Biochemically induced diagenesis of Jurassic micrite: evidence from phase analysis, carbon, oxygen, and strontium isotopes (Franconian Alb, Germany)

Abstract

The marine Upper Jurassic rocks of the Franconian Alb consist largely of micritic carbonate of partly dolomitized reef mounds and bedded basinal limestone. All carbonates were lithified in the shallow (centimeters, meters) subsurface and have a wide range of ∂¹³C (≤ + 3‰ to − 10‰VPDB) but always negative ∂¹⁸O (− 1 to − 6‰VPDB). Dolomite and reef limestone show the highest ∂¹⁸O and ∂¹³C values. The most negative ∂¹³C (≥ − 10‰) occurs mainly as cement in dolomite of a basinal, partly dolomitic, biostrome interval. Basinal limestone shows intermediate ∂¹³C values. Because freshwater diagenesis and elevated temperatures cannot explain the observed isotope values, pH is here considered a major factor influencing the isotope signal of micritic limestone. The bulk sediment isotope signal was reset to lower values, from an original lime mud with ∂¹³C ≥ 3‰ and a ∂¹⁸O of ≥  + 1‰, as a result of biochemically induced diagenesis. Carbonate, probably mostly aragonite but occasionally including dolomite, was dissolved in a zone where low pH developed as a result of organic matter degradation. Dissolved carbonate was translocated by diffusion and re-precipitated as cement (ca. 50vol%) in a zone with elevated pH where all in situ lime mud ∂¹⁸O was reset. Imported cement carbonate precipitated in equilibrium with the pore fluid with negative isotope values, whereas ∂¹³C of the in situ lime mud remained unmodified. The negative shift of the bulk ∂¹³C and ∂¹⁸O is variable and depends on pH and the contribution of ¹²C from anaerobic sulfate reduction in the zone of cement precipitation. This produced an ubiquitous covariance of ∂¹⁸O and ∂¹³C. Incorporation of seawater-derived Mg²⁺ during recrystallization of carbonate can account for the local dolomitization. Elevated ⁸⁷Sr/⁸⁶Sr ratios are explained as a result of interaction of clay minerals with the stationary pore fluids. This study shows that the isotopic signal produced by biochemically induced shallow submarine subsurface carbonate diagenesis can be indistinguishable from freshwater diagenesis, that ∂¹⁸O and ∂¹³C of the bulk rock are always reset, and that carbonates can show, in the presence of clay minerals, elevated ⁸⁷Sr/⁸⁶Sr ratios even when the pore fluids were never exchanged.

Graphical Abstract

Introduction

Micritic limestone is very common in the geological record, but when and how the originally muddy sediment was transformed into a solid rock has been controversial (discussion in Munnecke et al. 2023). Based on observations from recent sediments, it has often been assumed that micritic carbonate lithifies under meteoric conditions which provide the cement necessary to transform a lose carbonate mud into a solid rock. Micritic limestone, which has not been (intermittently) subaerially exposed, must have lithified in a submarine environment (discussion in Munnecke et al. 2023). This is best exemplified by slabs of intraformational micritic limestone underlain and overlain by entirely submarine micrite, and these slabs preserve the fabric of the rocks in which they are embedded (Fig. 1). The Upper Jurassic rocks of Germany, which consists largely of micritic limestone, are an ideal natural laboratory to assess when, how, and why these limestones formed.

Fig. 1

An example for submarine lithification of micritic carbonate. An interval of contorted beds (“Krumme Lage”) between two micritic deep water units of the Tithonian Plattenkalk succession in southern Germany. The contorted interval consists of folded slabs of Plattenkalk with the characteristic planar stratification and thin, slightly argillaceous, interbeds characteristic of the undisturbed rocks and proves that lithification occurred under submarine conditions. Southern Franconian Alb, Solnhofen area (Horstberg quarry), for location see Fig. 2b

The term micritic limestone is used here for carbonate which are classified as mudstone or wackestone. Micritic limestone, at high magnification and under the scanning electron microscope (SEM), mostly is a microcrystalline rock with different allochems, and this is not different for the studied rocks. Current techniques do not permit the sampling of micron-sized components, and most analyses of this study are bulk rock data.

Previous work

The diagenesis of lime mud “..involves a change from solid carbonate phases within pore solutions, to a rock composed of low-Mg-calcite.. (associated with) a drastic reduction of the primary porosity to 2–3%..” (Flügel 1982). Because micritic limestone normally lacks signs of early compaction, the provenance of the enormous amounts of cement, which reduce porosity from several tens of % to very low values, is difficult to explain. Cement precipitation from seawater percolating the sediment appears, in view of the ca. 100.000 pore volumes necessary, unlikely because of the low depositional permeabilities (discussion in Munnecke et al. 2023). Equally unlikely is large-scale pressure solution as cement source, because the micritic limestones were cemented before pressure solution set on. It must be concluded, therefore, that cementation occurred in the shallow subsurface in a submarine environment (Melim et al. 2002; Munnecke et al. 2023).

The depositional polymorph of calcium carbonate plays a critical role in whether and how lime mud transforms into a solid rock, or remains poorly consolidated. Depositional low-Mg calcite (LMC), best exemplified by the Upper Cretaceous–Lower Cenozoic chalk, a micrite consisting of coccolith remains, is a largely uncemented rock with very high preserved residual primary porosity (Fabricius 2007). Pure high-Mg calcite (HMC) muds likely undergo incongruent transformation to LMC and microdolomite, but would not provide cement necessary to transform a lose mud into a hard limestone. Aragonite dissolution and re-precipitation as LMC cement is necessary to lithify lime mud (Munnecke et al. 2023). Microbial processes and products in the uppermost layers of the sediment column are now held responsible for the lithification. The importance of diagenesis in the shallow subsurface, particularly in the sulfate reduction zone, has been proposed decades ago (Coniglio 1989; Joachimski et al. 2002). The present paper has the purpose to provide evidence from a geochemical and phase analysis perspective.

Regional setting and stratigraphy of the study area

The south German Jurassic rocks consist of clastic and carbonate rocks deposited on a shallow shelf north of the Tethys ocean (Schmid et al. 2005). The Jurassic strata are a few hundred meters thick and best preserved on, and in the periphery of, the flat-lying karstic plateaus of the Franconian and Swabian Alb (Fig. 2), continuing southwestward into Switzerland and France as the so-called Faltenjura (folded Jurassic). Thickness of the south German Upper Jurassic carbonates is 300–600 m. The stratigraphically most complete succession occurs on the southern part of the Franconian Alb and includes the Lower Tithonian (Schweigert 2015). The Upper Jurassic is an entirely submarine succession without intraformational exposure, but a final period of Upper Tithonian to Cenozoic subaerial exposure. Lower Cretaceous and Cenozoic rocks form relics of continental clastics and document a long period of subaerial exposure, but a thin sequence of Upper Cretaceous marine clastic deposits occurs in the Regensburg area. The exact burial history in not well constrained, but measurements of vitrinite reflectance from Lower to Upper Jurassic rocks suggest that organic matter preserved in some intervals is immature for hydrocarbon generation (Kreiensiek 2019; for similar conclusions see also Freitag et al. 2022). Maximum burial temperature was, therefore, < 60 °C.

Fig. 2

a Simplified geological map of Germany with the main areas of Jurassic rocks. Inset rectangle shows the map area of (b), RO7324-B3 is well used for correlation in Fig. 13c. b Carbonate facies of the Franconian Upper Jurassic, modified after Meyer and Schmidt-Kaler (1989), also showing the localities mentioned in the text. The reef facies established in the Oxfordian and Early Kimmeridgian (Malm α and β) and expanded into the Late Kimmeridgian Malm δ) at the expense of the intervening basinal areas, where the bedded facies was deposited

The Upper Jurassic rocks of southern Germany have been subdivided into six intervals named with Greek letters alpha to zeta (α–ζ; Quenstedt 1856–1858). This practical and simple subdivision is now supplanted by a plethora of formations and members with locality names difficult to comprehend but generally not deviating from the old subdivision (Niebuhr and Pürner 2014; Schweigert 2015). For comparison, the old and new subdivisions are illustrated in Fig. 3, but for reasons of simplicity, the old subdivision will be used in this paper. The terms Malm and Upper Jurassic are used synonymously.

