Introduction

The Jurassic-to-Paleogene Adriatic Carbonate Platform (AdCP: Vlahović et al. 2005) is dominated by shallow-water carbonates. However, Kimmeridgian-to-Tithonian deep-water (~ 200 to 300 m) carbonates, siliceous rocks, and black shales occur within the Lemeš and Gorski Kotar Basins (Fig. 1).

Fig. 1
figure 1

a Location of the study area in southern Croatia. (GPS: 44°13′59.94″ N 16° 6′23.79″ E), b geological map of the area of Mt. Poštak (after Hrvatski geološki institut 2009a; b), and c map of the present-day position of the Kimmeridgian/early Tithonian AdCP, the estimated deep-water foreland (radiolarite) basins in front of the obducted Neo-Tethys ophiolites (after Velić et al. 1994, 2002a, b), modified according to new results

Two contrasting tectonic models exist for the genesis of the Lemeš Basin.

In one model, the elongated southeast–northwest trending Lemeš Basin formed in the platform interior (see also Fig. 1c) at the onset of intense block-faulting and subsidence at the northeastern AdCP margin during the early Kimmeridgian. This is believed to have been influenced by oceanic spreading in the Vardar Ocean (Bucković et al. 2004, following Herak 1986). Normal-faulting and extensional subsidence resulted in the deposition of the deep-water bituminous limestones in this basin during a time of otherwise global eustatic sea-level fall (Bucković et al. 2004; Haq et al. 1988).

In contrast, Velić et al. (2002a, b and references therein) assumed that the newly formed intraplatform basins reflect a transpressional stress regime with the formation of small pull-apart basins and nearby uplift zones with emersion and facies differentiation.

In the type area of the Lemeš Basin in the Svilaja Mts. (Fig. 1), the deep-water Lemeš Formation (Lemeš-Schichten: Furlani 1910) records a continuous deepening above shallow-water limestones with the “stromatoporid” sponge Cladocoropsis sp. (Kerner 1907). The transition between the shallow-water limestones of the AdCP and the deep-water Lemeš Formation was dated by the presence of Kimmeridgian aged ammonoids in the siliceous deep-water limestones (Furlani 1910).

In the study area (Fig. 1), the Lemeš Formation has a slightly different lithological evolution due to the more central palaeogeographic position in the newly formed Late Jurassic basin. Furthermore, the transition from shallow- to deep-water limestones is older than in the type region. Based on the depositional history and facies of the Lemeš Formation as defined by Furlani (1910), the roughly time-equivalent siliceous bedded limestones of the Gorski Kotar Basin do not belong to the Lemeš Formation (see also Velić et al. 1994). Inclusion of the Gorski Kotar Basin limestones in the Lemeš Formation led to erroneous palaeogeographic and tectonic reconstructions (Blažeković Smojić et al. 2009).

Siliceous and organic-rich deposits are widespread within the sedimentary succession of the AdCP (see also Jacob et al. 1983; Sebecic and Ercegovac 1983; Barić et al. 1988; Moldowan et al. 1992; Mattavelli et al. 1993; Jerinic 1994; Cota and Baric 1998; Radovčić et al. 1983; Fiket et al. 2008). There is no evidence to correlate these organic-rich strata with global or Tethys-wide oceanic anoxic events (OAEs) (Jenkyns 2010). Therefore, they represent deposits in local, periodically restricted basins. This necessitates an in-depth investigation of the geodynamic, palaeoclimatic, and eustatic sea-level controls leading to deposition of the organic-rich sedimentary rocks in the Upper Jurassic Lemeš Basin.

The black shales within the Lemeš Formation were subject to preliminary sedimentological and geochemical investigations (Bulić and Trutin 1996; Fiket et al. 2008; Blažeković Smojić et al. 2009). However, detailed investigations into the vertical variation of the depositional environment during the basin evolution have not previously been undertaken.

To fill this gap, the organic-rich part of the Lemeš Formation was sampled at high resolution and various geochemical methods were applied to the sample set. The geochemical assessment was accompanied by organic petrographic analysis. The geochemical and petrographical results are complemented by microfacies analysis, which also include the under- and overlying strata. The comprehensive interpretation of all data provides information on the depositional environment, with significant implications for a new depositional model. Additional biostratigraphic results improve the age constraints of the main stages of the Lemeš Basin evolution. These data also provide new insights into the regional, Upper Late Jurassic, Dinaridic tectonic regime, which was caused by ophiolite obduction, and its effects on the evolution of intra-platform basins of the AdCP. Additionally, the hydrocarbon potential of the Lemeš Formation is quantified.

Geological setting

The investigated succession is located at the slope of Mt. Poštak (Fig. 1) in southwestern Croatia. The region consists of the Triassic-to-Lower Jurassic AdCP basement and Toarcian-to-Lower Cretaceous AdCP stratigraphic units. These represent the majority of the Outer Dinarides carbonate succession (Vlahović et al. 2005).

The Late Jurassic AdCP interior is known for its diverse facies and deep-water depositional environments (i.e., Lemeš and Gorski Kotar Basins; Vlahović et al. 2001; Fig. 1), which followed a long-lasting Middle Jurassic-to-Oxfordian shallow-water depositional environment (Tišljar et al. 1994; Velić et al. 1994, 2002b; Lawrence et al. 1995; Vlahović et al. 2005). The Lemeš Basin sediments are siliceous and organic-rich carbonate deposits restricted to deep intraplatform areas with low oxygen exchange (Tišljar et al. 1994; Velić et al. 1994, 2002b; Vlahović et al. 2005; Blažeković Smojić et al. 2009; Bucković and Markić 2016). The Kimmeridgian aged lower Lemeš Formation, known as the Bänderkalk, is characterized by thin-bedded and strongly fractured grey cherty limestone, intercalated by very fine-grained white platy limestones (Furlani 1910). The middle Lemeš Formation, known as the Fleckenkalk, is represented by thick-bedded, grey, and bioturbated limestones with irregular chert nodules. The highest stratigraphic unit at the type locality is described as a 20–30 m-thick succession of Tithonian, bedded, and bioturbated grey limestone with aptychi. The lateral continuation of the Lemeš Formation shows a gradual change from deep basin limestones to peri-reefal and reefal facies (Furlani 1910). The age dating of the Lemeš Formation at the type locality at Mt. Lemeš is based on (late) Kimmeridgian and early Tithonian ammonites, i.e., Oppelia sp., Perisphinctes sp., Aspidoceras sp., Haploceras sp., and Belemnoidea sp. (Furlani 1910).

Based on dasycladalean algae and foraminifera from the Lemeš Formation and age equivalent deposits in the Gorski Kotar Basin, Velić et al. (1994, 2002a, b), Bulić and Trutin (1996), and Velić (2007) confirmed a Kimmeridgian-to-early Tithonian age. However, these authors also emphasized the difficulties to define strict boundaries for distinct biozones.

The Lemeš Basin extends over 42 km2 (Fig. 1c). The 250–450 m-thick basin fill has been subdivided into different mappable layers (Blažeković Smojić et al. 2009 and references therein; Fig. 2). The lower part of the basin fill comprises thick limestone layers with chert lenses and nodules (layer 1) overlain by platy siliceous and cherty limestones (layer 2, Fig. 2). An overlying tuff layer (layer 3) records volcanic activity (Blažeković Smojić et al. 2009). The presence of bentonite suggests an average pH ~ 8 during diagenesis, typical for deep-marine intraplatform settings (Šegvić et al. 2006). Similar remnants of volcanic activity were evident from pyroclasts at the base of the Gorski Kotar Basin (Bucković and Markić 2016).

