The C Record

Fig. 1A presents an up-to-date version of the Neoproterozoic composite 513C record, modified from Halverson et al. (2005). This record includes new data from the Little Dal Group (Mackenzie Mountain Supergroup), northwest Canada, and incorporates new radiometric ages, most importantly, two U-Pb dates from the basal Doushantuo Formation—the cap carbonate sequence to the Marinoan-aged Nantuo glacials in south China (Condon et al., 2005) and ages on the Sturtian glaciation that suggest an age closer to 710-700 Ma (Fanning and Link, 2004). The principal difference between this compilation and that presented by Halverson et al. (2005) is the assumption that the Petrovbreen Member diamictite—the older of two glacigenic units in Svalbard—may be Sturtian in age rather than Marinoan, as argued in Halverson et al. (2004). However, this difference does not change significantly the overall structure of the 513C curve. More problematic to this compilation are persistent uncertainties regarding the timing, nature, and global correlation of the Sturtian glaciation. The Sturtian glaciation is here assumed to span from ca. 715 to 700 Ma (Fig. 1A) based on radiometric constraints from pre-Marinoan glacial deposits in Oman (Brasier et al, 2000; Allen et al, 2002) and Idaho (Fanning and Link, 2004) (Table 1). Although the glaciations are treated as discrete events for the sake of constructing the compilation (i.e., no data are included within the Sturtian and Marinoan glacial intervals), it is becoming inceasingly apparent that the pre-Marinoan record is much more complex.

Figure 1. (on Page 237) (A) Composite 813C record based on correlations shown in (B) and modified from Halverson et al. (2005) with new data included from the Little Dal and Coates Lake groups in NW Canada. This compilation is based on the correlation (B) of the Petrovbreen Member diamictite in Svalbard with the Sturtian glacials in Namibia (Chuos) and northwest Canada (Rapitan). The implication of this correlation is that negative 813C anomalies precede both the Sturtian and Marinoan glaciations. Symbols on the top line in (A) indicate prescribed ages used in constructing the timescale: star = direct age constraint; triangle = age constraint correlated from other succession with high degree of confidence; X = age constraint correlated from other succession with a moderate degree of confidence; diamond = arbitrary age constraint. The time scale is interpolated linearly between all imposed ages. Solid horizontal lines indicate duration of the contribution of carbon isotope data each from each of the four successions used in this compilation (NW Canada: Little Dal and Coates Lake Group; Svalbard: Akademikerbreen Group; N Namibia: Abenab and Tsumeb Subgroups; Oman: Huqf Supergroup). Solid + dashed lines show inferred time span of the Neoproterozoic sedimentary succession at each location (note that although the Oman sequence extends below the Sturtian, the interglacial record is almost completely absent; Le Guerroue et al., 2005). (B) Simplified stratigraphic sections of successions from which the carbon isotope data in (A) are derived, showing the correlations used as a basis for the compilation. U-Pb age constraints (in Ma) are shown in boxes. CLG = Coates Lake Group; RG = Rapitan Group; Om = Ombombo Subgroup; Ug = Ugab Subgroup.

Notwithstanding the ambiguity remaining in some correlations, the advantage of these compilations over previously published S13C records for the Neoproterozoic is that they are constructed from a limited number of thick, carbonate-rich successions for which high-resolution isotopic data are available. For all carbon isotope data, ages were assigned a posteriori through linear interpolation of fixed ages from successions from which the data is derived and assumed ages for the beginning and end of the Sturtian glaciation and the beginning of the Marinoan glaciation. Unfortunately, firm radiometric ages from these successions are few, and most of the calibration dates are correlated into the composite record from other successions, which unavoidably entails the risk of miscorrelation.

This method is not ideal and the resulting time scale is surely inaccurate in places, but the relative position of the data should be correct (apart from some mismatch across the intervals where correlations are made). Additional radiometric ages from other successions can then be applied to the record with varying degrees of confidence, based on correlation with the carbon isotope record and other considerations (such as other isotopic data).

