Australia: The Land Where Time Began

A biography of the Australian continent 

Elatina Glaciation, Late Cryogenian (Marinoan Epoch), South Australia – Sedimentary Facies and Palaeoenvironments 

The Elatina Glaciation from the late Cryogenian (Marinoan Epoch), South Australia, is named after the Elatina Formation of the glaciogenic Yerelina Subgroup, which covers ~200,000 km² in the Adelaide Geosyncline as well as on the cratonic Stuart Shelf and is up to ~1,500 m thick. Many facies that are like those of the glaciations in the Phanerozoic mark the Elatina glaciation:

·       Basal diamictite that displays glacitectonites with penetrative deformation of preglacial beds, which indicates grounded ice or scouring by icebergs.

·       Glaciomarine diamictites that contain many faceted and striated clasts up to several metres across of intrabasinal and extrabasinal origin.

·       Laminated siltstone and mudstone that contains scattered, ice-rafted dropstones in outer marine shelf environments.

·       Sandstones that were deposited in fluvial, deltaic and inner marine shelf settings.

·       Tidalites that were deposited in estuaries and tidal deltas during interstadial rise of sea level, with cyclic tidal rhythmites that recorded the annual oscillation of sea level and displaying ripple marks that were wave-generated, which together indicate long-lived, extensive open seas.

·       Siltstone that contains acicular crystal pseudomorphs which implies the formation of evaporite minerals in littoral deposits.

·       Permafrost regolith of frost-shattered quartzite breccia ≤20 m deep, that displays large-scale periglacial structures that include primary sand wedges, which are 3 m deep that indicate a frigid seasonal climate near sea level.

·       Periglacial sandstone that covers 25,000 km² and contains primary sand wedges near its base.

A spectrum of settings is recorded in these deposits that range from permafrost regolith and periglacial sand sheet on the Stuart Shelf in the west, to fluvial, deltaic, and inner marine-shelf in the western and central parts of the Adelaide Geosyncline, and outer marine-shelf in the north and southeast. It is indicated by the widespread and persistent rainout of fine-grained sediment and ice-rafted debris that the sea was not frozen over during the Elatina Glaciation. At present the only age that can be applied to the Elatina Glaciation is a maximum and minimum age limits of between ~604 Ma and ~580 Ma, respectively, as no direct determination is available. Deposition within 10^o of the palaeoequator is indicated by high quality palaeomagnetic data for red beds from the Elatina Formation, which is supported by positive fold tests on soft sediment slump folds that demonstrate the early acquisition of magnetic remanence. The Nuccaleena Formation “cap carbonate” overlies disconformably to unconformably the Yerelina Subgroup in the Adelaide geosyncline and on the Stuart Shelf.

The presence of Glaciomarine deposition, grounded ice and permafrost near sea level at near palaeolatitudes, a strong seasonal palaeoglacial climate, and widespread open seas implies a paradoxical palaeoclimate and palaeogeographic setting for the Elatina Glaciation.

A stimulus for worldwide multidisciplinary research on Cryogenian glaciogenic successions has been the strong evidence for a non-atualistic climate in the late Cryogenian in South Australia. In recent years the term “Marinoan Glaciation” for the younger of the 2 Cryogenian glaciations that left their mark in the record of the Neoproterozoic of most continents. In order to correlate these glaciations of the late Cryogenian with the glacial episode of the Marinoan Epoch in South Australia various lithostratigraphy and chemostratigraphic arguments have been employed (Coats & Preiss, 1987; Preiss, 2000), though the correlations have not been tested by direct dating. Unfortunately, there is no presentation that is available that discusses the current understanding of the stratigraphy, sedimentary facies, and palaeoenvironments of the of the late Cryogenian glaciogenic succession in its de facto “type region” of the Adelaide Geosyncline. In volumes of the Geological Survey of South Australia and the Geological Society of Australia (Coats & Preiss, 1987; Lemon & Gostin, 1990; Preiss, 1993) that have limited distribution beyond Australia and these do not include diverse stratigraphic and sedimentological studies and detailed palaeomagnetic analyses of this succession that were carried out over the past 2 decades. According to Williams et al. it is highly desirable that a synthesis that is up-to-date is in the international literature. This paper has the aim of meeting this need by bringing together the findings of pioneering stratigraphic studies by many workers from South Australia and recent research in order to demonstrate the wide range of glacial facies and palaeoenvironments that are associated with the Cryogenian glaciation in the Adelaide Geosyncline region. According to Williams this is timely as there is currently widespread interest in glacial environments from the Neoproterozoic and the Ediacaran System and Period with its global Stratotype Section and Point (GSSP) purportedly placed at the base of the Nuccaleena Formation “cap carbonate” which overlies glacial deposits from the late Cryogenian in the Flinders Ranges in South Australia (Knoll et al., 2004, 2006; Preiss, 2005). Also discussed are the appropriate terminology for the glaciogenic succession in the late Cryogenian in South Australia, and its palaeomagnetism and limited geochronology. The glacial palaeoclimate and Palaeogeography of the region, with important implications for the global climate of the late Cryogenian, are illuminated by the sedimentological and palaeomagnetic data.

Geological setting and stratigraphy

The glaciogenic succession from the late Cryogenian that is in this paper mostly accumulated in the sedimentary basin that dates to the Neoproterozoic to the Early Palaeozoic age that is now marked by folded strata of the Flinders Ranges and Mount Lofty Ranges in South Australia. There are several terms that have been applied by various authors to this basin such as “Adelaide Geosyncline”, “Adelaide Rift”, and “Adelaide Rift Complex”. The Adelaide Geosyncline concept was discussed (Preiss & Forbes, 1987) which stated (p. 15) that “although rifting played an important role in the early stages of development, it had little relevance during the later stages, and terms such as the Adelaide Rift     and Adelaide Aulacogen are not considered applicable to the whole basin throughout its evolution.” In this paper Preiss & Forbes, (1987, p.15) employing “Adelaide Geosyncline” herein “as a formal term in the broad and non-genetic sense of geosyncline as originally defined, i.e., a complex basin of very thick sediment accumulation.”

Rifting of a craton from the Precambrian in the Early Neoproterozoic initiated the Adelaide Geosyncline, which is now represented in the west by the Gawler Craton and by the Curnamona Province in the east. A number of inliers in the margins of the basin have exposures of the basement. Sm-Nd isochrones 867 ± 17 Ma and 802 ± 35 Ma (Zhao & McCulloch, 1993; Zhao, 1994) and a U-Pb baddelyite age of 827 ± 6 Ma (Wingate et al., 1998), were yielded by mafic dykes that are regarded as feeders for volcanic rocks near the base of the succession from the Neoproterozoic. The subsequent development of the Adelaide Geosyncline as a deeply subsiding basin complex from the Neoproterozoic to Cambrian was marked by a succession of major rift cycles early in the development, though rifting on a large scale was less important in its later history, including the Late Cryogenian (Preiss, 2000).

In the Adelaide Geosyncline the Neoproterozoic-Cambrian succession was deformed by the Delamerian Orogeny at 514 to 490 Ma (Drexel & Preiss, 1995; Foden et al., 2006). The following major subdivisions of the Adelaide Geosyncline region as was identified by Preiss (2000) are based on differing responses to deformation, though they also partially reflect primary sedimentary facies belts:

·       Cratonic platforms on the Stuart Shelf in the west and the Curnamona Province in the east, with thin, little deformed Neoproterozoic to Cambrian cover.

·       Meridional Torrens Hinge Zone of gentle folding.

·       Central Flinders Zone of broad dome and basin structures.

·       North Flinders Zone of arcuate, open to tight folds.

·       Nackara Arc of long, arcuate, folds that are relatively upright.

·       Fleurieu Arc marked by major orogenic shortening by thrusting and tight folding.

The Delamerian Origen is the term used for the zone of Delamerian deformation.

The term “Adelaide Fold Belt” has sometimes been used in recent literature for an orogen, though it should not be used for the sedimentary basin.

The Neoproterozoic stratigraphy of the Adelaide Geosyncline regions has been discussed by Preiss (1987a, 1993 and 2000) and Preiss et al. (1998). According to Williams et al. the full sequence can be divided into 12 sequence sets, and the lithospheric classification can be refined so that group and subgroup boundaries coincide with the most significant sequence boundaries. Each subgroup is, therefore, a discrete depositional entity, whereas formations that are within subgroups are characterised by particular facies associations and may have boundaries that may be gradational and/or diachronous. Boundaries of sequences may represent significant time gaps in the record and may have origins that are tectonic or eustatic, or both.

