Abstract
In the sub-Alpine chains of Haut Provence, SE France, a very well-exposed Mesozoic sequence showing rapid thickness and facies changes associated with Jurassic and Cretaceous extension on the margin of the Ligurian Tethys has been deformed by ‘Alpine’ compression which occurred from the Late Cretaceous to the Pliocene. Although the geology has been very well known for decades, aspects of the structure remain enigmatic and cannot be explained by either Mesozoic extension or Alpine shortening alone. We infer that some deformation resulted from salt tectonics. A completely overturned, highly condensed Jurassic section near Barles village resembles the elevated roof of a Triassic salt body in a deep-marine basin. This carapace became overturned as a flap in the Middle Jurassic when salt broke out at the seafloor and overran the inverted flap as an allochthonous extrusion, comparable to those in the deepwater Gulf of Mexico or Angola. Later, Alpine compression exploited the weakness of the salt sheet as the Digne Thrust moved over the inverted flap. Although the flap is in the footwall of the thrust, evidence of soft-sediment deformation and other anomalous structures within the flap suggest that it did not originate as an overturned footwall syncline.
It is our conviction that seismic data generally cannot be interpreted reliably without knowledge of, and reference to, geological structures in the field. For many types of salt-related structures however, one could argue that the opposite is true. A wealth of seismic images and physical models shows many types of salt structure in different tectonic settings and at different stages of structural maturity. Knowledge of these analogues helps us interpret what salt structures might originally have looked like, even though they are now deformed in an orogenic belt. We are by no means the first to try to do this. Previous attempts include Mascle et al. (1988), Dardeau & de Graciansky (1990), de Ruig (1992), Jackson et al. (2003), McClay et al. (2004), Canerot et al. (2005), Jackson & Harrison (2006) and Sherkati et al. (2006). The first two of these papers specifically addressed a number of the general Alpine problems discussed here.
In this paper we attempt to be more specific and re-interpret a piece of field geology that, although well-known, still has some puzzling anomalies. This paper may be the first documentation of truly allochthonous salt that extruded pre-orogenically but left a sedimentary record still recognizable in a present-day mountain belt.
Tectonic setting and history
The sub-Alpine chains of Haute Provence have been studied for many years, so the stratigraphy and structural geology are very well known. Lemoine (1973), Gigot et al. (1974), Gidon (1975), Lemoine et al. (1986), Fry (1989), Lickorish & Ford (1998) and Ford & Lickorish (2004) have all documented the evolution of the Tethyan passive margin of this part of the world and its subsequent deformation during Alpine compression. Gidon's authoritative web site (Gidon 2010) compiles sections, diagrams and photographs which illustrate in great detail the geology of this part of the French Alps.
The Chevauchement de Digne (the Digne Thrust) system is an important part of the sub-Alpine arc in Haute Provence (Fig. 1). A detailed account of its evolution is given by Lickorish & Ford (1998) while Vann et al. (1986) describe the phased evolution of the mountain front south of the city of Digne-les-Bains.
(a, b) Location and simplified geological map of the sub-Alpine chains of Haute Provence and the Digne Thrust system. White dotted line is the approximate edge of the Valensole platform in the Early–Middle Jurassic. Grey dashed fault line is the trace of the buried Durance Fault. Rectangle shows the location of (c). (c) More detailed geological map of the area of interest. Dotted axial traces are those of the youngest (long wavelength) Tertiary folds. White arrow shows the minimum displacement and transport direction of the Authon Thrust sheet, yellow arrow shows the same for the Digne Thrust sensu stricto. Red arrow indicates Jurassic olistoliths. The locations of section lines in Figures 4 and 5 are indicated by the numbered black lines. Key to colours is given in Figure 2. Maps simplified from BRGM Carte Geologique de France 1:1 000 000 and BRGM Carte Geologique de France 1:250 000 Sheet 35 – Gap.
The detachment surface of the Digne Thrust system lies in the Upper Triassic (Keuper) gypsiferous shales (Fig. 2), which typically form the base of the hanging-wall stratigraphy The transport direction was towards the SW judged from the ‘Bow and Arrow’ rule of Boyer & Elliott (1982) and from previous work (e.g. Fry 1989; Lickorish & Ford 1998). In its central section around Digne-les-Bains, two distinct thrust sheets override the foreland. The lower unit, the Authon sheet, has a minimum displacement of 20 km measured in the transport direction from its onlapped front NW of Digne-les-Bains (around Melan) to its root south of Rochebrune near Remollon (white arrow on Fig. 1c). The upper unit, the Digne Thrust (sensu stricto), has a minimum of 20 km slip measured from the eroded mountain front near Digne-les-Bains to its most eastern outcrop in the Clue de Verdaches NE of Barles (yellow arrow on Fig. 1c).