Clearly, the composite record is far from a finished project, and just as the version here differs from alternatives presented in Halverson et al. (2005), so too will this version give way to improved compilations as new data become available and correlations are tested. In order to facilitate construction of improved records and integrations this record with other data sets, all 513C data from NW Canada, Svalbard, and Namibia included in the record are available at as composite sections.

2.3 Bases for Correlation

Due to the recognition of glacial deposits of clearly Sturtian and Marinoan affinity (Hoffman and Prave, 1996; Kennedy et al. 1998, Hoffman et al. 1998b) and the abundance of carbonate section spanning the two glacial horizons in the Otavi Group, the Neoproterozoic succession of northern Namibia serves as the backbone of the composite carbon isotope record (Fig. 1). The correlations between Cryogenian sequences used here fundamentally rest upon the assumption that the Chuos and Ghuab diamictites in Namibia are equivalent to the Sturtian and Marinoan glacials in Australia and the Rapitan and Stelfox glacials in NW Canada (Kennedy et al., 1998; Hoffman and Schrag, 2002; Halverson et al., 2005), although, as discussed below, new radiometric ages (including a 607.8 ± 4.7 Ma Re-Os age on shales from the purported equivalent of the upper diamictite in the Mackenzie Mountains; Kendall et al., 2004) have challenged this model. Since most of the data shown in the compilation are indubitably pre- and post-Cryogenian, the uncertainties in correlation do not profoundly affect the overall structure of the 513C record.

A U-Pb zircon age of 635.5 ± 1.2 Ma on the Ghaub glacials in central Namibia (Hoffmann et al, 2004) provides a key time constraint on the Marinoan glaciation. The thick (< 2 km) Tsumeb Subgroup, overlying the Ghaub glacials, presents an unrivaled post-Marinoan carbonate record. The age of the top of this passive margin sequence is poorly constrained, but is presumed to approximate (Halverson et al., 2005) the ca. 580 Ma onset of continental collision on the western margin of the Congo craton (Goscombe et al., 2003). Two pre-Sturtian U-Pb ages from the Naauwpoort Volcanics (746 ± 2 Ma; Hoffman et al., 1996) and the Ombombo Subgroup (760 ± 1 Ma; Halverson et al., 2005) are useful time markers within the Otavi Group but are not applied to the 513C compilation due to difficulty in correlating the fragmentary pre-Sturtian record from Namibia with the much more complete but virtually undated records in Svalbard and northwest Canada.

Whereas Halverson et al (2004, 2005) suggested that the Polarisbreen diamictites (Petrovbreen Member and Wilsonbreen Formation) collectively correlated with the Marinoan glaciation, more recent data suggest instead that the lower of these diamictites predates the Marinoan glaciation (Halverson et al., in review). If the Petrovbreen Member represents the Sturtian glaciation in Svalbard (e.g. Kennedy et al., 1998), then it follows that both the Marinoan and Sturtian glaciations were preceded by negative 513C anomalies of similar magnitude, thus minimizing the use of a pre-glacial anomaly as a correlation tool. Furthermore, purported glendonites between the two glacial intervals (Halverson et al., 2004) could be roughly coeval with recently discovered glendonites in strata between the Rapitan and Stelfox glacials in NW Canada (James et al., 2005), and perhaps account for the growing body of evidence for glaciation at ca. 680 Ma (e.g. Lund et al., 2003; Zhou et al., 2004; Fanning and Link, 2004; Kendall et al., 2005).

Although this correlation does not dramatically alter the shape of the 513C record, it does have important implications for the ages of other North Atlantic glacial deposits and the duration and completeness of the pre-Sturtian records in Svalbard and northwest Canada, as discussed below. Irrespective of whether the Petrovbreen Member is Sturtian, Marinoan, or something in between, the Akademikerbreen Group in Svalbard is entirely pre-Sturtian in (Halverson et al., 2005), meaning that the Hekla Hoek Series preserves a very complete (2 km) carbonate record (Knoll and Swett, 1990) for a period within the Neoproterozoic that is not well understood (Figs. 1-2).

