Introduction

Limestone-marl alternations have a striking appearance because of the lateral continuity of beds and interbeds, and the repetitive bimodal intercalation of the two distinct lithologies. While they traditionally have been interpreted as archives of bimodally fluctuating environmental conditions (Seibold 1952; Einsele 1982; Elrick and Hinnov 2007), doubt has long been shed regarding their environmental origin, and the effects of diagenesis have been discussed controversially (Sujkowski 1958; Hallam 1964, 1986; Ricken 1986; Möller and Kvingan 1988; Munnecke and Samtleben 1996; Boulila et al. 2010, 2011; Mattioli et al. 2011). The main reason for controversy is that the two intercalated lithologies in such rhythmites have undergone distinctly different post-depositional processes (Ricken 1986), so-called differential diagenesis (Reinhardt et al. 2000; Westphal et al. 2000). Specifically, limestone beds have been cemented during early diagenesis by import of calcium carbonate cement, and they are preserved largely uncompacted (Munnecke and Samtleben 1996). The interbeds, in contrast, at the same time have been subject to dissolution of calcium carbonate and were being compacted by increasing sediment loading (Ricken 1986; Munnecke and Samtleben 1996; Westphal et al. 2000). Differential diagenesis is initiated by changes in pore-water chemistry caused by microbially mediated decay of organic matter (Raiswell 1987, 1988; Walter and Burton 1990; Canfield and Raiswell 1991; Walter et al. 1993) which takes place in defined depth intervals, causing the resulting diagenetic patterns to develop parallel to the sea floor, and thereby in most cases also parallel to the sedimentary layers. According to Munnecke and Samtleben (1996) and Westphal et al. (2000, 2008a), the carbonate cement in calcareous rhythmites such as limestone-marl alternations that fills the pore space of the (later) limestones derives from selective dissolution of aragonite in the non-cemented layers (marls). This process takes place close to, and thus parallel to, the seafloor during early marine shallow-burial diagenesis.

The parallel orientation of sedimentary and diagenetic patterns generally makes the distinction between the sedimentary and diagenetic bedding difficult or even impossible. At the same time, differential diagenesis alters rock parameters such as carbonate content, color, and magnetic susceptibility differently in the cemented and uncemented lithologies. This is most apparent for carbonate contents that diagenetically increase in the cemented beds whereas at the same time they decrease due to dissolution in the interlayers. This is also the case for other parameters including porosity, permeability, palynomorph concentration, and stable isotope composition (Ricken 1986; Munnecke et al. 2001). These parameters can then no longer be used as the basis for cyclostratigraphic interpretation on the couplet scale (Westphal et al. 2008a, 2008b, 2010). In order to determine whether calcareous rhythmites reflect primary sedimentary differences, rock parameters that are inert to differential diagenesis need to be studied, such as palynomorph associations (as opposed to simple concentrations), and the ratios of diagenetically stable elements such as Ti/Al (Westphal et al. 2000, 2008a, 2008b, 2010).

An example is presented here that shows a distinct angle between two sets of coeval diagenetic bedding, one following sedimentary bedding, the other oriented oblique to sedimentary bedding. This example verifies and demonstrates in striking clarity that diagenesis can create bedding independent of primary depositional layers.

Materials and methods

The Miocene (Burdigalian to Lower Langhian) Banyalbufar Turbiditic Formation consists of 350 m of sediments filling an actively subsiding trough (Pomar et al. 1990). These sediments have been interpreted as calcareous turbidites deposited during a time of rapid deepening of the Valencia Trough related to Alpine tectonics (Rodríguez-Perea 1986–1987). The Valencia Trough is a NE–SW-oriented extensional basin between the Iberian Peninsula created during the late Oligocene to middle Miocene and the Balearic Promontory that forms the northeastern prolongation of the Betic thrust belt (Fontboté et al. 1990; Torres et al. 1993).

The Banyalbufar turbidites are exposed on the western coast of Mallorca (Balearic Islands, Spain) close to the village of Banyalbufar (39°41′29′′N, 2°30′50′′E; Fig. 1). They consist of mixed carbonate-siliciclastic silt to sand-grained sediment sourced from the rising Mallorcan orogens and from marine carbonate production. The outcrop studied here has been created during enlargement of a parking lot in 2009.

