Larger benthic foraminifera content, geochemical characteristics and paleoenvironmental interpretation of the Bartonian-Priabonian Kovancılar section of the Kırkgeçit Formation (Elazığ, Eastern Türkiye)

1 Introduction

The middle-late Eocene transition (Bartonian–Priabonian) represents a pivotal interval in the middle-late Eocene period, characterized by a shift from greenhouse to transitional climatic conditions, global cooling, and associated eustatic sea-level fluctuations (Zachos et al., 2001; Cramer et al., 2011; Westerhold et al., 2020). The middle-late Bartonian and early Priabonian witnessed the onset of ephemeral Antarctic glaciation, a ∼50 m sea-level drop, and marked changes in marine and continental ecosystems (Shackleton and Kennett, 1975; Zachos et al., 1994; Katz et al., 2010). These climatic shifts were related by reorganizations in oceanic circulation and sedimentation patterns, particularly in tectonically active regions such as the Neotethys (Jovane et al., 2009; Pearson, 2012; Hoogakker et al., 2024). Within the Neotethyan domain, these changes are recorded across a broad spectrum of depositional settings, from marginal marine shelves to foreland and back-arc basins, including those in present-day Anatolia (Akkiraz et al., 2006; Rego et al., 2018; Bozkurt and Görmüş, 2021).

In Anatolia, the closure of the southern Neotethys and the northward subduction of oceanic crust created a complex array of basins influenced by both regional tectonics and global sea-level changes (Türkmen et al., 2001; Aksoy et al., 2005). The Elazığ basin, situated in eastern Anatolia, formed within this tectonically active framework during the middle Eocene to middle Miocene and preserves thick successions of marine and mixed siliciclastic-carbonate sediments. This basin experienced syndepositional block faulting, producing an irregular seafloor topography that controlled facies distribution and sedimentation patterns (Özkul and Kerey, 1996; Cronin et al., 2000a; Cronin et al., 2000b). During the Bartonian–Priabonian, several basins in Anatolia, including those in western and central regions, were characterized by transgressive–regressive cycles, coastal swamp and carbonate platform environments, and faunal communities typical of subtropical to tropical conditions (Akkiraz et al., 2006; Özcan et al., 2010; Yücel et al., 2020).

Within this context, the Kırkgeçit Formation is of particular significance (Figures 1A,B). This lithostratigraphic unit crops out extensively in the Elazığ basin and consists of a thick and variable succession of shallow to deep-marine deposits, ranging from fossiliferous limestones and marls to siliciclastic turbidites (Özkul, 1988; Aksoy et al., 2005; Alkaç and Aksoy, 2022). The formation unconformably overlies the upper Cretaceous Elazığ Magmatics/Yüksekova Complex and is succeeded by the upper Miocene-lower Pliocene Çaybağı Formation and Quaternary Palu deposits (Perinçek, 1979; Sungurlu et al., 1985; Kerey and Türkmen, 1991; Koç-Taşgın and Türkmen, 2009). Biostratigraphic studies based on LBF such as Nummulites, Chapmanina, Silvestriella, and orthophragminid have dated the Kırkgeçit Formation to the middle and late Eocene, with some interpretations extending into the Oligocene (Sirel, 2000; Sirel, 2003; Sirel, 2015; Hoşgör and Okan, 2009; Sirel and Deveciler, 2018). The presence of Nummulites fabianii, Nummulites hormoensis, and associated taxa allows the correlation with shallow benthic zones SBZ 18–20 (Less and Özcan, 2012; Özcan et al., 2019a), supporting a latest Bartonian to Priabonian age. These faunal assemblages provide valuable insights into paleoenvironmental gradients, water depth variations, and sedimentary conditions during a time of major paleoceanographic reorganization (Hottinger, 2000; Beavington-Penney and Racey, 2004; Briguglio and Hohenegger, 2009; Henry et al., 2025a; Henry et al., 2025b).

Although numerous paleontological studies have been conducted on the Kırkgeçit Formation across the Elazığ basin (Sirel et al., 1975; Avşar, 1983; Avşar, 1991; Avşar, 1996; Özcan et al., 2006; Özcan et al., 2019a; Kayğılı, 2021; Kayğılı, 2022), investigations combining sedimentological, mineralogical, and geochemical proxies remain scarce. In particular, the Kovancılar area, located approximately 46 km northeast of Elazığ, has seen limited research focused on its LBF (Çetin-İç, 2004; Kayğılı and Aksoy, 2017), and no integrated geochemical approach has yet been applied to elucidate its depositional environment. This represents a significant gap in understanding how local sedimentary systems responded to Bartonian–Priabonian global changes, especially given the potential for high-resolution paleoenvironmental reconstruction using multi-proxy data sets (Dinçer, 2016; Dinçer, 2022; Rodelli et al., 2018; Rego et al., 2018).

