Cyclic depositional model
The Upper Triassic Lofer cyclothems in the Northern Calcareous Alps and their elementary Lofer cycles were originally defined and described as unconformity-bounded, deepening-upward (i.e. transgressive) sedimentary cycles that are made up of three characteristic lithofacies types, referred to as Lofer cycle members A, B, and C (Fischer 1964). Member A consists of red or green marl or claystone that commonly contains pebble-sized clasts of penecontemporaneous limestone. Member B is characterized by stromatolites and/or fenestral laminites (loferites) with common occurrence of desiccation features. Member C is made up of carbonates with wackestone or packstone texture that contain shallow marine fossils, including locally abundant megalodontids and benthic foraminifera. Originally, members A and B were interpreted as tidal flat deposits that were formed in the supratidal and intertidal zones, respectively, whereas member C is characteristic to the subtidal zone (Fischer 1964). Subsequently, however, it was recognised that the ideal elementary Lofer cycle shows a symmetrical A-B-C-B’(-A’) pattern (Haas 1982). The regressive member B’ is similar to the transgressive member B but it is commonly mud-cracked, leached, and, in some cases, more intensely dolomitized (Haas 1982, 1994).
The typical Lofer cyclic successions were deposited on the inner (i.e. more coastal) part of the platform, characterized with a very gently sloping (~1°) ramp-like topography. The adjacent outer platform had a slightly steeper slope, stabilized by primarily microbial, and subordinately metazoan mounds ("microbial reefs") and oncoidal mounds accumulated by vigorous currents. Thus, although typical Lofer cycles did not form on the outer part of the platform system, nevertheless the outer platform could also influence the deposition on the inner platform via the so-called basket effect (Haas 1994, 2004).
During a lowstand, large areas of the inner platform become subaerially exposed and are subjected to karstification and pedogenic processes, while most of the outer platform remains submerged. At the onset of the transgression, as the sea level starts to rise and reaches the inner platform, first the terrestrial sediments and the erosional products are reworked and redeposited (member A). Subsequently, if the sea level rise is slow, a microbial mat-covered tidal flat environment is established (member B). In this scenario, there is no basket effect and the platform behaves as a ramp-like platform, leading to the development of an ideal Lofer cycle. On the contrary, when the basket effect applies and/or the sea level rise is too rapid, most of the inner platform immediately transitions into the subtidal zone where member C is deposited. This model may explain the occasionally missing A and/or B members. After reaching the maximum flooding, during the late highstand, normal regression occurs and, with the reduction of the available accommodation space, the inner platform transitions into a tidal flat covered by microbial mats again (member B’). If this process is too fast, member B’ may be missing entirely. In both cases, however, member B’ could also be removed by erosion, when forced regression takes place and the inner platform becomes subaerially exposed again at the end of the sequence. This leads to the development of a disconformity, occasionally with the presence of member A’ (Haas 1982, 1994, 2004).
These cyclic sea level fluctuations are hypothesized to be of a few metres to 15 m in amplitude (Fischer 1964; Haas 1994; Balog et al. 1997) and commonly associated with orbitally forced eustatic sea-level variations (Sander 1936; Schwarzacher 1954, 1993, 2005; Fischer 1964, 1975, 1991; Haas 1982, 1991, 1994; Schwarzacher and Haas 1986; Balog et al. 1997; Cozzi et al. 2005; Hinnov and Cozzi 2020). From the Late Carnian to the Rhaetian, the Dachstein platform was not affected by any significant tectonic deformation and only a prolonged, slow and gradually accelerating subsidence is assumed (Haas 1994).
Previously, the thermal expansion of sea water was hypothesized as the most likely driving force of orbitally-driven sea-level fluctuations during greenhouse climate states, supported by modeling that showed that this process can generate sea level variations up to 2 m (Schulz and Schäfer-Neth 1997). More recently, however, aquifer- and limno-eustasy is increasingly accepted as a more potent driving force as several studies suggested that the water-bearing potential of groundwater aquifers and lakes in greenhouse climate state is about equal to that of polar ice caps during icehouse intervals (Southam and Hay 1981; Hay and Leslie 1990; Wagreich et al. 2014; Peters and Husson 2015; Wendler et al. 2016).