The Neoproterozoic succession of the Adelaide Geosyncline was divided (Mawson & Sprigg, 1950; Sprigg, 1950), which they termed the Adelaide(an) Syncline system, in 4 chronostratigraphic units or series rank: Willouran, Torrensian, Sturtian and Marinoan. These series, like the lithostratigraphic subdivisions, have boundaries that are now recognised as major sequence boundaries and therefore the chronostratigraphic significance that was implied by the original definitions remains valid.

In the Adelaide Geosyncline region the glaciation of the late Cryogenian is recorded by diverse tectonic and glaciosedimentary settings, which span permafrost regolith and periglacial-aeolian on the cratonic platform in the west, through fluvial, deltaic and inner marine shelf further east to outer marine shelf to the north and southeast. The deposits of the late Cryogenian are all included in the Yerelina Subgroup (Preiss et al., 1998) which represents the 10^th of the 12 sequence sets. Within the Yerelina Subgroup the stratigraphy and suggested correlations a possible division into 3 sequences marked by erosional boundaries and a proposed sea level curve (Preiss et al., 1998; Preiss, 2000). During the glaciation of the late Cryogenian there are no volcanism events that are known of, and active rifting on a large scale, such as characterised deposition in the Willouran, Torrensian, Sturtian times, was not important. Source regions include the Gawler Craton and the Curnamona Province, though the Provence of much glacial detritus remains to be identified.

Terminology

“Marinoan Glaciation” is a term that has been shown to be useful in South Australia, but based on recent stratigraphic evidence and geochronological data from other parts of Australia as well as other continents it is now not appropriate. The Marinoan Series was defined chronostratigraphically (Mawson & Sprigg, 1950) for strata between the top of the Brighton Limestone (~650 Ma) and the base of the Cambrian. The corresponding term for time is the Marinoan Epoch (Preiss, 1987b), therefore includes the Ediacaran Period. The Cryogenian glaciogenic succession has attracted worldwide attention in recent years. As a result of this the “Marinoan Glaciation” being applied indiscriminately to glacial deposits dating to the late Cryogenian from elsewhere in the world that are perceived, rightly or wrongly, to correlate with the glaciation from the late Cryogenian in South Australia. The Ediacaran Glaciation in Western Australia (Corcoran & George, 2001) and on other continents (Fairchild & Kennedy, 2007) also occurred during the Marinoan Epoch. Therefore, Williams et al. agree with Preiss (2000) that usage of the term “Marinoan” for glaciation in the late Marinoan is invalid.

In order to avoid confusion with terminology Williams et al. employed the term “Elatina glaciation” as was proposed originally by Mawson (1949a) following his discovery of many faceted and striated clasts that contained diamictite, which he termed “Elatina tillite”, at Elatina along Elatina Creek in the central Flinders Ranges. The Elatina Formation was defined by subsequent mapping which showed it to be a widespread lithostratigraphic unit of the glaciogenic succession in the Marinoan Series (Coats & Preiss, 1987), and it was recommended by Williams et al. that the term “Elatina glaciation“ should be adopted in South Australia and the term “Marinoan glaciation” should be discarded. Presumptions and circularity of arguments concerning the global extent, correlations and nature of those glaciations would be helped if naming Cryogenic glaciations on other continents after the principal respective glaciogenic formations.

In the description and naming of diamictite units of glaciogenic successions from the Proterozoic in the Adelaide Geosyncline (Howchin, 1928, Mawson, 1949a, b), as it has for diamictites in many other countries (Hambrey & Harland, 1981). As applied in South Australia at the present formal usage of the term “Tillite” for the lithostratigraphic terminology of Cryogenian formations that contain diamictites (Coats & Preiss, 1987; Preiss et al., 1988) reflects the early practice. It implies that all diamictites have a glaciogenic component, rather than that they were deposited from ice. It is common for faceted and striated clasts, many of extrabasinal origin, to be contained in Cryogenian diamictites in the Adelaide Geosyncline. Together with many dropstones in associated laminated sediments that are fine-grained, confirm that glacial processes were involved in their genesis.

Age of glaciation

Difficulties of correlation are compounded by the Elatina glaciation not being accurately dated. U-Pb zircon dating of igneous rocks that are associated with diamictites in the Namibia (Ghaub Formation) and on King Island, Tasmania (Cottons Breccia), respectively, that were assumed to be correlative with the Elatina glaciogenic succession, were the basis for suggested ages of glaciation of 635 ± 1.2 Ma (Hoffmann et al., 2004) and 575 ± 3 Ma (Calver et al., 2004). Cottons Breccia is, however, of uncertain genesis and it is not clear what its relationship to the Elatina formation is. An older age limit of 663 ± 4 Ma (Zhou et al., 2004) and a younger age limit of 635.2 ± 0.6 Ma was given (Condon et al., 2005) for the Nantuo glaciation in China, which they equated with the Elatina glaciation.  An estimated age of ~600 Ma for the Elatina glaciation was based on chemostratigraphy (Walter et al., 2000). The correlation that has been claimed between the glacial deposits in Namibia and South Australia (Hoffmann et al., 2004) assumes synchronous glaciations during the Cryogenian. It is still to be securely determined whether the glaciations were synchronous or diachronous (e.g., Lund et al., 2003; Zhou et al., 2004; Kendall et al., 2006) and how many glaciations in the Neoproterozoic that are represented in the geological record.

Constraints have been placed on the age of the Elatina glaciation by geochronological studies of the strata in the Adelaide Geosyncline. A detrital zircon U-Pb age of 657 ± 17 Ma was obtained for a single grain from the Marino Arkose Member in the upper part of the underlying Upalinna Subgroup (Ireland et al., 1998; Preiss, 2000). An age of 643 ± 2.4 Ma was given by Re-Os dating for black shale from the Tindelpina Shale Member (Lowermost Tapley Hill Formation) that overlies glacial deposits from the Sturtian glaciation (Kendall et al., 2006).  Furthermore, an age of ~658 Ma was given by U-Pb SHRIMP dating of a tuffaceous layer from the upper part of the Sturtian glaciogenic succession (Fanning & Link, 2006). The age of the Bunyeroo Formation from the Ediacaran may be approximated as 590-580 Ma based on Rb-Sr whole rock data and chemostratigraphy (Compston et al., 1987; Walter et al., 2000).

It is indicated by the above findings that only maximum and minimum age limits of ~640 and 580 Ma, respectively, can be applied at present to the Elatina glaciation. Accepting their age of 643 Ma for the Tindelpina Shale Member, as has been pointed out (Kendall et al., 2006), and also an age of 635 Ma for the Elatina glaciation requires high rates of sedimentation for the >4 km of interglacial strata in the Central Flinders Zone. The Elatina glaciation and the beginning of the Ediacaran Period could, alternatively, be younger than 635 Ma.

Glaciogenic deposits

Great variation is shown by the Yerelina Subgroup in the Adelaide Geosyncline. The Elatina Formation is 30-130 m thick in the western and central areas         , though in the North Flinders Zone the subgroup is thickest and more complete, >1,500 m, and the Nackara Arc, >1,000 m thick. Lithological descriptions and sections that have been measured showing the suggested facies relationships and correlations are given by (Coats & Preiss, 1987; Lemon & Gostin, 1990; Preiss, 1993; Preiss et al., 1998). Broad correlations are possible, though there is no continuity between these areas. The waxing and waning of glacial conditions are recorded by the thicker successions of the Yerelina Subgroup, whereas about the time of the main glacial stage, the deposition of the Elatina Formation began, and the glacial conditions persisted to the end of the deposition of the Elatina Formation. Therefore, in the west the margins of the basin were initially subjected to exposure, periglacial conditions and erosion, while the deeper more distal areas recorded the gradual onset of glacial conditions and accumulated a substantial thickness of glaciogenic deposits.

The Nuccaleena Formation overlies the Yerelina Subgroup, the contact being a disconformity to very low angle unconformity (Preiss, 2000; Knoll et al., 2006). The Nuccaleena Formation forms an important marker unit and is discussed below.

Yerelina Subgroup, North Flinders Zone and Nackara Arc

Fortress Hill formation (1070 m)

The onset of the glacial conditions is marked by the fortress Hill Formation. It is comprised of grey-green (weathering reddish brown) pyritic siltstone, that is finely laminated, that contains scattered pebbles and cobbles and gritty lenses. The formation is ~1,070 m thick at the type section located 1 km north of Mount Saturday. There is a progressive increase upwards in the material that has been ice-rafted, and some of the clasts are faceted and have been interpreted as drop stones. Included among the clast lithologies are granite, pink and grey quartzite, pale grey limestone, oolitic limestone, and dolostone.

Gumbowie Arkose (45-90 m)

The Gumbowie Arkose overlies the Fortress Hill Formation disconformably, at a possible sequence boundary, in the Nackara Arc. The Gumbowie Arkose is a white to pale yellow cross-bedded, coarse-grained feldspathic sandstone and arkose that contains some interbeds of pebbly sandstone, sandy siltstone, and pebbly mudstone.

