Introduction

Terrestrial carbonates can form in different environmental conditions in continental basins. There is an exhaustive individualized treatment for each environment (Alonso-Zarza and Tanner 2010). Thereby, the calcrete and carbonate palustrine facies are extensively described in the literature for the Paleogene–Neogene time (Freytet and Plaziat 1982; Alonso-Zarza 2003). Fluvio-lacustrine carbonates dominated by microbialite facies show contrasting environmental significances with respect to calcretes and carbonate palustrine and laminar crusts. Their facies are found in fluvial settings (Arenas et al. 2007, 2015) and littoral areas within the photic zone of carbonate-rich lakes with limited input of siliciclastic grains (Dean and Fouch 1983; Gierlowski-Kordesch 2010). Finally, travertines and tufas represent carbonate chemical precipitates formed around seepages, springs and along streams and rivers, occasionally in lakes, within a vadose or occasionally shallow phreatic environment (Pentecost 2005).

The climatic and paleohydrological constraints of the terrestrial carbonate spectrum are varied. Whereas calcrete-palustrine facies indicate restricted water conditions under varying periods of drought (Freytet and Plaziat 1982), tufa and travertine require a constant supply of water, and wet and warm conditions favor their development (Viles and Pentecost 2007). Microbialites (also present in tufas and travertines), for their part, require stable water levels for encrusting microbial communities to thrive (Casanova and Hillaire-Marcel 1992). The unusual conjunction of facies found in the Borj Edouane Unit (BEU), filling the Pleistocene El Gara Basin, is an opportunity to understand the coexistence of terrestrial carbonates with such different water requirements in the same basin. Although this unit has been recently studied (Ghannem 2018; Ghannem et al 2019), new fieldwork and additional sedimentological analyses improve the interpretation of the unit. A first aim focuses on defining the facies associations, analyzing their relationships and stabilizing the depositional model. Special emphasis has been placed on understanding the circumstances that explain the asymmetrical distribution of facies along the basin, the paleohydrological and paleoclimatic constraints and tectonic control that favored the development of these depositional carbonate systems.

Geological setting and field area

The study area is situated at about 450 m asl and is a part of the North-West of Tunisia near the Tunisian-Algerian border in the northwestern of the Tunisian Atlas region (Fig. 1a). In the study area, the Quaternary deposits lie on the uppermost Lower and Upper Cretaceous formations (Fig. 1b), represented by a succession of carbonate facies from the Aptian (clays, marls and sandstones), Albian (Allam limestones, Mouelha limestones and clays) and Cenomanian-Coniacian (marls, clays and limestone beds). The Quaternary deposits consist of conglomerates, sandstones and limestones of Mid-Late Pleistocene age, and Holocene deposits (terraces, alluviums, alluvial cones) and carbonate soils (Fig. 1b). The Mid-Late Pleistocene carbonates that form the BEU (Burollet and Sainfeled 1956; Ben Haj Ali et al. 1985), object of this study, crop out through the center of El Gara Basin (Figs. 1b and 2). This fini-Neogene basin extends along a NE-trending fold (El Gara anticline), parallel to most Tunisian Atlassic folds (Fig. 1b, c), and was mainly filled with Quaternary deposits. At present, its watershed discharges into the Mediterranean Sea through Mellegue and Sarrat rivers. It has an elliptical contour covering an area of ca. 20 km2 between Sarrat, Mellegue and Ezarga Rivers. The carbonates of the BEU occupy the central area of the basin.

Fig. 1
figure 1

Geological setting of the Borj Edouane Unit (BEU) and location of the sedimentary logs displayed in other figures of this work. The box in 1b is the outline of Fig. 2

Full size image
Fig. 2
figure 2

Modified from Ghannem et al. 2019

Map of facies distribution. Between brackets, a reference of figures where the stratigraphic logs Gr (3), F (4), D (5), CCL11 (6), TJCB (7), U (9a), 16 (9c) and 25 (10) are represented.

Full size image

The tectonic framework of this area resulted from three successive compressive tectonic events that affected the region since the Cenozoic (Frifita et al. 2016; Hamdi et al. 2019). The two first events occurred during the Eocene and Upper Miocene respectively, with a NW–NNW shortening direction (Frifita et al. 2016; Hamdi et al. 2019). The third event occurred during the Quaternary (Upper Pleistocene) with NNW, NW to N–S shortening direction (Hamdi et al. 2019). These events reactivated a pre-existing conjugate fault-system and resulted in NE–SW trending folds, E–W to NW–SE dextral strike-slip faults, N–S sinistral strike-slip faults and NE–SW reverse faults (Frifita et al. 2016; Hamdi et al. 2019), (Fig. 1a, c).

Methodology

The terminology of Alonso-Zarza (2003) has been used for the description of calcretes, laminar crusts and palustrine carbonates. The description of microbialites follows the terminology of Grey (1989), and for travertines, the terminology of Pentecost (2005), and Gandin and Capezzuoli (2014). This work follows the distinction between tufas and travertines after the definition of Ford and Pedley (1996): groundwater-fed deposits precipitated under surface temperature and hydrothermal conditions, respectively. Throughout this work, the term fabric refers to the textural features that imply an internal ordering of the rock. Field work consisted of the study of 27 stratigraphic logs (up to 11 m thick) and more than 80 outcrops throughout the carbonate unit (Fig. 1). About a hundred of collected samples were analyzed by the Bernard calcimeter to measure the total carbonate (CaCO3). Twenty selected samples of the unit carbonates were analyzed by XRD through PANalytical X’pet X-Ray Diffraction spectrometer (System operating at 40 kV and 30 mA and reflection geometry 2θ) at the XRD Analysis Services of the Faculty of Sciences of Bizerte. The textural and petrographic analysis of more than 100 polished hand-specimens and about 90 thin sections were studied under binocular stereo microscope vision and transmitted-light-microscopy. Petrographic differentiation of carbonate phases in thin section has been carried out by the dual stain (Dickson 1965). Nine polished thin sections were examined under Cathodoluminescence (CL) with a unit model CCL 8200 mkIII (12–20 kV, 200–400 mA and 0.2–0.1 torr).