The thrust system has a polyphase deformation history. Thrusting must have started before the Eocene because the east or SE trending tip folds of a splay from the Digne Thrust near Chateauredon (Fig. 3) are overlain unconformably by nummulitic limestone of Priabonian age. Within the Authon thrust sheet north of Digne-les-Bains, beyond the depositional limit of the nummulitic limestone, east–west trending synclines (classically ‘Pyrenean-Provençal’ folds) were infilled and buried by Oligocene red-bed molasse (A on Fig. 1c). During or after the Oligocene, the Digne and Authon thrust sheets (probably then a single thrust sheet) advanced over bedding-parallel footwall ‘flats’ in Oligocene red beds. Jurassic olistoliths eroded from and overridden by advancing thrusts on these flats suggest that the flats were land surfaces. Such features are visible above the village of Baudinard (red arrow on Fig. 1c), below the Digne Thrust east of Esclangon (Fig. 1c) and (probably) above the village of Le Caire (4°30′,49°16′).
Folds at the tip of the upper splay of the Digne Thrust are overlain unconformably by Eocene nummulitic limestone near Chaudon Norante (CN). See Figure 2 for legend. Map simplified from BRGM Carte Geologique de France 1:250 000 Sheet 35 – Gap.
Onlapping Miocene marine sediments mark the cessation of slip on the Authon Thrust. The Digne Thrust, on the other hand, has at least 12 km of Pliocene ‘out-of-sequence’ displacement over the Mio-Pliocene molasse of its foreland basin (Valensole Plateau). This displacement also seems to have been over a land surface because olistoliths of resistant Tithonian limestones, overridden by the thrust, are visible all along the outcrop of the thrust front near Digne. During both the Oligocene and the Pliocene, sedimentary infilling of the topography seems to have created a level surface which favoured displacement of a major thrust sheet. A good basal detachment would have been provided by Triassic evaporitic shales (and probably salt).
Finally, a post-Pliocene pulse of compression gently folded the Digne Thrust to form the La Robine syncline (B on Fig. 1c) and the tectonic half-windows of Barles and Verdaches (indicated on Fig. 1c by their axial traces). Beyond the present-day erosional limit of the thrust the Mirabeau anticlines (C on Fig. 1c), SW of the thrust front at Digne-les-Bains, also fold the Pliocene.
Our previously unpublished cross-sections (Figs 4 & 5) and the generalized structural restoration (Fig. 6) illustrate some of these structural relationships. The Digne Thrust disrupted Jurassic extensional structures of the Tethyan margin and translated thick basinal units over the much thinner platform sequence of the Valensole Plateau. The Liassic (Lower Jurassic) section at Barles is no more than 250 m thick and mostly composed of limestones, compared to a maximum Liassic thickness of at least 1700 m of mostly shales in the hanging wall of the thrust system (Haccard et al. 1989). The section within the hanging wall becomes thinner and less basinal southwestwards in the transport direction. At the mountain front near Digne-les-Bains (A on section 4, Fig. 5) the Liassic section is therefore only a few hundred metres thick, dominated by limestones and comparable with the section at Barles (B on section 4, Fig. 5). Indeed, one can make the case that the footwall cut-off of the Digne Thrust is close to the southern erosional limit of the overturned section of the Barre de Chine, the main focus of this paper (at B on section 4, Fig. 5), and the hanging-wall cut-off is close to the mountain front (at A on section 4, Fig. 5). The thrust therefore telescopes the transition from basin to platform.
Interpretive sections 1 and 2 from the northwestern part of the area of interest. ‘A’ is a cargneule-cored upright isoclinal anticline with a reduced Liassic and Middle Jurassic envelope. It is probably a compressed salt diapir. Sections are at 1:1 scale. See Figure 2 for legend.
Interpretive sections through the central part of the area of interest. Section 4 shows the Barre de Chine (the flap marked ‘B’ on the section) in regional context (see Fig. 1c for location). The two versions of section 4 show alternate interpretations of the deep structure, one as a basement-involved thrust (section 4a) and the other a footwall shortcut (section 4b). Sections are at 1:1 scale. See Figure 2 for legend.
Sequential restoration of section 4a (constructed using 2D Move) showing the translation of basinal rocks over the platform of the Alpine foreland in the hanging wall of the Digne Thrust. See Figure 2 for legend.
The northeastern limit of the outcrop of the Digne Thrust is bedding parallel in Triassic in both its hanging wall and footwall. It is folded by a deeper thrust sheet which is probably basement involved because it elevates Carboniferous coal measures and basal Triassic sandstones which stratigraphically underlie the main Upper Triassic detachment surface. The two different interpretations of section 4 (Fig. 5) are essentially just different interpretations of this lower sheet: one as a basement- involved thrust and the other as a footwall shortcut of a Mesozoic extensional fault.
The entire hanging wall of the Digne Thrust is between 7 and 10 km thick and comprises the post-rift Mesozoic sequence (Upper Jurassic–Upper Cretaceous) and the older part of the foreland-basin megasequence (Eocene marls and the Eo-Oligocene turbidites of the Grès d'Annot).
The overturned section of the Barre de Chine (a flap associated with allochthonous salt?)
