Introduction

Significant events in Earth history are often associated with major changes in the carbon isotopic composition of marine carbonates (δ13Ccarb) and co-occurring sedimentary organic matter (δ13Corg). Globally correlatable excursions in marine δ13Ccarb records are often thought to be related to global changes in the carbon cycle, such as those induced by snowball Earth events in the Neoproterozoic1,2,3, the oxygenation of Earth’s atmosphere4,5,6,7, the evolution of Ediacaran metazoans8,9, as well as marine and terrestrial extinction episodes10,11,12,13,14. A common approach used to establish whether the variations in a δ13Ccarb record reflect changes in the isotopic composition of the ancient dissolved inorganic carbon pool is to assess the covariation between coeval carbonate and sedimentary organic carbon isotope records15,16,17,18,19,20,21,22. Classically, covariant δ13Ccarb and δ13Corg records are interpreted as evidence that both the carbonate and organic matter were originally produced in the surface waters of the ocean, and that they have retained their original δ13C composition10,19,20,21,22,23,24,25, while decoupled δ13Ccarb and δ13Corg records have been interpreted as evidence for diagenetic alteration16,19,20,26, the ‘Rothman ocean’ model27, or that local syn-sedimentary processes have made the δ13Corg record noisy8. The application of covariance between δ13Ccarb and δ13Corg is based upon the theoretical assumption that ‘….no secondary processes are known (or for that matter, conceivable) which always shift the isotopic composition of carbonate and organic carbon in the same direction at the same rate…’20. This assumption has been widely used to establish the original nature of Precambrian and Palaeozoic δ13Ccarb records derived from shallow platform and marginal marine carbonates10,15,16,19,20,22. Consequently, the analysis of coeval δ13Ccarb and δ13Corg values have been a fundamental approach in studies of Precambrian and Palaeozoic carbon cycling, because it is thought to distinguish geologically meaningful records of significant biogeochemical changes in Earth history from those records that have been altered by diagenesis.

Since shallow marine deposits may have been periodically subaerially exposed during sea-level oscillations, it is important to address the possibility of diagenetic alteration. This is particularly important since freshwater alteration has been shown to generate negative δ13Ccarb excursions that are similar in magnitude to those observed in early Earth history28. Although a variety of other diagnostic tools have been employed to assess the degree of alteration, including trace element ratios29,30, cathodoluminescence31,32, as well as the relationship between δ13Ccarb and δ18Ocarb records33,34,35, the covariance between δ13Ccarb and δ13Corg records is thought to prove that the system is ‘rock buffered’ and that the records have retained their initial δ13C values10,15,16,19,20,22. Remarkably, however, the effects of diagenesis on the relationship between carbonate and organic δ13C records have never been directly investigated.

Here we evaluate the effects of post-depositional alteration on paired δ13Ccarb and δ13Corg values from a core that has unequivocally been altered by freshwater and marine diagenetic processes (‘Clino’, Fig. 1). More than 470 paired δ13C measurements were conducted on this ~700-m core that was drilled into the margin of the Great Bahama Bank36. During the late-Pleistocene, multiple sea-level oscillations exposed the upper 120 m of the platform to the influence of meteoric waters. Ten subaerial exposure surfaces have been identified in the top 100 m of the core, each of which is proposed to have been related to a Pleistocene glacial period37,38. Evidence of both meteoric and marine diagenesis has been recorded within this 5.3 Ma record of marginal and shallow marine carbonates39,40,41, including the development of caliche crusts, blocky spar cements, large-scale dissolution and soil development37,38,40. Evidence of marine burial diagenesis includes non-depositional surfaces in the core42, which in some cases are associated with dolomites containing negative δ13Ccarb signatures43. Paired δ13Ccarb and δ13Corg analyses in these altered sediments are strongly covariant throughout the length of the core, particularly in the Plio–Pleistocene section of the record. These results demonstrate how post-depositional processes, linked in time by subaerial exposure, have shifted the isotopic composition of carbonate and organic carbon in the same direction at the same time.

Figure 1: Location of Clino core.
figure 1

Clino was drilled on the platform top of the Great Bahama Bank, and the water depth at the time of drilling was 7.6 m (ref. 36).

Full size image

Results

Bulk geochemical relationships from Clino

The Clino core has previously been separated into three diagenetic zones based upon petrographic characteristics and δ13Ccarb and δ18Ocarb values39. Although the bulk δ13Ccarb and δ18Ocarb (Fig. 2a, r2=0.44, P<0.05, n=465) and δ13Ccarb and δ13Corg values (Fig. 2b, r2=0.59, P<0.05, n=465) show statistically significant positive correlations, the relationships are variable within each of the different diagenetic environments (Fig. 3a,b). The δ13Ccarb and δ13Corg data (Supplementary Table 1) are considered within this framework.