Pepuarta Tillite (120-197 m)

This is a sparse diamictite that is comprised of scattered pebbles, cobbles and boulders in a grey massive, laminated, calcareous siltstone matrix. Faceted and striated boulders are up to 2.5 m across (Binks, 1971). Included in cast lithologies are grey and pink granite, granite gneiss, grey porphyry, quartz granule conglomerate, various quartzites, and vein quartz. Scattered clasts up to boulder size, are contained in the laminated siltstone facies, which is consistent with deposition from floating ice. The glacial maximum of the Yerelina Subgroup is marked by the Pepuarta Tillite and the correlative Mount Curtis Tillite.

Mount Curtis Tillite (90)  

The Fortress Hill Formation in the North Flinders Zone is overlain sharply by sandstone and conglomerate that is comparable with the Gumbowie Arkose which is suggested by Williams et al. may record a lowering of relative sea level with a sequence boundary at their base. The Mount Curtis Tillite, which is a diamictite which is comprised of sparse erratics of pebble to boulder size in grey-green, massive and laminated, sandy siltstone matrix, follows these units. Some of the clasts are faceted and striated. Clast morphologies are mostly quartzite, limestone and dolostone, and less commonly granite and porphyry. Granite boulders are up to 3 m x 8 m (Dickson et al., 1951).

Grampus quartzite (45-60 m)

The Grampus Quartzite overlies unconformably the Pepuarta Tillite; possibly at a sequence boundary that defines a 3^rd genetic sequence of the Yerelina Subgroup. It is comprised of pinkish-grey medium grained and coarse-grained feldspathic quartzite, with local thin dolomite beds.

Balparana Sandstone (130 m)

This is a pale grey and brownish grey medium-grained feldspathic sandstone, that has interbeds and lenses of calcareous siltstone and pebble conglomerate. The pebbles are mostly comprised of vein quartz, with some quartzite, siltstone and granite clasts.

Ketchowla Siltstone (≥170 m)

This is a pale grey (weathering brown), laminated to ripple cross-laminated, siltstone that is highly calcareous (Preiss, 1992). It contains scattered granules and pebbles and local concentrations of cobbles and boulders, mostly quartzite, including some up to 1 m across. The clasts were interpreted as glacial erratics that had been deposited from floating ice. The outer marine shelf deposition under glacial conditions that are generally waning is represented by the Ketchowla Siltstone.

Elatina Formation, Central Flinders Zone (30-200 m)

The lithostratigraphy and sedimentology of the Elatina Formation in the Central Flinders Zone have been described (Lemon, 1988; Lemon & Gostin, 1990). In a peripheral sink around the active and emergent Enorama Diapir (Lemon, 1985), deposits that were relatively thick (~100 m) accumulated. Glacial advance to sea level that was followed by a punctuated glacial retreat is attested to by the nature of the basal unconformity and the main sedimentary facies. The characteristic reddish hue of the Elatina Formation is the result of haematite pigmentation that is finely dispersed, which is important for palaeomagnetic studies, though it is absent from the more distal deposits of the Nackara Arc and North flinders Zone.

Basal Unconformity

The Elatina Formation overlies unconformably the Trezona Formation and the Yaltipena Formation, which is preglacial, in the Central Flinders Zone. At outcrop, little relief is shown by the sub-Elatina surface, with the exception of some possible Karst solution of the Trezona Formation. Frost shattering may be indicated by a regolith of breccia that is ≤1 m thick that overlies the limestone surface in places. The Trezona Formation, about 50 km north of Enorama Creek, is >450 m thick, has been eroded and the Elatina Formation rests on the Enorama Shale (Brenchley-Gaal, 1985).

The Yaltipena Formation (≤100 m) is a clastic red bed unit of limited extent (Lemon & Reid, 1998) that largely coarsens upwards. It displays languid and symmetrical marks, which commonly have ladder-like interference patterns, desiccation cracks and shale-clast conglomerate. A shallow water environment is indicated by these features, which imply the lowering of sea levels that heralds the approaching glacial advance.

Basal Boulder Diamictite

The top few metres of the Yaltipena Formation, that was poorly lithified, was truncated and deformed the Trezona Bore by boulder conglomerate and a massive boulder diamictite that was up to 5 m thick. The basal diamictite contains extrabasinal boulders that were forced into the top of the Yaltipena Formation, here, as well as other places. These structures that display evidence of penetrative deformation were interpreted as Type A glaciotectonics (Evans et al., 2006)    which indicate the advance of grounded ice (Lemon & Gostin, 1990; Lemon & Reid, 1998) or scouring by icebergs.

Sandstone Facies

Overlying the basal diamictite and conglomerate is about 5 m of channelled, cross-bedded, coarse-grained sandstone that had a pebble conglomerate at its base, which were interpreted as fluvial deposits (Lemon & Gostin, 1990). The sandstone grades upwards into flaser-bedded, muddy and silty sandstone, which was followed by several sandstone beds up to 1 m thick. Large ball and pillow structures are shown by the Sandstone ascribed (Lemon & Gostin, 1990) to rapid influx of the sand bodies and their collapse into the underlying muds. In the Elatina Formation the main arenaceous unit is comprised on 20-40 m of feldspathic sandstones that are white, pink to red-brown, and poorly sorted that extend beyond the peripheral sink around the Enorama Diapir to areas where they overlie directly the basal unconformity. Contained within these sandstones are subangular to subrounded quartz grains (60%), about (20%) feldspar grains (sodic plagioclase, orthoclase, and microcline) as well as lithic fragments and heavy mineral grains (Coats & Preiss, 1987). The absence of distinct bedding and the presence of thin discontinuous granule layers is a characteristic feature of the sandstones. Slumped cross-bedded sets up to 1 m thick that contain fining-upwards beds, are shown by some exposures. There are small indistinct, dewatering structures as well as rare pebbles to outsize boulders. The sandstones are generally bimodal, being marked by a dominance of coarse silt to fine sand and sand that is very coarse to granule fractions and depletion of medium and coarse sand fractions, which suggests there was aeolian winnowing in the source area. It is suggested there was subaqueous deposition in the source area (Preiss, 1987b; Lemon and Gostin, 1990).

Diamictite Facies

The extensive sandstone unit grades upwards into reddish grey, laminated siltstone and red mudstone containing rare dropstones that pass laterally into diamictite with sparse clasts. No evidence of slumping is shown by the diamictite, which together with the random scattering of large clasts suggests that deposition was from floating ice and a return to glacial conditions.

To the west of the Enorama Diapir, in the upper third of the Elatina Formation, reworking by currents of diamictite in shallow marine conditions it is indicated by the common occurrence of lag gravel layers that are capped by cross-laminated sandstones within what were otherwise massive units. It is common to find in the diamictite at the top of the Elatina formation in the Trezona Bore-Enorama Creek area, faceted and striated clasts of vesicular basalt. The whole of the Elatina Formation being deposited under conditions that were glacially influenced is attested to by the presence of a basal diamictite with glacitectonites and a glaciogenic diamictite immediately below the Nuccaleena Formation.

Deposits east of diapiric islands

East of the Elatina Formation it is ~120 m thick and displays some similarities with the formation that is described above west of the Diapiric islands. There is a thin basal conglomerate that is overlain by massive diamictite carrying boulders that are of extrabasinal origin. Succeeding this is ~50 m of pink sandstone that has indistinct bedding. Ripple cross-laminated fine to very fine-grained sandstone that has rare lonestones dominates the remaining section. Further east, in the Chambers Gorge area, the sandstone facies contain cross-bedding of large scale that thickens to ~200 m, which suggests a delta was present at the western margin of the Curnamona Province (Coats & Preiss, 1987).

Nature of clasts in diamictite facies

In the diamictites many of the clasts are faceted and striated (Mawson, 1949a; Dalgarno & Johnson, 1964; Coats & Preiss, 1987; Lemon & Gostin, 1990), with the striations typically paralleling the long axis of the clasts. Linear series of chattermarks occur at the Bulls Gap section on a faceted boulder of quartz that is 75 cm in diameter. In the diamictites, about 60% of the clasts, including those composed of basalt, dolostone and heavy-mineral banded sandstone and other lithologies were derived from the          underlying formations and diapiric islands within the sedimentary basin. At the top of the Elatina Formation the common occurrence of faceted and striated clasts of vesicular basalt in the diamictite is an indication that glacial erosion of rafts of identical basalt that were up to 1 km across were present within the diapiric islands a few kilometres to the east (Coats & Preiss, 1987; Lemon & Gostin, 1990). It is therefore suggested by Williams et al., that grounded ice persisted on the diapiric islands to the end of Elatina deposition.