Despite all this complexity, the structural style in the Digne Thrust system is comparable to that in the external zones of mountain belts elsewhere in the world. The thrust-related folds are typically asymmetric with long gentle limbs and short steep or slightly oversteepened limbs, without the wholesale overturning associated with nappes of the more internal parts of mountain belts. There are two exceptions to this conventional style. One is an isoclinal upright anticline in stratigraphically thinned Lower and Middle Jurassic rocks near Nibles (Chateaufort) (A on Fig. 4). The other, the subject of this paper, is the Barre de Chine just SW of Barles. Here in the footwall of the Digne Thrust is a horizontal but inverted section of anomalously thin Lower and Middle Jurassic limestone and shale.
Traced up the hillside from the Barles road, subvertical Rhaetian–Middle Jurassic strata overturn to horizontal (Fig. 7) to form an impressive limestone cliff, the Barre de Chine, making the skyline above rolling meadows underlain by poorly exposed, easily weathered Callovian–Oxfordian black shale or Terres Noires (Fig. 7). The entire geometry is clearly seen in the Google Earth images and geological map (Figs 8 & 9).
Panorama of the Digne Thrust trace (now eroded) above the overturned flap of the Barre de Chine and the Triassic cargneule breccia in the valley to the right (the presumed site of the diapir).
Oblique Google Earth image of the flap on the Barre de Chine. View direction is towards the west, and the relief has a vertical exaggeration of 1.5×. The flap is overturned to the SW and overturned strata are marked conventionally with a Y.
This overturned limb is presently 3 km long in the tectonic transport direction and 2.5 km wide along-strike, but before recent erosion it may have been up to 4 km long and at least 7 km along-strike (the presumed eroded area is between the solid and dashed white lines on Fig. 9a).
(a) Vertical Google Earth image and (b) simplified geological map of the flap on the Barre de Chine showing its present-day extent (solid white line) and its probable original extent before erosion (dotted white line). Trace of the Digne Thrust is dashed red on the image, solid black on the map. Red is autochthonous Carboniferous and Lower Triassic sandstone, orange (t) is Triassic shale, dolomite cargneule and evaporite, mauve (l) is Lias and part of the Middle Jurassic, pale blue (jn) is Callovian–Oxfordian Terres Noires (black shales), dark blue is Upper Jurassic limestone, green (ki) Lower Cretaceous and fawn (o) is Oligocene red beds. See Figure 2 for further details. Map simplified from BRGM Carte Geologique de France 1:50 000 Sheet 3440 La Javie.
The stratigraphic succession of the flap is complete, though only a seventh of the thickness of equivalent strata in the hanging wall of the Digne Thrust about 1.5 km away. The flap contains hard grounds and death assemblages of ammonites, crinoids and bivalves, suggesting hiatuses in an open-marine or pelagic setting (Fig. 10a). Although they are completely overturned in the flap, the Liassic and Middle Jurassic limestones are not internally deformed. There is no cleavage, the stylolites are bedding-normal and show no sign of shear strain (Fig. 10b). Carbonate grains have recrystallized and twinned, but there is no evidence of significant internal strain (Fig. 10c). Because these completely inverted rocks lie in the footwall of the Digne Thrust (which has a displacement of at least 20 km), this lack of strain is anomalous. Reference to the fundamental models of simple shear (Ramsay 1967; Ramsay & Graham 1970) suggests that if the limestone had been indurated before overturning and thinned by horizontal ductile simple shear associated with the thrust, the stratigraphic thinning in the Lias and Middle Jurassic would imply a shear strain of more than 6, equivalent to a strain ratio of more than 40:1. At the depths implied by the overlying hanging wall of the Digne Thrust (7–10 km), a strain this high would have produced a carbonate schist, mylonite or cataclasite. No such rocks exist here.
(a) Hard-ground death assemblage within the Liassic section at Barles indicating depositional hiatus. (b) Undeformed bedding-normal stylolites resulting from compaction in Liassic limestones at Barles. (c) Thin section of Liassic limestone from the Barre de Chine suggesting a lack of significant internal strain.
The minor structures in the overturned flap are also anomalous. The asymmetry of minor folds (Fig. 11) is not the Z-shape expected in an overturned footwall syncline produced by Alpine shortening; we have not seen the Z-shaped folds suggested by Gidon on his cross-section of the Barre de Chine (Gidon 2010) and are unsure whether or not this was merely diagrammatic. All of the folds we have seen there are S-shaped and verge to the SW.
Panorama of the Barre de Chine outcrop showing S-folds (blue dotted lines) and an array of minor extensional faults (black lines). These structures have the wrong sense to have been generated by strain in the Digne Thrust footwall, and are interpreted here as gravitationally driven (either sliding downslope before overturning or spreading after overturning on a soft sub-stratum).
The cliff exposures of the flap are also cut by a number of SW-dipping normal faults (Fig. 11). These trace in relatively planar fashion through the competent limestones, although they may sole out in the poorly exposed inverted Oxfordian shales beneath them if the faults postdate overturning. Like the folds, these faults cannot be associated with orogenic deformation in an overturned footwall syncline. They must be interpreted either as gravitational break-away faults that formed before overturning, or as faults associated with spreading or layer thinning under load after overturning. Either way, taken together, the minor folds and faults suggest deformation of strata that were soft and cool enough to undergo major distortion by folding under low temperatures and pressures, yet were brittle enough to fault.
