Australia: The Land Where Time Began

A biography of the Australian continent 

The Marinoan Glaciation

The Polarisbreen Group¹

In the northeast of the Svalbard Archipelago there are Neoproterozoic deposits that are 7 km thick, the top 1 km of which is the Polarisbreen Group, that includes mixed carbonate-siliciclastic rocks in 2 units that are separate and distinct and both are rich in diamictite. The accumulation of the platformal succession occurred on the windward side of Laurentia that was tropical to subtropical at the time. The Petrovbreen Member, a thin glacimarine diamictite lacking a carbonate cap, containing clasts that were derived locally, overlying a regional karstic disconformity that was preceded directly by a large δ¹³C anomaly, more than 10‰, in the shallow marine carbonates beneath it. In Australia, Canada and Namibia there are anomalies that are homologous with this anomaly, that precede the Marinoan Glaciation. Extra-basinal clasts are contained within the Wilsonbreen Formation, that is younger, and contains ice-contact deposits of terrestrial origin, and the formation is draped by a transgressive cap dolostone that 3-18 m thick. There are may sedimentary features within the cap dolostone that are associated globally with caps dating to post-Marinoan times, its isotopic profile being virtually identical to other Marinoan cap dolostones. The authors² suggest the 2 diamictite-rich of the Polarisbreen Group should represent the fist and final phases of the Marinoan Glaciation from an inter-regional perspective. The MacDonaldryggen Member above the Petrovbreen diamictite is about 200 m thick, is comprised of dark-olive rhythmites that are finely laminated, that have been interpreted as representing suspension deposits that were beneath multi-annual sea ice (sikussak). The Slangen Member is a regressive sequence of dolomite grainstone and evaporitic supratidal microbialaminite that is less than 30 m thick situated between the Wilsonbreen diamictites above and the suspension deposits below. The authors¹ interpreted this sequence, that they described as Sabkha-like lagoonal, as an oasis deposit resulting from precipitation as local ice melted under forcing by greenhouse conditions, though the tropical ocean remained covered as a result of seaglaciers inflowing from higher latitudes. The Polarisbreen Groups appears to represent a record of the Marinoan snowball glaciation that is unusually complete.

A Caledonian (Silurian) fold belt in the northeastern part of the Svalbard Archipelago has exposed a 7-km-thick succession of sedimentary strata covering the time period from the Neoproterozoic to the Middle Ordovician. A platform that was contiguous with the windward tropical to subtropical margin of Laurentia was formed by the Hecla Hoek Succession (Harland & Wilson, 1956; Knoll & Swett, 1990; Harland et al., 1993) and the similar succession in the East Greenland Caledonides (Hambrey & Spencer, 1987; Sønderholm & Tirsgaard, 1993; Fairchild & Hambrey, 1995). There are 2 separate and distinct units that are glacial diamictites-rich (Harland et al., 1993) with in the mixed terrigenous and carbonate sequences of the Akademikerbreen Group that underlies the Polarisbreen Group from the Late Neoproterozoic.

According to the authors¹ it has been accepted widely that there was at least 1 large ice sheet that reached tidewater near the equator in the Neoproterozoic (Embleton & Williams, 1986; Schmidt & Williams, 1995; Sohl et al., 1999; Evans, 2000), the authors¹ suggesting that this scenario cannot be understood in terms of analogues from the Quaternary. It has been proposed that during the Sturtian Glaciation and the Marinoan Glaciation that entire ocean was frozen over (Kirschvink, 1992; Klein & Beukes, 1993; Hoffman et al., 1998), based on various lines of geological and geochemical evidence and simple energy-balance climate models (e.g. Budyko, 1969; Sellers, 1969; Caldeira & Kasting, 1992).