About 40% of the clasts are of extrabasinal origin, and are comprised of granite gneiss, red porphyritic dacite, schist, metaquartzite, vein quartz, iron formation, gritty quartzite, as well as other sedimentary rocks. It is indicated by the wide variety of clasts that derivation was from a cratonic and metamorphic basement. Some may have been derived from the Curnamona Province, and lithologies have been identified (Lemon and Gostin, 1990) that may be matched with basement rocks in the Iron Knob area of the Gawler Craton. According to Williams et al. it is not likely that derivation was from the west of the Flinders Zone, however, as the presence on the Stuart Shelf of a preglacial palaeosols, a permafrost regolith, and it is implied by the presence of a conformably overlying periglacial-aeolian sand sheet that the cratonic region to the west was not eroded by the glaciers of the Cryogenian.

Elatina Formation, Hallett Cove area (120 m)

Pale brownish grey sandstone of the preglacial Wilmington Formation is overlain disconformably by the Reynella Siltstone Member of the Elatina Formation in the Hallett Cove area of the Fleurieu Arc near the southwestern margin of the Adelaide Geosyncline (Thomson, 1966; Dyson & von der Borch, 1986; Coats & Preiss, 1987). The Reynella Siltstone Member is well exposed in a 120 m thick coastal section of Marino Rocks, which is 2.5 km to the north of Hallett Cove and is comprised of the following facies in ascending stratigraphic order:

·       A massive dark red siltstone that contains rare granules and angular fragments of dolostone.

·       Siltstones and fine-grained sandstone including Tidalites that display herringbone cross-bedding, flaser bedding and cyclic tidal rhythmites.

·       Calcareous and dolomitic sandstone that contains angular, granular- to pebble-sized intraclasts of limestone and dolostone that are obviously derived from calcareous interbeds. Tepee-like structures and dolostone interbeds with stromatolytic laminae are also present. Acicular crystal pseudomorphs after evaporite minerals occur in some beds and in small vugs within red siltstone.

·       At the top of the section there is massive, granule-bearing siltstone.

Marino Rocks the Reynella Siltstone Member displays no conclusive evidence of glacial deposition, though it includes a siltstone with a quartzite erratic that has been interpreted as a dropstone, 7 km to the south-southwest (Coats & Preiss, 1987). The gritty, calcareous Marino Arkose Member in the Hallett Cove area was correlated with the glaciogenic deposits at Elatina 400 km to the north (Mawson, 1949a, p. 120), who regarded the southern deposits to be “the echo, so to speak” of “distant glaciation”, though he did not recognise the Reynella Siltstone Member. No evidence of glacial influence is contained in the Marino Arkose Member, and this member is the basal part of the Wilmington Formation. Williams et al. concurred with Forbes & Preiss (1987) that the Saecliff Sandstone which, in the Hallett Cove area, overlies the Reynella Siltstone Member, is a partial lateral equivalent of the Nuccaleena Formation.

Tidal Deposits

Evidence of tidal activity

In the Elatina Formation Tidalites are common, notably in a rhythmite unit at Pitchi Richi Pass that is ~10 m thick, and represents ~70 years of deposition, and not in the Reynella Siltstone Member at Marino Rocks (Williams, 1989, 1991, 1998a, 2000). Tidal deposition is indicated by the following features:

1.    Reynella Siltstone Member, herringbone crossbedding in sandstone.

2.    Reynella Siltstone Member, flaser bedding, consisting of trough cross-laminated, fine-grained sandstone with flasers of mudstone in the troughs.

3.    Cyclic rhythmites that are accreted vertically in which most cycles are 5 to 60 mm thick and are comprised of up to 16 laminae of fine-grained, feldspathic sandstone and siltstone that are normally graded and thicker near the centre of each cycle. Red mudstone bands form the boundaries of the cycles. The laminae have been interpreted as increments that represent the lunar day, and the tidal ranges (or greater tidal heights) are reflected by the thicker laminae, and faster tidal currents marking the spring phase of the fortnightly tidal cycle. The thin laminae are a reflection of the small tidal ranges, and the mudstone bands the slack water around the times of neap tides. The cycles are comparable with the neap-spring (fortnightly) cycles that have been identified in modern tidal deposits, including those on Glaciomarine settings (Smith et al., 1990; Cowan et al., 1990).

4.    Sublaminae of unequal thickness  within the rhythmites represent semidiurnal increments, best seen in the Reynella Siltstone Member. Such alternation of thick and thin laminae are present in tidal deposits of the present, and record the “diurnal inequality” of successive semidiurnal tides and tidal currents, which reflects a mixed semidiurnal-diurnal tidal pattern (de Boer et al., 1999; Dalrymple et al., 1991).

5.    The similarity of long term (years to tens of years) patterns and laminae and neap-spring cycle thickness with the pattern of modern tidal records. Regular changes in the thickness in the neap-spring cycles in the Elatina rhythmites, e.g., are comparable with variations in fortnightly tidal heights for tides at Townsville, Queensland. The “monthly inequality” of heights and ranges that result from the elliptical lunar orbit is reflected in this pattern.

Facies model for tidal rhythmites

At Pitchi Richi Pass the rhythmite unit overlies diamictite and sandstone, and is succeeded by fine- to medium-grained sandstone and diamictite that contains scattered granules and small pebbles, with rare clasts that are up to 10 cm across. At Warren Gorge 20 km to the north a correlative rhythmite is overlain by a similar diamictite that is 6 m thick (Jablonski, 1975). Very thin neap-spring cycles (0.5-3.0 m) are shown by continuous cores from 3 drill holes through the Elatina rhythmite unit (Williams, 1991) to be confined to the base of the unit and grade downwards into muddy fine-grained sandstone. Within the limits of exposures and drill holes, the neap-spring tides can be traced laterally over several hundred metres. It is indicated by rare directional structures such as ripple cross-lamination and slumped beds that there was movement to the present east, away from the presumed nearby western margin of the basin.

Deposition on prograding, distal ebb-tidal delta is involved in a facies model for the Elatina rhythmites (Williams, 1989, 1991, 2000).  Fine-grained sediment is entrained in such a setting by ebb-tidal currents and is transported mainly in suspension by periodic turbid currents and jets (özsoy, 1986) to deep water offshore by the main ebb channel. Such jets transform to plumes with increasing distance of transport (Powell, 1990) and suspended sediments settle out to form normally graded laminae in distal settings. The potential sediment load of ebb-tidal currents and the effectiveness as an agent of sediment entrainment and deposition of the tide are related directly to tidal range (or maximum tidal height) and speed of the current (FitzGerald & Nummedal, 1983; Boothroyd, 1985). The relative thickness of successive laminae and neap-spring cycles, however provide a proxy tidal record.

The adjacent hinterland on the Stuart Shelf was a periglacial region with minimal fluvial activity interfering with periodic tidal processes, during the deposition of the Elatina rhythmites. Tidal estuaries and inlets that favoured the deposition of cyclic tidal rhythmites in nearshore settings and on distal tidal deltas may have been produced by the drowning of coastal areas through rising sea levels as was discussed (Williams, 2000). At Pitchi Richi Pass the rhythmite unit that was underlain and succeeded by the glaciogenic deposits was interpreted as marking a high stand of relative sea level during temporary glacial retreat. That the glaciation occurred close to sea level is confirmed by the structural proximity of Tidalites, and that glaciation occurred close to sea level is confirmed by diamictite.

At Marino Rocks the Reynella Siltstone Member contains some neap-spring cycles that are relatively thick and a variety of sedimentary structures which suggest a setting that is more variable, and at times a more proximal setting than that that has been envisaged for the Elatina tidal rhythmites at Pichi Richi Pass. The tidalites in the Reynella Siltstone Member may have accumulated on upper and lower delta slopes as well as in estuarine settings like those in the Bay of Fundy, Canada (Dalrymple et al., 1991).

Slump structures and ripple marks

At Pichi Richi Pass and Warren Gorge the Elatina rhythmites display slightly sinuous, discontinuous, near cuspate folds that are near-parallel and are spaced 13-50 cm apart and with amplitudes of 3-5 cm (Williams, 1996) that are draped by symmetrical and interference ripple marks that are generated by waves. Some cuspate folds have crests that are truncated or show delicate scouring, which indicates that the folds were present during deposition. At Warren Gorge the ripple marks are more common than at Pichi Richi Pass, which implies there was a more distal, deep water setting for Pichi Richi Pass.

