Geochemistry, Geophysics, Geosystems

High-resolution estimate for the depositional duration of the Triassic Latemar Platform: A new magnetostratigraphy and magnetic susceptibility cyclostratigraphy from basinal sediments at Rio Sacuz, Italy

Authors


  • This article was corrected on 12 August 2014. See the end of the text for full details.

Abstract

[1] A new magnetostratigraphy and magnetic susceptibility cyclostratigraphy from Middle Triassic basinal sediments at Rio Sacuz, northern Italy, suggests a depositional duration of ~1 myr for most of the 670 m carbonate Latemar Platform, giving a high-resolution estimate for the duration for the Latemar's deposition. The new magnetostratigraphy from Rio Sacuz clarifies the ambiguities in a previous magnetostratigraphic study of the Latemar caused by lightning strike remagnetizations. Our Rio Sacuz study was unaffected by lightning. Using thermal demagnetization, we show a Normal-Reversed-Normal-Reversed sequence at the Latemar-correlated locality of Rio Sacuz. With a polarity interval duration of ~0.25–0.5 myr for the Middle Triassic, this gives a depositional duration of ~1–2 myr. Rock magnetic experiments show that the magnetic carrier is magnetite, suggesting a primary, depositional-age remanence. Measurement of magnetic susceptibility at 1 m intervals from Rio Sacuz reveals eight short eccentricity cycles (~95–125 kyr) bundled into slightly more than two 405 kyr long eccentricity cycles indicating a more precise duration for the Rio Sacuz section of 0.8 to 1 myr. The shorter depositional duration required by this study forces a sub-Milankovitch time scale of 1–2 kyr for the Latemar platform's meter-scale rhythmic bedding and provides strong evidence for nonorbitally driven climate change in the Middle Triassic.

1 Introduction

[2] The Middle Triassic Latemar platform in the Dolomites of northern Italy has been the subject of a cyclostratigraphic controversy centered on the order of magnitude difference in estimates of the duration of deposition for the 670 m thick carbonate platform [Brack et al., 1996; Preto et al., 2001; Zühlke et al., 2003; Zühlke, 2004; Kent et al., 2004]. The cyclostratigraphic interpretation suggests that the meter-scale shallowing upward cycles, bundled in 5:1 packages, are the result of precessional-scale orbital forcing, giving a depositional duration of ~12 myr [Hinnov and Goldhammer, 1991; Preto et al., 2001]. U-Pb single zircon dating of tuff layers in the sequence suggest a depositional duration of just ~2 myr [Mundil et al., 2003]. In addition, the entire sequence contains not much more than one ammonite zone, also suggesting a depositional duration of ~1–2 myr [Zühlke, 2004, 2003]. A previous magnetostratigraphy suggested a depositional duration of <1 myr; however, while the data interpretation was reasonable, the data were plagued by lightning strikes [Kent et al., 2004]. Resolving this discrepancy has significant implications for our understanding of important paradigms in the Earth sciences. If the ~12 myr duration suggested by cyclostratigraphy is correct, then our understanding of the accuracy of U-Pb single zircon dating is not complete. Alternatively, if the ~1–2 myr duration suggested by the U-Pb single zircon dating is correct, then we must rethink our understanding of climate change in the Middle Triassic [Preto et al., 2004]. Our new magnetostratigraphic and rock magnetic cyclostratigraphic study resolves the discrepancy in favor of the geochronologic/biostratigraphic interpretation and further refines the duration of the Latemar platform's deposition. The new results also support evidence of strong lithologic cyclicity in the Triassic at frequencies higher than orbitally forced frequencies.

2 Geologic Background

[3] Since Kent et al.’s [2004] lightning-plagued magnetostratigraphy forced the shortest duration estimate, we chose to test it by selecting a study area that could be easily correlated stratigraphically to the Latemar but would not be plagued by lightning strikes. Previous studies have integrated tephrostratigraphy and ammonoid biostratigraphy to correlate between the basinal section at Rio Sacuz and the carbonate platform deposits at the Latemar [Brack and Muttoni, 2000; Preto et al. 2005] (Figure 1). The Late Anisian-Ladinian succession at Rio Sacuz correlates to approximately 540 m of the 670 m Latemar platform sequence. Additionally, Rio Sacuz's location in a river valley, below the tree line, greatly reduced the probability of lightning strikes affecting our study.

Figure 1.

Location map showing the section studied at Rio Sacuz and its location relative to other sections (Latemar, Seceda, and Frotschbach) correlated to Rio Sacuz using tephrostratigraphy and biostratigraphy. Gray areas in the bottom map show Late Anisian-Ladinian carbonate platform locations. The lighter the gray, the greater the degree to which the platform was drowned. Arrows indicate progradation directions. Figure modified from Preto et al. [2005].

