Neoproterozoic sedimentary basins with glacigenic deposits of the East Greenland Caledonides
Published:January 01, 2008
- PDF LinkChapter PDF
Martin Sønderholm, Kasper S. Frederiksen, M. Paul Smith, Henrik Tirsgaard, 2008. "Neoproterozoic sedimentary basins with glacigenic deposits of the East Greenland Caledonides", The Greenland Caledonides: Evolution of the Northeast Margin of Laurentia, A.K. Higgins, Jane A. Gilotti, M. Paul Smith
Download citation file:
Two major Neoproterozoic sedimentary basins that probably formed in response to an early pulse of Iapetan rifting along the Laurentian margin are well exposed in the East Greenland Caledonides. The Hekla Sund Basin is exposed at the northern termination of the East Greenland Caledonides, and it is represented by the Rivieradal and Hagen Fjord Groups, which attain a cumulative thickness of 8–11 km. The evolution of this basin reflects deposition during active rifting and a postrift thermal equilibration stage. The Eleonore Bay Basin of East Greenland includes the deposits of the Eleonore Bay Supergroup of early Neoproterozoic age overlain by Cryogenian (mid-Neoproterozoic) glacial deposits of the Tillite Group, which have a combined thickness in excess of 14 km. Four stages of basin evolution may be distinguished based on paleogeographic reorganizations of the shelf and a change from siliciclastic to carbonate deposition, and the final stage was dominated by glacigenic deposition. Major regional stratigraphic breaks seem to be absent, as is other evidence of rift-related sedimentation, suggesting deposition in one or a series of connected ensialic basins. A comparison with other Neoproterozoic basins along the Laurentian margin of the Iapetus Ocean shows similarities between the Eleonore Bay Basin and coeval deposits on Svalbard and the Central Highlands of Scotland. The development of an extensive carbonate platform during the later stages of both the Eleonore Bay and Hekla Sund Basins testifies to a period of tectonic stability prior to onset of Iapetus rifting. The extent of this carbonate platform may have been even larger, since similar successions are present in the Caledonides of Scotland and Ireland.
In the North Atlantic region, extensive basin development was initiated during the Neoproterozoic, and it continued until Ordovician times. This basin formation was associated with the disintegration of the Neoproterozoic supercontinent Rodinia (Piper, 1982) and the subsequent creation of the Iapetus Ocean (e.g., Harland and Gayer, 1972; Winchester, 1988). The actual opening of the Iapetus Ocean occurred in the Ediacaran (late Neoproterozoic) around 570–535 Ma (Cawood et al., 2001). Evidence of Neoproterozoic basin formation comes from thick successions of siliciclastic and carbonate sediments, locally overlain by Marinoan tillites, preserved in the Caledonian terrane within Svalbard, western Scandinavia, the British Isles, and from North and East Greenland, where the well-exposed Hekla Sund and Eleonore Bay Basins form some of the major architectural elements of the East Greenland Caledonides (Fig. 1). In general, the sediments of these basins are only weakly metamorphosed, and only their lower parts show high grades of metamorphism and tectonic overprint. This paper reviews the present knowledge relating to these two sedimentary successions exposed in East and North-East Greenland, and we finally present a comparison with other Neoproterozoic successions along the Laurentian margin of the Iapetus Ocean.
HEKLA SUND BASIN
Sediments of the Neoproterozoic Hekla Sund Basin are exposed at the northern termination of the East Greenland Caledonides. The region constitutes a key area for studies of the western border zone of the orogen (Fig. 2), since it exposes continuous sections from the undisturbed foreland in the west, across parautochthonous foreland affected by folding and thin-skinned thrusting, to allochthonous thrust sheets in the east (Higgins et al., 2001a, 2001b).
The Hekla Sund Basin as used in this paper incorporates both the deposits of the Rivieradal Group (Smith et al., 2004a) and the Hagen Fjord Group (Clemmensen and Jepsen, 1992). The Rivieradal Group has a cumulative thickness of up to 7.5–10 km and is exposed in a major Caledonian thrust sheet (the Vandredalen thrust sheet) in eastern Kronprins Christian Land. The Hagen Fjord Group, which has a maximum thickness of 1000–1100 m, is mainly exposed in the Caledonian foreland to the northwest in western Kronprins Christian Land, Mylius-Erichsen Land, J.C. Christensen Land, and Heilprin Land, but it also occurs within the Vandredalen thrust sheet (Figs. 2 and 3). The Hagen Fjord Group unconformably overlies the Mesoproterozoic Independence Fjord Group (including the Midsommersø Dolerite Formation) and the Zig-Zag Dal Basalt Formation (Clemmensen and Jepsen, 1992; Collinson et al., this volume).
The Rivieradal Group represents a pre-Iapetus, synrift, deep-water succession deposited in an eastward-facing extensional half-graben that was originally situated at least 40 km east of the present outcrop area. The Hagen Fjord Group is composed of fluvial and shallow-marine sandstone overlain by a succession of carbonate platform deposits. It represents a period of postrift thermal reequilibration during the Neoproterozoic, where the youngest units in the group overstep the original rift shoulder (Higgins et al., 2001a, 2001b). The present-day distribution of remnants of the Rivieradal and Hagen Fjord Groups indicates that the original depositional basin was more than 200 km long (parallel to bounding rift faults). During the early stage of basin development (Rivieradal Group), the basin was at least 50 km wide (perpendicular to the rift faults), and it was >300 km wide during the later stages.
Deposits of the Fyns Sø Formation, representing the youngest part of the Hagen Fjord Basin, are at most localities unconformably overlain by sandstone of the Lower Cambrian Kap Holbæk Formation. These are in turn unconformably overlain by the Lower Ordovician Wandel Valley Formation (see Peel and Sønderholm, 1991; Smith et al., 2004b).
