Stable Isotope Records from Paleocene Rocks of the Scaglia Rossa Formation in the Apennines Mountains of Italy

This study documents a convincing correlation of stable oxygen isotopic variation extracted from Paleocene aged bulk sediment carbonates of the Scaglia Rossa formation in the Apennines Mountains of Italy with the variations found in Pacific, benthic organism derived carbonates presented by Zachos et al. (2005). Using δO variation as a proxy for sea-level change, a first-order correlation can also be made with a proposed curve by Hardenbol et al. (2000) representing global fluctuations in sea-level (eustacy) based on sequence stratigraphy data. The Late Cretaceous and Paleocene climate has been characterized as warm and incapable of sustaining appreciable global ice. As the database of regional sea-level changes that are in agreement with eustatic curves over the Late Cretaceous and Paleocene grows, so does the argument for glacio-eustacy in a hothouse world since it is the only mechanism yet identified to produce large magnitude (20-100m), rapid (10ky-1My) fluctuations in sea-level (Miller et al., 2005). 2 Introduction Paleoclimate studies and paleontological evidence of the Cretaceous and Paleocene indicate that elevated temperatures existed from the equator through high latitudes (Berner et al., 1983; Arthur et al., 1985; Krassilov, 1981; Habicht, 1979; Vakhrameev, 1975; Colbert, 1973) often leading to the interpretation of an Earth incapable of sustaining appreciable global ice. However, fluctuations in global sea-level, eustasy, have been mapped over this time period and show rapid, high magnitude changes on the order of 100 meters over intervals less than 1 My (Haq et al., 1987; Hardenbol et al., 1995; Zachos et al., 2001). These fluctuations are difficult to account for on a global scale except from the transgression and regression of ice sheets (Miller et al., 2005; Immenhauser, 2004; Stoll and Schrag, 1996). Hardenbol and colleagues based their eusatic curves primarily on seismic sections, well log data, outcrops, and basin studies; while Zacho’s group analyzed the isotopic variation of deep-sea carbonates derived from benthic organisms from various oceans of the world as a proxy for sea-level change. In this investigation, bulk sediment carbon and oxygen stable isotope data extracted from surface exposures of pelagic carbonates of the Paleocene age Scaglia Rossa section in Italy are consistent with dramatic changes in local sea-level that can be correlated with the record established by Hardenbol and Zachos thereby supporting the proposal that the conditions are eusatic and driven by global ice. 3 Background The two studied outcrops have been previously dated within the uncertainties of biostratigraphic correlation (Alvarez and Lowrie, 1984) and sampled from the Paleocene's upper Danian through Selandian stages (57.4 62.5 Ma). Although adjustments have been made in placement and magnitude of fluctuations, eustasy curves consistently characterize this time period (beginning in the Late Cretaceous and continuing through the Eocene) with large-scale (20-100m), rapid (<1Ma) sea-level changes (Miller et al., 2005). Eustasy results from either a change in the volume of water in the ocean, or a change in the volume of the basin holding the ocean (Figure 1). Figure 1. Timing and amplitudes of geologic mechanisms of eustatic change. (Adapted from Miller et al., 2005) Water volume is primarily affected by global ice, thermal expansion, and to a lesser degree the amount of water locked in groundwater and lakes. The basin is affected by the processes of seafloor spreading, sedimentation rates, and continental collisions. In Kenneth Miller and colleagues' summary of sea-level changing processes, it is ice4 volume change that is primarily seen (and to a lesser extent thermal expansion) in the rapidly fluctuating oxygen isotope signal through this time (Miller et al., 2005). Since eustasy shifts are contributed to the afore mentioned processes, it is not a surprise that evidence from the last 65 Ma contained in deep-sea sediment cores have been linked to Earth's orbital geometry and plate tectonics as the perpetual forces responsible for these changes (Zachos et al., 2001). Orbital parameters of eccentricity, obliquity, and precession oscillate at periodic (and quasi-periodic) frequencies and determine the distribution and amplitude of incident solar energy. Zachos explains that: "orbitally related rhythms" found in the deep-sea data move about a climatic mean that is constantly drifting mainly in response to conditions dictated by plate tectonics, such asshifts in continental geography and topography, oceanic bathymetry and gateway locations, and concentration of atmospheric greenhouse gases. The idea that rapid eustasy before the Oligocene is primarily linked with global ice is in conflict with palenontological evidence of poleward expansion of floral provinces, coral reefs, as well as the occurrence of high-latitude dinosaurs. (Miller et al., 2001). Miller addresses this by exploring other lines of evidence included in tills, diamictons, and ice rafted detritus (IRD). Most compelling of this research is the discovery of IRD in easily dated pelagic sediments from the early Oligocene; however IRD has been sparse in retrieved cores. Immenhauser (2005) came to a similar conclusion when he attempted to untangle the processes of global eustacy by approaching the problem from a regional scale. From looking at sea-level curves from Britain, Russia, and Oman, he found that: although absolute numbers of sea-level varied, even with large error-bars he was able to correlate the regional to the eustatic. He notes that from all the known causes of sea-level change, only changes of mid-ocean ridge systems, regional