Fig. 3

Stratigraphic subdivision of the Upper Jurassic (Malm) of the Franconian Alb, after Schweigert (2015). Because the formation names by and large correspond to the Quenstedt subdivision with Greek letters for a to z (α–ζ), this older subdivision will be kept throughout the paper for reasons of simplicity

Upper Jurassic carbonate rocks of Germany consist of two major facies, namely a regularly, decimeter bedded sequence of carbonate alternating with thinner marly interbeds, here called bedded facies or basinal facies, and a reef facies consisting of massive limestone and dolomite with little macroscopically detectable fabric. The bedded facies consists of a dense, usually structureless, light colored lime mudstone with isolated, partly broken, fossil remains such as ammonoids, brachiopods, and bivalves. The Franconian Alb shows a variable facies development including extensive, partly dolomitic, reef tracts in addition to the Plattenkalk (platy limestone) occurrences, both of which are very rare on the Swabian Alb. Plattenkalk is distinguished from the bedded limestone facies in the planar bedding planes (Munnecke et al. 2008; Mäuser 2015; Viohl 2015a, b)(Fig. 1) frequently with internal millimeter lamination. The Oxfordian to Kimmeridgian rocks are well correlatable on a bed-by-bed basis, partly into the Swabian Alb, so for about 200 km (Freyberg 1966; Meyer and Schmitt-Kaler 1989).

Two major basinal facies areas can be distinguished. One is the so-called Fränkisch-Schwäbische Mergelfazies (literally Franconian-Swabian marl facies, here called marly basin facies, Fig. 2b), which characterizes the northernmost part of the Franconian Jurassic and the Swabian counterpart. The marly basin facies is limited in the south by a persisting reef tract extending from the Gräfenberg area roughly eastward (Fig. 2b). South of this reef tract the kalkreiche fränkische Fazies (limestone-rich Franconian facies, here called the calcareous basin facies, Fig. 2b), dominates and has fewer and thinner argillaceous interbeds in the Malm β, γ and δ (depicted as Southern Franconian Alb in Fig. 4). The marly basin facies of the Northern Franconian Alb (Fig. 4a) is characterized by limestone intervals called Bankgruppe (group of beds) A–D separated by more argillaceous intervals called Zwischenmergel (marl interbed) (local German stratigraphic terms used in this paper are in Fig. 4, and are translated if used in the text). A marker interval in both facies units is the Platynota-Mergel (Platynota Marl, Fig. 4). In the calcareous basin facies, the Crussoliensis-Mergel (Crussoliensis Marl) and the Treuchtlinger Marmor (Treuchlingen Marble) form readily correlatable intervals.

Fig. 4

Stratigraphic sections of the northern (a) and southern (b) Franconian Alb showing the facies of the basinal carbonates and oxygen and carbon isotopes for the intervals Malm β–ζ. Original German local stratigraphic terms and index intervals are, when used, translated in the text. Dashed vertical lines in the isotope logs are for orientation at + 2 and – 3.5‰. Logs are flattened at the base of the Malm γ

The Franconian reefs originated in the Late Oxfordian and Early Kimmeridgian (Malm α and β) as tracts (Fig. 2b) where isolated and laterally coalesced reefs prevail (e.g., Lang 1989). These reefs display two different geometries (Fig. 5). One consists of a massive reef core or patch reef (or reef mounds, the terms are here used synonymously), a few meters thick, with abundant sponges which is overlain and grades laterally into the bedded facies (Fig. 5a). The flank angles are very variable and range from a few to 60°, but appear symmetrical. The beds overlying the massive reef core form a drape with consistent bed thicknesses (Fig. 5a). This type of stratification has been called Girlandenschichtung (literally “garland bedding”; Geyer and Gwinner 1968) which can cover laterally extensive areas documenting the existence of patch reefs below (Figs. 5b, 6a). In many instances, these isopachous beds contain abundant sponges similar to the reef core facies. Although these draping geometries are no reefs in terms of a thickness anomaly, they are characteristic features of Upper Jurassic carbonates and included here in the term reef. The submarine topography of these reef areas was up to about 60 m. In map view, the reefs are linear at the hundred meter scale (Lang 1989), but also show bifurcations and changes in orientation (Fig. 6b). Reefs are sites of an extensive network of subvertical and bedding parallel karst dikes filled with unconsolidated siliciclastic sediment (Figs. 5, 6), which is undated but most likely formed during Late Tithonian to Cenozoic uplift and karstification. Many reefs and reef complexes consist of dolomite (Meyer and Schmidt-Kaler 1989).

Fig. 5

Geometries of Upper Jurassic reefs. a Reef (Gräfenberg, Deuerlein&Bärnreuther quarry), developed on a flat base, shows an unstratified reef core draped by isopachous beds similar to the geometry of (b), which is the most common reef geometry in the south German Jurassic and characterized by isopachous beds with mounded shapes, the so-called Girlandenschichtung (garland bedding). It is an entirely dolomitic reef (Ludwag quarry). Both reef types have in common that they are cut by numerous vertical dikes and sills filled with siliciclastics of Tithonian to Cenozoic age

Fig. 6

a Line drawing of the north face of the Ludwag quarry showing the dolomitic reefs of the Malm β draped by the Platynota Marl of the lower Malm γ and the sample points of this nearly entirely dolomitic outcrop. The reefs consist exclusively of draping beds with the typical Girlandenschichtung suggesting the existence of unexposed patch reefs in the lower Malm β and/or α. b Map of the top of the Malm β, contour increment is 2 m. The reef geometries in the outcrop section (a) are partly artifacts caused by varying intersection angles with the quarry wall. Mapping results in a linear, branched reef with E–W to ENE–WSW orientation. Based on geological map data (Antonianidis et al. 1972), the reef complex is elevated by about 60 m above the adjacent (calcareous) basinal facies

A peculiar feature of Franconian Upper Jurassic rocks is the development of a largely dolomitic Upper Kimmeridgian biostrome facies (mainly Malm δ 3 + 4). This facies marks the peak of reef expansion (Fig. 3), which also affects the basinal areas. This thick bedded facies consists of grey, sucrosic dolomite, calcite-cemented dolomite, and dedolomite, which preserve layered microbial crusts similar to laterally linked hemispheroids of stromatolites (e.g., Meyer and Schmitt-Kaler 1994; Fig. 7a). Many beds contain an unquantified amount of organic matter, easily identified in the field by the odor of kerosene on freshly broken surfaces. Calcite-cemented dolomite occurs as beds, decimeters to meters thick, which can be traced across entire small basinal areas (Fig. 8a). It is strictly speaking not exactly stratiform, but forms an anastomosing geometry within stratified sucrosic dolomite at the meter scale. The biostrome dolomite shows decimeter to meter white or light grey, calcite-cemented patches with diffuse boundaries, and a tendency to disintegrate into dolomite sand (Mäuser et al. 2002). Laterally, the biostrome beds grade updip into grey reef dolomite within a few meters (Fig. 8a).

Fig. 7

Facies types and diagenetic modifications associated with the δ¹³C anomaly. a Calcite-cemented biostrome dolomite with preserved stromatolitic, light colored parts in the upper half similar to cauliflower stromatolites. b Vug partly filled with blocky calcite containing micrite clasts overlain by geopetal micrite. c Replacive calcite concretion consisting of fibrous calcite with coloration mimicking the grain size changes in graded layer. The wavy lamination in the dark, organic rich layers suggests stromatolites. White triangles show graded detrital limestone. d The isotope anomaly from a composite drill core and outcrop log of the Wattendorf quarry. e Reduced thickness of the anomaly in a location at ca. 700 m distance from (d). Dashed vertical lines in the isotope logs are for orientation at + 2 and – 3.5‰, vertical scale is in meters

Fig. 8

Reef-to-basin transition in outcrop (a; Kümmersreuth quarry, Fig. 2b) and in isotope logs (b). The drill cores 1 and 2 of the Wattendorf area document reef limestone, capped by dolomite, with upward increasing δ¹³C and ∂¹⁸O values, in which isotope log correlation is possible. The adjacent core 3 (cf. Figure 7d) documents basinal facies probably underlain by reef limestone with isotope values similar to those of core 1 and 2. The basinal facies is recognized by the more negative ∂¹⁸O values and the serrated δ¹³C log. Differences in log depth are actual differences in elevation in a structurally horizontal situation. Vertical lines in the isotope logs are for orientation at + 2 and − 3.5‰

The basinal facies above the biostrome facies, up to 500 m thick, consists of bedded, very pure limestone, occasionally slightly bituminous, with Plattenkalk intervals (Fig. 4).

Methods

Mineral phase analysis

For the purpose of better quantifying the mineralogical composition, X-ray diffractometry (XRD) was carried out on 266 bulk rock samples. For the X-ray phase analyses, a PhilipsPW1710 with Co or Cu K-radiation, respectively (1 nm; Ni filter, under 30 kV and 30 mA with a step increment of 0,004°), was used. Approximately 1 cm³ of each sample was powdered to analytical fineness in a vibratory disc mill using a tungsten carbide grinding set. For the semi-quantitative determination of the content of different carbonates and the exact location of the dolomite and calcite d(104) peak in the samples, CaF₂ was added as spike mineral in a ratio of 1:5. To achieve randomly oriented aggregates, which allow to compare relative intensities for a series of peaks, the powder was carefully scraped into the cavity of the aluminum holder and only gently pressed with a glass slide to counteract the tendency of particles to lie parallel to the glass surface. The evaluation was performed using the software program Xpert High Score (by Panalytical). Calculation of dolomite stoichiometry was performed with the equations of Kaczmarek and Sibley (2011).