The conditions that initiated the low-latitude glaciation have been well-studied in marine sedimentary records and isotopic proxy records (Halverson et al., 2002; Schrag et al., 2002) and the conditions leading to the termination of the glaciation have been studied (Fairchild, 1993; Grotzinger & Knoll, 1995; Kennedy, 1996; Hoffman et al.,, 1998; Kennedy et al., 1998; James et al., 2001; Hoffman & Schrag, 2002; Higgins & Schrag, 2003). A unique carbon isotope anomaly (McKirdy et al., 2001; Halverson et al., 2002; Hoffman & Schrag, 2002) lies directly beneath deposits associated with the Marinoan Glaciation in the 3 best-known regions, the Adelaide Rift Complex of South Australia, the northern Canadian Cordillera, and in Namibia, the Congo cratonic margin. The Marinoan Glaciation deposits are overlain by a cap-carbonate sequence that is lithologically unique (Kennedy, 1996; Soffer, 1998; James et al., 2001; Hoffman & Schrag, 2002). There is no premonitory isotopic anomaly associated with the Sturtian Glaciation, and it has a cap-carbonate sequence that bears no similarity, isotopically or lithologically, to the cap of the Marinoan Glaciation sequences (Kennedy et al., 1998). The authors¹ suggest that this should allow for a more reliable correlation basis than the glacial deposits alone, which are usually discontinuous and usually vary spatially.

In their paper the authors¹ present detailed stratigraphic and geochemical evidence covering a period of time over which the evidence of glaciation has previously been interpreted as 2 separate, distinct glacial periods in the Polarisbreen Group (Hambrey, 1982; knoll et al., 1986; Harland et al., 1993; Fairchild & Hambrey, 1984, 1995; Kaufman et al, 1997). The glacial units differ in their lithology, and there is a cap-carbonate sequence (basal Dracoisen Formation) associated with the Wilsonbreen Formation (the younger one) that is not distinguishable from that of other Marinoan cap carbonates. An isotope anomaly of the type that is consistently present beneath Marinoan Glaciation deposits in other regions is present beneath the Petrovbreen Member, the older glacial unit, though it is never found beneath the younger glacial deposits. The authors¹ suggest there are 2 interpretations possible to explain this, either the correlation criteria that were developed for the Australian, Canadian and Namibian glacial deposits are faulty, or in Svalbard both glacial units, as well as the intervening strata, that include deposits of open-water lagoonal origin, are associated with a single Marinoan Glaciation. This paper is concerned with the authors' ¹ attempt to reconcile the evidence from Svalbard with that from elsewhere the Marinoan Glaciation was a snowball cycle (Kirschvink, 1992; Hoffman et al., 1998; Hoffman & Schrag, 2002. They also discuss the implications of their proposed correlation with regard to the Sturtian Glaciation in Svalbard.

Background geology

The Elbobreen, Wilsonbreen and Dracoisen Formations are all mixed siliciclastic-carbonate strata, from the uppermost Neoproterozoic of the Hecla Hoek Succession from northeastern Spitsbergen, comprise the Polarisbreen Group (Wilson & Harland, 1964; Fairchild & Hambrey, 1984). It overlies conformable the Akademikerbreen Group, a thick carbonate platform, and grades downwards into the Veteranen Group, siliciclastic strata beneath it. The Polarisbreen Group is separated from the Tokammane Group, of Cambrian age, by a cryptic unconformity (Harland et al., 1993) but a significant hiatus is suggested by biostratigraphy Knoll & Swett, 1987). Extending from southern Olav V Land in Spitsbergen to western Nordaustlandet the Polarisbreen Group is exposed in a north-south trending belt. According to the authors¹ in Nordaustlandet a different stratigraphic nomenclature was defined but they decided to use the Spitsbergen nomenclature in this paper as correlation with Spitsbergen is unambiguous (Harland & Gayer, 1972; Knoll et al., 1986; Fairchild & Hambrey, 1995). They have proposed 3 informal names for established members.