[4] The Late Anisian Dolomites had an extremely complex paleotopography, dominated by an episode of extensional tectonics that created structural highs and fault-bounded basins with a limited areal extent of a few square kilometers [Masetti and Neri, 1980]. During this phase of extensional tectonics, there was a period of strong regional subsidence, in which isolated carbonate platforms nucleated on structural highs and developed aggradational or backstepping geometries [Bosellini, 1984; De Zanche et al., 1995]. As part of the regional subsidence, the basins widened and deepened rapidly [Preto et al., 2011]. Therefore, the Upper Anisian-Lower Ladinian of the Dolomites had two types of depositional environments, carbonate platform, and basinal. The carbonate platform deposits are thick, with up to 700 m successions of shallow-water carbonates. Basinal areas were sediment starved and have much thinner deposits (few tens of meters) of shale, limestone, and cherty limestone successions. The coeval basinal and platform successions lie on a Late Anisian carbonate bank (Contrin Formation) in the Western Dolomites, similar to that observed at the Latemar and its surrounding areas, but are stratigraphically on top of basinal successions in the Eastern Dolomites. At Rio Sacuz, where our magnetostratigraphic study was conducted, a basinal Late Anisian-Ladinian succession lies on Late Anisian deep-water marls with thin-shelled bivalves and ammonoids, which alternate with cm to dm-scale calci-turbidites. This unit is called the Ambata Formation and laps on the Late Anisian carbonate bank (Contrin Formation) to the west (Figure 2).

Figure 2.

Late Anisian-Early Ladinian stratigraphy of the Dolomites [modified from Manfrin et al., 2005]. The stratigraphic relationship between the three magnetostratigraphic records (Latemar, Seceda and Rio Sacuz) is shown. The gray lines indicate the stratigraphic coverage of the magnetostratigraphic records, the actual stratigraphic section may be longer (e.g., the Seceda core also drilled part of the Contrin Formation). Sequence stratigraphy (white boxes on left and right) from De Zanche et al. [1993]. Lam: lower Ambata Formation; Uam: Upper Ambata Formation; Mo: Moena Formation; Plk: Plattenkalke; Knk: Knollenkalke.

[5] Rio Sacuz is located on the flank of the Cernera platform; the samples of this study encompass the upper Ambata Formation and part of the Late Anisian-Ladinian basinal succession that includes the Lower Plattenkalke and part of the Knollenkalke [Blendinger et al., 2004; Preto et al., 2005]. The Lower Plattenkalke is an approximately 15 m thick succession of black, well-bedded silicified argillite, limestone, and dolostone with cherty beds. The Knollenkalke is composed of cm to dm-scale nodular beds of gray-greenish limestones with chert nodules. In contrast to what is observed in the Seceda drill core [e.g., Muttoni et al., 2004], the boundary between the Plattenkalke and Knollenkalke is not sharp, but rather represented by a transitional interval of siliceous shales and limestones approximately 8 m thick. Crystal tuffs up to 50 cm thick are found interbedded in both the Plattenkalke and Knollenkalke.

[6] While Rio Sacuz represents a Late Anisian basinal setting, the coeval Latemar is an aggrading carbonate platform that formed on an isolated structural high [Preto et al., 2011]. The platform's interior is a succession of peritidal carbonate cycles capped by subaerial exposure horizons [e.g., Goldhammer et al., 1987; Preto et al., 2001; Christ et al., 2012] that include primary ashfalls or crystal tuffs [Zühlke et al., 2003; Preto et al., 2005]. Subaerial exposure intervals often include tepee structures that acted as sedimentary traps for ammonoid shells, thus, providing a detailed ammonoid biostratigraphy for the platform [Manfrin et al., 2005].

3 Sampling Strategy and Magnetic Measurements

[7] The sampling strategy for the magnetostratigraphic study was devised by using a stratigraphic sampling interval dense enough to define polarity intervals in both the longer cyclostratigraphy-based and shorter geochronologic-based depositional duration estimates for the Latemar platform. Correlating and extrapolating Preto et al.’s [2001] Latemar cyclostratigraphic depositional duration to Rio Sacuz suggests that the Rio Sacuz sequence could have been deposited in ~10.9 myr. Therefore, at least 20 polarity intervals would be observed in the entire Rio Sacuz section, assuming a polarity interval duration of ~0.25–0.5 myr in the Middle Triassic [Hounslow and Muttoni, 2010], each having a stratigraphic thickness of ~3 m. Assuming that at least three consecutive horizons are needed to unambiguously resolve a polarity interval, the Rio Sacuz section was sampled every ~1 m over ~70 m of section, collecting 65 horizons of at least three oriented samples per horizon. The shorter geochronologic-based depositional duration would predict only 2–4 polarity intervals at Rio Sacuz. Meter-scale sampling would be more than sufficient to detect this predicted reversal stratigraphy. Oriented, 25 mm diameter cores were collected with a gasoline-powered, diamond-bit coring drill.

[8] The magnetic susceptibility of the unoriented chips trimmed from the paleomagnetic cores was measured and normalized by mass. The magnetic susceptibility for each horizon, collected at 1 m intervals, was the average of independent susceptibility and mass measurements of about three to four chips from the tops and bottoms of the multiple cores collected from each horizon.