The Neoproterozoic sediments in eastern North Greenland were examined by several geological expeditions in the period from 1947 to 1958 (Troelsen, 1949; Adams and Cowie, 1953; Fränkl, 1954, 1955), combined with systematic aerial reconnaissance (e.g., Haller, 1971). During 1978–1980, sedimentologic and stratigraphic studies were carried out on the Hagen Fjord Group sediments as part of Geological Survey of Greenland mapping (Clemmensen and Jepsen, 1992). The Rivieradal Group succession was studied as part of reconnaissance field work in 1980 by Hurst and McKerrow (1981), who recognized three discrete thrust sheets in the Kronprins Christian Land region, and it was suggested that the allochthonous units had been transported ∼150 km to the west during the Caledonian orogeny (Hurst and McKerrow, 1985).
Detailed studies on the sedimentologic, stratigraphic, and structural setting of the Rivieradal and Hagen Fjord Groups were carried out as part of regional mapping of the southern Kronprins Christian Land area in 1994 and 1995 (Sønderholm and Tirsgaard, 1998; M.P. Smith et al., 1999, 2004a, 2004b; Higgins et al., 2001a, 2001b). During this work, the main concepts developed by Fränkl (1954, 1955), Hurst and McKerrow (1980, 1981, 1985), and Hurst et al. (1985) were shown to be accurate and largely valid, although the Rivieradal Group succession was confirmed to be confined to a single thrust sheet, the Vandredalen thrust sheet (Higgins et al., 2001b).
The concept of the Hagen Fjord Group was introduced by Haller (1961), mainly on the basis of aerial reconnaissance work. As a result of field work carried out in 1978–1980, the group was redefined by Clemmensen and Jepsen (1992) to include only Neoproterozoic sediments of mainly shallow-marine origin in the region between Heilprin Land and Kronprins Christian Land (Fig. 2).
The Hagen Fjord Group unconformably overlies Mesoproterozoic strata that are intruded by 1380 Ma dolerite intrusions (Fig. 4; Upton et al., 2005), and it is composed of a lower, siliciclastic part (Jyske Ås, Campanuladal, and Catalinafjeld Formations) and an upper, carbonate-dominated part (Kap Bernhard and Fyns Sø Formations). Previously, the Kap Holbæk Formation was thought to constitute the top of the group (Clemmensen and Jepsen, 1992). However, a major unconformity has been demonstrated beneath the Kap Holbæk Formation, and the deep, decimeter-scale, burrows of the ichnogenus Skolithos that occur within it led M.P. Smith et al. (2004b) to remove the unit from the Hagen Fjord Group and to interpret it as a Lower Cambrian correlative of the Buen Formation of the Franklinian Basin.
The first systematic work on rocks of the Rivieradal Group was carried out by Fränkl (1954, 1955) in the area around Centrumsø in Kronprins Christian Land. He recognized that the Neoproterozoic succession could be divided into an autochthonous and an allochthonous succession separated by a major thrust, upon which his “main nappe” was transported to the west. The metasediments of the nappe were divided into a lower, more metamorphosed part (>1100 m) and an upper, less metamorphic part that also included the upper part of the Hagen Fjord Group succession (2000–3000 m; see Smith et al., 2004a).
Based on Geological Survey of Greenland reconnaissance field work, the allochthonous clastic succession underlying the Hagen Fjord Group was assigned to a single unit, referred to as the “Rivieradal sandstones” by Hurst and McKerrow (1980). Following the mapping of the region in 1994–1995, this unit was formally defined as the Rivieradal Group by Smith et al. (2004a).
Age constraints on the Rivieradal Group are poor and rely upon the ages of the underlying Midsommersø Dolerite Formation and the overlying Hagen Fjord Group, which together indicate that the Rivieradal Group was deposited in the later part of the interval between 1380 Ma and ∼700 Ma (cf. Smith et al., 2004a). The age of the overlying Hagen Fjord Group is also poorly constrained; it is based on a correlation between the Hagen Fjord Group and the Eleonore Bay Supergroup in eastern Greenland (see later section), which both show a transition from siliciclastic to carbonate deposition in the middle Neoproterozoic (Sønderholm and Jepsen, 1991; Sønderholm and Tirsgaard, 1993). The Eleonore Bay Supergroup is overlain by tillites of probable Marinoan age, so a pre–600 Ma age is thus suggested for the Hagen Fjord Group. Reworked glacial deposits of the Morænesø Formation are also present in the outcrop area of the Hagen Fjord Group (Fig. 2). They are of possible Marinoan age and occur in a depositional setting similar to the Tillit Nunatak Formation of the Gåseland, Charcot Land, and Målebjerg tectonic windows (see following; Collinson et al., 1989); however, their exact stratigraphic relationship with the Hagen Fjord Group is not clear.
Two major stages of basin evolution are recognized in the Hekla Sund Basin, along with a phase of initial rifting that created two major east-facing half-grabens. The two half-grabens show considerable differences in subsidence rates, since the deep-marine Rivieradal Group is only represented in the eastern half-graben, and the half-grabens have contrasting stratigraphies. The later postrift stage includes the deposits of the Hagen Fjord Group, which is represented in both the western and the eastern half-grabens (cf. Higgins et al., 2001b).
Stage 1: Rifting and Basin Initiation
Eastern half-graben. In the eastern half-graben of the Hekla Sund Basin, the initial stage of rifting is represented by the Rivieradal Group. The group is lithologically very variable and possesses a strong proximal to distal polarity in which the most distal sediments and stratigraphically lowest levels occur in the east, whereas proximal sediments and higher levels occur to the west (Figs. 4 and 5).