To address this gap, the present study combines detailed micropaleontological, mineralogical, and geochemical analyses to investigate the depositional history of the Kırkgeçit Formation in the Kovancılar region. The primary objectives are: (1) to reconstruct the paleoenvironmental evolution of the basin during the middle-late Eocene based on fossil assemblages, facies characteristics, and geochemical indicators; (2) to assess the relative influence of tectonic activity, climate, and eustatic sea-level fluctuations on sedimentation; and (3) to compare the local stratigraphic record with regional and global patterns associated with the Bartonian–Priabonian transition. By integrating foraminiferal biostratigraphy with goechemical and mineralogical proxies, this study aims to contribute to a broader understanding of paleoenvironmental dynamics within the closing Neotethys realm during a key interval of the Cenozoic Earth history.

2 Geological setting

The Kovancılar section is located approximately 46 km northeast of Elazığ, in eastern Türkiye, where the middle-upper Eocene Kırkgeçit Formation is well exposed (Figures 1A,B). This formation represents one of the main stratigraphic units of the Elazığ basin and provides key insights into marine sedimentation within the Neotethyan domain. The stratigraphic succession in the region also includes the upper Cretaceous Elazığ Magmatics/Yüksekova Complex, the upper Miocene–lower Pliocene Çaybağı Formation, and the Quaternary Palu Formation (Perinçek, 1979; Sungurlu et al., 1985; Aksoy et al., 2005). The type section of the Kırkgeçit Formation, located near Van (approximately 25 km southeast of the city), was first defined by geologists from the Turkish Petroleum Corporation (Perinçek, 1979; Sungurlu et al., 1985).

The Kırkgeçit Formation extends broadly across the eastern Taurus region, with NE–SW oriented outcrops around Elazığ (Perinçek, 1979; Sungurlu et al., 1985; Aksoy et al., 2005). It is interpreted as having been deposited within a back-arc basin formed under an extensional tectonic regime, where block faulting strongly influenced basin morphology and sediment dispersal (Cronin et al., 2000a; Aksoy et al., 2005). The formation exhibits diverse facies ranging from shallow-marine limestones and clastics to bathyal bioclastic limestones, marls, and deeper marine siliciclastic deposits (Özkul, 1988; Özkul and Kerey, 1996; Türkmen et al., 1999; Türkmen et al., 2001; Alkaç and Aksoy, 2022).

Chronostratigraphic studies based on fossil assemblage have dated the Kırkgeçit Formation primarily to the middle and late Eocene (Avşar, 1983; Avşar, 1991; Avşar, 1996; İbilioğlu, 2010; İbilioğlu, 2011; Özcan et al., 2006; Özcan et al., 2019a; Kaya-Çağlar, 2009; Kayğılı, 2021; Kayğılı, 2022), although some interpretations extend into the Oligocene (Sirel, 2000; Sirel, 2003; Sirel, 2015; Hoşgör and Okan, 2009; Sirel and Deveciler, 2018). Numerous works have examined the paleontological and biostratigraphic characteristics of the formation (Sirel et al., 1975; Sirel and Gündüz, 1979; Sirel, 2000; Sirel, 2003; Sirel, 2015; Rodelli et al., 2018; Rego et al., 2018; Özcan et al., 2019a). However, in the Kovancılar district, studies specifically focusing on LBF are limited (Sirel, 2000; Sirel, 2003; Sirel, 2015; Çetin-İç, 2004; Kayğılı and Aksoy, 2017; Sirel and Deveciler, 2018), and no prior geochemical investigations had been conducted before this study.

The stratigraphic section measured at Kovancılar reaches a total thickness of 602 m and comprises alternating intervals of marl, conglomerate, sandstone, sandy limestone, bedded limestone, and massive limestone (Figures 2 and 3). A total of 405 thin sections were prepared from 45 samples collected throughout the sequence—including both loose and rock materials—for the taxonomic identification of LBF.

3 Materials and methods

3.1 Microfacies analyses and larger benthic foraminifera

335 oriented thin sections from loose samples and 70 thin sections were prepared from rock samples collected along the Kovancılar stratigraphic section measured (The starting and ending coordinates are as follows: 37 S 574991.95 E, 4288331.57N; 574731.81 E, 4289355.20 N) in the Kırkgeçit Formation. The identification of Shallow Benthic Zones (SBZ) was based on the taxonomic composition of LBF observed in these thin sections (Figures 4–7). Zonation followed the schemes proposed by Less and Özcan (2012) and Papazzoni et al. (2017). Carbonate rock textures were classified according to the system of Dunham (1962) and Embry and Klovan (1971). Facies analyses and paleoenvironmental interpretations were based on criteria from Hottinger (1997), Beavington-Penney and Racey (2004), Flügel (2004), Nebelsick et al. (2005), and Schlager (2005) (Table 1; Figures 8 and 9).

TABLE 1. Facies description.

3.2 Mineralogical and geochemical analyses

Stable isotope analyses (δ¹³C and δ¹⁸O), SEM-EDS, XRD, XRF, and ICP-MS analyses were performed on samples taken from the Kovancılar section (Tables 2–7; Figures 10–14).