[9] Paleomagnetic measurements were made on a 2G Enterprises 755R superconducting rock magnetometer in a magnetically shielded room with a nominal field of 350 nT. Thermal demagnetization was conducted with a TD-48 ASC Thermal Demagnetizer in up to 15 temperature steps between 100°C and 500°C. Rock magnetic measurements were made to determine the magnetic mineralogy that carried the characteristic remanence isolated by thermal demagnetization. Isothermal remanent magnetizations (IRMs) were applied using an ASC Impulse Magnetizer. Magnetic susceptibility measurements were made with a KLY-3s Kappabridge susceptibility meter that applies a 300 A/m amplitude field that alternates at 875 Hz.

[10] Spectral analysis of the magnetic susceptibility data series was conducted with the SSA-MTM Toolkit, Version 4.4 [Ghil et al., 2002]. Spectra were measured using the multitaper method (MTM), and all magnetic susceptibility data were treated using three 2Π prolate tapers. Using the SSA-MTM Toolkit, robust red noise models [Mann and Lees, 1996] were fit to all of the spectra and displayed with confidence limits (90%, 95% and 99%).

4 Paleomagnetic and Rock Magnetic Results

[11] Based on their thermal demagnetization behavior, the samples can be broken down into three different categories (Figure 3). The first category has samples with a low unblocking temperature component that is removed by ~200˚C and a higher unblocking temperature component that is southerly directed and has intermediate upward inclinations (Figure 3b). The second category has samples with univectoral directions trending into the origin of the vector endpoint diagrams that are northerly directed with intermediate downward inclinations (Figure 3a). The third category includes samples with univectoral directions that are southerly directed with intermediate upward inclinations (Figure 3e). Most samples were fully demagnetized by 400°C; however, about 10% of the samples were not fully demagnetized until 500°C.

Figure 3.

Representative vector endpoint plots [Zijderveld, 1967] for thermal demagnetization results from the Rio Sacuz rocks. Black squares are the horizontal component, and white squares are the vertical component. Demagnetization temperature steps are labeled on the horizontal component. Sample position within the Rio Sacuz stratigraphy is labeled below sample names. (a) Example of a normal polarity sample. (b) Example of a reversed polarity sample with a low-temperature ((<200°C) overprint component. (c) Example of a normal polarity sample from a horizon with multiple polarities (40 m). This sample is completely normally overprinted. (d) Example of a reversed polarity sample from a horizon with multiple polarities (58 m). This plot also shows some of our poorest data. (e) Example of a reversed polarity sample with a univectoral decay into the origin. All characteristic remanences were determined by principal component analysis [Kirschvink, 1980] with a MAD cutoff of 15˚. Plots created with PaleoMag X [Jones, 2002].

[12] The virtual geomagnetic pole (VGP) latitudes used for determining the magnetostratigrapy are the great circle distance from the mean VGP (Table 1) of the entire dataset [Kent et al., 1995] (Figure 4). VGP latitudes have either high positive values between +45˚ and +90˚ (normal polarity), intermediate values between −45˚ and +45 (indeterminate polarity), or high negative values between −90˚ and −45˚ (reversed polarity). The mean Rio Sacuz normal polarity direction is D=19.1˚, I=49.4˚ (N=146, k=17.1, α95=2.9˚), and the mean reversed polarity direction is D=208.5˚, I=−22.5˚ (N=36, k=2.5, α95=19.6˚), and although declinations are separated by close to 180˚, the data fail a reversals test (McFadden and McElhinny, 1990) probably due to extensive normal polarity overprinting. The mean paleolatitude of the Rio Sacuz locality was 27.3˚ + 3.6˚/−3.3˚ which is statistically the same, at the 95% confidence level, as the mean paleolatitude of the Seceda, Belvedere, Pedraces, and Frostbach localities of coeval Triassic age strata in the Dolomites based on paleomagnetic studies reported by Muttoni et al. [2004] (mean paleolatitude=19.0˚ +10.7˚/−8.6˚).

Table 1. Mean Paleomagnetic Direction and VGP
 DeclinationInclinationNKα95
Geographic354.6˚59.1˚1838.93.7˚
Stratigraphic19.7˚49.4˚1838.03.9˚
Mean VGPLat.=66.4˚NLong.=142.3˚EA95=4.1˚  
Figure 4.