Further evidence of soft-sediment deformation is near the stratigraphic top of the Middle Jurassic limestone-shale section, in the subvertical limb of the flap. Competent limestone units form wide lenses possibly equivalent to the limestone lentils that formed next to diapirs in the La Popa Basin, Mexico, as sea level fell and diapiric shoals emerged (Giles & Lawton 2002). One of the faults cutting these lentils (Fig. 12) is particularly significant. Its surface must be parallel to bedding for much of its exposed length, but towards the structurally lower part of the outcrop it cuts up stratigraphic section into Callovian black shales. Here the hanging wall has degenerated into chaotic debris. This sort of structure indicates a gravity-driven slump in freshly deposited sediments which disintegrated at the seafloor. The precise transport direction cannot be specified, but the trace of the ramp cuts up stratigraphy towards what is now the SW. The fault must have formed on an unstable slope, the dip-slope of beds which were later overturned by folding.
Steep limb of the flap near Barles (the Barre de Chine is the skyline). The competent limestone beds tend to be lensoid (white arrows). The thrust is a gravity-driven synsedimentary slump cutting up stratigraphic section with chaotic debris in the hanging wall.
Triassic evaporites
The regional Triassic stratigraphy consists of a basal quartzitic sandstone (Werfenian) overlain by Muschelkalk dolomite, then by variegated gypsiferous shales of the Keuper (Fig. 2). In tectonically disturbed areas, outcrops of a cavernous chaotic breccia of carbonates and evaporite (known locally as ‘cargneule’) have been interpreted as being formed by de-dolomitization associated with gypsum-derived sulphate (Grandjacquet & Haccard 1975).
The Digne and Authon thrust sheets follow the gypsiferous shale. The shale section is only some 20 m thick where exposed in the hanging wall of the Digne Thrust, east of Barles (Fig. 1). In stark contrast, in the immediate footwall of the thrust just north of Barles, an almost 2 km wide accumulation of chaotic cargneule breccia and gypsiferous earth (Fig. 13) stratigraphically underlies (but structurally overlies) the anomalous inverted section. The cargneule resembles the chaotic caprock breccia on sub-aerially exposed evaporite diapirs we have seen in the Great Kavir and Hormoz Basins (Iran), Paradox Basin (USA), La Popa Basin (Mexico) and the Maritimes and Sverdrup basins (Canada).
(a) Chaotic cargneule breccia and (b) highly deformed, foliated gypsum both near Barles in the Triassic outcrop north of the flap.
Geological anomalies of the Barles flap
The flap is not strained despite being overturned by 180°, and is only one-seventh the thickness of adjoining equivalent strata. It is therefore unlikely to have been formed as the completely inverted limb of a footwall syncline of the Digne Thrust during Alpine shortening.
Minor folds and faults in the overturned limb are geometrically incompatible with the structures expected on the overturned limb of a thrust-related footwall syncline. The normal faults and the asymmetry of minor folds are, however, compatible with gravitationally induced creep and slumping of tilted seafloor during the Mid-Jurassic.
The rocks in the overturned flap were deposited in an open-marine setting and, although anomalously thin, are stratigraphically complete between the Early Lias and the Callovian.
Hard grounds within the sequence record non-deposition, suggesting an elevated position in the basin.
The anomalous flap adjoins a mass of chaotic, brecciated material resembling the dissolution residue common above sub-aerially weathered evaporite diapirs.
Nearly all the features described above indicate a setting close to the seafloor and assuredly not beneath the 7–10 km overburden carried by the Digne Thrust. We suggest that the association of the anomalous features strongly implies salt tectonics. No halite is exposed in the Barles area, but this is common in sub-aerially exposed evaporite diapirs. Our interpretation includes the possibility that overlying salt was welded out or dissolved after the flap formed.
Interpretation
We interpret the overturned Jurassic section of the Barre de Chine as a salt-related flap that was originally the thin roof of a broad, gentle, shallowly buried salt body. These roof sediments were then overturned as the salt body inflated and rotated strata along its peripheral fold scarp before breaking out and flowing over the inverted flap (Fig. 14).
Cartoon evolution of the Barles flap (left-hand column) and its secondary structures (right-hand column). (a, b) Middle Jurassic folds and faults form by gravity-driven instability on the steepening limb of a salt swell. (c) Outer-arc extension on the diapir shoulder thins the roof and causes reactive diapirism. (d) Salt breaks out from the salt swell as continued rotation of the roof sediments rotates older structures. (e) Flap rotates to horizontal, aided by compaction of the Upper Jurassic Terres Noires black shales, inverting older structures. In the Palaeogene the allochthonous salt canopy or weld is exploited by the Digne Thrust.
The anomalously thin Lower and Middle Jurassic section is explained as a condensed carapace that accumulated slowly on the elevated roof of a salt structure in an open marine basin in the Ligurian Ocean (Fig. 14a). The uniform thinness of the flap suggests that in the Early and Middle Jurassic the crest of the underlying salt body was flat and several kilometres broad, forming a salt-supported plateau which presumably had been inflated by expulsion of salt from beneath thicker sedimentary load elsewhere. The original salt body may have been either a diapir discordant to its roof or a salt pillow concordant to its roof (in our evolutionary diagrams it is depicted as a pillow for simplicity). The roof stratigraphy suggests that the diapir was initiated in the Early Liassic and grew though Liassic rifting and the Late Liassic and Middle Jurassic post-rift subsidence. Regionally the salt body was located within the transition from the Valensole Platform to the Alpine (Dauphinois) Basin, which was later telescoped by the Digne Thrust.