Isotope analysis

To quantify the isotopic properties of the Jurassic carbonates, oxygen, carbon, and strontium isotopes were measured. One thousand five hundred sixty one bulk rock samples and four fibrous cements were analyzed for stable isotopes of oxygen and carbon at the Geozentrum Nordbayern in Erlangen. Sample material (0.5–1 g) was drilled with a 1 mm bit from fractured rock samples and slabbed surfaces, respectively, and reacted with 100% phosphoric acid at 70 °C using a Gasbench II connected to a ThermoFisher Delta V Plus mass spectrometer. Reproducibility and accuracy were monitored by replicate analysis of laboratory standards calibrated by assigning ∂¹³C values of + 1.95‰ to NBS19 and − 47.3‰ to IAEA-CO9 and ∂¹⁸O values of – 2.20‰ to NBS19, and – 23.2‰ to NBS18. Reproducibility for ∂¹³C and ∂¹⁸O was ~  ± 0.05 ‰. The ∂¹⁸O data were reported for calcite relative to the Vienna PDB (VPDB) standard in the usual delta notation. Potential phosphoric acid fractionation of dolomitic samples was not corrected. Uncertainty of the exact fractionation factor has not improved since Land (1980), but it is usually assumed that dolomite ∂¹⁸O is about 1.1‰ more negative than calcite. One hundred twelve carbonate samples were analyzed for ⁸⁷Sr/ ⁸⁶Sr ratios at the Institute for Material Analysis and Science, Göttingen. Prior to digestion, all samples were spiked with a tracer solution (Sr 4-AD) and dissolved in Teflon flasks by various steps of addition and evaporation of 2.5 N HCl until complete dissolution. The column procedure using Micro Bio-Spin Columns (Bio-Rad) with Sr-Resin-B (Triskem) was applied to separate Sr. The isotope measurements were done with a ThermoFinnigan TIMS Triton TI. Instrumental mass bias was corrected with ⁸⁸Sr/⁸⁶Sr of 0.1194 using exponential law. Concentrations were calculated using ID-TIMS technique. Within this study, the average value for NBS 987 (n = 9) is 0.71027 ± 7, which is consistent with the long-term average ⁸⁷Sr/⁸⁶Sr value of the laboratory (0.71027 ± 5, n = 466).

Results

Petrography and phase analysis

Upper Jurassic micrite is a microcrystalline (5–20 μm) rock with peloids, crinoid fragments, sponge spicules, and other inderterminable shell fragments. Occasionally detrital quartz and mica grains can be recognized, particularly in the Platynota Marl. Pyrite is present, estimated at < 1% of the rock volume in thin section and is also indicated by covarying Fe and S contents of < 1% measured by X-ray fluorescence (Blendinger and Mäuser 2021). Clay minerals are enriched in laterally discontinuous flaser seams, < 1 mm thick in thin section (Supplementary Appendix 1a), and interbeds, millimeters to centimeters thick, with millimeter to centimeter subglobular to flattened micritic components. Dolomite rhombs < 0.1 mm in size occur either as isolated crystals or as nests in micrite. Plattenkalk has intervals of bituminous, partly dolomitic, micrite with stromatolitic lamination (Fig. 7c) consisting of crystals 10–20 mm in size (Scharfenberg 2011). The bedded limestone contains ca. 10% HCl-insoluble residue (clay minerals, rare quartz grains, silicified microfossils), whereas the Platynota and Crussoliensis Marls contain 15–20% insoluble residue. The Upper Kimmeridgian bedded limestone and Plattenkalk have very low clay contents (< 5%) and are, therefore, quarried as highly pure limestone for industry purposes. The insoluble residue of (Malm β) reef rocks is 10–15%, similar to that of the bedded facies. Macroscopic signs of compaction are lacking. This is best documented in the Treuchtlingen Marble, which contains large sponges, belemnites, and ammonites which are undeformed. In the marly facies (such as the Platynota Marl), large fossils (sponges, ammonoids) are not compacted but the macroscopic rock fabric is more nodular similar to Supplementary Appendix 1a. Hard or firm grounds are lacking, too, except for indirect evidence, in the reefs, from sessile organisms such as Diceratid bivalves, brachiopods and crinoids.

Reef micrite is macroscopically mostly indistinguishable from basinal micrite, but locally shows a pelletoidal-clotted fabric (e.g., Meyer and Schmidt-Kaler 1989). The pelletoidal micrite consists of densely packed peloids, micritic ooids and small oncoids with diffuse outer boundaries, and microsparitic cement (Koch and Schorr 1986; Meyer and Schmitt-Kaler 1992). Sponge calcification occurred early in diagenesis (Brachert 1986) replacing the siliceous skeleton by dark, microcrystalline calcite with peloids and locally preserves the microstructure of the sponges. Millimeter thick overgrowths of dark peletoidal micrite occur on many sponges and lead to the classification as sponge-microbial reefs (e.g., Meyer and Schmitt-Kaler 1989, 1991, 1992). The reef facies show no signs of compaction, because large sponges have preserved their cup or discoid shape. Dolomite preserves the fabric in large thin sections (Supplementary Appendix 1b) and consists of a cloudy groundmass containing sponges with stromatolitic crusts. Under the microscope, dolomite consists of idiomorphic crystals < 1 mm in size, with a turbid core and a clear cement rim (Supplementary Appendix 1c). The cores of the dolomite crystals can show a clotted–peloidal relic fabric. Residual primary voids are normally lined by a thin equigranular, dolomitic rim cement or a transparent, planar dolomite overgrowth cement.

Calcite-cemented dolomite of the biostrome facies forms irregular patches surrounded by grey dolomite with little calcite cement (Supplementary Appendix 2a). The calcite-cemented dolomite preserves areas with high intercrystalline porosity and consists of idiomorphic dolomite rhombs with calcitised cores (Supplementary Appendix 2b) and idiomorphic, clear rims. Clear, blocky calcite is the cement. Curiously, neither the outer dolomite rims nor the calcite crystals are corroded, but corrosion is restricted to the inner parts of the dolomite crystals (Supplementary Appendix 2b). Dedolomite occurs in micritic limestone, recognized by abundant hollow dolomite rhombs causing a patchy distribution of high dissolution porosity, and are only locally supplanted by calcite (Supplementary Appendix 2c).

In all facies, centimeter to decimeter sized concretions and vugs occur (Mäuser 2015; Fig. 7b, c). The geometry of these is generally subglobular to elongated in outcrop, but very irregular in drill core (Fig. 7b) and fresh quarry walls. The concretions and voids typically parallel bedding in irregular, meter to ten meter spaced intervals. In slump folds of Plattenkalk, these voids are included in the deformation. The vugs are either empty or filled with black to transparent, fibrous calcite crystals, and yellow micritic internal sediment. Concretions of fibrous calcite in Plattenkalk can be replacive mimicking the fabric of graded layers (Fig. 7c). Blocky calcite overlain geopetally by yellow micrite is also observed (Fig. 7b). White silica concretions with concentric layering, structureless silica nodules and nodules of unconsolidated silica powder (Brachert 1986) are common.

The XRD analyses show that two mineral phases dominate the Upper Jurassic carbonate. One is a Mg-free calcite, the other is dolomite. Some limestone samples show a subordinate dolomite peak (Supplementary Appendix 3a, b), quantifiable as a few percent dolomite. Dolomite occurs also in reef limestone (Supplementary Appendix 3c). A quartz peak, indicating low percentages (< 1%) of SiO₂, is present in most samples and most likely documents silica from silicified fossil remains and/or silica nodules, because detrital quartz is practically absent in thin section. Dedolomite can be pure calcite, with no detectable dolomite remains (Supplementary Appendix 3d). The difference between calcite and dolomite is that a peak shift, indicative of the presence of Mg in the calcite phase, is not observed in calcite, but does occur in many dolomite samples, mainly in the Lower Kimmeridgian dolomite (Fig. 9a). The dolomite peak is split in two allowing the determination of dolomite stoichiometry (Kaczmarek and Sibley 2011) for the two dolomite phases. These show a bimodal distribution for the Mol% CaCO₃ and the amount of non-stoichiometric dolomite in dolomite (Fig. 9b). The bimodality appears typical for dolomite (Sperber et al. 1984; Searl 1994). Similar to Cenozoic and Pleistocene marine dolomite of the Bahamas (Vahrenkamp and Swart 1994), nearly stoichiometric and Ca-rich dolomite coexist in the Upper Jurassic rocks.