(b) Magnetostratigraphic results for the Rio Sacuz section shown in the center column. VGP latitudes for individual samples at one horizon are connected by horizontal lines. Rio Sacuz reversal stratigraphy is based on the rules discussed in the text. Black is normal polarity, white is reversed, gray is indeterminate. Half-bars indicate a submagnetozone based on only one horizon. The correlation of the Rio Sacuz with the Latemar platform and the (c) Triassic geomagnetic polarity timescale of Hounslow and Muttoni [2010] is shown along with the (a) biochronostratigraphic timescale of Brack and Rieber [1993]. Latemar sedimentary facies are based on Egenhoff et al. [1999]. “Trans” marks the gradational boundary between the Plattenkalke and the Knollenkalke (see text). The Latemar magnetostratigraphy in Figure 4a is based on Kent et al. [2004]. The boundaries of biochronozones are determined from ammonoid distributions in Zühlke et al. [2003], Manfrin et al. [2005], and Preto et al. [2005]. Uncertainty about the base of the curionii zone is due to the uncertain identification of the chiesense groove (a lithological marker) at Rio Sacuz. Tuff correlations (Te, Td, Tc, and T5) are based on Preto et al. [2005]. Tuff paleomagnetic results are shown with green triangles. The magnetostratigraphy of Kent et al. [2004] is reinterpreted here with a proposed additional reversed polarity interval nearly coincident with the Middle Tepee Facies (MTF). Correlations between the Rio Sacuz reversal stratigraphy and the Hounslow and Muttoni [2010] timescale are shown by dotted lines and are strongly constrained by the biochronozones, and geochronologic ages. The correlation of the base of RS1n (dashed line with question mark) within the MT6 reversed interval is based on biostratigraphy. Note that our results at Rio Sacuz have not observed the base of RS1n. Red squares indicate the location of the samples used for rock magnetic measurements.

[13] The mean direction of the low unblocking temperature 0–200˚C component was calculated from the thermal demagnetization vector endpoint diagrams so it could be compared to the modern field direction (D=2˚, I=62˚) and the geocentric axial dipole direction (D=0˚, I=64.5˚, site latitude=46.4˚N) for the Brunhes normal polarity chron. When the low-temperature component is plotted in geographic coordinates (D = 8.1˚, I=58.7˚, α95=6.1˚), its 95% confidence interval includes the present-day field direction, but when it is plotted in stratigraphic coordinates (D=27.4˚, I=47.7˚, α95=6.5˚), it does not include the present-day field direction within its circle of 95% confidence. The geocentric axial dipole field direction is not contained within the low-temperature component's 95% confidence intervals in either geographic or stratigraphic coordinates. These results suggest that the 0–200˚C component is most likely due to present-day viscous magnetization overprinting. It is also demonstrates the predominance of normal polarity overprinting in the Rio Sacuz section. This observation will inform the development of the magnetostratigraphy.

[14] The possibility of high-intensity natural remanent magnetizations (NRMs) due to lightning strikes similar to that observed by Kent et al. [2004] for the Latemar was checked by plotting NRM intensity versus VGP latitude (Figure 5) similar to Kent et al. [2004]. There was no evidence of the high-intensity NRMs observed by Kent et al. [2004]; thus, our study has avoided lightning-induced overprinting of the rocks’ magnetization.

Figure 5.

NRM intensity, normalized by mass, plotted as a function of VGP latitude determined from paleomagnetic analysis. The dashed line is the arbitrary cutoff used by Kent et al. [2004] to separate samples from the Latemar that were not affected by lightning (intensities less than 10−7 Am2/kg) from those that were affected by lightning (intensities greater than 10−7 Am2/kg). Most of the Rio Sacuz samples, both reversed and normal polarity, have NRM intensities below this arbitrary cut off suggesting that lightning has not affected the Rio Sacuz samples.

[15] All specimens showed similar rock magnetic behavior. In IRM acquisition experiments, samples became saturated by ~300 mT (Figure 6a). The acquisition curves were modeled [Kruiver et al., 2001] with one coercivity component for most samples, but were modeled with two components for a small subset of the samples measured. The dominant component contributes to ~95% of the magnetization with a mean coercivity of 50 mT (maximum coercivity =160 mT) (Figure 6b). The second component contributes to ~5% of the magnetization with a mean coercivity of 500 mT (maximum coercivity=1.0 T). The fields for the application of orthogonal IRMs in a coercivity-unblocking temperature test [Lowrie, 1990] were based on the maximum coercivity values derived from the IRM modeling. The results of the Lowrie test show that the magnetic moment of the 160 mT component generally decreases rapidly to 450˚–500˚C, and then decays more gradually to 600˚C (Figure 6c). SIRM/χ ratios range between 0.71 and 3.18 kA/m, indicating that the magnetization is carried predominately by magnetite, which has SIRM/χ ratios of 0.2–20 kA/m [Peters and Dekkers, 2003]. Bcr values are estimated from the IRM acquisition modeling and fall within the range expected for magnetite. A plot of SIRM/χ versus Bcr indicates that the magnetic mineralogy of our samples is dominated by magnetite (Figure 7). The reapplication of orthogonal IRMs after heating yielded values two to four orders of magnitude higher than before heating, ruling out the presence of maghemite typically a secondary magnetic mineral formed by low-temperature oxidation, which tends to lose its magnetization near to 400°C, since it would have inverted to low magnetization hematite during heating [Tauxe, 2010]. However, this could indicate either an iron sulfide or an Fe-rich silicate oxidizing to magnetite during heating [van Velzen and Zijderveld, 1992].