As the salt structure continued to inflate and rise, it arched the accumulating Liassic and Middle Jurassic roof (Fig. 14b). Along the periphery of the rising salt, a bathymetric fold scarp formed as a monocline separating the elevated roof from the lower-lying peripheral strata. Such fold scarps in the seafloor are visible today in the Gulf of Mexico wherever salt has inflated below the Sigsbee Escarpment (Fig. 15). On the Sigsbee Escarpment, fold scarps overlying inflated salt are several kilometres wide and up to 1 km tall.
Break-out of allochthonous salt through the upper hinge of a monoclinal fold scarp. (a) Schematic cross-section and (c) seismic example from the Sigsbee Escarpment, northern Gulf of Mexico before break-out. (b) Schematic cross-section and (d) seismic example from the Santos basin, Brazil of a completely overturned flap formed by rotation of the monoclinal fold limb as salt breaks out overhead. In (d) the flap and the overlying allochthonous nappe are both part of the Aptian layered evaporites, but the scale is similar to that of the Barles flap. Cross-sections based on Hudec & Jackson (2006); seismic from CGG Veritas, courtesy of Carl Fiduk.
Fold scarps of unlithified sediment cannot steepen indefinitely without failing. The Barles fold scarp sheared downslope to form the minor S-folds whose asymmetry resembles that of gravity-induced knee folds or cascade folds. The scarp also extended by normal faulting. As noted, some normal faults cut the strata at a high angle such as those in the cliffs of the Barre de Chine (Fig. 11) and other normal faults formed as the fold scarp delaminated as slumps or slides that tore away extensionally at a headscarp, forming the synsedimentary fault illustrated in Figures 12 and 14. At some point salt must have broken out through the roof and escaped at the seafloor, most likely at a crestal graben formed by outer-arc extension (Fig. 14c–e). Formation of such crestal grabens and associated reactive diapiric walls are illustrated by physical models of drape monoclines (Vendeville et al. 1995; Withjack & Callaway 2000). The possible location and mechanisms of salt break-out at the Barles flap are discussed below.
Although recent erosion has removed all sign of it, we infer that during the deposition of the Callovian–Oxfordian black-shale (Terres Noires), the salt body began its crawl across the seabed with the overturned limestone flap folding beneath it (Figs 14 & 15b). The limestone was a strong competent unit between salt and mud, and compaction of structurally underlying Oxfordian mud may have had a role in tightening and overturning the flap. Exposure within Callovian–Oxfordian strata below the Barre de Chine is not good enough to display the discordances and folded onlaps that probably formed as the flap was overturned.
The seafloor salt break-out must have been at least 4 km×7 km in area and therefore comparable to one of the allochthonous lobes that comprise the Sigsbee Escarpment of the present-day Gulf of Mexico. The area from which the salt was derived welded and collapsed to form the cargneule breccia (our assumed cap-rock) north of Barles. After halite was removed by viscous flow and dissolution, a subhorizontal weld overlay the overturned section. We suggest that this weak surface was exploited by the Digne Thrust during Alpine deformation.
Analogous flap structures on seismic profiles in the field and in models are described in the following section.
Discussion
A flap of limestone overturned in a marine halokinetic setting leads to a number of geological questions which can be answered with varying degrees of confidence.
Can a limestone slab be completely inverted by gravity tectonics?
The answer is yes. Flaps analogous with the Barles structure were described by Harrison & Falcon (1936) from the Dezful Embayment of the Iranian Zagros foreland (Fig. 16); indeed these authors coined the term ‘flap’ for this type of structure. They described how erosionally resistant Asmari limestone buckled up through the evaporitic Gachsaran Formation then gravitationally collapsed to form various secondary structures, three of which are relevant here. (1) On the limb of a tightening anticline, limestone strata detach on weaker beds like shale as parasitic folds. Forming in response to gravity like a bending knee, these parasitic knee-folds mature to an asymmetry where their shorter limb faces away from the main anticline. Such asymmetry in a knee-fold would be the same as the S-fold shape in the Barles flap and is different from the asymmetry of parasitic folds formed by buckling. (2) A flap of Asmari limestone forms as the hinge zone of a tightening knee-fold breaks and its steep limb bends backwards away from the anticline crest to become completely overturned without breaking, as for the Barles flap. (3) Slip-sheets of Asmari limestone delaminate and slide down a dip slope as intact slabs that remain right-side-up. Here there are some parallels with the synsedimentary fault in the vertical limb at Barles (Figs 12 & 14). The Asmari limestone flaps were overridden by allochthonous sheets of Gachsaran gypsiferous marl, just as we think the Barles flap was overridden by Triassic evaporites. Much larger structures (10–20 km scale) are described from the Argille Scagliosa of the Apennines by Maxwell (1959) and Hsu (1967) (Fig. 17). All these examples indicate that even limestone can remain intact while responding to gravity while being rotated 180° to form an isoclinal recumbent fold.