The folds were interpreted as gravity slides on a tidal delta that was triggered by stresses generated by storm waves (Williams, 1996). Similar cuspate folds, which were also interpreted as gravity slides, that underlie deposits that were generated by wave action are displayed by deposits in Newfoundland that date to the Neoproterozoic which were generated by wave action (Myrow & Hiscott, 1991). On the lower slopes of submarine fans, fan deltas and deltas (e.g., Prior & Coleman, 1980; Prior & Bornhold, 1989), were marked by gravity slides that were slightly sinuous, discontinuous and en echelon compressional ridge    s of dimensions that were similar to the soft-sediment folds at Elatina are ubiquitous.

Quantitative analysis of tidal rhythmites

Time series analysis of stratigraphic logs of lamina and neap-spring cycle thickness that were obtained from 3 drill cores from the Elatina rhythmite unit, that were supplemented by data from the Reynella Siltstone Member rhythmites, has provided the most complete palaeotidal set that is available for the Precambrian (Williams, 1989, 1991, 1998a,  2000). In late Cryogenian times it is indicated by the rhythmites that there were 13.1 ± 0.1 lunar months/year, 400 ± 7 solar days/year and 21.9 ± 0.4 h/solar day, with a mean Earth-Moon distance 96.5 ± 0.5% of the present distance.

The non-tidal, annual oscillation of sea level is also recorded by the Elatina rhythmite unit. This oscillation is the result of the interaction of the physical processes that include changes in the temperature of the water, pressure changes of winds and atmosphere (Komar & Enfield, 1987). The annual oscillation of sea level in low and moderate latitudes mostly results from seasonal changes in the heat content of the sea, with winds being a minor importance (Roden, 1963; Pattullo, 1996; Wunsch, 1972; Mellor & Ezer, 1995). It is indicated by the Record of the annual oscillation in the Elatina data that the late Cryogenian global ocean was not frozen over during the interstadial of the Elatina glaciation, because if there was complete ice cover it would have isolated the sea from seasonal changes of temperature and winds (Williams, 2004; Williams & Schmidt, 2004).

Permafrost regolith

During the Neoproterozoic, the arenaceous Pandurra Formation that dates to the Mesoproterozoic (1,424 ± 51 Ma; Fanning et al., 1983) formed a broad structural high, the Pernatty Upwarp, in the central part of the Stuart Shelf. On the irregular surface of silicified Pandurra Formation (quartzite), and locally on the Tapley Hill Formation, from the Cryogenian, the Cattle Grid Breccia developed (Williams & Tonkin, 1985; Williams, 1986; Coats & Preiss, 1987). Silicification of the Pandurra Formation predates the Elatina glaciation, and as it is not evident in the poorly lithified sandstone that is encountered in deep drill holes appears to be a superficial effect. It is implied by this probable palaeosol, though it has been modified by cryogenic processes during the Elatina glaciation, that the cratonic Stuart Shelf was not eroded by the glaciers of the late Cryogenian. In the now abandoned and partially filled Cattle Grid open pit, as well as other pits near Mt Gunson, where it hosts a stratabound copper deposit the Cattle Grid Breccia is exposed (Tonkin & Creelman, 1990). A periglacial origin for the Cattle Grid Breccia and structures that were associated with it was first proposed (Williams, 1981) during an inspection of the Cattle Grid mine in January 1981, after which Williams continued to study the structures that were revealed by mining until the late 1980s when excavation ended.

The Cattle Grid Breccia is overlain conformably by the Whyalla Sandstone, which is capped by the Nuccaleena Formation. It is demonstrated by the stratigraphic relationships that the Cattle Grid Breccia and the Whyalla Sandstone both formed during the Elatina glaciation (Coats, 1981; Coats & Preiss, 1987; Preiss, 1993; Preiss et al., 1998). The post-glacial marine transgression that submerged most of the Stuart Shelf (Forbes & Preiss, 1987) is marked by the Nuccaleena Formation, which indicates that the Cattle Grid Breccia and the Whyalla Sandstone formed near sea level.

Brecciated bedrock

The Cattle Grid Breccia is 0.5 to 20 m thick, and averages 5 m. The top ≤2 m is comprised of angular to subrounded quartzite clasts mostly 2-10 cm across, with some blocks up to 40 cm, which are oriented randomly in a sandy matrix. In places there is crude horizontal bedding. Channels form that are up to 10 m wide and 2-3 m deep, which contain     fragments and well-rounded quartzite pebbles, that occur locally at the contact with the overlying Whyalla Sandstone. The sedimentary breccia passes downwards into in situ breccia in which there are tabular clasts of cobble to boulder size that are roughly horizontally oriented, marking the original bedding of the Pandurra Formation, and clasts which are adjacent have matched sides. The in situ breccia has little or no sandstone matrix and passes downwards into the solid bedrock.

It is common to find anticlinal structures 0.5 to 4 m high and 0.3 to 4.0 m wide in the in situ breccia (Williams & Tonkin, 1985). There are near vertical axial planes in most anticlines, though there are some structures that are asymmetrical, and overturned in extreme cases. Associated with some of the asymmetrical anticlines are moderate- to high-angle reverse faults. Locally, there are diapiric-like mounds that are truncated at the contact with the Whyalla Sandstone.

The Cattle Grid Breccia was compared with block fields (Williams & Tonkin, 1985), which are level or sloping gently (<10^o) surfaces that are covered with angular blocks, that are distributed widely in polar regions (Embleton & King, 1975; White, 1976; Washburn, 1980). The block fields are generally agreed to have developed by frost shattering in situ, of adjacent, well-jointed bedrock and the expansion that results exerts pressure in all directions. According to Williams et al. both horizontal and vertical expansion of the breccia profile that is developing and the frost heaving and frost thrusting that results is attested to by the anticlines and reverse faults within the Cattle Grid Breccia.

The in situ breccia is composed of tabular fragments that are near-horizontal of bedrock that extends to depths of a few metres to 11+ m below the surface of the ground, occurs in southeast England, Spitzbergen, and Melville Island in the Canadian Arctic (Murton, 1996). In southeast England small open, folds are exhibited which are capped by an involuted layer 0.5 – 2.0 m thick comprised of clayey beds and chalk diamicton with sub-rounded to rounded clasts of chalk. These features are similar to those of the Cattle Grid Breccia. Continuous permafrost more than 500 m thick with a mean annual ground temperature of about -17^oC underlies the breccia on Melville Island.

Primary sand-wedge polygons

At the top of the Cattle Grid Breccia V-shaped sandstone wedges that are downwards pointing, 15 cm – 4 M in apparent maximum width, and 3.5 – 3.0  m deep are common (Williams & Tonkin, 1985; Williams, 1986, 1994; Williams  Schmidt, 2004). The wedges consist of fine- to very coarse-grained sandstone displaying a divergent or fan-shaped lamination with laminae near the margins of the wedge paralleling the breccia contact and laminae near the centres of the wedge is dipping steeply. In the breccia the relict bedding is turned upwards adjacent to most wedges. The surface of the Cattle Grid Breccia has grooves that are several metres wide and up to 1 m deep that overlie large wedges which outline polygons ~1 – 30 m across has been revealed by mine workings. There are 2 generations of wedges that were recognised.

At the present sand wedge polygons that are closely comparable with those in the Cattle Grid Breccia and Whyalla Sandstone are forming in the arid Arctic and dry valleys of the Antarctic (Péwé, 1959; Washburn, 1980). In the upper part of the permafrost such wedges develop from vertical, thermal contraction cracks that are ~1–5 mm wide and several metres deep that outline polygons ~10–30 m across. Following rapid temperature drops during repeated severe winters the contraction cracks form. In polar periglacial regions seasonal changes of temperature may reach depths of more than 15 m, whereas the upper permafrost diurnal temperature fluctuations may affect the upper permafrost only to a depth of ≤1 m (Embleton & King, 1975). It has been confirmed by observations over several decades that under climates that are strongly seasonal the wedges are forming actively at high latitudes. The contraction cracks are filled by windblown sand in arid periglacial areas, and the primary sand wedges that result show lamination that is deeply dipping. Such wedges can be distinguished from ice wedges, which form in periglacial areas that are relatively humid where water freezes in the cracks which may then be replaced on thawing by sediment that commonly exhibits horizontal layering. During summer expansion lateral pressure causes upturning of permafrost adjacent to the wedges.

At Mount Gunson it is indicated by the geometry, dimensions and internal fabric of wedges that they are primarily sand wedges. Further features that are shared by the sand wedges in the Cattle Grid Breccia and in Antarctica (Péwé, 1959; Black, 1973) include their development in bedrock rubble that was produced by the action of frost and the presence of vertical wedge-in-wedge. Structures that are interpreted as periglacial sand wedges or fossil-ice wedges are known to occur on other continents (Deynoux, 1982; Spencer, 1085; Williams, 1986; Moncrieff & Hambrey, 1990).