Figure 6.

Rock magnetic test results for Rio Sacuz. Specimens were selected for rock magnetic tests in order to include all lithologies and all polarity groups. Their stratigraphic position is shown in Figure 4 and their stratigraphic position is listed. (a) IRM acquisition results for Rio Sacuz specimens. Note that all specimens, regardless of lithology or polarity, show similar behavior in this test, becoming saturated with ~300 mT of applied field. Coercivities of 300 mT or less are suggestive of magnetite as the primary magnetic carrier. (b) Gradient of IRM acquisition curve for a representative sample showing that one coercivity component with a mean magnetization of 50 mT can be used to model the IRM acquisition data. (c) Coercivity-unblocking temperature spectrum for an IRM applied in a 160 mT field. The magnetic moment of most specimens decreases rapidly through ~450˚–500˚C, then more slowly up to ~600˚C.

Figure 7.

SIRM/χ plotted as a function of Bcr for the seven samples used in the IRM acquisition experiments and coercivity-unblocking temperature measurements (Figure 6). The Bcr is determined from modeling the IRM acquisition curves (Figure 6). The yellow, blue, and red fields are from Peters and Dekkers [2003] showing the characteristic values for magnetite, greigite, and hematite. The Rio Sacuz samples SIRM/χ versus Bcr values show that magnetite is the dominant magnetic mineral in the Rio Sacuz samples.

[16] The results of the rock magnetic tests indicate that the characteristic remanence is predominately carried by magnetite. The IRM acquisition experiments show that the samples become saturated at fields close to 200–300 mT, suggesting magnetite (Figure 6a). The SIRM/χ versus Bcr plots fall into the magnetite field of Peters and Dekkers (2003) (Figure 7). The coercivity-unblocking temperature tests show that the 160 mT coercivity component has an unblocking temperature component that typically has a maximum temperature of ~ 450˚–500°C but can persist to temperatures no greater than 600˚C (Figure 6c). The thermal demagnetization results combined with the SIRM/χ versus Bcr observations indicate that the characteristic magnetization is most likely carried by magnetite, most likely a primary depositional magnetic mineral.

5 Magnetostratigraphy

[17] The following rules were used to develop the magnetostratigraphy from the individual sample VGP latitudes at a horizon.

  1. Sample VGP latitudes that are within 45° of the mean normal polarity paleopole (65.9˚N, 147.9˚E, α95=4.0˚, N=195), or its antipode, are normal or reversed polarity, respectively. VGP latitudes between 45˚ N and 45˚ S are considered to be intermediate, and their polarity is indeterminate. The choice of 45˚ from a given polarity paleopole follows Kent et al. (2004) in the development of their polarity stratigraphy for the Latemar.
  2. If only one sample in a horizon is reversed, then the horizon is interpreted to be reversed polarity. The strong evidence of present-day field overprinting affecting these rocks suggests that the overprinting is biased toward normal polarity, so reversed polarity samples are given more weight in the development of the magnetostratigraphy (Kodama, 1979).
  3. Based on the evidence for strong present-day field overprinting, if only one sample at a horizon is intermediate, then the horizon's polarity is intermediate.
  4. Two or more consecutive horizons of the same polarity indicate a magnetozone. One horizon surrounded by horizons of opposite polarity is a submagnetozone.

[18] Our interpretation of the VGP latitudes from the Rio Sacuz section yields a Normal-Reversed-Normal-Reversed (N-R-N-R) reversal stratigraphy up section, with a few submagnetozones. The magnetozones are labeled RS1n, RS1r, RS2n, and RS2r, upsection (Figure 4). At stratigraphic levels 56–58 m and 40 m, horizons with both reversed and normal polarity samples probably indicate strong normal polarity present-day field overprinting on reversed polarity horizons (Figure 4). The polarity zone boundaries do not correspond to lithologic transitions, and there is no correlation between rock magnetic properties and polarity, all indications of a robust magnetostratigraphy.

[19] Because the main goal of this study is to determine a depositional duration for the Rio Sacuz section, and hence the Latemar platform, it is important to compare our results to the previous magnetostratigraphic results of Kent et al. [2004]. The high-intensity NRM samples that Kent et al. [2004] rejected, based on strong evidence that they were affected by lightning strikes, indicate reversed polarity VGP latitudes mostly within the Middle Tepee Facies (MTF in Latemar stratigraphy in Figure 4), as well as in the upper part of the section (UTF in Latemar stratigraphy in Figure 4). These stratigraphic intervals correlate to reversed polarity zones in the Rio Sacuz rocks, with the MTF of the Latemar roughly correlating to the Plattenkalke at Rio Sacuz and the UTF at the Latemar correlating to the top of the Knollenkalke at Rio Sacuz. This correlation suggests that some of the specimens that Kent et al. [2004] rejected may not have been totally overprinted by lightning strikes. If the Latemar magnetostratigraphy is reinterpreted with reversed polarity intervals in the MTF and the UTF, the magnetostratigraphies of Rio Sacuz and the Latemar correlate exactly, both showing a N-R-N-R pattern that is in agreement with biostratigraphic and ash-bed correlations (Figure 4). This correlation would be in agreement with that suggested by Preto et al. [2005] on the basis of ammonoid biostratigraphy and tephrostratigraphy.