Stratigraphy and facies associations

The fill of the Pleistocene El Gara Basin consists of two units that are laterally transitional to each other (Fig. 2): 1) a marginal alluvial unit, consisting of an alluvial association, extends through the western and northeastern margins of the basin; and 2) the BEU, formed by terrestrial carbonates throughout the basin. The BEU consist of two facies associations: calcrete-palustrine and composite microbialite-travertine. The carbonate facies show an asymmetrical distribution through the basin: the first association extends across the western strip and the second covers the central and eastern strips (Fig. 2). The carbonate mineralogy of the BEU consists of low magnesian calcite and the insoluble residue content is always below 31%. The contents of quartz, phyllosilicates, and feldspar, ranging from 0–2%, 0–10, and 0–9%, respectively, are higher in the calcretes and very low in the travertines.

Alluvial association

The alluvial association crops out along a marginal fringe around the BEU (Fig. 2). It consists of conglomerates and sandstones with common intercalations of carbonate-rich mudstones in the transition to the carbonate unit (Fig. 3a). The conglomerates are clast-supported and stratified in tabular to lenticular beds about 0.5 m thick that can show basal erosive surfaces, internal structures, and clast fabrics (Table 1; Fig. 3a–c). Clasts are subrounded (from a few cm to more than 40 cm) and dominated by Cretaceous carbonates (up to 75%) and, in a lesser amount, quartz and quartzite within a matrix of sand-sized grains and micrite cemented by sparite (Fig. 3d). The sandstones form tabular to lenticular, massive to cross-stratified beds a few decimeters thick (Table 1, Fig. 3a). They are fine to coarse-grained litharenites consisting of carbonate (10–60%; calcrete and Cretaceous clasts) and quartz (20–80%) grains cemented by sparite. (Fig. 3e). Calcretized mudstones form decimetric massive beds with diffuse boundaries. Fining upward sequences from conglomerates through sandstones to carbonate-rich mudstones are common along the western border of the unit (Fig. 3a).

Fig. 3
figure 3

Alluvial association. a Sedimentary log Gr showing the transition from alluvial to palustrine facies (other symbols in Fig. 4, and Fig. 2 for situation). b Crudely stratified carbonate conglomerates with clast imbrication passing upward to cross stratified conglomeratic sandstones. c Detail of the lower part of (a) with internal erosion surfaces (arrowed), cross stratification (dotted lines) and alternation of pebbly (Pb) and cobbly beds (Cb). d Subrounded carbonate clasts (Ls) from the Cretaceous and quartz grains (Q) surrounded by a prismatic isopachous calcite rim; dispersed patches of internal micrite sediment (M). PPL (plane polarized light). e Mixed sandstones with Cretaceous lithoclasts (Ls) and quartz (Q) grains with corrosion gulfs (G) filled with clay-rich microsparite; patches of light sparite occupy the remaining porosity. Recycled Cretaceous planktonic foraminifers (arrowed). XPL (crossed polarized light)

Full size image
Table 1 Alluvial and calcrete-palustrine-laminar crust associations
Full size table

Interpretation. The organized conglomerates represent the sedimentation by sheet flows and locally developed stream channels, such as those documented in distal areas of stream-flow-dominated alluvial fans (Nichols 2009). Poorly organized beds indicate the action of more viscous hyperconcentrated flows. Sandstones represent channel fillings and sheet flood deposits in the distal areas of alluvial fans, where they pass basinward to distal alluvial plains encrusted by secondary carbonate and dominated by distal sheetfloods.

BEU associations

Calcrete–palustrine association

It forms the westernmost strip of BEU that edges and overlies the alluvial association (Fig. 2).

Calcrete facies

The calcretes predominate at the top of the fining-upward alluvial sequences (Table 1, Fig. 4a, b) and are interdigitated with palustrine facies, being difficult to separate from each other. Calcrete beds are tabular, up to 0.5 m thick, with a variable carbonate content ranging from calcareous paleosoils, through nodular to massive calcretes. Their microfabric consists of homogeneous micrite with floating sediment particles and secondary porosity (Table 1, Fig. 4c).

Fig. 4
figure 4

Calcrete facies. a Sedimentary log F (situation in Figs. 1 and 2) showing alluvial to calcrete upward transition. b Field view of (a): interbedding of crudely stratified, conglomeratic sandstones and mudstones crowned by a massive calcrete. c Quartz-rich (Q), micrite matrix with Cretaceous lithoclasts and foraminifera (arrowed); rootlet porosity (Rt) filled with sparite. PPL

Full size image

Interpretation. Calcrete microfabrics share features from both the alfa and beta types of calcretes (sensu Wright and Tucker 1991) and represent the calcite precipitation on carbonate-rich muds within the distal alluvial plains. The precipitation of carbonate took place in surficial soils within the shallow vadose-phreatic interface through drying-wetting processes in the transition (calcrete belt) from distal alluvial plain to palustrine environments (Huerta and Armenteros 2005).