Overturned flaps in the Zagros mountains, Iran: (a) Harrison & Falcon's (1936) explanation of how a flap detaches from an anticlinal limb and overturns under gravity. (b) Photograph of recumbent isoclinal syncline of Asmari Limstone equivalent to the final stage of flap overturning in (a) from Sherkati et al. (2006), courtesy of Jean Letouzey.
Cross-section from Maxwell (1959) showing a 10 km long flap isoclinally overturned in the Apennines, Italy.
In non-orogenic settings, some of the largest overturned flaps overridden by salt are those within the Aptian evaporites of the Santos Basin, Brazil (Mohriak et al. 2008) (Fig. 15d). In places, lower parts of the evaporite unit have pierced upper, more-layered evaporites. Some of these intrasalt diapirs have asymmetrically spread above overturned flaps of layered evaporite 0.5 km thick and 5 km wide.
How did the flap become overturned?
We are uncertain. The Alpine overprint destroyed the inferred diapir and its allochthonous outflow, so the evidence is circumstantial. Several theoretical possibilities depend on where and when the roof of the salt body broke. The rupture that freed the flap could have been either on the shoulder (Fig. 14) or the crest (Fig. 18a, b). In each of these models the flap remains attached to the flank of the salt body – it never breaks away to become a raft. This folding that rotated the flap could have occurred either before or after salt break-out from the salt body (Fig. 18a, b), and the mechanism could have involved either a fixed (Fig. 14) or rolling hinge (Fig. 18a, b and discussion below).
Cartoon showing alternative rolling-hinge models for overturning of the flap. (a) A rolling anticline hinge overturns roof strata prior to salt break-out, analogous to the front of a tank-tread. (b) A rolling syncline hinge, analogous to the rear of a tank-tread that overturns the flap after early crestal salt break-out.
A steep, narrow shoulder typically surrounds a flat-topped diapir that is inflating. Structurally the shoulder is a peripheral monocline whose middle limb defines a fold scarp as along parts of the Sigsbee Escarpment (Fig. 15a, c). Outer-arc extension in the monocline crest can ultimately tear the shoulder. The inflating salt that escapes from this vent extrudes down the steep fold scarp (Figs 14, 15b & 19). Freed of the anchoring roof, the monocline's middle limb rotates outwards past the vertical to become the inverted flap.
Three different physical models showing break-out of diapiric salt from extensional grabens in either (a) crestal or (b) peripheral locations, seen from obliquely overhead. (c) Cross-section showing coherent overturned flaps of the original roof, shedding thin debris flows on the periphery of the diapir. On the left, allochthonous salt extruded over the overturned rim onto the surrounding plain. On the right, the roof was thrust over the overturned flap and debris flows, along with allochthonous salt in its hanging wall. Models courtesy of Tim Dooley, Applied Geodynamics Laboratory, Bureau of Economic Geology, University of Texas.
Alternatively, salt may escape from the crest of a peaked diapir, where outer-arc extension thins the roof. If so, one half of the roof steepens and rotates outwards before ultimately overturning as a flap. For the same-sized diapir, this flap is much larger than the flap inverted from a narrow shoulder.
Both the above hypotheses imply a fixed hinge in which a flap of fixed length swings upwards and outwards like a trapdoor. The present-day flap is 3 km long, so the original salt body must have had significant elevation particularly if the flap represents its peripheral fold scarp. During trapdoor-style rotation the flap might be vulnerable to erosion if not protected by in-filling sediments.
Harrison & Falcon (1936) favoured this sort of fixed hinge for their flaps in Iran (Fig. 16). They envisaged that, once the upper limb of the syncline rotated past vertical, it sagged under its own weight. Gravity spreading ejected the stratigraphically overlying Gachsaran gypsiferous marl from the core of the recumbent syncline. Lateral erosion then removed the extruded marl.
An alternative hypothesis for flap rotation involves a rolling hinge. If the salt did not break out until after significant lateral displacement (Fig. 18a), the anticlinal hinge would have migrated as the roof strata rotated (as if being fed into the front of a tank-tread). If the salt broke out early, the western margin of the flap would have become separated from the rest of the roof and been overrun by an allochthonous evaporite sheet. The upper limb of this syncline would then have become the overturned flap by rolling forwards (westwards) like the back of a tank-tread (Fig. 18b). In either the anticlinal or synclinal type of rolling hinge, the inverted flap need not rise much above regional.
How was the flap protected during overturning?
Protection was provided by overlying evaporites and adjoining onlap. Stratigraphically below but structurally above the Lower–Middle Jurassic flap, Triassic evaporites would have buried and protected the flap as it rotated. The weight of extruding Triassic evaporites would have added to the load, flattening the overturned flap.