TABLE 2. Stable isotope analysis results.

TABLE 3. Summary of the main statistically significant associations identified in the Spearman correlation matrix.

TABLE 4. EDS analyses result for larger benthic foraminifera.

TABLE 5. Major oxide analysis results (%).

TABLE 6. Trace elements analysis results (ppm).

TABLE 7. REE elements analysis results (ppm).

Stable isotope analyses (δ¹³C and δ¹⁸O) were performed on 24 loose samples and 3 rock samples from the Kovancılar section (Table 2; Figures 10A,B). The δ¹³C and δ¹⁸O isotope ratio determination experiment in carbonate samples was carried out using the Gas Bench-Continuous Flow Isotope Ratio Mass Spectrometry (DeltaPlus XP Isotope Ratio Mass Spectrometer - ThermoFinnigan) technique. The measurements were done at the Stable 1sotope Laboratory of the Middle East Technical University, Ankara, Türkiye.

Scanning Electron Microscopy with Energy Dispersive X-ray Spectroscopy (SEM-EDS) was conducted on one LBF specimen at the General Directorate of Mineral Research and Exploration, in Ankara, Türkiye (Table 4; Figures 11 and 12).

For SEM-EDS analysis, the samples were first coated with gold. Under high vacuum conditions, secondary electron, backscattered electron, or mixed signal images were taken from the sample surface and photographed (https://www.mta.gov.tr/ucretli-isler/). Non-standard semi-quantitative point analysis or line analysis was performed at the desired location. Photographs and EDS analysis results were recorded.

X-ray Diffraction (XRD) analyses were performed on 10 samples from the stratigraphic section (Figure 13), at the Centro Oceanográfico de Registros Estratigráficos (CORE), of the Oceanographic Institute of the University of São Paulo, in São Paulo, Brazil. Approximately 0.2 g of bulk sediment were pulverized in an opal mortar, then sieved into 150 μm. The samples were analysed with an Olympus® BTX diffractometer using Co-K radiation, operated at 30 kV and 0.326 mA, over the 2θ range of 5°–55°, with a step size of 0.05° and 80 exposures (Figure 13). Mineral identification was carried out using the Highscore Plus software, which uses the PDF-2 International Centre for Diffraction Data (ICDD) database. XRD identification criteria was based on Biscaye (1965); and Moore and Reynolds (1989).

In addition, XRD and X-Ray Fluorescence (XRF) analyses of LBF were carried out on one (Kov3) and two samples (Kov4 and Kov39), respectively, at the Geosciences Application and Research Center, in Ankara Türkiye (Figure 14A; Table 5). XRD measurements of the samples were taken with Inel brand Equinox 1000 model X-ray diffractometer (XRD) device at 28mA and 29Kv with cobalt (Co.) anode. The measurement data taken together are given in this report by converting to copper (Cu) note. X-Ray Fluorescence Spectrometry (XRF) analyses of the samples were performed on a Spectro brand X-Lab 2000 model device. Measurements were carried out by preparing powder/press pellets at 15.0 kV and 11.4 mA conditions.

Six additional LBF samples were analysed by XRF (Kov1, Kov2, Kov5, Kov13, Kov19, Kov37) (Table 5), and nine by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) (Kov1, Kov2, Kov4, Kov5, Kov13, Kov19, Kov36, Kov37, Kov39) (Tables 6 and 7), at Wuhan Sample Solution Analytical Technology Co., Ltd., in Wuhan, China (https://www.samplesolution.cn/en/download/35.html). Rare Earth Element (REE) concentrations were normalized against Post-Archean Australian Shale (PAAS), following Taylor and McLennan (1985).

The sample pretreatment of whole rock for major element analysis was made by the melting method (https://www.samplesolution.cn/en/download/35.html). The flux is a mixture of lithium tetraborate, lithium metaborate, and lithium fluoride (45:10:5). Ammonium nitrate and lithium bromide were used as oxidants and release agents, respectively. The melting temperature was 1050 °C and the melting time was 15 min. Zsx Primus II wavelength dispersive X-ray fluorescence spectrometer (XRF) produced by RIGAKU, Japan was used for the analysis of major elements in the whole rock. The X-ray tube is a 4.0 Kw end window Rh target, the test conditions are voltage: 50 kV, current: 60mA, and all major element analysis lines are kα, the standard curve uses the national standard material: rock standard sample GBW07101-14, soil standard sample GBW07401-08, stream sediment standard sample GBW07302-12. The data were corrected by the theoretical α coefficient method. The relative standard deviation (RSD) is less than 2%.