Palustrine facies

They extend along the W and SW of the unit, forming the transition from calcrete to microbialite facies (Fig. 2), and occur as orange-colored tabular massive beds from 0.15 to 0.80 m thick (Table 1; Fig. 5a,b). Original textures are principally mudstones and wackestone with silt-sized quartz grains (1–5%), disperse intraclasts, and Cretaceous lithoclasts and fossils. The fossils consist of scattered fragments of gastropods and ostracods (Fig. 5c), and more locally, small oncoids and charophyte circular sections. An Helicidae gastropod, 4 mm wide [Helicella (Helicopsis) gigaxii barcinensis (Bourguignat)] and scattered micro-shells (≈ 60 µm in diameter) of gastropods have been identified; pedodiagenetic textures and porosities are common (Fig. 5c, d). Palustrine beds can intercalate intraclastic and oncoid-rich layers a few cm thick and can pass upward to carbonate laminar crusts at the top of alluvial fining-upward sequences (Fig. 6a). Laminar crusts occur in tabular beds 0.01–0.40 m thick that can show tepees and desiccation polygons (generally pentagons) from a few cm to 30 cm in diameter (Fig. 6b). They show an alternation (1–4 mm thick) of light and dark micritic laminae that conform to the irregularities of the substrate and can form centimeter-thick sets, locally discordant to each other (Table 1, Fig. 6c). Dark laminae show filamentous films. The laminated layers are commonly associated with massive intraclastic-peloidal-ooidal layers (a few mm thick) that have quartz-grains, gastropods, and aggregates of calcite prims (spherulites) (Table 1, Fig. 6c–g).

Fig. 5
figure 5

Palustrine facies. a Massive beds with diffuse boundaries and nodular brecciation orange colored by iron-rich mottles. Section D. b Hand specimen (PL in Fig. 3a) of micrite mudstones with orange ferruginous mottles and irregular cracks. c Mudstone (fine-grained limestone) with disperse quartz grains (Q), channel porosity (Ch, rootlet traces enhanced by brecciation), gastropod remains (G) and ostracods (O). Sample C1, PPL. d Detail of (B) showing a crack section (Cr), abundant silt-sized quartz grains (Q), ferruginous mottles (white-dotted contours), an articulated ostracod (arrow). PPL

Full size image
Fig. 6
figure 6

Laminar crusts. a and b Log CCL showing the laminar crust bed at the top crowning a sequence of chalky (sample 10) and nodular limestones (sample 11); tepee structure (arrowed). c Polished section of ooidal-intraclastic layers with thin intercalations of poor laminated crusts (Lc). d Erosional surface on the basal ooidal-intraclastic layer (dotted line), covered by a layer rich in spherulite prisms (see g) with dispersed ooids (O) and intraclasts (I). At the top, intraclastic texture rich in gastropods (G) that covers two micrite laminae (Dm) separated by a lamina rich in spherulite remains (white arrow); PPL. e Poorly-sorted ooidal texture and scattered quartz grains (Q), joined by a calcisiltite mass. Cortical ooidal laminae show clay birefringent fabrics (arrows). PPL. f Arch-like arrangements of prismatic calcite crystals (arrows). g Scattered calcite prisms resulting from the reworking of spherulites (arrows)

Full size image

Interpretation. The fossil components and the scarcity of quartz grains are useful to differentiate palustrine from calcrete facies. The presence of gastropod protoshells is a distinguished element in carbonate palustrine settings, characterized by harsh conditions in ephemeral ponds (Pérez Jiménez 2010; Bustillo et al. 2017). The presence of Helicella is an indicator of dry steppe conditions (Haas 1991), which combined with exposure features support the carbonate palustrine origin of this facies. The presence of desiccation features and laminar crusts suggest the existence of dry conditions (Armenteros and Daley 1998). The crusts show spherulites and dark filamentous films, which evidence cyanobacterial activity resulting in the stromatolitic-like structure of this type of laminar crust (Verrecchia et al. 1995).

Composite microbialite–travertine association

It is made up of three association facies: 1) microbialite (oncoids and stromatolites), 2) clastic microbialite-travertine, and 3) laminated travertine. The microbialites mainly occur in the center of the study area where they form a north–south strip that passes eastward into association 2, which, in turn, gradually evolves into travertines in the easternmost strip zone of the unit (Fig. 2; Table 2).

Table 2 Composite microbialite-travertine association
Full size table
Microbialites: oncoids and stromatolites

Oncoid facies form brownish orange, well lithified, tabular to wedge-shape beds, 0.1–0.6 m in thickness, separated by sharp surfaces, that are grouped into sequences up to 1 m thick (Fig. 7a, b). Oncoid beds show abundant microbialite fragments and dispersed charophyte fragments, ostracods and gastropods. In south-center outcrops, beds can be interstratified with stromatolites (Table 2, Fig. 7d–f), forming small fining upward sequences 1–2 cm thick (Fig. 7c). In central outcrops, beds are massive to crudely laminated (Fig. 7a, b). In the northeastern area, tabular, crudely stratified, oncoid-rich layers are associated with conglomerate-sandstone channel fill deposits that have Cretaceous clasts and horizontal and cross stratification. Oncoid size ranges from submillimetric to a few centimeters, defining the grain size of the oncoid facies. Two main textures are observed corresponding to oncosparite (laminated) and oncomicrite (massive) layers (Table 2; Fig. 7e–h). The oncoids have varied nuclei (intraclasts, mollusk bioclasts and microbialite fragments; Fig. 7f) and morphologies (Fig. 7g). Their coating consists of couplets (a few hundred µm thick) of dark micritic and light sparite laminae (Fig. 7g, h).

Fig. 7
figure 7

Oncoids. a and b Sedimentary log TJCB and field view of the oncoid facies. c Small fining-upward laminated sequences (arrow) on erosion surfaces (dotted lines). d Oncoids coated with a planar stromatolite; polished section. e Horizontally stratified oncoids with intercalations of planar stromatolites (Ps). f Intercalation of a pseudocolumnar stromatolite (Ps) between intraclastic, oncoid-rich layer (O); PPL. g Oncoids with different nuclei in a fine intraclastic matrix with scattered silt-sized quartz; an intraclast with an ostracod (arrowed); PPL. h An encrusted thallus section of charophyte (Ch) and oncoids with intraclastic nuclei in a fine intraclastic matrix; PPL

Full size image

Stromatolites form orange to brown laminated tabular beds up to 40 cm thick and have a limited extent in the south central areas of the unit where they are interdigitated with oncolitic and clastic microbialite levels (Fig. 7d–f). Beds can show irregular undulations and convex morphologies of a wavelength up to 0.5 m (Fig. 8a). Lamination is given by the alternation of dark (dense micrite) and lighter laminae (spongy micrite texture filled with sparite) forming couplets at different scales (Table 2, Fig. 8b, c). There are intercalations of millimetric diffuse alternations of irregular wavy micrite laminae and thicker clotted-peloidal spongy laminae with fenestral porosity, and irregular burrows that destroy the lamination and are filled with pellets and sparite (Fig. 8d).

