithosphere was probably weakened and extension was preferentially localized in the thickened areas. However, the Pyrenean thrusts were E-W trending, implying that the NE trending extensional crustal ramps cannot correspond to reactivated thrusts. The frontal ramps of the thrusts may have been reactivated as oblique extensional ramps [Sdranne et al., 1995], except in the Corbib. res transfer zone (Figure lb) where the local SW-NE trending of thrusts allowed their reactivation as frontal extensional ramps [Gorini et al., 1991].

Since the Mesozoic faults were probably steeply dipping

basement faults, as discussed above for the Nimes fault, we

conclude that the extensional low-angle crustal ramps were most probably newly formed at the onset of the Oligocene- Aquitanian rifting; the new formation has probably been favored by crustal weakening associated with the previous Pyrenean thickening.

However, the Vistrenque area that forms a triangular zone between the N•mes fault, the Ar16sienne transfer zone and the

Pyrenean frontal thrust, stands as an exception in this structural framework (Figure 11). Although the graben was formed by an extensional basement ramp, it stands outside the area of previous Pyrenean thickening and relief (Figure 9). This suggests that development of the extensional system beyond the thickened area was more stable than reactivation of the frontal thrust in an extensional regime. The extensional system developed 10 to 20 km northwestward until it reached preexisting major discontinuities in the basement; namely, the steeply dipping precursor Nimes fault and the Ar16sienne transfer zone.

6. Discussion

The Vistrenque graben, the deepest depocenter of the Gulf of Lion passive margin, is located (1) at the boundary between

NW

g a)

Castries type series Albaron 101 type series

Triassic

Mesozoic Nimes Fault

Pyrenean relief

N-vergent Pyrenean thrust

Okm

2

3

4

5

6

?

8

9

10

SE 1209

o o

o

b)

c)

d)

c-

e)

c-

f)

Okra

1

2

3

4

5

6

?

8

N-Montpellier basins Vistrenque graben I I

precursor Nfmes fault [• extension 750m displaced and deformed

ramp linking ";"•'• Basin fill cover d•col.-crustal•

• extension 3000m

Basin fill rid;r I

t• extension 750m displaced and deformed

Basin fill

t• extension 5000m .

Petit Rh6ne graben I

[•extension 1100f

• extension 1100m

Okra

2

3

4

5

Okm

1

Okra

1

2

3

4

5

6

7

6

9

10

'•,-----"• .........Basin fill

breakup

rider 2

postrift tectonic movement = 500 m

Figure 12. Model of kinematic relationships between the low-angle basement faulted and cover ddcollement domains.

1210 BENEDICTO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY

the basement faulted and the cover d6collement domains of the

margin, (2) close to the transfer zone separating the continental margin frora the Rh6ne valley intracontinental rift, and (3) in the foreland of Pyrenean crustal thickening (Figure lb). These characteristics make the Vistrenque graben a key structure for understanding major aspects of the origin and kinematics of the margin extensional structures.

6.1. Kinematic Relationships Between the Basement Faulted and Cover D6coilement

Domains of the Margin

The grabens of the north Montpellier area are mainly filled by lower Oligocene sediments (Stampian), but the upper Oligocene (Chattian) is also present in some of them (e.g., the Sommi•res basin, Cavelief et al. [1984]). Although the age of the earliest synrift sediments in the Vistrenque graben is not precisely known, it is Stampian or older, and synrift sedimentation continued through to Aquitanian. This implies (1) that the cover d6collement in the external part of the margin was active during the Oligocene period of activity of the Nimes fault crustal ramp, and (2) that it was rooted in the latter. However, in the Vistrenque graben transect, the total Oligocene extension between the Nimes and C6vennes faults calculated from cross section restoration is only 1.5 km, while it is at least between 9000 and 9500 m on the Nimes fault

