© C. Thomasset et al., Published by EDP Sciences 2024

1 Introduction

On 11th November 2019, the Le Teil earthquake in southern France produced a 5 km long surface rupture with a 10 cm average reverse displacement (Ritz et al., 2020) and strong ground accelerations (Causse et al., 2021). The event occurred along the Rouvière fault, one of the faults belonging to the Northeastern Cevennes Faults System (NCFS) that bounds the Southeast Basin in France (Fig. 1). Prompt investigations from the scientific community after the event provided new insights that elucidated the earthquake's origin and associated deformation (Ritz et al., 2020; Ampuero et al., 2020; Cornou et al., 2021; Vallage et al., 2021; Causse et al., 2021; Delouis et al., 2022).

One of the critical questions that arose was whether any other faults of the NCFS could localize the same kind of event. This question strongly depends on the accuracy of available surface geological maps, especially the interpreted geometry and history of the NCFS. Recent studies in the epicentral area updated these geological parameters and provided cross sections based on detailed analysis of geological maps locally (Ritz et al., 2020; Vallage et al., 2021). Marconato et al. (2022) improved these initial cross-sections by using well and outcrop data allowing integration into a GeoModeller, producing a 3D geological-fault model within the epicenter area. At the same scale, a second fault model was proposed by Burnol et al. (2023) using well and seismic reflection data.

Although these two structural models provided crucial information about the geometry of major faults within the epicentral area, their scale remains too local to constrain the geometry of the NCFS and reconstruct its complex evolution at the scale of the Southeast Basin. Moreover, the numerous sedimentological studies that characterize the age and thicknesses of depositional sequences are also insufficient for reconstructing the NCFS tectonic evolution at the scale of the Southeast Basin (Sornay, 1950; Pascal, 1959; Cotillon et al., 1979; Malartre, 1994; Elmi et al., 1996; Razin et al., 1996; Ferry et al., 2022).

In this paper, we provide new interpretations of deep seismic reflection profiles to constrain the geometry of the different faults composing the segmented NCFS. These data allow updating the interpreted geology of the Vivaro-Cévenol area (Fig. 1C) and the structures that have affected the Southeast Basin since the Mesozoic.

Fig. 1 (A) Location of the Southeast Basin in France. (B) Tectonic map showing main structures and depth of the Southeast Basin. (C) Geological map of the northern part of the Southeast Basin (modified from the 1/50,000 map of the BRGM) and the segmentation of the NCFS. CF: Cevennes fault, NF: Nîmes fault, MDF: Moyenne Durance fault, NCFS: Northeast Cévennes Fault System, VF: Valence fault, VT: Vercors thrust, LT: Lance thrust, VLT: Ventoux-Lure thrust, LuT: Luberon thrust, DN: Digne nappe, S.S: Saou syncline, S.A: Saou anticline, D.S: Dieulet syncline, P.A: Puygiron anticline, L.A: La Lance anticline, D.A: Donzère anticline, R.A : Roussas anticline and SR.A: Saint Remèze anticline.∗: dashed line for unknown location of fault. (A) Localisation du Bassin du Sud-Est en France. (B) Carte tectonique montrant les principales structures et la profondeur du Bassin du Sud-Est et sa profondeur. (C) Carte géologique de la partie nord du Bassin du Sud-Est (modiée à partir de la carte au 1/50,000 du BRGM) et la segmentation du NCFS. CF: faille des Cévennes, NF: faille de Nîmes, MDF: faille de la Moyenne Durance, NCFS: Système de failles Nord-est des Cévennes, VF: faille de Valence, VT: chevauchement du Vercors, LT: chevauchement de la Lance, VLT: chevauchement du Ventoux-Lure, LuT: chevauchement du Luberon, DN: nappe de Digne, S.S: synclinal de Saou, S. A: anticlinal de Saou, D.S: synclinal de Dieulet, P.A: anticlinal de Puygiron, L.A: anticlinal de La Lance, D.A: anticlinal de Donzère, R.A : anticlinal de Roussas et SR.A: anticlinal de Saint Remèze.∗: ligne pointillée pour localisation inconnue de la faille.

2 Geological setting

2.1 Geological evolution of the Vivaro-Cévenol and Vocontian domain

Within the Southeast Basin, our study focuses on the Rhône river valley around the city of Montélimar in France, at the transition between the Vivaro-Cévenol and Vocontian geological domains (Fig. 1). These domains form the northern part of the Southeast Basin. They are bounded to the west by the French Massif Central and to the east by the Alpine foreland thrusts (Fig. 1A and B). The geological and tectonic history of this region comprises several major phases of deformation that produced a complex network of faults, including the NCFS, and fault remobilization over the last 320 million years (Fig. 2).

At the end of the Palaeozoic, the tectonic collapse of the Variscan orogenic belt was characterized by extensional tectonics. During the Stephanian (∼300 Ma), continental lacustrine sediments rich in organic matter (coal deposits) were deposited in basins along major faults whose regional orientation called the “Cévenol strike” varied between N020°-40°E (Arthaud and Matte, 1975; Quenardel et al., 1991, Ballèvre et al., 2018). From the Upper Triassic to the end of the Upper Jurassic, the evolution of the Southeast Basin was intricately shaped by Tethysian rifting, unfolding across two major tectonic phases (Bonijoly et al., 1996). From the Upper Triassic (204 Ma) to the Lower Jurassic (175 Ma), the extensional reactivativation of NE-SW (Cévenol direction) and NW-SE (Vellave direction) faults affected the Vivaro-Cévenol margin (Razin et al., 1996; Elmi et al., 1996). The activity of these faults led to a syntectonic sedimentation of evaporites and then dolomites. During the Lower Jurassic (200 Ma), the margin is structured into a mosaic of tilted blocks that begin deepening the basin towards the east (Dumont, 1988; Elmi, 1988). The peak of this tectonic subsidence occurred between the Toarcian and Bathonian (late Lower Jurassic), with a sharp increase in sedimentation between Aubenas and Alba-la-Romaine (Fig. 1) where a thick sequence of limestones and marls was deposited (Fig. 2). The Callovian (165 Ma) marks the transition from the syn-rift phase to the post-rift phase, and thermal subsidence then controlled the dynamics of the basin (Razin et al., 1996). During the Kimmeridgian, a carbonate platform (Portlandian facies, 160–150 Ma) develops throughout the Vivaro-Cévenol domain with a constant thickness and caps the previous fault (Delfaud et al., 1975, Elmi et al., 1996).

The Lower Cretaceous deposits (145–112 Ma) correspond to most of the geological units observed outcropping in the Vivaro-Cévenol domain. They are characterized by variable palaeoenvironments. This variability is notably observed within the eastern part of the Vivaro-Cévenol domain where the Vocontian Trough, a diverticulum of the Tethys Ocean initiated in the Upper Jurassic, developed during this Lower Cretaceous period (Curnelle and Dubois, 1986; Ferry and Flandrin, 1979). Within this region, the Vivaro-Cévenol domain recorded significant marly pelagic sedimentation during the Valanginian (140 Ma). Within the western part of the Vivaro-Cévenol domain, a progradation process made the sedimentation evolving from marly limestone deposits during the Hauterivian (135 Ma) to the Urgonian limestone facies during the Upper Barremian (125 Ma; Cotillon et al., 1979; Ferry and Flandrin, 1979; Elmi et al., 1996). However, this period appears to have been affected by a potential extensive tectonic reactivation until the Barremian (Elmi et al., 1996). Marine regression continued during the Albo-Aptian period (120–100 Ma), during which black sandy marls and then coastal sands were deposited over the entire northern part of the Southeast Basin. It is associated with a series of unconformities up to the Cenomanian that are sensitive to vertical movements in the area. These movements occurred following the formation of the Durancian Isthmus associated with the opening of the Gulf of Biscay on the Atlantic margin (Figs. 1A, 2; Ferry, 1999; Schettino and Turco, 2011; Schreiber et al., 2011; Angrand and Mouthereau, 2021).

