© M. Zaagane et al., Published by EDP Sciences 2025

1 Introduction

One of the purposes of fold-and-thrust belt studies lies within the reconstruction of reliable geometries in order to unfold these systems, to ultimately access shortening amounts that will be used to constrain the restoration of the initial configuration. In compressional systems where salt layers are involved, these reconstructions are two-fold because understanding the fold-and-thrust belt geometries implies also to understand the pre-orogenic inheritance given by the prior salt motions (Rowan and Vendeville, 2006; Callot et al., 2007; Jahani et al., 2009; Célini et al., 2020, 2024). On one hand, in orogenic systems, if previous sedimentary architectures of the rifted margin were salt-rich, the pre-orogenic geometries likely already showed complex 3D patterns due to salt structures and they can be mistaken with shortening structures (Jackson and Hudec, 2017 and references therein). On the other hand, the presence of salt within the orogenic prism will have strong impacts on the mechanical behavior of the prism and its structural evolution (e.g., Hudec and Jackson, 2017 and references therein). For instance, various mechanical, modeling and field studies demonstrated the effects of weak basal decollement with variable geometries for the orogenic prism and geometries of tectonic structures (Davis and Engelder, 1985; Letouzey et al., 1995; Costa and Vendeville, 2002; Bonini, 2003; Jourdon et al., 2020). Field examples and analog models have shown that a complex mechanical stratigraphy with intermediate salt decollement levels can be responsible for variable structural styles and fold wavelength (e.g., Sherkati et al., 2006 in the Zagros). In addition, the 3D patterns of inherited salt structures add vertical heterogeneity that will influence the location of compressive deformations (Rowan and Vendeville, 2006; Callot et al., 2007, 2012).

The issue of interpreting correctly the influence of pre-orogenic salt structures requires upscaling from field observations and interpretations (from hundreds of meters to few kms) to the fold-and-thrust belt scale (10-100 kms). This work has been handled for many circum-Mediterranean belts located on former salt provinces (Soto et al., 2017), in order to discriminate between inherited pre-orogenic and orogenic effects of the salt to get better estimates of shortening: in the Alps (Graham et al., 2012; Célini et al., 2020, 2021; Brooke-Barnett et al., 2023), in the Pyrenees (among others: Ford and Vergés, 2021; Parizot et al., 2023), in the Betics (e.g., Flinch and Soto, 2022), in the Rif (Gimeno-Vives et al., 2019, 2020), or in the Atlas belt (Michard et al., 2011; Saura et al., 2014; Vially et al., 1994; Jaillard et al., 2017).

Within this overview, a blind spot remains for the Algerian and Tunisian Tell. Only a few publications exist trying to handle this issue in Tunisia (e.g., Khomsi et al., 2009; Khelil et al., 2020). Salt is known to be of crucial importance in the western external Tell in Algeria, either as diapiric structures or as the principal decollement level in the external nappes of the system (Dalloni, 1952; Mattauer, 1958; Wildi, 1983; Midoun, 1989). However, its importance for picturing consistently the geometries of the external Tell belt has been rarely taken into account (Michard et al., 2023). Considering how important salt affected the rifting and post-rift structures even before inversion in other Alpine belts around the Mediterranean, it is thus important to check it for the Western Tell where the salt has been systematically interpreted as basal decollement of nappes, when it was not diapiric. Hence, the possibility of this salt structures (diapir, canopies) to become squeezed during deformed during the Cenozoic compressional stages has never been taken into account, which is of importance for shortening rates for instance.

In this contribution, we focus on the Ouarsenis Massif that belongs to the Western Tell, which is part of the Maghrebides or Tell-Rif, that is to say an Alpine belt resulting from the consumption of the Tethys domain from the Late Cretaceous up to present (see Durand-Delga and Fontboté, 1980; Wildi, 1983; Leprêtre et al., 2018; Fig. 1A). The Ouarsenis massif belongs to the external zones of Western Tell (Fig. 1B). Embedded within mainly Cretaceous-Paleogene rocks, the Culminating Range of the Ouarsenis massif (Fig. 2) is famous for its Jurassic outcrops, elsewhere either absent or reduced. In addition, the fact that the culminating point, the Great Peak (GP; ‘Grand Pic” of French authors) is made of an 800 m-thick reversed section, has offered a structural challenge to geologists since more than a century.

This work is a critical review of the evolution of opinions about the Ouarsenis Culminating Range and must be seen in a historical perspective on this understudied Mediterranean Alpine segment. We detail first a review of the available stratigraphy and structural information of the Ouarsenis Culminating Range. In order to bring a new model, we have re-interpreted key outcrops of the Culminating Range, considering salt tectonics concepts. We thus discuss how these re-interpretations allow to significantly revise the previous propositions. Accordingly, we then try to replace its evolution within the regional context and question the consequences of this new interpretation (1) for the Maghrebian Tethys evolution and (2) for the Western Tell orogenic structure. An updated geological map and cross-sections are proposed to illustrate the structure of this area.

Fig. 1 Geological context of the studied area. (A) Position of the Tell belt within the W. Mediterranean context of Africa-Europe convergence: BF: Baleares Fault; PH: "Pays des Horsts"; V.Tr: Valencia Trough. (B) Structural map of the Western Tell belt with the location of the Ouarsenis massif and the Culminating Range (yellow star). (C) General cross-section of the central Ouarsenis massif, encompassing the new interpretations proposed in this paper. It is importantly modified after the general cross-section n°8 of Mattauer (1958). Location of Figure 17 cross-section is indicated.

Fig. 2 Geological map of the Culminating Range of Ouarsenis (based and modified after Calembert, 1952; Mattauer, 1958; Tchoumatchenco, 1994). Location of salt outcrops shown in Figure 4 and flute-casts observation of Figure 8 are indicated. Also, panoramic views of Figures 3, 4d and 12 are indicated.

2 Geological setting

2.1 The Western Tell

Located in the southern part of the Western Tell, in Algeria (Fig. 1), the Ouarsenis massif is part of the external nappes of the Tell system. These nappes, made of Cretaceous-Paleogene rocks, show former Tethys passive margin sedimentary successions that have been transported southward during the inversion of the margin since Eocene times, and mainly during lower-middle Miocene (Leprêtre et al., 2018). To the north, the Flyschs units thrust over the external nappes and represent remnants of the Maghrebian Tethys sedimentary cover (Bouillin, 1986). The Flyschs units are then overthrust by the internal zones of the Tell-Rif system, represented, in the Tell, by the Kabylies (Fig. 1B). The Kabylies are mainly basement units with complex evolution at least since Paleozoic times that recorded Variscan events, Mesozoic Tethys opening and its closure during the Cenozoic (Michard et al., 2006 and references therein).

In the external Western Tell, Mesozoic-Cenozoic series recorded the story of the southern passive margin of the western Maghrebian Tethys (e.g. Kireche, 1993). There, the future Maghrebian Tethys domain suffered extension as soon as the Triassic, during the latest Early Jurassic/Middle Jurassic with a likely oceanization by the Late Jurassic/Early Cretaceous (Leprêtre et al., 2018 for a review). The nature of the basement of this domain is still discussed. However, some evidence shows that oceanic crust likely existed below the Flysch units (e.g. in Texenna: Boukaoud et al., 2021 and references therein), whereas in other parts, serpentinized mantle is suggested to be the substratum of the oceanic series (e.g. in Beni Malek in the eastern Rif; Gimeno-Vives et al., 2019). The closure of this oceanic domain began as soon as the Cretaceous, with subduction geometries that are still disputed (Guerrera et al., 2021). An important compressional phase was recorded during the Eocene (Delteil et al., 1971), in the external Tell. Yet, the major inversion occurred in Burdigalian-Langhian times, when the Kabylies and Alboran domains (i.e., the northern margin of the Maghrebian Tethys; Leprêtre et al., 2018) collided with the southern Tethys margin.

2.2 The ouarsenis massif and its culminating range

The Ouarsenis massif emerges as a tectonic window below the highest nappes of the external Tell. It is bounded to the west by the Mina and Beni Chougrane Mountains, to the north by the Cheliff Basin, to the east it is relayed by the Biban Massif. South of the Ouarsenis, external Tellian nappes thrust the Miocene foredeep. Within the Ouarsenis massif, the Culminating Range corresponds to an E-W elongated zone, which extends to 10 km long and 5 km wide (Fig. 2).