At the same time, the Jurassic black shales would likely have onlapped the flap as it rotated, providing a protective buttress to its flank. Older Oxfordian onlaps would have been at a low angle, recording when the fold scarp was still gentle. Younger onlaps are likely to have been at higher angles as the flap steepened. As folding progressed, the shifting ratio between aggradation rate and rotation rate would create other discordances in the Upper Jurassic black shales. The right balance is needed for these rates. If aggradation was too rapid, it would infill slopes and remove much of the gravitational instability that drove overturning. If aggradation was too slow, the exposed and unbuttressed flap would be degraded or destroyed by erosion and disaggregation. The black shales themselves would deform as they rolled up in the core of the evolving syncline. None of these features are readily visible in the hillslope meadows.
Did the minor folds and faults in the flap form before or after overturning?
We are uncertain. The minor folding and faulting could have taken place during either early or late stages of flap rotation (either gravity gliding down a palaeoslope before overturning or gravity spreading of an already overturned limb).
At the early stage, the fold scarp would creep downslope under gravity, as depicted in Figure 14. Comparable folds are present in analogous settings above some active salt diapirs in the North Sea (Davison et al. 2000).
Davison et al. (1996) proposed shear traction for the formation of recumbent folds in a smaller but otherwise similar overturned flap next to the Jabal al Milh diapir in coastal Yemen (Fig. 20a). However the ability of weak evaporites to shear underlying sediment requires the limestone flap to be unrealistically weak. We suggest this makes a gravity-driven mechanism more likely.
Cross-sections of overturned flaps on the flanks of salt diapirs that extruded salt laterally: (a) Jabal el Milh diapir, Yemen (Davison et al. 1996) and (b) Riedel mine in Hänigsen diapir, Germany. Both of these flaps are significantly smaller than the flap at Barles.
Some flaps associated with diapirs appear to be orogenically driven rather than gravity driven. For example, Hänigsen Diapir in Germany has a synclinal flap whose upper limb is overturned by 170° (Fig. 20b). The Hauterivian synclinal core dates the overturn as no earlier than Early–Mid-Cretaceous. The upper part of the flap is truncated by an overhang of allochthonous salt and by an unconformity, both of Late Cretaceous age. Thus the allochthonous salt extruded after the flap was overturned and could not have caused the overturn. In addition, the 1 km thickness of the flap and 100 Ma time span of indurated rocks would require more than shear traction by overlying salt to overturn the flap. Instead, regional compression in the Late Cretaceous probably overturned the strong, thick flap and squeezed the diapir so that it expelled salt. Many other examples of contractional salt tectonics in the Late Cretaceous are present in the same basin (Baldschuhn et al. 2001).
The minor normal faults in the Barles flap represent brittle layer-parallel extension. As for the minor folds, a case could be made for the faulting before overturn (as depicted in Fig. 14) driven by the gravitational instability of a slope or, after overturn, driven by the weight of the overlying allochthonous sheet. For these normal faults, it may be possible to distinguish between pre-rotation and post-rotation origin by examining the downwards extension of the normal faults in the poorly exposed hillslopes below the flap. If the faults end abruptly against the onlapping Upper Jurassic black shale, the faults formed early before overturn. If on the other hand the faults of the flap sole out downwards into the black shales, they are likely to have formed (or at least been reactivated) after overturn.
Is the advance direction of allochthonous salt compatible with basin architecture?
This is arguable. In the Gulf of Mexico, Brazil and other passive margins affected by allochthonous salt tectonics salt sheets advance mostly seawards, driven by prograding sediment load on the continental margin and guided by regional dip. In the Lower and Middle Jurassic, the platform beneath the Valensole Plateau was a large fault block bounded to the west by the Durance Fault (Fig. 21). West of the fault is the SE France Basin which, from Triassic to Middle Jurassic time, may have been deeper than the Ligurian Tethyan basin to the NE of the Valensole Block (Debrand-Passard & Courbouleix 1984). Barles is close to the northern limit of the block where it plunges down into a common basin; the inferred westwards advance of allochthonous salt is possible in such a setting. After the Middle Jurassic the basin deepened to the east (Fig. 21). Allochthonous salt had already started to advance westwards by then, having been initiated in the previous setting of westwards deepening.
Schematic cross-sections showing the sequential evolution of the Barles diapir and the Digne Thrust in regional context. Sections are at 1:1 scale. See Figure 2 for legend.
Did the flap influence the local Alpine shortening?
Yes, by supplying a weak, flat footwall for the major thrust. The Digne Thrust overlay the flap but was removed from the Barre de Chine itself by recent erosion. The thrust does not reappear in outcrop until the SW side of the Barles half-window, where it cuts down into the sub-surface beneath the broad La Robine syncline and emerges in the mountain front at Digne-les-Bains (Fig. 1c). Here a train of moderately tight Triassic-cored anticlines overlie a gently dipping, planar section of the thrust (section 4, Fig. 5). We noted earlier that these structures effectively represent the hanging-wall cut-off of the Digne Thrust, and their stratigraphy closely matches that of the flap at the Barre de Chine (section 4, Fig. 5). Their decapitated geometry leads us to speculate that these structures are the crests of original salt-cored anticlines within the Barles salt body.