Fig. 2
figure 2

Subdivision of the Lemeš Formation into layers 1–9 and the original lithological description after Blažeković Smojić et al. (2009 and references therein). The authors do not provide a vertical scale, but the entire formation is 250–450 m thick

The succession changes upward into organic-rich mudstones (layer 4, Fig. 2), which are 2–70 m thick. The calcareous shales are dark brown–grey, olive–grey, dark grey, brown–black, and/or grey–black, and show mm-scale lamination. Bed thickness typically varies from 10 to 65 cm, but may reach 3 m. According to Blažeković Smojić et al. (2009), the dominant fossils assemblages include radiolarians, juvenile ammonoids, aptychi, calpionellids (sic! in Oxfordian–Kimmeridgian sedimentary rocks), and Saccocoma. The organic-rich layer 4 is overlain by siliceous limestones with ooids reworked from shallow-water areas (layer 5), grey laminated limestones (layer 6), and whitish laminated chert (layer 7, Fig. 2). The upper part of the Lemeš Formation includes thick, red, nodular, partly cherty limestones with detrital components and ooids (layer 8), and thin-bedded siliceous platy limestones (layer 9, Fig. 2) This reflect a shallowing upward trend (Blažeković Smojić et al. 2009 and references therein).

Geochemical investigations of the organic matter-rich layer 4 have been performed by Blažeković Smojić et al. (2009) using sample material from eight sites in the Lemeš Basin. The most important results are briefly summarized below: total organic carbon (TOC) contents up to 27 wt.% (typically 2–6 wt.%) and hydrogen index (HI) values between 253 and 637 mgHC/gTOC (average: 518 mgHC/gTOC) prove a good petroleum source potential. High sulphur (S) contents (8–11.5 wt.%) of kerogen concentrates suggest the presence of sulphur-rich kerogen type II-S. Based on Tmax values, layer 4 is thermally immature in the southern (411–416 °C) and northern part of the Lemeš Basin (424–425 °C) but reaches early oil window maturity in its middle part (435–437 °C). Biomarker analysis suggests marine microbes as dominant organic matter source (Bulić and Trutin 1996; Fiket et al. 2008; Blažeković Smojić et al. 2009). According to Blažeković Smojić et al. (2009), uniform geochemical features reflect similar depositional conditions and organic input in most parts of the Lemeš Basin.

Studied succession and samples

The investigated succession of the Lemeš Formation is located at the southeastern slope of Poštak mountain (GPS: 44°13′59.94″ N 16° 6′23.79″ E, see also Fig. 1), where it overlies Oxfordian lagoonal limestones (Tišljar and Velić 1993; Veseli et al. 2012). Here, the Lemeš Formation is about 270 m thick (Figs. 3, 4) and comprises the entire fill of the Lemeš Basin (Fig. 4a). Its lower part, the Rastičevo member (Veseli et al. 2012) is approximately 100 m thick and is characterized by silicified pack- and wackestones with open-marine organisms (ammonoids and radiolaria) in its upper part. The transition to the overlying Dimici member (Veseli et al. 2012; Figs. 3, 4a) is marked by two distinct claystone layers, interpreted in the field as silicified ash layers. Unit 2 is 55 m thick and consists of siliceous limestones with intercalated organic-rich layers (black shales). The overlying 110 m-thick Lemeš member (Veseli et al. 2012) above the black shales consists of limestone with ammonoids and cherty nodules, and is increasingly intercalated by calciturbidites. Near the top of the Lemeš member, siliceous packstones with shallower water organisms become dominant.

Fig. 3
figure 3

Lithological column of the Lemeš Formation in the investigated area and the macroscopic delineation of the suggested stratigraphic Units 1–3. The subdivision of Unit 2 (geochemical profile) is based on geochemical results

Fig. 4
figure 4

a Outcrop near Rastičevo, unit boundaries are indicated as well as a young fault, b intercalated bituminous limestone and siliceous organic-lean limestone of Unit 2, and c trace fossil (burrow) within Subunit 2a, showing that high-energy events may replace the anoxic conditions shortly and allow opportunistic species to invade

The main focus of the present study is on the organic matter-rich Dimici member (Unit 2). Hence, 185 samples (geochemical profile, Fig. 3), representing each of the exposed layers have been taken for geochemical analysis. Unfortunately, outcrop conditions prevented continuous sampling. Two additional samples were collected from the basal volcanic ash layers for XRD analysis. Rastičevo member (Unit 1) is recorded with 17 samples (geochemical profile, Fig. 3) for geochemical analysis and nine samples for microfacies analysis. The sample set of Unit 3 comprises four ammonite samples and eight samples for microfacies analysis. Two claystone layers have been selected for XRD analysis.

Analytic methods

Total carbon (TC) and sulphur (S) were determined for 202 samples using a Leco 300 CS analyser. Samples treated with hydrochloric acid to remove carbonate-bound carbon were analysed for TOC. The concentration of total inorganic carbon (TIC = TC − TOC) was used to determine the percentage of calcite equivalent (Ceq. = TIC × 8.33). Rock–Eval pyrolysis was performed on a Delsi Rock Eval II + instrument to determine the amount of free hydrocarbons (HCs) (S1: mg HC/g rock), HCs generated during gradual heating (S2: mg HC/g rock), and the temperature of maximum HC generation (Tmax). S1 and S2 were used for the calculation of the hydrogen index (HI = [S2/TOC] × 100), the petroleum potential (PP = [S1 + S2]), and the production index (PI = S1/[S1 + S2]) (Lafargue et al. 1998).

Thirty-one rock samples selected for biomarker analysis were extracted with dichloromethane in a Dionex ASE 200 accelerated solvent extractor (75 °C; 50 bar). After evaporation of the solvent in a Zymark TurboVap 500 closed cell concentrator, asphaltenes were precipitated from a hexane–dichloromethane solution (80:1) and separated using centrifugation. The hexane soluble fraction was separated into NSO compounds, saturated hydrocarbons, and aromatic hydrocarbons by medium-pressure liquid chromatography with a Köhnen–Willsch instrument (Radke et al. 1980). The saturated and aromatic hydrocarbon fractions were analysed with a gas chromatograph (GC) fitted with a 30 m DB-1 fused-silica capillary column (i.d. 0.25 mm; 0.25 μm film thickness) coupled to a Finnigan MAT GCQ ion trap mass spectrometer for a first sample set and with a 30 m DB-5MS fused-silica column (i.d. 0.25 mm; 0.25 μm film thickness) coupled to a Thermo Fischer IS dual-quadrupole mass spectrometer for a second set. The analysis was performed with helium carrier gas and a heating rate of 4 °C/min between 70 and 300 °C, and a final isothermal period of 15 min. The scan range of the GC in the electron ionisation mode was set from m/z 50–650 (0.7 s total scan time). Absolute concentrations normalized to TOC, and relative percentages of distinct saturated and aromatic compounds were identified from retention time in the total ion current chromatograms and calculated from peak areas in accordance to the internal standards (saturated HC: deuterated n-tetracosane, aromatic HC: 1,1′-binaphthyl) or by peak areas in the according mass chromatograms, corrected by response factors for the intensities of the fragment ion.

Inorganic carbon (δ13Ccarb) and oxygen isotope (δ18Ocarb) ratios were carried out for 202 samples. Measurements were performed by adding 100% H3PO4 to samples heated at 70 °C in an online system (Gasbench II with carbonate option). Analysis was performed using a Thermo Fisher DELTA V isotope ratio mass spectrometer (Delta V IRMS). The results were calibrated to the Vienna-Pee Dee Belemnite (VPDB) standard for δ13Ccarb (+ 1.95‰ to NBS19 reference material) and δ18Ocarb (− 2.20‰ to NBS19 reference material), with analytic reproducibility better than 0.2‰.

For organic carbon isotope (δ13Corg) analyses of organic-rich limestones, the samples were weighed into tin foil shuttles and combusted in an excess of oxygen using an elemental analyser (Flash EA 1112) at 1020 °C. The resulting CO2, after separation by column chromatography, was analysed online by the DELTA V IRMS. The 13C/12C isotopic ratios of the sample CO2 were compared with a monitoring gas calibrated to the V-PDB standard using NBS-19 reference material. Reproducibility of the whole analysis is in a range of 0.1–0.2‰.

All isotope ratios are reported in the delta (δ) notation in parts per mil (‰) in reference to the VPDB standard.

