Introduction

Collisional orogens and associated sedimentary basins are of great scientific and economic significance because they track information about the collisional processes, refining the geodynamic models, and are related to economically viable resources (Ding et al. 2005; Weislogel et al. 2006; Najman et al. 2010; Maravelis et al. 2012; Nance et al. 2014; Hu et al. 2015; Bega 2015; Velić et al. 2015; Critelli 2018; Tserolas et al. 2019; Critelli and Martín-Martín 2022). Foreland basins are developed in each side of the orogen (in the pro- and retro side) and result from lithosphere flexure, caused by the enormous mass produced by crustal thickening associated with mountain belt development (Beaumont 1981; Sinclair and Allen 1992; Garzanti et al. 2007; Catuneanu 2018; Critelli and Criniti 2021). Further, local factors like basement tectonics and relative eustacy of sea level have an influence on subsidence (Catuneanu 2018). The sedimentary successions in foreland basins provide insights about the growing and development of orogens and adjacent depocenters (White et al. 2002; Garzanti et al. 2004a, 2004b; Garzanti et al. 2005). These attributes can be further evaluated by using sedimentary and sequence stratigraphic data (Cantalamessa and Di Celma 2004; Catuneanu 2004; Martins-Neto and Catuneanu 2010; Maravelis et al. 2018; Melehan et al. 2021), geochemical (Verma and Armstrong-Altrin 2013; Zaid and Gahtani 2015; Tawfik et al. 2017; Maravelis et al. 2021) and structural data (Ryan and Williams 2007; Maravelis et al. 2017). Processes, such as orogenic unroofing, recycling and erosion can be also investigated by studying the provenance of the detritus that comprise the sedimentary fill in foreland basins. The sedimentary provenance in the two foreland basins (pro- and retro foreland basin) reflects the orogenic evolutionary stages, with the pro foreland displaying an up-section increase in lithic rock fragments, and the retro foreland exhibiting an up-section increase in metamorphic rock fragments (Dorsey 1988; Chen et al. 2001; Kirstein et al. 2009; Nagel et al. 2014; Critelli and Criniti 2021).

Reconstructions of tectonic and paleogeographic history are benefited by provenance investigations that consider the petrographic and geochemical composition of clastic deposits. Petroleum and geochemical studies have identified fundamental elements influencing sediment composition in depositional systems such as climate, drainage system, source-rock composition, and basin geometry (Johnsson 1993; Critelli et al. 2003; Garzanti et al. 2007). Petrographic and geochemistry analyses are effective tools in the study of basin provenance and tectonic setting because sediment composition is directly impacted by transportation processes, source rock composition, tectonic and paleoclimatic conditions of the source area (Zimmermann and Spalletti 2009; Ghazi and Mountney 2011; Garzanti et al. 2013; Maravelis et al. 2015; Fathy et al. 2018; Iqbal et al. 2019). Source rocks for sedimentary successions are identified based on the relative immobility of elements under various tectonic settings, such that their concentration reflects their original composition. Because these components (trace and rare earth elements) are thought to be immobile and inactive while in transit, they are suitable proxies for provenance research (Von-Eynatten et al. 2003).

The examined succession consists of Upper Eocene to Lower Oligocene sub-marine fan deposits. These sedimentary rocks deposited in the central Pindos Foreland Basin (PFB, Fig. 1) are described primarily in terms of sedimentology-stratigraphy that includes lithofacies and depositional environment (Piper et al. 1978; Pavlopoulos 1983; Avramidis and Zelilidis 2001; Botziolis et al. 2021; Kovani et al. 2023; and references therein), structural deformation (Faupl et al. 1998; Kamberis et al. 2000; Sotiropoulos et al. 2003; Piper 2006; Konstantopoulos and Zelilidis 2012; and references therein) and oil-gas potential (Karakitsios and Rigakis 1996; Zelilidis et al. 2003; Zelilidis and Maravelis 2015; and references therein). This work presents an integration of geochemical and petrographic results that further constrain the origin and tectonic setting of these deposits, emphasizing to sedimentary processes, such as weathering, sorting, and recycling that affected the deposition of the central PFB strata. This research elaborates provenance data to link sediment composition to particular stratigraphic units of Pindos orogen, and to highlight the impact of orogenic growth on the sedimentation of Pindos foreland basin.

Fig. 1
figure 1

Geological map of Western Greece (modified from Zelilidis et al. 2015), illustrating the major thrusts and strike-slip fault zones. The red box shows the study area and the profile at the A-A’-A” cross section refers to the schematic diagram of Fig. 19

Full size image

Geological setting

The complicated geotectonic evolution of western Greece was caused by changes in the type and relative movements of the main tectonic plates. Pindos Orogen (Fig. 1) is a thrust system that separates external from internal Hellenides and is the result of the collision of Apulian and Pelagonian continental blocks, after the closure of Pindos Ocean (De Graciansky et al. 1989; Doutsos et al. 1993, 2006; Karakitsios 2013; Zelilidis et al. 2015).

The sedimentological-stratigraphic and structural assessments of the external Hellinides indicate different evolutionary phases that can be broadly defined as an early extensional and a latter collisional phase (Fig. 2) (Karakitsios 1990, 1995, 2013; Zelilidis et al. 2015; Bourli et al. 2019a, 2019b; Zoumpouli et al. 2022). The early phase is characterized by deposition of carbonates in an extensional setting (Ionian Basin), whereas the sedimentation in the latter phase is dominated by clastics in a foreland basin (PFB). The early pre-rift stage that predates tectonic processes is testified by Lower to Middle Triassic evaporites and Upper Triassic to Lower Jurassic shallow marine limestones (Fig. 2). This stage is followed by the syn-rift stage, during which the sedimentation is represented by late Lower Jurassic to Lower Cretaceous carbonates (with a few rare mudstone successions) with a general deepening upwards sedimentation (Fig. 2). During this stage, syn-sedimentary extensional tectonics develop half-grabens and result in notable thickness fluctuations of the syn-rift deposits (Zelilidis et al. 2015; Bourli et al. 2019a, 2019b). The sedimentation in the final syn-rift stage includes carbonates from the Middle Cretaceous to Lower Eocene (Bourli et al. 2019a, 2019b; Fig. 2).

Fig. 2
figure 2

Lithostratigraphic column of the Ionian zone (modified after Bourli et al. 2019a and Zoumpouli et al. 2022)

Full size image

In western Greece, the Middle Eocene marks the transition from crustal stretching to crustal shortening (Lutetian to Bartonian, Zoumpouli et al. 2022). The Pindos orogen developed because of this plate motion, and it is defined as an elevated plug with both a pro- and a retro-wedge domain in the idea of a doubly vergent thrust wedge (Doutsos et al. 2006). The Pindos orogen suture zone is characterized by ophiolitic rocks (remnants of the former oceanic crust) in the orogen central sections (Robertson 2004). Progradation occurred over Triassic evaporites representing the preferential décollement zone, as evidenced by their proximity to thrusts (Underhill 1988; Karakitsios 1995; Kamberis et al. 1996). The Mesozoic to Eocene carbonate thrust tip anticlines were formed by many intrabasinal thrusts that impacted the basin in the Late Oligocene and Early Miocene (Jenkins 1972; Clews 1989).