Structural studies in Rivieradal itself suggest that the maximum cumulative thickness of the various units within the entire Rivieradal Group is on the order of 7.5–10 km (Higgins et al., 2001b). However, restoration of the Vandredalen thrust suggests that the Rivieradal profile represents a >50-km-long E–W cross section of the half-graben (Smith et al., 2004a) and that several of the units thus may at least be partly lateral equivalents to each other. The actual stratigraphic thickness of the basin fill is therefore probably less than the 7.5–10 km thickness, and may be on the order of 5–6 km, comparable to the thickness of the Rivieradal Group in northern Vandredalen, where a total thickness of 4500 m has been measured (of which 3000 m are in a continuously exposed section). The basal 200 m of this section lie above a thrust contact with Ordovician carbonates and consist of strongly sheared conglomerates. The conglomerates are overlain by a 500-m-thick phyllite-dominated unit, and then by more than 2200 m of strata dominated by thick-bedded (30–250 cm), structureless to vaguely laminated sandstone turbidites (Ta–c,) interbedded with dark pyritic mudstone, possibly deposited in a fan-delta environment (Figs. 6 and 7A). Laterally and vertically, this succession grades into deep-marine, basin plain mudstone and equivalent phyllitic rocks that crop out over a large part of southeastern Kronprins Christian Land. On the west side of Vandredalen, to the northwest of innermost Ingolf Fjord, a 1000-m-thick succession consisting of storm- and tide-dominated shallow-marine deposits overlies the deep-marine succession. Current directions are generally toward the northeast parallel with the inferred paleocoastline (Figs. 5, 7A, and 8; Higgins et al., 2001b; Smith et al., 2004a).
The most proximal sediments are found along the leading edge of the Vandredalen thrust sheet. In this region, coarse-grained channelized, clast-supported conglomerate units occur along strike in three discrete areas (Figs. 5 and 9). The conglomerate clasts are well rounded and vary in size from a few decimeters to over a meter in the thicker beds, but boulders up to 3–4 m have been recorded. The clasts are derived from the Independence Fjord Group and the Midsommersø Dolerite Formation, which can be presumed to have been exposed to active erosion along the western margin of the Rivieradal Group basin. In a section west of Romer Sø, the thick conglomerate units can be traced eastward (basinward) over a distance of 1–2 km into coarsening- and thickening-upward sandstone units that sometimes grade into conglomerate. Farther eastward, these pass over a similar distance into coarsening- and thickening-upward mudstone–sandstone units. In the southern area around Blåsø, granite and quartz pebbles indicate an additional metamorphic basement source, suggesting that a deeper erosional level was reached here.
Toward the end of Rivieradal Group deposition, a general decrease in the rate of generated accommodation space is recorded by the presence of ∼350 m of fluvial sediments that may be lateral equivalents to some of the conglomeratic units (Fig. 8; Sønderholm and Tirsgaard, 1998). Variations in fluvial style within this unit have been attributed to changing rates of generated accommodation space and a shift toward a more arid climate.
The more distal representatives of the Rivieradal Group are seen at the eastern end of the Rivieradal valley. Pelite with quartzite is exposed at the mouth of Rivieradal and is overlain to the west by pelite, semipelite, and calcareous pelite with prominent metacarbonate units (the “Stenørkenen Phyllites” and “Sydvejdal Marbles” of Fränkl, 1955). This unit is overlain by a thick succession of phyllite and quartzite (the “Taagefjeldene Greywackes” of Fränkl, 1955) similar to the turbidites seen in the section between Ingolf Fjord and Vandredalen (Higgins et al., 2001b).
Overall, the Rivieradal Group basin is characterized by point sources of sediment input (Figs. 5 and 7A). The geometry of the conglomerate deposits, together with their discrete occurrences, suggests the presence of at least three discrete fan-delta systems that acted as major feeder distributary systems. The three conglomerate developments are all in the upper part of the Rivieradal Group succession, but not necessarily at the same stratigraphic level. The repeated cycles of coarsening-upward conglomerate deposition, recorded in the frontal parts of the thrust sheet, may have been controlled by seismic activity on the basin-margin fault systems. Substantial conglomerate fan-deltas accumulated in the immediate vicinity of the source areas and are associated with sandy, proximal turbidites. Between the fans and in the eastern distal part of the basin, sedimentation was dominated by mud and calcareous mud. As the basin filled, the depositional style switched from deep to shallow marine, and the upper part of the group is dominated by less localized and more laterally persistent, tidal- and storm-dominated shallow-marine and locally fluvial deposition (Sønderholm and Tirsgaard, 1998; Higgins et al., 2001b).
Western half-graben. During the initial phase of basin development, sedimentation in the western half-graben was probably very slow and restricted to fluvial and shallow-marine deposition represented by the Jyske Ås Formation, which forms the lower part of the Hagen Fjord Group.
The Jyske Ås Formation is up to 500 m thick, and it records marine transgression following the long hiatus represented by the sub–Hagen Fjord Group unconformity. Although the basal, red part of the formation may include some fluvial sandstone, the main part consists of trough- or tabular, large-scale cross-bedded sandstone, which is interpreted to be of beach and shallow-tidal shelf origin (Clemmensen and Jepsen, 1992). The dominance of sediment-transport directions toward the northeast probably reflects ebb-tidal or storm-enhanced offshore-flowing tidal currents. The northwesternmost exposures of the formation appear, however, to be entirely of fluvial origin.