Trace element and REE analysis of whole rocks was conducted on an Agilent 7700e ICP-MS at the Wuhan Sample Solution Analytical Technology Co., Ltd., Wuhan, China (https://www.samplesolution.cn/en/download/35.html). The detailed sample-digesting procedure was as follows: (1) Sample powder (200 mesh) were placed in an oven at 105 °C for drying of 12 h; (2) 50 mg sample powder was accurately weighed and placed in a Teflon bomb; (3) 1 mL HNO₃ and 1 mL HF were slowly added into the Teflon bomb; (4) Teflon bomb was putted in a stainless steel pressure jacket and heated to 190 °C in an oven for >24 h; (5) After cooling, the Teflon bomb was opened and placed on a hotplate at 140 °C and evaporated to incipient dryness, and then 1 mL HNO₃ was added and evaporated to dryness again; (6) 1 mL of HNO₃, 1 mL of MQ water and 1 mL internal standard solution of 1ppm. In were added, and the Teflon bomb was resealed and placed in the oven at 190 °C for >12 h; (7) The final solution was transferred to a polyethylene bottle and diluted to 100 g by the addition of 2% HNO₃.

4 Results

A diverse assemblage of LBF was identified in the Kovancılar section, including Nummulites hormoensis, N. fabianii, N. ex. gr. perforatus, N. maximus, N. striatus, Operculina sp., Heterostegina armenica, Discocyclina dispansa, D. augustae, Orbitoclypeus varians, O. zitteli, Asterocyclina stellata, A. alticostata, Sphaerogypsina globulus, Gypsina cf. mastelensis, Asterigerina rotula, Halkyardia minima, Chapmanina gassinensis, Silvestriella tetraedra, Biplanispira sp., Nummulites sp., Gyroidinella sp., Planorbulina sp., as well as miliolid, textularid, algae, orthophragminid, stomatorbinid, and planktonic foraminifera (Figures 4–8). Shallow benthic zones were identified as SBZ18 according to Nummulites hormoensis, SBZ19- SBZ 20 according to N. fabianii, SBZ18-SBZ 20 according to Chapmanina gassinensis, Silvestriella tetraedra, and SBZ 17-SBZ18a according to N. ex. gr. Perforatus, N. maximus, N. striatus (Less and Özcan, 2012; Papazzoni et al., 2017). Based on the LBF, shallow benthic zones SBZ 17, SBZ 18 and SBZ 19–20 were identified (Less and Özcan, 2012; Papazzoni et al., 2017), assigning a Bartonian-Priabonian age to the sequence.

Eight carbonate lithofacies have been identified in the studied section. Their descriptions and respective interpretation of the depositional processes are presented in Table 1, and representative examples of each one are shown in Figures 8, 9.

4.1 Stable isotopes

Stable carbon and oxygen isotopes (δ¹³C, δ¹⁸O) are available for 27 samples (24 loose samples; 3 rock samples) across multiple facies. Both δ¹³C and δ¹⁸O values display higher variability in the lower part of the section, where the density of the samples is also higher (Table 2; Figures 10A,B). The δ¹³C–δ¹⁸O cross-plot coloured by lithofacies shows that most samples cluster within a range with δ¹³C values between ∼0.5‰ and 2.1‰ and δ¹⁸O values between −4.6‰ and −3.0‰ (Figure 10B). Importantly, this main cluster is characterized by substantial overlap among lithofacies, and no facies occupies an exclusive or clearly separated field. A second feature is the presence of a small number of isotopic outliers relative to the main cluster. One end-member corresponds to a sample with strongly negative δ¹³C (∼−4.7‰) accompanied by δ¹⁸O around −4.9‰, which deviates sharply from the rest of the dataset. A second outlier shows comparatively less negative δ¹⁸O, ∼−1.5‰.

The correlation coefficient obtained between the stable carbon and oxygen isotope values (Figure 10B) is poor (correlation coefficient = 0.28).

4.2 Elements concentrations

Elemental concentrations by ICP–MS are available for 9 samples (Kov1, Kov2, Kov4, Kov5, Kov13, Kov19, Kov36, Kov37, Kov39) spanning five facies (NOP, ONP, PFB, NPG, BG). Major oxides by XRF are available for 8 samples (Kov1, Kov2, Kov4, Kov5, Kov13, Kov19, Kov37, Kov39). Spearman correlation analysis of ICP–MS elemental concentrations (n = 9) identify a few major covariance groups in the dataset. The strongest first-order pattern is the marked negative correlation between Si and Ca. A second major association is defined by Al, K, and Ti, together with Rb, Ta, Nb, Ga, Sc, Zn, and other positively correlated lithogenic elements. Zr and Hf form a distinct positively correlated pair. Among transition metals, Co, Ni, and Cr define another positively correlated group. Mn is strongly associated with REE and Y, especially within the MREE–HREE range, whereas P shows positive correlations with Mg, Na, and U. In contrast, Sr is negatively correlated with Fe. Ca shows no significant correlation with Al, Rb, Ti, or Sr, and Zr is not significantly correlated with Al (Table 3).

