Orbital pacing and secular evolution of the Early Jurassic carbon cycle

Significance Cyclic variations in Earth’s orbit drive periodic changes in the ocean–atmosphere system at a time scale of tens to hundreds of thousands of years. The Mochras δ13CTOC record illustrates the continued impact of long-eccentricity (405-ky) orbital forcing on the carbon cycle over at least ∼18 My of Early Jurassic time and emphasizes orbital forcing as a driving mechanism behind medium-amplitude δ13C fluctuations superimposed on larger-scale trends that are driven by other variables such as tectonically determined paleogeography and eruption of large igneous provinces. The dataset provides a framework for distinguishing between internal Earth processes and solar-system dynamics as the driving mechanism for Early Jurassic δ13C fluctuations and provides an astronomical time scale for the Sinemurian Stage.


ROCK-EVAL PYROLYSIS
Rock-Eval analysis was performed on 667 samples at the Department of Earth Sciences, University of Oxford. The equipment used was a Rock-Eval 6 Standard Analyzer unit by Vinci Technologies. The analyzer includes a pyrolysis and oxidation oven in which homogenized, powdered bulk rock samples are incrementally heated to 300-650 and 300-850°C for the pyrolysis and oxidation ovens, respectively. The flame ionization detector measures the quantity of free volatile hydrocarbons that are present in a sample before cracking (S1 peak) and the hydrocarbons and CO2 produced during the thermal cracking of insoluble organic matter (kerogen, S2 and S3 peak). After pyrolysis, the residual organic carbon is heated under air in the oxidation oven to determine residual organic carbon and mineral carbon content (S4 and S5 peak), also using the flame ionization detector. The total organic carbon (TOC, in wt%) is calculated as the sum of pyrolyzed organic C and residual organic C. The Hydrogen Index (HI, in mg HC/g TOC) represents the pyrolyzable organics relative to TOC. Oxygen Index (OI, in mg CO 2 /g TOC) represents the amount of oxygen relative to TOC. Tmax (°C) is the temperature with the highest rate of hydrocarbon generation during pyrolysis temperature (recorded at S2 peak), and is used as a thermal maturity parameter. The mineral carbon (wt%, TIC -total inorganic carbon) is calculated from the sum of pyrolysed and oxidized mineral carbon and was converted to calcium carbonate (CaCO 3 ). Laboratory procedures after Behar et al. (2) were followed. The in-house standard SAB134 (Blue Lias organic-rich marl) was regularly measured (every 8 to 10 samples). The standard deviation on TOC and HI of the in-house standard (SAB134) was 0.11% and ±29.95 mg HC/g TOC, respectively.
The standard deviation for Tmax, OI and TIC were at 1.9°C, 2.8 mg CO2/g TOC and 0.25%, respectively. The long-term average TOC of the international reference standard IFP 160000 is 3.27%, with a standard deviation of 0.04% (the TOC content of the IFP 160000 standard is referenced at 3.28±0.14 wt%).

ORGANIC PETROGRAPHY
Optical petrographic rock analysis was performed at the Department of Earth Sciences, University of Oxford, UK. A total of 14 samples from Hettangian, Sinemurian and upper Pliensbachian strata were analyzed. The samples were mounted in epoxy resin and polished according to the method described by Gorbanenko (3).
The polished blocks were analyzed with a Leica DMRX -MPVSP microscope photometer, using both reflected white light and fluorescence illumination under oil immersion at a magnification of ×500. Maceral analysis was undertaken with the point-count method (Peclon counting system) based on 1000 individual determinations per sample. The nomenclature of Taylor (4) was used for the macerals of the vitrinite and inertinite groups, and that of Hutton (5) for the liptinite group macerals. Results are presented in vol% relative to the total organic carbon content.

δ 13 C TOC ANALYSIS
About 1-2 g of homogenized powdered bulk rock sample were treated with ~40 ml of 3 M hydrochloric acid (HCl) and left in a warm water bath (~60°C) for about 2 hours.
Samples were then centrifuged and the HCl decanted. Carbonate-rich samples were treated with 3 M HCl a second time. The samples were then rinsed with deionized water until neutral pH was reached, and dried in an oven overnight at ~40°C. Dried samples were homogenized in an agate pestle and mortar and weighed into tin capsules (~10 mg of sample, aiming for 25µg pure carbon). The δ 13 CTOC analysis was performed at the NERC Isotope Geosciences Facilities, British Geological Survey, Keyworth, Nottingham (United Kingdom) by combustion in a Costech Elemental Analyser (EA) online to a VG TripleTrap and Optima dual-inlet mass spectrometer.
The δ 13 CTOC values were calibrated to the VPDB scale using in-house standards, which have been calibrated against international standards (NBS-18, NBS-19 and NBS-22). Replicate analysis of the in-house standards gave a precision of ± <0.1‰ (1 SD).