The Upper Cretaceous (100–85 Ma) is poorly represented in the Vivaro-Cévenol region. It is mainly composed of conglomeratic detrital sediments resulting from the erosion of proximal reliefs (Sornay, 1950; Pascal, 1959). Those reliefs correspond to E-W trending anticlines (Puygiron and Donzère, Fig. 1C) due to a N-S shortening throughout the Southeast Basin (Fig. 1C) associated to the Pyreneo-Provencal phase (Balansa et al., 2022). On a local scale in the Couijanet syncline, discordant sequences, such as the Ceno-Turonian (95 Ma) on the Lower Cretaceous occurred during this compressional stage (Elmi et al., 1996). Triassic salt units and diapirs seemed to have played an important role during the inversion of the formation of the Saou and Dieulefit synclines associated to the Pyrenean shortening (Fig. 1C; Martinod, 1988). The total shortening of the sedimentary cover associated with this compressional regime is estimated at 20–25 km around Pierrelatte and the Valréas basin (Ferry, 1999). Note that Bergerat (1982) considers the compressional deformation observed in the region to be mainly associated with the paroxysm of Pyrenean tectonics during the Eocene. This interpretation is consistent with the shortening measurements performed in the Provencal part of the South-East basin. A total horizontal shortening of 38 km has been estimated, but this is strongly attenuated to the north at the level of the Lubéron thrust, which is far to the south of our study area (Balansa et al., 2022).

After a long period of emersion from the Turonian (90 Ma) to the Upper Eocene (40 Ma, Fig. 2), the western margin of the Southeast Basin underwent an ESE-WNW extensional phase associated to the Gulf of Lion rifting (Seranne et al., 2002). During this period, faults involved in the NCFS controlled the formation of Oligocene basins of unknown age (35–25 Ma). Thick sedimentary sequences deposited in Valence and Alès, but they are found only locally between Le Teil and Saint Montan (Fig. 1C). This feature could be due to a transfer zone between Alès and Valence basins, along which the deformation was less intense (Bergerat, 1982). In the Montélimar area, Oligocene deposits correspond to fluvial deposits (conglomerates) and ocher sands.

During the Upper Miocene (15 Ma), the Alpine tectonics induced the inversion of many pre-existing faults in the Vocontian domain, under an overall E-W compressional regime. The frontal thrust of the Vercors and the Lance anticline (Fig. 1C) deform the marine Miocene sediments (Chenevoy et al., 1977; Bergerat, 1982; Elmi et al., 1996; Debelmas et al., 2004). In the area of Montélimar and further west, it is difficult to identify clear Alpine tectonic structures (Marconato et al., 2022). At this time, the alkaline volcanism of the Coirons developed along a network of fractures oriented N120°E, in the Vellave direction. This could be consistent with a NW-SE shortening, favoring the opening of fractures (Bandet et al., 1974, Thierry et al., 2014). During the Upper Miocene (5–6 Ma), a large canyon along the N-S Rhône river axis incised throughout the Southeast Basin due to the Messinian salinity crisis. Between Saulce-sur-Rhône and Pierrelatte (Fig. 1C), the canyon reached about 300 m deep (Beaudoin et al., 1997; Colson et al., 2000), but its exact geometry is not known. The canyon was rapidly filled up with clays and sands during the Lower Pliocene (4 Ma). This stage was followed by a marine regression during which a fluvial terraces system emplaced along the Rhône river valley between the Upper Pliocene (∼3 Ma) to the present. Morphotectonic work within a Rhône river tributary cutting through the NCFS suggests a Quaternary uplift rate of ∼0.6 mm/yr during the last 60 ka, based on the age obtained from a terrace (Cathelin et al., 2024; in review). According to Mandier (1988), this uplift was also recorded on the Rhône terraces and would be controlled by neotectonic deformation of the Donzere anticline.

Fig. 2 Stratigraphic chart and geodynamic context of the study area. Charte stratigraphique et contexte géodynamique de la zone d’étude.

2.2 Geometry and kinematics of the NCFS

Geological structures involved in the NCFS are somewhat complex. They are characterized by numerous fault segments extending from the Vivaro-Cévenol domain to the Rhône river valley (Fig. 1C). In contrast to its basin-scale significance in the Alès area, where it forms the Southeast Basin boundary, the NCFS cuts exclusively Cretaceous terrains in the region under study. Previous studies allowed describing the multistage activity of the NCFS since the formation of the Southeast Basin. The local increase of sedimentation in the hanging wall of the faut system observed in wells shows that the NCFS was already active during the Middle Jurassic (Razin et al., 1996). Similarly, in Upper Barremian deposits, strong variations in the thickness of the limestones between Alba-la-Romaine and Viviers are interpreted as the result of the NCFS activity during this period (Elmi et al., 1996). At the same time, the Nîmes fault (Fig. 1B), running in the same NE-SW direction as the Cévenol fault system, was active and affected the Urgonian platform with left-lateral apparent displacement (Ferry, 1999, Homberg et al., 2013). Note that according to Cotillon et al. (1979), these thickness variations would be related to the structure of the margin, at the interface between the outer and deep domains. There is no large-scale evidence that the NCFS was reactivated during the Pyreneo-Provencal N-S convergence. However, the fault system being directly connected to the EW trending Donzère anticline where Turonian (90 Ma) sediments are strongly folded (Roure et al., 1992) suggests that it could have been reactivated during the Pyreneo-Provencal tectonic activity. In contrast, the NCFS shows clear evidence of its reactivation during the Oligocene with large-scale non-eroded morphotectonic feature as well as fault slip data demonstrating normal displacements (Bergerat, 1982; Ritz et al., 2020; Marconato et al., 2022). Since the Miocene, no evidence of tectonic activity has been described along the NCFS, except along the Marsanne fault between the village of Marsanne and Crest (Fig. 1C), where Lower Miocene sediments are tilted against a fault plane striking N040°E (Pascal, 1959; Chenevoy et al., 1977). The occurrence of M_w 4.9 Le Teil earthquake of 11 November 2019, along the La Rouvière fault, shows that the NCFS must be considered presently-active (Ritz et al., 2020).

This multi-stage tectonic history suggests that the NCFS is controlled by different deformation mechanisms. In a first stage, notably within the Vivaro-Cévenol margin, it would be associated to the Tethysian extension along high-angle faults associated with intra-Carboniferous detachments (Bonijoly et al., 1996). In a second stage, during the Pyreneo-Provencal compression and the Oligocene extension, the formation of large detachments in the Triassic evaporites would have controlled a thin-skinned tectonic process (Roure et al., 1992, Sanchis and Seranne, 2000). This suggests that there is no evidence of crustal fault reactivation during these periods in the study area. According to Roure et al. (1994), the whole NCFS joins at depth within a SSE dipping basement fault between the village of Marsanne and the edge of the Valence basin (Fig. 1).