The western section of the external Hellenides underwent compression from the Middle Eocene to the Early Miocene (Fig. 2), whereas the eastern section is dominated by extensional tectonics (Aubouin 1959; Jacobshagen 1986; Pavlides et al. 1995; Doutsos and Koukouvelas 1998; Doutsos et al. 2000). Throughout the Late Eocene to Early Oligocene, deep-sea fan sediments were deposited westward in the PFB as a result of the tectonic uplift and erosion of the Pindos orogen (Koch and Nicolaus 1969; Skourlis and Doutsos 2003; Sotiropoulos et al. 2003; Botziolis et al. 2021), evolving up-section to deltaic and alluvial deposits (Fig. 2) (Avramidis and Zelilidis 2001; Piper 2006; Piper et al. 1978; Konstantopoulos and Zelilidis 2012; Botziolis et al. 2021).

The PFB trends parallel to the Pindos orogen and is bounded by the Pindos and the Ionian thrust to the east and the west, respectively (Fig. 1) (Aubouin 1959). The activity of the Pindos thrust was more important during the evolution of the PFB, compared to the Gavrovo, internal Ionian and middle Ionian thrusts (Avramidis et al. 2000). Additionally, and despite the absence of precise chronologic information, structural features, such as thrust-related fault bend and fault progradation folds, as well as the absence of growth strata, suggest that these thrusts postdate the sediment deposition (Zygouri et al. 2021). PFB has also been affected by strike-slip faults (King et al. 1993; Avramidis et al. 2000; Avramidis and Zelilidis 2001; Konstantopoulos et al. 2013). The Pindos thrust is cut by strike-slip faults that operated independently in several locations and epochs, namely in Ioannina in the Early Eocene, Arta in the Late Eocene, and Mesolongi in the Early Oligocene (Fig. 1) (Zelilidis et al. 2008). The strike-slip faults acted as pathways, affected the paleocurrent system, causing sediment to discharge at far-off locations (Zelilidis et al. 2008).

Stratigraphy of Pindos foreland basin

PFB is thought to have formed as a foredeep during the Late Eocene to Early Oligocene, receiving sediments from the rising Pindos Orogen. The investigated succession consists of ten facies and sub-facies associations and thirteen sedimentary facies, according to Botziolis et al. (2021). Central PFB deposits can be divided into three distinct depositional environments (Botziolis et al. 2021; Figs. 3, 4, 5). The study area consists of a submarine fan system overlying Eocene carbonates (Fig. 5a). There is a general trend from west to east from the carbonates, through the abyssal plain (Fig. 5b), to the outer (Fig. 5c), and inner (Fig. 5d) fan deposits, testifying system progradation and temporal shallowing of the PFB (Botziolis et al. 2021). Between the enclosing topography of the levees, conglomeratic channel-belt facies consisting of limestone, chert, sandstone, and shale clasts may be discovered. The sediments were deposited within the Pindos foredeep and belong to the system underfilled stage. The examination of the sediments reveals deposition during the onset of the Pindos orogen, when sedimentation was constrained by the accommodation space provided by lithospheric flexure (Botziolis et al. 2021).

Fig. 3
figure 3

Geological map of the studied area depicting the analysed outcrops as well as the spatial distribution of the various environments and sub-environments of deposition (modified by Botziolis et al. 2021)

Full size image
Fig. 4
figure 4

Representative stratigraphic log of the studied deposits (modified by Botziolis et al. 2021) displaying the relevant analysed samples for geochemical and petrographic study. The outer fan deposits (with a general upward increase in the thickness and abundance of sandstone beds) overlying abyssal plain deposits and in turn overlain by inner fan deposits (with a general upward decrease in the thickness and abundance of sandstone beds). Take note of the consistent paleoflow direction to the northwest. The purple arrow points the samples with the highest V concentration, the green arrow points the samples with the highest Cr concentration and the blue arrow points the samples with the highest Ni concentrations

Full size image
Fig. 5
figure 5

Photographs of outcrops depicting the stratigraphic evolution of the southern PFB. The outer fan deposits are overlain by abyssal plain deposits, which are overlain by inner fan deposits

Full size image

Data from sole markings show two distinct directional flows. Across the stratigraphic strata of the examined sections, the NE-SW orientation of the measurements suggests that the axial flow pattern predominated during sediment deposition (Fig. 4) (Botziolis et al. 2021). Additionally, a flow-spread tendency has been linked to all perpendicular flows and associated with (distal) lobe-fringe or levee deposits. The relationship between space accommodation and sediment transport volume on a subsiding tectonic setting is shown to be essential in shaping subsurface architecture and the subsequent stacking pattern. These controls can be detected by the system geometry and architecture. PFB lobe complexes were developed in an unconfined environment and are distinguished by a compensating stacking pattern, resulting in widespread deposits (Botziolis et al. 2021). This is because of the availability of sufficient space for compensation and a sediment supply insufficient to exceed the accommodation space.

Analytical methods

Prior to sampling, extensive fieldwork was carried out to identify the study area’s sedimentary processes, depositional conditions, and stratigraphic development to ensure that the samples cover the entire sedimentary succession (Figs. 3, 4). For petrographic investigation, 35 samples of very fine- to coarse-grained sandstone were collected (Figs. 3, 4). Thin-sections were cut perpendicularly to the structureless or parallel-laminated sandstone bedding and were petrographically analysed using a B-810 Series Optika Italy polarizing microscope. The detrital assemblages in the samples were identified using the Gazzi-Dickinson point-counting approach by Dickinson (1970).

From classifications derived by just using a petrographic microscope, the origin of quartz, the main constituent of most sandstones, still cannot be ascertained (Götze and Zimmerle 2000). Thus, stricter limitations have been imposed using technically improved cathodoluminescence. Six quartz types can be distinguished, which serve as a guide to provenance. By distinguishing between various feldspar and rock fragment types, cathodoluminescence petrography (CL) may also offer quantitative estimates of their abundance. Moreover, CL enables the estimation of the original grain size and roundness characteristics and enhances the separation of detrital grains from syntaxial overgrowths in well-cemented sandstones. In this work, a Canon Powershot A630 digital camera was used in conjunction with a Reliotron III Cathodoluminescence device linked to a Leitz Wetzlar Orthoplan Microscope to investigate the CL (Liritzis et al. 2019). This CL configuration enables observation of low luminous minerals on exposed thin sections as well as viewing of a reasonably broad area of the sample under cathodoluminescent, plane-polarized, and cross-polarized light (Sippel 1965, 1968; Zinkernagel 1978). In this technique, the thin sections are blasted with electrons using an electron gun that operates at an accelerating voltage of 10 kV and a current of 0.200 mA at an operating vacuum of 100 mTorr, resulting in an electromagnetic phenomenon that may have visible spectrum wavelengths.

Additionally, 35 samples of mudstone were collected (Figs. 3, 4) for geochemical analysis at Bureau Veritas Commodities Canada Ltd. (formerly ACME Analytical Laboratories Ltd., Canada). Inductively coupled plasma optical emission spectrometry (ICP-OES), which has a high degree of accuracy and a low detection limit, was used to analyse major elements, while inductively coupled plasma mass spectrometry (ICP-MS) was used to analyse trace and rare earth elements (REE). To assure appropriateness for comparison with contemporary, well-known tectonic environments, geochemical investigations of sedimentary rocks are recalculated dry (Rollinson 1993). The loss on ignition (LOI) was calculated, and the contents of the principal elements were utilized in several plots after recalculating to an anhydrous (LOI-free) basis and adjusting to 100% for statistical coherence. Major oxide ratios and discriminant functions were calculated to distinguish tectonic setting. Dickinson et al. (1983) ternary diagrams and Verma and Armstrong-Altrin (2013) discriminant-function-based multi-dimensional diagrams for sediments with high and low silica concentrations, respectively, were employed to establish the tectonic setting of the sediments.