(Figure 8). This implies that only a fraction of the Oligocene extension on the crustal ramp was transmitted to the d6collement, whereas the other fraction was transmitted to the

emergent part of the Nimes fault. On the basis of these remarks, a schematic model of

relationships between the Nimes fault and the cover d6collement is proposed on Figure 12. In the NW part of the section, the listric faults detaching into the Triassic level are synthesized as a single "north Montpellier graben" (total extension: 1,5 km). The initial structure (Figure 12a) is characterized by (1) a steeply dipping Mesozoic fault which separates a thin Mesozoic cover in the NW from a thick one in the SE, and (2) an hypothesized Pyrenean thrust responsible for the prerift relief. During extension, the footwall remains undeformed and the hangingwall deforms by antithetic shear at an angle of 60 ø . The amounts of extension applied at each stage are adjusted in order to fit the finite extension determined from the present-day section. The timing of the successive stages is constrained by the recognition of syntectonic sedimentation in the different basins. Stages b, c, and d correspond to the Oligocene period, stage e corresponds to the Aquitanian, and stage f corresponds to a possible postrift (post-Aquitanian Miocene) movement. For simplicity, we have considered that the Petit Rh6ne graben was formed independently from the Nimes fault during the early stages of rifting. In Figure 12, formation of the Petit Rh6ne graben is arbitrarily distributed in stage b.

We infer that during the Oligocene, distribution of the movement between the d6collement and the upper part of the Nimes fault was achieved by alternating episodes: (1) when the d6collement was active and ramped down into the low-angle basement ramp, the Vistrenque graben acted as a ramp basin (hangingwall syncline), without an emerging active fault at its boundary (b and d in Figure 12); (2) when the d6collement was inactive, the Nimes fault propagated upward and breached the

surface and, consequently, the Vistrenque graben acted as a classical half graben (c and e in Figure 12). The new formed upper segment of the fault propagated NW of the previous high-angle segment (through-going normal fault, Withjack et al. [1990]), generating successive "riders" [Gibbs, 1984] (c and e in Figure 12). This is necessary because d6collement activity implies southeastward displacement, deformation and abandonment of the upper part of the Nimes fault. For clarity, only two episodes are represented in Figure 12. However, the lack of visible riders on the seismic lines suggests that actual deformation involved a greater number of episodes, resulting in a greater number of smaller riders, unresolvable by seismic reflection.

The model implies that (1) the NW limit of the basement- faulted domain (Nimes crustal ramp) of the margin remained fixed during the whole rifting period, and (2) this limit formed the landward boundary of the deformation domain during most of the time. Cover d6collement corresponded to short events when a small amount of extension was transmitted to more

external areas. This may have been induced by gravitational instability on the basement slope, possibly associated with changes of regional tilting during margin development (isostatic or thermal adjusttnents).

6.2. Why is the Vistrenque Graben the Deepest Synrift Depocenter of the Margin?

The regional section (Figure 2) shows that the Vistrenque graben is the major synrift basin of the margin along this transect. Southeast of the Vistrenque graben, the major extensional detachments identified in seismic lines under the

continental shelf have not generated large grabens on their hangingwall, whereas the highly stretched areas underlying the continental slope support only thin and discontinuous synrift sedimentation [de Voogd et al., 1991]. This is due to the prerift high surface elevation inherited from the Pyrenean orogeny. If continental basins were formed on the collapsing mountain belt, they were rapidly cannibalized. There is thus little stratigraphic record of the early stages of rifting in the areas of prerift relief [Sdranne et al., 1995]. This may explain the discrepancy between the amount of extension estimated from crustal thickness and that estimated from basin geometry [Burrus, 1987; Steckler and Watts, 1980]. On the other end of the section, in the zone of the north Montpellier basins, sediment thickness was limited by the small amount of extension (1.5 km) and shallow depth of the d6collement.

In contrast, the Vistrenque graben is (1) controlled by a major crustal ramp active during the whole rifting history (9 to 9.5 km of extension), and (2) located in the zone where initial surface elevation progressively sloped down to sea level at the front of the Pyrenean belt (Figure 9), which allowed the accumulation of a complete and exceptionally thick synrift sequence.