Fig. 8
figure 8

Stromatolites. a Planar stromatolite bed with a clastic microbialite layer (Cl) and irregular wavy lamination in the lower level. Outcrop 433, Fig. 2. b Micro-wavy to pseudocolumnar stromatolite consisting of alternation of light and dark laminae; above (arrow pointing top), uneven-sized microoncoids with irregular bumpy lamination. TJCB4 sample; PPL. c Alternation of microlaminated (L) layers with microdomes (arrows), and massive spongy laminae (S) with vesicular porosity filled with sparite; PPL. d Clotted-micropeloidal fabric (≈ thrombolite-like fabric) with a broken intercalation of a microlaminated layer; irregular pores (arrowed), likely due to larval activity; PPL

Full size image

Interpretation. Microbialite structures, such as those studied here, are microbial reefs produced by calcification, attributed to microbial activities dominated by cyanobacteria (Burne and Moore 1987; Riding 2002). Both oncoidal and stromatolite microbialites have been recognized in fluvial tufas (Ordóñez and García del Cura 1983; Arenas-Abad et al. 2010), carbonate springs systems (Chafetz and Folk 1984; Pentecost 2005) and in shallow lakes and nearshore lacustrine systems (Schäfer and Stapf 1978; Dean and Eggleston 1984; Arp 1995; Martin Bello et al. 2019). They are also described in mixed fluvial-lacustrine environments (Arenas et al. 2007).

Although oncoids can develop without movement (Golubic and Fischer 1975; Hägele et al. 2006), their sphericity depends on the morphology of the nucleus and movement by waves and currents (Shäfer and Stapf 1978; Brook et al. 2011). The laminated interbedding of oncosparite and stromalite layers (Fig. 7d, e), graded bedding (Fig. 7c), and absence of exposure features suggest deposition in shallow lacustrine environments affected by currents. On the other hand, massive oncoid-rich beds, usually with charophytes, may represent sedimentation in restricted low-energy areas of the lake. Similar facies develop from moderate currents in shallow lacustrine conditions (Schäfer and Stapf 1978). The oncosparites at the top of conglomerate-sandstone channel fill deposits were deposited within unconfined channels and palustrine-flood plain environments, developed in distal alluvial plains. Since the oncoids in the channel fills are mixed with lithoclasts, it has been suggested that oncoids were reworked and formed in inactive channel areas during low-discharge alluvial events (Parcerisa et al. 2006).

The morphology of tubular pores, which partially destroy the lamination, and their pellet fillings (Fig. 8d), are likely burrows made by insect larvae in carbonate fluvial-lacustrine environments (Janssen et al. 1999; Pentecost 2005).

Clastic microbialite-travertine association

This association is transitional between microbialites and travertines along a north–south zone, sharing features with both of them (Fig. 2). It consists of brownish-red tabular beds (0.10–0.5 m thick) that are horizontal and cross-laminated (Table 2, Fig. 9a–c), and stack in sections up to 2.5 m thick (Fig. 9a–c). These contain millimetric to centimetric thick layers of stromatolites, oncoids, and travertine deposits (Fig. 9f), and show intercalations of lenticular beds up to 0.6 m thick with erosive base, covered by clasts and/or laminar crusts (Table 2, Fig. 9b, c). This facies shows grain- to mud-supported granular texture (sensu Gandin and Capezzuoli 2014), ranging from coarse silt- to pebble-sized poorly-sorted components (Table 2, Fig. 9d–f). Original porosity was high (Fig. 9e, f).

Fig. 9
figure 9

modified from Ghannem et al., 2019

Clastic microbialite-travertine association. a Change upward from laminated microbialites (lower part) to the clastic association (log U; see Fig. 7 for symbols). b U outcrop showing inner discontinuity surfaces (D); note the large intraformational clasts (Ic) above the surface crossing the notebook. c Laminated aspect of the facies in outcrop 16, 200 m westward from log U (Fig. 2); a set with cross lamination is arrowed. d Detail of previous figure showing horizontally arranged oncoids. e Clastic microbialite texture with reversed position of stromatolitic hemispheroids (St) in the nucleus of an elongated oncoid. Outcrop 27 in Fig. 2; PPL. f Thrombolite-like texture with poorly preserved microbialite fragments. M section, PPL. Figure 9a, b

Full size image

Interpretation. Clastic facies associated with microbialites structures have been recorded in lacustrine (Arenas et al. 2007), fluvial (Ordóñez and García del Cura 1983; Pedley 1990; Arenas et al. 2010) and thermal springs (Chafetz and Folk 1984; Gandin and Capezzuoli 2014) environments. The common interlayered oncoids may be due to oscillation between allochthonous (movement) and autochthonous (accretion) sedimentation following Pentecost (2005). The intraclastic texture, internal structures and discontinuities, and intense staining by iron oxide components suggest sedimentation in unconfined small streams and shallow ponds through the bench linking the travertine springs to the central lake, dominated by microbialite structures. The composition of the clasts and the palaeogeographic position suggests this connection (Fig. 2). A facies of sand- and gravel-sized fragments of travertine tufas has been described associated with spring-fed small streams that surrounded the playa-lake system of the Eocene Wilkins Peak Member (Smoot 1978).