δ 13 C wood ANALYSIS
Macroscopic fossil plant material was extracted from reserve bags using a metal preparation needle or scalpel, resulting in 97 samples. Extracted wood and leaf fragments were treated with ~1 ml of dilute nitric acid (2 % vol/vol) for approximately 12 hours to dissolve any carbonate remains from attached rock matrix and calcite and other carbonate minerals that might have impregnated the wood or precipitated in cracks. The samples were washed with deionized water at least 5 times. The residual liquid remaining after the washing steps was evaporated in an oven at 50°C within ~12 hours.
The resulting pure wood fragments were gently crushed where necessary and a target amount of 400 to 600 µg of material weighed (at 1 µg precision) transferred into tin capsules for mass spectrometry. Resulting sample weights were 53 to 830 µg, with amounts of < 400 µg in 18 of 110 analyzed samples dictated by sample availability. Bovine Liver and 32 Alanine standards is 0.06 ‰ for δ 13 C wood . Analyses of Alanine yielded a much wider spread of carbon content, because Alanine was pipetted onto an absorbent material rather than weighed like the other unknowns and standards used for mass spectrometry.

SPECTRAL AND TIME-SERIES ANALYSIS
Spectral and time-series analysis on the Mochras δ 13 C TOC record was aimed to test whether the medium-amplitude carbon isotope excursions (CIEs) were paced by longeccentricity (405-ky) orbital parameters. Data preparation and spectral and time-series analysis were performed using the R Package for Astrochronology, version 0.3.1 (6) on δ 13 C TOC data from the Hettangian to Pliensbachian δ 13 C TOC data of this study, combined with data from the upper Pliensbachian and lowermost Toarcian (7) of the core, together covering the stratigraphy between 1890 and 855 mbs. The lower part of the planorbis Zone (1906.7 to 1890 mbs) was excluded from the analysis, as it is comparatively thin. Moreover, there is a possible stratigraphic gap and associated hiatus at the sharp lithological change at the inferred Triassic-Jurassic boundary.
Analysis of the full-length data set appears unsuitable for defining the dominant frequencies of the δ 13 C TOC shifts in the Mochras record as the frequency range of dominant spectral components is changing to shorter frequencies along the dataset as shown by wavelet analysis ( Figure S2). This trend is ascribed to sedimentation rate changes. Furthermore, interpolation of the full data set to equal sample spacing adds a bias to the data, creating a considerable number of artificial data points within the intervals of low data resolution (Hettangian to upper Sinemurian), and removing data in the higher-resolution intervals in the lower and upper Pliensbachian (see Figure S1 for sample resolution).
In order to precisely determine the dominant spectral peaks in the δ 13 C TOC record and to avoid sample biases such as aliasing, the δ 13 C TOC record has been divided into three individual segments based on the wavelet analysis of the full data set ( Figure S2, S3), as well as on differences in the original sample spacing, with the aim being to reduce the number of artificial data points during interpolation to equal sample spacing. The individual segments overlap slightly to avoid artificial peaks at the end-points during spectral analysis. Each segment has been manipulated to the average sample spacing of the grouped data set using linear interpolation.
Spectral analysis on the individual segments was performed using 3π multi-taper spectral analysis with robust red-noise model. Each segment shows a dominant spectral peaks in a wide range of frequencies, some of these overlap with those identified in elemental Ca concentrations (45.4-8 m, 8) and the digitized gamma-ray log (32 m, 9), of the uppermost Sinemurian to Pliensbachian strata of the Mochras core, which have been interpreted to correspond to long eccentricity (405-ky) cycles.
The average spectral misfit (ASM) method for astrochronologic testing (10, 11) was used to evaluate the dominant spectral peaks for each segment (Null-hypothesis test of orbital influence). For segment 1 and 2 all spectral peaks ≥95% MTM harmonic F-test confidence level were evaluated for ASM, utilizing 10,000 Monte Carlo simulations.
For segment 3 spectral peaks ≥90% MTM harmonic F-test confidence level were evaluated. Predicted orbital periods (E1=404.8 ky, e1=132.4 ky, e2=99.8 ky, o1=39.9 kry, p1=24.1 ky) were determined from (Ref. 12). Precession (p1) was not used in ASM testing of segment 1 due to lower sample spacing. Average sedimentation rate for the studied interval is estimated to ~5.7 cm/ky, based on the total thickness of the studied section and the duration of the studied interval based on radioisotoic ages (13)(14)(15)(16)(17)(18). ASM results were tested for sediment accumulation rates between 2 and 10 cm/ky for segment 1 and 2, and 1 and 6 cm/ky for segment 3, which corresponds to a stratigraphic interval in which a lower sedimentation rate is expected (8). Results with Null Hypothesis (H 0 ) significance level of ≤0.5% were identified ( Figure S3).
The ratios of statistically significant peaks identified as E1, e1, o1, and p1 (see Figure   S3) are similar to the ratio 20:5:2(:1) associated with orbital long eccentricity, short eccentricity, obliquity (and precession) target frequencies (Segment 1: 18:4 Pliensbachian are in comparable range to the spectral components found in elemental calcium concentrations and gamma ray logs in previous studies (8,9).
In order to generate a relative (floating) timescale, individual 405-ky orbital cycles were allocated to the δ 13 C TOC record based on the dominant spectral components ( Figure S2). The data within each defined cycle was interpolated to represent 405-ky using linear interpolation.
Uncertainties occur at the stratigraphic interval corresponding to the oxynotum Zone (~1380 to 1280 mbs) where the spectral signal is relatively weak. Several coring and associated sampling gaps, transition between samples referring to a depth-interval to core-slab samples, and the presence of a fault potentially cutting out some strata may inhibit the determination of a strong spectral signal in this interval.