Data on Moho depth [Chamot-Rooke et al., 1995; Gaulier et al., 1995] indicate the lack of Moho uplift below the Vistrenque graben (Figure 2) and its offshore continuation [de Voogd et al., 1991]. This is in agreement with the asymmetric structure of the extensional system controlled by low-angle detachments [Lister et al, 1991]. Upper crustal extension on the low-angle ramps transferred basinwards to the zone of continental breakup by an intracrustal detachment or/and distributed in lower crust extension across the whole margin.

BENEDICTO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY 1211

7. Conclusions

Detailed structural interpretation of the subsurface data in the Vistrenque graben demonstrates that the Nimes fault, considered previously as a steeply dipping basement fault, was at depth a low-angle basement ramp during the Oligocene- Aquitanian rifting. This case study of low-angle crustal fault associated with the formation of extensional sedimentary basin adds to the on-going discussion of the activity of normal faults with a shallower dip than predicted by mechanics. The present study emphasizes the role of lateral changes of fault shape in controlling the variations of hangingwall geometry: a listric geometry of the fault in the southern part of the graben generated a rollover and divergent Oligocene-Aquitanian basin fill, while a two-segments planar geometry of the fault in the northern part resulted in the development of a pseudo-rollover and compensation graben. Areas with different structural style are separated by transfer zones superimposed onto preexisting faults inherited from Mesozoic extension and Pyrenean compression. The origin of along-strike varying geometry of the low-angle Nimes fault is difficult to establish. It probably reflects along-strike varying structural patterns across the steeply dipping precursor of the Nimes fault.

In the studied transect, synrift extensional deformation of the margin was achieved by several large basement ramps, the Nimes fault being the landwardmost one. Small amounts of extension were transmitted landward during short intervals of time to cover d6collement probably resulting from gravitational instability during margin collapse.

Kinematic modelling shows that the preexisting (Mesozoic) steeply dipping Nimes fault controlled (1) the location of the ramp linking the landward cover d6collement and the low-angle crustal ramp, and (2) the emergence of the Oligocene-Aquitanian fault; but its upper part was not active

during rifting because it was involved (displaced and deformed) in the hangingwall of the cover d6collement. The lower basement part of the steeply dipping precursor was reactivated during the postrift period, reutilizing the upper part of the emergent Nimes fault, allowing deposition of thick subhorizontal units above the tilted synrift units.

The low-angle crustal ramp of the Nimes fault, as well as the other ramps of the margin of similar orientation, are interpreted as newly formed structures whose formation was allowed by crustal weakening resulting from the previous Pyrenean thickening, rather than reactivated Pyrenean thrusts. Upper crustal extension corresponding to the graben formation was not locally compensated, but transmitted basinward to the zone of continental breakup by an intracrustal detachment, or/and distributed in the lower crust

across the margin. Within the margin, the Vistrenque graben is a distinctive

feature due to the Mesozoic extension and Pyrenean orogeny structural heritage. The SE deepening of the basement across the precursor Nimes fault controlled the emergence of the newly formed extensional ramp; the crustal thickening and surface elevation in the Gulf of Lion made the Vistrenque graben, located outside the orogenic front, the only one whose sedimentary fill records the whole rifting episode, resulting in the deepest depocenter of the margin.

Acknowledgments. This work was supported by the European Community DGXII (contract JOULE II - CEC Project n ø PL 920287 "Integrated Basin Studies"). We are grateful to Coparex, Elf-Aquitaine, Elf-Atochem and ESSO-Rep for providing data for this study and especially to ESSO-Rep and Coparex for authorized publication of line drawings of seismic lines. IFP has reprocessed lines SG2 and SG4. Ideas developed in this contribution benefited from discussions with A. Mascle. We are grateful to A. Gibbs, R. Collier, and an anonymous reviewer for their insightful and detailed comments which improved the paper.