Figure 2: Relationship between isotope records from the whole core subdivided by lithology.
figure 2

(a) Correlation between δ13Ccarb and δ18Ocarb values from the entire length of the core Clino (r2=0.44, P<0.05, n=465) subdivided by published lithological assignments37,42. (b) Relationship between δ13Ccarb and δ13Corg values from the entire length of the core Clino (r2=0.59, P<0.05, n=465) subdivided by published lithological assignments37,42.

Full size image
Figure 3: Relationship between isotope records subdivided by the diagenetic zone.
figure 3

(a) Correlations between δ13Ccarb and δ18Ocarb records from the diagenetic zones defined by Melim et al. (ref. 39): meteoric (blue circles, r2=0.01, P>0.05, n=53), mixing (purple circles, r2=0.81, P<0.05, n=58) and marine burial (orange circles, r2=0.22, P<0.05, n=354). (b) Relationship between δ13Ccarb and δ13Corg values subdivided by diagenetic zones: meteoric (blue circles, r2=0.21, P>0.05, n=53), mixing (purple circles, r2=0.87, P<0.05, n=58) and marine burial (orange circles, r2=0.06, P>0.05, n=354).

Full size image

Geochemical relationships within each diagenetic zone

The uppermost portion of Clino (0–100 m), corresponding to the vadose and freshwater phreatic zones35, is characterized by large variations in the δ13Ccarb record and rather constant, but negative δ18Ocarb values. Throughout this interval there are abundant subaerial exposures (Fig. 4), which have more negative δ13Ccarb values (but constant and negative δ18Ocarb) and high concentrations of trace metals such as Fe and Mn37. The δ13Ccarb and δ18Ocarb values are not statistically significantly correlated (Fig. 3a), the concentration of total organic carbon (TOC) is <0.1% (Fig. 4) and there is no statistically significant correlation between the δ13Ccarb and δ13Corg within the meteoric zone (r2=0.21, P>0.05, n=53, Fig. 3b). Between 100 and 200 m, there is a transition from negative to positive δ13Ccarb and δ18Ocarb values (the ‘mixing zone’35) (Fig. 4). This portion of the core exhibits a very strong positive correlation between the δ13Ccarb and δ18Ocarb values (Fig. 3a), as well as between the δ13Ccarb and δ13Corg values (r2=0.87, P<0.05, n=58, Fig. 3b). Below the ‘mixing zone’ there is a region in which there are relatively positive δ13Ccarb and δ18Ocarb values, an area interpreted as having been affected only by marine diagenesis39. The marine burial zone shows no statistically significant relationship between the δ13Ccarb and δ18Ocarb (Fig. 3a) or between δ13Ccarb and δ13Corg values (r2=0.06, P>0.05, n=354, Fig. 3b). The concentration of TOC increases through the ‘mixing zone’ and is an order of magnitude higher in the marine burial zone reaching values up to 1.2% (Fig. 4).

Figure 4: Geochemical records and lithostratigraphy of Clino.
figure 4

Total organic carbon (TOC) content, carbonate δ13C values, organic δ13C values and carbonate δ18O values (n=465 for each record) produced by this study from the Neogene carbonates in the core Clino. The data used to construct the simplified stratigraphic column presented in this figure were obtained from Kenter et al. (ref. 42) and Kievman (ref. 37). Diagenetic zones (‘meteoric’, ‘mixing’ and ‘marine burial’) were defined by published interpretations of both petrographic40 and isotopic constraints35. The record of percent dolomite was determined by X-ray diffractometry39.

Full size image

Discussion

The zone of meteoric alteration in Clino has the lowest δ13Ccarb (−2 to +2‰) and δ13Corg (−29 to −17‰) values, and no statistically significant covariance between δ13Ccarb and δ13Corg (Fig. 3b). Both the δ13Ccarb and δ13Corg values are significantly lower than those reported for modern shallow marine sediments from Great Bahama Bank, which average +4.5 and −12‰, respectively44,45. We suggest that these low δ13Ccarb values arise from the oxidation of organic matter, which imparts a low δ13C value to the dissolved inorganic carbon, along with cementation, mineralogical stabilization and recrystallization of the carbonate33,39. Concurrently, δ13Corg values became more negative as the labile marine organic matter was oxidized, and additional organic material was contributed from terrestrial C3 plants and freshwater algae, which colonized the newly exposed platform top. Multiple subaerial exposures of the platform top during the Pleistocene have superimposed the effects of such diagenetic processes on the carbonates and organic matter preserved in Clino, and as a result, the records currently observed are the cumulative product of these post-depositional changes.