Aquifer- and limno-eustasy continuously redistributes water between the oceans and continents. During the maximum of seasonality, the hydrological cycle strengthens, and water is delivered from the oceans to the continents via precipitation. The continental surface and subsurface reservoirs such as lakes and groundwater aquifers recharge, the water levels of the lakes rise while, at the same time, the global sea level falls. During the minimum of seasonality, the hydrological cycle weakens, and the continental reservoirs discharge via fluvial runoff. This would lead to a lowering of water level in lakes and concomitant sea level rise (Wagreich et al. 2014; Li et al. 2016; Wendler et al. 2016; Li et al. 2018). Several recent studies document the antiphase relationship between variations of sea level and lake water levels linked by orbital forcing in the Mesozoic (e.g. Wagreich et al. 2014; Li et al. 2018). One such study is especially relevant as it demonstrated the relationship between sea level variations observed in Lofer cycles in the Southern Alps and lake level variations in the Newark Basin (Hinnov and Cozzi 2020). In addition, ˝megamonsoon˝-driven aquifer- and limno-eustasy is proposed from the partly coeval intraplatform Csővár basin that was located on the outer part of the Dachstein platform system in the Transdanubian Range (Vallner et al. 2023).
Our analysis of the core Po-89 reveals that the thickest members C were deposited consistently during long eccentricity minima, when the climate was more balanced, and seasonality could reach its absolute minima. This finding is in agreement with the concept of aquifer- and limno-eustasy that predicts that the long eccentricity minima are associated with generally higher sea level and increased accommodation space, promoting the deposition of thick members C. In the opposite phase, during the long eccentricity maxima, although conditions may favour the formation of members A, they do not become thicker or more common in the record. As member A is pedogenic and redeposited in origin and prone to erosion, this observation does not contradict the model of aquifer- and limno-eustatic control.
Orbital forcing on the completeness of the Lofer cycles
The common occurrence of symmetrical A-B-C-B’(-A’) Lofer cycles was first described from the core Po-89, but incomplete cycles with depositionally missing members and/or truncated cycles that were affected by a higher degree of subaerial erosion were also encountered (Haas 1982). This observation raises the possibility of orbital forcing of the different completeness of the cycles preserved in the sedimentary record. To test this hypothesis in general, and to assess if the dominant cycle types and average cycle completeness differs at long eccentricity minima and maxima, here we apply time-series analysis on the cycle completeness. First, a completeness index was developed to categorize each cycle, based on the subdivision described by Haas (1998) (Table 3). Then, time-series analysis was carried out using the same data preparation steps and settings for the power and evolutionary spectra generation as for the analyses of other time-series.
Index value
|
Description
|
Cycle types
|
1
|
Ideal, symmetric, complete cycles
|
dABCB'A'd; dABCB'd
|
2
|
Incomplete cycles with depositionally missing members
|
dBA'd; dCB'd; dCA'd; dBCB'd; dACB'd; dABCA'd
|
3
|
Truncated cycles with higher degree of subaerial erosion at their top
|
dABCd
|
4
|
Incomplete and truncated cycles
|
dACd; dCd
|
Table 3 Completeness index used to test the possible orbital forcing of the different completeness of the Lofer cycles. d – disconformity, i.e. erosion between the different cycle members
We found that during the long eccentricity maxima the average of the completeness indices of the cycles was consistently lower than during long eccentricity minima, with only one exception (at a long eccentricity minimum between -147 and -177 m). The average value of all long eccentricity maxima is 1.5, in stark contrast to the average of 2.7 of all long eccentricity minima. The complete, symmetrical cycles are more common during the maxima, whereas the minima are dominated by the incomplete and truncated cycle type. The purely incomplete and purely truncated Lofer cycle types do not dominate any orbital cycle segments; they are nearly evenly distributed along the long eccentricity curve (Fig. 7).
The results of the time-series analysis of the completeness index are noisier but they are in agreement with the other time-series analyses performed. However, due to the lower resolution of the completeness index, the majority of the sub-Milankovitch cycles are not detectable by this method (Fig. 8).