Travertines

They crop out in the south east border, where the unit reaches the highest altitude at 490 m, near the fault system (OOT, Fig. 1). Travertine deposits are terraced basinward to the northwest with a gradient of 2.6%. In the opposite direction, toward the Ezarga River (Fig. 2), there is terracing of discontinuous travertine outcrops inclined (≈ 3.5–4%) towards the SSE, SE, and ESE, descending to 445 m. These facies form tabular, well laminated bedsets 0.30–1.5 m thick, stacked in sequences up to 3 m thick (Fig. 10a, b). The beds (5–20 cm thick) are often inclined and can show domic morphologies and irregular wavy lamination (Fig. 10b, c). There are internal erosive surfaces covered by lags of travertine clasts (Fig. 10a).

Fig. 10
figure 10

Travertine association. a Travertine section at 18–25 log; see (b) for location and Fig. 7 for other symbols. b Travertine structure with slight inclination of laminae and beds; note the irregularly wavy lamination and the elongated porosity. c Field view of (a); see (b) for location. Note the boundary interlayers (long dashed lines), lamination of the beds and domal morphologies (short dashed lines). d Wrinkled lamination with pointed ridges (arrowed) and fenestral-like porosity (larval galleries?) enlarged by dissolution. Polished section. e Alternations of irregularly wavy dark micrite and light sparite laminae, including micrite-encrusted possible bubbles (arrowed). Outcrop ST; PPL. d and e are like the “flaky puff pastry-like fabric” of microbial affinity (Gandin and Capezzuoli 2014)

Full size image

At meso- and micro-scale, the lamination consists of wrinkled centimicron-thick micrite laminae or alternating micrite-sparite laminae that can be broken (Fig. 10d, e). It is common the intercalation of broken and distorted laminated crusts (≈ 1 centimicron thick) consisting of a thin micrite lamina coated by an alternation of micrite and sparite laminae that have a serrated tracing (Fig. 11a). They resemble the raft structures common in travertines (Gandin and Capezzuoli 2014). This lamination is associated with decimicron to centimicron stromatolitic couplets of dark micrite and light sparite laminae that commonly show wavy to zigzag tracing; in some levels, micrite laminae are faint/diffuse and thin, whereas those of sparite are thicker (Fig. 11b). These fabrics can include circular sections 200 μm in diameter (Fig. 10e).

Fig. 11
figure 11

Travertine association. a Micrite crusts coated with dentated calcite crystals on their lower side. Outcrop ST; PPL. b Alternation of dark (micrite) and white (sparite) laminae with zigzag tracing and variable thickness of the couplets. Outcrop ST; PPL. c White laminae consisting of upright, aligned shrub-like forms (arrowed) that alternate with dark (brown) laminae enriched in iron oxyhydroxides. Outcrop 18–26, polished slab. d, e Shrub-like forms consisting of large bladed sparite crystals embedding decimicron thick alternations of micrite and sparite laminae (L); dark iron oxyhydroxides line pores; see location in (c). Outcrop 18–26; PPL. f Micrite bacterial shrubs (arrowed) surrounded by white columnar crystals of sparite; outcrop ST2; PPL. g Dark micrite remnants of cyanobacterial filaments (Fi) embedded in dirty microsparite (Mc). Blocky calcite mosaic (Bc) fills bioturbation voids coated with a dark layer of iron oxyhydroxides. White arrow points to the top. Outcrop OT; XPL. h Schizothrix-like colony (Sc) and iron oxide-rich micropeloids (Mp) on an erosion surface (dotted line), both embedded in white blocky calcite. White arrow points to the top. Outcrop ST2; PPL. i Single to compound micro-oncoids with laminated coatings surrounded by an outer rim of lighter ferrous calcite. White arrow points to the top. Outcrop R; PPL

Full size image

Some laminae show dendritic (shrub-like forms) and botryoidal dark micrite masses, commonly embedded in sparite (Fig. 11c–f). Filaments (≈ 15 μm wide), arranged in fan-like dispositions (Fig. 11f), and scattered hemispherical Schizothrix-like colonies (600–800 μm wide) have been seen (Fig. 11h). There are layers of single to composite oncoids with complex morphologies (knobby, botryoidal, mammellonar and pustular), around 1 mm in diameter (Fig. 11i).

Interpretation. Although it is not easy to assimilate ancient calcareous spring deposits to water temperature or specific source water, there is some consensus on their adscription to travertines (warm to hot waters) and tufas (environmental waters) based on sedimentological and geochemical features (Ford and Pedley 1996; Pentecost 2005; Jones and Renaut 2010; Gandin and Capezzuoli 2014; among others). Therefore, we use the term travertine due to the textural similarities with fabrics of travertine deposits, such as the presence of raft and bubble structures (Figs. 10e and 11a) and the abundance of coarse crystalline mosaics (cf. Chafetz and Folk 1984; Folk et al. 1985). Under the microscope, the characteristic alternations of thin micrite and thicker sparite laminae with wavy to zigzag tracing (Fig. 11b) resemble the lamination observed in stromatolites, which has been attributed to seasonal microbial doublets (Casanova 1994; Freytet and Plet 1996). The origin of light sparite crystals is varied and subject to debate (Freytet and Verrecchia 1998, 1999; Chafetz and Guidry 1999). The hemispherical arrangements of filaments look like colonies of Schizothrix fasciculata (Freytet and Plet 1996), and the shrub-like and botryoidal forms of micrite masses represent the bacterial shrubs observed in travertines (Chafetz and Guidry 1999). It is worth mentioning that calcite dendrite crystals and abiotic crystalline crusts are poorly distributed in the association (cf. Jones and Renaut 2010; Gandin and Capezzuoli 2014). Bacterial shrubs in travertine systems reflect evaporation and water level fall, whereas the intervening micrite laminae are due to the settling of lime-mud during flooding conditions (Gandin and Capezzuoli 2014). Meanwhile, in the wrinkled alternation of thin micrite and thick sparite laminae (Fig. 11b), latter represent the fill of shelter porosity with crystalline cement. This lamination is similar to the “microbial, flat to curled laminites and puff-pastry-like” microbialite fabrics of thermal spring systems (Gandin and Capezzuoli 2014), where they were linked to the micrite precipitation on bacterial mats on the bottom of puddles; the space between them would be filled later with sparite. The associated circular sections are likely gas bubbles that nucleate on microbial surfaces in low energy travertine environments, commonly associated with floating rafts (cf. Chafetz et al. 1991, Gandin and Capezzuoli 2014). The formation of raft-like platelets (Fig. 11a) occurs at the water surface and requires low energy conditions favored by stagnant water and evaporation (Jones and Renaut 2010; Gandin and Capezzuoli 2014). These conditions have been found in pools behind rimstone dams in recent travertines (Cook and Chafetz 2017).