X-ray diffraction (XRD) was performed by a Philips X’Pert3 diffractometer with ten hand-mortared bulk samples on a side and top filled mount with a sample thickness of ~ 10 μm. The measurement was set from 2 to 70° 2θ. A rough qualitative analysis of diagenetic calcite 29.84° and 31.24° 2θ, quartz 20.84° and 36.54° 2θ, and clay minerals was established, and a semi-quantitative analysis was performed according to an intensity factor in counts per second per 100 percent after Schultz (1964).

Fifty-five thin sections and 12 polished blocks were analysed using a Leica transmitted light microscope for microfacies and a Leica MPV microscope for qualitative maceral analysis and vitrinite reflectance measurements. Both microfacies analysis (Flügel 2004) and reflectance measurements (Taylor et al. 1998) followed established procedures.

All analysis was performed at Montanuniversitaet Leoben. For standard deviation of different methods, see Agahayeva et al. (2021).

Results

Lithology, microfacies, and mineralogy

The Lemeš succession in the study area is composed of three different units, distinguished on the basis of their lithology, microfacies (Fig. 5), and mineralogy.

Fig. 5
figure 5

a Bioturbated wackestone with crinoids, brachialia of Saccocoma Agassiz, foraminifera, and carbonate clasts, b bioturbated wackestone with few recrystallized radiolarians (lower part) and peloidal packstone (upper right) from the basal part of Unit 3, c bioturbated radiolaria wackestone with organic-rich layers and a concretion filled with organic matter, d bioturbated radiolarian wackestone with rare thin shells, e laminated radiolaria wacke- to packstone with bigger OM components, and f packstone with ammonoid shell and shell fragments

Lemeš Unit 1 (0–104 m) is dominated by brownish to light-greyish, thin-bedded (up to 15 cm) wackestones and thick-bedded (up to 50 cm) packstones with shallow-water debris, large benthic foraminifera, and shell fragments in the basal 60 m. Strong bioturbation and absence of shedding features are evident within all samples. The samples are free of deep-marine assemblages (i.e., radiolarian and ammonoids) and have low OM contents. One sample shows accessory amounts of pyrite. The facies of this section is typical for intra-platform environments (Flügel 2004). Upsection, a transition to radiolarian-rich, thick-bedded (up to 50 cm) pack- and wackestones, accompanied by ammonoids (Fig. 5f) and shell fragments is observed. The majority of the samples have better OM preservation compared to the basal packstones. Bioturbation is still prevalent. The radiolaria and ammonoid assemblages suggest a deep-marine setting (Flügel, 2004).

Lemeš Unit 2 (104–154 m) consists of dark-brown, organic- and radiolarian-rich, siliceous, turbiditic packstones, grainstones, and bioturbated wackestones with rare thinned-shelled organisms and ammonoid fragments. OM is dominantly bound to thin-bedded (1–46 cm) and laminated (1–10 mm) limestones consisting of fine-grained, low-density turbidites (Fig. 5c–e). OM is also found within larger bioclasts. The silicified parts of the succession are relatively poor in OM due to early diagenetic silicification. These organic matter-rich deposits are typical of restricted basins (Flügel 2004).

XRD analysis at the base of Unit 2 shows the presence of 100 wt.% calcite in the lower layer and 81 wt.% quartz and 19 wt.% calcite in the upper layer. Because clay minerals are below the detection limit, the nature of volcanic ashes could not be determined. Upsection, the mineral composition of the siliceous limestones is represented by 93–99 wt.% calcite and 1–7 wt.% quartz. One sample shows a small ankerite peak, which could not be quantified.

Lemeš Unit 3 (154–257 m) displays crinoid- and foraminifera-rich, greyish, thin-bedded (up to 25 cm) wackestones with gradually increasing shallow-water influence. Radiolaria and filaments are still present but are less abundant than in the underlying Unit 2. Moreover, the organic matter content is reduced. The onset of greater resediment input, calciturbidites, and increasing bioturbation is noted upsection. The facies is accompanied by more frequently preserved ammonoids. Ataxioceras sp. (Fig. 6) was found about 20 m above the base of Unit 3. Ataxioceratinae are specific for the lower Kimmeridgian Tethyan ammonite zone A. hypselocyclum (Atrops 1982; Moliner and Olóriz 2009, 2010; Wierzbowski 2017). Ataxioceras sp. is rare in the uppermost layers of the underlying S. platynota zone (Atrops 1982). Additionally, a Saccocoma Agassiz brachialia/arm element (Fig. 5b) typical for Late Jurassic strata (Lefeld and Radwanski 1960; Flügel 2004) was identified. This unit shows typical sedimentary and stratigraphic features of slope or toe of slope environments (Flügel 2004). In some samples, large pores are filled with migrated, light-grey bituminous matter.

Fig. 6
figure 6

Ataxioceras sp. dated to the lower Kimmeridgian A. hypselocyclum zone from the lowermost part of Unit 3

Bulk geochemical parameters

The vertical variation of geochemical parameters is shown in Fig. 7. The full geochemical dataset is presented in the Appendix. Initially, the geochemical data for siliceous limestone and bituminous limestone samples are presented separately. Afterwards, the vertical variation of parameters will be presented for each unit.

Fig. 7
figure 7

Vertical plots of bulk geochemical parameters: a Cceq. (calcite equivalent percentages), b TOC (total organic carbon), c HI (hydrogen index). wt.% weight percent, d S (sulphur) content, and e TOC/S ratio; the defined units and the lithotypes within the profile are indicated on the left

Siliceous limestone

Siliceous limestone samples typically contain very high calcite contents (avg. Cceq. 95.8 wt.%; Fig. 7a) and low amounts of organic matter (average TOC: 0.43 wt.%; max. TOC: 1.1 wt.%; Fig. 7b). HI values vary strongly (7–609 mg HC/g TOC; Fig. 7c). HI variations in organic-lean sediments are often either due to the mineral matrix effect (e.g., Langford and Blanc-Valleron 1990) or due to the presence of inert and reactive organic matter (e.g., Dahl et al. 2004) (Fig. 8).

Fig. 8
figure 8

Cross-plot of HI vs. Tmax after Espitalié et al. (1984)

A plot of Rock–Eval S2 versus TOC shows a good correlation (r2 = 0.95; Fig. 9a). The slope of the regression line (5.39) and its x-intercept (0.14 wt.%) determine either the true HI (539 mg HC/g TOC) and the mineral matrix effect (e.g., Langford and Blanc-Valleron 1990) or the HI of the reactive organic matter and the amount of inert organic carbon (e.g., Dahl et al. 2004). In either case, Fig. 9a shows that the (reactive) organic matter is a hydrogen-rich type II kerogen. Sulphur content is very low in all samples (0.03–0.20 wt.%, Fig. 7d). Consequently TOC/S ratios are relatively high (average 4.6, max 8; Fig. 7e), except the uppermost part of the succession. Tmax values (average 426 °C) indicate that the organic matter is immature. Within this context, PI values are surprisingly high (average: 0.19).

Fig. 9
figure 9

Cross-plots of S2 vs. TOC after Espitalié et al. (1984) with regression lines for a siliceous limestones, and b. bituminous limestones. The slopes of the regression lines (5.29 and 5.58) show that both siliceous and bituminous limestones contain hydrogen-rich organic matter with HI values in the order of 550 mg HC/g TOC

Bituminous samples

Bituminous samples consist of limestones with 75–98 wt.% Cceq. and some mudstones, calcareous mudstones, and marlstones with 0.7–48.8 wt.% Cceq. Average TOC contents (3.57 wt.%) and HI values (502 mg HC/g TOC) are high. The slope and x-intercept of the regression line in the S2-TOC diagram (Fig. 9) show that the HI of the reactive organic matter (true HI) is 557 mg HC/g TOC and the amount of inert organic carbon is 0.24 wt.%. This indicates that there is no difference between the reactive organic matter of bituminous and siliceous limestone samples. Sulphur contents (average 0.60 wt.%) are slightly higher than in siliceous limestone samples, but TOC/S ratios are in the same order (average 6.1, from 3.8 – 7.9). Both, Tmax (average 429 °C) and PI values (average: 0.03) indicate low thermal maturity.