The Petrovbreen Member is the older unit of glaciogenic origin (Harland et al., 1993) in the Elbobreen Formation of the Polarisbreen Group. The Wilsonbreen Formation is comprised on glaciogenic diamictites that are younger and typically and much thicker. Sedimentological studies have been carried out on the glacial units (Chumakov, 1968; Hambrey, 1982; Fairchild & Hambrey, 1984; Harland et al., 1993), and geochemical studies have been carried out (Fairchild & Spire, 1987, 1990; Fairchild et al., 1989). There is no geochronological confirmation for them often being found to correlate with diamictite pairs of Neoproterozoic age in the North Atlantic region (e.g. Hambrey, 1983; Nystuen, 1985; Fairchild & Hambrey, 1995). According to the authors¹ the most widely accepted correlation is that with the Tillite Group in East Greenland (Kulling, 1934; Harland & Gayer, 1972; Knoll et al., 1986; Hambrey & Spencer, 1987; Fairchild & Hambrey, 1995). There is a consensus that the successions in East Greenland and eastern Svalbard formed in sedimentary basins that were contiguous, both of Neoproterozoic age, as they are similar in their entirety An early record of the secular variation of δ¹³C in carbonate and organic matter carbon in both successions, supporting the correlation, and attention is drawn to the connection between large negative δ¹³C anomalies and glaciations in the Neoproterozoic (Knoll et al., 1986). For the carbonates of the Polarisbreen Group additional isotopic data has been provided that includes stromatolitic limestones in the Wilsonbreen glacials. The data provided in this paper is said by the authors¹ to confirm and add to previous work.

Tectonic setting

Geochronology

Conclusion

There are 2 units that are rich in diamictite in the Polarisbreen Group of northeastern Svalbard that have conventionally been interpreted as 2 distinct glaciations, The Petrovbreen Member, the older, and the Wilsonbreen Formation, the younger. The diamictites of the Petrovbreen, that are derived from the Russøya Member carbonates, that lies directly beneath, that is overall thinner than the polymictic diamictites of the Wilsonbreen that include extrabasinal basement material. In the Russøya Member, directly beneath the erosional surface where the diamictites of the Petrovbreen had been deposited, there is a well-defined, 10‰, δ¹³C anomaly that is homologous with those present beneath the Elatina glacials of South Australia (McKirdy et al., 2001), the Ghaub glacials of northern Namibia (Halverson et al., 2002), and the Stelfox glacial of northwestern Canada (Hoffman & Schrag, 2002), all having been correlated with the Marinoan Glaciation in South Australian type section Kennedy et al., 998; Hoffman & Schrag, 2002). There is no well-defined cap carbonate in the Petrovbreen diamictites, though there is a cap dolostone (basal Dracoisen Formation) in the diamictite of the younger Wilsonbreen diamictites that is compositionally, sedimentologically and isotopically similar to the cap dolostones of Marinoan age in Australia (Kennedy, 1996), James et al., 2001) and Namibia (Hoffman & Schrag, 2002). The 2 diamictite-rich units in Svalbard are suggested by the authors² to apparently represent a beginning and end of a single Marinoan glacial cycle, from the perspective of the bounding strata and global correlations.

The authors² have proposed a new interpretation of the glacial history of the EGES platform consistent with these correlations as well as with the snowball Earth hypothesis for the Marinoan Glaciation. The EGES marine platform was located on the margin of Laurentia facing south, represented by the hinterland of Greenland of the present, in the Polarisbreen times, and at this time was in the subtropics of the Southern Hemisphere. At the end of the sedimentation of carbonate of the Russøya Member a base level drop of more than 50 m heralded the glaciation. An ice dome formed in East Greenland adjacent to the platform from which cold-based ice extended to tide water. Thick tills were deposited at the ice front by the ice dome and also released icebergs carrying large amounts of debris drifted across the platform dropping their debris as they melted that became the Petrovbreen glacimarine diamictite. The MacDonaldryggen Member, that overlies the ice-rafted debris, is a thick unit of suspension deposits, that are finely laminated, lacking dropstones but speckled with calcite pseudomorophs that the authors¹ interpret as glendonites. The authors¹ regard the onset of the full-blown snowball earth as being the contact between the Petrovbreen and the MacDonaldryggen, a time when sea-glaciers, thick marine ice covering the ocean surface that flows towards the melting or ablation zone, extended as far as the tropical oceans (Goodman & Pierrehumbert, 2003). They suggest the flux of ice-rafted debris stopped when the EGES platform became choked with sikussak ice, multi-annual sea ice that was shorefast. The ice dome glaciers were prevented from calving by backstopping sikussak, and the authors¹ suggest it may also have been pinned to the platform by impinging of the sea glacier on the edge of the shelf. The MacDonaldryggen suspension deposits that continued to accumulate beneath the sikkusak ice resulted from hypopycnal plumes and density flows that resulted from ice streams beneath the ice dome that was growing slowly.