Direct lithostratigraphic correlation between the eastern and western half-grabens of the Hekla Sund Basin cannot be established during this phase of basin evolution since deposition in the two half-graben segments reflects very different sedimentary and structural environments. This suggests that the rift shoulder between the two half-graben segments—now seen in the area between Harefjeld and Marmorvigen—separated two subbasins that had significant differences in tectonic evolution, with subsidence rates varying by an order of magnitude (Fränkl, 1955; Higgins et al., 2001b).
Stage 2: Postrift Thermal Equilibration
The transition from the rift phase to the postrift thermal equilibration phase of the Hekla Sund Basin probably occurred during early Campanuladal Formation time, since this formation is recognized in both the western and the eastern half-graben segments, indicating that the rift shoulder of the eastern half-graben did not form a barrier at this time (Figs. 4 and 7B).
The Campanuladal Formation (and the correlative Catalina-fjeld Formation in the extreme northwest) overlies the Jyske Ås Formation in the foreland region and the Rivieradal Group in the Vandredalen thrust sheet (Fränkl, 1955; Clemmensen and Jepsen, 1992; Smith et al., 2004a).
The Campanuladal Formation (110–175 m) consists mainly of a variegated succession of fine- to medium-grained sandstone and siltstone units, including a distinctive stromatolitic dolostone. The units are arranged in a characteristic sequence that is recognizable at most localities (Fig. 10). The fine-grained sandstone and siltstone are horizontally laminated and show small-scale wave- and current-formed cross-lamination. Desiccation cracks and gutter casts showing a preferred NE–SW orientation are common. The succession records a transition from inter- and supratidal deposition to more offshore conditions (Clemmensen and Jepsen, 1992). The stromatolitic dolostone horizon in the upper part of the formation probably developed in shallow subtidal environments as a response to reduced clastic influx. Current indications are similar to those observed in the Jyske Ås Formation, suggesting that the open sea was to the northeast (Clemmensen and Jepsen, 1992). In the allochthon, the Campanuladal Formation is represented by an ∼80-m-thick, strongly tectonized, marly, variegated mudstone-dominated unit, which overlies the fluvial sandstone seen in the uppermost part of the Rivieradal Group (Sønderholm and Tirsgaard, 1998).
The Catalinafjeld Formation (260–350 m) mainly consists of laminated mudstone, including thin sandstone turbidite beds that indicate transport directions toward the east. The sediments are considered to represent flooding and the establishment of deeper-marine environments in the northwestern part of the region. Locally, coarsening- and thickening-upward successions record episodes of shoreline progradation (Sønderholm and Jepsen, 1991; Clemmensen and Jepsen, 1992).
The Campanuladal Formation is rather abruptly overlain by the Kap Bernhard Formation (150–215 m), which marks a change from siliciclastic, shallow-shelf deposition to incipient carbonate-platform deposition (Fig. 10). The Kap Bernhard Formation mainly consists of reddish-brown limestone. In the lower part of the formation, soft-sediment deformation structures are abundant, and intraformational breccias are locally conspicuous. Upward, the degree of soft-sediment deformation decreases, and intervals with edge-wise breccias and stromatolitic units occur (Clemmensen and Jepsen, 1992). The carbonate platform may have been initiated along the former rift shoulder and subsequently spread into the eastern and western segments of the basin. The rather sudden shift from siliciclastic shallow-marine deposition to incipient carbonate-platform deposition seen in both basin segments is probably a result of reduced siliciclastic influx related to a climatic change toward more arid conditions during a period of general retrogression. The thicker development of the Catalinafjeld Formation in the westernmost exposures compared to the underlying Jyske Ås Formation of stage 1 could suggest that siliciclastic deposition prevailed for a longer time along the western margin of the basin due to the presence of a larger catchment area in this region. The upper part of the Catalinafjeld Formation may thus correlate with the lower part of the Kap Bernhard Formation farther to the east.
Overlying the incipient platform deposits of the Kap Bernhard Formation, the Fyns Sø Formation (up to ∼325 m) records the establishment of a well-developed prograding carbonate platform (Fig. 10). The formation consists of generally massive dolostone, which, in the upper part, is commonly interbedded with siltstone. Sedimentary structures and textures in the dolostone are locally preserved, including slump structures, intraformational breccias, and rare ripple marks. Stromatolitic horizons occur throughout the formation and are especially common in the uppermost part, where they locally form spectacular linked mounds with a relief of up to 2 m (Fig. 11; Sønderholm and Jepsen, 1991; Clemmensen and Jepsen, 1992). Conical columnar stromatolites (“conophyton”) have also been reported from the formation. These stromatolites were probably restricted to subtidal environments, and they often form the only stromatolitic component in basinal and slope deposits (Donaldson, 1976; Hoffman, 1976; Grotzinger, 1989). The co-occurrence of “conophyton” and slumped horizons suggests subtidal deposition on the slope of a prograding carbonate platform (Sønderholm and Jepsen, 1991).
The Fyns Sø Formation, which forms the uppermost unit of the Hekla Sund Basin, is at most localities overlain by the Early Cambrian Kap Holbæk Formation, which represents early Iapetus passive-margin sedimentation (see Smith and Rasmussen, this volume). Although thought at one time to be conformable, it is now known that this boundary marks a significant hiatus, across which much of the Neoproterozoic is absent, and a substantial paleokarst development occurs in the upper parts of the Fyns Sø Formation (Figs. 7B and 12). Near the crest of the rift shoulder, seen at Sæfaxi Elv, the Kap Holbæk Formation is present only as the fill of paleokarst within the Fyns Sø Formation (Fig. 12), which, at this locality, is overlain by Ordovician carbonates (cf. M.P. Smith et al., 1999, 2004b).