tectonism, and glacio-eustasy are 5 capable of changing volume on the scale of meters to tens-of-meters per My. By using particular sample ranges based on the wavelength of local/regional tectonism and making positive correlations to the oscillating, rapid, large magnitude sea-level changes, he was able to exclude tectonism as a driver for the regional eustasy observed. Thermal anomalies causing rapid spreading at mid-ocean ridges were also excluded because they are characterized by rapid sea-level rise and a much slower fall. This behavior does not match the rapid regressive-transgressive cycle. Immenhauser (2005) came to the same conclusion as Miller et al. (2005) and found that on a regional scale that it is only global ice volume (or some unknown mechanism) that is responsible for the intense eustasy curves found in the greenhouse world of the late Mesozoic. Both Miller and Immenhauser used climate models to incorporate the new data and produce a scenario where ephemeral glaciers substantial enough to affect sea-level formed in the higher altitude interior of Antarctica during cold snaps in the greenhouse climate. Ice was restricted to inland Antarctica so it didn’t affect the water circulation or temperature around the continent. Very few, maybe only one, of these glaciers grew large enough to reach the coast. This hypothesis leaves the northern hemisphere ice-free as well as coastal Antarctica to support the flora and fauna found in the fossil record (Miller et al., 2005; Immenhauser, 2005). In this study, I continue to challenge the eustatic sea-level curves by investigating Paleocene aged pelagic carbonate rocks of the Scaglia Rossa formation in the Apennines Mountains of Italy by using bulk sediment oxygen and carbon stable isotopic variation as a proxy for sea-level change. Results will be compared to the eustatic curves to see if a positive correlation exists and if they support the hypothesis of glacio-eustacy occurring in a hothouse world. 6 Site Selection Even though global ice is proposed as the major factor in rapid eustasy on the geologic scale of the time period sampled, it is not the only factor that can affect the isotopic variation of the oceans. To reduce the primary influence on the oxygen isotopic variations to that of Paleocene climate variability, consideration needs to be given to the characteristics of the studied sequence and the environment in which they were deposited. A similar study has been conducted by Stoll and Schrag (2000) in Gubbio, Italy using the following guidelines: The section should be stratigraphically continuous over the interval studied, free of faults slumps, or folds. The section should be well dated with a resolution of biostratigraphic/magnetostratigraphic control high enough to allow correlation of multiple sections and to compare the isotope data to eustatic curves. The section should be homogeneous, which would be compromised by the presence of turbidites and biogenic shelf communities. The section should be of a relatively deep, exclusively pelagic environment because in shallow environments during sea-level falls, the oxygen isotope signal may be influenced by meteoric water input, which can precipitate more negative cement in the exposed sediment. Carbonate diagenesis is the final consideration. It is unlikely to find Paleocene carbonates that have not undergone at least some diagenesis, particularly the dissolution of primary calcite and the recrystallization of secondary calcite from pore fluids. This recrystallization may shift the mean δO by several per mil (Stoll et al., 2000). This possibility makes an interpretation of absolute temperature suspect; however the frequency variation in δO will be preserved, although partially attenuated. The formation studied here lay in the bathyal depths during the Late Cretaceous and Paleocene and is now exposed in roadcuts and outcrops in the Apennines Mountains of Italy. The outcrops at Pietralata and Bottaccione were originally deposited in the 7 deep-sea basin of Umbria well away west from any influence of the Adriatic carbonate shelf but well above the CCD (Kuhnt, 1990) at a depth of 1000-3000m. The Pietralata (Furlo) section was deposited in a sub-basin and although this makes it ideal for uninterrupted sediment accumulation, the geometry of it made it susceptible to turbidite flows thought to have occurred during underlying normal fault reactivation (figure 2). Figure 2. Present day geologic map (a) and paleogeographic reconstruction (b) of the Umbria-Marches Apennines. Rocks at Bottaccione were originally deposited in the Umbria pelagic basin and those at Pietralata were deposited in the Furlo sub-basin. (Adapted from Bice, D. 2006) The primary composition of undisturbed beds at both sites is 85-95% carbonate, which at Bottaccione is ~10% foraminifera (mostly planktonic) mixed with smaller calciferous nanofossils, and at Pietralata a range from nearly pure micrite to >70% (Montanari, 1980). The site sampled at Bottaccione meets all the considerations mentioned aboveit is composed of entirely pelagic limestone derived from sediments that settled in a relatively flat area of the Umbria pelagic basin. At Pietralata, however, as mentioned 8 above, it collected in a sub-basin created by movement on several underlying extensional faults. These faults remained active during deposition periodically mobilizing sediments. This action is recorded in the early lithology (Cenomanian-Coniacian) by slump folds and slide masses and in the late lithology (Campanian-Paleocene) by intercalation of white calcarenitic turbidites, composed almost entirely of planktonic foraminifera (Alvarez and Lowrie, 1984). There is no evidence of sediment removal by turbidity flows (e.g., rip-up clasts, channeling); however, the section from Pietralata is not homogeneous through the section studied. It will be assumed that sampling from the undisturbed beds at Pietralata represent continuous deposition; however, it is understood that the sampling may not represent an uninterrupted window. Magnetostratigraphic studies previously carried out on both sections correlate the undisturbed beds of Pietralata to those at Bottaccione (Alvarez and Lowrie, 1984). 