Fig. 1
figure 1

a Location of the village of Banyalbufar (black dot); Island of Mallorca, b close-up; outcropping Banyalbufar Turbidites are shaded in grey; outcrop studied is marked with a black dot

Outcrop description was complemented by detailed study of eight samples by means of light-microscope petrography and scanning electron microscopy (SEM). For the latter, the samples were mounted on SEM stubs perpendicular to the stratification, polished, and treated with hydrochloric acid (0.5 %) for 20 s for the cemented and 10 s for the uncemented samples, and afterwards coated with gold. Carbonate content was determined by the calculated difference between the measured TC and TOC values, determined using a LECO Carbon Analyzer. Carbonate is expressed as calcite (CaCO3 = (TC − TOC) × 8.33).

Results

The outcrop exposes 3 m of the stratigraphic succession along a width of 25 m. The succession is preserved as cemented and uncemented beds with an average couplet thickness around 10 cm (Fig. 2). The outcrop features two sets of cemented beds, one following the sedimentary layers, and a second one at an acute angle to the bedding: Whereas the sedimentary layers dip with 5–10° (average 8°, n = 5) in a northwestern direction (300–350°, average 342°), the oblique cemented beds dip with 20°–35° (average 27°, n = 7) in a southeastern direction (110°–150°, average 134°).

Fig. 2
figure 2

a Outcrop showing two bed-sets of rhythmic cementation; blue arrow: direction of sedimentary layers; black arrow: direction of oblique bedding; white arrow: later joint; dashed line: joint parallel to oblique cemented beds (height of wall ca. 3 m). b Detail of sedimentary layering running horizontally through both sets of cemented limestones (pencil of 18 cm length for scale); blue arrow: direction of sedimentary layers; black arrow: direction of oblique bedding. c Outcrop close to (a) (height of wall ca. 1.3 m). In the lower part of the outcrop, diagenetic bedding follows the sedimentary layers of the turbidite succession; in the upper part, a second, oblique direction of diagenetic bedding occurs. d Detail of sedimentary layering and small-scale cementation in direction of the joints (black arrows). Cemented layers show calcite mineralization on broken surface

The succession shows three distinct sets of joints (i.e., tectonically induced disturbance without visible displacement). The most conspicuous joints exposed in the outcrop are oriented parallel to the oblique beds (dotted line in Fig. 2a). Close to these joints, the oblique beds are most continuous and become less prominent with distance from the joints. Another prominent joint direction unrelated to cementation lies almost orthogonal to the latter (white arrow in Fig. 2a).

The sedimentary stratification is clearly visible in the uncemented layers as well as in cemented ones where weathered (Fig. 2b, d). For both sets of beds, the color of the cemented beds is whitish-grey, whereas the uncemented interbeds show a yellow-brownish tone.

Generally, the sediment composition of the succession is rather uniform and consists of well-sorted grains dominated by detrital quartz, planktonic foraminifera, and sponge spicules in a mud matrix for both cemented and uncemented beds (Fig. 3a, b). The couplets are characterized by a slightly coarser-grained base and a finer-grained top layer. Carbonate contents of cemented beds average at 81.2 % (range, 80.8–81.4 %; n = 3), those of the uncemented beds at 60.4 % (range, 56.7–63.2 %; n = 4).

Fig. 3
figure 3

a Thin-section photograph of cemented bed (packstone) consisting mainly of detrital quartz grains and planktonic foraminifera. b Thin-section photograph of non-cemented interbed (packstone) also consisting of detrital quartz grains and planktonic foraminifera. Compaction is indicated by texture. Higher porosity is indicated by blue resin (all samples where impregnated with blue resin before preparation of thin-sections). c SEM micrograph of cemented non-deformed foraminiferal test in a cemented bed. Pyrite framboids are present in the microspar matrix. d SEM micrograph of broken foraminifera test in a non-cemented interbed

Petrographically, the cemented beds of both bed-sets (parallel to sedimentary layering and oblique ones) are indistinguishable from each other, and the same holds true for the uncemented beds of both sets. All uncemented beds are classified as packstones, and all cemented ones as packstones to poorly washed grainstones.

Diagenetically, the two bed-sets are also indistinguishable. For both sets, the cemented beds contain low amounts of interparticle and intraparticle sparry cement, whereas the uncemented layers show a higher porosity, compactional features, and dissolution seams (Fig. 3). Where the horizontal and oblique cemented beds merge (Fig. 2b), they do not show a lithological boundary that would imply that one set postdates the other; they appear to have undergone cementation synchronously. The absence of significant compaction features in the cemented beds (Fig. 3c) implies that the cementation process was early and prior to significant sedimentary overburden.