The presence of erosion surfaces coated by travertine clasts results from diversion and/or intermittence of the currents fed by the springs. The processes that originate these gravelly beds suggest desiccation cracking or synsedimentary tectonic fracturation or karstic alteration during inactivity periods of the travertine (Koban and Schweigert 1993; Gandin and Capezzuoli 2014).

Diagenesis of microbialites and travertines

Original primary fabrics of stromatolites, oncoids, and travertines, such as submillimetric lamination, dark inclusions, bacterial shrubs, and bushes of filaments (≈15 μm wide, > 100 μm long), are commonly “embedded” in leaf-shaped to prismatic sparite crystals perpendicular to the lamination (Figs. 11d, e, g and 12b–e). Under cathodoluminescence, lamination is highlighted by the alternation of bright/dull and dull/non-luminescent colors (Fig. 12f, g), corresponding to the dark micrite/sparite-rich light laminae. Sparite cement filling moldic and interparticle pores is non-luminescent or shows diffuse non-luminescence/dark dull zoning (Fig. 12g); interparticle micrite is dark dull. The presence of labyrinthine pores with corrosive walls on the travertine and microbialite structures is common. Their contour is coated by iron oxyhydroxides and they end up filling up with sparite (Figs. 11d, g and 12a, b).

Fig. 12
figure 12

Diagenetic features. a Bulbous stromatolitic morphologies showing a black laminated external coating. Channel-like pores (arrowed) filled with white sparite and black micropeloids. Outcrop ST2; XPL. b, c Cauliflower-like microbialite with wavy microlamination, embedded in large columnar sparry calcite crystal with undulose extinction (c). Note the external corrosion surface lined by iron oxyhydroxides, followed by a blocky calcite cement. White arrow points to the top. Outcrop R; PPL, XPL. d, e Dirty sparite mosaic ‘superimposed’ on the laminated structure of a travertine fabric. PPL and XPL, respectively. f, g Compound ooid under PPL and CL (cathodoluminescence). Note the intense luminescence of the micritic lamina around the polynucleus (arrow). The sparite cement filling the pores shows dull and non-luminescent bands

Full size image

The preservation of the micrite structures or their remnants (clots, laminae, shrub- and filament-like morphologies), embedded in sparite mosaics, shows that the sparite was posterior. The sparite cement that occupies the space between the filaments (cf. Pentecost 2005; Golubic et al. 2008) could have an early abiotic origin (cf. Chafetz and Guidry 1999). The absence or rarity of microbial evidence could be due to recrystallization (Freytet and Verrecchia 1998, 1999; Golubić et al. 2008) or abiotic precipitation (Casanova 1994; Gandin and Capezzuoli 2014). The dissolution of sparite mosaics and larval burrowing are associated, which evidences the early origin of sparite. Later, a limpid cement of sparite filled porosity (Figs. 11g and 12a).

Discussion

Environments and depositional model

The calcrete-palustrine-laminar crust and microbialite-travertine associations of the unit show an asymmetrical distribution throughout the basin (Figs. 2 and 13). The calcrete and palustrine facies association overlaps the distal alluvial fan deposits along the western strip of the unit, resulting in a retrogradation of the alluvial fan system (Fig. 13). The calcretes developed on the muddy distal areas of the alluvial fans, where their genesis alternated with sedimentary aggradation (cf Cojan 1999.). The palustrine environment developed basinward of the calcrete fringe on periodically flooded shallow areas, where the lacustrine precipitation of muddy carbonate and intermittent exposure took place (cf. Freytet and Plaziat 1982; Alonso-Zarza 2003). Development of laminar crusts linked to exposure features (tepees, spherulites) mark periodic harsh conditions (cf Verrecchia et al. 1995).

Fig. 13
figure 13

Sedimentary model of the Plesitocene filling of the El Gara Basin. Other symbols in Figs. 3, 4, and 7

Full size image

Microbialite facies were deposited in a shallow lacustrine system extending along a S–N strip in the center of the basin, between the calcrete-palustrine environment to the west and the travertine springs and frontal bench to the east (Figs. 2 and 13). The microbialite development was favored by a low terrigenous influx (cf. Gierlowski-Kordesch 2010) and the constant supply of water coming from the alluvial runoff influx, in the NE, and, to a greater extent, from the springs that emerged along the eastern margin of the basin. As a result of this configuration of water sources, oncomicrites were deposited in restricted and central lacustrine areas, surrounded by a fringe of oncosparites and interbedded stromatolites deposited on a low-gradient ramp. Oncoid-bearing deposits are formed in modern shallow lakes oversaturated for calcite (Schäfer and Stapf 1978; Dean and Egglestone 1984; Davaud and Girardclos 2001). In the northeastern and, locally, southwestern areas, the oncoid-bearing layers are associated with conglomerate and sandstone facies. They may represent channel fills and splay deposits in the distal alluvial plains; Paleogene fluvio-lacustrine deposits in Mallorca, Spain, record a similar association (Arenas et al. 2007).