Other periglacial forms

The Cattle Grid Breccia and the Whyalla Sandstone are both involved in the following structures (Williams & Tonkin, 1985; Williams, 1986):

·       In the top of the Cattle Grid Breccia sags of the Whyalla Sandstone form that are as much as 2 m deep.

·       Rare diapiric tongues of breccia penetrate upwards into the lowermost few metres of the Whyalla Sandstone.

·       Tabular blocks of quartzite that are up to 1 m long that are inclined steeply within the Whyalla Sandstone 1-3 m above the top of the Cattle Grid Breccia.

These structures were interpreted as periglacial involutions, periglacial injections and frost-hardened blocks, respectively Embleton & King, 1975; Washburn, 1980). It is suggested by the breccia tongues that liquefaction of the topmost breccia as a result of permafrost thawing in the Cattle Grid Breccia. In the Whyalla Sandstone probable water-escape structures include sinuous trails that are near vertical of disrupted bedding above sand wedges, and vertical columns of sandstone that are well cemented that are 30-40 cm in diameter. In the basal Whyalla Sandstone the presence of these structures, together with primary sand wedges, shows that periglacial Cryogenic activity continued for some time following the burial of the Cattle Grid Breccia.

The presence of liquid water and liquefied sediment at the top of the permafrost horizon is implied by many features of the topmost Cattle Grid Breccia and the lowermost Whyalla Sandstone (Williams & Tonkin 1985; Williams, 1998b):

·       Reworked breccia horizon.

·       Channel forms that contain well-rounded pebbles.

·       Thermokarst microtopography.

·       Periglacial involutions.

·       Diapiric breccia injections.

·       Water-escape structures.

An active layer at the   top of the permafrost is attested to by these features. The active layer was defined (Washburn, p. 57) as the “layer of ground above permafrost which thaws in summer and refreezes in

winter”.

Periglacial aeolianite

The Whyalla Sandstone, which is flat lying and 165 m thick (Coates & Preiss, 1987; Preiss, 1993; Williams, 1998b) covers 25,000 km² in outcrop and subcrop on the Stuart Shelf. The formation consists mostly of medium- to very coarse-grained quartzose sandstone that is moderately well assorted, bimodal, with rounded, frosted quartz grains that are 1-2 mm across and subrounded to subangular, fine (50 -200 μm) grains of quartz, feldspar, and lithic fragments. Bands of rounded granules, and small pebbles of quartz and lithic fragments, that occur locally. Commonly, the sandstone is intercalated with gritty or granule bearing siltstone, and in places at contacts with underlying or overlying formations and in the east the transition to the Elatina Formation, there are argillaceous beds. The sandstone weathers yellowish grey and light grey. Reddish brown and pale greyish red to pale red colours predominate below depths of 60 m, which results from ultrafine haematitic pigment in the matrix and haematitic staining on detrital grains. In the formation, regional fining is towards the southeast of the present.

Sedimentary structures

The Whyalla Sandstone displays a variety of structures (Williams, 1998b):

·       The principal stratification type is represented by low angle (≤15^o) strata.

·       Large-scale cross-bedded sets that are up to 7 m thick occur mainly in the central area of the sandstone body. Cross-stratal inclinations are mostly up to 20-23^o and rarely up to 30^o, with most being inclined towards what is the southeast of the present.

·       Inversely graded subcritically climbing translatent strata, which are diagnostic of aeolian deposition (Hunter, 1977). Grainfall laminae and grainflow deposits and pin-stripe lamination.

·       2 generations of primary sand wedges at the base of the formation.

·       Drop involutions, periglacial injections and frost-heaved blocks that involve the Cattle Grid Breccia and the Whyalla Sandstone.

·       Water-escape structures, thermokarst microtopography, and small-scale, soft sediment tensional faults which suggest collapse caused by snow melt.

Depositional environment

An origin that is predominantly aeolian is indicated by the primary sedimentary structures, and it is implied by the periglacial structures that the climate was cold and arid. The preponderance of low-angle strata that display locally periglacial cryogenic structures that accord with a periglacial sand-sheet environment (Koster, 1988; Dijkmans & Törnqvist, 1991), with most dune forms occurring near the centre of the sand sheet.

Cross-bedding attitudes and the regional fining direction indicated by palaeomagnetic data for the Elatina Formation that the winds were palaeowesterly to palaeonorthwesterly surface winds near the palaeoequator during the Elatina Glaciation. It is implied by the subdued topography and the absence of glaciation in the late Cryogenian of the Gawler Craton in the palaeowest and palaeonorthwest (Coats & Preiss, 1987; Preiss, 1983) that the wind direction, rather than the katabatic winds blowing off an ice cap or ice sheet has been recorded.

The contact between the Yerelina Subgroup and the Nuccaleena Formation was interpreted as a sequence boundary that was marked by a disconformity to unconformity of very low angle (Preiss, 2000). Throughout the Central Flinders Zone, the Elatina-Nuccaleena contact varies from generally sharp to gradational over a thickness of ~0.5. In Enorama Creek, at the Ediacaran GSSP the gradation is from silty dolostone downwards into reddish, fine-grained sandstone that is ~0.2-0.3 m thick, which has an irregular, probably erosional base over reddish siltstone ~0.3 m thick. The siltstone is poor in clasts, though grades down into diamictite of the Elatina Formation. This interface between the sandstone and the siltstone may be a sequence boundary that marks the base of the Nuccaleena Formation, with the sandstone possibly being a local manifestation of the Seacliff Sandstone.

Cap carbonate

Contact with the Yerelina Subgroup

A persistent marker through the Adelaide Geosyncline region is the Nuccaleena Formation, and particularly its dolostone member (Forbes & Preiss, 1987) that represents the cap carbonate of the Yerelina Subgroup. The contact between the Nuccaleena Formation and the Yerelina Subgroup is interpreted as a sequence boundary that is marked by a disconformity to a very low angle unconformity (Preiss, 2000). The Elatina-Nuccaleena contact varies from generally sharp to what appears to be gradational over a thickness of ~0.5 m. At the Ediacaran GSSP the gradation if from silty dolostone downwards into reddish, fine-grained sandstone ~0.2-0.3 m thick, which has an irregular base that is probably erosional, over reddish siltstone ~0.3 m thick. The siltstone is poor in clasts, but grades down into diamictite of the Elatina Formation. Williams et al. suggest the interface between the sandstone and the siltstone may be a sequence that marks the base of the Nuccaleena Formation, with the sandstone possibly being a local manifestation of the Seacliff Sandstone.

To the east, at Chambers Gorge, the Elatina-Nuccaleena contact is an unconformity that is of very low angle. At this sequence boundary erosion is very pronounced in the North Flinders Zone where several different stratigraphic units of the Yerelina Subgroup are overlain by the dolostone member (Ambrose, 1973; Preiss, 2000). The primary stratigraphic evidence is consistent with isostatic rebound of the Curnamona Province in the east and possibly an appreciable delay between the end of the Elatina glaciation and the following marine transgression, though the erosional contact seems to be a peneplained surface that lacked major relief (Knoll et al., 2006).

Sedimentary facies

The Nuccaleena Formation, at its type section near Mount Saturday, is comprised of 10 m of cream, well-bedded dolostone that is overlain by up to 60 m of purple shale that contains rare dolostone bands. The Nuccaleena Formation, however, usually lacks the upper shale member. The range in thickness of the lower, dolostone member is from several metres along the western margin of the Adelaide Geosyncline, and on the Stuart Shelf, 5-17 m in the Central Flinders Zone and the Curnamona Province where it forms part of the cratonic cover succession. Along the southern Torrens Hinge Zone, in the Hallett Cove area, and in some areas of the Nackara Arc, laminated, pale bluff dolostone that is characteristic of the Nuccaleena Formation intertongues with the Seacliff Sandstone, which is up to several hundred metres thick (Forbes & Preiss, 1987; Dyson, 1992; Preiss, 2000). The dolomitic member that is characteristic of the Nuccaleena Formation typically commences with a pink, silty dolostone that passes upwards into buff, well laminated to massive, micritic dolostone, which in places, forms linear tepee-like structures that are up to 1.5 m high and are associated with growth faults and sheet veins (Plummer, 1979; Gammon et al., 2005). In turn, this dolostone grades upwards through silty dolostone, flaggy lenticular dolostone that is interbedded with red shale, to red shale (blue Shale in the eastern Nackara Arc). The top of the Nuccaleena Formation is defined by the last appearance of dolostone (Lemon & Gostin, 1990). The uppermost shale, which is believed to possibly record the point of maximum flooding (Preiss, 200), is succeeded by the Brachina Formation.