3 New constraints on the geometry and tectonic history of the NCFS

In the studied region, the NCFS can be described as a set of 5 faults whose strikes are comprised between N030°E and N060°E (Fig. 1C and Tab. 1). These faults are themselves segmented, and the mapping of their segmentation is derived from the 1/50,000 geological map of Aubenas and Montélimar (Elmi et al., 1996; Lorenchet de Montjamont et al., 1978) and the study by Marconato et al. (2022). At the surface, all segments show an apparent normal displacement. The NCFS defines in map a fan-shaped fault pattern with the easternmost faults, such as the Marsanne and Saint Montan faults, displaying a more easterly trend than the Alba and Pontet-de-Couloubre faults. Segmentation seems to decrease towards the northeast. To the southwest of Viviers (Fig. 1C), all the faults converge in a highly segmented corridor less than 5 km wide. Note that a fault not connected to the NCFS, located northwest of the Alba Fault, which we named the “St Pons” fault, is included in the study area. The La Rouvière, Marsanne, and Saint Montan faults most likely cross the Rhône river valley, but they do not show morphological signatures within the plio-quaternary alluvial deposits. In Figure 1C, we therefore mapped them as dashed lines. Other authors already proposed extending the Marsanne fault to the north (Bergerat, 1982; Roure et al., 1992) and this section has been considered as a potential active fault from the historical seismicity recorded within Montélimar (Jomard et al., 2017).

At depth, it is difficult to give a precise value of displacements along the faults of the NCFS, due to the strong thickness variations within the different geological units. However, using data from Marconato et al. (2022), and our own geological cross-sections, we estimated minimum cumulative offsets (Tab. 1). The Pontet-de-Couloubre fault would record the largest displacement, with an apparent normal displacement of about 1 km.

3.1 Data

The aim of this study is to build up a 3D fault model of the NCFS by (1) mapping individual faults and their segmentation beneath the Rhône river valley, (2) characterizing their geometry and relationships, and (3) reconstructing the evolution of the NCFS through time. To do this, we have performed and interpreted several new seismic lines across the entire NCFS. These seismic data have been calibrated using surface geology, well information, and reprocessed industrial seismic lines.

3.1.1 Surface data

To constrain our surface model, we used data from 6 geological maps (Fig. 3). This dataset was supplemented by stratigraphic and dip measurements in the field (see Database 1). The digital elevation model (DEM) used for all topographic settings is the IGN Scan 1 m model with a maximum horizontal resolution of 8 meters.

Fig. 3 Presentation of seismic lines performed by EDF for this study with location of well and reprocessed ancient seismic data. Présentation des lignes sismiques réalisées par EDF pour cette étude avec l’emplacement des puits et les données sismiques antérieures retraitées.

3.1.2 Well data

For the deep drilling data, we used 4 exploration wells and 1 core drill from the Minergies.gouv.fr database (Fig. 3). Except for the core drill, the 4 exploration wells VB-1, VAL-1, SAV-1 and MSN-1 are deep and contain detailed geology (lithology, dip, fault), logs, diagraphies and seismic well velocities (VTS) (see Supp. Mat. 1). The VTS are time/depth data that allow geological markers initially at depth to be converted to two-way travel time (TWT) so that they can be visualized in seismic interpretation (Tab. 2). When the seismic profiles were performed, the Datum Plane (DP) was set at a high elevation of 50 meters. SU-1 and MDG-1 wells, which are not located in the study area, were also included because they contain VTS data for the Upper Cretaceous in addition to the other 4 wells. Using the Bureau de Recherches Géologiques et Minières (BRGM) borehole database (BSS), all shallow boreholes in the Rhône valley were collected to provide the most accurate mapping of the geological units beneath the Quaternary sediments. As a result, these geotechnical boreholes provide only depth-related lithological information, supplemented by a very general description of the formations encountered. The density of these borehole data provides an overall picture of geology of the Rhône river valley around Montélimar beneath the Quaternary (see Supp. Mat. 2A).

3.1.3 Geophysical data

Near the study area, before our own investigations, three historical oil exploration campaigns had been carried out. First, between 1958 and 1961, the Société Nationale des Pétroles d’Aquitaine (SNPA) carried out several oil exploration campaigns. The results of these campaigns have not been published. BRGM reprocessed all these lines in 2008, including line 60-GR-11 (Fig. 3). From these data, only a synthetic map of the ante-Triassic basement structures has been produced (Debrand-Passard et al., 1984). Second, in the framework of the Géologie Profonde de la France Program (GPF), seismic profiles were performed within the region of Balazuc (BA-1, Fig. 3), and published in Roure et al. (1992). Lastly, in 1981–1982, two seismic lines (81-SE-3 and 82-SE-4) (Fig. 3) were performed by the Institut Français du Pétrole (IFP) in the framework of a project called “South-East Synthesis-DHYCA”. In addition, we also used a map of gravity and magnetic anomalies performed by the BRGM in 2022 to better constrain the areas where the Urgonian platform outcrops beneath the Mio-Plio-Quaternary sediments (see Supp. Mat. 2B).

We used these existing data, and 27 new seismic lines carried out by EDF and the “S3” company between 2020 and 2022 (Manchuel et al., 2022). In this study, we present 4 of these seismic lines for a total length of 82.8 km. Their locations are shown in Fig. 3, with 3 of the lines (21-CR-101b/104 and 22-CR-107) intersect the NCFS orthogonally. Line 22-TRI-104 trends N-S and intersects lines 21-CR-104 and 22-CR-107. These two directions were chosen to constrain in 3D the deformations within the NCFS and potentially differentiate the non-coaxial tectonic phases at the origin of the observed 3D structure (Pyreneo-Provençal compression, Oligocene extension, Alpine compression). The acquisition parameters (geophone spacing, recording frequencies) and the seismic source were designed to image structures up to 10 km depth. We also used the 81-SE-3 and 60-GR-11 lines to help us constrain the interpretation of the 4 above-mentioned lines. The processing of our lines and 81-SE-3 was carried out by CDP Consulting.

3.2 Constraining a 3D structural model of the NCFS

3.2.1 Seismic-well correlation

The 3D structural model of the NCFS was constructed using Kingdom Suite® seismic and geological interpretation software. The chosen coordinate system was the French Lambert 93 III geographic projection (ESPG: 2154). We describe hereafter the 3 main steps that we followed. First, we picked geological horizons from reflectors on the seismic lines using geological cross-sections we made from surface, well, and geophysical data. The cross-sections were especially useful within the first kilometers and in areas where the seismic signal was highly degraded (see Supp. Mat. 3). Second, we used the stratigraphic column constrained within the wells and seismic facies (e.g. continuous high amplitude triplet) to pick up seismic horizons at depth. The Upper Jurassic seismic facies (Kimmeridgian/Tithonian) was the main reference for our interpretation because it defines a very regular and continuous high amplitude triplet (Sanchis and Séranne, 2000). Other characteristic reflectors were used, such as the top of the Middle Jurassic, which is defined by a high-amplitude signal separating two transparent facies corresponding to the Middle and Upper Jurassic (Fig. 4). Finally, when available, well data with diagraphy synthetic films were projected on the seismic lines maintaining the orientation of the dip planes between the well and the seismic line (Fig. 4). This operation provided consistent calibration points to interpret the reflectors. In the framework of our study, we used these data obtained within two wells (VAL-1 and SAV-1, located 3.8 km from line 21-CR-104 and 2.3 km from line 21-CR-101b, respectively; Fig. 3). Note that one of the SNPA seismic lines (i.e., 60-GR-18), which passes through the VAL-1 well, intersects our 21-CR-104 line, providing another reliable calibration (Figs. 3 and 4). Finally, we used previous interpretations of lines 81-SE-3 and 82-SE-4 (Martinod, 1988; Roure and Colletta, 1996; Rangin et al., 2010; Chabani, 2019) for comparison with our calibration within the eastern part of our study area.