MINERAL MATRIX EFFECT
TOC, HI, and OI values obtained by Rock-Eval analysis from organic-lean samples (<0.5 wt% TOC) are considered to be unreliable due to possible analytical uncertainties (no clear S2 and S3 peaks are generated). Hence, TOC, HI and OI values corresponding to samples with TOC <0.5 wt% (indicated as red squares in Figure S4) were excluded from interpretation.
A common factor artificially reducing TOC and HI values during analysis is the matrix retention of hydrocarbons in the presence of clay minerals (mineral matrix effect) (21), and CO 2 generated from carbonate may lead to elevated OI values (22).
The mineral matrix effect is largest in organic-lean samples and those containing abundant clay minerals (especially illite) in the mineral matrix (23). Based on lithological descriptions (22,24) and observations during sampling for this study, higher amounts of clay minerals are likely to occur in the mudstones present in the lower succession (Hettangian to lower Sinemurian) of the Mochras core where the sediments appear softer and darker compared to the carbonate-rich upper Sinemurian and Pliensbachian strata. This difference suggests a likely greater mineral matrix effect in the stratigraphically lower part of the Mochras core. There, low TOC and HI values (~1.5wt% and ~80 mg HC/g TOC on average, respectively), and the gradual increase in TOC and HI values along the Sinemurian-Pliensbachian transition (2.6wt%, up to 380 mg HC/g TOC, respectively), could therefore be a function of retention within the mineral matrix.

THERMAL MATURITY
Thermal maturity can affect TOC and HI values due to the migration of generated hydrocarbons. Both Rock-Eval pyrolysis derived T max values (428 °C on average) and Vitrinite reflectance (R 0 max = 0.38-0.63)(25) indicate moderately low maturity for the Hettangian to Pliensbachian strata. The average T max is furthermore likely to be slightly overestimated due to the matrix-retention effect (5-6 °C for Type II kerogen and 10-12°C for Type III kerogen) (26). The generally lower HIs recorded from Hettangian and Sinemurian strata are therefore not related to higher maturity levels in the deeper part of the core.