Our results imply that the different completeness of the Lofer cycles is also driven by Milankovitch-scale orbital forcing, possibly through the changing pace of sea level variations controlled by the available water mass in the oceans. As a consequence of aquifer- and limno-eustasy, long eccentricity maxima are associated with less available water mass in the oceans and thus generally lower sea level. As a result, slower changes of sea level are expected that, together with the effect of lower sea level, would lead to a complete development of each facies types within the Lofer cycle. On the contrary, during long eccentricity minima, when the available water mass is greater, the sea level is generally higher, and subsequently the sea level changes are faster, the inner platform would remain mostly in the subtidal zone and enter the intertidal zone for only a short period of time that is insufficient for the members B and B’ to develop properly. In this case, commonly only a disconformity surface is formed that marks a brief subaerial exposure. These conditions also explain the detected higher thickness of members C during long eccentricity minima.
The time-series analysis of the completeness index revealed another peculiar phenomenon. A comparison of the evolutionary spectra (Fig. S2) allows the identification of three segments (from -94.4 m to -140 m, -240 m to -330 m, and -480 m to 502.1 m, respectively) where the original time-series yields lower levels of spectral power, as opposed to the higher levels of spectral power in the completeness index time-series. On the contrary, all other segments with higher levels of spectral power in the original time-series are found to exhibit less spectral power in the completeness index time-series. A plausible explanation of this antiphase relationship calls for oppositely changing variance in the original time-series and the completeness index time-series. Thus, when the variability among the different cycle types is higher, i.e. more types are alternating more frequently, then the variability within the Lofer facies types is lower. Then, fewer facies types and a more limited range of lithology, colour, and greyscale values are alternating, and do so with a lower frequency. The changing variability among these segments exhibited by the completeness index time-series is also apparent in Fig. 7.
We suggest that this pattern stems from the inherent properties of the genesis and preservation of Lofer cycles. The maximum variance in the Lofer facies types, as also reflected in the broadest range of lithology, colour, and greyscale indices, occurs when most of the Lofer cycles are complete and there is only limited deviation from the ideal type, i.e. minimal variance among the cycle types. It follows that changes in variance may also be paced by the orbital cycles as the dominance of the complete cycle type is proved to be controlled by the long eccentricity cycle. Since there are only three complete segments of higher and lower variability within the core, time-series analysis cannot offer a rigorous test for this hypothesis, yet the similar thicknesses of these segments may still hint at regular cyclicity. The thickness of these segments is 100, 90, and 150 m, respectively, which may form parts of larger cycles with periods of 190–240 metres. Notably, Schwarzacher (2005) also reported the presence of 190–240 m thick cycles from the bundles of Lofer cycles in the Leoganger Steinberge section in the Northern Calcareous Alps. A conversion of the thickness of this, as yet hypothetical cycle to time domain, using a sedimentary rate of 16.5 cm/kyr as determined in this study, yields a period of 1.15–1.45 Myr. Remarkably, there is a striking coincidence of this period with that of the ~1.2–1.3 Myr grand orbital cycle (Lourens and Hilgen 1997; Li et al. 2018; Boulila 2019), pointing to another level in the hierarchy of orbital control exhibited in Lofer cyclic successions.
Sub-Milankovitch cycles
Another notable outcome of our cyclostratigraphic analyses is the identification of 19 cycles in the sub-Milankovitch band, i.e. with periods of less than ~17–20 kyr (Table 2). The most prominent among them are the cycles of 13.5, 7, 5, 3.4, 2.4, and 1.48 kyr. Currently there are no established target cycles for the Mesozoic in this scale, which makes their matching to known cycles challenging. However, several recently published cyclostratigraphic studies report sub-Milankovitch cycles from the Mesozoic, therefore a comparison with their results may provide further insights, despite the lack of agreement on the driving forces of these cycles.