Sandstone petrographic analysis

The samples are very fine- to coarse-grained sandstones, poorly- to very well-sorted, with angular to sub-rounded grains (Figs. 6, 7). According to Garzanti (2016), the majority of samples are feldspatho-quartzo-lithic to feldspatho-litho-quartzose (Fig. 8).

Fig. 6
figure 6

Representative photomicrographs illustrating the petrographic features of the sandstones from the Upper Eocene-Lower Oligocene PFB deposits. a Presence of sandstone fragment, monocrystalline quartz (Qm) and tectonic quartz (Qt) displaying undulose extinction; b schist fragment, along with plagioclase (P) showing repeated twinning; c chert and polycrystalline quarts (Qp), d tectonic quartz (Qt) displaying fracturing; e detrital quartz and limestones with oolite fragments; f detrital oolite and foraminifera fragments in calcite matrix. All are taken under crossed nicols. Abbreviations: Qp: polycrystalline quartz; Qm: monocrystalline quartz; Qt: tectonic quartz; Qz: quartz in matrix; P: plagioclase; K: K-feldspar; C: chert; B: Biotite; M: muscovite; Bf: benthic foraminifera; Pf: planktonic foraminifera; Ol: oolite and Lc: lithic calcite

Full size image
Fig. 7
figure 7

CL photomicrographs (a-d) illustrating the petrographic features of the sandstones from the Upper Eocene-Lower Oligocene PFB deposits, under cathodoluminescence. Abbreviations: Q: quartz; P: plagioclase; K: K-feldspar; Lc: lithic calcite

Full size image
Fig. 8
figure 8

Ternary diagrams (after Garzanti 2016) showing provenance fields defined by total quartz, feldspar and unstable (nonquartzose) lithic fragments. Abbreviations: Q = quartzose; F = feldspathic; L = lithic; lFQ = litho-feldspatho-quartzose; lQF = litho-quartzo-feldspathic; qLF = quartzo-litho-feldspathic; qFL = quartzo-feldspatho-lithic; fQL = feldspatho-quartzo-lithic; fLQ = feldspatho-litho-quartzose

Full size image

Monocrystalline quartz in the samples range from 21.67 to 37.67% (mean value: 28.22%), whereas lithic fragments range from 2.67 to 12.67% (mean value: 8.77%) (Table 1). Quartz frequently exhibits undulose extinction, whereas polycrystalline quartz, feldspar, biotite, and muscovite are less common (Figs. 6, 7). Tortosa et al. (1991) observed a distinct pattern in sediment composition. They found that sediments originating from granitic and gneissic formations tend to have Qp grains with a limited number of crystals, typically five or fewer. On the other hand, sediments sourced from low-rank metamorphic environments exhibit a prevalence of polycrystalline quartz grains, characterized by more than five fine-very fine crystals. This discrimination based on grain characteristics provides valuable insights into the provenance and geological history of these sediments and for this reason. While the concentration of tectonic quartz (Qt) (including the polycrystalline quartzs with more than 5 splits and undulosed quartz) ranges from 15.33 to 37.67% (mean value: 28.41%), that of polycrystalline quartz without tectonic fabric (Qp) (less than 5 splits) ranges from 2.67 to 6.00% (mean value: 2.91%). Only a few of the quartz grains are well-rounded, with the most being sub-rounded to sub-angular (Figs. 6, 7). All samples contain higher content of plagioclase (ranges between 2.67 and 14.33%, with a mean value: 11.50%), compared to K-feldspar (3.33 to 8.33%, mean value: 5.60%). Many plagioclase crystals exhibit multiple albite twinning (Fig. 6). The content of chlorite and mica (shared by muscovite and biotite), varies from 0.33 to 8.00% (mean value: 4.60%). Mica crystals are mostly elongated (Fig. 6). Rare chlorite and composite grains made of quartz, K-feldspar, plagioclase, and mica are also present. Lithic clasts include fragments of sedimentary rocks such as chert, sandstone, shale, and limestone as well as low-grade metamorphic rocks such as slate and quartz-schist (Fig. 6). Sandstone fragments are mostly sub-angular to sub-rounded and range from 2.34 to 11.00% (mean value: 7.62%) (Fig. 6). Chert fragments (Fig. 6) are mostly sub-angular to sub-rounded, but rounded grains also occur and their abundance ranges from 2.67 to 18.00% (mean value: 9.05%). Slate and schist fragments are mostly sub-angular and vary between 0.33 and 1.67% (mean value: 1.05%) (Fig. 6).

Table 1 Point-counting data (% percentage) for the PFB system. Three hundred points were counted per thin-section by utilizing the Gazzi-Dickinson method
Full size table

Cathodoluminescence can be used to investigate the differentiation and origin of sandstone detrital grains more effectively in sedimentary rocks (Götze and Zimmerle 2000). High-temperature crystallization and rapid cooling produce red or bright blue luminescing quartz, which is frequently found in volcanic rocks or rocks that have experienced contact metamorphism (Boggs et al., 2002). Slower cooling and lower crystallization temperatures cause the CL signal in quartz to be less intense, turning the grain from typically bright to dark blue. This kind is common in plutonic rocks because they cool more slowly than volcanic rocks (Boggs et al., 2002). According to Zinkernagel (1978), non-luminous quartz is formed diagenetically or crystallizes below around 300 °C. The authigenic quartz CL colours range from light blue to green to reddish brown (Ramseyer et al. 1988; Neuser et al. 1989). Quartz grains may lose their CL colour and often have a brown colour due to local metamorphism and acquire a blue hue by high-grade metamorphism (Boggs et al., 2002). In the studied samples, quartz crystals with red and brown tones were observed. Optical classification of the quartz crystals revealed that 22% of them were brown quartz and 78% were red quartz (Fig. 7). Additionally, plagioclase crystals emit a greenish yellow colour, feldspar crystals emit a vivid blue colour and carbonate minerals an orange-red colour (Fig. 7).

Mudstone geochemical analysis

Major elements abundances

The major elements abundances document that the studied sediments have slightly lower SiO2, Al2O3, K2O, TiO2, P2O5 contents (on average 59.56 wt.%, 14.70 wt.%, 2.68 wt.%, 0.77 wt.% and 0.12 wt.%, respectively) relative to the Post-Archaean Australian Shale (PAAS, 62.80 wt.%, 18.90 wt.%, 3.70 wt.%, 1.00 wt.% and 0.16 wt.%, respectively, McLennan et al. 1993) (Table 2). On the other hand, the average abundances of MgO, CaO and Na2O (4.08 wt.%, 9.18 wt.% and 1.50 wt.%, respectively) are enriched relative to PAAS (2.20 wt.%, 1.30 wt.% and 1.20 wt.%, respectively, McLennan et al. 1993) (Table 2). The mean contents of Fe2O3 and MnO are 7.01 wt.% and 0.09 wt.%, respectively, similar to PAAS (7.10 wt.% and 0.11 wt.%, respectively, McLennan et al. 1993) (Table 2). LOI values range from 7.6 to 17.9 wt.% (Table 2).

Table 2 Major elements (in wt.%) after LOI correction, trace elements, rare earth elements (REE) (in ppm) and elemental ratios for the PFB system
Full size table

The relationship between minerals and the distribution of major elements can be documented by the Pearson correlation coefficient variations of selected major elements (SiO2, TiO2, and K2O) versus Al2O3 (Bauluz et al. 2000). Al2O3 was selected because Al is immobile throughout weathering and diagenesis (Bauluz et al. 2000). In the PFB samples, SiO2 displays a weak negative linear correlation with Al2O3 (r = −0.17) (Fig. 9). In contrast, K2O, TiO2, and Al2O3 demonstrate a strong positive linear correlation (r = 0.95 and 0.95, respectively) (Fig. 9). All samples display K2O/Al2O3 ratios below 0.21.