ELEONORE BAY BASIN
Deposits of the Neoproterozoic Eleonore Bay Basin are assigned to the Eleonore Bay Supergroup, which is more than 14 km thick, and to the overlying 800–1300-m-thick Tillite Group, which occurs in an outcrop belt (the Franz Joseph allochthon) that can be followed for 500 km from N to S along the strike of the orogen (Fig. 13). Recent work on the tectonic architecture of this part of the orogen has demonstrated that, although the maximum E–W width between Canning Land and Scoresby Land in the south to Bessel Fjord in the north is currently a maximum of 100 km, the total westward displacement of the thrust sheets was around 200–400 km, with estimated shortening of 40%–60% (Higgins et al., 2004a).
The lower ∼12 km of the Eleonore Bay Supergroup are made up mainly of shallow-marine siliciclastic sediments (Nathorst Land and Lyell Land Groups), whereas the upper 2 km are dominated by carbonate-platform deposits (Ymer Ø and Andrée Land Groups; Fig. 14; Sønderholm and Tirsgaard, 1993). The overlying Tillite Group includes five formations, the lower three of which are Cryogenian (mid-Neoproterozoic) in age and include diamictites and marine deposits. The upper two formations are of Ediacaran age, and they consist of dolomitic mudstone and sandstone of shallow-marine to supratidal origin (Figs. 14 and 15; Hambrey and Spencer, 1987). The Tillite Group is unconformably overlain by the Cambrian–Ordovician Kong Oscar Fjord Group, which represents early Iapetus passive-margin sedimentation (see Smith and Rasmussen, this volume).
The nature of the lower boundary of the Eleonore Bay Supergroup has been widely debated. The oldest sediments of the Nathorst Land Group lie in contact with metasediments of the Krummedal supracrustal succession (Figs. 13 and 16). This boundary has been variously interpreted as transitional (Higgins et al., 1981), an extensional shear zone (White and Hodges, 2002; White et al., 2002), and an unconformity (Friderichsen and Thrane, 1998), although most workers are now agreed that it is an extensional detachment with a significant increase in metamorphic grade between footwall and hanging wall (e.g., Watt et al., 2000). It is uncertain whether the Nathorst Land Group was once in unconformable contact with the underlying Krummedal succession, but it is probable. The relationship is further complicated by extensive anatexis and presence of Caledonian granites that yield dates of 435–425 Ma (Hartz et al., 2000; Watt et al., 2000; Kalsbeek et al., 2001; White and Hodges, 2002; Andresen et al., 2004; Higgins et al., 2004a). The Krummedal succession of the Hagar Bjerg thrust sheet, which forms the footwall of the extensional detachment that bounds the Nathorst Land Group, is also intruded by a suite of older leucogranites that have yielded emplacement ages of 950–920 Ma (Kalsbeek et al., 2000; Watt et al., 2000; Watt and Thrane, 2001; Higgins et al., 2004a). Metamorphic zircons from nonmigmatitic paragneisses of the Krummedal succession have also yielded dates of 955–938 Ma, albeit with larger error bars (Watt and Thrane, 2001). Farther south, in Renland, granites have yielded slightly younger protolith ages of 915–910 Ma that postdate a period of deformation and upper-amphibolite-facies metamorphism at 1000 Ma (Leslie and Nutman, 2000).
The relationship of the Nathorst Land Group to the underlying Krummedal succession and the 950–940 Ma granites is critical to an understanding of the timing of basin formation and the onset of deposition in Eleonore Bay Basin. The Krummedal succession contains detrital zircon populations that are dominated by grains of Paleoproterozoic and Mesoproterozoic age, clustering in the range 1700–1200 Ma, but the youngest grains are 1070 Ma (Kalsbeek et al., 2000; Watt et al., 2000; Higgins et al., 2004b). This provides a tightly constrained window for deposition of the Krummedal succession between 1070 Ma and the metamorphic event and intrusion of granites at 955 Ma onward. Watt et al. (2000) noted that detrital zircon suites in two samples from the base and top of the Nathorst Land Group did not contain a component derived from the 950–920 granites and concluded that the Krummedal sediments and the Nathorst Land Group were part of the same depositional sequence, rather than the latter being part of the Eleonore Bay Supergroup. We consider this to be highly improbable on a number of grounds: (1) the sedimentary facies, depositional style, and architecture of the Nathorst Land Group are strikingly similar to those of the overlying Lyell Land Group; (2) in terms of regional metamorphism, there is a significant contrast between the metasediments of the Krummedal succession and those of the Nathorst Land Group; and (3) the 950–920 suite of granites is widely present within the Krummedal succession of the Hagar Bjerg thrust sheet but is entirely absent from the Nathorst Land Group. We therefore consider the absence of 950–920 granites to reflect the absence of unroofing of the granites at the time of deposition of the Nathorst Land Group. Furthermore, additional data from detrital zircons from the Nathorst Land Group suggest a maximum age for the group of 987 ± 18 Ma (Dhuime et al., 2007). Thus, the upper age limit for the initiation of subsidence in Eleonore Bay Basin can be placed at ∼900 Ma.
The upper age limit of the Eleonore Bay Supergroup is better defined since it is conformably overlain by the diamictites and associated sediments of the Tillite Group (Fig. 15). Although the correlation and age of this glacial episode have been the subject of some debate (see following discussion), there is now a considerable body of evidence in support of a Marinoan age (for reviews, see Fairchild and Hambrey, 1995; Halverson et al., 2004, 2005), corresponding to a 663 Ma age for the top of the Eleonore Bay Supergroup (Halverson et al., 2005). The younger of the two diamictites in the Tillite Group is overlain by a cap carbonate, which defines the base of the Ediacaran and which thus lies at the base of the Canyon Formation, corresponding to a date of 636 Ma (Hambrey and Spencer, 1987; Halverson et al., 2004, 2005). Up to 345 m of Ediacaran sediments overlie the youngest diamictite and are overlain with very low-angle unconformity by the Lower Cambrian Kløftelv Formation of the Kong Oscar Fjord Group (Fig. 15; Henriksen and Higgins, 1976; Hambrey, 1989; Hambrey et al., 1989; Smith et al., 2004b). However, the distinctive suite of fossils that encompasses the Ediacaran fauna has not been recovered from the upper Tillite Group, leading Halverson et al. (2004) to suggest a minimum age limit of 575 Ma for the group, the age of the oldest Ediacaran faunas. The unconformable break between the Tillite and Kong Oscar Fjord Groups would thus span an interval of at least 30 m.y.