References

Arthaud, F., and P. Matte, Les d6crochements

tardi-hercyniens du SW de l'Europe. G6om6trie et essai de reconstitution des

conditions de la d6formation,

Tectonophysics, 25, 139-171, 1975. Arthaud, F., M. Ogler, and M. S6guret,

G6ologie et g6ophysique du Golfe du Lion et de sa bordure nord, Bull. Bur. Res. Geol. Min. Fr., s. 2 (1), 175-193, 1980.

Arthaud, F., and M. S6guret, Les structures pyr6n6enes du Languedoc et du Golfe du Lion (sud de la France), Bull. Soc. Ggol. Fr., s. 7 (23), 1-63, 1981.

Auzende, J.M., J. Bonnin, and J.L. Olivet, The origin of the Western Mediterranean basin, J. Geol. Soc. London, 129, 607-620, 1973.

Beaudrimont, A.F., and P. Dubois, Un bassin

m6sog6en du domaine p6ri-Alpin: le sud-est de la France, Bull. Cent. Rech. Explor.- Prod. E.If Aquitaine, 1, 261-308, 1977.

Beaufort, L., J. Bruneau, A. Crepin and Y. Julian, Ampleur de l'6rosion pontienne et du comblement plioc•ne en Camargue, Bull. Soc. Ggol. Fr., s. 6 (4), 157-184, 1954.

Bessis, F., and J. Burrus, Etude de la

subsidence de la marge du Golfe du Lion (M6diterran6e occidentale), Bull. Cent. Rech. Explor.-Prod. Elf-Aquitaine, 10 (1), 123-141, 1986.

Biju-Duval, B., J. Letouzey, and L. Montardet, Variety of margins and deep basins in the Mediterranean, AAPG Bull., 29, 293-317, 1979.

Burrus, J., Contribution to a geodynamic synthesis of the Provencal basin (North- western Mediterranean), Mar. Geol., 55, 247-269, 1984.

Burrus, J., Review of geodynamic models for extensional basins; The paradox of stretching in the gulf of Lions, Bull. Soc. Gdol. Fra., s. 8 (5), 377-393, 1987.

Cavelier, C. (coord.) et al., Pa16og•ne, in Synth•se Gdologique du Sud-Est de la France, edited by S. Debrand-Passard, S. Courbouleix, and M.J. Lienhardt, PP. Bull. Res. Geol. Min. Fr., 125, 389-468, 1984.

Chamot-Rooke, N., F. Jestin, R. Vially, and J.M. Gaulier, Finite kinematics of the

Liguro-Proven•al Basin, AAPG lnt. Con. f, Nice, 1995.

Closs, E., Experimental analysis of Gulf Coast fracture patterns, AAPG Bull., 52, 420-444, 1968.

Crochet, J.Y., G6ologie et pa16ontologie de la partie septentrionale du foss6 oligocb•ne des Matelles (H6rault, sud de la France), Ggol. Fr., 1, 91-104, 1984.

Curnelle, R., and P. Dubois, Evolution des

grands bassins s6dimentaires fran•ais' bassins de Paris, d'Aquitaine et du Sud-Est, Bull. Soc. Ggol. Fr., 8, 4, 526-546, 1986.

Davison, I., Listric normal fault profiles: calculation using bed-legth balance and fault displacement, J. Struct. Geol., 8, 209- 210, 1986.

de Voogd, B. et al., First deep seismic reflection transect from the gulf of Lions to Sardinia (ECORS-CROP profiles in western Mediterranean), in Continental Lithosphere: Deep Seismic Reflections, Geod. Ser., vol. 22, edited by PP. AUG. Washington, D.C., 1991.