In this context, the high elevation of the GP of Ouarsenis (almost 2000 m), consisting mainly of Jurassic rocks appears as a surprising element. However, the perched location of Jurassic rocks is not the only puzzling fact. The general structure leaves wondering: it is consensual since the studies of Calembert (1952) and Mattauer (1958) that the GP is made of more than 800 m of overturned Jurassic and Cretaceous series. This observation is even much older: after the nice observations of Ficheur (1891), Bertrand (1891) suggested and then Répelin (1895) demonstrated that the series of the GP of Ouarsenis were overturned.

The Ouarsenis Culminating Range (Figs. 2 and 3) has been subdivided into four distinct sub-units, which are:

1. The GP, which occupies the central position and culminates at 1985 m, exhibits a ca. 800 m-thick reversed Mesozoic succession, going from Early Jurassic to Albian. Associated to it, the smaller Aïn Sour Massif is also characterized by overturned stratigraphic series;

2. The Sra Abdelkader (SAK), culminating at 1776 m, is a spectacular E-W monocline structure extending over 4 km and showing almost sub-vertical layers all along;

3. The southeastern mountains represented by Batha, Fartas and Belkheiret (BFB) Massifs, which culminate between 1420-1615 m. These three mountains stand as monoclinal structures whose Mesozoic layers dip towards E to SE and cut-out by strike-slip faults (N120° and N140°);

4. The Rokba Atba (RA) and Aïn el Hadjela (AH) isolated mounts stand farther, to the NW of the GP (Fig. 2). They both show a general monocline structure. The RA appears in faulted contact with Aptian-Albian rocks. The AH Mesozoic series thrust toward the NE, on top of the nappes stack (external Tellian nappe; Mattauer, 1958). Overall, these four main Jurassic-Neocomian massifs are surrounded by Aptian-Albian turbidites (formerly “Flysch Aptien-Albien” of French authors).

An outlying Jurassic massif also exists 6-7 km to the SSE of the GP, which is the Kef N’Hal outcrop. This small Jurassic-cored massif presents steep Jurassic layers, unconformably covered by Aptian-Albian rocks.

Fig. 3 Pictures of the GP area. (A) View from the northwest. (B) View from the south.

2.3 Tectono-stratigraphical record of the ouarsenis culminating range

The outcrops of the sedimentary series of the Culminating Range show a striking contrast in age and altitude compared to the neighboring softer formations. Indeed, the Jurassic limestone masses form reliefs above the surrounding countryside (Fig. 3). Previous studies have detailed the lithological succession ranging from Early Jurassic to Albian-Aptian (Mattauer, 1958; Polvèche, 1960; Atrops et al., 1991; Atrops and Benest, 1994; Benest, 1985; Tchoumatchenco and Khrischev, 1992; Tchoumatchenco, 1994; Tchoumatchenco et al., 1995; Benhamou, 1996; Alméras et al., 2007).

The Triassic is represented by (i) variegated clays (greenish and reddish) enclosing pale white to pink spots of salts, (ii) gypsum with saccaroid forms (Fig. 4a) and (iii) “cargneules”, which constitute peeled and arid sections on the rugged ground. Limestones and dolomites can be found associated with these rocks (Mattauer, 1958). This Triassic succession is never autochthonous and is found in different structural positions. It can either form the decollement of the external Tellian nappes playing the decollement role (Fig. 2) or “be embedded” within the younger rocks, either within faulted zone (Figs. 4b and 4c) or without evidence of fault (former canopies ?). Fieldwork shows large areas where the salt is “intruding” within the Culminating Range as already described by Calembert (1952) and Mattauer (1958) in their geological maps.

Forming a significant part of the reliefs, the Jurassic rocks are mainly limestone. The Jurassic sequence is the more complete on the reversed sequence preserved in the western GP (Fig. 5). The base of the Early Jurassic is not known, but a highly dolomitized level is widely represented below the first dated layers, attributed to the initial Early Jurassic levels (“pre-Sinemurian”). From Sinemurian to Pliensbachian, the series is represented by limestone formations topped by marly-calcareous levels, Toarcian in age. A conglomeratic level is intercalated between the Toarcian and older series. It reworks basement rocks and is associated to the activity of normal faults, now overturned (Mattauer, 1958). The lower Jurassic succession reaches up to 300-350 m in the GP, similarly to the SAK, 500 m in RA, and ca. 130 m in Kef N’Hal (Mattauer, 1958; Tchoumatchenco and Krishev, 1992). The Aalenian is not well-defined. The Bajocian is represented by a siliciclastic formation topped by limestone with flint. The extensional activity started during Early Jurassic with evidence of protracted activity from the Toarcian up to the Bajocian (Mattauer, 1958). The Bathonian and Callovian are not dated in the Culminating Range, but on the field, between the dated Bajocian and Oxfordian levels, the continuous sedimentation brought authors to usually consider that the limestone level below the Oxfordian characterizes the Bajocian to Callovian/basal Oxfordian (e.g. Tchoumatchenco and Krishev, 1992; Atrops and Benest, 1994). The middle Jurassic successions are rather reduced compared to the lower Jurassic ones: ca. 110 m in the GP and thicker in the SAK (ca. 140-180 m), up to 10 m in RA, 20-25 m in the Kef N’Hal. Overall, lower and middle Jurassic successions reach between 530-700 m of thickness. On top of it, an Ammonitico-Rosso level is Middle Oxfordian in age (Atrops et al., 1991). The end of Oxfordian is argillaceous with few limestone levels. Kimmeridgian and Tithonian show a limestone dominance, ending with marly-calcareous alternations in the Late Tithonian, continuing within the Berriasian up to the Hauterivian and possibly the Barremian (Calembert, 1952; Tchoumatchenco and Krishev, 1992). Thicknesses of the upper Jurassic successions are variable, between 130-170 m in the GP, 50-80 m in SAK, 15-20 m in the RA and 40-45 m in the Kef N’Hal. The Neocomian shows a thickness reduction from the NW flank of the GP (∼150 m) toward the eastern flank (∼60/70 m), and the southern part of Aïn Sour (50-70 m). It was also identified in the BKB and RA massifs (Tchoumatchenco et al., 1995; no thickness estimates) whereas it is absent in the Kef N’Hal massif.

On top of the Neocomian, the Aptian-Albian is made of turbidites that surround the GP and other Jurassic massifs (Calembert, 1952; Mattauer, 1958; Figs. 2 and 5e). The contact between the Aptian-Albian and underlying series is difficult to analyze, but has generally been considered as stratigraphic, sometimes with an unconformity between them (Mattauer, 1958; Kireche, 1993). In addition, at the base of the Aptian-Albian layers, conglomeratic levels or lenses rework Jurassic limestones (Calembert, 1952), but also Neocomian rocks up to the base of the Aptian (Mattauer, 1958). The conglomeratic levels are expressed through an unconformity that is clearly identified in the Kef N’Hal area, though with an angle < 10°.

The presence of younger rocks in the GP is defended by Mattauer (1958) evidencing the presence of a small Senonian outcrop on the SE flank of the GP, in a position difficult to assess (Fig. 2). Otherwise, Senonian to Eocene and Miocene rocks belong to the nappes stack of the external Tell, generally representing the top of the nappes system. They crop out to the NW, surrounding the RA small massif (Fig. 2).

A particular aspect of the Jurassic to Lower Cretaceous formations is their variable thickness across the Culminating Range (Calembert, 1952). In fact, and for most of the Jurassic formations, the WNW part of the GP always shows thicker rock layers compared to any other massifs of the area. Moreover, this change in thickness is described as gradual, with only some more rapid changes close to the faults that affected the Early and the base of Middle Jurassic in the northwest.

Fig. 4 Examples of evidence of salt presence around the GP, locations on Figure 2. (A,B) Senan pass; (C) Northwest of Boucaid, intruding within the Aptian-Albian turbidites (also located on Fig. 12 on a panorama); (D) Panorama from the south, on the Senan pass, showing the location of the salt outcrop detailed on (A,B). bbnb bx.

Fig. 5 General stratigraphy of the GP massif. On the top left (A): Panorama of the western GP (from the north). Numbers indicate some key stratigraphical levels given on the stratigraphic column of the right panel. (B) to (E) pictures below the western GP panorama (A) shows some of these key levels: (B) Toarcian − Senan formation; (C) Callovian-Oxfordian − Ammonitico-Rosso; (D) Marl/clays alternations − Neocomian; (E) Aptian-Albian turbiditic levels. These letters are localized against the stratigraphic column too. Letter numbering on the right side of the stratigraphical column for the different rock types. (a) dolomite; (b) dolomitic limestone with tidalites; (c) compact oolithic limestone; (d) limestones with karstic caves; (e) conglomerates; (f) dark limestones; (g) foraminifera-rich massive limestones; (g) marl-limestone alternations; (i) sandstones with silicified woods; (j) silex limestones; (k) Zoophycos limestones; (l) bioclastic limestones with filaments; (m) Ammonitico-Rosso; (n) quartz-clay alternations; (o) major stratigraphic discontinuities. The faults are presented after Mattauer (1958) who considered them as normal lower Jurassic faults.