HI versus OI
The van Krevelen-type plot (HI versus OI, Figure S5) of the Mochras bulk sediments indicates relatively mature Type III (0-200 mg HC/g TOC, typically sourced from terrestrial organic matter) and Type II (200-400 mg HC/g TOC mixed marine and terrestrial) organic matter (27). Only four samples can be classified as Type I (marine, 400-600 mg HC/g TOC).
Degraded marine organic matter commonly shows a pyrolysis signature almost indistinguishable from terrestrial material (21). Oxygen availability during early diagenesis in the sediment (and likely also in the water column during sinking of the organic particles) can greatly compromise the HI (28) Figure   S6, S7). Liptodetrinite is thought to be produced from the physical breakdown of different marine algae or terrestrially derived liptinites such as sporinite and cutinite derived from the outer wall of spores and pollen, as well as leaves and stems (21) and is associated with high levels of water-column oxidation (29). The highly degraded shape of liptodetrinite macerals observed in Mochras ( Figure S6) suggests fragile marine organic matter as primary source. Furthermore, the fluorescence of terrestrial organic matter should decrease with greater biodegradation (30). The fluorescence of liptodetrinites in the Mochras core instead appears higher compared to intact terrestrial liptinites, indicating primarily marine precursors.
The highly oxidized nature of the suggestively primarily marine organic components is confirmed by the lack of correlation between the relative amounts of marine versus terrestrial organic matter as determined by microscopic evaluation and δ 13 C TOC and HI values ( Figure S7). Even samples with high relative amounts of liptodetrinite (72-94vol.%, e.g., samples 1747.2, 1656.3 and 1392.9 mbs, Figure S7) comprise HI values in a range generally associated with terrestrially derived organic matter (65-83 mg HC/g TOC). Based on the organic petrography it is thus evident that HI values generated by Rock-Eval pyrolysis are highly compromised and not indicative of the source of organic matter.
Samples corresponding to the upper Pliensbachian strata contain higher relative amounts of bituminite, and some samples include higher amounts of alginite compared to the samples corresponding to Hettangian and Sinemurian strata ( Figure   S7). Bituminite, also known as amorphous organic matter (AOM), is the product of bacterial decomposition of algae and faunal plankton (31) under low-energy anoxic conditions (29). The presence of bituminite in samples therefore indicates low-oxygen bottom water conditions during the Pliensbachian.

δ 13 C versus HI
A negative linear relationship between hydrogen indices indicative of the approximate source and source mixture of organic matter and compound-specific isotope values can generally be used to determine the impact of organic-matter changes on shifts in the δ 13 C TOC record (21,32,33), given that the HI values are well preserved and thus indicative for the primary source. Isotopic variations corresponding to compositional changes of bulk organic matter will plot on, or close to a mixing line defined by the δ 13 C TOC and HI values for the marine and terrestrial end-members (21,32,33). The HI-δ 13 C TOC plot of the composite Mochras dataset indicates two main trends, one of which comprises paired HI-δ 13 C TOC data plotting along the suggested background values and might indicate that shifts in δ 13 C TOC may be associated with variations in the primary source of the organic matter. The second group, corresponding to paired HI-δ 13 C TOC data related to the main large-scale CIE intervals, plots on a steeper slope, indicating that changes in δ 13 C TOC not being related or only marginally related to changes in HI, with the exception of the upper Pliensbachian margaritatus zone positive excursion ( Figure S5). As discussed above, the HI values are likely not indicative for relative changes in the source and source mixture of the bulk organic matter but rather compromised by oxic degradation. A similar plot testing a link between δ 13 C TOC and the relative amount of marine/terrestrial components in the bulk organic matter based on microscopic evaluation ( Figure S7) does not show a correlation between the relative amount of marine/terrestrial organic constituents and isotope values. Although data are available for the Sinemurian-Pliensbachian transition, which is marked by a large ~4‰ negative CIE, it appears unlikely that an increase in marine organic matter contributes to the magnitude of shift as marine organic matter comprises the predominant amount of the bulk organic matter in both, the underlying and above strata. It is, however, noteworthy that the available data points (14 samples analyzed) may not be representative enough to resolve the impact of relative changes in the primary organic matter source, or preferential preservation of more refractory terrestrial organic matter (signature characterized by more positive Hettangian to upper Sinemurian stratigraphic interval, which, based on the organic matter assemblage, is associated with better-oxygenated bottom waters compared to the upper Sinemurian to Pliensbachian (34, and this study). As similar CIEs have been recorded in δ 13 C carb (see discussion in the main text) it appears unlikely that the medium-amplitude CIEs in δ 13 C TOC are solely driven by changes in preservation.

CONCLUSION
The

Triassic
Rha.  and band-pass filter of dominant spectral peaks identified as 405-ky signal for each segment (red, bandpass parameter given in figure). Key to ammonite subzone numbering given in Figure S1.   Samples with TOC values < 0.5 wt% and corresponding HI and OI are marked in red. δ 13 C TOC samples shown as white squares indicate samples taken from reserve collection bags, depth refers to mid-point of reserve bag sampling interval. For ammonite subzone nomenclature, see Figure S1.