Unit 1 The uppermost 25 m of the Unit 1 have been included in the geochemical study. Siliceous limestones with low TOC contents (0.15–0.95 wt.%) prevail in the studied interval with subordinate layers of bituminous limestone with relatively high TOC contents (max. 2.68 wt.%).

Unit 2 Based on the geochemical parameters, three subunits were defined within the Lemeš Unit 2 (from base to top: Subunit 2a, 2b, and 2c).

Subunit 2a (24.5–33.5 m) includes a high number of organic matter-rich rocks (bituminous limestones and some low-carbonate rocks) with TOC contents ranging from 1.22 to 12.63 wt.% and relatively high sulphur contents (0.05–2.01 wt.%). HI values are very high as well (346–609 mg HC/g TOC), indicating the presence of type II kerogen. Siliceous limestone layers are rare. Consequently, average TOC content (4.60 wt.%) of Subunit 2a is higher than that of any other unit. The strongly varying carbonate contents reflect the alternating lithotypes of this unit.

Subunit 2b (33.5–61.5 m) contains a high number of siliceous limestone layers. TOC contents of bituminous samples are generally more uniform and range from 1.09 to 5.88 wt.%. The average TOC of Subunit 2b is only 1.71 wt.% reflecting the high number of siliceous limestone layers.

Subunit 2c (61.5–75.5 m) contains more bituminous limestone layers than the underlying Subunit 2b, especially in the lower part. The TOC content of the organic-rich layers is similar to that in Subunit 2b (0.71–6.31 wt.%; avg.: 1.98 wt.%). TOC/S ratios of siliceous limestone layers show an upward decrease from values above 5 to values in the order of 0.94. In contrast, TOC/S ratios of bituminous limestone layers show an opposite trend and increase from 5.0 to 7.6.

Organic petrology

Representative photographs of polished blocks from Units 1 and 2 are shown in Figs. 10 and 11. Rocks in the upper part of Unit 1 contain small amounts of organic matter dominated by detrovitrinite and inertodetrinite (Fig. 10). Pyrofusinite is present in very low amounts (Fig. 10a).

Fig. 10
figure 10

Reflected white light microphotographs from a, b Lemeš Unit 1. vit vitrinite, fus fusinite

Fig. 11
figure 11

Reflected white light (left) and fluorescence mode (right) microphotographs from Lemeš Subunits 2c (a, b), 2b (c), and 2a (d, e). lam lamalginite, tel telalginite, vit vitrinite, in inertinite, py pyrite, bit secondary bitumen, fus fusinite

Organic matter in the bituminous limestones of Subunit 2a is dominated by lamalginite, and contains minor amounts of telalginite and inertodetrinite (Fig. 11d,e). Fusinite and accessory detrovitrinite are also present. Organic matter in Subunit 2b is also dominated by lamalginite, but shows increased amounts of detrital vitrinite and inertinite (Fig. 11c). Organic matter in the upper Subunit 2c contains high amounts of lamalginite and telalginite (Fig. 11a, b), and also has greater amounts of terrestrial macerals (dominant fusinite and subordinate detrovitrinite). All Lemeš Unit 2 samples include small amounts of micropyrite (diameter: 2–8 µm) crystals, which do not show any signs of weathering (e.g., Fig. 11a–c).

Apart from primary macerals, small amounts of secondary bitumen are noticed along joints, in pores and parallel to lamalginite layers in Unit 2 samples (e.g., Fig. 11a–c, e). All samples show strong background fluorescence. This may result from a high amount of amorphous OM.

Vitrinite reflectance measurements were difficult because of the low amount of vitrinite, its often very small size, and the presence of bituminite. Moreover, vitrinite reflectance may be reduced in liptinite-rich samples (e.g., Peters et al. 2018). Vitrinite reflectance of samples Cro 40 (0.39 ± 0.06; n = 36), Cro 93 (0.47 ± 0.07; n = 21), and Cro 223 (0.52 ± 0.08; n = 29) vary significantly, and have a high standard deviation. Particles interpreted as secondary bitumen show reflectance values between 0.28 and 0.38%.

Stable isotope composition of carbonate minerals and organic matter

The vertical trend of stable carbon (δ 13C) and oxygen (δ 18O) isotope ratios of bulk carbonate and the carbon isotope ratio of organic matter are presented in Fig. 12.

Fig. 12
figure 12

Vertical variations of a the inorganic carbon isotope composition (δ13Ccarb; VPDB), b the oxygen isotope composition (δ18Ocarb; VPDB) of carbonate minerals, and c the isotope composition of organic matter (δ13Corg; VPDB) of bituminous limestone samples. The defined units and the lithotypes within the profile are indicated on the left

Stable carbon isotope ratios of carbonate minerals (δ13Ccarb) range from − 1.36 to + 3.50‰VPDB. They are typically lower in siliceous limestone samples than in bituminous limestone samples. The highest differences (up to 1.5‰) are observed in Subunit 2a and the lower part of Subunit 2b, as well as near the border between subunits 2b and 2c. Despite these differences, the vertical trend of both lithotypes is similar. The upper part of Unit 1 shows an upward decrease in δ13Ccarb values (3.50–1.76‰ VPDB) independent of lithology. Subunit 2a is characterized by the lowermost δ13Ccarb values, which is especially valid for siliceous limestone, which show negative and relative uniform values (− 0.17 to − 1.36‰VPDB). In contrast, values from laminated limestones are strongly oscillating (1.59 to − 1.18‰VPDB). Subunit 2b has low values in the lower and upper part (minimum: − 0.66‰ VPDB) and higher values in the middle part (maximum: 2.13‰ VPDB). Subunit 2c starts with low δ13Ccarb values and shows an increasing upward trend (− 0.48 to 2.43‰ VPDB). Similar values for both lithotypes are observed in the upper part of Subunit 2c.

In contrast to δ13Ccarb, δ18Ocarb values are relatively uniform (− 4.18 to − 1.70‰ VPDB). Moreover, with the exception of Subunit 2c, where δ18Ocarb values in siliceous limestones are less negative that in bituminous limestones, differences between values from both lithotypes are minor. Two negative outliers are observed in bituminous samples in the lower part of Subunit 2a.

Organic carbon isotope values (δ13Corg) of Unit 1 are consistently high (− 26.50 to − 26.20‰ VPDB). The basal bituminous layer of Subunit 2a shows the minimum δ13Corg value (− 29.90‰ VPDB) as well as the minimum δ13Ccarb value. Above the minimum, an abrupt increase in δ13Corg values is visible with high values often between − 27.00 and − 26.00‰ VPDB. This interval is followed by more negative δ13Corg values (− 28.20 to − 27.00‰ VPDB). δ13Corg values near the base of Subunit 2b are low (≤ − 28.00‰), but the main part of Subunit 2b is characterized by a decreasing upward trend from − 27.0 to − 28.4‰. Subunit 2c is characterized by rather uniform δ13Corg values (− 28.50 to − 27.70‰ VPDB).

Molecular composition of hydrocarbons

The full data set of absolute concentrations and derived ratios (Fig. 13) are presented in the Appendix. The n-alkane distributions are characterized by high amounts of short-chain (avg.: ∑n-C15–20/∑n-C15–35 = 0.41) and mid-chain n-alkanes (avg.: n-C21–25/∑n-C15–35 = 0.36), whereas long-chain n-alkanes typically occur in lower concentrations (avg. ∑n-C26–32/∑n-C15–35 = 0.22). However, long-chain n-alkanes dominate in one sample from Subunit 2c (0.75). The carbon preference index (CPI; as defined by Bray and Evans 1961) varies in the narrow range from 1.09 to 1.45.