Fig. 9

Dolomite stoichiometry and isotopes. Reef dolomite often shows a split dolomite peak in XRD (a), which allows the distinction between a Ca-rich dolomite and a nearly stoichiometric dolomite. The quantification results in a bimodal distribution of the two dolomite polymorphs and that dolomite consists of bimodally distributed amounts of Ca-rich dolomite (b). Lateral stoichiometry and isotope changes along the dolomitic north face of the Ludwag quarry (c) (cf. Figure 6a) are not pronounced. d shows the same data as in (c) plotted against the percentage of stoichiometric dolomite in dolomite, suggesting that dolomite stoichiometry does not affect oxygen and carbon isotopes

Carbon and oxygen isotopes

General: Carbon isotopes range from ∂¹³C ~ − 10‰ to ~  + 3‰, ∂¹⁸O from ~ − 6‰ to ~ − 1‰ (Fig. 10a). Bedded limestone, reef limestone, and dolomite cover slightly different fields (Fig. 10b–d). Bedded limestone, calcite-cemented dolomite, and dedolomite can show negative ∂¹³C values. Pure dolomite always has positive ∂¹³C. The most negative ∂¹³C values were measured in fibrous calcite cement (Fig. 10a). The isotope logs of the basinal facies show that Lower Kimmeridgian (corresponding to the Malm β and γ in Fig. 4) bedded limestone ∂¹³C is generally more positive, ~ 2‰, than Upper Kimmeridgian and Tithonian ∂¹³C, where the values fluctuate between ca. − 3 and + 2‰ and a relatively consistent ∂¹⁸O around − 5 to − 6‰. By contrast, Lower Kimmeridgian bedded limestone shows less negative ∂¹⁸O, ~  − 2 to − 3‰. The shift to more negative ∂¹⁸O is stratigraphically somewhat lower in the southern Franconian Alb (near the top of the Crussoliensis Marl) than in the northern Alb (in the Malm δ, Altenberg section, Fig. 4a). It is also associated with a minor but consistent negative shift of ∂¹³C of ca. 0,2‰.

Fig. 10

Oxygen and carbon isotopes. a Cross-plot of all carbon and oxygen isotope data, split into reef limestone data (b), reef dolomite and biostrome dolomite (c), and bedded limestone data (d). The data cover similar fields, but bedded limestone data are shifted to slightly more negative ∂¹⁸O values with respect to reef limestone data. The overall data distribution (a) is similar to the “inverted J” thought to be typical for freshwater diagenesis (Lohmann 1987). Carbon and oxygen isotopes and the histograms for dolomite δ¹³C from the Lower Kimmeridgian reefs (e, f, data are mainly from the Ludwag quarry), and Upper Kimmeridgian reefs (g, h) show that Upper Kimmeridgian dolomite shows somewhat less positive δ¹³C values than Lower Kimmeridgian dolomite. Negative dolomite δ¹³C data are exclusively dedolomite and calcite-cemented dolomite. (i) Isotope cross-plot for Upper Kimmeridgian reef limestone, based on drill core data (see Fig. 8b), show that ∂¹⁸O and δ¹³C co-vary in a ratio of ~ 4:1–3:1

Dolomite–limestone relationships: Dolomite shows the most positive/least negative ∂¹³C and ∂¹⁸O values of the studied rocks. Lower Kimmeridgian reef dolomite (and dolomitic Platynota Marl, Fig. 6) cover a field which is slightly more positive than that of equivalent reef limestone (Fig. 10e, f). The reef limestone data show an inclined upper limit of the most positive data points, whereas dolomite has a horizontal upper limit. Oxygen isotopes in dolomite show a comparatively small spread of 2‰ only and a nearly complete overlap with limestone. The most positive isotope values of Upper Kimmeridgian reef limestone and dolomite have an inclined limit, ∂¹³C and ∂¹⁸O values completely overlap (Fig. 10g, h). The inclined trend is particularly well documented in the limestone data from two drill cores (Figs. 8b, 10i). The trend line forms a slope with a ratio of 1:3 to 1:4 (∂¹³C/∂¹⁸O) and fair correlation coefficients. The ∂¹⁸O values have a spread of ca. 3‰, thus slightly more than for Lower Kimmeridgian reefs.

Lateral trends in dolomite isotopes are not very pronounced at the hundred meter scale. In the Ludwag quarry, dolomite ∂¹³C slightly increases eastward within ca. 200 m, associated with a slight decrease in ∂¹⁸O (Fig. 9c), but the correlation coefficients are poor. Better defined is a change in the content of stoichiometric dolomite which prevails in the eastern part of the outcrop, but dolomite stoichiometry shows no relation with oxygen and carbon isotopes (Fig. 9d).

Lateral limestone–dolomite transitions: The isotope data of stratigraphically equivalent dolomite and limestone intervals show that dolomite is always slightly more positive in both ∂¹³C and ∂¹⁸O. The difference is ~ 1‰ in an outcrop of the biostrome facies (Fig. 11a), and the dolomite and limestone logs show comparable curve shapes (Fig. 11b). The difference between limestone and dolomite is higher in two hand specimens from Plattenkalk and its dolomitic, finely crystalline equivalent (Fig. 11c, d). Here, the difference is ~ 1.5‰ for ∂¹³C and ~ 3‰ for ∂¹⁸O, but the ∂¹³C difference of this interval can be up to 4‰ (Fig. 7d).

Fig. 11

Dolomite–limestone pairs at the outcrop scale (a, b) and hand specimen scale (c, d) (a) show a dolomite and laterally equivalent limestone of the biostrome facies (Malm δ3 and 4). The superimposed isotope logs (b) for the two facies show that dolomite is shifted to slightly more positive values with respect to limestone, but the overall curve shapes are comparable. Hagenacker. Dashed lines in the isotope logs are for orientation at + 2 and – 3.5‰. c, d is a pair of a calcareous and a microcrystalline dolomite sample (from inclined beds) from the Plattenkalk interval of the Wattendorf quarry; for location, see Fig. 2b

Vertical limestone–dolomite transition: These show three types. The first shows no significant changes in ∂¹³C and ∂¹⁸O from limestone to overlying dolomite, but slight positive anomalies in the downdip termination of reef tongues (Supplementary Appendix 4a). By contrast, the isotope log of an isolated Upper Kimmeridgian reef overlain by (unexposed) basinal sediments (Supplementary Appendix 4b) shows an upward decrease of ∂¹³C and ∂¹⁸O, terminating in the serrated ∂¹³C anomaly (see below) and finally into the more positive ∂¹³C of dolomite. The reef facies of a nearly time-equivalent reef, but part of the persistent reef belt (Fig. 2b) on the northern Franconian Alb (Fig. 8b), shows the opposite trend, namely upward increasing ∂¹³C and ∂¹⁸O values in the calcareous part, and the thin cap of dolomite does not show isotopic differences with the underlying reef limestone.

Centimeter-scale isotopic variation: Fragments of microbial mats in bedded limestone and the micritic matrix show trends similar to that of dolomite and calcite-cemented dolomite (Fig. 12a, b). In the biostrome dolomite, ∂¹³C becomes gradually more negative along a slightly curved trend. Although unquantified, this is the effect of increasing amounts of calcite cement in the samples. Microbial mats and matrix show indistinguishable values in the lower part of the Treuchtlingen Marble, but in the upper part, ∂¹³C becomes slightly less positive. This stratigraphic trend is detailed in an isotope log of the upper part of the bedded limestone of Treuchtlingen Marble (Fig. 12c, d).

Fig. 12

Oxygen and carbon isotopes from polished hand specimens of the Malm δ. (a, b) Calcite-cemented dolomite and dolomite show a curved covariance of ∂¹⁸O and δ¹³C. The bedded limestone below microbial crusts (Algenkrusten-Lagen (algal crust layers) in Meyer and Schmitt-Kaler 1994) and matrix have similar values, but slightly more negative crust δ¹³C in the upper beds. This trend is mirrored in a bulk rock isotope log (c) of the transition from limestone to dolomite (abandoned quarry in the municipal building yard, Eichstätt). d Standard section of the Treuchtlingen Marble with bed numbers and thicknesses as used in the quarry industry. For localities, see Fig. 2b. Index beds (OMP = Obere Mergelplatte (Upper Marl Plate), UMP = Untere Mergelplatte (Lower Marl Plate)) are marked in red. Colors and stratigraphic terms as in Fig. 4

The ∂¹³C anomaly: A conspicuous feature of the studied Upper Jurassic carbonates is a distinct ∂¹³C anomaly, restricted to the basinal facies, in which values of nearly − 10‰ occur alternating with more “normal”, positive values and causing a serrated log shape (Fig. 4, 13a). The anomaly is invariably associated with the biostrome facies and can be up to 60 m thick (Fig. 4b). The lowest negative ∂¹³C values occur regionally at slightly different stratigraphic positions (Fig. 4). Several somewhat less negative, isolated ∂¹³C peaks occur in different stratigraphic intervals below and also in the Plattenkalk intervals (Fig. 4, 13b). Up-section, the anomaly gradually changes to ∂¹³C values of around 0‰ within about 10 m. The anomaly is lacking in the persistent reef tracts (cf. Figure 8b).