Fig. 9
figure 9

Selected major and trace element Pearson’s coefficient correlation variations presented in table for the samples from the southwestern PFB. Major element data are calculated on an anhydrous normalized basis

Full size image

Trace element abundances

The trace element values were normalized against the PAAS to evaluate the degree of enrichment or depletion and to determine the source of the sediments (Fig. 10). Certain Large Ion Lithosphere Elements (LILEs) like Rb, Cs, Ba, and Sr, some High Field Strength Elements (HFSEs) like Nb, Zr, Th, and U, and other trace elements like Ga and Ta have lower concentrations when compared to PAAS (Table 2). Other trace elements, such as Cr, Ni, Cu, Pb and Zn are enriched compared to PAAS (Table 2). Elements like Sr (LILE), Sc, Hf (HFSEs), and a few other trace elements (Co, V, and Y) are comparable to those of PAAS (Table 2).

Fig. 10
figure 10

Post-Archaean Australian Shale (PAAS)-normalized (McLennan et al. 1993) multi-element diagram for the trace element concentrations of PFB

Full size image

Pearson correlation coefficient variations of some trace elements (Ba, Rb, Th, V and Ni) against Al2O3 and Zr against SiO2 document the relationship between the minerals and the trace elements distribution (McLennan et al. 1993; Chen et al. 2014; Ali et al. 2014). In the PFB, Ba displays moderate positive linear correlation with Al2O3 (r = 0.76) (Fig. 9). Rb, Th and V demonstrate a strong positive correlation with Al2O3 (r = 0.96, 0.93 and 0.99, respectively) (Fig. 9). Ni exhibits a weak negative linear correlation with Al2O3 (r = −0.09) while, Zr displays a moderate positive linear correlation with SiO2 (r = 0.65) (Fig. 9).

Rare earth elements (REE)

The samples average total REE content is 120.82 ppm, with a range of 90.86 to 168.42 ppm (Table 2). Light REE (LREE, La, Ce, Pr, Nd, Sm, and Eu) to heavy REE (HREE, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, and Y) ratio is high (7.34 on average) (Table 2). The average (La/Yb)N and (La/Sm)N ratios of 7.41 and 3.46, respectively, where subscript N refers to chondrite-normalized values, support the moderate enrichment in light REE (LREE) patterns. The samples also exhibited relatively flat heavy REE (HREE) patterns (Fig. 11), confirmed by the (Gd/Yb)N = 1.50 ratio. N-MORB normalised patterns for the PFB samples reveal high contents of Rb, Th and Ba, along with Nb, Sr and Ti depletion (Fig. 12). The samples display a negative Eu anomaly [(Eu/Eu* = (Eu)N/((Sm)N × (Gd)N)1/2)] (Eu/Eu* = 0.71, average), and any Ce anomaly [(Ce/Ce* = (Ce)N/((Pr)N^2 / (Nd)N))] (Ce/Ce* = 0.88, average) is lacking (Table 2).

Fig. 11
figure 11

Chondrite-normalized rare earth element patterns. Chondrite normalization values are from Taylor and McLennan (1985). REE patterns of Post-Archean Australian Shale (PAAS) and Upper Continental Crust (UCC) are included for comparison

Full size image
Fig. 12
figure 12

Mid-ocean ridge basalt (MORB)-normalised spider diagrams for the PFB samples. Normalising values after Pearce (1982)

Full size image

Discussion

The type of source rock, the rate of sediment supply, the degree of sorting during transit and deposition, and the degree of weathering all impact the petrographical and chemical composition of clastic deposits (McLennan 1989; Cox et al. 1995; Roddaz et al. 2006). Therefore, before making inferences about the provenance of sediments and the regional tectonic setting, each of these characteristics must be evaluated.

Source area weathering

SiO2 and Al2O3 in the PFB samples exhibit a moderate negative correlation, indicating that quartz and aluminous clays during deposition hydrodynamically separate (Purevjav and Roser 2013). The strong positive correlation between K2O and Al2O3 and the low K2O/Al2O3 ratios (less than 0.3) suggests that K is found in phyllosilicates or clay minerals (Cox et al. 1995). Ba was absorbed into phyllosilicate minerals in addition to K-feldspar, as shown by the exceptionally strong positive correlation of Ba with Al2O3 and K2O (Fig. 9) (McLennan et al. 1993). According to Fig. 9, the moderate positive correlation between Rb and Al2O3 indicates that the Rb is mostly present in phyllosilicate minerals (McLennan et al. 1993). Th is concentrated in clay minerals rather than accessory minerals, according to the strong positive correlation of Th with Al2O3 (Fig. 9) (Armstrong-Altrin et al. 2015; Etemad-Saeed et al. 2015; Amendola et al. 2016). The weak positive correlation of Zr with SiO2 (Fig. 9) suggests that Zr is present in the rock silicate component. The strong positive correlation between Al2O3 and V (Fig. 9) suggests that the V content increases with increasing Al2O3 content. This correlation can be attributed to the association of V with Al-rich phases, such as clay minerals. Because Al2O3 and Ni have a weak negative correlation, the concentration of Ni in heavy minerals or relict mafic minerals is most likely what controls how the element is distributed (Fig. 9).

The PFB mudstones of the foredeep depozone compared to the PAAS mudstones are depleted in SiO2, Hf and Zr and enriched in Cr and Ni (Table 2, Fig. 10). Quartz dilution caused by the loss of unstable phases and relative zircon concentrations is the most likely cause (Roddaz et al. 2006). In contrast to PAAS, Na2O, CaO, MgO, and Sr are more enriched than K2O, Rb, and Ba. Plagioclase, a rapidly weathering mineral with more Na, Ca, and Sr than minerals containing Rb, Ba, and K, is responsible for this trend (White and Brantley 1995). Furthermore, strong positive correlations have been observed between Rb and Ba with K2O (Fig. 9), indicating that the distribution of these elements might be influenced by the presence of alkali-feldspar (Armstrong-Altrin et al. 2014). Another explanation is the immobilization of smaller cations such as Na, Ca, and Sr rather than bigger cations such as K, Cs, Rb, and Ba because of adsorption and absorption on clay mineral surfaces (Nesbitt et al. 1980; Wronkiewicz and Condie 1990; Bauluz et al. 2000; Roddaz et al. 2006).

The Chemical Index of Alteration (CIA index) is a widely used method for estimating the degree of weathering at the source location (Nesbitt and Young 1982; Fedo et al. 1995, 1996; Bauluz et al. 2000; Lee 2002; Hofmann et al. 2003). The formula CIA = [Al2O3 / (Al2O3 + CaO*, Na2O, and K2O)] * 100 is used to calculate the amount of feldspar weathering in comparison to unaltered protoliths (molar proportions). CaO* stands for the concentration of the silicate component. Because we do not have CO2 contents for the samples, we are unable to account for Ca in carbonates. As a result, CaO values are accepted if the mole fraction of CaO is equal to or less than Na2O and the moles of CaO are equal to Na2O if CaO values are greater than Na2O (McLennan et al. 1993; Bock et al. 1998; Jian et al., 2013). According to Garzanti and Resentini (2015), no matter what correction method is used, CIA control remains resilient. In Fig. 13a, the CIA values are displayed. Unmodified upper crustal rocks and unaltered plagioclase and K-feldspar have CIA values that are nearly equal to 50. Higher CIA values imply increase in weathering.