Fig. 13
figure 13

Vertical variations of a n-alkane distribution of short-, mid-, and long-chain n-alkanes, b CPI carbon preference index, c pristane/phytane (Pr/Ph) ratio, d Σsterane/Σ17-hopane ratio, e C28-18α-trisnorneohopane/C27-17α-trisnorhopane (Ts/Tm) ratio, and f dibenzothiophene (DBT)/phenanthrene ratio (empty squares indicate samples with high concentrations of dibenzothiophene, but no phenanthrene). The defined units and the lithotypes within the profile are indicated on the left

The ratio of the acyclic isoprenoids pristane (Pr) and phytane (Ph) is very low in Unit 1 (≤ 0.2) and higher, as well as more variable, in Subunit 2a (0.35–1.33). In Subunit 2b, Pr/Ph ratios decrease upwards from 0.96 to 0.20. Variable, but generally small ratios (0.17–0.61) are observed in Subunit 2c.

C16–C18 aryl isoprenoids were detected in low amounts in all samples (0.1–4.1 µg/g TOC). Some samples also contain C14–C21 aryl isoprenoids. Concentrations of C16–C18 aryl isoprenoids tend to increase upwards, although some high values also occur in the Unit 1 and Subunit 2a.

C29 steranes (30.8–52.5%) are more abundant than C27 (25.0–39.2%) and C28 steranes (20.4–39.5%) (Fig. 15). The percentage of C29 steranes shows a subtle upward decrease. Hence, the difference between the amounts of C27, C28, and C29 steranes is relatively low in Subunit 2c.

Diasteranes and 4-methylsteranes are generally absent. C21-pregnane and C22-homopregnane are present in significant amounts (max. 19.61 μg/g TOC and 18.09 μg/g TOC, respectively), but do not show any stratigraphic trend. Norcholestane (up to 1.02 μg/g TOC) has only been observed in some samples. Monoaromatic steroids (MAS) are present in all samples (up to 46.64 μg/g TOC), whereas triaromatic steroids (TAS) occur in small concentrations and are absent in many samples. The ααα-C29 sterane 20S/(20S + 20R) isomerization ratios (average: 0.49) are slightly below the equilibrium value 0.54. Relatively low isomerization ratios are observed in Subunit 2a. The ααα/(ααα/αββ) C29 steranes ratio is high (0.52–0.73).

The steranes/hopanes ratios are low (0.14–0.56; average: 0.27). In Subunit 2b, the steranes/hopanes ratio decreases upwards, whereas an increasing upward trend is observed in Unit 2c (Fig. 13). Similar to the sterane isomerization ratio, the αβ-C32 hopane 22S/(22S + 22R) isomerization ratios are close to the equilibrium value of 0.61 (average 0.58).

Gammacerane is absent. The maximum concentration of 28,30-bisnorhopane is 39.06 μg/g TOC, but most samples contain low amounts. Benzohopanes (max. 28.09 μg/g TOC) and C31-hopene (max. 23.53 μg/g TOC) are found in some samples only. The ratios of C28-18α-Trisnorneohopane (Ts) and C27-17α-Trisnorhopane (Tm) are very low (avg. 0.06), even though Subunit 2a and Subunit 2b show values up to 0.18 (Fig. 13).

The methylphenanthrene index (MPI-1) (Radke and Welte 1981) could be determined for some samples and ranges from 0.30 to 0.49, with a single exception (0.67). A slight upward decrease is observed. Dibenzothiophenes (DBT) are recorded in all samples in significant amounts. In contrast, phenanthrenes (Ph) occur generally in low concentrations and could not be quantified in all samples. Consequently, DBT/Ph ratio are very high, especially in Unit 2 samples. Samples with high concentrations of methyldibenzothiophenes show methyldibenzothiphene ratios (MDR) (Radke 1988) between 0.60 and 1.12 (average 0.88). Alkylbenzenes are noted in some samples with concentrations up to 66.31 μg/g TOC.

The gymnosperm specific biomarkers bis-nor-abietatriene (max. 2.48 μg/g TOC) and nor-abietatriene (max. 1.91 μg/g TOC) are found in some samples from Unit 2.

Discussion

Stratigraphic evolution

The first appearance of open-marine organisms (i.e., radiolarians, filaments, and ammonites) is recorded in Unit 1 (Fig. 14) above Oxfordian shallow-water limestones. These Oxfordian shallow-water deposits, dated by foraminifera (Velić 2007), are of lagoonal facies with diverse benthic organisms, strong bioturbation, and relatively high micritic carbonate production (Dragičević and Velić 1994; Velić et al. 1994, 2002a; Tišljar et al. 1994).

Fig. 14
figure 14

The depositional evolution of the Lemeš Basin, tectonically controlled in the uppermost Oxfordian and eustatic driven from the Kimmeridgian transgressive system tract (TST) and Kimmeridgian-to-Tithonian highstand system tract (HST)

The radiolarian packstones of this microfacies contain components derived from shallow-water environments, which, as seen in the type region (Furlani 1910), must have been within the proximity of the Lemeš Basin. Furthermore, evidence for nearby vegetated land is given by the occurrence of gymnosperm biomarkers and vitrinite particles in Unit 2. Unit 1 records the first accumulation of high amounts of OM (and sparse pyrite), which is probably due to periodically decreasing oxygen contents within the basin. The changes within Unit 1 reflect rapid rise of relative sea level due to increasing subsidence. The shallow-water organisms at that time were no longer able to counterbalance the newly created accommodation space by carbonate production. Hence, shallow-marine carbonates were restricted to the rims of the newly formed basin (Furlani 1910). The basin received little carbonate material from these shallow-water areas and became periodically restricted near the bottom.

The transition to Unit 2 is marked by a change to organic-rich bituminous limestones intercalated into often-laminated, siliceous limestones. This transition could be related to a volcanic event at the base of Unit 2 (Blažeković Smojić et al. 2009). The organic-rich Unit 2 consists of radiolarian wacke- and packstones with high organic matter content. Most of the samples are non-bioturbated and their depositional structures, organic-rich layers, and pyrite are well preserved. These are indications for a further decrease of the free oxygen, and the evolution of an anoxic depositional environment, which was likely due to water column stratification. The sedimentary structures indicate deposition below the fair-weather wave base. However, rare trace fossils in organic-rich bituminous limestones in Subunit 2a (Fig. 4c) indicate at least periodical open basin conditions with increased oxygen.

An abrupt decrease in preserved organic matter is noted in Unit 3, accompanied by a more open-marine microfacies and more oxygenated depositional conditions. This transition is dated by Ataxioceras sp. to the early Kimmeridgian Hypselocyclum zone (Moliner and Olóriz 2010), which agrees with the presence of Saccocoma Agassiz indicating a Kimmeridgian-to-early Tithonian age. The decrease in organic matter, ammonite fragments and radiolarians, and an increase in shallow-water carbonate components, accompanied by a crinoid and foraminifera fauna, reflect a change in the depositional environment with increased shallow-water carbonate production near the rims of the basin. Enhanced shedding of shallow-water carbonates resulted in a shallowing of the basin and more oxygenated conditions. However, a connection to a deep-marine realm must have existed during the early phase of the deposition of Lemeš Unit 3, which served as migration pathway for ammonites and radiolarians.

Source of organic matter and its amount

Despite varying HI values in siliceous limestone samples (Fig. 8), reconstructions of true HI values (Langford and Blanc-Valleron 1990) or HI values of the reactive organic matter (Dahl et al. 2004) show that the organic matter in siliceous and bituminous limestone samples is similar and contains high HI values (~ 550 mg HC/g TOC). This classifies the organic matter as hydrogen-rich type II kerogen.

This classification is in agreement with maceral composition, which is dominated by alginite and liptodetrinite (Fig. 11) derived from algal and bacterial biomass. Lamalginite reflects algal and microbial mats, whereas telalginite is typically derived from planktonic algae. Very low steranes/hopanes ratios (Fig. 13d) show that microbial biomass contributed significantly to the aquatic organic matter (Moldowan and Fago 1986; Thiel et al. 1999; Peters et al. 2005). High amounts of aquatic organic matter are also reflected by n-alkane distributions dominated by short- and mid-chain n-alkanes (Ficken et al. 2000; Fig. 13a).