The most characteristic sub-Milankovitch cycles identified in other studies for the Mesozoic are the cycles of ~9–13, ~7–8, ~4–5, and ~1.5–2 kyr (Rodríguez-Tovar and Pardo-Igúzquiza 2003; Zühlke et al. 2003; Kent et al. 2004; Friedrich et al. 2005; Vollmer et al. 2008; Boulila et al. 2010; Wu et al. 2012; De Winter et al. 2014; Chu et al. 2020; Boulila et al. 2022; Hasegawa et al. 2022; Ma et al. 2022; Zhang et al. 2023a). Several recent studies propose a link of the ~9–13 kyr cycle with the so-called hemi- or semi-precession cycle (Rodríguez-Tovar and Pardo-Igúzquiza 2003; Vollmer et al. 2008; Wu et al. 2012; Chu et al. 2020; Zhang et al. 2023a). Theoretically, a cycle with half the period of the precession may arise from the insolation maxima due to the biannual passage of the Sun across the intertropical zone (Berger and Loutre 1997; Berger et al. 2006). In the context of the present study, however, an alternative explanation, that these cycles may arise from the occurrence of incomplete or truncated Lofer cycles, cannot be ruled out. The ~7–8 kyr cycle was linked to the Heinrich events (e.g. De Winter et al. 2014), even though the Heinrich events were originally described from the last glacial period and are thought to correspond to ice sheet melting (Heinrich 1988). However, recent studies suggest that this cycle may originate in the tropical areas, as the nonlinear response to precession-induced insolation variations and subsequent changes in El Niño frequency (e.g. Turner 2004; Ziegler 2009). The the ~4–5 kyr cycles are also linked to either the Heinrich events (Hasegawa et al. 2022) or the harmonics of the precession (Chu et al. 2020), whereas the ~1.5–2 kyr cycles are associated with either the 1.3–1.6 kyr Dansgaard–Oeschger (DO) cycles modulated by the precession cycle and its harmonics (Boulila et al. 2022; Hasegawa et al. 2022) or the ~1 kyr Eddy and ~2.3 kyr Hallstatt or Bray solar activity cycles (Hasegawa et al. 2022; Ma et al. 2022). The DO cycles were originally described from records of the last glacial period (Dansgaard et al. 1982) but recent studies also established the presence of a cycle with similar period in the Mesozoic (Boulila et al. 2022; Hasegawa et al. 2022; Ma et al. 2022). Moreover, the controversial Latemar couplets also appear associated with a ~1.7–2.2 kyr cycle (Kent et al. 2004).
Our results provide additional support to the conclusions of the studies cited above, as we suggest that the 8.7–13.5 kyr cycles likely represent the semi/hemi-precession cycles, the 7.1–7.77 kyr cycles the Heinrich events, the 3.4–5.6 kyr cycles also the Heinrich events or harmonics of the precession cycle, the 2.06–2.4 kyr cycles the Hallstatt or Bray solar activity cycles, and the 1.4–1.6 kyr cycles the DO cycles (Table 2). However, to explain the origin of the 1.72–1.84, 2.73–2.95, and 6 kyr cycles remains challenging. Tentatively we propose that the 1.72–1.84 kyr and 2.73–2.95 kyr cycles may be similar to the Latemar couplets with an as yet unknown driving force or, alternatively, may be related to either the DO cycles or solar activity cycles, whereas the 6 kyr cycle may represent another expression of the Heinrich events or harmonics of the precession.
Comparison with other Lofer cyclic successions
As the allocyclic vs autocyclic nature of the Lofer cyclothems has long been debated, a regional comparison is warranted and may yield additional insights. From the Alps to the Hellenids, we comprehensively collected the available data on the thickness of the individual Lofer cycles, the thickness and proposed period of the other identified cycles, and sedimentary rates within the Lofer cyclic successions (Table 4). To validate the hypothesis of orbital forcing, a high degree of similarity in the cyclic pattern of geographically distant study areas is expected.