3 A historical review

The GP structure attracted the attention of geologists since the end of XIX^th century at a time when its geometry was still debated (normal succession for Ficheur, 1891; reverse for Bertrand, 1891) but finally considered to be overturned (Répelin, 1895) with the general idea of an overturned limb of a recumbent fold. Dalloni (1936) did not really modified the general interpretation, adopting an autochthonous point of view. Indeed, at the time the fashionable idea was to consider the Jurassic pieces as autochthonous ones when associated to the salt, but as part of diapiric expressions. He nonetheless proposed an interesting opinion on the various Jurassic outcrops emerging throughout the Cretaceous rocks by considering them rooted. Most of these opinions will be kept by Calembert (1952) before being challenged by Mattauer (1958).

Calembert (1952), in line with Bertrand (1891), proposes a structural interpretation involving a succession of recumbent folds cut out by faults (Fig. 6A). The general structure would result from a succession of three important S-verging anticlines, each anticline being separated from each other by squeezed and highly deformed synclines. For instance, the SAK would represent the remnant of the steep limb of a south-verging anticline (Fig. 7A). In this view, the GP position should be the inverted flank of a south-verging recumbent syncline. The hinge of this syncline would be localized between the GP and the Senan Fault zone. This structural solution still appeared difficult to Calembert since he could not easily explain why only reversed flanks of anticlines were preserved. Indeed, for the most spectacular structure of the GP, the normal flank is nowhere to be found northward (Figs. 2 and 7A).

Later in the same decade, Mattauer (1958) proposes another structural interpretation. In spite of the dominant autochthonous views about the structure of the Tell during the first half of the XX^th century, few authors proposed allochthonous interpretations (Glangeaud, 1932). Anyway, after the 1950’s, according to the allochthonous views developed by academics (Caire et al., 1953; Caire, 1957; Mattauer, 1958; Polvèche, 1960 under Glangeaud’s ideas) but also based on industrial hydrocarbon exploration (Kieken, 1956), the Culminating Range was incorporated into the southward nappes system of the Tell belt. The Jurassic rocks would have been revealed through deep erosion within a wide anticlinorium structure, in addition to a reduced thickness of sediments. Mattauer (1958) considers that the whole Culminating Range is not fully autochthonous but consists in both rooted and displaced units (Fig. 7B). Particularly about the western GP, he underlined that the basal contact of the reversed serie was quite flat, within the Albian. Nonetheless, on the field only few sedimentary figures enable to recognize the reversed Aptian-Albian (Fig. 8), supposedly on top of normally organized series, with the youngest ones being also Albian in age (Fig. 6B). His revised mapping shows that rather than a reverse syncline flank, the reverse GP series can be best understood as a klippe. Finally, Mattauer (1958) proposes that the reverse series of both the GP and the smaller Aïn Sour massif, in addition to the AH nappe, all represented unrooted structures, whose origin should be found north of the RA-SAK line, below the Cretaceous series (Fig. 7B; pp.404-405 in Mattauer, [1958]). The rooted Jurassic massifs are “extrusions” from lower structural levels that pierced through the Cretaceous cover, around the unrooted units, after the emplacement of the latter. Mattauer (1958) did not explain why the GP is in reverse polarity; his cross-sections are clear, but no sequential explanation nor sequential restoration enable understanding what was his view on the structural evolution that led to this original structure.

Regarding the timing of deformation, Mattauer (1958) used the westward AH nappe that lies on top of a nappe stack post-dating Lower Miocene unconformities above deformed Cretaceous rocks, to infer that the GP nappe was emplaced after the Early Miocene. Anyway, although this could be acceptable for the AH nappe case, nothing permits to extrapolate this to the GP. There, the observations only show the reversed Albian series directly on top of normal Albian rocks below. Calembert (1952) is very elusive on giving a precise age for his “folded” deformation and suggest that it amplified ancient structural trends, “sometime” after the Neocomian.

The Calembert (1952) and Mattauer (1958) interpretations were clearly dependent on both the field elements they could get and the conceptual frame during the period when they worked on this topic: the allochtony vs. autochthony debate. The main novelty of Mattauer (1958) against Calembert’s model probably lies in the re-interpretation of the basal contact of the GP.

The last works on the Culminating Range are from Aïfa and Zaagane (2014, 2015). In these works, the authors claim that most of the observed deformations are related to Miocene to post-Miocene tectonic phases, mainly compressional because they attempted to correlate them to known tectonic orientations shown elsewhere (e.g. Meghraoui et al., 1986). The authors used fault data interpreted in a Neogene time frame only. Unfortunately, this was likely a reducing approach to understand the complexity of the studied area. Furthermore, faults cannot be properly dated and we must better follow the geometrical relationships between the different structures to propose a new structural understanding.

Fig. 6 Former cross-sections of the Great Peak by Calembert (1952) and Mattauer (1958). (A, top right) The main cross-section of Calembert (1952) showing the fold hinge used for his “recumbent fold model”. (B1 and B2, bottom right) Two main cross-sections of Mattauer (1958). On top left corner, a small zoomed map (after Fig. 2) locates the cross-sections precisely, as well as our own cross-section presented in Figure 15.

Fig. 7 Conceptual model after the two former structural interpretations. On these two sketches, the black continuous lines show the observable features that are extrapolated in dashed thin lines to show sub-surface and “over-surface” consequences of the structural models of the two authors. (A) Interpretation of Calembert (1952): this author identified a hinge, in jurassic and lower cretaceous levels that he interpreted to be the hinge of a recumbent syncline, whose northern flank represents the reverse serie of the GP. There is no allochthony of the GP. (B) Interpretation of Mattauer (1958): the GP allochthonous reverse package is issued from the same structural unit than the para-autochthonous in normal polarity below, i.e., the “complex A” of Mattauer (1958). As lateral equivalent of Aïn el Hadjela nappe, he considered it to come from an undefined area north of the line RA-SAK. This “complex A” has been further thrust by the upper Tellian nappes during the Miocene.

Fig. 8 Sedimentological evidence of reverse polarity within Aptian-Albian turbiditic levels on the western slope of the GP (see location on Fig. 2).

4 The culminating range of Ouarsenis: a sightseeing of the main outcrops

The geological map presented in our study (Fig. 2) is a synthesis of the different works by Calembert (1952), Mattauer (1958) and the Tchoumatchenco’s team in the 90s (in particular, Tchoumatchenco (1994)). Further fieldwork will be obviously still required to precise the geometries.

4.1 The “rooted” Jurassic outcrops

4.1.1 The Kef N’Hal massif

To the south of the area, the Kef N’Hal Jurassic outcrop forms a high E-W cliff surrounded by Aptian-Albian turbidites with a sudden northward virgation on its NE side (Figs. 2 and 9A). Along its southern side, the Aptian-Albian turbidites rest unconformably on Jurassic rocks. From west to east, the Aptian-Albian rocks are covering younger and younger rocks (Fig. 9C, D; spanning end of Lower Jurassic to Upper Jurassic ages; see Figs. 24 and 165 in Mattauer [1958]) showing an angular unconformity. This unconformity is materialized through: (1) levels in turbidites showing lower Jurassic olistolithes (Fig. 9E) and (2) some conglomeratic levels that reworked most of the pre-Aptian-Albian rocks (Mattauer, 1958; Fig. 9D with coarser beds emerging within the turbiditic levels) also show basement rocks, like basic amphibolite (Fig. 10C). On its northern side, the contact between Aptian-Albian and older rocks has been mapped as a “classical” nappe contact lubricated by salt by Mattauer (1958) (Fig. 10A). Along this Triassic “nappe contact”, remnants of Triassic rocks appear mainly as disorganized clayey levels relatively rich in gypsum and pebbles of reworked rocks (Fig. 10B). Triassic remnants surround the structure, with the classical rock suite of “Infra Lias” breccias of black dolomite, gypsum and cargneules but also reworked basement rocks (Fig. 10C, with amphibolites found in the south but in also in the northern flank of the structure).

Fig. 9 Re-interpretation of the Kef N’Hal structure. (A) Simple geological map (modified from Mattauer, 1958). (B) Cross-section of Kef N’Hal modified after Mattauer (1958). The main modification is the re-interpretation at depth with a megaflap. Given our re-interpretation, the “nappe” contact is more likely to be the scar of a former salt canopy. (C, D) Pictures of the unconformity between the Aptian-Albian and the upper Jurassic or the lower Jurassic series, respectively. (E) Focus on the unconformity between lower Jurassic and Aptian-Albian series, showing the conglomeratic formation and the stratigraphic contact.