Fig. 13

Correlation of Upper Kimmeridgian isotope logs in the northern (a, flattened at top Malm δ) and southern (b, flattened at top Malm δ2) Franconian Alb. In both cases, the δ¹³C anomaly is clearly identified, but has varying thicknesses and a stratigraphically inconsistent base. The correlatability from Switzerland to northern Germany (c, flattened at top Divisum zone) is poor, but the minimum ∂¹⁸O and maximum δ¹³C values are comparable in all logs. Blue log = ∂¹⁸O, red log = δ¹³C. The North German section shows neither correlation nor consistent covariance of the isotopes. Vertical lines in the isotope logs are for orientation at + 2 and – 3.5‰, location of the Swabian section given in Fig. 2a

Covariance: The isotope logs show a ubiquitous ∂¹³C/∂¹⁸O covariance (Fig. 4, 7, 8, 12, 13). It occurs at all scales including the hand specimen scale (Fig. 11c, d), irrespective of the mineralogy. This covariance is difficult to quantify, except for the Upper Kimmeridgian reef facies (see above), because the absolute amounts differ from one locality to another. The change is normally more pronounced for ∂¹⁸O (Fig. 4), but can also amount to very similar values for ∂¹³C and ∂¹⁸O (Fig. 11c, d). The negative ∂¹³C peaks in the biostrome and bedded limestone facies have only minor corresponding negative ∂¹⁸O excursions (Fig. 4, 8b).

Belemnite–micrite relations: The oxygen and carbon isotopes of LMC skeletal material such as belemnites, brachiopods, and inoceramid bivalves are normally used to assess a marine isotope signal. Diagenetic processes can alter the oxygen isotopes (e.g., Vickers et al. 2022), but belemnite ∂¹⁸O is not considered in detail in this study. Material for calibration is scarce in south German Jurassic rocks. Belemnite ∂¹³C data exist for the Upper Kimmeridgian Pseudomutabilis (Blendinger and Mäuser 2021) and Beckeri-Zones (Malm ε and lower ζ) (Stevens et al. 2014). These are 0–1‰, values which are similar to material from the northwestern Boreal realm (England: Vickers et al. 2022) but slightly higher or lower than the micrite ∂¹³C in which they are embedded (Figs. 11c, 14). For the Oxfordian and Lower Kimmeridgian, no shell material ∂¹³C is available from the Franconian Alb, but in the Swabian Alb Upper Oxfordian belemnite, ∂¹³C is about 3‰ more negative than the micritic limestone in which they are embedded. The micrite values are similar for the few bulk rock samples obtained from the same interval of the Franconian Alb and show a very consistent ∂¹³C of ca. + 2,5 to + 3‰ (Fig. 14).

Fig. 14

Comparison between belemnite isotope data and the micritic matrix. Samples from Plettenberg (5 belemnites: Wierzbowski 2004; 5 matrix: Ruf et al. 2005) and Müllerfelsen (2 matrix) are from the Oxfordian (Malm α), the Wattendorf data from the Upper Kimmeridgian Pseudomutabilis Zone (2 belemnites, 21 matrix: Blendinger and Mäuser 2021), the Nusplingen data are from the Upper Kimmeridgian Beckeri Zone (25 belemnites, 6 matrix; Stevens et al. 2014). For location, see Fig. 2

Karst dikes–host rock relationships. Limestone samples from the periphery of, and clasts within, a karst dyke filled with unconsolidated red clay and fine quartz sand show isotope values indistinguishable from the undisturbed stratigraphic equivalent (Supplementary Appendix 5, cf. Fig. 4).

Isotope log correlation is possible even for ∂¹⁸O at the hundred meters to kilometer scale in the Swabian Alb area (Ruf et al. 2005), but less so at a larger scale. Within the reef facies of the Wattendorf area, intervals can be correlated based on the isotope log and calibrated at a cored lithologic marker interval. The reef shows an internal topography at the ten meter scale (Fig. 8b). The presence of reef limestone below the adjacent basinal facies is indicated by the less negative ∂¹⁸O, comparable to that of the adjacent reef and no surprise in view of the palaeogeographic position at the edge of a reef belt (Hegenberger and Schirmer 1967; Fig. 2b). The biostrome facies shows the typical serrated ∂¹³C log, which allows mapping of lateral thickness changes, but a peak-by-peak correlation is not possible (Fig. 13a). A similar negative ∂¹³C peak is developed on the Swabian Alb (Fig. 13c) as a widely correlatable interval (Ruf et al. 2005) consisting of stromatolitic facies in a biostrome limestone facies and which is probably a slightly older calcareous equivalent of the Franconian biostrome facies (Fig. 13b). Correlation within the Malm d on the southern Franconian Alb is not possible at the bed level, but shows local ∂¹³C anomalies (Frankenschotter section, Fig. 13b).

The hundred kilometer scale shows that a correlation based on absolute values is not feasible (Fig. 13c), but the south German and Swiss sections have similar thickness and the covariance of ∂¹³C and ∂¹⁸O in common. The north German Langenberg section (Zuo et al. 2018) is considerably thinner and the covariance is less well developed. In all cases, the oxygen log shows similar minimum ∂¹⁸O around -6‰ and maximum ∂¹³C around + 2‰. The logs show the highest values in the Lower Kimmeridgian and the lower part of the Upper Kimmeridgian. In the Alps (Oxfordian-Tithonian Quintner Kalk (Quinten Limestone) of the Helvetic nappes in Switzerland), thicknesses of the Kimmeridgian-Tithonian carbonates are similar to those of the study area. ∂¹³C values are around 2‰ in the Kimmeridgian followed by an a gradual decrease to ca. 0.5‰ in the Tithonian (Weissert and Mohr 1996). Unfortunately, oxygen isotope data were not reported in this study.

Strontium isotopes

Strontium isotope ratios range from 0.70667 to 0.71146 and show many values differing from the presently best estimate for Kimmeridgian seawater (0.707685–0.707696; Wierzbowski et al. 2017) (Fig. 15a, c). Six samples with ratios below the lower estimate for Kimmeridgian seawater are from the Malm ε and one sample from the Malm δ of the northern Franconian Alb, the samples with the highest ratios are from the Malm γ (Platynota Marl and Bankgruppe A) from the Ludwag quarry. Strontium concentrations (Fig. 15a, c) are nearly always < 100 ppm in dolomite, which appear as typical values (Vahrenkamp and Swart 1994), but can reach ca. 500 ppm in limestone. Dolomite and limestone Sr concentration data only overlap in the case of bedded, partly microcrystalline dolomite (from the Plattenkalk and biostrome intervals). These relationships are no surprise and are in line with findings from other diagenetic carbonates (e.g., Bathurst 1971). Sr isotope values do not co-vary with oxygen and carbon isotopes (Fig. 15b, d) or dolomite stoichiometry (Fig. 15e).

Fig. 15

Strontium isotopes of Lower Kimmeridgian (a) and Upper Kimmeridgian carbonates (c) of the Franconian Alb. A large portion of the samples shows values slightly elevated with respect to Kimmeridgian seawater. The values > 0.07080 are from the Malm γ of Ludwag. Sr isotopes and oxygen and carbon isotopes show no covariance (b, d). Lacking correlation indicates that the content of near-stoichiometric dolomite in dolomite samples has no effect on stable isotope and Sr isotope values (e, see also Fig. 9d). Kimmeridgian seawater after Wierzbowski et al. (2017)

Discussion

Provenance of lime mud

Because the analyzed Jurassic limestone is largely micritic, the origin of the lime mud first needs to be discussed. Here, particularly stratigraphic parameters are of importance. Several potential sources exist for the Jurassic lime mud. Reef-derived lime mud winnowed by currents and storms have been invoked (Viohl 2015a, b, c), but the reefs completely lack current induced sedimentary structures, and this places them largely into calm waters. Most importantly the bedded basinal limestone does not show any thickness asymmetries but laterally constant thickness including individual beds, even in small depressions in the larger reef tracts (e.g., Malm β of the Ludwag area: Antonianidis et al. 1972; Treuchlingen Marble: Meyer and Schmidt-Kaler 1994). This makes the reef tops as dominant source for lime mud unlikely. Another option is that the Upper Jurassic lime mud originated as a nanoplankton ooze. Calcareous nanoplankton does occur and has been documented mainly from the Plattenkalk intervals (Viohl et al. 2015a), but is questionable as far as rock forming quantities are concerned, in which case an uncemented ooze similar to chalk would form. Such chalky beds do exist in the Malm ε and ζ (Meyer and Schmidt-Kaler 1994), but they are stratigraphically subordinate. This makes it more likely that the Jurassic micrite is largely an inorganic precipitate, similar to what has been proposed recently for Bahamian lime mud (Geyman et al. 2022). The higher thickness of the reefs was probably caused by automicrite produced by microbes (e.g., Meyer and Schmidt-Kaler 1994) and the abundant sponges.