4.3 PAAS-normalized REEs

PAAS-normalized REE patterns for nine Kovancılar samples (Kov1, Kov2, Kov4, Kov5, Kov13, Kov19, Kov36, Kov37, Kov39) spanning five facies (NOP, ONP, PFB, NPG, BG) were identified. Across the full suite from La to Lu, most values fall between 0.04 and 0.30 REE/PAAS, with most points clustering around ∼0.10–0.25. The curves are subparallel and define a broadly coherent pattern from LREE to HREE, with limited divergence among samples (Figure 10C). Among the dataset, Kov37 (NPG) exhibits the highest overall normalized values (e.g., La–Lu predominantly ∼0.13–0.30), whereas Kov5 (NOP) displays the lowest REE/PAAS values (e.g., La–Lu ∼0.04–0.09). Intermediate patterns are shown by the remaining NOP, ONP, and NPG samples (Kov1, Kov2, Kov4, Kov13), and by Kov19 (PFB), Kov36–Kov37 (NPG), and Kov39 (BG), which generally plot at higher normalized levels than the lowest NOP sample. The spider plot indicates that differences among samples are expressed mainly as changes in overall amplitude rather than sharp changes in pattern shape.

4.4 XRD and SEM-EDS

SEM–EDS analysis of LBF revealed that the specimens are composed predominantly of Ca and O, with minor amounts of C, Si, Al, Mg, Fe, and K (Table 4; Figures 11 and 12).

X-ray diffraction (XRD) analyses performed on 10 samples from Kov23 to Kov32 from the marl-dominated interval of the Kovancılar section (Figure 13) indicate a consistent mineralogical assemblage primarily composed of calcite and quartz, with subordinate amount of clay. Calcite is the dominant phase across all samples, identified by strong peaks at 3.05 Å and 2.28 Å. Quartz is also consistently present, with a peak at 3.36 Å, suggesting a significant siliciclastic contribution. Ca-Montmorillonite is identified in a few samples, particularly Kov27 and Kov30, with a characteristic peak at 14.92 Å, indicating minor clay mineral input. The possible contribution of clay appears in moderate amounts (notable peaks at 4.28 Å and 3.22 Å). The mineralogical composition reflects the marl lithology dominating this interval and confirms a mixed origin with both biogenic carbonate and terrigenous components.

X-ray diffraction (XRD) analyses performed on one LBF sample (Kov3) of the Kovancılar section indicate a consistent mineralogical assemblage primarily composed of calcite and quartz (Figure 14A).

X-ray fluorescence (XRF) analysis of LBF samples indicated high CaO contents (39.54%–56.57%), with variable amounts of SiO₂ (3.50%–25.29%), Al₂O₃ (0.19%–1.64%), MgO (0.16%–1.19%), Fe₂O₃ (0.19%–0.77%), K₂O (0.03%–0.28%), Na₂O (0.05%–0.16%), TiO₂ (0.02%–0.09%), MnO (0.006%–0.029%), and P₂O₅ (0.006%–0.030%) (Table 5).

5 Discussion

5.1 Depositional environment

The micropaleontological, mineralogical, and geochemical data from the Kovancılar section offer valuable insight into the depositional evolution of the Kırkgeçit Formation during the Eocene, particularly across the Bartonian–Priabonian transition. The integration of LBF assemblages, sedimentary facies, and geochemical proxies allows the reconstruction of the paleoenvironment and its controlling factors, shedding light on regional and broader Tethyan paleoceanographic patterns.

The bioclastic assemblages are typical of a warm to temperate water carbonate platform of the Eocene, with oligotrophic conditions (e.g., Flügel, 2010; Coletti et al., 2021; Mariani et al., 2024; Morabito et al., 2024). The presence of various types of fragmented bioclasts occurring in all the studied facies indicates high reworking, whereas intergranular micritic matrix can deposit in low energy conditions. Therefore, the facies association of the Kovancılar section represents an inner to outer ramp environment, probably between the fair weather and the storm wave bases, characterized by oscillating energy (Table 1). The variable proportion between Nummulites and orthophragminids indicates a slightly variable depth range, with deeper conditions when the proportion of the latter’s increases (Hottinger, 1997; Sinclair et al., 1998; Pomar, 2001a; Pomar, 2001b; Ćosović et al., 2004; Kayğılı and Aksoy, 2019; Coletti et al., 2021). The decrease in micritic matrix and the higher degree of fragmentation observed up section may indicate a shallowing upward trend, shifting from a middle ramp to an inner open marine ramp, where the fair-weather waves removed the micritic mud and enhanced the fragmentation of the bioclasts (Table 1) (Romero et al., 2002; Beavington-Penney et al., 2006; Kayğılı and Aksoy, 2019). The prominent increase in micrite, alongside the higher abundance of bioclasts of deeper dwelling organisms, such as bryozoans, occurring at the top of the section, may express an abrupt marine transgression (Table 1) (e.g., Schlager, 2005). On the other hand, the marly interval in the lower part of the section, with no skeletal bioclasts despite being encompassed in a succession of bioclastic packstones, can represent an episode of crisis of the carbonate factory, possibly due to a climatic shift or an increase in nutrients level (e.g., Schlager, 2005). The presence of silt size quartz extraclasts is consistent within most of the section, indicating relative proximity to the continent, however, the sandstones and conglomerates in the upper part of the section testify to a strong increase in siliciclastic input. Considering the geological context and the paleogeographic position of the Kovancılar section, this event may be related to a tectonic uplift and a change in the configuration of the basin, as also proposed for adjacent sections in the Elazığ and Malatya basins (Cronin et al., 2000a; Türkmen et al., 2001; Aksoy et al., 2005). This tectonic re-arrangement may also be responsible for the abrupt deepening observed at the top of the section. The sudden deepening of the environment can be explained by this tectonism, however, a partial contribution could be due to the global transgression that occurred during the Priabonian period, preceding the major global regression at the Eocene-Oligocene transition (Miller et al., 2024).