The travertine deposits in the easternmost strip of the carbonate unit lie onto the faulted Cretaceous substrate (Frifita et al. 2016; Hamdi et al. 2019). Apart from the characteristic fabrics analyzed above, the origin from springs is supported by the slight inclination of bedsets toward the basin (i.e. westward, Fig. 10); disposition of outcrops along a diffuse terraced slope, inclined basinward to the NW with a gradient of 1–4% (Figs. 2 and 13); and the close relationship with a system of vertical faults (Figs. 1and 2). Based on geophysical prospecting and hydrogeochemical data from groundwaters, the Central-Western Tunisia region, where the study area is located, is a geothermal province with water temperatures reaching 39 ℃ at a depth of 199 m (Inoubli et al. 2006). Springs nourished the lacustrine central basin (westward), where microbialites developed. Away from the basin (towards the paleovalley of the Ezarga River to the SE), the springs formed sheet streams that flowed down on terraced levels, crowned with travertine constructions (Fig. 2). The slight inclination of the tabular layers has led to consider that they are linked to springs and sheets of water flowing on a slope system (cf. Erthal et al. 2017). Travertines are conditioned by the availability of water and are commonly influenced by tectonically driven groundwater and elevated groundwater table (Capezzuoli et al. 2014). For the study case, the travertine association was most likely linked to spring waters related to a fault area limited by two NNW–SSE faults. There is evidence of one mound-shaped buildup with gentle slopes and about 150 m of diameter in the line of one of the recognized faults. The discharges from this spring line could have fostered westward currents to the eastern margin of the central lake basin, helping to keep a steady lake level.

The lateral change basinward from the travertine to the microbialite association occurs through the strip of the clastic microbialite-travertine association (Fig. 9a, c), deposited along a gentle slope bench between the travertine-depositing springs and the central lacustrine areas with microbialite development (Fig. 13). The presence of internal discontinuities, covered with intraformational travertine-microbialite clasts and thin laminar crusts (Fig. 9), indicates that exposure episodes resulted from phases of spring inactivity. This could be due to drier periods and/or water current diversion from the travertine springs to the basin center.

Paleohydrologic and paleoclimatic reconstruction

The development of travertine and microbialite associations required a constant water supply, while that of calcrete and palustrine carbonates involved more restricted conditions during the deposition of the unit. Therefore, the coexistence and coeval development of microbialite-travertine and calcrete-palustrine facies could depend on the variable balance of water along the basin.

The development of microbialite facies requires relatively stable lake levels since encrusting benthic microbial communities are very sensitive to small differences in paleowater chemistry and hydrological regime (Casanova and Hillaire-Marcel 1992). Stromatolites are indicators of wetter climatic conditions in East Africa (Hillaire-Marcel and Casanova 1987). These constraints may explain the scarcity of exposure features in the microbialite association of the present study. The lamination in stromatolites might reflect a seasonal climatic influence (Casanova 1994), with the spongy light laminae formed in spring–summer and the dark dense one in winter (Golubić and Fischer 1975); in other cases, the lamination represents an alternation of dense summer-autumn laminae and porous winter-spring laminae (Kano et al. 2003). Tufa constructions depends closely on climatic factors and, although it is difficult to generalize, their development increases in humid and warm periods (Pentecost 2005). For travertine, however, a minor influence of climate is required (Pentecost 2005), although fluid circulation through carbonate formations, steady discharge of subterranean waters and precipitation of travertines have been related to climate conditions (Minissale 2004).

Characteristic positive 13C values are attributed to thermogen travertines (Pentecost 2005; Gandin and Capezzuoli 2014). In the unit of study, however, high and positive values have also been found in paleogeographically and genetically different carbonates (i.e. calcrete, palustrine, microbialite, travertine) with 13C values from − 6.4 to 11.7‰ and 18O values around − 5.8‰ for all facies associations (Ghannem et al. 2019). The relatively homogeneous isotopic composition of these carbonates remains their interpretation somewhat conjectural (Ghannem et al. 2019). These high 13CPDB values, especially for the microbialite-travertine carbonates (all positive from 0.42–11.68‰) and microbialites (7.22–11.68‰), suggest that waters from which the carbonates precipitated had the imprint of the Cretaceous substrate (0.77–1.90‰; Ghannem et al. 2019) along with the combined degassing by evaporation and photosynthesis, especially in the central lacustrine areas. Only, weathered samples of microbialites and travertines have negative 13C values (− 3.9, − 7.9), which reflects the contribution of soil-derived CO2. The calcrete-palustrine-laminar crust association shows relatively high values (− 4.34 to + 2.9‰), which reflect an increased contribution of respired CO2. The restricted range of δ18O values around − 5.8 ‰ for all facies associations indicates that carbonate precipitation had the imprint of the paleoprecipitation of Atlantic influence (Ghannem et al. 2019). The presence of cyanobacterial vestiges and nets of insect larvae argues for not too hot waters. Furthermore, the absence of original aragonite precipitation points out that the spring waters were not warm (James and Jones 2016), which is also evidenced by the rapid changes between microbialite and travertine fabrics at all scales.

The above indicators make it clear that climate in the region was wetter than today, even though if only palustrine facies had been developed. Drier short periods in between are evidenced by the presence of discontinuities, somewhere accompanied by laminar crusts, in the clastic microbialite-travertine association. Also, travertines have been revealed as useful indicators of tectonic activity that, together with climate, control the paleohydrology (Hancock et al. 1999). Indeed, they can reflect in surface the mix of meteoric with ascending magmatic, metamorphic, and geothermal fluids, as revealed in Quaternary travertine records of Central Italy (Minissale 2004), where the extensional tectonics favored the hydrothermal activity and travertine precipitation also during the Messinian (Croci et al. 2016). These authors propose climatic oscillations to explain the recharge of aquifers that fed hydrothermal springs and ensuing travertine precipitation.