Fig. 4 Picking of characteristic seismic horizons of the stratigraphic column based on interpolation with the synthetic film computed from the well VAL-1. Pointage des faciès sismiques caractéristiques de la colonne stratigraphique et projection du puit VAL-1 à l’aide d’un film synthétique de vitesse.

3.2.2 Time/Depth chart

To convert the interpreted horizons from time (T) to depth (D), we extracted each horizon from the model and applied the T/D chart separately for each horizon. This allowed faster modification of the depth values and easier integration of the surface calibration points. We constructed the T/D chart using time-depth correlation data from 6 wells close to the study area (Tab. 2). The depths of the wells have been redefined according to the DP. The time-depth correlation values extracted from MAR-1, SAV-1, VAL-1 and VB-1 wells line up according to a second-degree polynomial law given by Eq. (1) (Fig. 5). This law applies for values between the Barremian and the base of the sedimentary sequence. The values of the Upper Cretaceous units given by wells MDG-1 and SU-1 follow a linear relationship given by Eq. (2).

(1)

(2)

Concerning Pliocene deposits found in the Messinian valley, we suggest a velocity of 2000 m/s, which is consistent with recent results from Do Couto et al. (under review, pers. communication).

Fig. 5 Time-depth chart used to convert seismic lines from oil drilling with Eqs. (1) and (2), derived from the Table 2. Loi temps-profondeur utilisée pour convertir les lignes sismiques à partir des forages pétroliers avec l’équation 1 et l’équation 2.

3.2.3 Quality factor

The response of the surface geological units to the vibrating source varies greatly depending on the lithology of the strata that are crossed by the waves during the seismic survey (Fig. 4). For each of the seismic lines, 3 different acoustic responses were observed depending on the outcropping unit. (1) A relatively good response was obtained for Valanginian, Hauterivian and Barremian marls and calcareous marls (Fig. 4, green to yellow zone). Similarly, a good seismic response was obtained for the sands and clays of the Albo-Aptian and the Upper Cretaceous. (2) No seismic response was obtained for the massive and karstic Urgonian limestone (Fig. 4, red zone). (3) A response is observed, but with a significant loss of signal and artefacts for the thick Pliocene clays filling the Messinian canyon. These information allowed us to discuss the reliability of certain reflectors.

3.3 Interpretation

In this section, we present our interpretation of the data in terms of sedimentary sequence and structural pattern within the NCFS. Figure 6 shows depth and thicknesses maps of the geological surfaces encountered along the seismic lines. This helps illustrating the major tectonic events that have shaped the study area.

Fig. 6 Isohypse map of the (a) Top Carboniferous, (b) Top Sinemurian, (c) Top Bajocian, (e) Top Jurassic, (f) Top Barremian and the thickness map of the (d) Lower Jurassic sequence, (g) Lower Cretaceous sequence. Location on Figure 1. 1.AP: Anticline Puygiron, AL: Anticline Lance. 1: Alba Fault, 2: Pt Couloubre Fault, 3: La Rouvière Fault, 4: Marsanne Fault, 6: St Pons Fault, Carte des isohypse du (a) toit du Carbonifere, (b) toit du Sinemurien, (c) toit du Bajocien, (e) toit du Jurassique, (f) toit du Barremien et carte d’épaisseur des couches du (d) Jurassique inférieur, (g) Crétacé inférieur. Localisation sur la Figure 1. AD: Anticlinal de Donzère, AP: Anticlinal de Puygiron, AL: Anticlinal de la Lance. 1: Faille d’Alba, 2: Faille de Pt Couloubre, 3: Faille de La Rouvière, 4: Faille de Marsanne, 6: Faille de St Pons.

3.3.1 Sedimentary sequence

The top of the Palaeozoic coincides with the top-Carboniferous observed within wells BA-1 and VAL-1 (Tab. 2). It corresponds to the base of energetic reflectors (yellow horizon) that are observed between 3000 m at Balazuc (Roure et al., 1992), 5000 m west at Alba-la-Romaine to more than 8000 m east at Montélimar (Figs. 6A and 7). Note that the Permian is not intersected by drilling in the survey area. In the Vocontian area, the numerous tectonic folds degrade the resolution of the seismic lines, increasing the uncertainty of the top-Palaeozoic horizon picking. However, locally, the upper part of this unit shows well-organized reflectors, allowing the interpretation of a Carboniferous basin (Fig. 7). In addition, the reflectors visible at 4 s TWT in Figure 7B are consistent with the interpretation presented by Roure et al. (1992) on line 82-SE-4. At a regional scale, the Carboniferous shows a regular deepening towards the southeast (Fig. 6A). It is not possible to have a precise interpretation to define the thickness of the Carboniferous deposit, which is why the interpretation of the profiles stops at this horizon.

Within the Vivaro-Cévenol domain, the Triassic salt remains thin (<150 m at VAL-1) and therefore does not show a transparent facies. In the eastern part of the depocenter of the Alba sub-basin, the Triassic appears as an alternation of energetic reflectors, which are sometimes difficult to discriminate from the Lower Jurassic (Fig. 7A, between the distances 10,000 and 13,000 m). The top-Triassic horizon is less visible in the eastern parts of profiles 21-CR-104 and 21-CR-101b (Fig. 7), but is thicker in the Vocontian area (81-SE-3; Fig. 7B). Martinod (1988) also interprets a diapir at this level, established during the Middle Jurassic.

The Lower Jurassic (Hettangian-Sinemurian) dolomitic limestones along the Vivaro-Cévenol domain show a homogeneous facies and constant thickness throughout the model. They appear tilted in a few areas (Fig. 7A). During the Middle Jurassic, marly deposits succeeded the dolomitic limestones and their thickness varied considerably during the Pliensbachian-Toarcian period. From west to east, between the VB-1 and VAL-1 wells, the marly Lower-Middle Jurassic increases by more than 1000 meters, before exceeding 3000 meters within the MSN-1 well (Fig. 6D). A NE-SW trending basin is located at Alba-la Romaine, bounded on either side by structural heights (Fig. 6B). Despite the limited overlap of the seismic lines, we also observe this thickening below the Donzère anticline and in the Vocontian domain (Fig. 6B and 6C). Sedimentation seems stabilizing from the Upper Callovian period. This is followed in the Upper Jurassic by a homogeneous and regular sequence characterized by the establishment of the Kimmeridgian reef platform. The Upper Jurassic series, which is very well picked thanks to the characteristic Upper Jurassic facies (high amplitude triplet), is crossed by all the wells in the area (Figs. 6E and 7).