Terrestrial macerals (vitrinite and inertinite) occur in minor amounts and are important only in samples with low TOC contents (e.g., in Unit 1; Fig. 10). Hence, percentages of long-chain n-alkanes and CPI values are low (Eglinton and Hamilton 1967; Fig. 13a, b). Despite low land plant input, relative proportions of C29-steranes are high (Fig. 15). This shows that land plants are not the only precursor of C29-steranes. Specific bacterial assemblages may contribute to C29-steranes (e.g., Moldowan et al. 1985; Volkman 1988; Affouri et al. 2013), especially in carbonate source rocks (e.g., Grantham 1986; Buchardt et al. 1989). Gymnosperm-derived biomarkers (nor-abietatriene, bis-nor-abietatriene) are present in only a few samples. Rare pyrofusinite may have a wind-blown origin.

Fig. 15
figure 15

Ternary plot of relative proportions of C27, C28, and C29-steranes

HI values in Unit 2 show only minor variation (Fig. 7c). This indicates that the organic matter type is uniform and that preservation conditions did not vary. Hence, very high TOC content in Subunit 2a probably reflects increased bioproductivity. For samples with carbonate content exceeding 80 wt.% Cceq., there is a general negative relation between TOC and Cceq. (r2 = 0.62), suggesting that apart from bioproductivity, organic matter richness is also controlled by dilution by carbonate minerals.

Redox conditions and sulphur supply

Pr/Ph ratios are a widely used redox parameter (Didyk et al. 1978). The observed very low ratios, which exceed 1.0 in only two samples from Subunit 2a (Fig. 13b), suggest strictly anoxic conditions. These varied only temporarily to dysoxic conditions during deposition of Subunit 2a and the basal part of Subunit 2b. Anoxic-to-dysoxic conditions are also suggested by a cross-plot of the pristane/n-C17 vs phytane/n-C18 ratios (Fig. 16). Moreover, the presence of aryl isoprenoids suggests that anoxic conditions, at least temporarily, even reached into the photic zone.

Fig. 16
figure 16

Plot of pristane/n-C17 alkane vs. phytane/n-C18 alkane ratios

Very high DBT/Phenanthrene ratios indicate high concentrations of free H2S in the water column (Hughes et al. 1995), especially during deposition of Unit 2. However, it has to be emphasized that high ratios are caused by low phenanthrene rather than by very high DBT concentrations. In the Pr/Ph vs DBT/Phenanthrene diagram (Hughes et al. 1995; Fig. 17), the samples plot into the field of strongly oxygen-depleted marine carbonate environments. These environments are often iron-limited. Therefore, sulphur is preferentially incorporated into the organic matter. This is also supported by high sulphur content in kerogen concentrates from other Lemeš locations (Blažeković Smojić et al. 2009).

Fig. 17
figure 17

Cross-plot of pristane/phytane vs. DBT/phenanthrene ratios after Hughes et al. (1995). The position of the samples agrees with a strongly oxygen-depleted marine carbonate environment

Total sulphur content and resulting TOC/S ratios provide a proxy for the distinction between marine and (sulphate-limited) freshwater and between oxic and anoxic conditions (Berner and Raiswell 1983). TOC/S ratios are in average 5.46 and exceed 2.8, the value considered typical for oxic marine environments (Berner and Raiswell 1983) considerably. Only some siliceous limestones in Subunit 2c; Fig. 7) are close to 2.8. Considering the presence of radiolarians and ammonites, a brackish or freshwater environment can be excluded for the Lemeš Basin.

Weathering of the outcrop samples is a possible explanation for low sulphur contents in the studied sections. However, the presence of fresh pyrite in Unit 2 samples (Fig. 11) does not support strong weathering. Thus, it is likely that sulphate concentration in the depositional environment was not very high, but because Fe-availability in the carbonate environment was limited, most available sulphur was incorporated in the organic matter. The dominance of very small-sized framboidal pyrites (below 4 µm) may indicate its formation within the water column (Wilkin et al. 1996; Liu et al. 2019).

Gammacerane, a proxy for salinity stratification (Sinninghe Damsté et al. 1995), is absent in all samples. However, some samples from Units 1, 2a, and 2b contain pregnane (max. 19.61 μg/g TOC) and homopregnane (max. 18.09 μg/g TOC) in significant amounts, which have been considered as possible markers for hypersaline environments (Ten Haven et al. 1986; Peters et al. 2005).

In contrast to geochemical proxies, significant bioturbation in Unit 1 indicates oxygen concentrations that allowed life at the sea floor. Bioturbation is also observed in some layers in Unit 2 (e.g. Figure 4c). This suggests significant variations of redox conditions. Only organic matter-rich layers, which were deposited in dysoxic to anoxic environments, have been selected for geochemical analysis.

Despite varying oxygen conditions during deposition of Subunit 2a, TOC contents reach a maximum of 12.63 wt.% in this unit. This reflects high organic productivity of marine organisms in the photic zone, whereas the good preservation reflects oxygen depletion in the deeper part of the water column.

Isotope composition, correlation, and paleoenvironmental effect

δ13Ccarb values (− 1.36 to + 3.50‰VPDB) are in the normal range of data from other Upper Jurassic AdCP sections (− 3.9 to + 4.3‰VPDB; Fiket et al. 2008). They also fit within the gross global average of open-marine carbonates (Marshall 1992; Jenkyns et al. 2002) and follow the trend in Kimmeridgian datasets (Weissert and Erba 2004).

Vertical δ13Ccarb trends of organic matter-rich and organic matter-poor limestones are similar, but δ13Ccarb values from organic matter-rich limestones are up to 2‰ VPDB higher (Fig. 12). The enrichment of 13C in these limestones may be a result of 12C depletion in the surface water due to strongly increased photosynthesis (Saltzman and Thomas 2012). Thus, the difference in δ13Ccarb values may reflect variations in bioproductivity. Interestingly, the difference reaches maxima in Unit 2a and the lower part of Unit 2c, which contain very high percentages of bituminous limestones.

In the following discussion, δ13Ccarb trends are used to determine the age of the studied succession. The basal part of Unit 3 has been dated into the A. hypselocyclum ammonite zone. Therefore, the isotope trends of the underlying units (upper part of Unit 1 and Unit 2) are compared in Fig. 18 with those from the uppermost Oxfordian and lower Kimmeridgian hemipelagic limestone from the Alpine Atlantic (Missoni and Gawlick 2011) in Spain (Sierra de Lugar and Cortijo del Cardador, O’Dogherty et al. 2018). Indeed, the δ13Ccarb trends in the Lemeš Basin match reasonably well with those from the western Alpine Atlantic margin in Spain (Fig. 18). However, δ13Ccarb values in the Lemeš Formation show much stronger variations (− 1.36 to + 3.50‰VPDB) than in the hemipelagic sections, where δ13Ccarb values fall into the narrow range of 2.3–2.9‰VPDB. The stronger negative excursions are probably an effect of limited exchange with the open oceanic carbon reservoir in the (semi-) restricted intraplatform Lemeš Basin. As pointed out by Saltzman and Thomas (2012), the restricted circulation allows larger amplitude and/or higher frequency in δ13Ccarb variability than the global ocean reservoir.

Fig. 18
figure 18

Correlation of the δ13C isotope excursion trends of ammonite dated lower Kimmeridgian sections in Spain: a Sierra de Lugar, b Cortijo del Cardador after O’Dogherty et al. (2018), to the studied geochemical profile below the A. hypselocyclum zone, c δ13Ccarb of siliceous limestones, d δ13Ccarb of bituminous limestones, and e the eustatic sea-level scaled to the correlated ammonite zones after Haq (2018)

The best fit between the δ13Ccarb trends in the well-dated pelagic sections (Fig. 18a, b) and the Lemeš Basin (Fig. 18c, d) is observed in the lower Kimmeridgian succession. In the pelagic sections, the lower Kimmeridgian includes an interval with decreasing upward δ13Ccarb values (Bimammatum ammonite zone), a minimum near the Bimammatum/Planula zone boundary, a relative maximum within the Planula ammonite zone (Planula/Galar subzones boundary), and an upward trend with decreasing δ13Ccarb values. This suggests that the Unit 2a was deposited during the late Bimammatum (Hauffianum subzone) and the early Planula ammonite zones (Planula subzone). Unit 2b represents the remaining part of the Planula ammonite zone. Unit 2c is tentatively attributed to the Platynota ammonite zone, although the observed trend with increasing upward δ13Ccarb values is not clearly recorded in the pelagic sections. The upper part of the underlying Unit 1 shows a distinct decreasing upward trend. This may correlate with a similar trend in the Sierra de Lugar section (Berrense subzone of the Hypselum zone; Fig. 18a).