As weathering couldn't keep pace with outgassing of carbon dioxide Snowball Earth warmed slowly, vapour pressure increased and there was more rapid flow of the sea glacier. The tropical oceans continued to be covered by thick ice because of the inflow of sea ice from higher latitudes, even when the temperature of the surface water reached melting point. An area of open water formed over the EGES as the sikkusak ice melted, as well as other subtropical marine platforms and inland seas, though the world ocean remained ice-covered. The Slangen Member, a sequence of sabkha-like evaporitic dolomite that is situated above the MacDonaldryggen Member, has been interpreted by the authors¹ to be the result of sediment deposition beneath this area of open water. Once this EGES oasis opened  the downwind ice dome expanded rapidly, advancing across the shelf and deposited the Wilsonbreen ice-contact deposits. Outwash streams redeposited the sands of the Bråvika Member in front of the advancing ice. According to this interpretation the Wilsonbreen ice sheet advanced almost to the northern limit of the Polarisbreen outcrop belt of the present when the sea glaciers collapsed resulting to a cataclysmic deglaciation and associated rise of sea level. The signature of this deglaciation is the basal Dracoisen cap dolostone. According to this interpretation there are 2 units, the suspension deposits of MacDonaldryggen Member and the sabkha cycle of the Slangen Member, that are not overtly of glacial origin, are associated with the full-blown Snowball Earth.

As both units rich in diamictite in Svalbard originated in the Marinoan Glaciation, the underlying strata should represent the Sturtian Glaciation (Knoll, 2000). There are no know glacial deposits beneath the Petrovbreen/Ulvesø diamictites. As the conformable nature of the Akademikerbreen Group, a vast shallow marine carbonate platform, would not provide the topography necessary for the nucleation of terrestrial glaciers, the authors¹ suggest it is not surprising there are no glacial deposits where glacial deposits from the Sturtian Glaciation would be expected to be found. It is possible that a cap carbonate sequence of Sturtian age is present in the Russøya Member, sedimentologically similar to the Rasthof Formation of Namibia (Hoffman & Prave, 1996; Hoffman et al., 1998; Kennedy et al., 2998; Yoshioka et al., 2003). There is no negative δ¹³C anomaly in the basal Russøya Member, though it does have a positive δ¹³C trend as is present in other cap carbonates from the Sturtian (Kennedy et al., 1998).

Evidence has been found in other regions for open water on shallow platforms during the Marinoan Glaciation (e.g. Williams, 1996; Prave, 1999; Williams & Schmidt, 2000; Kennedy et al., 2001; Condon et al., 2002). This evidence has always been cited by opponents of the Snowball Earth hypothesis, since Kirschvink (1992). Coming to the opposite conclusion, the authors¹ suggest open water should be exposed on shallow shelves and inland seas by the melting of shorefast sea ice in the relative ablative subtropics, long before the final collapse of the sea glacier. As a natural consequence of the snowball cycle oases should develop, and the authors¹ highlight the role of sea glaciers in critical phases of the cycle (Goodman & Pierrehumbert, 2003). They remind the reader that the snowball Earth cycle is fundamentally an ocean phenomenon the stratigraphic records relied upon for evidence of the events during the Neoproterozoic are all epicontinental.

Sources & Further reading

1. Halverson, Galen P., Adam C. Maloof, and Paul F. Hoffman. "The Marinoan Glaciation (Neoproterozoic) in Northeast Svalbard." Basin Research 16, no. 3 (2004): 297-324.