The Valanginian and the Hauterivian layers (marls and limestones) together form the upper part of the interpreted seismic lines (0.25–1 s TWT; Fig. 7). The top of the Hauterivian corresponds to the base of Urgonian limestones (Barremian). These horizons (Valanginian and Hauterivian) are relatively well picked on the seismic lines from surface and well data. In the Vivaro-Cévenol domain, these two formations appear to be rather isopach. On the other hand, in the deep part of the Pierrelatte and Valreas Basin, the Hauterivian and Barremian become 1.5 times thicker with respect to the Vivaro-Cévenol domain (Fig. 6F and G). Using the T/D law, the depth of the top Valanginian and Hauterivian interpreted by Rangin et al. (2010) has been projected to the level of well BO-1 in Figure 8. This also shows a thickening of 1000 meters between well MSN-1 (Vocontian domain) to the north and well BO-1 (Valréas basin) to the south. It is difficult to determine the age of the subsidence that caused the thickening that also affects the Barremian and Urgonian platforms. The interpretation of the top-Barremian, when it is not eroded, gives us information on the thickness variations of this Urgonian reef platform. In the Vocontian domain, this reef limestone facies (Urgonian) does not exist, while it reaches 250 m thick at Viviers, and more than 1000 m within the MDG-1 well (Fig. 8). The rest of the Cretaceous series is poorly observed in the study area (Fig. 7).

The Oligocene deposits are not thick enough to be detected and picked on the seismic lines. On the other hand, the Base-Pliocene unconformity unravels the Messinian incision inside the pre-existing Cretaceous series. It appears to be the most visible reflector on the seismic lines and corresponds to an erosion surface located between 0.15 s TWT south of Marsanne, 0.25 s TWT east of Viviers (Fig. 7). Moreover, the analysis of the gravimetric data available in the region brings accurate information about the geometry of the Messinian Canyon through the study area (see Supp. Mat. 2B). The Upper Miocene-Pliocene filling of the canyon is difficult to pick on seismic line 22-TRI-104 due to the low seismic wave velocity contrast with the Albian units. We used BSS drilling data to adjust our interpretation along this line (Fig. 9 and see Supp. Mat. 2A, B).

Fig. 7 (A) and (B) correspond to seismic lines interpreted and uninterpreted with horizons layers and faults along sections 21-CR-104, and 21-CR-101b - 81-SE-3 transect, respectively. Locations of the seismic lines on the geological map (Supp. Mat. 3). Vertical scale is expressed in second two-way travel (TWT) and horizontal scale is in meter (m), The horizontal/vertical scale has been respected. See cross section on Supplementary material 3. (A) et (B) correspondent aux lignes sismiques interprétées et non-interprétées avec des couches d’horizons et des failles le long des sections 21-CR-104, et 21-CR-101b - 81-SE-3 transect, respectivement. Localisation des lignes sismiques sur la carte géologique. L’échelle verticale est exprimée en secondes temps double (TWT) et l’échelle horizontale est exprimée en mètres (m). L’échelle horizontale/verticale a été respectée.

Fig. 8 Evolution of sedimentation in the Southeast Basin between the Vivaro-Cévenol border and the Valréas basin. The depth setting of borehole BO-1 comes from the interpretation of line 82-SE-4 by Rangin et al., 2010. 1: Alba Fault, 2: Pt Couloubre Fault, 3: La Rouvière Fault, 4: Marsanne Fault, 6: St Pons Fault, 7: Donzère Basement Fault. Evolution de la sédimentation dans le bassin du Sud-Est entre la frontière Vivaro-Cévenol et le bassin de Valréas. La profondeur du forage BO-1 provient de l’interprétation de la ligne 82-SE-4 par Rangin et al. 2010. 1: Faille d’Alba, 2: Faille de Pt Couloubre, 3: Faille de La Rouvière, 4: Faille de Marsanne, 6: Faille de St Pons, 7: Faille du socle de Donzère.

Fig. 9 Seismic line interpreted and uninterpreted with horizons layers and faults along section 22-TRI-104. Locations of the seismic line on the geological map (Supp. Mat. 3).Vertical scale is expressed in second two-way travel (TWT) and horizontal scale is in meter (m), The horizontal/vertical scale has been respected. See cross section on Supplementary material 3. Ligne sismique interprétée et non interprétée avec les horizons, les couches et les failles le long de la section 22-TRI-104. L’échelle verticale est exprimée en secondes de déplacement dans les deux sens (TWT) et l’échelle horizontale est exprimée en mètres (m), l’échelle horizontale/verticale a été respectée. Voir la coupe transversale sur Supplementary material 3 .

3.3.2 Structural pattern

The structural pattern within the NCFS is interpreted from the seismic lines and surface data. The top of the Palaeozoic (top-Carboniferous) shows a general gentle slope towards the southeast with local variations associated with high-angle crustal faults (Fig. 6A). One of these faults called Alba basement fault is located bellow the Alba-sub-Basin and shows a large normal displacement. Two high amplitude reflectors located in Stephanian units or ante-Stephanian units called Cévennes schists (Debrand-Passard et al., 1984), are shifted between 2.5 and 3.2 s TWT along a straight 60°SE dipping plane (Fig. 7A). This pattern was also proposed by Bonijoly et al. (1996) further west at Balazuc. The basement fault appears to be responsible for the over-thickening of the Carboniferous formations in its hanging-wall. However, this data is poorly constrained due to the weak description provided for the bottom of the VB-1 well (Fig. 8 and Supp. Mat. 1). A second basement fault, the Donzère basement fault, is visible Figure 9, characterized by a strong deepening of the top-Carboniferous, which shows a minimum offset of 1500 m below the Donzère anticline (Fig. 6A and B). Assuming that the Donzère basement fault corresponds to that shown in Figure 7B, it is oriented N045°E and defines the boundary between the Vivaro-Cévenol and Vocontian domains at depth. The fault controls the strong thickening observed in the Lower Jurassic and in the Hauterivian/Barremian sedimentary sequence (see Fig. 6D and G) as also observed in Rangin et al. (2010). Beneath Alba la Romaine, the top-Carboniferous is marked by a NE-SW trending depression, bounded to the north-west by the Alba basement fault (Figs. 6A and 7A) and to the south-east by an uplifted structure. All these structures follow a Cevenol direction (N025°-030°E).

Above the Carboniferous surface, we observe numerous faults trending SW-NE and dipping SE (Figs. 7 and 9). Except for a few of them, they all show a normal fault kinematics. They form a set of connected listric faults, affecting all the Mesozoic layers. They are all rooted in the Triassic decollement, which they reached at different depths, with the exception of the St Pons fault that roots in the Stephanian deposits (Fig. 7A). According to surface and seismic data, the St Pons fault only slightly affects the Lower Cretaceous sediments (Fig. 7A). However, the St Pons listric fault strongly controls syn-tectonic sedimentation between the Hettangian and the Callovian (Fig. 6B–D). Another issue is that the central deposit of Lower Jurassic sediments corresponds to the depression zone of the top-Carboniferous. Therefore, the St Pons listric fault corresponds to a Tethysian roll-over structure that is rooted in Carboniferous sediments, and the structural high marked by the top-Carboniferous is the result of this anticline roll-over.

Above the Alba sub-basin and slightly towards the southeast, the Alba and Pontet-de-Couloubre faults affect the Mesozoïc sedimentary sequence and merge at depth to form a main decollement in the Triassic layers. However, the respective offsets observed along the Alba and Pontet-de-Couloubre faults vary in an opposite ways from northeast to southwest. To the northeast, the Alba fault corresponds to the main fault of the NCFS, recording a 700 meters minimum offset, while the Pontet-de-Couloubre records less than 200 m offset. To the southwest, the Alba fault show an offset lower than 100 m, while the Pontet-de-Couloubre fault shows a 500 m offset (Figs. 6E and 7).