Although the Neoproterozoic succession in northwest Canada is not well dated, close similarities between the Sturtian and Marinoan cap carbonate sequences, the interglacial 513C record, and strontium isotope data support the correlation between the Rapitan and Ice Brook (Stelfox) glacials in northwest Canada and the Chuos and Ghaub glacial in Namibia (Kennedy et al., 1998; Hoffman and Schrag, 2002). It follows from this correlation that the Coates Lake Group in northwestern Canada is pre-Sturtian in age (Figs. 1-2). The Rapitan and Coates Lake groups are separated by an unconformity (Jefferson and Ruelle, 1986), which means that the latter likely does not preserve a complete record leading into the Sturtian glaciation. The contact between the Coates Lake and Little Dal groups is also unconformable (Fig. 2), and given that the former was deposited during a phase of regional extension (Jefferson and Ruelle, 1986), the time span between the top of the Little Dal carbonates and the base of the Coates Lake carbonates could be significant. Locally, the Little Dal Basalt, which is inferred to be ~780 Ma based on geochemical similarity to mafic dikes and sills that intrude the Mackenzie Mountain Supergroup (Jefferson and Parrish, 1989, Harlan et al., 2003), occurs at this contact and appears to be conformable with the top of the Little Dal carbonates (Aitken, 1981). The Little Dal Basalt thus provides a potentially useful calibration point in the 513C record.

The Huqf Supergroup in Oman is one of the best documented and most complete stratigraphic sections spanning the Ediacaran Period (Gorin et al., 1982), and the carbonate-rich, latest Neoproterozoic section is superbly preserved in outcrop and drill core (Amthor et al, 2003, Le Guerroue et al., 2006). Radiometric ages from the Precambrian-Cambrian boundary interval pin the age of the boundary at 542 Ma and constrain the duration of the negative 513C anomaly associated with the boundary to < 1 m.y. (Amthor et al., 2003). Oman was also one of the first places (along with South Australia) where the large, post-Marinoan Shuram (or Wonoka) negative 513C anomaly (Halverson et al, 2005) was first documented; the 513C record from the Huqf Supergroup (Burns and Matter, 1993; Amthor et al., 2003; Cozzi et al, 2004; Le Guerroue et al., 2006) is among the most complete spanning this anomaly.

The Fiq glacials and overlying Masirah Bay Formation cap carbonate sequence are equivalent to the Ghaub-Maieberg in Namibia (Leather et al., 2002, Hoffman and Schrag, 2002, Allen et al., 2005) and constitute one tie point between these two successions. Unfortunately, since the Masirah Bay Formation (cap carbonate sequence) is predominantly siliciclastic above the Haddash cap dolostone (Allen and Leather, 2006) and the Tsumeb Subgroup in Namibia appears to be truncated beneath the Shuram/Wonoka anomaly

Figure 2. (on Page 241) Pre-Sturtian composite stratigraphic and 813C records from Northeast Svalbard (Halverson et al., 2005), the Mackenzie Mountains (this paper), and central Australia (Hill et al., 2000). The correlation shown implies that the succession in the Mackenzie Mountains preserves a significantly older record of 813C than found in Svalbard and Australia. G1 and S1 designate the isotopic shifts and associated sequence boundaries (in Svalbard), that define the so-called Bitter Springs Stage (Halverson et al., 2005). COATES L = Coates Lake Group; RR = Redstone River Formation. Note the change in scale between the Coates Lake and Little Dal Groups.

(Halverson et al., 2005), it is impossible to tie the complementary Nafun and Tsumeb 513C records precisely. However, the compilation of 513C data from the Nafun Group supports the argument that there was only one major 513C anomaly in the middle Ediacaran period (Le Guerroue et al, 2006). Thus, the correlation between a sharp downturn in 513C in the upper Kuiseb Formation (basin facies equivalent of the upper Tsumeb Subgroup) proposed by Halverson et al. (2005) is maintained here. It should be noted, however, that Condon et al. (2005) proposed a significantly different time scale for the Wonoka/Shuram anomaly, based on radiometric and carbon isotopic data from south China, indicating instead that the nadir of this anomaly significantly post-dates the Gaskiers glaciation and is perhaps as young ca. 555 Ma.

Was this article helpful?

0 0

Post a comment