Reference
|
Location of the studied section(s)
|
Thickness of an individual Lofer cycle
|
Identified cycles
|
Calculated sedimentary rate
|
Method(s) used
|
Schwarzacher (1954)
|
Loferer Steinberge, Northern Calcareous Alps, Austria
|
average 3.5 m
|
3.5 m (~20 kyr) & 15–18 m (~100 kyr)
|
—
|
suggested by bundling pattern
|
Fischer (1964)
|
Dachstein, Leoganger Steinberge, Steinernes Meer & Loferer Steinberge sections, Northern Calcareous Alps, Austria
|
5–6 m
|
~20, 50 & 100 kyr
|
~11 cm/kyr*
|
thickness divided by time & Fischer-plot
|
Schwarzacher and Haas (1986)
|
Loferer Steinberge, Steinernes Meer & Dachstein sections, Northern Calcareous Alps, Austria and cores T-5, Po-89 & Ut-8, Transdanubian Range, Hungary
|
average 3.5 m (Loferer Steinberge), 5.69 m (Steinernes Meer), 4.84 m (Dachstein), 2.21 m (T-5 core), 3.1 m (Po-89 core) & 4.29 m (Ut-8 core)
|
2.5–4 m (~20 kyr), 5–7 m (~40–45 kyr), 12–15 m (~100 kyr), 20–27 m (~150–200 kyr) & 45 m (~300 kyr)
|
~13–23 cm/kyr, most likely within 13.3–16.5 cm/kyr
|
thickness divided by time, Fischer-plot & Walsh power spectra on Lofer ABC indices from the Loferer Steinberge and the Hungarian core sections
|
Haas (1982, 1994, 2004)
|
Transdanubian Range
|
2–5 m, average 3.1 m
|
—
|
~15–16 cm/kyr
|
thickness divided by time
|
Balog et al. (1997)
|
Cores Ut-8, Zt-62, E-5, T-5, Td-4 & Po-89, Transdanubian Range, Hungary
|
1–5 m, average 3.1 m
|
2–3 m (~20 kyr), 5–7 m (~35–45 kyr), 12–14 m (~90–100 kyr), 33 m, 40–50 m (~400 kyr) & 99 m
|
~10–16 cm/kyr
|
thickness divided by time, Fischer-plot on cores Ut-8, Po-89, T-5, Zt-62 & Td-4 & Walsh and Fast Fourier Transform (FFT) power spectra on Lofer ABC indices from core Po-89
|
Cozzi et al. (2005)
|
Monte Canin section, Julian Alps, Italy
|
average 2.41 m
|
1.6–3.5 m (~20 kyr), 13.7 m (~100 kyr) & 54 m (~400 kyr)
|
~8–18 cm/kyr*
|
multiple power spectra and evolutionary spectra on greyscale indices created from field photographs
|
Schwarzacher (2005)
|
Loferer Steinberge, Leoganger Steinberge & Steinernes Meer sections, Northern Calcareous Alps, Austria
|
2–3.5 m
|
2–3.5 m (~20 kyr), 5–7 m, 9–13 m, 15–27 m (~100 kyr), 60–80 m (~400 kyr), 190–240 m
|
~15–20 cm/kyr
|
Lomb-Scargle analysis on greyscale indices created from field photographs from the Leoganger Steinberge
|
Haas and Pomoni-Papaioannou (2009)
|
Argolis Peninsula, Hellenids, Greece
|
1–5 m
|
—
|
—
|
—
|
Todaro et al. (2017)
|
Monte Sparagio section, Sicily, Italy
|
0.8–3 m
|
—
|
—
|
—
|
Hinnov and Cozzi (2020)
|
Monte Canin section, Julian Alps, Italy
|
average 2.41 m
|
2–3.1 m (~17–23 kyr), 4.46–8.3 m (~30–50 kyr), 17.3 m (~100 kyr) & 21.3–39.5 m (~400 kyr)
|
~15–17 cm/kyr
|
MTM power spectra, FFT evolutionary spectra, and tune and release method on the data from Cozzi et al. (2005)
|
This study
|
Core Po-89, Transdanubian Range, Hungary
|
average 3.1 m
|
2.9–3.6 m (~17.5–21.6 kyr), 3.85–7.1 m (~23.3–42.9 kyr), 13–20 m (~78.7–121.2 kyr), 32.3 m (~200 kyr) & 66.6 m (~404 kyr)
|
~16.5 cm/kyr
|
See Chapter 3
|
Table 4 Summary of published cyclostratigraphic analyses of Lofer cyclic successions, with the estimated thickness of individual Lofer cycles, detected periods and their matching with orbital cycles, calculated sedimentary rates, and the method(s) used in each study. If the sedimentary rate was not given in the original study, it is calculated here (denoted by an asterisk) from the known and matched periods of the detected cycles
The thickness of the Lofer cycles consistently falls between 1 and 6 m, with an average thickness of 2.5–3.5 m, whereas the inferred sedimentary rate is also consistent between 10 and 20 cm/kyr. Most studies suggest the precession cycle as the driving force for the basic cycle, in accordance with the earlier assumption (Sander 1936; Fischer 1991; Schwarzacher 1993), except one study that suggested obliquity forcing, although based on limited data only (Fischer 1964). Each study that was based on detailed cyclostratigraphic analysis successfully identified the short and long eccentricity cycles with periods of ~12–17 m and ~50–70 m, respectively, whereas five out of six studies also found cycles in the frequency band of obliquity (with periods between ~4–7 m) and a cycle with a ~30 m period. The similarity between the periods is remarkable. All the studies reported a robust short eccentricity signal exhibited in the bundling pattern of the individual Lofer cycles, whereas the long eccentricity signal appears weaker, more fragmented, distorted, or ambiguous. The latter finding is consistent with the observation of less clear and obvious 20:1 bundling pattern of the Lofer cycles (Haas 1982; Haas and Schwarzacher 1986; Goldhammer et al. 1990; Balog et al. 1997).