In this same southwestern area, in addition to the two above-mentioned faults, several other faults dipping steeply (60°) to the southeast, affect the entire sedimentary cover forming a dense faults system with a fault interval lower than 1 km (Figs. 6E, F and 7). Among them, the N044°E trending La Rouvière fault roots into the Triassic decollement that is reached at 1.6 s TWT in its southwestern part and extends beneath the Rhône river sediments to the northeast where it roots into the Triassic decollement at 1.5 s TWT. The offset along the fault decreases towards the northeast, from 230 m in its southwestern part, to 70 m in its northwestern part, where it affects the sandy sediments of the Aptian (Fig. 9). Further southeast, the N048°E trending Marsanne fault is visible throughout the study area and appears as one of the major faults of the NCFS (Figs. 6, 7 and 9). It roots into the Triassic decollement between 2 and 2.5 s TWT. Within its southwestern part, the Marsanne fault shows a 500 m offset, which is recorded between Barremian and Oligocene deposits. This offset decreases to the northeast, reaching 450 m south of Montélimar, and then 250 m further north (Fig. 6E). The fault appears to control the geometry of the boundary between Urgonian limestones and the Vocontian domain during the Upper Barremian (Fig. 8A). Its orientation is similar to the Donzere basement fault. The Marsanne fault could therefore be associated to three extensional phases: the Tethysian extension, the Lower Cretaceous extension, and the Oligocene extension. The Saint-Montan fault is difficult to characterize in the seismic data, owing to the fact that Urgonian limestones outcrop between Viviers and Saint-Remèze that diffract the signal (Figs. 1C and 9; see Sect. 3.2.3). The fault affects the Donzère anticline that folds the Upper Cretaceous formations (Fig. 9) showing that the fault has been active after the folding. The fault also appears to control thickening during the Middle Jurassic. The occurrence of the Donzère basement fault may explain the strong morphological signature of the St Montan fault at the surface (Cathelin et al., 2023) and its significant offset (∼200 m) compared to its small length.

Seismic data allow also distinguishing a few compressional features such as folds and thrusts. As mentioned above, the tilted horizon corresponding to the Aptian sediments of the Couijanet syncline partly defines the Donzère anticline (Pascal et al., 1989). This anticline corresponds to a fault propagation fold highlighted in the field by very steep dips to the north and a much gentler dip to the south (Fig. 9). To the south, the stratigraphy shows a sudden deepening inside Donzère. It is difficult to know whether this deepening is structural or whether it corresponds to the incision of the Messinian canyon. Overall, the structure resembles a pop-up structure (Figs. 6F and 9). It is difficult to identify the root of this fault propagation fold, and therefore to decide between an interpretation of thick-skinned or thin-skinned shortening deformation. A few other compressional structures can be observed within the Vocontian domain. They correspond to the NW-SE trending La Lance anticline and the E-W trending Puygiron fold that both affect the Upper Cretaceous (Fig. 7B). Note that the La Lance anticline has been interpreted as being associated with Jurassic-Cretaceous diapirism (Debrand-Passard et al., 1984a; Martinod, 1988).

4 Discussion

The above-interpretations are consistent with the well data and previous geological studies of the Vivaro-Cévenol domain and the Vocontian area, affirming the robustness of our horizon calibration methodology. The approach that we applied in this study was challenging due to complex structural pattern and multiphase non-coaxial deformation process both within the basement and the sedimentary sequence. Our new seismic data, combined with the available well data allowed us to constrain well-defined horizon picking. In addition, the combination of surface data and previous seismic interpretations of seismic line 81-SE-3 (Martinod, 1988; Roure and Colletta, 1996) and seismic line 81-SE-4 (Rangin et al., 2010) improved the horizon picking in areas where the seismic signal was degraded or where there was no geophysical data.

The synthesis of our interpretations of sedimentary sequences and structures within the Vivaro-Cévenol border allows us to reconstruct the tectonic evolution of the NCFS, and highlight the main events (Fig. 10). A first point concerns the structural inheritance related to the Upper Carboniferous basin development. Our data show that the presence of Carboniferous basins associated with basement faults (Fig. 10A) controlled the sedimentary sequence and the deformation of the Vivaro-Cévenol domain that emplaced later. These results are in agreement with previous interpretations on the relationship between basement faults, the Carboniferous basin and the Mesozoic basin within the Vivaro-Cévenol border (Arthaud and Matte, 1975, Bonijoli et al., 1996).

During the Triassic, continental sediments were uniformly deposited all over the domain (Fig. 10A), but it was during the Lower Jurassic (Hettangien), at the beginning of the Alpine Tethys ocean rifting (Dumont, 1988), that the syn-tectonic sedimentation started in the basin, as also previously documented (i.e.Razin et al., 1996; Bonijoly et al., 1996; Martin and Bergerat, 1996). In the northeastern part, the sedimentation is controlled by the re-activation of the Alba basement fault associated with a major decollement that are rooted in the Carboniferous shales beneath Alba-la-Romaine. It is difficult to interpret the same mechanism in the south-eastern part of the basin, as the Pyrenean and Oligocene phases have reshaped the structure of the Vocontian boundary. The Marsanne fault could therefore interact in the same way as the St Pons fault interacts with the Alba basement fault, but rooted in the Triassic detachment level (Fig. 7B).

Middle and Upper Jurassic (Callovian-Tithonian) correspond to post-rift periods during which sedimentation is controlled by thermal subsidence (although localized thickness variations are observed within the future Vocontian domain that could be associated to diapirism process within the Triassic). During the Lower Cretaceous (Hauterivian-Barremian), the Marsanne fault forms and controls the palaeoenvironment of the basin (Fig. 10C). As described in previous works, the fault separates a rather thick hemi pelagic/reef zone (Urgonian) to the west from a very thin pelagic zone in the Vocontian domain (Pascal, 1959; Cotillon et al., 1979; Ferry and Flandrin, 1979). From the seismic profiles, it is difficult to identify at this period whether the Marsanne fault connects directly at depth to the Donzère basement fault.

The Upper Urgonian period is marked by an unconformity between the reef facies limestones of the Urgonian and the Aptian detrital sediments. This result is consistent with the denudation ages recorded on the Cévenol crystalline margin, indicating an exhumation of the margin about 130 million years ago leading to a terrigenous nature in the sedimentation (Contensuzas, 1980; Barbarand et al., 2001; Séranne et al., 2002; Ferry et al., 2022). It can be noted here that this is during this same period that basement faults—under the far field extension related to the rifting of the Bay of Biscay and the Valaisan Basin—would have been reactivating inducing the northward migration of Southeast Basin (Cotillon et al., 1979; Curnelle and Dubois, 1986; Ferry, 1999, Beltrando et al., 2007; Homberg et al., 2013; Angrand and Moutheraud, 2021). We propose here that the Donzère basement fault corresponds to one of these faults, relayed by the Saint-Montan and the Marsanne faults within the sedimentary sequence (Figs. 9 and 10C).