Interpretation

As was pointed out earlier (Jenkyns et al. 1990, p. 57), in the Banyalbufar Formation, “the composite layering […] appears to be largely due to differential cementation, but this probably reflects original bedding.” This interpretation of a bedding-parallel cementation has been confirmed here for one set of cemented beds. The new discovery of the additional oblique bed-set reveals an unequivocally diagenetic feature that is not predefined by sedimentary layering.

As mentioned above, cemented beds in rhythmic calcareous successions in most cases follow the sedimentary layers and are oriented parallel to the sea floor of the time of deposition (Canfield and Raiswell 1991). Combined with sedimentary anisotropies, e.g., in porosity and permeability, diagenetic zones caused by microbially mediated decay of organic matter in many cases forces diagenetic beds to match the orientation of the sedimentary layers. The layer-parallel bedding of the Banyalbufar succession also follows primary anisotropies. At the same time, however, the contemporaneously formed oblique bedding also follows another anisotropy present during early diagenesis. A process that potentially produces such oblique pathways for fluids are syn-sedimentary to early post-sedimentary tectonic movements affecting the consolidated, but not yet cemented, sediment (Fig. 4). The stiffness of the sediment must have allowed small joints to form as opposed to ductile deformation. This is in accordance with the silty to sandy sediment making it more prone to brittle than to ductile deformation, the latter being more typical for muddy sediment. The joints observed in the outcrop studied have not displaced the sediment (Fig. 2d) but have had enough of an effect on the sediment properties to add an additional direction of the physical anisotropy of the sediment. This dual anisotropy led to synchronous cementation, resulting in cemented beds that intersect each other (Fig. 4). The microbially defined diagenetic zones were not tilted, neither were the sediments. They were oriented in their original horizontal position (parallel to the sea floor). The physical properties, however, were changed in the tectonic direction of the joints, thus distorting the diagenetic (microbial) zones according to the now bidirectional anisotropy of physical properties (Fig. 4). Dissolved calcium carbonate (likely derived from dissolution of aragonitic constituents in the uncemented layers, cf. Munnecke and Samtleben 1996) migrated by diffusion into a cementation zone, the shape of which is influenced by petrophysical properties (e.g., porosity and permeability), here determined by both sedimentary layers and tectonic joints (Fig. 4).

Fig. 4
figure 4

Sketch of the process forming the intersecting diagenetic bedding

The scenario requires a syn- to early post-sedimentary tectonic movement, which is in accordance with the tectonic situation in the Burdigalian. During deposition of the Banyalbufar Turbiditic Formation, the area was subject to extensional faulting with local rapid subsidence (Pomar 1979; Ramos-Guerrero et al. 1989; Wadsworth and Adams 1989). This interpretation is supported by the fact that oblique-oriented cemented beds are only observed in the vicinity of tectonic joints and they disappear with distance.

The question about the mechanisms or sediment properties that define the exact position of the cementation zone has not yet been conclusively answered (see discussions in Westphal et al. 2008a, 2008b, 2010). Canfield and Raiswell (1991) and Munnecke and Samtleben (1996) speculated that the zone of anaerobic methane oxidation represents the cementation zone. In modern sediments, this zone is located in shallow burial depths between the sulphate reduction zone and the methanogenesis zone and is characterized by an abrupt increase in carbonate supersaturation. A discussion of this hypothesis, however, is not within the scope of the present paper.

The Miocene turbidite succession exposed near Banyalbufar illustrates how anisotropies of different origin can influence early diagenesis in a way to produce intersecting sets of cemented beds. Cemented beds thus do not necessarily represent primary sedimentary layers, and, in an extreme case, can even mimic sedimentary layers at an angle to the true sea-floor orientation.

As the long-lasting controversy has demonstrated, bedding without such relation clearly cannot be the basis for astrochronology (see references in Westphal et al. 2010). This example of diagenetic bedding entirely unrelated to sedimentary layers demonstrates that creating a high-resolution, highly precise time-framework needs to be undertaken with great caution. Successions without clear proof of a primary sedimentary signal should not be chosen for GSSPs and other binding stratigraphic data points. Calcareous rhythmic successions are only reliable for sophisticated methods such as astrochronology if a primary, sedimentary origin has been proven independently and unequivocally.