[20] Not only does the Rio Sacuz magnetostratigraphy presented here correlate to our reinterpretation of Kent et al.’s [2004] Latemar magnetostratigraphy, it can also be correlated to a modified version of Hounslow and Muttoni’s [2010] integrated geomagnetic polarity timescale for the Triassic (Figure 4). The modification results from the recognition of a new normal polarity interval designated RS1n in the Rio Sacuz section. RS1n encompasses the entire Ambata Formation at Rio Sacuz, which has an ammonoid association including Aplococeras avisianum, Hungarites spp., and Latemarites sp. [Manfrin et al., 2005; Preto et al., 2005]. This association is typical of the late Anisian avisianum subzone (i.e., upper reitzi zone of Brack and Rieber, [1993]) and was also observed in the Lower Tepee Facies of the Latemar platform [Brack and Rieber, 1993; Zühlke et al., 2003; Manfrin et al., 2005]. A. avisianum and Latemarites are only found in the reitzi zone; thus, RS1n cannot be correlated with the normal polarity interval at the base of MT6 of Hounslow and Muttoni [2010], which is within the older trinodosus zone.

[21] The magnetostratigraphy of the Illyrian (upper Anisian pp.) stratigraphic interval, and especially the portion that corresponds to the trinodosus and reitzi ammonoid zones of Brack and Rieber (1993), is poorly documented in marine Tethysian successions. In particular, the long magnetozone MT6 of Hounslow and Muttoni [2010] is mostly defined by the continental records of Spain and central-northern Europe [e.g., Dinarès-Turrell et al., 2005; Nawrocki and Szulc, 2000; Hounslow and McIntosh, 2003] and by the poorly dated carbonate platform succession of the Badong Formation in South China [Huang and Opdyke, 2000]. The correlation of these sections to the Tethysian geomagnetic polarity record is mostly based on magnetostratigraphy, with very loose control from palynomorphs and one conodont species a great distance away in China, whose first and last occurrences are most likely facies-dependent because of the extremely shallow water depositional environment (often peritidal) of the Badong Formation. Magnetozone MT6 is only observed in three deep marine sections; Felsoors in Hungary [Márton et al., 1997], Seceda [Muttoni et al., 2004], and Rio Sacuz (this work) in the Dolomites. Only the very top of MT6r was recovered at Seceda. While the full magnetozone should be present at Felsoors, based on ammonoid biostratigraphy, the magnetic record of Felsoors is poor quality with many samples of ambiguous polarity. Furthermore, Felsoors is a succession of condensed limestones with thick ash beds intercalated indicating episodic deposition [Brack et al., 2005]. Because of the low correlation potential of continental or shallow water sedimentary sequences and the lack of evidence from deep-water successions, Rio Sacuz fills a gap in the Triassic geomagnetic polarity time scale below the base of the Seceda drill core.

[22] Similarly, RS1n cannot correspond to the normal interval at the base of MT7 of Hounslow and Muttoni [2010], because the lower MT7 correlates to the base of the Knollenkalke, or, in some localities, to the uppermost part of the Plattenkalke, and not the Ambata Formation.

[23] The stratigraphically sequential Plattenkalke and Knollenkalke are coeval throughout the Dolomites, as determined by tephrostratigraphy. A prominent ash bed (Tc) is present at the base of the Knollenkalke in the Lombardian Alps, throughout the Dolomites and in Hungary (see Brack et al. [2005] for a review). Tc was also identified at Rio Sacuz [Preto et al., 2005] and falls within the RS2n polarity interval (Figure 4). This ties RS2n to the MT7n polarity interval of Hounslow and Muttoni [2010]. Another prominent ash bed (T5) is present in the upper part of the Plattenkalke at Rio Sacuz, within the very bottom of the RS2n polarity interval. We correlated T5 with the major ash bed in the upper part of the Plattenkalke at Seceda. Ash bed T5 nearly coincides with the top of the avisianum ammonoid zone of Manfrin et al. [2005], i.e., to the top of the reitzi zone of Brack and Rieber [1993] [Manfrin et al., 2005; Preto et al., 2005]. At Seceda, the Plattenkalke below T5 within polarity interval MT6r of Hounslow and Muttoni [2010] yielded Lecanites misanii, Aplococeras avisianum, and Parakellnerites aff. rothpletzi [Preto et al., 2005; Brack et al., 2005], which is a typical ammonoid association of the upper reitzi zone. Thus, the polarity interval RS1r of Rio Sacuz corresponds to MT6r. This polarity interval was defined in the Seceda drill core and in the Frötschbach section [Muttoni et al., 2004] and mostly coincides with the Plattenkalke in both sections. RS1r at Rio Sacuz is also fully within the Plattenkalke, which is thus confirmed to be roughly coeval throughout the Dolomites.