Deposition and diagenesis

The sandstone above the Elatina Formation, and the dolostone of the Nuccaleena Formation were interpreted as a transgressive tract (Gammon et al., 2005). It is believed that the primary carbonate was calcite that was deposited in a marine shelf setting up to several hundred metres deep (Knoll et al., 2006). Thin beds of dolomicrite that has red brown shale partings and current ripples may be turbidites (Kennedy, 1996). Contrary to the impression that was conveyed by others (Halverson et al., 2004; Knoll et al., 2006), however, none of those features are present in the Nuccaleena Formation.

Based on δ¹³C values in 10 different stratigraphic sections throughout the Adelaide Geosyncline, the distinctive carbon isotopic profile of the Nuccaleena Formation decreases upwards from -1 to -2.5‰ at its base to -2 to -3.5‰ at its top (McKirdy et al., 2001, and in preparation). The profile in its shape and absolute values is broadly similar to profiles of carbonates from elsewhere that date to Early Ediacaran, notwithstanding its uncertain significance of its lateral variability across the basin (1.5‰) that is yet to be determined more certainly. Knoll et al. (2006, p. 23) was led by this attribute to interpret the Nuccaleena Dolostone Member as representing a “short-lived chemical oceanic event accompanying Marinoan deglaciation and the sea level rise, and to be of global extent”. Recent work has shown, however, that more than 90% of the Nuccaleena Cap Carbonate is comprised of dolomicrospar with a radical geochemical and isotopic zonation that could have formed only via early diagenetic organogenic dolomitisation (McKirdy et al., 2995; Gammon, 2006). It does not appear likely that the Nuccaleena Formation records the carbon-isotopic composition of the post-glacial ocean.

As was first noted (lemon, 1988) and (Lemon & Gostin, 1990), the unusually large tepee-like structures in the Nuccaleena Formation do not contain many of the characteristic features of peritidal tepees (e.g. Assereto & Kendall, 1977) and evidently formed at a time when the host sediment was submerged entirely. It was indicated by recent study of such structures (Smith, 2001; Gammon et al., 2005) that they developed as much as 7 m below the sediment-water interface, and that they as well as their associated sheet veins were the result of fluid overpressure that was generated by the expansive crystallisation of dolomicrite during early diagenesis.

Palaeoenvironment of the Elatina glaciation

Palaeogeography

A wide range of depositional environments, which together cover an area of ~200,000 km², that mark glaciation in the late Cryogenian that occurs immediately before that Ediacaran:

·       In the west, a permafrost regolith that was blanketed by a periglacial aeolian sand sheet that formed on the Stuart Shelf at the eastern margin of the Gawler Craton. There was little or no glaciation in this cratonic region, possibly because of its aridity and low relief, and wide areas remained free of ice. Palaeowesterly to palaeonorthwesterly winds prevailed and a climate that was frigid and strongly seasonal prevailed near sea level.

·       To the east, in the Central Flinders Zone, fluvial, deltaic and inner marine shelf settings of the Elatina Formation prevailed, with deposits that were glaciomarine and local glaciers on emergent diapiric islands extending to sea level.

·       In the North Flinders Zone, the thick (>1,500 m) glaciogenic succession of the Yerelina Subgroup accumulated in an intracratonic sub basin in an outer marine-shelf setting. To the east detritus was derived from glacial erosion of crystalline basement, where basement rocks of the Curnamona Province may have been exposed (Preiss, 1987b). To the north, the Muloorina Ridge, that is now buried, may also have been a source area.

·       To the southeast, in the Nackara Arc, an outer marine shelf formed by subsidence along a passive margin and accumulated the glaciogenic Yerelina Subgroup (>1,000 m), which was derived from glacial erosion of crystalline basement to the northeast and east.

Major rifting or compressional tectonics, or any bordering terrain that was highly elevated, have left no record. Only subdued extensional tectonics, with tilting and down-to-basin faulting involved in the northern sub-basin and southeastern outer marine shelf (Plummer & Gostin, 1976; Preiss, 1993). Because outer marine settings that were bordered by cratons mark these regions, ocean currents and waves had little influence. Under the protection of an ice shelf and icebergs, during maximum glaciations, wave activity was reduced further, which allowed the accumulation of massive and laminated calcareous diamictites of the Mount Curtiss Tillite and Pepuarta Tillite under anoxic conditions. The presence of wet-bedded glaciers with rainout of fine-grained sediments and ice rafted debris, attested to widespread deposition of such glaciomarine diamictites that contained extrabasinal erratics and the thick mudstone and siltstone deposits that contained scattered dropstones. The Fortress Hill Formation and Ketchowla Siltstone are representative of outer marine shelf deposition under glacial conditions that were waxing and waning, respectively, with the presence of icebergs indicated by large glacial erratics in the Ketchowla Siltstone.

Grounded ice or icebergs deformed the preglacial Yaltipena Formation, that were poorly lithified, and depositional diamictites at the base of the Elatina Formation. Fluctuations of ice margins and sea level are implied by the deposition of fluvial, deltaic and inner marine shelf sandstones, and tidalites that originated in estuarine and tidal-delta environments, then alternated with glaciomarine diamictites that were deposited from floating ice.

Glacial conditions persisted, and within the basin, grounded ice on diapiric islands, ice rafting and diamictite deposition continuing to the top of the Elatina Formation.

It is indicated by the widespread occurrence and persistence of ripple marks that were generated by waves in the Elatina tidal rhythmites, that there was a continuously open sea over several decades, the signal of the annual oscillation of the sea level in these rhythmites implies the presence of extensive open seas for at least the ~70-year interval of rhythmites deposition (Williams, 2004; Williams & Schmidt, 2004). Importantly, it is indicated by the widespread occurrence of diamictites at the top of the Elatina Formation that glacial conditions returned after rhythmite deposition. Further evidence that the sea was not frozen over during the deposition of the Yerelina Subgroup, which permitted the rainout of fine-grained sediment and ice rafted debris, is provided by the extensive glaciomarine diamictites with clasts of extrabasinal origin, and thick formation of laminated siltstone with dropstones.

According to Williams et al. the various igneous and high grade metamorphic erratics in the glaciogenic succession were derived from adjoining cratonic areas, had probably been transported by glaciers or ice streams that emanated from ice caps that bordered the eastern margin of the Adelaide Geosyncline. That the extensive cratonic region to the west of the Adelaide Geosyncline was essentially free of ice is implied by the presence of silcrete-type palaeosols, a permafrost regolith and an aeolian sand sheet on the Stuart Shelf and the lack of evidence for glaciation of the Gawler Craton during the late Cryogenian. It is suggested for most of the extrabasinal clasts, there was a provenance to the east of the present day, which includes the Curnamona Province. Their source terrain has not yet been confirmed, however, and it is implied by some of the sparsely pebbly nature of the diamictites that this facies was deposited in settings that were relatively distant.

Palaeoclimate

The Periglacial structures are among the most reliable of palaeoclimate indicators, as they formed through physical weathering processes and, therefore, their interpretation avoids uncertainties that arise from the former nature of the atmosphere and biosphere and later diagenetic alteration. Quantitative information on palaeoclimate that is provided by periglacial structures include indications of mean annual air temperature (MAAT), seasonal temperature range and mean annual precipitation (Washburn, 1980; Karte, 1983), as well as longer-term oscillatilons of climate, Primary sand wedges that have dimensions and structure that are comparable with Mount Gunson of the present are confined to arid periglacial regions that are marked by a climate that is strongly seasonal (Péwé,  1959, Washburn, 1980; Karte, 1983). Such primary wedges are indicative of a MAAT of <-12 to <-20^oC and a mean monthly air temperature (MMAT) that ranged from <-35^oC in mid-winter to +4^oC in mid-summer (Karte, 1983). The Cattle Grid Breccia that was 20 m thick attests to frost shattering at the top of the permafrost zone and a frigid climate.

A mean horizontal growth rate of ~1mm/year that was determined for certain modern sand wedges and ice wedges (Black, 1973; Washburn, 1980) suggests that at Mount Gunson the sand wedges took up to ~4 ky to form (Williams & Tonkin, 1985). In the Cattle Grid Breccia and the lowermost Whyalla Sandstone the development of at least 4 generations of sand wedges, punctuated by intervals that were marked by permafrost deformation or aeolian sand deposition which suggests that there were climate cycles of several thousand years duration that involved large fluctuations of MAAT, that ranged from -12 to -20^oC or lower up to >0^oC.