Fig. 10 Kef N’Hal massif re-interpretation. (A) NE contact between verticalized middle Jurassic layers and overlying Aptian-Albian turbidites. (B) zoomed view on the poorly preserved Triassic remnants laterally along the contact between Aptian-Albian and older rocks. (C) Thin section views of one of the amphibolitic samples we found within Aptian-Albian levels south of the Kef N’Hal massif. On the left, plane-polarized light, on the right cross-polarized light.

4.1.2 The SAK and RA massifs

Monoclinal structures are represented in two main locations: the SAK and the RA (Fig. 2). The SAK wall is the most prominent topographical feature, extending from the GP northern tip toward the east. It is relayed to the W by the RA massif, showing very similar orientation and structural organization, though spatially restrained.

The SAK always shows vertical or southward steeply dipping layers. Morphologically, the western termination presents a vertical-axis bending towards the SW (Fig. 2). This bending is ensured by intervening radially distributed faults with directions evolving from N160° to N/S. In present-day position, these faults generally present a right-lateral strike-slip displacement (Aïfa and Zaagane, 2015). In terms of stratigraphy, the steeply S-dipping layers show Early Jurassic age to Late Jurassic/Neocomian outcropping on its southern face. Aptian-Albian deposits are unconformably covering the southern side, whereas the northern contact is a faulted one, putting highly verticalized Aptian-Albian rocks below the massive lower Jurassic limestones (Calembert, 1952). Thanks to mining and field data, this contact is materialized by a 10 to 100 m-thick breccia level dipping to the south, within which remnants of Triassic and lower Jurassic rocks (of the northern SAK flank) are deformed and mixed with pieces of the Aptian-Albian dismembered footwall. In places, Triassic salt is also found as intrusions within the Aptian-Albian turbidites along the northern front of the SAK (Fig. 4A). Geometrically, along the southern flank of the SAK, the rock layers quickly go back to sub-horizontal dips, suggesting a relatively local steepening (in less than 1 km).

The RA, to the NW of the GP, outcrops as a NW-SE trending Jurassic small relief within the Aptian-Albian turbidites. Jurassic formations show 45° to 60° dips towards the SE. The lithological succession begins to the NW with the formations of basal Early Jurassic and goes up to the Neocomian to the SE (Figs. 2 and 6). The RA massif could look like an isolated westward extension of the SAK. In spite of its reduced area, first order observations are similar here. If we consider the Jurassic-Neocomian layers, similarly to the SAK, the oldest Jurassic layers to the NW are in faulted contact with the Aptian-Albian, with the exception of its SE border where the Aptian-Albian seems in stratigraphical contact on top of reduced upper Jurassic/Neocomian rocks. This contact is not clearly identified as an unconformity at the surface (Calembert, 1952). In addition, on the southwestern and northeastern faces of the RA, the massif is in fault contact with highly deformed Aptian-Albian deposits where a few meter-thick zone can be visible thanks to the mining activity (Fig. 11; Calembert 1952). This small outcrop is difficult to integrate in the general structure we have described until now, because of the absence of lateral continuity.

Fig. 11 (A) view of the southwestern face of the RA massif. (B) Zoom on the fault contact between the Jurassic limestones and the Aptian-Albian turbidites. Location of (B) on Figure 2.

4.1.3 The BFB complex

Regarding the BFB hills to the E/SE of the GP, they generally present flanks inclined towards the east with important dips: 80° at Belkheiret, 30 to 40° at Fartas and 70° in Batha. This zone is cut out by steep faults striking between N120−140° with large fault planes preserved on the limestones (Aïfa and Zaagane, 2014, 2015), confirming at least a faulted contact between the Jurassic-Neocomian “rooted” massifs against the Aptian-Albian turbidites. In particular, at the southernmost extremity of the BFB (Belkheiret sub-massif), the contact between the Jurassic rocks and the Aptian-Albian turbiditic series is documented through mining data (Calembert, 1952). It shows a 35−40° N-dipping shear zone against which the Liassic rocks are brecciated along a contact made of laminated gypsum. Here again, on the northeastern face of this massif, the Aptian-Albian series are lying unconformably onto the Jurassic-Neocomian ones, reminding the same structural setting than the SAK.

4.2 The GP reversed series

Following the map of Figure 2, the GP appears as one main reversed package, whose base is difficult to follow on its eastern side. The reverse series seem duplicated in two locations, in the SW in the Aïn Sour massif and between the Batha and GP in the east (Figs. 2 and 6). In both cases, the duplicated reverse series lie at much lower elevations, than the GP summit. The reverse/normal polarities contact is easily visible on the western slope of the GP. There, the inverted Aptian-Albian series lie on top of Aptian-Albian layers showing normal polarities (Figs. 2 and 8; Mattauer, 1958; Tchoumatchenco, 1994). The Aïn Sour Massif shows a similar structural configuration, with inverted series, going from Aptian-Albian at the base “down to” Pliensbachian at the top of the massif. At its base, the Aïn Sour Massif also shows the reverse/normal polarities contact between Aptian-Albian levels. This contact is lost below screes and its continuity is not obvious toward the eastern slope, further complicated with the BFB massifs.

In the landscape, the GP reverse package appears south of an erosional gap that separates the vertical rock walls of the SAK and the RA massifs (Figs. 3C and 12), questioning the possible continuity of both structures. Between the city of Boucaid and the northern flank of the GP, Calembert (1952) described Neocomian and middle Oxfordian levels bending by up to 30° toward the north (Figs. 6 and 13). Following the Aïn Mora path along the NW GP flank, the horizontal layers in reverse polarity are bending toward the N/NNW, with the bending affecting progressively younger rocks when approaching the Aptian-Albian contact. Downdip, the layers disappear, cut at their base against Aptian-Albian turbidites.

Fig. 12 Picture showing the topographical gap between the SAK and the RA flaps. The former Aptian-Albian main weld is indicated from the north of the SAK to the RA, probably reactivated as a steep thrust in the Cenozoic. The GP is shown over a gliding surface representing the contact of the ancient overturned megaflap over the former adjacent halokinetic basin. Location of Figure 14 is indicated.

Fig. 13 (A) Folded northern flank of the GP overturned series. A walking path parallel to the E-W cliffs of the northern west GP enables one to observe that the overturned series from Callovian-Oxfordian to Neocomian are gently folded toward the north, with a maximum bend of ca. 30°. (B, C) Callovian-Oxfordian Calpionella marl-limestones alternations. (D, E) Neocomian marls/limestones alternations. (F) Callovian-Oxfordian Ammonitico-Rosso limestones.

5 Reinterpreting the Ouarsenis culminating range

5.1 A collection of “simple” structures

The Kef N’Hal and northern monoclinal structures of the SAK and RA offer a first sight of the type of deformation that we need to account for. There, remnants of Triassic rocks are everywhere to be found with its classical “suite” of rocks, but also the presence of basement rocks, dragged upward by the former evaporites.

In the simple structure of the Kef N’Hal, the relationship between Aptian-Albian and Jurassic rocks is the key point. First, on the southern side of Kef N’Hal, the angular unconformity between Aptian-Albian and Jurassic rocks (Fig. 9) accounts for tilting of pre-Aptian-Albian rocks. This cannot be explained by the Cenozoic compression and requires deformation mechanisms in pre-orogenic times. Second, hectometric Triassic lenses are found either within Aptian-Albian rocks or specifically at the contact between Aptian-Albian and older rocks (Figs. 9A and 10A-B). This means that the Triassic salt pierced throughout the whole stratigraphical series, at least up to the Aptian-Albian. This is further confirmed by the conglomeratic levels reworking sedimentary pre-Aptian-Albian rocks but also by the reworking of basement rocks in Aptian-Albian (Fig. 10C). We thus suggest that the angular unconformity resulted from halokinesis and that the overall geometrical structure could fit well with a halokinetic weld where we only see the southern half, with a steep flap (Fig. 9B). The Triassic salt piercing is in agreement with the tilting of the whole pre-Aptian-Albian stratigraphical sequence enabling in situ erosion, hence generating conglomeratic levels and olistoliths in basal Aptian-Albian levels together with few basement rocks brought to the surface by the salt ascent (Figs. 9C-E and 10C). With this halokinetic weld hypothesis, we can also account for the Triassic lenses on the northern side of the Kef N’Hal. On the surface where salt pierced, it then could spread on top of the structure before being wiped out, possibly only leaving scars and discontinuous levels (Figs. 9A and 10B). On this northern Kef N’Hal side, instead of nappe remnants (Mattauer, 1958), the Aptian-Albian rocks could be better understood as Aptian-Albian series overlying remnants of a small salt glacier (Fig. 9B).