δ18Ocarb values in Units 1 and 2 typically fall into the narrow range of − 2.5 to − 3.5‰ VPDB. These values are similar to those recorded for other locations with Lemeš Formation (Fiket et al. 2008), but about 1.5‰ lower than in Oxfordian and Kimmeridgian rocks in other Tethyan localities (Weissert and Erba 2004). Two outliers with highly negative values near the base of Unit 2 are assumed measurement errors. Apart from those, δ18Ocarb values show a subtle decreasing upward trend. It remains speculative, if this trend reflects the warming following a late Oxfordian-to-early Kimmeridgian cooling phase (Weissert and Erba 2004), or if other factors (e.g., salinity variations, paleoclimate, volcanic activity, and diagenetic processes; Hallam, 1984; Weissert and Erba 2004) contributed to the observed δ18Ocarb trend.

The average δ13C value of organic matter (δ13Corg) is − 27.63‰VPDB conforming to other Upper Jurassic AdCP data (Fiket et al. 2008). Unit 1 has uniform δ13Corg values (avg. − 26.28‰VPDB), whereas Units 2a and 2b show a trend towards more negative values. Low, but uniform δ13Corg values (avg. − 29.55‰VPDB) are observed in Units 2c. A slight trend towards more negative values is also recognized on a global scale (Jenkyns et al. 2002). In the Lemeš Basin, the onset of this trend correlates with the base of the organic matter-rich Unit 2. Largely parallel δ13Corg and Pr/Ph trends in Unit 2 suggest that δ13C depletion might be related to redox conditions and carbon recycling (e.g., Küspert 1982; Hollander and Smith 2001; Jenkyns et al. 2002; Luo et al. 2014). However, changes in the organic source and specific organism assemblages may have contributed to the observed trend.

Thermal maturity

Tmax values of siliceous (average 425 °C) and bituminous limestone (avg. 429 °C) (Fig. 8) suggest that the organic matter is immature. The light yellow fluorescence colour of liptinite macerals (Fig. 11) supports the low maturity. The average CPI value (1.22) is in agreement with an immature or early mature stage (Allan and Douglas 1977). Vitrinite reflectance of three samples (0.39, 0.47, and 0.52%Rr) varies significantly, but shows that the oil window has not yet been reached.

Low maturity is also corroborated by very low Ts/Tm ratios (avg. 0.06), the presence of thermally instable benzohopanes in some samples (Peters et al. 2005), and very low ratios of tri- and monoaromatic steroids (average 0.22) (Mackenzie et al. 1980; Riolo et al. 1986; Peters et al. 2005). MPI-1 ratios could be calculated for 17 samples (avg. 0.41). Vitrinite reflectance (Rc) has been calculated using the equation (Rc = 0.60 × MPI-1 + 0.40; Radke and Welte 1981) and is on average 0.65%. This value is slightly higher than the measured reflectance values. Vitrinite reflectance values derived from MDR values (Rc = 0.4 + 0.3 × MDR − 0.094 × MDR2 + 0.011 × MDR3; Radke, 1988) range from 0.55 to 0.63% (average 0.60%). However, all aromatic maturity parameters may be biased by very low contents of aromatic compounds in many samples.

The average 22S/(22S + 22R) isomerization ratios of αβ-C32 hopane (0.58) are close to their equilibrium value (~ 0.6), whereas the average 20S/(20S + 20R) ratios of ααα-C29 sterane (0.49) did not yet reach equilibrium (~ 0.55). This suggests early oil window maturity (e.g., Seifert and Moldowan 1980, 1986).

Hence, parameters considered as most reliable (Tmax, fluorescence colour) suggest that the organic matter is immature. However, early oil window maturity, suggested by some biomarker parameters, cannot be ruled out.

Petroleum potential

Cross-plots of petroleum potential (PP; S1 + S2) vs TOC are widely used for the assessment of source rock quality (Fig. 19) (Peters and Cassa 1994). Most bituminous limestones show a good-to-very good HC potential based on the PP (6–70 mg HC/g TOC) and TOC contents (1–13 wt.%). In contrast, siliceous limestones hold poor-to-fair source potential with low PP and TOC.

Fig. 19
figure 19

Cross-plot of petroleum potential (S1 + S2) (mg HC/g rock) vs. TOC% after Peters and Cassa (1994)

The amount of hydrocarbons (in tons), which can be generated beneath 1 m2 of surface area, can be estimated using the Source Potential Index [(SPI = h × (S1 + S2) × ρ/1000); Demaison and Huizinga 1994], where h is the thickness and ρ the average rock density. In the present paper the density of siliceous and bituminous limestones was assumed as 2.6 and 2.4 t/m3, respectively. Demaison and Huizinga (1994) suggested neglecting intervals with a PP below 2 mg HC/g TOC). Only a small percentage of the samples in the organic matter-rich Unit 2 is characterized by PP values below 2 mg HC/g TOC. Moreover, they are evenly distributed within the studied section. Hence, these values have been also taken into account.

The SPI of Unit 2, which is 50.95 m thick, was calculated assuming that the percentage of bituminous limestones is similar, in both the non-exposed and exposed part of the succession (46%). Note that the obtained SPI value (1.32 t HC/m2) is a conservative estimate as it is likely that organic-rich bituminous rocks are more easily weathered than siliceous limestones.

With respect to the SPI, given the wide lateral extent of the organic-rich strata (42 km2; Blažeković Smojić et al. 2009) and their maximum thickness (70 m; Blažeković Smojić et al. 2009), it is evident that the Lemeš Formation could be a valuable source rock, provided that it reaches oil window maturity.

Early hydrocarbon migration

Although Tmax values are significantly below the threshold value for oil generation from typical type II kerogen (435 °C; Tissot et al. 1987; Peters et al. 2005), bituminous matter in thin sections (Fig. 20) and polished blocks (Fig. 11) is a strong indication for generated and migrated hydrocarbons. In addition, a cross-plot of PI vs TOC (Fig. 21) shows a relative enrichment of free S1 hydrocarbons in the siliceous limestones of Unit 2, which have low TOC contents (PI up to 0.84) compared to the organic-rich laminated limestones (PI up to 0.18).

Fig. 20
figure 20

Thin section microphotograph from the basal meters of Unit 3 with bit bituminous pore fills. Sample Kro 3

Fig. 21
figure 21

Plot of Production Index (S1/(S1 + S2) vs TOC% visualizing early hydrocarbon migration

Early hydrocarbon generation may be related to the presence of type II-S known for HC generation at low Tmax values (< 435 °C, e.g., Lewan and Ruble 2002). This interpretation is supported by high DBT/Phenathrene ratios in the studied samples. Moreover, based on very high organic sulphur contents in kerogen concentrates, Blažeković Smojić et al. (2009) suggested the presence of type II-S kerogen in other Lemeš locations.

Formation of the Lemeš Basin

The Lemeš Basin is a deep-water basin that was formed in the Late Jurassic evolution of the Adriatic Carbonate Platform, and was surrounded by shallow-water or emerged areas (Vlahović et al. 2005). Existing models explain the formation of the elongated southeast-to-northwest-striking basin by either (1) extensional tectonics with formation of steep escarpments (Herak 1986; Bucković et al. 2004; Bucković and Markić 2016), (2) unspecified synsedimentary tectonics (Fiket et al. 2008), or (3) transpressional strike-slip tectonics with formation of pull-apart basins (Velić et al. 2002a, b; Vlahović et al. 2005 and references therein).