Fig. 13
figure 13

a Chemical Index of Alteration (CIA) against Index of Compositional Variability (ICV) diagram for the sedimentary rocks in the central PFB (after Nesbitt and Young 1984; Cox et al. 1995). PAAS composition taken from Taylor and McLennan (1985). b CIA and ternary A-CN-K diagram (after Nesbitt and Young 1982; McLennan et al. 1993) for the sedimentary rocks of PFB. PAAS (Taylor and McLennan 1995) is plotted for comparison. Average tonalite (T), granodiorite (Gd) and granite (G) compositions shown are taken from Condie (1993). Predicted weathering trends (arrows) of T, Gd and G are also shown. c 15Al2O3-Zr-300*TiO2 ternary diagram (Garcia et al. 1991)

Full size image

The compositional maturity of mudstones is assessed using the Index of Compositional Variability (ICV) (Cox et al. 1995). It can be utilized to determine if the rocks in the provenance experienced sediment recycling (Cullers and Podkovyrov 2000). The ICV formula uses the following chemical ratios: ICV=(Fe2O3+K2O+Na2O+CaO+MgO+MnO+TiO2)/A12O3. Compositionally immature mudstones are typically deposited in tectonically active settings and have ICV values higher than one (1) (Cullers and Podkovyrov 2000) due to a large percentage of non-clay silicate minerals (Van de Kamp and Leake 1985). By contrast, compositionally mature mudstones generated in a passive margin (Weaver, 1989) have ICV values smaller than one (1) (Cullers and Podkovyrov 2000). Figure 13a depicts the ICV values. The samples ICV and CIA values (mean 1.73 and 72.18, respectively) indicate that they originated from immature source rocks with low to moderate source weathering (Fig. 13a). The Al2O3 - CaO* + Na2O - K2O (A-CN-K) ternary diagram can be used to determine the level of weathering (Nesbitt and Young 1984). The samples in this diagram are near the Pl-Ks tie line and progress towards the Al2O3 apex, following and parallel to the projected tonalite-granodiorite weathering pattern (Fig. 13b). These characteristics provide a weathering trend that deviates from the fundamental composition of tonalite (Fedo et al. 1995). The degree of weathering across the various samples is mostly similar, resulting in a compact group throughout the tonalite weathering trend (Fig. 13b), indicating weathering conditions that remained constant (Nesbitt et al. 1997) for the Late Eocene to Early Oligocene PFB mudstones.

Sedimentary processes, sorting and recycling

Sedimentary processes may significantly alter mineral abundances, which in turn affects the concentrations of certain elements. Our sediments include a variety of grain sizes, from mudstone to very fine sandstone. Contrary to coarser-grained sediments, fine-grained sediments have a chemical composition that is identical to that of their source (Cullers 1994, 2000). The SiO2/Al2O3 ratios can be used to measure the textural maturity (McLennan et al. 1993), ranging from 3.19 to 6.85 and are typically greater than those of PAAS (3.3). Plagioclase may be incorporated into mudstones because of sorting, which would reduce the Eu anomaly (McLennan et al. 1993). This is especially accurate for mudstones created in tectonically active regions (McLennan et al. 1990), where mudstones lack a constant Eu enrichment because of the low percentage of plagioclase in sediments (Nathan 1976; Bhatia 1985; McLennan et al. 1993). Mudstones from the foredeep depozone in the PFB had similar Eu anomalies to those in the PAAS, demonstrating the absence of plagioclase concentration because of sediment physical sorting. In contrast to mature sedimentary rocks, immature source rocks with limited degree of sorting and recycling have a narrower range of TiO2/Zr variation (Garcia et al. 1991). The samples plot close to the PAAS and show a narrow range of TiO2/Zr variation, indicating an immature source with low source sorting and sediment recycling (Fig. 13c).

Provenance

The results of this study indicate multiple sources that contributed to the sedimentation of the PFB. Detritus such as monocrystalline and polycrystalline quartz (Qp < 5 splits), plagioclase, K-feldspar, composite quartz-feldspar, and quartz-mica grains provide evidence that the PFB originated from a magmatic source. These results are supported by cathodoluminescence petrography of quartz grains, which shows that detrital monocrystalline quartz is of volcanic origin (red quartz crystal CL color) and, less frequently, of metamorphic origin (brown quartz crystal CL color) (Götze and Zimmerle 2000). Chlorite can be created by the transformation of mafic minerals and volcanic glass. Nevertheless, the presence of detrital minerals of the chlorite group which are well known in low temperature and prograde metamorphic rocks (Deer et al. 2013), provide evidence of a low-grade metamorphic source-rock, which is further confirmed by the occurrence of polycrystalline quartz with more than five splits, as well as slate and schist. Sedimentary sources are another significant contributor that provided debris to the PFB. This source is also made of shale, sandstone, limestone, and chert, according to the thin sections and conglomerates of the study area.

Chemical features reflect the composition of the sources involved in the sedimentation. The homogeneous mixing of the central PFB sediments (Fig. 13c) is attributed to the efficient mixing of the source rock fragments during transit and deposition (McLennan 1989; Vital and Stattegger 2000). The A-CN-K plot pattern indicates that it originated from less felsic parent rocks such as granodiorite (Fig. 13b). TiO2 vs. Al2O3 graph and moderate to high Al2O3/TiO2 ratios (mean value: 19.02) indicate that the detritus came from felsic rocks similar to that suggested in the A-CN-K plot (Fig. 14a). The REE, Th, and HFSE can be used to evaluate the composition of the rocks in the source areas (Taylor and McLennan 1985; McLennan et al. 1990). These elements become insoluble and immobile during weathering, metamorphism, and they have a low sensitivity to post-crystallization alteration (White et al. 2002; Grosch et al. 2007; Koralay 2010). Furthermore, they are more abundant in felsic rocks than in mafic rocks (Etemad-Saeed et al. 2011) and Th is widely used as a proxy for terrigenous sediment input in marine environments (McManus et al. 2004).

Fig. 14
figure 14

a TiO2 against Al2O3 diagram for the PFB samples. The “granite”, “granodiorite” and the “3 granite + 1 basalt” lines are from Schieber (1992), b plot of the studied samples on the La-Th-Sc ternary diagram proposed by Cullers (1994). Abbreviations: PM: Passive margin; ACM: Active continental margin; CIA; Continental Island arc; OIA: Oceanic island arc, c Th/Sc against Zr/Sc diagram (McLennan et al. 1993) that evaluates the source rock composition of the studied sediments. Abbreviations: BAS, AND, GRD and GR refer to average gabbro, andesite, granodiorite and granite, respectively (Le Maitre 1976)