During the Upper Cretaceous, the influence of the Pyreneo-Provençal tectonics leads to the formation of E-W trending folds associated with basal detachment of inverted faults that is rooted in the Triassic in the Valreas basin. This compressional event is characterized in the Southeast Basin by the uplift of the Durance High (Fig. 1B). Within the Vivaro-Cévenol domain, the structures associated with the NCFS, which were initially N040°E trending normal faults, are re-activated as left-lateral strike-slip fault in the field, associated with E-W trending folds (Fig. 10D). From the Upper Cretaceous to Eocene, the structural pattern within the NCFS becomes more complex and the fault system acts as a transfer zone between the stable Vivaro-Cévenol domain and the deeper Vocontian and Valreas domains (Fig. 10D). The formation of the Donzère fold is laterally connected to the NCFS at Viviers. This gave rise to the Couijanet syncline (Figs. 9 and 10D). Numerous previous works placed this transpressional tectonics during the Albian within the Valreas Basin (Masse and Philip, 1976; Bergerat, 1982; Masse et al., 1990; Malartre, 1994; Roux and Brulhet, 1999; Rangin et al., 2010; Ballas et al., 2012). However, within the study area, this transpressional tectonics mainly occurred during the Turonian (Elmi et al., 1996). This phase is characterized by conglomeratic deposits in the Couijanet syncline, resulting from direct erosion of the surrounding relief. The complex geometry of the Donzère fold appears to be correlated with the tectonic thickening of the Hauterivian/Barremian layers induced by the inversion of the normal Donzère fault. This leads to the formation of a steeply dipping reverse fault associated with a long-wavelength fold with a short overturned forelimb. A balanced cross-section of the Donzère anticline is proposed in Supplementary material 4 based on seismic line calibrations. These geometries have largely been observed in analogue inverted oblique normal fault models (e.g.Roure et al., 1992; De Vicente et al., 2009; Jagger et al., 2018; Heja et al., 2022; Ferrer et al., 2022). However, the origin of the inversion by thin-skinned or think-skinned deformations is debatable, and the seismic profiles do not allow answering this question. Our interpretation is based on models that have already been proposed with a major detachment in the Triassic sequence active since the Pyreneo-Provencal shortening phase (Roux and Brulhet, 1999; Ballas et al., 2012, Balansa et al., 2022).

During the Oligocene, after a general emersion of the region in the Upper Eocene at the end of the Pyreneo-Provençal compression, a general NW-SE trending extension associated with the opening of the Lion Gulf (Fig. 1) reactivated all segments of the NCFS as normal faults. The Alba Fault controls the deformation within the Vivaro-Cévenol domain, rooting into the Triassic decollement until the Donzère basement fault (Fig. 10E). It shows the largest displacement (1000 m). However, very few Oligocene deposits associated with this normal movement are observed on the seismic lines, while thick sedimentary sequences are observed in the Valence Basin to the north and in the Alès Basin to the south (Séranne et al., 1995; Chabani, 2019). Two interpretations can be proposed to explain these observations. A first one considers that the Oligocene deposits accumulated at the footwall of the Alba fault have been subsequently eroded. This erosion might be associated to an uplift of this part of the Vivaro-Cévenol domain after the Oligocene. A second interpretation considers that a part of the 1000 m normal offset recorded along the Alba fault would correspond to a previous extensional episode, which most likely occurred during the Lower Cretaceous (Fig. 8).

During the Miocene, no alpine deformation is observed on seismic data within the NCFS. However, the Pyrenean structures of La Lance and the Dieulefit syncline were reactivated in the Tortonian, during which the marine molasse is affected by the NE-SW shortening (Bergerat, 1982; Casagrande, 1989; Roure et al., 1994a; Debelmas et al., 2004). The occurrence of the Messinian incision can be characterized by a canyon feature followed by its filling during the Pliocene (Fig. 10F). Figure 10F also shows our interpretation of the present situation with reactivation of the NCFS, at least part of it, contemporaneously with the regional uplift of the Vivaro-Cévenol domain (Ritz et al., 2020; Malcles et al., 2020; Cathelin et al., 2023; Thomasset et al., 2023).

Fig. 10 3D structural models illustrating major tectonic events leading to the formation of the NCFS from the Carboniferous to the present. P.A: Puygiron syncline, C.S: Couijanet Syncline, D.A: Donzère Anticline, A.f: Alba fault, D.f : Donzère fault, Pt. f: Pontet de Couloubre fault, LR.f: La Rouvière fault M.f: Marsanne fault. Principaux événements tectoniques ayant conduit à la formation du NCFS depuis le Carbonifère jusqu’à l’actuel. P.A: Synclinal de Puygiron, C.S: Synclinal de Couijanet, D.A: Antilcine de Donzère, A.f: faille d’Alba, D.f : faille de Donzère, Pt. f: faille du Pontet de Couloubre, LR.f: Faille de La Rouvière M.f: Faille de Marsanne.

5 Conclusion

This study about the geometry and tectonic history of NCFS highlights the fundamental role of the basement faults in the tectono-sedimentary evolution of the Vivaro-Cévenol domain. The Donzère basement fault, in particular, has been controlling the NCFS for the past 210 Ma. The interaction between thin-skinned and thick-skinned deformations, the occurrence of two levels of detachment and the combination of tectonic phases help to explain the complex structural scheme of the NCFS. The Donzère basement fault interacts with the Marsanne fault, influencing its orientation. However from the Cretaceous, the Triassic sequence appears to decouple the Mesozoic series from the top-Carboniferous, avoiding any direct connection between the NCFS and basements faults (i.e. Alba and Donzères).

Our study also provides information that greatly improves the geological knowledge within the Southeast Basin. At a larger scale, it brings an insight into the role of the NCFS in the regional geodynamics, especially concerning the interactions between Alpine and Pyrenean domains. For instance, it suggests that during the Lower Cretaceous, there were some connections between the extension observed in the Vocontian basin of the Alpine Tethys Ocean to the east, and the Pyrenean rift to the southwest. These connections have probably occurred via the Donzère basement fault, which we demonstrate is responsible of the subsidence between the Vivaro-Cévenol domain and the deep part of the Southeast Basin during the Lower-Middle Jurassic and the Lower Cretaceous.

Our study also establishes correlations between geophysical data acquired in the Vivaro-Cévenol domain and the Vocontian domain, and more generally within the Southeast Basin. These correlations will be useful to model the northern part of the Southeast Basin.

Finally, our sub-surface interpretation of the NCFS structural pattern constitutes an important insight in terms of seismic hazard in an area under active paleoseismological investigation following the 2019 Mw4.9 Le Teil earthquake. The mapping of the faults within the Rhône river valley will be particularly useful to investigate potential surface-rupturing events in Quaternary sediments.

Supplementary Material

Supplementary Material 1. Correlation of well data available on the vivaro-cévenole border. Stratigraphic detail, associated depth and velocity synthesis.

Supplementary Material 2. (left) sub-quaternary ecorché derived from field data, borehole data. (right) Vertical gradient of the Bouguer anomaly in the Montélimar region (-0.0029(blue)/0.0025(pink) mGal/m.

Supplementary Material 3. Various geological cross-sections showing the structure of the units based on outcrop data.

Supplementary Material 4. Palinspastic reconstruction of the formation of the Donzère Fold since the Lower Cretaceous according to several major tectonic phases.

Acknowledgments

This work gathers data acquired in the framework the EDF program for which CDP Consulting performed seismic lines processing, the CNRS-INSU Fremteil 2021–2023 project (PI J-F Ritz). Gravimetric and well data are from BRGM and were either downloaded freely or purchased by EDF. All data used in this study can be found in the tables and figures presented in the article and in Supplement Material (i.e., 1; 2A,B; 3; 4). We thank Gregory Ballas for having provided Camille Thomasset the access to the Kingdom software. We thank him also as well as Michel Séranne for their advises about many references about the Southeast Basin. We thank François Roure, Jean-Loup Rubino, Christophe Larroque, Florian Miquelis, Stéphane Baize and Nicolas Cathelin for fruitful discussions. We are grateful to Adrien Leroux and Christophe Vergniault for their help in the field and with the compilation of well data. Finally, we acknowledge Joseph Martinod and an anonymous reviewer for their constructive comments and suggestions, which greatly helped improving the original manuscript.