Outcrops of the Eleonore Bay Basin are cut by the “western fault zone,” a major normal fault zone that bounds the Devonian basin of North-East Greenland (Fig. 13; Larsen and Bengaard, 1991). Based on evidence from the Devonian succession, sinistral strike-slip movements in the order of 10–20 km along the western fault zone were estimated by Larsen and Bengaard (1991). Correlation of six stratigraphic markers in the Neoproterozoic Andrée Land and Tillite Groups east and west of the fault zone may, however, suggest a sinistral displacement of 80–90 km along the fault (Frederiksen, 2000b).
The spectacular Neoproterozoic succession of East Greenland (Fig. 17) has attracted interest since the earliest geological investigations in the 1870s, although the succession was not studied in detail until the period between the mid-1920s and the late 1950s, when systematic mapping efforts were undertaken on expeditions under the leadership of Lauge Koch (for accounts on previous studies, see Hambrey and Spencer, 1987; Sønderholm and Tirsgaard, 1993). During this time, many lithological and structural studies were carried out, and a lithostratigraphic scheme began to emerge.
During the mapping program of the Geological Survey of Greenland in 1968–1978, the lower part of the Eleonore Bay Supergroup in the Alpefjord region and the rather poorly known succession in Canning Land were investigated, resulting in the first sedimentologic studies of these deposits (cf. Bertrand-Sarfati and Caby, 1976). In the 1980s, several groups studied various aspects of the uppermost part of Eleonore Bay Supergroup and Tillite Group succession, resulting in some detailed sedimentologic work (Hambrey and Spencer, 1987; Herrington and Fairchild, 1989; Manby and Hambrey, 1989; Swett and Knoll, 1989; Moncrieff and Hambrey, 1990), descriptions of different microfossil assemblages (Knoll et al., 1986; Green et al., 1987, 1988, 1989), and a chemostratigraphic correlation of the Eleonore Bay Supergroup with equivalent strata on Svalbard (Knoll et al., 1986).
As part of mapping activities by the Survey in 1988–1990 and 1997–1998 (Higgins et al., 2004a; Henriksen and Higgins, this volume), sedimentologic and stratigraphic studies were carried out throughout the region (Sønderholm and Tirsgaard, 1993; Tirsgaard, 1993, 1996; Tirsgaard and Sønderholm, 1997; Henriksen, 1999; Smith and Robertson, 1999a; Frederiksen, 2000a, 2000b).
The Eleonore Bay Supergroup was formally defined and divided into five groups by Sønderholm and Tirsgaard (1993) using, in part, the more or less informal stratigraphic schemes presented by earlier workers (Fig. 15). The main outcrop area is the central fjord zone (72°N–74°30′N), where the Nathorst Land, Lyell Land, Ymer Ø, and Andrée Land Groups attain a combined thickness in excess of 14 km. However, in Andrée Land, 2.5–3 km of the Nathorst Land Group were eroded away at a well-exposed angular unconformity (Figs. 18 and 19) prior to the deposition of the Lyell Land Group (Henriksen, 1999; Smith and Robertson, 1999a), indicating the presence of at least one significant break in the succession. The Eleonore Bay Supergroup is also readily distinguished farther to the north in the Ardencaple Fjord–Bessel Fjord region (Nathorst Land Group and lower part of Lyell Land Group) and in southern Hochstetter Forland (Nathorst Land and Ymer Ø Groups; Sønderholm and Tirsgaard, 1993), although it is not known whether the unconformity seen at the base of the Lyell Land Group in Andrée Land is present in this area.
Approximately 50 km west of the central fjord zone, sediments crop out in an area centered on Petermann Bjerg (Fig. 13), and these sediments have long been correlated with the deposits of the Eleonore Bay Supergroup (cf. Wenk and Haller, 1953). Since a direct correlation with the main outcrop area could not be established with certainty, Sønderholm and Tirsgaard (1993) assigned these deposits to a separate unit—the Petermann Bjerg Group. More recent investigations (Smith and Robertson, 1999a) have suggested that the lower four units of the group can be correlated with the upper units of the Nathorst Land Group, and that the tentative correlation of the upper part of the Petermann Bjerg Group with the lower part of the Lyell Land Group suggested by Tirsgaard and Sønderholm (1997) is valid (Fig. 15).
The Tillite Group, originally defined by Haller (1971), includes five formations, two of which consist of glacigenic diamictites (Figs. 14, 32; Hambrey and Spencer, 1987). In the foreland to the west, correlatives of the Tillite Group directly overlie basement in the Målebjerg, Charcot Land, and Gåseland tectonic windows (Fig. 13; Higgins et al., 2001a, 2004a).
The development of depositional models for Proterozoic successions in order to elucidate basin evolution is strongly hampered by the lack of reliable biostratigraphic data. Correlations must therefore rely on lithostratigraphic principles. However, these may, if applied rigorously, often lead to false conclusions (van Wagoner et al., 1990), especially in correlations perpendicular to the general coastline trend.