The Kef N’Hal general structure looks like a partly verticalized flap, bounding the ascending salt, then unconformably covered by Aptian-Albian layers (Fig. 9B). This can be nicely illustrated by natural examples for instance on seismic cross-sections. Figure 14A could prefigure one possible geometry for pre-compression configuration. The geometry is asymmetrical, with a “flap” part to the SE showing the now visible part of the Kef N’Hal (Fig. 9B). In general, along an ascending salt structure, the flap is often more developed on one side (Rowan et al., 2016) and shows a strong asymmetry, like on Figure 14A. The northern flank of the Kef N’Hal being buried below Aptian-Albian series, its geometry remains consequently unknown.

The SAK geometry was interpreted by Calembert (1952) as a good example of thrust-related geometry but is difficult to hold. At last, two considerations can be put forward to argue against such an origin here. First, the thrust geometry would imply a northward vergence of the SAK structure, but this is difficult to reconcile with the S-ward recumbent folds model of Calembert (1952; Fig. 7A). Second, looking at the two flanks of the SAK, the structure shows two types of contact between Aptian-Albian and older rocks: (1) an unconformity of Aptian-Albian above older rocks on the southern flank quickly going back to sub-horizontal dips, suggesting a local steepening and (2) a thrust contact on the northern flank (Calembert, 1952). These two observations could fit an upper Neocomian/lower Aptian-Albian tilting event. We propose to re-interpret this geometry as a weld contact of an early squeezed diapir, whose active ascent during Neocomian/Aptian-Albian transition strongly tilted the diapir flanks. This interpretation is also in line with the fact that the southern face of the SAK is unconformably covered by Aptian-Albian, with an eastward younging trend of the rocks below the Aptian-Albian rocks (Mattauer, 1958). This kind of structure can be well-illustrated in some contractional systems, e.g., within the salt minibasins of Sivas Basin (Fig. 14B). Here, a former salt limit between two minibasins, shows a salt wall flanked by a flap on the northern side and subsiding younger sequences on the southern side. This salt limit has been then squeezed and reactivated as an inverse fault at the weld contact. This is typically a geometry that could explain the northern SAK contact, with similar scales.

Among these “simple” structures, the RA relationship with them and the GP is not obvious. We do not exclude a possible former continuation between the SAK and the RA, because the latter reproduces most of the observations made on the SAK. In this case, it could have been part of a formerly more significant megaflap bounding a past weld structure. It is nuanced in the next section with respect to the GP interpretation.

Be it the SAK or the Kef N’Hal ‘simple’ structures, the recorded geometries tend to evidence a general and necessary deformation or tilting that happened before or during the deposition of Aptian-Albian series. Moreover, this early deformation was sufficient to bring to the surface various stratigraphic levels from the Early Jurassic to the Barremian and maybe up to the base of the Aptian, but also basement rocks, reworked in Aptian-Albian levels. In these cases, the SAK and the Kef N’Hal could be better understood as remnants of flaps, whose relationships with the GP still need to be told.

Fig. 14 Seismic salt-analog structure. (A) Seismic cross-section from Rowan et al. (2016), showing a putative geometry for the Kef N’Hal structure, before compression and squeezing leading to weld formation. White arrow show the equivalent position of southern flank of Kef N’Hal weld. (B) Extract of cross-section by Kergaravat et al. (2016) showing the squeezing and inversion of a salt limit between two minibasins. This geometry could likely account for the geometry now observed at the SAK. Black arrow shows the equivalent position of the SAK flap, i.e., the pre-Aptian layers grouped in a flap south of the weld (see Figs. 2 and 15A). (C) Examples of recumbent geometries that could explain the GP deformation during Aptian-Albian (slightly modified after Dooley et al., 2015). On the left, the result of an analog modeling. On the right, the reproduction of a seismic line from the Santos Basin.

5.2 The GP and surroundings

As far as possible, explaining the general GP structure shall integrate it among its surroundings (BFB, SAK, RA), while taking into account (i) its nature as a gliding reversed package, (ii) the fact that no rocks younger than Aptian-Albian, nor evaporites are trapped between the nappe and the underlying autochthonous rocks (Fig. 2) and (iii) the bending of its northern flank (Fig. 13) with the presence of evaporites between it and the SAK, apparently in a lower structural position (Figs. 2 and 6). Following these pieces of evidence, the simplest timing for initiating the reversal and bending is the Aptian-Albian period. Here again, without any recorded contractional event at the time, we suggest for this reversal to result from halokinesis, generating folded structures and unconformities. This accounts for the absence of young rocks trapped below the southward gliding overturned series of the GP and allows the preservation of its northward bending with a later breaking of the hinge of this Aptian-Albian folded structure.

5.2.1 Overturned megaflap or carapace remnant?

Admitting the Aptian-Albian halokinesis, there remains the question of the initial position of the GP within this halokinetic structure. The GP could be the remnant of a megaflap flanking a diapir, or part of a diapir carapace that was overturned while the diapir pierced, dislocated and deformed its carapace (e.g., Kernen et al., 2021). Three lines of argument can be put forward for discussion.

First, the thickness of the GP series compares to the surrounding massifs (Tab. 1): with the exception of the NW part of the GP being particularly thicker than the other parts (280-320 m of Dogger to Neocomian), the rest of the GP massif shows similar thicknesses than the SAK (100-180 m of Dogger to Neocomian) and the RA (Neocomian is unknown in RA). Second, the position of the evaporites need to be considered. In particular, they seem structurally below the GP reversed serie in the Senan Pass (Figs. 2 and 4). In other places, evaporites are often related to faulted zones (RA, BFB) and the salt weld north of the SAK, reactivated as a fault (Figs. 2 and 4). Third, the recumbent folding accounting for the GP structure was likely largely completed by the Aptian-Albian. The fact that the thickness of the GP is mainly similar to the SAK could be in favor of it being the continuation of a megaflap previously attached to the SAK and overturned during the diapir piercing. Alternatively, we could consider the GP to be formerly part of a carapace, that could have been folded as an autosuture (Dooley et al., 2012 and 2015, their Fig. 23). In our case, pieces of the carapace would then be scattered within the Aptian-Albian with small pieces (e.g., the RA and BFB massifs) and bigger ones, possibly overturned (the GP). In this view, the SAK could still be seen as the rooted remnant of a megaflap previously in touch with the now-dislocated carapace. The Senan Pass evaporites being apparently located structurally below the gliding GP series are then more likely with this hypothesis, but their exact position at depth remain unclear (Fig. 6B1). At last, the timing consideration makes the GP recumbent folding contemporaneous with the recognized megaflaps development. If the GP is a leftover of a supra-salt carapace, there is no obvious reason to consider it was folded at the same time than the megaflaps and it could have been folded later. This would imply a different timing of folding with respect to the observed field relationships. With the present-day elements, we favor a megaflap explanation for the GP in the following (Fig. 15B). Let us note here that the RA massif, mainly surrounded by fault contacts, represents a different case. We propose for it to represent a scattered piece of either a dismantled megaflap or a former supra-salt carapace, embedded within the Aptian-Albian turbidites (in Fig. 15B presented as a small piece of the main GP-SAK megaflap).

In our N-S cross-section (Fig. 15A) slightly west of the SAK, we thus show the GP as a gliding nappe whose reverse base is Aptian-Albian in age, like the autochthonous series below. The northward bending at the rear of the GP emphasizes its origin within a recumbently folded package during the salt ascent in Aptian-Albian. The final breaking of the hinge with the gliding motion could occur much later, during the compressional stages of the Cenozoic. On map view, the Aïn Sour massif appears as a duplicate reverse serie (Fig. 2). The fault separating it from the rest of the GP could be a Cenozoic thrust or an extensional contact, inherited either from the Aptian-Albian halokinetic period, or from the Cenozoic with a massive collapse of this SW slope of the GP. Here, we favor a syn-halokinetic extensional feature for this duplicated characteristic, poorly reworked through later compressional events (Fig. 15A).

Considering this formation scenario, numerous seismic cross-sections have evidenced the possibility of having overturned thick packages of rocks in various geological contexts, while the salt is moving (e.g., Dooley et al., 2015 for natural and modeled analogs of megaflaps). In these cases, the overturning of the layers induces a fold geometry with the hinge developing at the passage of the salt, progressively overturning the layers, either pre- and syn-salt motion. In the end, the most recent layers are folded in the core of this structure (Fig. 14C). At this stage, it is easy to consider a subsequent faulting of this structure, detaching the overturned flank. This rupture might then preserve the bending at its rear, which is the structure preserved in the northern GP.