1212 BENEDICTO ET AL.: EXTENSIONAL CRUSTAL RAMP AND BASIN GEOMETRY

Demarq, G., Etude stratigraphique du Miocene rhodanien, Mdm. Bull. Res. Geol. Min. Fr.,

61,257 p, 1970. Dula, W.F., Geometric models of listric normal

faults and rollover folds, AAPG Bull., 75, 1609-1625, 1991.

Faure, J.-L., and J.-C. Chermette, Deformation

of tilted blocks, consequences on block geometry and extension measurements, Bull. Soc. Gdol. Fr., s. 8 (5), 461-476, 1989.

Gaulier, J.M., N. Chamot-Rooke, and F. Jestin,

Analysis of the isostatic equilibrium in the Gulf of Lion (France), Implications for the geodynamic and thermal evolution, AAPG Int. Conf., Nice, 1995.

Gibbs, A.D., Balanced cross section construction from seismic sections in areas

of extensional tectonics, J. Struct. Geol., 5, 153-160, 1983.

Gibbs, A.D., Structural evolution of

extensional basin margins, J. Soc. Geol. London, 141,609-620, 1984

Gorini, C., P. Viallard, and J. Deramond,

ModUle d'inversion structurale n6gative: La tectonique extensive post-nappe du foss6 de Narbonne-Sigean (Corbinres, SE France), C.R. Acad. Sci. Paris, s. 2 (312), 1013-1019, 1991.

Gorini, C., A. Le Marrec, and A, Mauffret, Contribution to the structural and

sedimentary history of the gulf of Lions (western Mediterranean), from the ECORS profiles, industrial seismic profiles and well data, Bull, Soc. Gdol. Fr., 164, 353-363, 1993.

Groshong, R.H., Half graben structures: Balanced models of extensional fault-bend

folds, Geol. Soc. Am. Bull., 101, 96-105, 1989.

Guennoc P, N. Debaglia, C. Gorini, A. Le Marrec and A. Mauffret, Anatomie d'une

marge passive jeune (Golfe du Lion - Sud France) : Apports des donn6es g6ophysiques, Bull. Cent. Rech. Explor.- Prod. Elf Aquitaine, 18, 33-57, 1994.

Ellis, P.G., and K.R. McClay, Listric extensional fault systems - results of analogue model experiments, Basin Res., 1, 55-70, 1988.

Jackson, J., and D. MCKenzie, The

geometrical evolution of normal fault systems, J. Struct. Geol., 5, 471-482, 1983.

Lefevre, D., Evolution morphologique et structurale du Golfe du Lion: Essai de

traitement statistique des donn6es, th•se doctorat 3 ø cycle, Univ. Paris VI, 1980.

Lister, G.S., M.A. Etheridge, and P.A. Symonds, Detachment models for the formation of passive continenetal margins, Tectonics, 10 (5), 1038-1064, 1991.

Maerten, L. and M. S6ranne, Extensional tectonics in the Oligo-Miocene H6rault basin (SE France), Gulf of Lion, Bull. Soc. Gdol. Fr., 6, 739-749. 1995.

Maillard, A., and A. Mauffret, Structure et volcanisme de la fosse de Valence, Bull. Soc. Gdol. Fr., 164, 365-383, 1993.

Masse, J.-P., P.-J.L. Masse, and G. Tronchetti, Variations s6dimentaires sous contr61e

tectonique durant l'Aptien sup6rieur- C6nomanien moyen ?• l'articulation des blocs provengal et languedocien (SE de la France): cadre pa16oc6anographique et implications pa16og6ographiques, Bull. Soc. Gdol. Fr., s. 8 (6), 963-971, 1990.

McClay, K.R., and P.G. Ellis, Analogue models of extensional fault geometries, in Continental Extensional Tectonics, edited by M.P. Coward, J.F. Dewey, and P.L. Hancock, Geol. Soc. Speci. Pub., 28, 109- 125, 1987.

Mitchum, R.M., Seismic stratigraphy and global changes of sea level, 1: Glossary of terms used in seismic stratigraphy, in Seismic Stratigraphy - Applications to Hydrocarbon Exploration, edited by C. E. Payton, AAPG Mem., 26, 205-212, 1977.