5.2 Biostratigraphic implications of LBF

Some new foraminiferal taxa, such as reticulate Nummulites, Heterostegina, Pellatispira, Silvestriella appear for the first time, while major groups of larger Nummulites and alveolinids disappear during Bartonian and early Priabonian (Less and Özcan, 2012; Serra Kiel et al., 2016; Özcan et al., 2018; Özcan et al., 2019a; Özcan et al., 2019b). Chapmanina gassinensis and Silvestriella tetraedra, which first appeared close to the Bartonian-Priabonian boundary and in the lower levels of SBZ 18A, continue their existence until the end of SBZ 20 (Less and Özcan, 2012; Özcan et al., 2018; Özcan et al., 2019a; Özcan et al., 2019b).

The reticulate Nummulites lineage exhibits consistent increases in inner proloculus diameter through time—a widely observed evolutionary pattern throughout the Tethys (Papazzoni, 1998; Less and Özcan, 2012; Less et al., 2018; Özcan et al., 2019a; Özcan et al., 2019b; Kayğılı, 2021; Kayğılı, 2022). In Kovancılar, N. hormoensis (154–199 µm) and N. fabianii (227–247 µm) (Table 8) fall within expected ranges for SBZ 18 and SBZ 19–20 respectively, supporting their use as temporal indicators (Özcan et al., 2010; Less et al., 2011; Less and Özcan, 2012; Özcan et al., 2019a; Özcan et al., 2019b; Zakrevskaya et al., 2020; Kayğılı, 2021; Kayğılı, 2022). This mirrors trends seen in Kutch (India), Armenia, and Western Europe, suggesting evolutionary responses to global climatic cooling and sea-level oscillations (Papazzoni, 1998; Sengupta et al., 2011; Cotton et al., 2015; Cotton et al., 2017; Saraswati et al., 2017; Less et al., 2018; Zakrevskaya et al., 2020). This study is based on N. hormoensis and N. fabianii; the biometrics of other Nummulites have also been determined and are given in Table 9. Discussions regarding these species continue in the literature (Table 9).

TABLE 8. The comparison of the mean inner cross diameter of the proloculus (Pmean (μm)) of reticulate Nummulites such as N. hormoensis and N. fabianii in previous studies and present study.

TABLE 9. The comparison of the range and mean inner cross diameter of the proloculus (Prange-mean (μm)) of other Nummulites in previous studies and present study.

In the Kovancılar section, the presence of taxa such as N. hormoensis, N. fabianii, N. ex. gr. perforatus, N. maximus, N. striatus, Chapmanina gassinensis, Silvestriella tetraedra, indicating a Bartonian–Priabonian age. These assemblages, typical of Eocene carbonate platforms in the Neotethys, reflect deposition in shallow marine setting across the inner to middle shelf (Racey, 2001; Romero et al., 2002; Colombié and Strasser, 2005; Vecchio et al., 2007; Adabi et al., 2008; Afify et al., 2016; Sarkar, 2017; Sarkar, 2019; Kayğılı and Aksoy, 2019; Gürsoy and Görmüş, 2020; Pleş et al., 2020; Bozkurt and Görmüş, 2021; Bozkurt et al., 2023).

5.3 Elements and stable isotopes geochemistry

Elemental and REE results are interpreted primarily as constraints on (i) compositional heterogeneity and mixing/dilution, and (ii) diagenetic alteration, with second-order attention to potential productivity-related Ba enrichment, rather than as a basis for high-confidence reconstruction of basin-scale environmental parameters (Algeo and Maynard, 2004; Tribovillard et al., 2006). This interpretive hierarchy is appropriate for LBF-dominated ramp carbonates, where small shifts in microfacies and early diagenesis can dominate bulk chemistry and must be evaluated before broader environmental interpretations (Hottinger, 1997; Flügel, 2010; Pleş et al., 2020).