The potential factors influencing the primary isotope signal of the Jurassic micrites

After deposition, all carbonates undergo some sort of diagenetic modification, which can, at least partly, be deciphered from isotope data. The oxygen and carbon isotopes of the studied rocks show the typical features of “diagenetic” carbonate: negative ∂¹⁸O, which cannot be a primary marine signal, coupled with a “marine” ∂¹³C. Negative oxygen isotopes are usually interpreted as freshwater influence during diagenesis, high temperatures, or brackish conditions during deposition (Table 1), or a combination of these. Freshwater diagenesis has been proposed for some reef rocks (Koch and Schorr 1986), but the basinal micrites, all with negative ∂¹⁸O, have never been subaerially exposed in the Kimmeridgian and Early Tithonian. It might be argued that the negative ∂¹⁸O is the result of wholesale isotope resetting by freshwater during post-Tithonian emergence and karstification, but this appears an unlikely scenario. The isotope signal of the host rock in the vicinity of karst dikes preserves similarly high ∂¹⁸O (Supplementary Appendix 5) as the unkarsted sections (Fig. 4). High burial temperatures can also be ruled out, because of the maximum temperatures of < 60 °C and the ∂¹⁸O log shape, indicating the reverse of what would be expected from temperature increasing with depth. In this case, more negative ∂¹⁸O would be expected in the lower part of the succession. Freshwater discharge would cause more negative isotopes in the shallowest parts, the reefs, but the opposite is observed. This means that alternatives for the alteration of the original sediment parameters have to be considered.

Table 1 Effects of diagenetic processes on oxygen and carbon isotopes, modified from Colombié et al. (2011)

The depositional CaCO ₃ polymorph

Micritic limestone can have different precursor polymorphs, and this is important when interpreting isotope data. Because co-precipitated CaCO₃ polymorphs have different ∂¹³C and ∂¹⁸O values, the isotopic differences could potentially be used to distinguish between aragonite, LMC, and HMC. Aragonite ∂¹³C is 1–2‰ more positive than co-precipitated LMC (such as that of the belemnites) (Swart 2015). High-Mg calcite is more positive than LMC, too (Jimenez-Lopez et al. 2004, 2006), but in recent sediments much less than aragonite (Swart 2015). Modern chemogenically precipitated aragonite has a ∂¹³C ca. + 4‰ (Geyman et al. 2022), but such fairly positive values might be influenced by photosynthetic effects during LMC micrite formation (cf. Table 1). The largest portion of the Jurassic ∂¹³C values is between + 3‰ and − 2‰ (Fig. 10a), which is a much larger spread than that expected from primary polymorph differences. The only argument against a pure LMC as depositional polymorph is the comparison with the chalk, but its diagenesis may have been controlled by mechanisms different from that of the Jurassic carbonate. High-Mg calcite, only occasionally preserved in the rock record (Stanieda 2016), would show incongruent recrystallization to calcite and (micro) dolomite (Sperber et al. 1984), but the dolomite content in limestone is often below detection limit in XRD. The belemnite isotopes could be taken as reference for a marine LMC, but belemnites have notoriously lower (1–3‰) ∂¹³C than contemporaneous LMC brachiopod and bivalve shells (Wierzbowski 2002, 2004; Alberti et al. 2022), and possibly neither of these fossil groups can provide a “true” marine LMC ∂¹³C due to unquantified vital effects. An Upper Jurassic LMC lime mud could have had a ∂¹³C of 2–3‰, an aragonitic lime mud 4–5‰. Consequently, the dominant precursor polymorph cannot be identified based on isotope data alone. According to Munnecke et al (2001) and Munnecke and Westphal (2005), it is likely that the Jurassic lime mud was mainly aragonitic but all potential aragonite was lost diagenetically. Quantitative ultrastructure studies (cf. Munnecke et al. 2023) could reduce the uncertainty.

Fluid flow versus diffusion

The next question is what process can cause diagenetic alteration of the original isotope signal, and in this respect, petrophysical sediment and rock properties are of importance. Micrite can have a very high depositional porosity, but has always very low permeabilities. Fluid flow, the conventionally assumed reason for diagenetic alteration of carbonates, is theoretically possible and advection of water through micrite dominated carbonates has in fact been documented in Kramer et al. (2000) and Staudigel and Swart (2019). The micron-sized particles are, in a (pore)water saturated state, surrounded by a film of water (the so-called pendant, connate or irreducible water). This film is at least 0.5–1 μm thick (Zhai et al. 2021). The very small pore spaces in the studied micrite, a few microns (e.g., Munnecke et al. 2008, 2023), suggest that the larger proportion of pore fluids was pendant water. The water–rock ratio between the connate water film and the crystals is very low and would, in the case of recrystallization, only allow for resetting of ∂¹⁸O and ⁸⁷Sr/⁸⁶Sr (Banner and Hanson 1991). For instance, in a pure fluid-flow scenario, dolomite could only form when continuous replacement of the pendant water takes place to provide enough Mg²⁺ ions. Pendant water can, however, be displaced only by high pressure differentials to overcome capillary (adhesive) forces. This makes it unlikely that fluid flow played a role in diagenesis of micritic carbonate including the unconsolidated state of the sediment, and diffusion is a more likely process in the dislocation of dissolved matter (Swart 2015; Munnecke et al. 2023).

Biochemical pore water zonation

Because fresh water and high temperature diagenesis cannot explain the isotopic parameters of the studied Jurassic rocks, an alternative needs to be developed to explain the overall trend towards more negative isotope values. Decomposition of organic matter in the sediment, derived from benthos and the water column, probably played an important role in the isotope shifts of the studied carbonates. The general process is that part of the organic material is decomposed by microorganisms in the oxic zone, another part in suboxic and anoxic pore water environments (Coniglio 1989; Joachimski et al. 2002; Munnecke et al. 2023). Aerobic organic matter degradation causes a negative pH shift and is associated with carbonate dissolution (e.g., Reuning et al. 2002; Freitas et al. 2022). It is assumed that this occurs in the upper few millimeters or centimeters of carbonate sediment, and also may occur deeper, in the zone of anaerobic oxidation of methane (Fig. 16a, b; Munnecke et al. 2023).

Fig. 16

Diagenetic processes affecting the isotope signal of carbonates. a Generalized pH-log and diagenetic processes in the shallow subsurface of the Arctic, from Freitas et al. (2022). b The presumed processes in a low latitude micritic carbonate (from Munnecke et al. 2023) and two options for the localization of aragonite dissolution and re-precipitation. c The effects of in situ recrystallization and translocation, respectively, on the isotope signal. It is here assumed that only aragonite dissolves and reprecipitates. Yellow bars for dissolution, green bars for precipitation zones. For explanation, see text

Dissolved carbonate can be translocated and re-precipitated in zones of elevated pH, but with uncertain depth in the sediment column. It can either re-precipitate adjacent to the aerobic zone of dissolution and partly be recycled to the seawater, or above and below the zone of anaerobic oxidation of methane (Fig. 16a, b; Munnecke et al. 2023). Because of the link between organic matter degradation in the oxic zone and the lowering of pH (Freitas et al. 2022), it can be assumed that the higher the biochemical activity, the more the pH is lowered, and the more carbonate can be dissolved.

The zone of carbonate precipitation is characterized by elevated pH, which causes a decrease of ∂¹⁸O and ∂¹³C of pore waters (Zeebe 1999, 2007; Zeebe and Wolf-Gladrow 2001). The higher the biochemical activity in this zone, the higher the pH, and the more ∂¹⁸O and ∂¹³C of the pore fluid are shifted to negative values. Carbonate minerals precipitated from such pore fluids will ideally be in equilibrium with the pore fluid and, therefore, the isotopic composition of the source material is lost. The two processes, translocation versus in situ recrystallization, result in different stable isotope signatures at the micron scale. Carbonate which is transformed into LMC in situ will, due to the very low porosity and resulting low water–rock ratio, have their ∂¹⁸O (and ⁸⁷Sr/⁸⁶Sr) reset, but not ∂¹³C (cf. Banner and Hanson 1991). The original ∂¹³C signal can so be preserved, and the diagenetic system for this part of the matrix is (almost) closed (Fig. 16c). Cement micrite will acquire the negative ∂¹³C signal derived from anaerobic sulfate reduction, which implies an (almost) open diagenetic system (Fig. 16c). The net result is that the higher the amount of translocated cement in carbonate is, the more the original bulk isotope signal will be shifted to more negative values. Because 50% depositional porosity and, consequently, cement volume, appear a realistic estimate (Munnecke et al. 2001), the high amount of isotopically negative cement will cause a covariance of oxygen and carbon isotopes of the bulk rock isotopes, and this is what is, in fact, observed in the studied rocks. Cement will shift the depositional ∂¹³C signal of the bulk rock to approximately the average value between the depositional and the diagenetic signal. The presence of pyrite, which is a diagenetic product of this process (Berner et al. 1985) in all studied rocks, suggests that this is a viable scenario.