References

All Tables

All Figures

Fig. 1 (A) Location of the Southeast Basin in France. (B) Tectonic map showing main structures and depth of the Southeast Basin. (C) Geological map of the northern part of the Southeast Basin (modified from the 1/50,000 map of the BRGM) and the segmentation of the NCFS. CF: Cevennes fault, NF: Nîmes fault, MDF: Moyenne Durance fault, NCFS: Northeast Cévennes Fault System, VF: Valence fault, VT: Vercors thrust, LT: Lance thrust, VLT: Ventoux-Lure thrust, LuT: Luberon thrust, DN: Digne nappe, S.S: Saou syncline, S.A: Saou anticline, D.S: Dieulet syncline, P.A: Puygiron anticline, L.A: La Lance anticline, D.A: Donzère anticline, R.A : Roussas anticline and SR.A: Saint Remèze anticline.∗: dashed line for unknown location of fault. (A) Localisation du Bassin du Sud-Est en France. (B) Carte tectonique montrant les principales structures et la profondeur du Bassin du Sud-Est et sa profondeur. (C) Carte géologique de la partie nord du Bassin du Sud-Est (modiée à partir de la carte au 1/50,000 du BRGM) et la segmentation du NCFS. CF: faille des Cévennes, NF: faille de Nîmes, MDF: faille de la Moyenne Durance, NCFS: Système de failles Nord-est des Cévennes, VF: faille de Valence, VT: chevauchement du Vercors, LT: chevauchement de la Lance, VLT: chevauchement du Ventoux-Lure, LuT: chevauchement du Luberon, DN: nappe de Digne, S.S: synclinal de Saou, S. A: anticlinal de Saou, D.S: synclinal de Dieulet, P.A: anticlinal de Puygiron, L.A: anticlinal de La Lance, D.A: anticlinal de Donzère, R.A : anticlinal de Roussas et SR.A: anticlinal de Saint Remèze.∗: ligne pointillée pour localisation inconnue de la faille.

In the text

	Fig. 2 Stratigraphic chart and geodynamic context of the study area. Charte stratigraphique et contexte géodynamique de la zone d’étude.
In the text

	Fig. 3 Presentation of seismic lines performed by EDF for this study with location of well and reprocessed ancient seismic data. Présentation des lignes sismiques réalisées par EDF pour cette étude avec l’emplacement des puits et les données sismiques antérieures retraitées.
In the text

	Fig. 4 Picking of characteristic seismic horizons of the stratigraphic column based on interpolation with the synthetic film computed from the well VAL-1. Pointage des faciès sismiques caractéristiques de la colonne stratigraphique et projection du puit VAL-1 à l’aide d’un film synthétique de vitesse.
In the text

	Fig. 5 Time-depth chart used to convert seismic lines from oil drilling with Eqs. (1) and (2), derived from the Table 2. Loi temps-profondeur utilisée pour convertir les lignes sismiques à partir des forages pétroliers avec l’équation 1 et l’équation 2.
In the text

Fig. 6 Isohypse map of the (a) Top Carboniferous, (b) Top Sinemurian, (c) Top Bajocian, (e) Top Jurassic, (f) Top Barremian and the thickness map of the (d) Lower Jurassic sequence, (g) Lower Cretaceous sequence. Location on Figure 1. 1.AP: Anticline Puygiron, AL: Anticline Lance. 1: Alba Fault, 2: Pt Couloubre Fault, 3: La Rouvière Fault, 4: Marsanne Fault, 6: St Pons Fault, Carte des isohypse du (a) toit du Carbonifere, (b) toit du Sinemurien, (c) toit du Bajocien, (e) toit du Jurassique, (f) toit du Barremien et carte d’épaisseur des couches du (d) Jurassique inférieur, (g) Crétacé inférieur. Localisation sur la Figure 1. AD: Anticlinal de Donzère, AP: Anticlinal de Puygiron, AL: Anticlinal de la Lance. 1: Faille d’Alba, 2: Faille de Pt Couloubre, 3: Faille de La Rouvière, 4: Faille de Marsanne, 6: Faille de St Pons.

In the text

Fig. 7 (A) and (B) correspond to seismic lines interpreted and uninterpreted with horizons layers and faults along sections 21-CR-104, and 21-CR-101b - 81-SE-3 transect, respectively. Locations of the seismic lines on the geological map (Supp. Mat. 3). Vertical scale is expressed in second two-way travel (TWT) and horizontal scale is in meter (m), The horizontal/vertical scale has been respected. See cross section on Supplementary material 3. (A) et (B) correspondent aux lignes sismiques interprétées et non-interprétées avec des couches d’horizons et des failles le long des sections 21-CR-104, et 21-CR-101b - 81-SE-3 transect, respectivement. Localisation des lignes sismiques sur la carte géologique. L’échelle verticale est exprimée en secondes temps double (TWT) et l’échelle horizontale est exprimée en mètres (m). L’échelle horizontale/verticale a été respectée.

In the text

Fig. 8 Evolution of sedimentation in the Southeast Basin between the Vivaro-Cévenol border and the Valréas basin. The depth setting of borehole BO-1 comes from the interpretation of line 82-SE-4 by Rangin et al., 2010. 1: Alba Fault, 2: Pt Couloubre Fault, 3: La Rouvière Fault, 4: Marsanne Fault, 6: St Pons Fault, 7: Donzère Basement Fault. Evolution de la sédimentation dans le bassin du Sud-Est entre la frontière Vivaro-Cévenol et le bassin de Valréas. La profondeur du forage BO-1 provient de l’interprétation de la ligne 82-SE-4 par Rangin et al. 2010. 1: Faille d’Alba, 2: Faille de Pt Couloubre, 3: Faille de La Rouvière, 4: Faille de Marsanne, 6: Faille de St Pons, 7: Faille du socle de Donzère.

In the text

Fig. 9 Seismic line interpreted and uninterpreted with horizons layers and faults along section 22-TRI-104. Locations of the seismic line on the geological map (Supp. Mat. 3).Vertical scale is expressed in second two-way travel (TWT) and horizontal scale is in meter (m), The horizontal/vertical scale has been respected. See cross section on Supplementary material 3. Ligne sismique interprétée et non interprétée avec les horizons, les couches et les failles le long de la section 22-TRI-104. L’échelle verticale est exprimée en secondes de déplacement dans les deux sens (TWT) et l’échelle horizontale est exprimée en mètres (m), l’échelle horizontale/verticale a été respectée. Voir la coupe transversale sur Supplementary material 3 .

In the text

Fig. 10 3D structural models illustrating major tectonic events leading to the formation of the NCFS from the Carboniferous to the present. P.A: Puygiron syncline, C.S: Couijanet Syncline, D.A: Donzère Anticline, A.f: Alba fault, D.f : Donzère fault, Pt. f: Pontet de Couloubre fault, LR.f: La Rouvière fault M.f: Marsanne fault. Principaux événements tectoniques ayant conduit à la formation du NCFS depuis le Carbonifère jusqu’à l’actuel. P.A: Synclinal de Puygiron, C.S: Synclinal de Couijanet, D.A: Antilcine de Donzère, A.f: faille d’Alba, D.f : faille de Donzère, Pt. f: faille du Pontet de Couloubre, LR.f: Faille de La Rouvière M.f: Faille de Marsanne.

In the text