Full size image

According to McLennan et al. (1993) and the references within, the Eu anomaly is widely assumed to have been acquired from the sediment sources. Large abnormalities are usually attributed to a felsic origin, whereas small Eu anomalies are frequently attributed to mafic debris (Taylor and McLennan 1985; Hassan et al. 1999; Cullers 2000). The PFB samples exhibit Eu anomalies (Eu/Eu* = 0.65–0.78) that are comparable to higher than the average value of PAAS (Eu/Eu* = 0.65), indicating that these sediments were eroded from a mixed source of felsic and mafic origin (Taylor and McLennan 1985). The chondrite-normalized rare earth element (REE) abundances of PFB (Figure 11) are similar to those of UCC and PAAS (enrichment in light rare earth elements (LREE), flat heavy rare earth elements (HREE), and negative Eu anomalies). This similarity suggests that PFB has a differentiated source resembling granite. This conclusion is further supported by the high LREE/HREE ratio (average 7.34), which is a characteristic feature of felsic source rocks (Cullers 1994; Taylor and McLennan 1985; Wronkiewicz and Condie 1990). Elemental ratios such as Th/Sc and La/Sc (Table 2) and the presence of the Eu anomaly are easily influenced by the average composition of the source rock (Taylor and McLennan 1985). The significant variations in element ratios among different regions indicate that local source rocks have an influence on sediment composition. The La/Sc ratios (1.33–2.07) and Th/Sc ratios (0.44–0.64) observed in the PFB sediments are not typical of recycled sediments, but instead suggest a higher proportion of felsic components in the source. They are also enriched in Th/U ratios and in HFSE compared to PAAS (Table 2). Therefore, it is probable that the samples are more likely derived from a felsic to upper continental source rather than a predominantly mafic source with little recycling (Taylor and McLennan 1985; Hassan et al. 1999; Bauluz et al. 2000).

The felsic to intermediate source rock composition is further suggested by (1) the La-Th-Sc ternary diagram proposed by Cullers (1994), in which the samples plot at the clay, silt, sand, and gravel from mixed sources field area, between the granitic gneiss to metabasic field (Fig. 14b); (2) LREE enrichment documented by the high (La/Yb)N ratios; (3) flat HREE segments documented by the (Gd/Yb)N ratios (Figs. 11) and 4) the negative Eu anomalies. The trace elements Th, Zr, and Sc are used in Fig. 14c because they are typically immobile. This figure is beneficial because passive continental margin sediments exhibit increased Zr/Sc ratios due to zircon enrichment via sediment recycling (McLennan et al. 1993). Active-margin samples, on the other hand, show a trend between mafic and continental origins. Fig. 14 demonstrates that the PFB samples follow this pattern, showing low sediment recycling, despite the presence of slight zircon addition in sandstones (Harper, 1980; Miller and Saleeby, 1995). Furthermore, in the plot of Th/Sc vs Zr/Sc (Fig. 14c), the samples show Th/Sc values that indicate origin from less felsic igneous sources, which is corroborated by the higher Cr and Ni values. Because of their low mobility during sedimentary processes, elements like V, Co, Cr, and Ni are regarded as being useful for determining the origin of sediments (Floyd and Leveridge 1987; McLennan et al. 1993; Rollinson 1993; Cullers 2000). Moreover, the higher Cr and Ni concentrations in the PFB samples (enriched relative to PAAS values) point to a mafic source input. Further evidence for this source type comes from the Cr/V and Y/Ni ratios, as well as the Y/Ni versus Cr/V and Ni against Cr diagrams (Fig. 15a, b).

Fig. 15
figure 15

Discrimination plots of (a) Ni against Cr after Garver et al. (1996) and (b) Y/Ni against Cr/V after McLennan et al. (1993). Samples that exceed the limits of 100 ppm Ni and 150 ppm Cr (dashed line) and that are enriched in the Cr/V ratio suggest mafic input. Abbreviations: PAAS: Post-Archean Australian Shale; UCC: Upper Continental Crust

Full size image

Tectonic setting

The QFL ternary diagram is used to define the tectonic setting (Garzanti 2019). PBF samples are concentrated in the recycled orogenic field (Fig. 16). Whole-bulk geochemistry, which has been employed by several authors, is used to document the tectonic setting using contemporary visuals (e.g., Verma and Armstrong-Altrin 2013; Zaid and Gahtani 2015; Tawfik et al. 2017; Maravelis et al. 2021). Therefore, in this study, the discriminant-function-based major-element diagram by Verma and Armstrong-Altrin (2013) is utilized to assess the tectonic setting for sedimentary rocks with high (SiO2 = 63–95%) and low (SiO2 = 35–63%) silica content. The samples plot in the collision field in both figures (Fig. 17a, b). The La-Th-Sc ternary diagram Cullers (1994) yields the same results, since all samples plot in the active continental margin field (Fig. 14b).

Fig. 16
figure 16

Ternary diagrams (after Dickinson et al. 1983) showing provenance fields defined by (a) total quartz, feldspar and unstable (nonquartzose) lithic fragments; (b) monocrystalline quartz, feldspar and unstable lithic fragments including extrabasinal carbonate lithic fragments plus polycrystalline quartz; monocrystalline plus polycrystalline quartz, feldspar and unstable lithic fragments including extrabasinal carbonate lithic fragments plus tectonite quartz and chert. The arrow illustrates the stratigraphic upward trend of the studied thin-sections

Full size image
Fig. 17
figure 17

Discriminant function multi-dimensional plots for both high-silica (a) and low-silica (b) clastic sediments (after Verma and Armstrong-Altrin 2013)

Full size image

However, stratigraphic and geochemical information should also be taken into account while analysing the tectonic framework (Ryan and Williams 2007; Maravelis et al. 2017). The region under investigation has sustained significant sediment deposition since the Triassic epoch, according to earlier stratigraphic data (Karakitsios 1995; Avramidis and Zelilidis 2001; Sotiropoulos et al. 2003; Papanikolaou 2009). Triassic evaporites that evolved upward into carbonates prevailed, followed by Upper Eocene siliciclastic sedimentation. The PFB system, which progresses from west to east from abyssal plain deposits to outer- and finally into inner-fan deposits, unconformably overlies the Eocene carbonates. The change from carbonate to siliciclastic sedimentation is related to the onset of the formation of Pindos orogen, the transition from a passive to an active continental margin, and the formation of the PFB (Fleury 1980; Degnan and Robertson 1998; Xypolias and Doutsos 2000). The succession’s stratigraphic evolution shows a comparable transition from muddy abyssal-plain deposits to sandy outer and sandy/conglomeratic inner fan deposits, illustrating system progradation and the temporal shallowing of the central PFB. In conclusion, the stratigraphic and geochemical data presented here are consistent with an end-Mesozoic and Cenozoic Alpine collision that led to development of the Pindos Orogeny and associated thrusting (Jones and Robertson 1991; Robertson et al. 1991).

Implications for the evolution of the Pindos Orogen — composition trends

During the Late Palaeozoic to Cenozoic, the Apulia microcontinent separated from Gondwana, leading to the formation of a NE-striking rift system and the opening of the Neo-Tethyan Ocean (Robertson et al. 1991; Ricou 1994; Frizon de Lamotte et al., 2011). This process resulted in the development of platforms and basins, as evidenced by sedimentary records. The domains that emerged during Early Jurassic rifting include the pre-Apulian platform, Ionian basin, Gavrovo platform, Pindos basin, and Parnassos and Pelagonian platforms (e.g. Aubouin 1965; Smith et al. 1979; Robertson et al. 1991; Karakitsios 1995, 2013; Zelilidis et al. 2003; Bourli et al. 2019a, 2019b). The distribution of shallow limestones and dolomites in the platform belts and pelagic marine carbonates in the deeper basin reflects these formations. In the Late Cretaceous and Early Cenozoic, the convergence of Africa and Eurasia caused the collision of the Apulia microcontinent with various continental fragments (e.g., Dewey et al. 1973). This collision resulted in the inversion of Mesozoic basins and the development of the External Hellenides thrust belt (Jacobshagen et al. 1978; Skourlis and Doutsos 2003; Kaplanis et al. 2013; Chatzaras et al. 2013). The thrusting in the external Hellenides occurred sequentially from east to west (Smith and Moores 1974; Robertson and Dixon and Robertson 1984). The Parnasos and Pelagonian units migrated westward, leading to the Late Eocene thrusting of the Pindos unit on top of the Gavrovo unit, causing flexural subsidence (Fig. 18). PFB is a depocenter formed by the Pindos Orogen, characterized by high sedimentation rates associated with the nascent orogen (e.g., Aubouin 1959; Konstantopoulos and Zelilidis 2012; Botziolis et al. 2021; Kovani et al. 2023).