Limestone and dolomite lack evidence for compaction, but it is possible that selective compaction of the matrix occurred before deformation of fossils set on and is the effect of aragonite dissolution in the uppermost part of the sediment column. It is not entirely clear, however, whether the marly interbeds were the primary cement source in the Jurassic (donor limestone in Munnecke et al. 2023). In this case, the marls are a diagenetic residue of insolubles and non-aragonitic carbonate. Alternatively, bedding records at least partly a primary alternation of clay rich and clay poor carbonate (cf. Ruf et al. 2005).

If the assumption that carbonate cement in micrite is entirely derived from intrastratal dissolution and re-precipitation is correct, then the observed rock thicknesses are significantly lower than the true sedimentation rates of lose sediment. The Jurassic accumulation rate was about 6 cm/kyr (ca. 3 m.a. for ca. 200 m Malm β including Malm δ, cf. Figs. 3 and 4), but actual sedimentation rates may have been (significantly) higher. In the case of 50% of depositional porosity and an intraformational source of micrite cement, the original sediment was twice as thick, but could have been significantly thicker if dissolved carbonate was recycled to the seawater (cf. Munnecke et al. 2023).

The effects of the zones of lowered and increased pH, respectively, on micritic sediment are summarized in Fig. 16a and b. The zones of dissolution and re-precipitation are here, including Fig. 17, depicted as vertically superimposed. The depths where these biochemically induced diagenetic modifications occurred are not directly recorded in the studied rocks, but probably were located centimeters to meters or decameters below the sediment–water interface, such as in modern tropical, temperate, and Arctic settings (Freitas et al. 2022; Reuning et al. 2022). The depth of the sulfate–methane transition zone largely depends on the sedimentation rate (Egger et al. 2018) and has a depth of < 20 m for sedimentation rates > 15 cm/kyr. The effects of the porewater zonation on the isotope signal will be discussed below.

Fig. 17

Diagenetic modifications of the isotope signal of Upper Jurassic micritic limestone, assuming that the overall trend toward negative values is mainly controlled by pH, see text for discussion. The depositional lime mud is considered to be mainly aragonite with an isotope signal shown by the dashed lines in the right diagram. (a–d) Different degrees of resetting of the original signal, related to different degrees of anaerobic sulfate reduction and changes in pH. All δ¹³C modifications are caused by imported cement carbonate, whereas the in situ part retains the original δ¹³C signal

Uncertainties

The assumption that pH is the only control on the oxygen and carbon isotopes of the studied rocks is certainly an oversimplification. Other processes operate as well. First, the Mg/Ca ratio in porewater is important for the type of carbonate mineral which is dissolved and precipitated (Reuning et al. 2006). Second, the ∂¹³C of the dissolved inorganic carbon (DIC) in porewater depends on the origin of the DIC. Dissolved inorganic carbon depleted in ¹²C can form in the zone of methanogenesis and can be incorporated in diagenetic carbonates resulting in very positive ∂¹³C values (e.g., Irwin et al. 1977). Third, dissolution of metastable carbonates and re-precipitation of more stable carbonates can occur simultaneously in the same zone and shift the isotopic composition of pore water DIC toward higher ∂¹³C (e.g., Hu and Burdige 2007). Fourth, the DIC ∂¹³C of pore water is strongly influenced by the ∂¹³C of the dissolved metastable carbonates. Observations and modeling of diagenetic processes show that oxidation of organic matter will normally not alter the ∂¹³C of bulk carbonates significantly (e.g., Swart and Eberli 2005; Ahm et al. 2018). The δ¹⁸O of diagenetic carbonates will strongly be influenced by the temperatures of the pore waters, which might be colder compared to sea surface temperatures, but this would cause a positive shift of δ¹⁸O. These counterarguments are mainly based on observations from modern environments, which might be an imperfect analog of the Jurassic carbonates studied. The expectations from modern analogs would imply minor modifications and possibly positive shifts of ∂¹³C and δ¹⁸O. The contrary is observed, however, in the studied carbonates. Because it is very unlikely that the negative isotope shifts were caused by elevated temperature and/or freshwater diagenesis, the effect of pH possibly played a more important role in ancient carbonates than previously assumed. Even microbes may have evolved and possibly caused a different biochemically induced diagenesis in the Jurassic than in the Recent.

The belemnite isotope record can provide information about temporal and spatial changes of the seawater DIC in the upper part of the water column (Vickers et al. 2022), but the similarity of Oxfordian and Late Kimmeridgian belemnite values (Fig. 14) indicates that these changes were only minor, possibly < 1‰. This makes it unlikely that secular variations of seawater DIC had a measurable effect on the studied carbonates.

pH-controlled micrite diagenesis

The uncertainties inherent to this model have been outlined above, and for reasons of simplicity, pH is considered the principal factor influencing the isotopic composition of the bulk micrite. Figure 17 is based on the assumption that Jurassic lime mud had a ∂¹³C of ca. + 3‰ and a ∂¹⁸O of ca. + 1‰, similar to a Modern Bahamian lime mud. Four types of diagenetic modification of the isotope signal can be distinguished. The first is seen in the reef dolomite and reef limestone, which preserve the highest values of ∂¹⁸O and ∂¹³C (Fig. 17a), and the lowest part of the bedded limestones (Fig. 4). The positive ∂¹³C indicates a moderate pH increase in the precipitation zone. In this case, possibly seawater sulfate was the only matter which could be utilized by the microbes, because all organic material (reefs) had decayed before it became relevant for diagenesis. Vital effects of microbes in reefs, which should result in negative ∂¹³C (e.g., Swart 2015) did either not develop or were not preserved.

The second type (Fig. 17b) is represented by the bedded limestone with only slightly lower ∂¹³C but ∂¹⁸O of -5 to -6‰, the lower range of values observed. The basinal Malm γ and parts of the Malm δ represent this type (Fig. 4). The only slightly lower ∂¹³C indicates that cement was precipitated under similar conditions as in the first case, but the shift to more negative ∂¹⁸O indicates resetting of all carbonate ∂¹⁸O under increased pore water pH with respect to case 1.

The third type (Fig. 17c) is represented by low ∂¹⁸O similar to case 2 and ∂¹³C values around 0‰. The similar ∂¹⁸O values indicate that pH in the precipitation zone did not change with respect to case 2 but shifted the ∂¹³C values of the bulk rock by ≥ 2‰. The original content of organic material could have been higher than in cases 1 and 2 leading to increased contribution of ¹²C from anaerobic sulfate reduction. This type is best represented in the drill core data of the Malm e (Fig. 4a), a transition from type 2 to type 3 is documented in the upper part of the Treuchtlingen Marble (Fig. 12).

The fourth type (Fig. 17d) is represented by the most negative ∂¹³C, but relatively constant ∂¹⁸O around − 6‰. This type occurs mainly in the interval of the ∂¹³C anomaly, but also in some Plattenkalk (Fig. 4). The isotope signal reflects the strongest alteration observed and the highest contribution of ¹²C from anaerobic sulfate reduction. The fibrous cement ∂¹³C from these rocks in larger cavities (as low as − 10‰; Fig. 10a) is considered an equivalent of the blocky crystals cementing dolomite. These cements probably provide evidence for the isotopic composition of the micrite cement in limestone with bulk ∂¹³C around − 2 to − 5‰ (such as in the upper part of the logs of Fig. 4). The blocky crystal habitus is in line with a shallow burial marine origin (Melim et al. 2002). Because dolomite is the least soluble carbonate, the pH in dissolution zone must have been so low that even dolomite could be dissolved (Supplementary Appendix 2c). Dolomite dissolution may have affected non-stoichiometric dolomite more than the somewhat less-soluble stoichiometric dolomite (Rodriguez-Blanco et al. 2015). The trends observed in the isotope diagrams (Fig. 10) are, therefore, most likely purely related to different degrees of biochemical diagenesis.