Regional unconformities (i.e., sequence boundaries) may be difficult to recognize in Precambrian successions, which, in consequence, are often falsely regarded as thick, conformable successions (e.g., Harris and Eriksson, 1990); sequence stratigraphic principles can therefore be applied only with difficulty (Christie-Blick et al., 1988). However, the formation of sequence boundaries, particularly on gently dipping shelves and ramps, is often characterized by an associated significant basinward translation of facies, which results in major geographic reorganizations of the basin (van Wagoner et al., 1988; Posamentier and James, 1993). In some successions, flooding surfaces are readily defined, and, where these can be confidently correlated and placed in a sequence stratigraphic framework, they may provide a better basis for the subdivision of a succession and the subsequent definition of genetic stratigraphic packages (e.g., Galloway, 1989).
Neither continental deposits nor horizons of fluvial incision have been observed in the Eleonore Bay Supergroup. Furthermore, there is only one unequivocal semiregional unconformity present within the entire 14 km succession, and it is of tectonic origin (Figs. 14, 18, 19, and 20). However, it is apparent that major paleogeographic reorganization took place several times during the deposition of the succession. It seems highly improbable, therefore, that the 14 km of sediments comprising the Eleonore Bay Supergroup form an entirely conformable succession (cf. Tirsgaard and Sønderholm, 1997).
The widespread nature of the major lithostratigraphic units (formations) and their component facies within the rather narrow window of exposure in the central fjord zone suggest that depositional conditions, including sediment influx and sub sidence, were uniform and parallel to basin strike for most of the time during deposition of (at least) the dominant siliciclastic part of the succession (Nathorst Land and Lyell Land Groups) and the lower part of the carbonate succession (Ymer Ø Group). During this time, the coastline was oriented approximately parallel to the window of exposure (present-day N–S; cf. Tirsgaard and Sønderholm, 1997). Therefore, units that show consistent stacking patterns within each sequence, and that are correlatable throughout the area, are assumed to have formed in response to regional changes in relative sea level and therefore have chronostratigraphic significance.
Within the Eleonore Bay Supergroup, large-scale cyclic sedi mentary patterns have been recognized, and, in order to divide the succession into genetically related packages, second-order sequence boundaries have been placed where the main regional unconformities are considered to be located (Tirsgaard and Sønderholm, 1997). Since only one hiatus is demonstrably present within the basin (at the base of the Lyell Land Group; Figs. 14, 18, 19, and 20), sequence boundaries have been inferred from the succession of facies. On this basis, sequence boundaries are considered to be located where the main regional basinward translations of facies have occurred, and these can be recognized throughout the central fjord zone and in the northern outcrops. They also appear to be associated with laterally extensive erosion of the underlying deposits, and in carbonate deposits with regional dolomitization and formation of dissolution collapse breccias (Fig. 20). The location of the maximum flooding surface is difficult to determine, but it is placed in the middle of the most fine-grained interval or at the shift from retrogradational to progradational deposition (Tirsgaard and Sønderholm, 1997).
In the Eleonore Bay Basin, four major stages of basin evolution may be distinguished based on major paleogeographic reorganizations of the shelf and a change from siliciclastic to carbonate deposition shown by the Eleonore Bay Supergroup, before glacigenic deposition eventually became dominant as recorded by the sediments of the Tillite Group.
Stage 1: Rapidly Subsiding Siliciclastic Shelf
Stage 1 of the evolution of the Eleonore Bay Basin is represented by the Nathorst Land Group, the type area of which is in the cliffs of Alpefjord and Forsblad Fjord (Fig. 21). Due to remoteness of outcrops, difficult access, and the vast thickness of the group, it has not yet been studied in the same detail as the overlying, but very similar, Lyell Land Group. The total thickness is 9 km, and it was divided into seven informal formations (NL1–NL7) by Smith and Robertson (1999a; Figs. 14, 18, and 20); these seven units correspond in turn to the three informal units (NL1–NL3) of Sønderholm and Tirsgaard (1993). Interpretation is further impeded by the tectonic overprint. Metamorphic grade increases rapidly toward the base of the group, attaining garnet zone commonly and sillimanite grade in places. Away from faults and more ductile detachments, deformation is, however, limited to the sporadic development of slaty cleavage, and sedimentologic and sequence stratigraphic studies are possible to some degree at most localities (Caby and Bertrand-Sarfati, 1988; Smith and Robertson, 1999a).
The nature of the lower boundary of the Eleonore Bay Supergroup has been debated for many years, since the contact of the original surface is obscured by abundant granite intrusions. The most recent studies (Watt and Thrane, 2001; White and Hodges, 2002; White et al., 2002; Higgins et al., 2004a) have suggested that the contact is associated with a major décollement surface rather than being a gradual transition to underlying amphibolitic gneisses. The apparent transitional nature of the contact zone resulted from movements on this surface during progressive Caledonian metamorphism (Higgins et al., 1981, 2004a; cf. Gilotti and McClelland, this volume).
Stage 1 may be divided into three second-order sequences that contain a similar range of lithofacies to the overlying stage 2 deposits, but they are distinctive in their exceptional thickness. Sequence 1a is a minimum of 2.4 km thick, 1b is 3.55 km thick, and 1c is 3.95 km thick. Lower-order sequences are not consistently recognizable, but they are clearly exhibited in the carbonate unit, NL5, and the overlying sandstone of NL6.