9 Methods In the past four decades, marine stable isotope data has given researchers a means of understanding past climate through the stratigraphic record. Oxygen isotope data are used to determine past water temperature, local salinity variations, and the extent of glaciations. Carbon isotope records are used to constrain ocean nutrient and water circulation patterns, and the concentration of atmospheric carbon dioxide. Linking data with magnetostratigraphic and biostratigraphic indicators has led to global correlations set to a geologic time-scale. The stable isotopes of oxygen used in isotope stratigraphy, and hence in eustasy reconstructions, are O and O. Approximately 99.75 percent of the oxygen atoms in the ocean/atmospheric system are O, while only about 0.2 percent is O. The ratio of O/O in an unknown sample is compared to the ratio of a known sample, and the variation is expressed in delta notation, δO, where: The original reference standard for carbonate samples was crushed belemnite from the Cretaceous aged Peedee formation in South Carolina. However this supply, known as PDB, has been depleted, so other standards have been calibrated to PDB and used as an intermediate reference. In this study, we measured stable isotopes of bulk carbonate instead of single species foraminifer components. Bulk study allows for rapid analysis and requires very little sample material. Even though the δOcalcite changes with different species of nanofossils, δO records have shown that δO values of bulk sediment accurately record shifts in sea-surface temperature and δO of seawater over glacial cycles (Shackleton et al., 1993; Schrag et al., 1995). It is unlikely to find unaltered calcite from the Paleocene. During deposition, lithification, and aging, primary calcite can dissolve and 10 secondary calcite precipitate from pore fluid. For this reason it is not possible to get an absolute temperature from analysis, but rather a relative change that may be slightly dampened. Sediment samples were collected as drill dust from the meter locations noted. Samples have a carbonate composition of >90%. Approximately 2mg of bulk sediment were loaded into a copper capsule. Samples were dissolved on line in a common acid bath at 90°C and precision averages are 0.015‰ for Carbon and 0.03‰ for Oxygen. The isotopic composition of generated CO2 gas was analyzed using a duel inlet, gas source mass spectrometer located at the Pennsylvania State Univerisity. The isotopic values were calibrated to PDB through NBS-19. At least 10% of samples from each run were duplicated for isotopic analyses. Average values were used whenever duplicates were done. 11 Stable Isotope Data The carbon and isotopic results for the Bottaccione and Pietralata sections are listed in the Appendix as Table 1 and displayed above in Figure 3. The carbon isotopic record at Bottaccione is characterized by an overall increase in δC from the base of polarity zone 27N at meter-355 to top of 26N at meter-368 (an interval from 61.3-57.5 Ma). Within the 26R there are two ~0.5‰ negative excursions at meter-356.8 and meter-362.6, a notable positive excursion at the very top of 26R ~0.4‰, a smaller negative excursion at the transition to 26N (57.9 Ma), and then a noticeable, nearly linear increase throughout the 26N. Pietralata showed a similar overall increase in δC of a similar magnitude over the same interval, two ~0.5‰ negative excursions within 26R, a positive excursion at top most part 26R, and the same nearly constant increase during the 26N. The data at Pietralata seems to lack the detail seen in the Bottaccione data with an especially pronounced lack of resolution over the turbidite laden zone 26R. The oxygen isotopic record at Bottaccione has a saw-tooth, increasing trend from meter-355.0 to meter-358.6 through the entire 27N and early 26R with δO excursions ranging from 0.25-0.5‰. There is a period of relatively constant δO (varying only 0.3‰ over the span) continuing from meter-358.6 to meter-363.4 interrupted by an anomalous negative excursion of ~2.0‰ from meter-262 to -362.8. The plot then returns to a saw-tooth trend for the rest of the sampled interval with excursions ranging from 0.3-1.0‰. In addition to the previously mentioned anomaly there is another large negative excursion at meter-355.6 of ~2.0‰. The oxygen isotopic record at Pietralata has two positive excursions of 1.4‰ and 0.7‰ beginning at the base of 27N to meter47.8 in 26R over the low resolution area mentioned earlier. The remainder of the sample section has a slight increasing trend with runs ranging from saw-toothed to sinuous in shape. Notable excursions are two small decreases (~0.03‰) that both happen to occur 12 just before the minor reversals within the 26N and a sinuous shape with a peak-to-trough difference of 0.8‰ that happens at the 26N/25R transition. Using magnetostratigraphic correlation established by Alvarez and Lowrie (1984) both data sets are able to be correlated and fit to a geologic timescale. The only common tie points are the base of 27N, the base of 26N, and the top of 26N. The remainder of the data is plotted assuming a common rate of deposition between tie points. The δO records are compared with the proposed eustatic sea-level curve of Hardenbol et al. (2000), and the smoothed (five-point running average) curve of Zachos et al. (2005) in