However, new basins rapidly formed in extensional or strike-slip settings are bordered by steep escarpments that trigger mobilization of mass transport deposits. Such mass transport deposits are normally found near the base of the basin fill, and the component spectrum should reflect the mobilized older sedimentary sequence suddenly exposed on the steep flanks of such basins (e.g., the Jurassic sedimentary sequence of the Adriatic Carbonate Platform and its basement). Moreover, in an extensional setting, the grain size and bed thickness should reflect a fining-upward trend (Einsele 1992 and references therein). In a pull-apart setting, the trend of bed thickness and grain-size should be irregular, because depositional centers, transport routes and source areas shift (Steel and Gloppen 1980; Cemen et al. 1985). None of these sedimentological trends are observed in the Lemeš Basin, and mass transport deposits are not known. Instead, around the Oxfordian/Kimmeridgian boundary, the depositional setting shifted in a very short time span from shallow-water to deep-water carbonates. The change is gradual, not abrupt, and no evidence of fault formation, such as resediments or seismites, is observed. In the Poštak region in the central part of the Lemeš Basin, rapid subsidence started earlier (around the Oxfordian/Kimmeridgian boundary) than in the more marginal areas, such as the type-area (late Kimmeridgian).

This rapid increase in subsidence resulted in a deepening event with deposition of deep-water carbonates with chert nodules and clayey and marly intercalations. Black shales formed mainly in the depositional center of this newly formed, elongated basin. The induced subsidence during the Kimmeridgian was higher than the carbonate production potential of shallow-water settings of the AdCP (i.e., 103−4 B; see Schlager 2005 for explanation).

The water depth in the basin can only be estimated on the basis of the thickness of the basin fill. As the basin fill does not exceed 450 m, the maximum water depth, which was during accumulation of the black shales in Unit 2, was probably not more than 250–300 m. In contrast, the assumption of ongoing subsidence during the entire Kimmeridgian would lead to a much higher water depth estimate. The sedimentary structures indicate deposition below the wave base.

We propose the formation of the Lemeš Basin took place during Middle-to-Late Jurassic west-directed ophiolite obduction above the wider Adriatic plate (Gawlick and Missoni 2019 and references therein). During the Middle (Bajocian-to-Callovian)-to-early Late Jurassic (Oxfordian; and not in the latest Jurassic or around the Jurassic/Cretaceous boundary as proposed by Schmid et al. 2008, 2020 and references therein), a west-directed thrust belt was formed in the Inner (Internal) Dinarides below the overriding ophiolites (i.e., the wider Adriatic plate attained a lower plate position). In contrast to the Inner Dinarides, the expression of this mountain building process was relatively mild in the Outer (External) Dinarides. Here, rapidly subsiding synclines (compare Picotti and Cobianchi 2017) were formed around the Oxfordian/Kimmeridgian boundary (Fig. 22), contemporaneous with the arrival of ophiolites in the area of the Drina-Ivanjica unit and its equivalents (Porkoláb et al. 2019; Gawlick et al. 2020 and references therein). Based on subsidence analysis along a cross-section from the northeastern AdCP margin to the Adriatic basin, Picotti and Cobianchi (2017) explained the formation of such synclines as an expression of the buckling of the Adriatic lithospheric mantle as a result of far-field compression from the Tethys/Vardar oceanic realm, coupled with formation of long wavelength (~ 80 km) fold structures.

Fig. 22
figure 22

a Position of the Lemeš Basin (red star) in the Kimmeridgian Western Tethyan/Vardar realm map based on Frisch (1979) and Missoni and Gawlick (2011). AA Austroalpine, ALB Albanides, BC Betic Cordillera, DIN Dinarides, DR-PEL Drau-Range-Pelsö, HEL Hellenides, SA Southern Alps, TZ Tisza, WC Western Carpathians; b schematic cross-section showing the buckling and folding of the AdCP as consequence of nappe stacking and west-directed ophiolite obduction (out of scale)

Furthermore, during Kimmeridgian-to-Tithonian times, the area of the Outer Dinarides was relatively shielded against the influence from the Tethys/Vardar Ocean, because of newly formed carbonate platforms on top of the obducted ophiolites and nappe fronts (Fig. 22b) (Schlagintweit et al. 2008; Gawlick et al. 2020). This is also expressed in the ammonite assemblage in the Lemeš Formation, which has practically no Tethyan influence (Furlani 1910).

Extensional tectonics related to the opening of the Alpine Atlantic west/northwest of the wider Adriatic plate can also be excluded in the area of Outer Dinarides for the Late Jurassic, because the oceanic break-up was much earlier, i.e., during the Toarcian (Ratschbacher et al. 2004). Consequently, during Middle–Late Jurassic times, this oceanic domain (Alpine Atlantic: Missoni and Gawlick 2011) was in a drifting stage, as documented by Bill et al. (2001).

Conclusion

The Lemeš Basin is formed during Late Jurassic time within the AdCP and is filled by the up to 450 m-thick Lemeš Formation. The Lemeš Formation is subdivided from base to top into Units 1–3. The investigation of the Lemeš Formation provides important new insights into the Late Jurassic evolution of the AdCP and deposition of organic matter-rich rocks in a restricted intraplatform basin.

  • Unit 1 overlies Oxfordian shallow-water deposits. The presence of radiolarian packstones and open-marine organisms suggest a fast relative sea-level rise and a connection to open-marine environments. Input of shallow-water debris into the oxygenated basin was minor.

  • Unit 2 consists of radiolarian pack- and wackegrainstones with high TOC contents (avg. 3.57 wt.%). The general absence of bioturbation, low Pr/Ph, and very high DBT/Phenanthrene ratios indicate strongly oxygen-depleted conditions in a (semi-)restricted basin. Water depth increased to 250–300 m. Organic matter input was dominated by marine organisms, whereas a minor contribution of land plants, recorded by rare vitrinite particles and gymnosperm biomarkers, indicates input from nearby land areas. High bioproductivity in the oxic photic zone increased δ13Ccarb values of organic matter-rich sediments.

  • Unit 3 reflects a return to oxygenated conditions and is accompanied by a gradual upward increase of open-marine crinoid- and foraminifera-rich microfacies.

  • The presence of Ataxioceras sp. in the basal part of the Unit 3 together with carbon isotope stratigraphy enables dating of the lower part of the Lemeš Formation. The upper part of Unit 1 accumulated during the late Oxfordian. Unit 2 represents the early Kimmeridgian E. bimammatum, I. planula and S. platynota, ammonite zones and the basal part of Unit 3 the early Kimmeridgian A. hypselocyclum ammonite zone and younger.

  • The absence of major basin-bounding faults and of a sedimentary succession typical for rift or pull-apart basins suggests that the basin did not form due to extension. Based on contemporaneous ophiolite obduction in the Inner Dinarides, we propose a basin model involving subsidence due to buckling of the AdCP in the Outer Dinarides as a far-field consequence of compressional processes in the Inner Dinarides.

  • Unit 2 contains good-to-very good, highly oil-prone source rocks, which are thermally immature (Tmax < 430 °C) at the study location. When mature (e.g., in sub-thrust position), these rocks can generate approximately 1.3 t HC/m2. High HI values (502 mgHC/gTOC) and high amounts of organic sulphur suggest the presence of kerogen type II-S. This kerogen type generates hydrocarbons at relatively low maturity. Migrated bitumen and high partially PI values are consequences of early hydrocarbon generation.

  • The sedimentary Lemes Basin fill with formation of source rocks in this newly formed far-field syncline with rapid subsidence in the frame of Middle-to-Late Jurassic ophiolite obduction onto the wider Adriatic plate clearly demonstrates a potential of undetected hydrocarbon reservoirs in the nappe stack of the Dinarides. Similar restricted basins may have formed to the east and later buried below the nappes formed during the polyphase thrust processes recognized in the Dinarides (Schmid et al. 2008, 2020) or even below the ophiolites, which were transported further to the west in the frame of latest Jurassic unroofing (Gawlick et al. 2020).