These similarities among sections in the Northern Calcareous Alps, the Transdanubian Range, the Southern Alps, Sicily, and the Hellenids imply that regular cyclicity was not an isolated phenomenon but rather a prevalent depositional feature in the entire Dachstein platform system (Fig. 9). Consequently, Milankovitch-scale orbitally forced eustatic sea level variations remain the only plausible cause for the near-uniform cyclicity that affected such a broad paleogeographic area for an extended period of time. Although the local effect of autocyclic processes on the sedimentation cannot be ruled out entirely, their role was much more limited.
Most of the studies that question the orbital forcing present arguments that the Lofer cycles have lateral thickness variations, occasionally they pinch out, and many cycles are not complete (Satterley and Brander 1995; Enos and Samankassou 1998, 2021; Samankassou and Enos 2019). In addition, questions were raised about the consistency of the stacking pattern and the viability of a driving force in a greenhouse world (Goldhammer et al. 1990; Satterley and Brander 1995; Satterley 1996).
To fend off these criticisms, it was pointed out that lateral variability was observed only in the Steinerness Meer section but elsewhere, in other sections studied in the Northern Calcareous Alps, lateral variations are rather rare (Schwarzacher 2005). Concerns of the consistency of stacking pattern were based solely on visual observation or Fischer-plots, rather than on detailed cyclostratigraphical analysis (Schwarzacher 2005). Although the core Po-89 used in this study obviously does not afford an opportunity to examine lateral variations, it is eminently suitable to detect orbital forcing due to the exceptional preservation of the cycles. As demonstrated in Table 4, our results are coherent with the other, originally distant parts of the Dachstein platform system (Fig. 9). It is not conceivable that random processes would produce highly similar, hierarchical sedimentary cyclicity across a wide geographic area and through an interval of millions of years, such as the spatial and temporal extent of the Dachstein platform system. The credibility of our results obtained using up to date cyclostratigraphic tools is also supported by the demonstration of the reliability and sensitivity of this methodology (Sinnesael et al. 2019).
An independent line of supporting arguments is provided by studies of the Upper Triassic to lowermost Jurassic carbonates deposited in the peri- and intraplatform basins of the Dachstein platform system. These deposits consist of calciturbidites that were sourced primarily from the adjacent platforms (Vallner et al. 2023). As Milankovitch cyclicity was identified in several of these deposits (Mesetti et al. 1989; Reijmer and Everaars 1991; Reijmer et al. 1993; Maurer et al. 2004; Vallner et al. 2023), it is highly likely that the deposition in the platforms and the adjacent basins were both paced by the same orbital cycles.
Our discussion of the results of this study is intended to help to settle the debate that has been going on for more than 70 years. The definitive evidence for orbital forcing of the sedimentation in the Late Triassic platforms validates the use of cyclostratigraphy and astrochronology as high-resolution stratigraphic tools for the Dachstein Limestone Formation and its analogues. In addition, comparison of the Upper Triassic cyclothems with the younger platform deposits that record orbitally forced cycles, as in the Arabian and the Adriatic Platforms, will further our understanding of the general patterns of cyclic sedimentation in the Mesozoic carbonate platform environments.