Calculations show that pore fluid ∂¹⁸O decreases with increasing pH and shows a negative plateau at a pH between ca. 7,5 and 10 (Swart 2015). In modern sediments, pH does not normally exceed slightly more than 8 (Soetaert et al. 2007), and the nearly vertical trend of ∂¹⁸O in the isotope diagrams (Fig. 10) is explained as effect of the negative ∂¹⁸O plateau rather than a “meteoric calcite line” (e.g., Lohmann 1987). The different diagenetic zones were laterally and vertically not consistent. This is indicated by isotopic variation and covariance at the hand specimen scale (Fig. 11b, c), positive anomalies in the vicinity of the reefs (Supplementary Appendix 4a) and the poor correlatability of the ∂¹³C anomaly of the biostrome facies (Fig. 13a, b). An extreme case of lateral changes in biochemically induced diagenesis can be seen in the local preservation of organic material in Plattenkalk, in which the black rock passes laterally within few tens of meters into light colored, organic free Plattenkalk (Munnecke, pers. comm., 2023). The poor correlatability on a peak-by-peak basis within the study area is probably also the result of development of small isolated basinal areas (cf. Figure 2b) with slightly differing avenues of biochemically induced diagenesis. By contrast, the more open shelf situation of the Swabian Alb can account for the better correlatability documented in Ruf et al. (2005). The almost identical thicknesses in Switzerland and the Franconian Alb and the covariance of oxygen and carbon isotopes (Fig. 13c) suggest that the primary mode of deposition—lime mud settling through the water column—and diagenesis was comparable, and that varying degrees of biochemically induced diagenesis (Fig. 17) account for the isotopic differences.

⁸⁷Sr/⁸⁶Sr can be reset at similarly low water–rock ratios as ∂¹⁸O (Banner and Hanson 1991) and the deviations of ⁸⁷Sr/⁸⁶Sr from the marine reference value do not forcibly indicate fluid flow. Instead, the elevated values can be interpreted as the result of the reaction of pore fluids with clay minerals, which are soluble at a wide range of pH (Werling et al. 2022). This also explains the lack of covariance between Sr, oxygen and carbon isotopes (Fig. 15b, d). Clay minerals are ubiquitous in the Jurassic micrites, particularly in the Platynota and Crussoliensis Marls, from which the highest ⁸⁷Sr/⁸⁶Sr ratios of this study come (Fig. 15a). The highly pure limestone of the Upper Kimmeridgian and Tithonian has very low clay content and, accordingly, normal marine values. The few values lower than seawater cannot be interpreted with the present data.

Dolomite

Because dolomite is an important constituent of the Franconian Jurassic, the measured isotope data require an explanation. If the assumption that the largest part of lime mud settled through the water column is correct, a different depositional polymorph in the laterally limited dolomite areas, dominated by draping beds (Figs. 5, 6b), is very unlikely. It is also unlikely that the dolomite is a microbially induced very-high-Mg calcite transformed into dolomite (Vasconcelos et al. 1995), because this type of dolomite should possess negative ∂¹³C (cf. Joachimski et al. 2002), but Upper Jurassic dolomite is always positive. Dolomite shows the lowest degree of ∂¹³C and ∂¹⁸O modification (Figs. 12, 13) and records, according to the model proposed here, very limited biochemical activity similar to case 1 of the limestone (Fig. 17a). Dolomite formation also precedes calcite cementation (in the biostrome facies, Fig. 10d, e). This could also be explained by the Mg/Ca ratios of pore waters that often decrease with depth (e.g., Rivers et al. 2012). The lack of compaction indicates that dolomitization occurred during or after cement import, and the fact that nearly all reef-derived debris is calcareous indicates that dolomitization was not a process at the sediment–water interface.

Because it is practically impossible to motivate a stable LMC to incorporate Mg²⁺ under ambient temperature and normal marine conditions (“the dolomite problem”), the most attractive interpretation of the Jurassic dolomitization process is the addition of Mg²⁺ during the transformation of metastable in situ carbonate (aragonite or HMC) and re-precipitation of translocated carbonate, when the CaCO₃ polymorph was still reactive. This is similar to experimental dolomitization of aragonitic ooids (Zempolich and Baker 1993), and this would mean that the diagenetic system was still open for Mg²⁺ and allowed diffusional (downward) transport from seawater into the sediment. This caused dolomitization of in situ sediment and dolomite cementation, where the translocated aragonite re-precipitated as dolomite cement (Supplementary Appendix 1c). Alternatively or additionally, Mg²⁺ migrated laterally and vertically (upward), via diffusion and fractures, and was derived from adjacent HMC stabilization. In fact, some small, isolated dolomite bodies are fracture related (Meyer and Schmidt-Kaler 1992), but the mass balance for the stratiform dolomite is uncertain. It is unknown how much if any HMC was in the original sediment to provide Mg²⁺, as result of Mg loss during stabilization to LMC (Fig. 16c), for dolomitization. The problem with this model is twofold. First, an originally mainly aragonitic reef composition is an untestable assumption, and HMC or LMC as precursor polymorph cannot be excluded. Second, experimental data indicate that aragonite recrystallization is impeded in an open diagenetic system exactly because of the abundance of Mg²⁺, and that aragonite recrystallization more likely occurs in a fairly closed system (Hashim and Kaczmarek 2021). The exact process of dolomitization, thus, remains obscure. The controls are, by contrast, more evident. It appears that the Upper Jurassic dolomite is restricted to reef complexes elevated by tens of meters above the surrounding basinal limestone (e.g., Ludwag; Fig. 6), whereas the Upper Kimmeridgian dolomitic biostrome facies also occurs in basinal positions (Fig. 7, 8), most likely caused by a general lowering of the relative sea level (e.g., Meyer and Schmidt-Kaler 1989). The Mg²⁺ content was possibly higher in the upper part of the stratified water column favoring dolomitization in the shallowest parts of the submarine topography.

The somewhat more positive ∂¹³C of dolomite in some reefs (Figs. 11a, b; 15a, b) is either a fractionation effect of HMC, due to the incorporation of Mg²⁺ into the calcite crystal lattice (e.g., Jimenez-Lopez et al. 2006), or an enhanced relic aragonite signal, either due to reduced volumes of imported cement or very limited pH increase during diagenesis. A similar effect for ∂¹⁸O, which is expected to amount plus 3–6‰ (e.g., Swart 2005), is not observed. Dolomite either was precipitated from pore fluids whose ∂¹⁸O differed from those causing CaCO₃ digenesis by exactly the amount of the presumed calcite-dolomite fractionation of 3–6‰. However, according to Swart (2015, p.19), “dolomites have equilibrium ∂¹⁸O values which are believed (italics by the author) to be about 3–6‰ heavier than LMC formed under the same conditions at 25 °C.…”. The exact fractionation, if any, is unknown because dolomite and calcite cannot be co-precipitated at low temperature (“the dolomite problem”), and the values for equilibrium fractionation with calcite largely derive from the extrapolation of high-T experiments (see discussion in Swart 2015, p. 19). Alternatively, dolomite and limestone show nearly the same oxygen isotope fractionation (Figs. 11b; 15a, b; 16a), which would support the theoretical calculations of Zheng (1999).

Conclusion

• 1. The Upper Jurassic micritic limestone of Southern Germany was deposited as lime mud settling through the water column with a ∂¹³C of ca. + 3‰, similar to modern chemogenically precipitated aragonite in tropical environments. Circumstantial evidence suggest aragonite as volumetrically important polymorph, but geochemical proof is lacking. The micrite lithified in a shallow burial environment (millimeters, centimeters, meters below the seabed) in which part of the lime mud, most likely aragonite, was dissolved, translocated by diffusion, and re-precipitated as low-Mg calcite cement.

• 2. Carbonate which recrystallized to LMC in situ retained its depositional ∂¹³C, but because cement amounts to ca. 50% and most likely precipitated from pore waters with elevated pH and the contribution of ¹²C from anaerobic sulfate reduction, the net result is a more or less strong negative shift of bulk rock ∂¹³C. Bulk rock ∂¹⁸O was shifted to negative values because the water–rock ratio was high enough to also allow resetting of ∂¹⁸O of the in situ carbonate during transformation to low-Mg calcite. The ubiquitous bulk rock covariance of ∂¹⁸O and ∂¹³C is the result of isotopically negative cement. Clay minerals interacting with the stationary pore fluids caused a locally elevated ⁸⁷Sr/⁸⁶Sr ratio of the bulk rock.

• 3. The exact process of dolomitization is not clear, but dolomite formed most likely when in situ metastable carbonate recrystallized, and dissolved carbonate re-precipitated as dolomite cement, incorporating seawater-derived Mg²⁺. Because the dolomitized beds represent the most elevated parts of the submarine topography, the Mg/Ca ratio was most likely higher in shallower depths than in deeper intervals of the stratified water column.