Monier, P., and S. Ferry, Mise en 6vidence d'un haut-fond pr6-urgonien dans le Barr6mien du mont Ventoux, r61e s6dimentaire de la faille de Nimes, Bull.

Soc. Gdol. Fr., s. 8 (3), 191-193, 1987. Olivet, J.L., P. Beuzard, J.M. Auzende and J.

Bonnin, Cin6matique de l'Atlantique Nord et Central, Publ. Cent. Nat. de la Rech. Sci., 54, 1-107, 1984.

Rehault, J.P., G. Boillot, and A. Mauffret, The western Mediterranean basin geological evolution, Mari.. Geol., 55, 447-477, 1984.

Route, F., J.-P. Brun, B. Colletta, and J. Van Den Driessche, Geometry and kinematics of extensional structures in the Alpine Foreland Basin of southeastern France, J. Struct. Geol., 14, 503-519, 1992.

Ryan W.B.F and M.B. Cita, The nature and distribution of Messinian erosional surface.

Indication of a several-kilometer-deep Mediterranean in the Miocene, Mar. Geol., 2 7, 193-230, 1978.

S6ranne, M., A. Benedicto, C. Truffert, G. Pascal, and P. Labaume, Structural style and evolution of the Gulf of Lion Oligo- Miocene rifting : Role of the Pyrenean

orogeny, Mar. Petrol. Geol., 12 (8), 809- 820, 1995.

Steckler, M.S. and A.B. Watts, Nature, The Gulf of Lion: Subsidence of a young continental margin, Nature, 28, 135-430, 1978.

Tempier, C., ModUle nouveau de mise en place des structures provenqales, Bull. Soc. Gdol. Fra., s. 8 (3), 409-628, 1987.

Valette, M., Etude structurale du gisement salif'ere oligoc•ne de Vauvert (Gard), th•se doctorat, Univ. Montpellier II, 1991.

Valette, M. and A. Benedicto, Chevauchements gravitaires halotectoniques dans le bassin distensif de Camargue (Marge du Golfe du Lion, SE de la France), Bull. Soc. Gdol. Fr., 166 (2), 137-147, 1995.

Verrall, P., Structural interpretation with application to North Sea problems, Joint Assoc. œor Petrol. Explor. Courses (UK), 3, 1981.

Wernicke, B., and B.C. Burchfiel, Modes of extensional tectonics, J. Struct. Geol., 4, 105-115, 1982. •

White, N.J., J.A. Jackson, and D.P. McKenzie,

The relationship between the geometry of normal faults and that of the sedimentary layers in their hangingwalls, J. Struct. Geol., 8, 897-909, 1986.

Withjack, M. O., J. Olson, and E. Peterson, Experimental models of extensional forced folds, AAPG Bull., 74, 1038-1054, 1990.

Withjack, M. O., Q. T. Islam, and P.R. La Pointe, Normal faults and their hangingwall deformation: an experimental study, AAPG Bull., 79, 1-18, 1995.

Xiao, H., and J. Suppe, Origin of rollover, AAPG Bull., 76, 509-529, 1992.

A. Benedicto, M. S6guret; and M. S6ranne, G6ofluides-Bassins-Eau (UMR 5569), CNRS-

Universit6 Montpellier II, cc.57, 34095 Montpellier Cedex 5, France. (e-mail: toni @ dstu.univ-montp2.fr; bassins @ dstu.univ- montp2.fr; seranne@ dstu.univ-montp2.fr)

P. Labaume, Laboratoire de G6ophysique Interne et Tectonophysique, CNRS-Universit6 Joseph Fourier, BP 53, 38041 Grenoble Cedex 9, France. (e-mail: plabaume@lgit.observ- gr.fr)

(Received September 12, 1995; revised March 29, 1996;

accepted April 1, 1996.)