In modern polar regions the shallow (≤1 m) influence of diurnal temperature changes on permafrost (Embleton & King, 1975) and at high elevations on the equator of the Pleistocene the lack of periglacial wedges, where MAAT was well below 0^oC for several millennia and fluctuations of just a few ^oC were mainly diurnal (Williams & Schmidt, 2004) militate against the claim (Maloof et al., 2002), that was based on numerical models that diurnal temperature fluctuations produced the 3 m deep sand wedges at Mount Gunson during the late Cryogenian. It was conceded (Maloof et al., 2002) that experiments would be required to ascertain if the strain that was caused by the variations of diurnal temperature were squandered in microcracking. Moreover, the length of day of 21.9 h in the late Cryogenian that derived from the Elatina-Reynella palaeotidal data set indicates that the diurnal cycle was shorter and therefore less effective as a geocryologic agent than the cycle of the present. An actualistic approach, which was based on more than a century of observations and research on periglacial geomorphology and processes argue strongly that the primary sand wedges as well as the associated suite of large-scale periglacial structures at Mount Gunson record a periglacial climate that was strongly seasonal.

In summary, by analogy with the climate significance applied to periglacial structures that are comparable in modern polar regions (Washburn, 1980; Karte, 1983), at Mount Gunson the suite of periglacial structures implies that the climate in South Australia in the late Cryogenian displayed the following features:

·       MAAT near sea level of -12 to -20^oC or lower.

·       Large seasonal temperature range from -35^oC or lower in mid-winter up to +4^oC in mid-summer (MMAT range of up to ~40^oC).

·       Aridity (<100 mm mean annual precipitation) and windiness.

·       Summer temperatures above freezing to produce an active layer at the top of the permafrost.

·       Fluctuations in MAAT that rage from as low as -20^oC or lower up to >0^oC on a 10³ year timescale.

This frigid climate, that is strongly seasonal prevailed at sea level at the margin of a marine basin within 10⁰ of the palaeoequator. The implications that MAAT rose above freezing, annually as well as through long-term temperature changes is consistent with other evidence that      long-lived open seas existed during the Elatina Glaciation.

Palaeolatitude

Currently, the most robust data that indicates a low palaeolatitude during the Cryogenian glaciation comes from detailed palaeomagnetic studies of red beds from the Elatina Formation in the Adelaide Geosyncline and equivalent strata on the Stuart Shelf (Embleton & Williams, 1986; Schmidt et al., 1991; Schmidt & Williams, 1995; Sohl et al., 1999; Williams & Schmidt, 2004). The early timing of magnetic remanence for the Elatina Formation is confirmed by fold tests that were executed on small folds of demonstrable soft-sediment origin (Williams, 1996), which yielded positive results at a high level of confidence (Sumner et al., 1987; Schmidt et al., 1991; Schmidt & Williams, 1995) and by the presence of sequential magnetic reversals (Schmidt & Williams, 1995; Sohl et al., 1999). The remanence was acquired during deposition through detrital remanent magnetisation (DRM) carried by silt-sized grains of high-temperature titano-haematite as well as soon after deposition during early diagenesis through chemical remanent magnetisation (CRM) carried by ultrafine haematite in the matrix and as coatings on grains. All the palaeomagnetic reliability criteria of Van der Voo (1990) are satisfied by the data for the Elatina Formation.

It is implied by the observation of mixed polarities within the specimens from Elatina that the CRM was acquired at the time scale of magnetic reversals; therefore, it is appropriate to employ sample directions to provide an overall formation mean direction (Schmidt & Williams, 1995). For the Elatina Formation, combined data (Schmidt & Williams, 1995; Sohl et al., 1999; Geological Society of America data Repository item) for 205 samples yielded a mean declination D = 208.3^o and a mean inclination of ­I = -12.9^o (α₉₅ = 4.2^o ) and a palaeolatitude  of 6.5 ± 2.2^o (P.W. Schmidt, personal communication, 2006). The palaeomagnetic remanence of Elatina has inclinations that are persistently low for different lithologies and for samples that were acquired from Basinal and cratonic tectonic settings. It is also shown by the remanence that anisotropy of magnetic susceptibility (AMS) of 4% for 44 samples that carry CRM and 2.1% of 21 samples that carry DRM (P.W. Schmidt & G.E, Williams, unpublished data) that indicates slight magnetic foliation (Enkin et al., 2003). It is implied by these findings that there is minimal shallowing of inclination. It was concluded that in the Adelaide Geosyncline the Elatina glaciation took place at a palaeolatitude of <10^o. This is consistent with the palaeolatitude of 8.4 + 6.2/-5.7^o that was determined for the Yaltipena Formation that is immediately periglacial (Sohl et al., 1999). Palaeomagnetic data that is available accord with the geomagnetic field approximating a geocentric axial dipole (GAD) of the Earth during the Proterozoic (Schmidt, 2001; Williams & Schmidt, 2004; McElhinny, 2004) concluded that the GAD model is a good first-order approximation probably for the whole of geological time. Therefore, the low palaeolatitudes that were determined for the Elatina Formation, as well as other glaciogenic successions in Australia dating to the Cryogenian (Pisarevsky et al., 2001, 2007) and other continents (e.g. Park, 1997; Evans, 2000; Kempf et al., 2000; Kilner et al., 2005; Weil et al., 2006) and the lack of palaeomagnetic data for glaciogenic deposits that indicate high palaeolatitude glaciation in the Cryogenian, present a major geophysical and geological enigma.

Discussion and conclusions

The Elatina glaciation occurred in the Adelaide Geosyncline-Stuart Shelf region of South Australia in the Cryogenian at a time that has been undetermined between about 640 Ma and 580 Ma. A wide range of facies displayed by the Yerelina Subgroup:

·       Permafrost regolith with many large scale cryogenic structures;

·       Periglacial aeolianite;

·       Littoral and neritic deposits that included tidalites and evaporites;

·       Basal diamictite;

·       Fluvial, deltaic and inner marine shelf sandstones;

·       Glaciomarine diamictites; and

·       Outer marine shelf mudstone and siltstone with drop stones that were ice rafted.

The Yerelina Subgroup with its variety of facies has much in common with the glaciogenic succession from the Cryogenian in the fjord region of East Greenland (Moncrieff & Hambrey, 1990).

Insight into the climate and the extent of sea ice during the glaciation of the late Cryogenian is provided by the various facies of the Yerelina Subgroup. An in situ cold climate near sea level (mean annual air temperature -12 to -20^oC or lower in mid-winter up to +4^oC in mid-summer) is implied by the cryogenic structures and changes in the longer-term (10³ years) in mean annual air temperature. The advance of grounded ice or icebergs is indicated by the presence of glacitectonites displaying evidence of penetrative deformation of preglacial deposits that are poorly lithified beneath the Elatina Formation. Ripple marks that were wave-generated and the signal of the annual sea level oscillation that are displayed by the Elatina rhythmites, and the widespread occurrence of glaciomarine diamictites and thick mudstone/siltstone with drop stone facies of the Yerelina Subgroup indicates the presence of extensive and persistent open seas during the Elatina glaciation. Glacial deposition continued to the top of the subgroup.

It is indicated by high quality palaeomagnetic data for red beds from the Elatina Formation that this frigid, strongly seasonal glacial-periglacial climate, in which there was grounded sea ice near sea level, occurred within 10^o of the palaeoequator. The recognition of such a non-actualistic glacial climate in South Australia in the late Cryogenian has been a catalyst for Cryogenian        glaciogenic successions around the world and has helped to stimulate the development of global models, accompanied by a lot of vigorous debate, to explain the underlying cause of such a paradoxical climate (e.g. Williams, 1993; Hoffman et al., 1998; Williams & Schmidt, 2004; Hoffman & Schrag, 2002; Jenkins et al., 2004; Fairchild & Kennedy, 2007).

The Nuccaleena Formation from the Early Ediacaran, which overlies the Yerelina Subgroup disconformably to unconformably, represents the cap carbonate that was deposited during the marine transgression in post-glacial times. The Nuccaleena Formation probably doesn’t record the carbon-isotope composition of the post-glacial ocean as >90% of its carbonate is comprised of dolomicrospar that formed through early diagenetic organogenic dolomitisation.

According to Williams et al. research on Elatina glaciation deposits in the future should aim at determining an accurate age for the Yerelina Subgroup in South Australia, preferably by identifying volcaniclastic layers, if present, that have yet to be discovered, in order to test correlations that have been proposed with other glaciogenic successions. Detailed study is also required on detrital zircons and on the nature and provenances of the extrabasinal clasts in order to illuminate the palaeogeography of the Elatina glaciation within Australia and adjoining continents. Such work will elucidate the temporal relationship of the Elatina glaciation to the glaciation of the late Cryogenian elsewhere and thereby establish whether it is part of a single, near-global event, or alternatively one of a series of glaciations that were more local that affected different parts of the Earth at slightly different times, some or all of which may have been recorded in South Australia only as boundaries of sequences.

Williams, G. E., et al. (2008). "The Elatina glaciation, late Cryogenian (Marinoan Epoch), South Australia: Sedimentary facies and palaeoenvironments." Precambrian Research 163(3): 307-331.