Fig. 18
figure 18

Evolutionary model of the external Hellinides during the Cretaceous–Lower Eocene extension (rifting period) and Lower Eocene-Oligocene compression (modified from Bourli et al. 2019a, 2022; Botziolis et al. 2021; Zoumpouli et al. 2022). Notice the gradual westward migration of the Pindos orogen. Abbreviations: eIb = external Ionian basin; mIb = middle Ionian basin; iIb = internal Ionian basin; Vm = Varasova mountain; Km = Klokova mountain

Full size image

The Pindos Orogen includes a variety of units that have been involved in the sedimentation of PFB. The Pelagonian unit consists of Variscan basement, unconformably overlain by an Early to Middle Triassic sedimentary succession that consists of sandstone, slate, schist, chert, marl, limestone, and a deformed succession of lavas (basalts, trachy-basalts, basaltic andesites, and basaltic trachy-andesites), tuffites, and welded tuffs of Early to Middle Triassic volcanism (Smith et al. 1975; Jacobshagen and Wallbrecher 1984; Pe-Piper and Panagos 1989; Smith, 1993; Pe-Piper et al. 1996; Stampfli et al. 1998; De Bono 1998; Sharp and Robertson 2006). This type of volcanism has been recorded from northern Italy to Turkey (Stampfli 1996) and may be related to an arc-related volcanism, most likely brought on by back-arc rifting (Pe-Piper 1982; Pe-Piper and Panagos 1989). Up-section, the rock units are represented by Middle Triassic to Jurassic carbonate succession followed by Lower Cretaceous conglomerates and carbonates, and finally Upper Cretaceous to Lower Eocene deep-sea fan deposits (Jacobshagen and Wallbrecher 1984). The Parnasos units are represented by a series of shallow carbonate platform deposits of Triassic to Cretaceous age (e.g. Mountrakis 1985; Pomoni-Papaioannou 1994), overlain by deep-sea fan deposits of Palaeocene to Eocene age. The Pindos older rocks units are Middle to Upper Triassic sandstone, chert, marl, limestone, and volcano-sedimentary material (Aubouin 1957; Aubouin et al. 1970; Wagreich et al. 1996). Up-section, Upper Triassic to Lower Jurassic mudstone, sandstone, siliceous limestone, and chest are followed by Middle to Upper Jurassic multi-coloured cherts with thin layers of clayey-siliceous material, pellites, and limestone. Up-section, Upper Jurassic to Lower Cretaceous limestones with thin layers of clayey-marly material and cherts, as well as brecciated limestones, calc-limestones, and cherts, are followed by Lower Cretaceous limestones with cherts, and a series of Upper Cretaceous to Lower Eocene transitional beds composed of mudstone, sandstone, and limestones. The Pindos youngest rock units are the Upper Eocene to Late Pliocene submarine-fan deposits (Fleury 1980; Ananiadis et al. 2004).

Sedimentation can be influenced by both uplift and erosion occurring simultaneously at different source locations, with the erosion of quickly uplifted sediments resulting in enormous volumes of debris that can travel hundreds of kilometres (Critelli 1993; Ingersoll et al. 2003; Garzanti et al. 2004a, 2004b). Each source may influence sediments composition by directly supplying detritus to the basins in an axial or longitudinal sediment delivery network and by reworking other sedimentary basins that are part of the same system, as they become part of the growing orogen prior to deposition in the developing foreland basin. The detrital signatures of the sediments may change over time because of two factors: (a) the progressive lateral growth of the external belts that shields the foreland basin from receiving detritus from the axial belt, and (b) the fluctuation along the basin of the entry points of the main draining system of the axial belt (Muttoni et al. 2003; Najman et al. 2003). Transit distance, the impact of strike-slip fault zones, and sediment distribution techniques are other elements influencing this trend (direct against longitudinal transport) (Schwab 1981).

Sediment deposition in the PFB is thought to be influenced by the Pindos Orogen thrust system and the related strike-slip faults (Zelilidis et al. 2008). The Pindos Orogen was the main source of sediment, according to the paleocurrent analysis, which shows a paleodispersal direction (NE-SW) axial to it (Botziolis et al. 2021). The quartz and low feldspar contents of the sandstones indicate that these sediments were likely generated by large rivers from collision orogen and foreland uplift sources (Dickinson and Suczek 1979). In the Pindos thrust zone, coarse-grained deposits (conglomeratic channels) have been found, and they are strongly connected to transform fault zones that crosscut the thrust (Fig. 19). This connection demonstrates that these faults acted as entrance points (Figs. 19, 20). The Pindos drainage system entering the basin quickly shifted to axial direction flows (because of its geometry) and channels were formed supplying and delivering the debris from the Pindos orogen source regions, depositing the sand-rich lobe system (Fig. 20).

Fig. 19
figure 19

Schematic diagram (check Fig. 1 for the A-A’-A” cross section profile) illustrating the proposed entry points of the sediments into the basin. Abbreviations: Clg = Conglomeratic channels; Scf-L = Sandy channels and levees; Lbs = lobe-axis, lobe off-axis, lobe fringe and distal lobe fringe; NCSS SFZ = North Corfu–South Salento strike-slip fault zone; AG SFZ = Agia Kiriaki strike-slip fault zone; P SFZ = Paramithia strike-slip fault zone; A SFZ = Alevrada strike-slip fault zone; T SFZ = Lake Trichonis strike-slip fault zone; PR SFZ = Patra-Rio strike-slip fault zone; S SFZ = Simopoulo strike-slip fault zone; PA SFZ = Pyrgos-Alfeios strike-slip fault zone; Z SFZ = Zacharo strike-slip fault zone; Ky SFZ = Kyparissia strike-slip fault zone; Ka SFZ = Kalamata strike-slip fault zone

Full size image
Fig. 20
figure 20

3D model reconstruction illustrating the depositional setting of the Pindos Foreland Basin during the (a) Late Eocene and (b) Early Oligocene

Full size image

The potential source rocks that contributed to the PFB sedimentation were identified by the mineral composition of the studied sandstones. Non-undulatory monocrystalline quartz grains and polycrystalline quartz (with < 5 splits) may have originated from tonalites found in Middle to Late Triassic volcanic-sedimentary strata of the Pelagonian unit. Moreover, PFB sandstone contains both plagioclase, feldspar, and alkali feldspar, although plagioclase is more prevalent. The presence of plagioclase, feldspar and alkali feldspar in sandstone can be indicative of a volcanic origin. Plagioclase and feldspar are typically more abundant in volcanic rocks compared to alkali feldspar. Therefore, as plagioclase is more prevalent in the PFB sandstone, it suggests a volcanic origin, most likely from a source rock related to the Early to Middle Triassic volcanic units (granite-tonalite rock) of the Pelagonian unit. These findings are consistent with sedimentation in a pro-foreland basin, which is distinguished by the presence of (recycled) sedimentary material from shallow crustal depth sources (Nagel et al. 2014). Polycrystalline quartz (with > 5 splits) and undulatory quartz grains were sourced from low grade metamorphic rocks and earlier sandstone provenances (Folk 1974; Basu et al. 1975), thus, can be ascribed to the sandstone, slate, and schist succession of the Middle to Upper Triassic formation of the Pelagonian and Pindos units, as well as the Upper Cretaceous to Lower Eocene mudstone and sandstone of Pindos unit. This interpretation is further suggested by the sandstone, slate and schist fragments, as well as by the presence of detrital chlorite in the thin sections. Sandstone source rocks include the Upper Cretaceous to Lower Eocene deep-sea fan deposits of Pelagonian units, the Palaeocene to Eocene deep-sea fan deposits of Parnasos units, and the Upper Eocene to Lower Pliocene submarine-fan deposits of Pindos units. The Middle to Upper Triassic chert beds, Upper Triassic to Upper Jurassic multi-coloured chert beds, and Upper Jurassic to Upper Cretaceous limestones that contain cherts are probable Pindos units that could have provided chert in PFB.