As shown in Figure 10B, the stable isotopes’ values do not display specific signatures according to the lithofacies. This indicates that their variability is not controlled by the relative abundance of specific carbonate constituents within the rock, such as the proportion of matrix vs. cement, or matrix vs. bioclasts (e.g., Huck et al., 2017). Nevertheless, shallow-water carbonates isotopic composition can be affected also by diagenetic modification, particularly concerning δ¹⁸O, whereas δ¹³C is often comparatively more resistant (Pearson, 2012; Dinçer, 2016; Dinçer, 2022). A particularly notable feature is the strongly negative δ¹³C value of sample Kov40 (−4.70‰; rock sample; ONP facies), which is probably not primary due to the large difference from the other values.

In carbonate-ramp sediments, bulk elemental composition is controlled first by the relative proportions of carbonate versus non-carbonate components, and second by the distribution of trace elements among minor constituents such as fine aluminosilicates, resistant accessory minerals, and minor reactive phases (Algeo and Maynard, 2004; Tribovillard et al., 2006). Ca-based dilution ratios (e.g., Ca/Si, Ca/Ti, Ca/Al, Ca/K, Ca/Fe) and terrigenous indicators (e.g., Al/Si, Ti/Al, Ti/Rb, Rb/Sr) therefore provide a necessary framework for describing mixing/dilution before any discussion of trace-element variability. Within the Kovancılar subset, CaO remains consistently high and major-oxide compositions indicate carbonate dominance, implying that even modest fluctuations in non-carbonate fractions can generate large relative changes in trace-element abundances. This reinforces the interpretation that a substantial fraction of the observed elemental variability is compositional rather than uniquely environmental (Algeo and Maynard, 2004; Tribovillard et al., 2006).

A defensible additional inference from the combined XRF and ICP–MS results is that “non-carbonate input” is not compositionally uniform in the sampled interval. The PFB sample (Kov19) is characterized by high SiO₂ coupled with very low TiO₂, Al₂O₃, and K₂O, and comparatively low Rb and Cs, a combination representative of quartz-rich and clay-poor siliciclastic fraction. In contrast, some NOP and ONP samples (notably Kov1–Kov2 in the XRF subset) show higher Al₂O₃–TiO₂–K₂O together with elevated Rb and Cs in the ICP–MS dataset, consistent with a more aluminosilicate-associated component typical of clays/micas/feldspars. This partitioning aligns with microfacies observations documenting variable abundance of quartz extraclast and allows for explaining the variability of terrigenous input across the section: some intervals are dominated by quartz, whereas others by a fine aluminosilicate component (Hottinger, 1997; Flügel, 2010).

Therefore, by looking at the various terrigenous proxies shown in Figure 14B, a relative decrease in siliciclastic input can be observed in the lower part of the section, corresponding also to a decrease in Al/Si and an increase in Si/Ti, indicating a shift towards coarser and more quartz rich terrigenous material. This occurs within an interval dominated by the PFB facies and characterized by slightly lower values of both δ¹⁸O and δ¹³C.

Strontium spans a wide range in the ICP–MS subset (hundreds of ppm). The absence of correlation with other terrigenous elements indicates that Sr is mainly bound to the carbonate phases in the studied section. In shallow-water carbonates, Sr is sensitive to primary mineralogy (e.g., aragonite versus calcite) and to diagenetic mobilization, therefore can help constrain diagenetic modification when interpreted alongside petrography (Brand and Veizer, 1980; Flügel, 2010; Pearson, 2012). Carbonate diagenesis generally removes Sr and enriches Mn, due to the very different partitioning coefficient of these two elements, therefore, the Mn/Sr ratio can be a useful proxy for diagenetic alteration (e.g., Derry et al., 1992; Narbonne et al., 1994; Jacobsen and Kaufman, 1999). Despite only five samples have data of both elements and stable isotopes, and that Mn/Sr values are quite higher than those expected from a primary signal (e.g., Narbonne et al., 1994), the Mn/Sr ratio does not show a clear match with δ¹⁸O, δ¹³C, or any lithofacies, suggesting that diagenesis is not the main controlling factor of these parameters (Figures 14Ca,Cb). Similarly, Mg/Ca ratio is independent of lithofacies and stable isotopic compositions, indicating no significant Mg loss or incorporation during diagenesis (Figures 14Cc,Cd) (Brand and Veizer, 1980).

The uniformly low PAAS-normalized REE levels are consistent with carbonate-rich facies, as the capacity of carbonate minerals to incorporate REE is limited (e.g., Sholkovitz and Shen, 1995). In such settings, REE variability often reflects small changes in the proportion of non-carbonate REE-bearing components (fine siliciclastics, resistant accessory minerals, and minor reactive phases) rather than large-scale shifts in external REE supply (Bau and Dulski, 1996; Lawrence et al., 2006). The Kovancılar patterns are broadly coherent and differ largely in amplitude, supporting an interpretation dominated by dilution/mixing effects. Amplitude contrasts between samples (e.g., higher overall REE/PAAS in Kov37 relative to very low levels in Kov5) can be explained by differences in the mass fraction of minor components relative to carbonate, as in shallow-water platform carbonates even small increases in fine siliciclastic material or other REE-bearing constituents can substantially increase the bulk normalized REE levels, while carbonate-rich intervals can show marked depletion by dilution.