Regarding the BFB, we believe that the fault contacts affecting the massif are likely related to late stages of deformation (likely Cenozoic). Yet, the relationships between on one hand the RA/GP/SAK and on the other hand the BFB massifs remain unclear. Are they the remnants of halokinetic welds of two initially separated halokinetic minibasins? Alternatively, is the BFB a structural southeastern extension of the single wide overturned GP megaflap or other pieces of a dismantled carapace, intensively deformed during the Miocene Tell formation? The field evidences are not yet sufficient to decide between the various options here.

Fig. 15 (A) Re-interpreted cross-section of the GP. Location of the map on Figure 6 in thick blue line. (B) Sketch showing a possible structural configuration of the system during Aptian-Albian halokinesis.

6 Discussion

6.1 Conditions of salt structure evolutions in the culminating range

As a whole, the Culminating Range of the Ouarsenis shows structures related to megaflaps (SAK, Kef N’Hal) and other more ambiguous possibly related to supra-salt carapace remnants (RA, BFB, GP ?). The putative carapace remnants bring less details for the halokinesis timing than the megaflap remnants, the latter being more helpful to discuss the tectonic context within which developed the main Culminating Range features.

Megaflap structures appear to be common structures in various salt-rich environment where minibasins develop (e.g., Callot et al., 2016 and references therein). We use the definition of Rowan et al. (2016) for the megaflap: it is “a panel of deep minibasin strata that extends far up the sides of a steep diapir or its equivalent weld; the width of folding and vertical relief both span multiple kilometers, and the maximum bedding attitude ranges from near vertical to completely overturned beneath an allochthonous salt sheet”. Some structures re-interpreted in the Culminating Range are thus consistent with such a definition.

Given the absence of detailed structural studies on this area, we must discuss these structures developments in the light of limited observations. First, the megaflaps are always made of lower Jurassic to upper Jurassic/Neocomian rocks, systematically unconformably covered by Aptian-Albian turbidites. Second, the basal levels of the Aptian-Albian rocks show beds containing olistholites or conglomeratic levels, reworking all older Mesozoic rocks but also basement pre-salt rocks (pre-Triassic). Third, in the GP case (as an overturned folded structure), Aptian-Albian beds showing both reverse and normal polarities are in direct contact without younger rocks pinched in-between.

These simple facts ascribe the main halokinetic development to Aptian-Albian times, and this is visible either in rooted units or in the GP. The presence of pre-Triassic basement rocks is not intriguing in the sense that ascending evaporites are known to be able to lift denser material through quick enough upward motions (Weinberg, 1993). It then only means that amphibolites blocks were available to be captured within the evaporites during their ascent. Given the absence of contractional tectonics at the time, while some extensional events are recorded in the eastern Tell during the Early Cretaceous (e.g., Frizon de Lamotte et al., 2011; specifically Aptian-Albian in Tunisia: e.g., Gharbi et al., 2022), we consider that an extensional context is likely between Neocomian and Aptian-Albian, so that “fresh” km-scale faults brought medium-grade metamorphic rocks in contact with the quickly ascending salt. Yet, extension is generally thought to be less likely for the development of megaflaps (see Rowan et al., 2016). Indeed, since megaflaps are thought to develop during the enlargement of the diapir width, as top-salt layers progressively incorporated within the minibasin flank, extension will naturally go against this: salt will be better used as accommodating the extension more than expending vertically the diapir, the latter mechanism being in large part responsible for megaflap formation. Anyway, various contexts have shown megaflaps can develop in extensional contexts (e.g. Alsop et al., 2015; Ford and Vergés, 2021; Parizot et al., 2023).

6.2 General evolution of the culminating range

Halokinesis is likely to have started as soon as the initial Jurassic stages, even though we have little data about it. Uneven distribution of sedimentary rocks on top of the evaporites are enough to trigger their motion (Fig. 16A). In addition, this uneven distribution could be enhanced through the extensional activity recorded in the NW GP side, during Early and Middle Jurassic (Mattauer, 1958) which is known at a larger scale (Escosa et al., 2021), in relationship with the Atlantic and Tethys opening. This extensional activity affecting northern Africa at the time helped generating basement motions that activated the salt motions. From the Middle Jurassic to the Neocomian/Aptian-Albian, subtle progressive thickness variations are still recorded across the GP but without a clear relationship with halokinesis. The major halokinetic stage occurred by the end of Neocomian/Aptian-Albian (sect. 6.1). The thickness and dips of Albian-Aptian series are unknown in the area, but it already witnessed the development of megaflaps either in the SAK or the Kef N’Hal areas (Fig. 16B). So far, no younger rocks have been described trapped in between the reverse and normally lying Aptian-Albian layers of the GP. Although this cannot be a definitive argument, the possibility of a younger timing for the full reversal in a compression setting would need a very complex interpretation. We thus propose that the GP megaflap already existed at the time likely enhanced thanks to the spreading of a salt body at the surface bending the megaflap toward the south (Fig. 16B).

The first shortening period occurred during the Paleogene, likely during the Middle-Late Eocene (Delteil et al., 1971; Leprêtre et al., 2018 for a review). Using it as a working hypothesis, we propose that the remnant of a Late Cretaceous external nappe trapped within the GP massif (see Fig. 2; described in Mattauer [1958]) could mean that most of the Aptian-Albian and Late Cretaceous cover in the Culminating Range was eroded since the Paleogene, so that at the time of the nappe emplacement it could be in tectonic contact with Jurassic/Neocomian rocks. This stage would have mainly signed the development of welds geometries (Fig. 16C) that will be enhanced during the main shortening of the Early Miocene (Fig. 16D). The progressive shortening could have enhanced the overturn of the megaflap, leading to the ‘detachment’ of the overturned part of the megaflap. The ongoing shortening enhanced the weld formation between the SAK Jurassic-Neocomian megaflap and the Aptian-Albian series to the north, triggering a slight reverse fault component northward (Calembert, 1952). Let us add that in our view, this interpretation confirms that the GP and other rooted units must be seen as autochthonous or para-autochthonous units that would have been only poorly affected by the external nappes emplacement.

Fig. 16 Model of evolution for the Culminating Range during the Mesozoic-Cenozoic evolution. (A) First stages from Jurassic to Neocomian, initial weak halokinesis, hypothetically at the vertical of inherited rifting normal faults. (B) Major halokinetic event in the Aptian-Albian, development of megaflaps in Kef N’Hal and GP, probable emplacement of salt canopies at least in the GP, overturning the older Mesozoic serie. An extensional activity is suggested. (C) Paleogene stage of compression, eroding partially, wiping out the canopies and welds formations. (D) Miocene stage of compression, during the main Tell phase, with the emplacement on top of the Ouarsenis massif of the external nappes with the maximum horizontal displacements.

6.3 Halokinesis and evolution of the southern Tethys margin in N. Africa

Halokinesis is evidenced, since a long time, in the Atlas system of North-Africa (Fig. 1), especially in Eastern Algeria and Tunisia (among others Vially et al., 1994; Vila, 1995; Perthuisot et al., 1998; Hlaiem, 1999; Jaillard et al., 2017). More recently, papers dealt with this topic in the Moroccan High and Middle Atlas (Michard et al., 2011; Saura et al., 2014; Vergés et al., 2017; Escosa et al., 2021; Skikra et al., 2021; Teixell et al., 2024).

In the Tell-Rif system, the role of salt has been considered as a major decollement level of thrust sheets (review in Wildi, 1983; Flinch et al., 2003) but it has also been largely reconsidered as an important decollement level during the passive margin stages. In fact, it is recognized in many places in the whole system (in the Rif, Gimeno-Vives et al., 2019, 2020; in the Western Tell: Midoun, 1989; Bracène and Frizon de Lamotte, 2002 [their Fig. 4]; Vially et al., 1994; in Tunisian Tell : Khomsi et al., 2009; 2019; Khelil et al., 2019). However, it is worth noting that it does represent one of the main characteristics of the system, helping to understand the Mesozoic evolution and the influence of this inheritance during the Cenozoic orogenic stages.

Maybe one of the nicest examples of the significant influence of this halokinetic activity is given by the evolution of the Central-Eastern Rif Mesozoic margin (Gimeno-Vives et al., 2019, 2020). There, the salt layer is responsible for:

• the presence of classical halokinetic features within the passive paleomargin (Gimeno-Vives et al., 2020, their Fig. 7).