Following the sandstone petrography, the geochemical analysis confirms that the studied sedimentary rocks were deposited in a collisional setting and that the PFB sediments were derived from a variety of rock types, including felsic to intermediate volcanic, sedimentary, and low-grade metamorphic rocks. Geochemical analyses of volcanic rocks from the Pelagonian unit, using rock/MORB and rock/chondrite REE spider diagrams and K2O-SiO2, N2O+K2O-SiO2, Ti-Zr-Y, Ti-Zr scatter diagrams reveal calc-alkaline to alkaline affinities (De Bono 1998; De Bono et al. 1998, 1999), similar to the studied sediments, further supporting the petrographical interpretation. Despite the absence of mafic minerals, the elevated Ni and Cr levels of the samples indicate that the Pindos ophiolite complexes also contributed detritus in the PFB. The stratigraphic location of the samples with the highest Ni, Cr, and V concentrations is shown in Fig. 4, indicating that the predominant contribution from a mafic source occurs in the upper-part of the outer fan deposits, above the Eocene-Oligocene boundary. The Jurassic Pindos Ophiolite complex in western Greece is a supra-subduction ophiolite with lavas and cross-cutting dykes representing a diverse spectrum of magma types (Dupuy et al. 1984; Pearce et al. 1984; Kostopoulos and Murton 1992; Pe-Piper et al. 2004). These rocks, which range in composition from high-Ti mid-ocean ridge basalt (MORB) to island-arc tholeiites (IAT) to boninites, are probable source rocks, explaining the high to extremely high Cr and Ni levels found in these samples. The lack of mafic minerals observed in the thin sections may be attributed to their transformation into secondary minerals, such as chlorite or carbonates. Consequently, Cr and Ni could potentially reside within these secondary minerals formed because of alteration of primary minerals. It is also conceivable that Cr and Ni are contained in fine-grained minerals that are not readily discernible through microscopic analysis under a polar microscope.

Sandstone composition can also give information about the sediment source and the unroofing history of the orogen, reflecting geographical and temporal variations of erosion. Detritus from the volcanic arc and subduction complex is present in the early phases of basin formation (Dorsey 1988; Garzanti et al. 1996; Najman and Garzanti 2000). The input of metamorphic debris steadily increases as the axial metamorphic core of the orogen expands during later collisional stages (White et al. 2002). At the early stage of the orogenic evolution, the PFB fills with sediments originating from the Pelagonian units as evidenced by the high proportion of monocrystalline and polycrystalline quartz (<5 grain splits), indicating a dominant igneous source rock. The up-section increase in the abundance of quartz with undulose extinction and/or having >5 grain splits, chert, sandstone and low-grade metamorphic lithic fragments (slate, and schist) are most likely associated with the progressive unroofing of the Pindos orogen. This tendency of temporal increase in lithic fragments throughout compositional transitions, marks the shift from stable craton (Pelagonian unit) to orogen (Pindos unit) and indicates the westwards migration of the Pindos orogen. At this later collisional stage, the contribution of the exposed metamorphic and sedimentary rocks of the orogenic axial core prevails (Parnasos and Pindos units), as indicated by the increase in lithic fragments and the decreasingly contribution of the igneous source rock (Pelagonian unit) is indicated by the decreasing trend of feldspathic, volcanic lithic, and monocrystalline quartz. This “shielding” can be further confirmed by the Cr vs Ni and Cr/V vs Y/Ni plots (Fig. 15a, b) that document a progressive upward decrease in the involvement of mafic rocks as a contributor to the basin sedimentation.

According to Kovani et al. (2023), provenance analyses of the conglomeratic matrix within the submarine fan deposits in the PFB have revealed similar trends. These trends have been interpreted as being influenced by the growth of the Pindos Orogen. Additionally, the progressive unroofing of the Himalaya Orogen throughout the Paleogene is demonstrated by an increase in the proportion of monocrystalline quartz and total lithic fragments from the lower Rud Faqirzai formation to the overlying Manzaki formation (Qayyum et al. 2001). On another example, the Sierra de Reyes foreland basin displays similar trends and indicates a typical unroofing history that resulted in the deposition of a succession with variable clastic composition, followed by an up-section enrichment in metamorphic rock fragments (Sagripanti et al. 2012). As a result, the detrital patterns of the deposited submarine fans in foreland basins collectively document the gradual unroofing of the orogen in a chronological and geographical context.

Conclusions

Petrographic and geochemical data, combined with previous sedimentological and stratigraphic analyses of the examined deposits, have provided insights into the PFB provenance and tectonic setting. The geochemical and petrographic composition of the sediments is principally affected by the different exhumation history of the source rocks and the distance of transportation. Petrographic analysis of sandstones revealed that PFB samples contain debris derived mostly from sedimentary and low-grade metamorphic rocks of the Pindos Orogen. The sedimentary sequence exhibits a general upward enrichment in sedimentary and metamorphic lithic fragments, as well as an upward reduction in monocrystalline quartz grains, feldspathic, and volcanic lithic clasts associated with the unroofing of the Pindos Orogen. The rapid exhumation of Pindos Orogen strata is demonstrated by the increasing trend in the quantity of lithic fragments in sedimentary layers, which is related with the Pindos Orogen’s unroofing and rapid uplift rates.

The samples are geochemically immature and originated from a source with a low to moderate degree of weathering, according to the ICV and CIA values. The PFB’s sediments exhibit PAAS-like chemical properties and are hence derived from a differentiated UCC. The Pindos orogen supplied sediments with calc-alkaline to alkaline affinities. The trace and REE concentrations in the samples, as well as the trace element ratios, revealed that the debris originated mostly from felsic to intermediate rocks, with a minor contribution from a mafic source. The formations exhibit PAAS-like affinities but are depleted in Si2O, Zr and Hf compared to PAAS, suggesting that they are generated from differentiated UCC, with low recycle. The Early to Middle Triassic volcanic rocks, the Upper to Lower Cretaceous carboniferous and radiolarite formations, and the Paleocene submarine-fan deposits, as well as the ophiolite complexes of the Pindos Orogen, are potential sources of detritus.

Based on the QFL and QmFLt ternary plots, the sedimentary rocks from the PFB originated from a recycled orogen. Multidimensional discrimination diagrams suggest sediment sources from a collisional setting and confirm the active continental margin setting. This explanation is in agreement with the geology of the PFB and is linked to the collision of the Apulia and Eurasian plates, which resulted in the development of the Alpine orogen. It is deduced that the studied area served as a foreland basin attached to the growing orogeny resulting from the continental collision.