Our hypothesis is that the path of the Digne Thrust was determined by the Barles allochthonous salt or by the weld that evolved from it (Fig. 21). This hypothesis is supported by the anomalously far advance of the thrust sheet. The strike length of the Digne Thrust, from its northern tip NE of Serres to the apex of Castellane arc, is no more than 80 km; its maximum displacement is 20 km, however, and that of the lower Authon sheet is similar. This is an abnormal ratio of displacement to strike length, much greater than the typical 7–10% (Boyer & Elliott 1982). Of course, some displacement is taken up in folding and minor associated thrusting beyond these tips, so the width could be greater (just as the displacement could be greater if the eroded hanging wall west of Digne-les-Bains is added). A lubricating sheet of allochthonous salt could well explain the anomalous efficiency of thrust advance. Butler et al. (1987) showed how thrust systems roll forward where the detachment improves. The limits of the Digne Thrust system may therefore mark the original extent of Jurassic allochthonous salt.
How does the velodrome fit in?
We are uncertain. In the southern part of the Barles window, immediately beneath the folded Digne Thrust, is a complex overturned syncline whose half-ellipsoidal non-cylindrical shape has prompted the geological nickname by which it is universally referred to today in French literature. It resembles a cambered cycle track.
In its oversteepened northern limb, overturned Oligocene red beds overlie Lower Cretaceous and Upper Jurassic limestones above a strongly folded unconformity. The Oligocene rocks are themselves unconformably overlain by marine Miocene section in which there is at least one further angular unconformity and much-folded onlap. Pliocene continental deposits unconformably overly the oversteepened Miocene (Fig. 22).
Panorama of the Velodrome and the Digne Thrust (looking NW). Progressively rotated unconformities (Oligocene on Upper Jurassic and Lower Cretaceous, Miocene on Oligocene, Intra Miocene, Pliocene on Miocene) record the Tertiary overturning of this syncline. The Digne Thrust clips the skyline where olistoliths (J) of resistant Upper Jurassic limestone testify to transport over a land surface.
The conventional (and probably the most reasonable) interpretation of these features follows the notion of progressive deformation at mountain fronts, comparable with that described by Riba (1976). However, the unconformities have much in common with those at the margins of minibasins as described by Giles & Lawton (2002) from the La Popa area of NE Mexico. It does not seem impossible to us that salt withdrawal at depth during the Tertiary might have contributed to the geometry of the Velodrome. The Authon thrust front lies beneath the unconformities of the Velodrome edge (section 4, Fig. 5). Is it conceivable that the long ‘flat’ of the Authon thrust sheet over the Oligocene land surface was facilitated by a terrestrial salt glacier in Oligocene time, and that this progressively deflated and welded out beneath what is now the Velodrome?
Conclusions
Mascle et al. (1988) seem to have been the first authors to document how it is possible to see through an overprint of Alpine contractional tectonics to reconstruct salt diapirism during the late synrift and post-rift stages of the evolution of the margin of the Mesozoic Tethys. Dardeau & de Graciansky (1990) and Dardeau et al. (1990) have documented in detail, and structurally restored, a number of such diapirs in the French Maritime Alps and sub-Alpine Chains. The evidence for diapirism cited by these authors is the same as that presented here: local reduced stratigraphy, local unconformities and facies changes, cap-rock breccia, debris flows, slumps and disconformable contacts and other anomalous structures unlikely to result solely by extensional faulting or the inversion of such faults.
This paper has gone slightly further than previous articles by suggesting that, in the Middle Jurassic in at least one part of the Alps, the Triassic salt actually became allochthonous and formed a large salt nappe on the seafloor, comparable with one of the lobes of the present-day Sigsbee salt allochthon of the Gulf of Mexico. We reached this conclusion by interpreting the Lower and Middle Jurassic outcrop of the Barre de Chines near Barles as the erosional remnant of a ‘flap’ overturned beneath the salt as it crawled out over the Jurassic seafloor. We have tried to understand how such a structure might initiate and grow.
When we start to look at Alpine geology in this way, it seems reasonable to wonder about other field relationships either nearby (e.g. the unconformities of the Velodrome) or further afield. We pointed to anomalous length–displacement ratios of the thrusts in the Digne Thrust system and wondered whether an allochthonous salt body had influenced thrust advance. If so, how many other Alpine thrust sheets could have been lubricated by a cushion of allochthonous extrusive salt? How many other upright isoclinal folds are cored by Jurassic salt walls, such as that near the Digne mountain front? How many salt walls evolved into Alpine thrust welds? How many other apparently ‘Alpine’ structures nearby are actually the deformed remains of older halokinetic structures? These are all questions we would dearly like to see answered.
Acknowledgments
We thank Hess Corporation for supporting our in-house training class and fieldwork in France. We would like to thank CGGVeritas and C. Fiduk for permission to use seismic examples from GoM and Brazil. Thanks to J. Letouzey for the colour photo of the flap in Iran included in Figure 16 and to T. Dooley and the Applied Geodynamics Laboratory for the physical models in Figure 19. Thank you to the excellent drafting departments at AGL and Hess London for the help with production of the figures. MJ thanks Hess for field support and the Applied Geodynamics Laboratory for supporting his research.
- © The Geological Society of London 2012
References
Allochthonous salt in the sub-Alpine fold–thrust belt of Haute Provence, France