Although preferential degradation of labile organic compounds can cause the δ13Corg value of the residual organic carbon to become more positive46, in the majority of cases degradation has been shown to produce residual organic carbon with more negative δ13Corg values47,48,49. However, these processes can only produce changes of up to 4–5‰ in the δ13Corg record46,48,49,50,51, and since the lowest δ13Corg value of sedimentary organic matter from the platform top is −17‰ (ref. 45), diagenetic reactions alone could not have produced the δ13Corg values of −29‰ observed in the upper 200 m of Clino. Consequently, a source of organic carbon with a δ13Corg value lower than −22‰ is required to produce the δ13Corg values observed in the top 100 m of the core. Such a source is likely to be terrestrial C3 plant matter, such as mangroves and freshwater algae, which have δ13Corg values ranging from −20 to −32‰ (ref. 52). Evidence of terrestrial plant contribution is provided by root casts observed in the subaerial exposure surfaces37 (Supplementary Figs 1 and 2). In addition, terrestrial organic matter is known to be preferentially preserved through time, especially in oxidizing settings where marine organic compounds have been found to be degraded twice as fast as terrestrial soil-derived organic compounds53. We suggest that these post-depositional processes may account for the low concentration of organic carbon (<0.2%), the negative δ13Corg values (Fig. 2) and the increase in the proportion of low-magnesium calcite40 in the section of the core affected by meteoric diagenesis.

The highest correlation between the δ13Ccarb and δ13Corg values (Fig. 3b) is observed between 100 and 200 mbmp, in the section of the core associated with a strong correlation between δ13Ccarb and δ18Ocarb. The strong positive correlation between δ13Ccarb and δ13Corg records in the ‘mixing zone’ (Fig. 3b) can be attributed to a gradient of post-depositional changes. The sediments and organic matter preserved closer to 100 mbmp exhibit low δ13Ccarb and δ13Corg values, because they have been repeatedly affected by freshwater diagenetic reactions and post-depositional contributions of terrestrial organic matter, as previously described. In contrast, the δ13Ccarb and δ13Corg values in the section of the core closer to 200 mbmp are comparatively more positive, and similar to those observed both on the modern platform top45 as well as those preserved in the marine burial zone (Fig. 4), suggesting that this section of the core has experienced fewer episodes of alteration and lower contributions, if any, of terrestrial organic carbon.

The δ13Ccarb and δ13Corg records in the marine burial diagenetic zone probably represent the least altered values within the entire core. The absence of covariance between δ13Ccarb and δ13Corg records, and the range of δ13Ccarb and δ13Corg values in the marine burial diagenetic zone (Fig. 3b) are similar to unaltered Pleistocene periplatform sediments deposited on the slope of the Great Bahama Bank45. Throughout the marine burial zone, there are minor variations in δ13Ccarb and δ13Corg that represent subtle changes in the source of the sediments through time, as well as diagenetic processes. An example of the influence that a change in source can have on both the δ13Ccarb and δ13Corg values is the synchronous change towards more positive values observed at 367 mbmp. At this depth, the background sediment type changes from a mixed peloidal-skeletal packstone with significant contributions from pelagic foraminifera, to a peloid-dominated chalky wackestone to packstone almost entirely devoid of planktic foraminifera42. The synchronous positive shifts in the δ13Ccarb and δ13Corg records are consistent with increased off-bank shedding as the platform prograded towards the Straits of Florida during the Pliocene54,55. Off-bank shedding would have contributed increasingly higher proportions of platform-derived carbonates and organic matter, which are characterized by relatively higher δ13Ccarb and δ13Corg values45. Minor fluctuations in the δ13Ccarb and δ13Corg records occur at marine hardgrounds (Fig. 4)42, and are likely associated with the oxidation of marine organic matter and the precipitation of dolomite below non-depositional surfaces43 within the marine burial diagenetic zone.

This data set clearly demonstrates how two post-depositional changes linked in time by periods of subaerial exposure, the diagenetic alteration of the carbonate and the post-depositional contribution of terrestrial organic carbon, can produce negative excursions with highly covariant δ13Ccarb and δ13Corg records. The excursion observed in the Neogene is similar in magnitude to those observed in Palaeozoic and Precambrian deposits. Whether or not those ancient deposits were exposed to the same degree of freshwater alteration as Clino is still a matter of debate28. In many cases, negative δ13Ccarb excursions have been interpreted to be pristine records of global carbon cycling15,19,20,21,56,57, because sedimentological evidence of subaerial exposure was not observed26,58. However, subaerial exposure surfaces can be cryptic in the rock record, and other workers have interpreted the same geochemical changes to be diagenetic in origin28,59. If the latter is true, and multiple sources of organic carbon contributed to the sedimentary organic matter preserved in the deposit, as was recently shown to be the case for the Ediacaran Shuram Formation in Oman60, then the model presented here could conceivably explain covarying trends in paired δ13Ccarb and δ13Corg records from the ancient geological record. Although higher level terrestrial plants were not present until the late Palaeozoic, the presence of terrestrial life in earlier time periods, including photosynthetic cyanobacteria, fungi and algae61,62,63,64,65,66, supports the possibility that ancient sedimentary organic carbon could have been composed of mixtures of marine and terrestrial organic carbon, in a situation analogous to the model of subaerial exposure proposed for the Neogene. In fact, the range in δ13Corg values of Precambrian sedimentary organic matter is the largest for any time period in Earth history67. Although the organisms were different in the ancient geological record, processes similar to those described here could have occurred.