[24] This modification to the Hounslow and Muttoni [2010] timescale does not change the average polarity interval duration of ~0.25–0.5 myr for this part of the Triassic.

[25] While RS1n cannot correlate to MT7 [Hounslow and Muttoni, 2010], the ammonoid associations, as well as the Plattenkalke-Knolenkalke transition, suggest that RS2n does correlate to MT7. The biostratigraphy shows that RS2n includes the upper reitzi, all of the secedensis, and the lower curionii ammonite zones. These ammonite zones correlate to Hounslow and Muttoni [2010]’s MT7, which is predominantly normal polarity but includes three short reversed subchrons. Our evidence for possible short reversed intervals that are strongly normally overprinted at 40 m and 56–58 m at Rio Sacuz therefore could correlate to the upper and lower short reversed intervals in MT7. The middle reversed subchron in MT7 corresponds to tuff layer Tc. We did not record a reversal at Tc; however, this reversal was only observed at Frotschbach and not at the correlated section at Seceda [Hounslow and Muttoni, 2010]. The N-R-N-R polarity sequence at Rio Sacuz and its correlation to the modified Triassic polarity time scale indicates a 1–2 myr depositional duration for Rio Sacuz and hence through correlation, the Latemar platform carbonates. A 1–2 myr duration is also supported by U/Pb single zircon dates of the tuff layers in the platform [Zühlke et al., 2003; Mundil et al., 2003] and correlated basinal successions [Mundil et al., 1996].

6 Magnetic Susceptibility Cyclostratigraphy

[26] A magnetic susceptibility cyclostratigraphy allows for further refinement of the depositional duration of the Rio Sacuz sequence. The magnetic susceptibility measured for the Rio Sacuz basinal sediments is positive, indicating that the positive susceptibility of ferromagnetics and paramagnetics in the rocks is stronger than the negative susceptibility of the diamagnetic calcite. The variations in susceptibility could be caused by fluctuations in the influx of detrital ferromagnetics and paramagnetics in the presence of relatively constant carbonate production as envisaged for rock magnetically recorded orbital frequencies in the Eocene Arguis Formation of Spain [Kodama et al., 2010]. The Rio Sacuz magnetic susceptibility shows cyclicity with amplitudes of approximately 3–4 m3/kg and periodicities of approximately 8–10 m (Figure 8a). The 8–10 m cycles appear to be bundled in 25–30 m cycles that each include three to four 8–10 m cycles. The MTM spectral analysis (Figure 8b) supports this interpretation showing spectral peaks at 1/7.1 cycles/m to 1/9.7 cycles/m frequencies and a longer wavelength peak at 1/23.3 cycles/m. However, the 1/23.3 m cycles/m peak does not rise above the 90% confidence limits for the robust red noise, likely due to the low number of cycles (only two) that are observed in our data series (Figure 8a). Given the magnetostratigraphic results that suggest an approximately 1–2 myr duration for the Rio Sacuz sedimentary section, the 8–10 m periods are interpreted to be short eccentricity and the longer, ~25 m bundling, is most likely 405 kyr long eccentricity. Counting the short eccentricity cycles (8) gives an estimate for the duration of the Rio Sacuz section of between 760 kyr (assuming short eccentricity ~95 kyr) and 1000 kyr (assuming short eccentricity ~ 125 kyr).

Figure 8.

(a) Magnetic susceptibility data series (top) showing eight short eccentricity cycles (labeled by “e”) that are 8–10 m in stratigraphic thickness. The eight short eccentricity cycles appear to be bundled in sequences of three or four into longer, 25–30 m, sequences that are probably the 405 kyr long eccentricity cycles. If each short eccentricity cycle is nominally 125 kyr in duration, the entire Rio Sacuz sequence is ca. 1000 kyr in duration. (b) (bottom), MTM spectral analysis of the Rio Sacuz magnetic susceptibility sequence. Peaks at about 1/7 to 1/10 cycles/m and 1/23.3 cycles/m are indications of short and long eccentricity cycles.

7 The Latemar Facies Cyclicities

[27] Geochronologic, biostratigraphic, and our new magnetostratigraphic and cyclostratigraphic data provide strong evidence that the meter-scale shallowing upward cycles at the Latemar were not forced by precession, the standard cyclostratigraphic interpretation of the Latemar platform's rhythmic bedding. There are two alternatives to the astronomically forced explanation for the meter-scale cycles and their apparent 5:1 bundling. One explanation is that these cycles are evidence for Quaternary-like sub-Milankovitch climate variability, such as Dansgaard-Oeschger cycles [Broecker and Denton, 1989], Heinrich events [Heinrich, 1988], and Bond cycles [Bond et al., 1992], in the Middle Triassic. This observation is particularly interesting since the Quaternary sub-Milankovitch climate variability is glacially related and there is no strong evidence for glaciations in the Triassic. Zühlke's [2004] interpretation that the Latemar cycles are the result of a combination of sub-Milankovitch and Milankovitch forcing is an example of this explanation. The Zühlke interpretation has recently gained support from Forkner et al.'s [2010] forward modeling of eustatic sea level oscillations at these different timescales.