• the strong decoupling between the Jurassic cover and the basement during middle Jurassic extension, provoking the juxtaposition of basement slices and Jurassic elements, later involved in a complex tectonic stack in this region of the Rif.

• the wide upper nappes of the Rif belt, that also exist in the Tell, have been suggested to result from Triassic salt in Cretaceous canopies that enhanced the mechanical decoupling between young Late Cretaceous-Paleogene upper nappes and older rocks below (Gimeno-Vives et al., 2020, their Fig. 12).

This example is inspiring for the Western Tell and the region between Oran, Tlemcen and the Chelif Basin, where many structural similarities with the central-eastern Rif have been underlined (Leprêtre et al., 2018). Salt is abundantly present (Midoun, 1989) and produced halokinetic features well-identified in sub-surface imaging (Vially et al., 1994; their Fig. 24 for Tlemcen area). Further east, in the Ouarsenis area, our study has completed this view by confirming the importance of halokinesis. Remnants of this activity, with development of spectacular structures, are quite clear in the Culminating Range. Based on the surface cross-sections of Mattauer (1958) we thus propose a re-interpretation of this part of the belt in Figure 17. Based on the fact that salt is widely present, either as an efficient decollement level but also likely as diapir or fossil canopies, it makes realistic the fact that numerous minibasins could compose the sub-surface structure of the Ouarsenis massif. In this case, it does not exclude the possibility of having far-travelled external nappes on the top part of the system, but all the allochthonous nappes detailed by Mattauer (1958) could have much less horizontal displacements than he supposed at the time. It is of course a first-order view based on this study but that shall be validated by further field works elsewhere in the Western Tell.

Broadening the discussion, there are obvious similarities between the GP and the structures described in the Haute-Provence Alps (France) by Graham et al. (2012) and Célini et al. (2020). We think, in particular, of the famous "Barre de Chine", also inverted over several kilometers. Other famous examples have been studied e.g., in the Sivas basin (Callot et al., 2016) or in the eastern Pyrenees (Ford and Vergès, 2021; Parizot et al., 2023). All these examples are intriguing structures, known for a long time, which have been reinterpreted using the salt tectonics concepts. On the basis of our observations, re-interpretations and existing surface cross-sections realized at the Ouarsenis scale by Mattauer (1958), we propose a new cross-section for this area, compared to the former one (Fig. 17). Since we focused on the Culminating Range, we have few field controls out of it and our cross-section is meant to give a qualitative view of what could the geometries be when taking into account the early halokinesis. First, this conceptual changes obviously reduces the amount of shortening compared to the former view. Indeed, most of the shortening is accumulated at the front of the structure and “nappes” are much less extended. In fact, their “continuity” across tens of km can be questioned when we now propose that some of the nappe contacts could only be welds, automatically reducing their importance. It is not reducing the importance of the role played by the salt in the Miocene nappes tectonics, but injecting more complexity by taking into account an early history that significantly influenced the structure of the belt in the Western Tell.

Fig. 17 New proposition for cross-section of the Ouarsenis massif, taking into account the important development of salt structures, consequently impacting the possible estimates of shortening. Location on Figure 1.

6.4 What is happening along the Maghrebian Tethys margin during the Early Cretaceous?

We have just seen that, in the general context of North Africa, the existence of halokinesis in the Ouarsenis massif is not surprising. Nevertheless, its age was less expected. There are discrete indications of halokinesis at the Lias-Dogger transition (with associated extensive faults, see Mattauer, 1958; recalled in section 2.2), but the major event occurred later (Early Cretaceous). It is known that at that time, the passive margin of the Tethys was already shaped (Leprêtre et al., 2018). Yet, we exclude that they could only result from passive halokinesis given arguments detailed in previous section 6.1. We favor the revival of extensional activity in the Early Cretaceous, accentuating extensional displacement on Jurassic faults (Fig. 13B). However, how to include such an undescribed event in the geodynamic context of North Africa?

We know that at a global scale, the Early Cretaceous is a major rifting period (review in Frizon de Lamotte et al., 2015). At the regional scale of the Western Mediterranean, extensional structures of this age are known both on the lower plate of the Tethys system (Eurasia) and on the upper plate (Adria, Africa). Thus, such structures are known on the Iberian plate, in the Pyrenees and the adjacent Aquitaine Basin (review in Saspiturry et al., 2021) but also in the Languedoc-Provence fold-thrust belt (Combes, 1990; Tavani et al., 2018; Hemelsdaël et al., 2021). A structural link with halokinesis has been put forward in the formation of many basins around the Western Mediterranean, like the Columbrets Basin (Etheve et al., 2018) but also in some of the Pyrenean orogenic system (Canérot et al., 2005; Ford and Vergés, 2021; Parizot et al., 2023).

In Africa, an Early Cretaceous rift system is described since a long time in central Africa (review in Guiraud and Maurin, 1992) as well as in Libya and Tunisia within the Sirte rift system (review in Frizon de Lamotte et al., 2009; 2011). The slab-pull mechanism has been proposed to explain the formation of rift systems in a very large area, sometime quite far from the Tethys realm (Frizon de Lamotte et al., 2015; Jolivet et al., 2016).

At the Maghreb scale, we know that there was a renewal of extensional activity during Early Cretaceous in the Tunisian Atlas (Soussi and Ben Ismail, 2000; Bouaziz et al., 2002). Westward in the Sahara Atlas (Algeria) evidence for Early Cretaceous halokinesis (Vially et al., 1994) also suggest extensional faulting at depth. In Morocco, there is no consensus about an Early Cretaceous rifting in the High Atlas. It still is a matter debated due to magmatic venues at the time (e.g., Skikra et al., 2021). Also, pre-100 Ma important tilting events are known, as demonstrated by paleomagnetic studies on various salt ridges in the High Atlas (e.g., Torres-López et al., 2018) although they cannot clearly be ascribed to a peculiar tectonic mechanism. On the contrary, in the Middle Atlas, strong arguments (Charrière et al., 2005; Haddoumi et al., 2008) show that at that time (and especially in the Barremian) subsidence was controlled by extensional tectonic activity, which is also attested by the halokinesis identified by Escosa et al. (2021). In the NE prolongation of the Middle Atlas, in the an “Pays des Horsts” (Fig. 1A) a similar timing of extension (Cattaneo, 1991; Chotin et al., 2000) seems to be recorded, extended up to NW Algeria (Benest, 1985), but without salt in this area, generating unconformities between upper Cretaceous series and below. It is worth noting that the “Pays des Horsts” is intersected by the Tell front, but the Ouarsenis massif remains in its exact continuation. Thus, we would highlight an extensive ENE-WSW structure of Early Cretaceous age slightly oblique on the E-W structures of the Tell. This Middle Atlas-Ouarsenis Early Cretaceous rift system (Fig. 1A) would constitute a new milestone in the framework of the Early Cretaceous rifts.

It is interesting to point out that the emplacement of salt canopies at the Early Cretaceous-Late Cretaceous transition have been highlighted along the Moroccan Atlantic margin (Hafid, 2008; Tari et al., 2000) and also in the Rif (Gimeno-Vives, 2020). According to this last author, these canopies served as a detachment level during the individualization of the Aknoul Nappe (element of the “Upper Nappes”). The link between rifting renewal and the development of the canopies has not been explicitly proposed but it seems likely. Early Cretaceous rifting may therefore be even more generalized than we suggest.

7 Conclusion

We propose in our study a new interpretation of the GP of Ouarsenis, which has intrigued geologists for 150 yr. This massif has the particularity within the Western Tell to show abundant Jurassic outcrops and among the oldest stratigraphic levels exposed, but with complex geometries that were challenging the former structural interpretations. The pioneer works of Calembert (1952) and Mattauer (1958) really made most of the key interpretations. However, they were lacking of the conceptual framework within which their observations could really make sense. In the conceptual framework used here, halokinesis of Triassic salt plays the major role. Beyond the fact that the GP represents an impressive reversed stratigraphic pile, the integration of the other pieces is here key to understand the general first-order organization of the whole Culminating Range. Within these various pieces, a variety of field elements points toward complex minibasin architectures now deeply reworked due to the two Cenozoic compression stages during Eocene and Miocene respectively. If we subtract these late stages, we can to propose a reconstruction of the pre-compression stages as it follows:

• First, an extensional period between Early and Middle Jurassic in relationship with the development of the Maghrebian Tethys. This phase offers weak evidences of halokinesis.

• Then a second extensional period by the end of Neocomian/Aptian-Albian, which reactivated the inherited faults of the margin and gave rise to the most amazing structure: the development of megaflaps with the superb example of the reversal of the GP sedimentary pile.