The generally higher values of HREE in the studied section reflect a typical shallow water carbonate signal (e.g., Elderfield and Greaves, 1982; Madhavaraju and Ramasamy, 1999). Within this framework, Y/Ho values (29–33) fall within ranges commonly reported for carbonate rocks, suggesting that carbonate mineralogy and detrital contribution influence the REE budget. Ce/Ce* values below 1 document a negative Ce anomaly in calculated indices; however, Ce anomalies in shallow-water carbonates can be modified by detrital mixing and early diagenesis and should not be used alone to infer redox conditions. Eu/Eu* values modestly above 1 can reflect mineralogical mixing (e.g., detrital components) and/or diagenetic effects (Bau and Dulski, 1996; Lawrence et al., 2006). Overall, the REE dataset shows the signature of a typical shallow water carbonate platform, influenced by fluctuating energy and detrital input, with minor diagenetic alteration.

Geochemical analysis results of samples taken along the examined section show a low positive correlation (r: 0.28) between δ¹³C and δ¹⁸O values. This positive correlation indicates diagenetic alteration (Fio et al., 2010). The general decreasing trend of the δ¹³C values may express the transition from a middle to an inner carbonate ramp, according also to the facies sequence (e.g., Frank and Bernet, 2000; Geyman and Maloof, 2021). The abrupt negative shift in both δ¹³C and δ¹⁸O occurring just below the marly interval may indicate a prominent environmental perturbation. Considering that this occurs close to the peak in siliciclastic quartz expressed by the low Ca/Si, Al/Si, and the high Si/Ti ratios in sample Kov19, this stable isotopes anomaly can be attributed to an increase in fresh water input from the continent, which introduced light C and O into the environment (e.g., Hudson, 1977; Oehlert and Swart, 2019). Such changes in salinity and nutrient contents were harmful for the main carbonate platform producers, inducing a temporary shutdown of the bioclastic sedimentation and a consequent decrease of the siliciclastic dilution, resulting the deposition of the marly interval (e.g., Westphal et al., 2010). The low compositional maturity of the siliciclastic fraction observed in the mineralogical data from this interval indicates that the sediments did not came from a proximal source. This evidence suggests that the increase in continental runoff could have been climatically controlled, rather than due to a paleogeographic rearrangement caused by tectonic.

The increase in δ¹⁸O at the top of the section is expressed by only one data point, so its representativeness may be questionable. However, considering the shift towards deeper facies in this interval, this signature may be due to the relative cooling related to the deepening of the environment, possibly enhanced by the global temperature decrease at the beginning of the late Eocene (middle Eocene-late Eocene transition) (Jovane et al., 2004; Jovane et al., 2006; Jovane et al., 2007; Jovane et al., 2009; Hutchinson et al., 2021).

6 Conclusion

The integrated micropaleontological, sedimentological, and geochemical analysis of the Kırkgeçit Formation in the Kovancılar section provides a detailed reconstruction of the paleoenvironmental evolution during the middle-late Eocene transition (Bartonian–Priabonian) in eastern Türkiye. This study successfully delineates a shallow marine carbonate ramp depositional environment, mainly characterized by inner-middle ramp facies, exhibiting a slight general shallowing upwards trend, and an abrupt transition to outer ramp at the upper levels of the section.

Biostratigraphic constraints provided by diverse LBF assemblages align the section with shallow benthic zones SBZ 17 through SBZ 19–20. Biometric data from N. fabianii and N. hormoensis further refine these regional stratigraphic correlations.

Geochemically, the stable isotope records (δ¹³C and δ¹⁸O) reflect a mainly primary marine signal with diagenetic overprinting. A significant negative excursion in both carbon and oxygen isotopes reveals a prominent environmental perturbation linked to increased continental runoff and freshwater input. This event likely triggered a temporary “carbonate crisis”, evidenced by the subsequent marly interval. Furthermore, the REE patterns, characterized by low overall concentrations and a typical shallow-water HREE enrichment, underscore the dominance of carbonate precipitation with subtle variations driven by siliciclastic pulses.

The transition from carbonate-dominated to siliciclastic-rich facies in the upper part of the section, followed by an abrupt shift to a deeper water carbonate setting, suggests a regional transgression due to tectonic activity. The observed sedimentation shifts are interpreted as responses to regional tectonic pulses and eustatic sea-level fluctuations within the evolving Neotethys basin and evolving global earth changes related to the middle-late Eocene transition. Ultimately, this research highlights the effectiveness of an integrated multi-proxy approach in deciphering the intricate depositional history and environmental dynamics of Eocene carbonate systems.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Michele Morsilli, University of Ferrara, Italy

Reviewed by: George Henry, University of Milano-Bicocca, Italy

Claudia Morabito, University of Padua, Italy