In contradiction to the assumption that coupled negative excursions in δ13Ccarb and δ13Corg values can only be produced by changes in the global carbon cycle, these results suggest that post-depositional processes can play an influential role in generating covariant δ13Ccarb and δ13Corg values. Consequently, interpretations of strongly correlated δ13Ccarb and δ13Corg values from the ancient geological record should reconsider the influence that similar post-depositional processes may have in generating some of the coupled negative excursions associated with noteworthy biogeochemical events in early Earth history.

Methods

Sampling strategy

Clino was sampled at roughly 1.5 m intervals to obtain at least 50 samples per diagenetic zone (meteoric zone, n=53, mixing zone, n=58 and marine burial diagenetic zone, n=354). However, major sedimentological features such as subaerial exposure surfaces and hardgrounds were avoided to preserve limited core material. Such features had been sampled in previous studies39,40,41, which accounts for the larger ranges in δ13Ccarb reported in those studies. For each paired carbon isotope data point, roughly a gram of bulk sediment was powdered and homogenized to provide subsamples for carbonate and organic carbon isotope analysis.

Carbonate δ13C and δ18O measurements

Carbonate δ13C and δ18O values were analysed via dissolution in phosphoric acid using the common acid bath method68. The CO2 gas produced by the reaction of phosphoric acid and carbonate was analysed on a Finnigan MAT 251 (Thermo Fisher Scientific, Bremen, Germany). In each run of 24 samples, four standards were processed at the start of the run and two at the end, followed by a measurement of the zero enrichment. Data were then corrected for any fractionation in the reference gas during the run and for the usual isobaric interferences modified for a triple collector mass spectrometer. Data are reported relative to the Vienna Pee Dee Belemnite (VPDB) scale, defined for carbonates by the δ13C value of NBS-19 (1.95‰ versus Pee Dee Belemnite (PDB)69). The error for these analyses is <0.1‰ as indicated by replicate analyses of internal standards.

Organic δ13C and TOC measurements

Co-occurring sedimentary organic material was separated via dissolution in 10% HCl acid overnight, followed by subsequent vacuum filtration onto glass microfiber filters (Whatman GF/C). The insoluble residue (IR) on the filter was allowed to dry for at least 48 h, or until a constant dry weight was achieved. The weights of the insoluble material were quantified by subtracting the weight of the empty filter from the weight of the dried insoluble material and filter after filtration. Samples of the insoluble material were scraped off of the filters, weighed and packed into tin capsules and loaded into a Costech ECS 4010 (Costech Analytical Technologies Inc., Valencia, CA, USA), where they were combusted. The resulting CO2 gas transferred for isotopic measurement to a continuous flow isotope-ratio mass spectrometer (Delta V Advantage, Thermo Fisher Scientific). For every run of 36 samples, 12 internal standards were analysed to calibrate the machine and to assess the precision of the measurements. An analytic blank and 6 internal standards preceded the first sample analysis, and two standards were run for every 10 samples analysed. The reproducibility of δ13C values is ±0.1‰ as indicated by the s.d. of replicate analyses of internal standards of glycine (n=54, δ13C value=−31.8‰ VPDB). All δ13Corg data are reported relative to the VPDB scale, defined for organic carbon as the δ13C value of graphite (USGS24)=−16.05‰ versus VPDB70.

To calculate weight percent carbon in the IR, a calibration line was established that related the peak area measured by the Delta V Advantage (Thermo Fisher Scientific) to the known weight of carbon in the internal standard, glycine. The weights of the standards were chosen to bracket the expected range of organic carbon in the samples. The s.d. of these analyses is 0.4% based upon repeated analyses of glycine (n=54). Delta V Advantage peak area measurements for each sample was transformed to mg of organic carbon in the insoluble residue using the equation of the calibration line. Organic carbon concentration in the insoluble residue in mg was converted to TOC by the following equation:

TOC=((Org C in IR (mg) × total IR weight (mg))/initial weight of the sediment (mg)) × 100

Statistical analyses

Pearson’s regression analysis was used to determine the relationship between isotope records. The r2, P and n values are listed in the main text for each analysis conducted.

Additional information

How to cite this article: Oehlert, A. M. and Swart, P. K. Interpreting carbonate and organic carbon isotope covariance in the sedimentary record. Nat. Commun. 5:4672 doi: 10.1038/ncomms5672 (2014).