It is interesting to note that this second episode of extension, only recently recognized in some localized parts of the Dorsale Calcaire (Michard et al., 2023), could fit well with the geodynamics at that time. We believe that in the northern part of the Ouarsenis massif, that is to say in the distal part of the margin, this same event was at the origin of the emplacement of salt canopies. As in the Rif, these former canopies were used as a decoupling and sliding surface during the shortening episodes, which gave rise to the Tell-Rif chain.

Acknowledgments

This work is initially part of a joint field work and collaborations between Mascara University, Oran II University and CY Cergy Paris University. We thank the two reviewers Naïm CELINI and André MICHARD for their very constructive reviews, that helped clarifying numerous points in the paper in addition of a welcome and necessary reorganization.

References

All Figures

Fig. 1 Geological context of the studied area. (A) Position of the Tell belt within the W. Mediterranean context of Africa-Europe convergence: BF: Baleares Fault; PH: "Pays des Horsts"; V.Tr: Valencia Trough. (B) Structural map of the Western Tell belt with the location of the Ouarsenis massif and the Culminating Range (yellow star). (C) General cross-section of the central Ouarsenis massif, encompassing the new interpretations proposed in this paper. It is importantly modified after the general cross-section n°8 of Mattauer (1958). Location of Figure 17 cross-section is indicated.

In the text

	Fig. 2 Geological map of the Culminating Range of Ouarsenis (based and modified after Calembert, 1952; Mattauer, 1958; Tchoumatchenco, 1994). Location of salt outcrops shown in Figure 4 and flute-casts observation of Figure 8 are indicated. Also, panoramic views of Figures 3, 4d and 12 are indicated.
In the text

	Fig. 3 Pictures of the GP area. (A) View from the northwest. (B) View from the south.
In the text

	Fig. 4 Examples of evidence of salt presence around the GP, locations on Figure 2. (A,B) Senan pass; (C) Northwest of Boucaid, intruding within the Aptian-Albian turbidites (also located on Fig. 12 on a panorama); (D) Panorama from the south, on the Senan pass, showing the location of the salt outcrop detailed on (A,B). bbnb bx.
In the text

Fig. 5 General stratigraphy of the GP massif. On the top left (A): Panorama of the western GP (from the north). Numbers indicate some key stratigraphical levels given on the stratigraphic column of the right panel. (B) to (E) pictures below the western GP panorama (A) shows some of these key levels: (B) Toarcian − Senan formation; (C) Callovian-Oxfordian − Ammonitico-Rosso; (D) Marl/clays alternations − Neocomian; (E) Aptian-Albian turbiditic levels. These letters are localized against the stratigraphic column too. Letter numbering on the right side of the stratigraphical column for the different rock types. (a) dolomite; (b) dolomitic limestone with tidalites; (c) compact oolithic limestone; (d) limestones with karstic caves; (e) conglomerates; (f) dark limestones; (g) foraminifera-rich massive limestones; (g) marl-limestone alternations; (i) sandstones with silicified woods; (j) silex limestones; (k) Zoophycos limestones; (l) bioclastic limestones with filaments; (m) Ammonitico-Rosso; (n) quartz-clay alternations; (o) major stratigraphic discontinuities. The faults are presented after Mattauer (1958) who considered them as normal lower Jurassic faults.

In the text

Fig. 6 Former cross-sections of the Great Peak by Calembert (1952) and Mattauer (1958). (A, top right) The main cross-section of Calembert (1952) showing the fold hinge used for his “recumbent fold model”. (B1 and B2, bottom right) Two main cross-sections of Mattauer (1958). On top left corner, a small zoomed map (after Fig. 2) locates the cross-sections precisely, as well as our own cross-section presented in Figure 15.

In the text

Fig. 7 Conceptual model after the two former structural interpretations. On these two sketches, the black continuous lines show the observable features that are extrapolated in dashed thin lines to show sub-surface and “over-surface” consequences of the structural models of the two authors. (A) Interpretation of Calembert (1952): this author identified a hinge, in jurassic and lower cretaceous levels that he interpreted to be the hinge of a recumbent syncline, whose northern flank represents the reverse serie of the GP. There is no allochthony of the GP. (B) Interpretation of Mattauer (1958): the GP allochthonous reverse package is issued from the same structural unit than the para-autochthonous in normal polarity below, i.e., the “complex A” of Mattauer (1958). As lateral equivalent of Aïn el Hadjela nappe, he considered it to come from an undefined area north of the line RA-SAK. This “complex A” has been further thrust by the upper Tellian nappes during the Miocene.

In the text

	Fig. 8 Sedimentological evidence of reverse polarity within Aptian-Albian turbiditic levels on the western slope of the GP (see location on Fig. 2).
In the text

Fig. 9 Re-interpretation of the Kef N’Hal structure. (A) Simple geological map (modified from Mattauer, 1958). (B) Cross-section of Kef N’Hal modified after Mattauer (1958). The main modification is the re-interpretation at depth with a megaflap. Given our re-interpretation, the “nappe” contact is more likely to be the scar of a former salt canopy. (C, D) Pictures of the unconformity between the Aptian-Albian and the upper Jurassic or the lower Jurassic series, respectively. (E) Focus on the unconformity between lower Jurassic and Aptian-Albian series, showing the conglomeratic formation and the stratigraphic contact.

In the text

Fig. 10 Kef N’Hal massif re-interpretation. (A) NE contact between verticalized middle Jurassic layers and overlying Aptian-Albian turbidites. (B) zoomed view on the poorly preserved Triassic remnants laterally along the contact between Aptian-Albian and older rocks. (C) Thin section views of one of the amphibolitic samples we found within Aptian-Albian levels south of the Kef N’Hal massif. On the left, plane-polarized light, on the right cross-polarized light.

In the text

	Fig. 11 (A) view of the southwestern face of the RA massif. (B) Zoom on the fault contact between the Jurassic limestones and the Aptian-Albian turbidites. Location of (B) on Figure 2.
In the text

	Fig. 12 Picture showing the topographical gap between the SAK and the RA flaps. The former Aptian-Albian main weld is indicated from the north of the SAK to the RA, probably reactivated as a steep thrust in the Cenozoic. The GP is shown over a gliding surface representing the contact of the ancient overturned megaflap over the former adjacent halokinetic basin. Location of Figure 14 is indicated.
In the text

Fig. 13 (A) Folded northern flank of the GP overturned series. A walking path parallel to the E-W cliffs of the northern west GP enables one to observe that the overturned series from Callovian-Oxfordian to Neocomian are gently folded toward the north, with a maximum bend of ca. 30°. (B, C) Callovian-Oxfordian Calpionella marl-limestones alternations. (D, E) Neocomian marls/limestones alternations. (F) Callovian-Oxfordian Ammonitico-Rosso limestones.

In the text

Fig. 14 Seismic salt-analog structure. (A) Seismic cross-section from Rowan et al. (2016), showing a putative geometry for the Kef N’Hal structure, before compression and squeezing leading to weld formation. White arrow show the equivalent position of southern flank of Kef N’Hal weld. (B) Extract of cross-section by Kergaravat et al. (2016) showing the squeezing and inversion of a salt limit between two minibasins. This geometry could likely account for the geometry now observed at the SAK. Black arrow shows the equivalent position of the SAK flap, i.e., the pre-Aptian layers grouped in a flap south of the weld (see Figs. 2 and 15A). (C) Examples of recumbent geometries that could explain the GP deformation during Aptian-Albian (slightly modified after Dooley et al., 2015). On the left, the result of an analog modeling. On the right, the reproduction of a seismic line from the Santos Basin.

In the text

	Fig. 15 (A) Re-interpreted cross-section of the GP. Location of the map on Figure 6 in thick blue line. (B) Sketch showing a possible structural configuration of the system during Aptian-Albian halokinesis.
In the text

Fig. 16 Model of evolution for the Culminating Range during the Mesozoic-Cenozoic evolution. (A) First stages from Jurassic to Neocomian, initial weak halokinesis, hypothetically at the vertical of inherited rifting normal faults. (B) Major halokinetic event in the Aptian-Albian, development of megaflaps in Kef N’Hal and GP, probable emplacement of salt canopies at least in the GP, overturning the older Mesozoic serie. An extensional activity is suggested. (C) Paleogene stage of compression, eroding partially, wiping out the canopies and welds formations. (D) Miocene stage of compression, during the main Tell phase, with the emplacement on top of the Ouarsenis massif of the external nappes with the maximum horizontal displacements.

In the text

	Fig. 17 New proposition for cross-section of the Ouarsenis massif, taking into account the important development of salt structures, consequently impacting the possible estimates of shortening. Location on Figure 1.
In the text