Although the sedimentation of the unit reflects climatic and tectonic influences, the interpretation of facies reveals a paleohydrological asymmetry. So, the distribution of facies (Figs. 2 and 13) shows that the water supply principally came from fault-controlled springs that deposited travertines in the southeast margin. This implied a progressive loss of water to the west and evaporation in the center of the basin, where microbialites developed.

The carbonate unit is loosely dated as Middle-Upper Pleistocene (Burollet and Sainfeled 1956; Ben Haj Ali et al. 1985) and, therefore, there is not enough precision to establish a correlation with climate changes in the region and/or at world scale. It is difficult, therefore, to attribute the observed environmental change in the succession to a time of the Pleistocene. Fontes and Gassé (1989) suggest for the northern Sahara that the last major Late Pleistocene humid period ranges from 80,000–150,000 yr BP (Table 3). The study on shells of Great Chotts in the southern of Tunisia deduced two late Pleistocene highstands corresponding with humid phases at about 90 and 150 ka (Causse et al. 1989). A later study of U/Th dating of lacustrine materials (shells) in central Tunisia recognized 4 distinct flood episodes of lacustrine deposits occurred at about 30, 95–100, 130–150 and 180–200 ka, highlighting the existence of large lakes in Northern Africa around 90–100 and 130–150 ka and evidenced a major lacustrine phase around 200 ka (Causse et al. 2003). In addition, two superposed sequences in raised marine deposits in southeastern Tunisia deliver ages between 147 and 110 ka and 141 and 100 ka that are assigned to the Marine Isotopic Substage 5e (MISs 5e, Last Interglacial) (Jédoui et al. 2003). Recently, travertine deposits in central Tunisia, the first documented in this region, have been linked to fault fissures and hydrothermal fluids. These travertines developed during a humid episode, not well constrained in time, during Early to Middle Pleistocene (Henchiri et al. 2017).

Table 3 Pleistocene humid records in North Africa
Full size table

A contemporaneous humid phase found in southern Egypt, Libya, and Tunisia is Th/U dated, respectively, at 141,000 ± 7000 (Szabo et al. 1989), 133,000 ± 10,000 (Gaven et al. 1981), and between 147,000 ± 15,000 and 134,000 ± 12,000 yr BP (Causse et al. 1989) (Table 3). In Egypt, the two younger episodes of travertine deposition occurred at near 100 and 200 ka and are correlated with interglacial conditions synchronized with pluvial episodes that recharged the aquifers (Crombie et al. 1997). In this southeast region of Egypt, more recent studies, based on tufa constructions and freshwater gastropods, also support the existence of a humid period of spring flow during the Marine Isotope Stage 6/5e transition. The δ18O low values of these components suggest that most of the Pleistocene pluvial precipitation came from an Atlantic source (Smith et al. 2004, 2007). In the nearby region of Libya, four lacustrine carbonate units were deposited during humid periods and correlated with the 100-eccentricity cycle of orbital forcing during the last 500 ka (at > 420, 380–290, 260–205 and at least 140–125 ka) (Geyh and Thiedig 2008). These authors match the humid episodes in North Africa with Middle Pleistocene interglacial phases in the temperate zone. In contrast, there are models based on field, isotopic, and geochronologic evidences in favor of glacial periods being wet and cold while interglacials were dry and warm (Abouelmagd et al. 2012, and references therein). These authors attribute earlier wet climatic periods over North Africa to the reinforcement of palaeowesterlies in glacial periods. Therefore, the recharge of the aquifers that maintained the development of lake carbonates, tufas and travertines in North Africa during the Middle and Upper Pleistocene may not be unequivocally linked to glacial or inter-glacial phases. Regardless of whether it was an interglacial or glacial phase, the shift to wetter conditions paralleled to the microbialite-travertine development in the study area. The abundance of examples around 125 ka, such as those described above along North Africa, would indicate a possible correlation with the warmer interglacial optimum (5e), close to the transition from Middle to Upper Pleistocene (Table 3).

Conclusions

Based on the stratigraphic and sedimentologic discussion above, the following conclusions can be drawn regarding the Pleistocene carbonate deposits from NW Tunisia:

  1. 1)

    The BEU consists of calcrete-palustrine-laminar crust and the composite microbialite-travertine associations, developed in the western and central-eastern strips of the unit, respectively.

  2. 2)

    The calcrete-palustrine-laminar crust association extends along the western and northern margins fringing the distal alluvial facies. A calcrete belt evolved in the distal alluvial plains passing upward and distally to carbonate palustrine environments under a climate with seasonal drought.

  3. 3)

    The microbialite association consists of massive and laminated oncoids with intercalated stromatolites, developed in the central lacustrine areas of the basin. They form the bridge between the palustrine facies to the west and the travertine facies to the east.

  4. 4)

    Travertines, cropping along eastern margin, pass laterally to the lacustrine microbialites toward the basin center through the clastic association (clastic microbialite-travertine). They developed from ambient temperature spring waters emerging along faults that nourished the central lacustrine environments. The clastic microbialite-travertine association developed on a bench that connected the travertine springs with the microbialite central lacustrine settings.

  5. 5)

    The coexistence of calcrete-palustrine and microbialite-travertine associations and their distribution through the basin reflect a hydrological asymmetry in the basin: periodic and scarce supply of water from the west margin and a constant water supply from the eastern springs that favored the microbialite development in the central lake.

  6. 6)

    The development of microbialites and travertines in the basin reflects wetter climatic conditions than today during the Middle/Upper Pleistocene in northern Tunisia.