[28] A competing explanation is that the cycles are not allocycles caused by the external forcing of climate variability, but, in fact, are autocycles that are controlled by carbonate production [Drummond and Wilkinson, 1993; Beerbower, 1964; Schwarzacher, 2000]. In this interpretation, the cycles are the result of varying accommodation space due to the relative rates of platform buildup (carbonate production) and tectonic subsidence. A key sedimentologic feature that would be suggestive of autocycles is the systematic presence of an intertidal and/or supratidal facies to form a full shallowing upward suite. The presence of an intertidal or supratidal facies in the Latemar sequence has been debated [Preto et al., 2004; Egenhoff et al., 1999; Forkner et al., 2009]. A supratidal facies is clearly present in a Latemar-correlated sequence with similar lithologic cyclicities at Mendola Pass [Forkner et al., 2009]. However, the similarity of cyclicities at the Mendola Pass locality to those at the considerably distant Latemar platform makes an autocyclic mechanism, which drives more localized system-specific sedimentary cycles, less likely. The support of the Zühlke interpretation by Forkner et al.'s [2010] modeling of eustatic sea level oscillations and the lack of an intertidal/supratidal facies all point to a combined sub-Milankovitch/Milankovitch forcing of the Latemar cycles.

[29] Our new magnetostratigraphy and magnetic susceptibility cyclostratigraphy allow for precise calculation of the duration of the Latemar cycles. Quantitative correlation between the Rio Sacuz section and the Latemar can be made between tuffs T5 and Te [Preto et al., 2005], which are stratigraphically separated by 18.3 m at the Rio Sacuz and 155.5 m at the Latemar. Based on our depositional duration for the Rio Sacuz section, the T5-Te interval is between 214 and 281 kyr in duration. With an average stratigraphic cycle thickness for the Latemar meter-scale cycles of 52 cm [Forkner et al., 2009] to 87 cm [Preto et al., 2001], the Latemar shallowing upward cycles are only ~0.8 to ~1.7 kyrs in duration. Kent et al. [2004] reached a similar conclusion for the duration of the meter-scale Latemar cycles (1.7 kyr duration) based on their lightning-plagued magnetostratigraphy. Forkner et al.’s [2010] more recent re-analysis using Milankovitch insolation forcing of high-frequency eustasy also indicates a 1.7 kyr duration for the meter-scale Latemar cycles. In comparison, Zuhlke et al.’s [2003] cyclostratigraphic re-analysis of the Latemar platform estimates the meter-scale cycles to be 4.2 kyr in duration. No results, from other localities, suggest such short suborbitally driven cycles in the Triassic. Wu et al. [2012] do observe suborbitally driven cycles in the Triassic Lower Daye Formation of South China that are recorded both lithologically and rock magnetically, but are estimated to be 4–5 kyr in duration. While the durations for the suborbitally driven cycles in China and northern Italy are not exactly the same, the presence of suborbitally driven cycles in such different geographic localities indicate global, suborbitally driven climate cycles in the Triassic.

8 Conclusions

[30] The new magnetostratigraphic results presented in this study show a N-R-N-R reversal stratigraphy for the Rio Sacuz section that can be correlated with a modified version of Hounslow and Muttoni’s [2010] geomagnetic polarity time scale for the Triassic. Based on a Middle Triassic polarity interval duration of ~0.25–0.5 myr [Hounslow and Muttoni, 2010], the new magnetostratigraphy indicates a depositional duration of ~1–2 myr for the Rio Sacuz section and hence the entire Latemar platform. A magnetic susceptibility cyclostratigraphy collected from the Rio Sacuz sequence further refines the duration of the Latemar's deposition to 760–1000 kyr. Our new magnetostratigraphy and magnetic susceptibility cyclostratigraphy from Rio Sacuz support the resolution of the Latemar controversy in favor of the shorter, myr-scale depositional duration of the Latemar platform. It also indicates the observation of ~1.7 kyr long sub-Milankovitch global climate cycles in the Middle Triassic. More generally, our work highlights the importance of having absolute, independent time control for identifying Milankovitch-scale climate cycles in sedimentary sequences.

Acknowledgments

[31] We thank Alessandro Marangon and Alessio Ponza for their assistance with the field sampling and Linda Hinnov for comments that improved the manuscript. This research was supported by a Geological Society of America Graduate Student Research Grant and by NSF grant EAR-0823477 to K.P. Kodama.

Erratum

  1. In the originally published version of this article, some of the data in the top part of Figure 8 was plotted in the wrong sequence, leading to a revised spectral estimate (bottom part of the figure). The figure has since been corrected and this version may be considered the authoritative version of record.

Ancillary