Sequence 1a. The lowest sequence of the Nathorst Land Group (Figs. 18 and 20) is the least amenable to sedimentologic and sequence stratigraphic study. The lowest lithostratigraphic unit, NL1 (750–950 m), is a quartzite unit that everywhere overlies the Franz Joseph detachment (Fig. 16; Higgins et al., 2004a). Toward the base, there is extensive anatexis and intrusion of granites, and the lowest exposed parts of the unit occur as xenoliths composed of quartzites and thin interbedded semipelitic gneissose rocks, both of which contain cordierite and sillimanite (Smith and Robertson, 1999a). However, toward the top of the unit, the metamorphic grade decreases, and the higher parts consist of a distinctive banded unit with medium-gray, ripple-laminated and cross-bedded sandstone together with very pale cross-bedded quartzarenites. Thinner, muddier, ripple-laminated beds occur throughout the unit and probably correspond to the gneissose semipelite in the lower part. NL1 is overlain by a mixed succession of heterolithic siltstone, fine-grained sandstone, and mudstone interbedded with quartzarenite and gray sandstone, the latter of which are large-scale cross-bedded or parallel laminated. This unit, assigned to NL2 by Smith and Robertson (1999a), is at least 2000 m thick.
Sequence 1b. The bipartite division seen in sequence 1a, which consists of sandstone (many of them of quartzarenite composition) overlain by heterolithic sediments, is maintained through sequences 1b and 1c, and indeed into the overlying Lyell Land Group. The progradational part of sequence 1b (Figs. 18 and 20) is NL3, a distinctive 1000-m-thick sandstone unit that consists of thick bedded sandstone with large-scale cross-bedding in the lower part overlain by quartzarenites in which strongly wedge-shaped units are present within massive gray sandstone and ripple-laminated sandstone. The high maturity, the marked absence of mud, and the presence of dunes and megaripples at a variety of scales are together interpreted as a shoreface facies association. The NL3–NL4 boundary is a prominent flooding surface, and there is evidence of a marked shift to outer-shelf, storm-dominated sediments consisting of normally graded, parallel- and ripple-laminated siltstone and fine-grained sandstone. This lower, 150-m-thick unit is overlain by 150 m of wavy bedded siltstone and 180 m of cross-bedded quartzarenites, but the remaining 1800 m is lithologically monotonous, composed of interbedded cross-bedded sandstone and wave-rippled, wavy bedded heterolithic siltstone/mudstone interpreted as a storm- and wave-dominated inner-shelf facies association. NL4 passes upward into the only carbonates that occur within the Nathorst Land Group (Figs. 18 and 20). NL5 is a 250 m carbonate unit composed of calcareous wavy-bedded heterolithic siltstone intensely veined by carbonates, which are probably pseudomorphed evapo rites. In the upper part of the unit, low-relief stromatolites are preserved and shallowing-upward parasequences are common. Typical parasequences have calcareous heterolithic sediments at the base, which become increasingly veined upward, with a microbialite cap made up of stromatolites and crinkly laminated mats with desiccation cracks and tepee structures. NL5 is capped by the quartzarenites of NL6 (sequence 1c), and the total thickness of sequence 1b is 3.55 km.
Sequence 1c. The base of sequence 1c (Figs. 18 and 20) is marked by the sharp erosive base of quartzarenites that constitute NL6. The quartzarenites are white, fine-grained, and channelized in places; trough cross-bedding, parallel lamination, and wave ripples are present. In cliff sections, NL6 (400 m) clearly exhibits three parasequences made up of pale quartzarenite interbedded with darker, finer-grained units. The quartzarenites, like those at the base, are dominantly trough cross-bedded with wedge-shaped beds and rippled tops. The darker units are composed of heterolithic, ripple-laminated very fine sand with muddy drapes. An abrupt boundary to NL7 is overlain by 130 m of graded siltstone to mudstones, in packets a few centimeters thick; these are interpreted as outer-shelf, sub–storm-wave-base deposits, where the NL6–NL7 boundary represents a major flooding surface. NL7 is 2000 m thick in total and consists of wavy-bedded silts, sands, and muds with interbedded sharp-based, tabular, parallel-laminated and current- and wave-rippled sandstone; the current ripples are frequently bidirectional. The sediments of NL7 are interpreted as an association of storm- and wave-dominated sediments.
The construction of a full sequence stratigraphic model for the Nathorst Land Group is hampered by the narrow, linear outcrop belt within the Franz Joseph allochthon. However, some lateral variability is evident northward from the type area, 170 km along strike, and between the Nathorst Land Group and the isolated klippe of coeval sediments, assigned to the Petermann Bjerg Group, which lie 110 km northwest of the type area of Alpefjord.
The lateral variability along strike is the result of a pronounced angular unconformity that occurs at the base of the Lyell Land Group in Eremitdal, at the junction with Snestormdal in Andrée Land, where it rests on NL4 (Fig. 20). The absence of NL5–NL7 indicates that at least 2.5–3 km of the succession seen farther south are missing at this boundary and provides evidence for intra–Eleonore Bay Supergroup uplift and erosion.
The Petermann Bjerg Group (Fig. 22) crops out in a klippe of the Franz Joseph allochthon, centered on Louise Boyd Land and Frænkel Land, that measures 80 km N–S and 60 km E–W. Although it has long been clear that the group is correlative with the Eleonore Bay Supergroup (Wenk and Haller, 1953; Haller, 1971; Sønderholm and Tirsgaard, 1993), a detailed correlation has not been available until recently (Smith and Robertson, 1999a). The metamorphic grade makes detailed sedimentology impossible in all but the quartzites, but it is now clear that the Petermann Bjerg Group is an attenuated correlative of the upper Nathorst Land Group and lowermost Lyell Land Group (Fig. 23). More specifically, formations PB1 and PB2 and the lower part of formation PB3 correspond to sequence 1b (Fig. 23), and the upper part of formation PB3 and formation PB4 correspond to sequence 1c; formations PB5 and PB6 are correlative with the Kempe Fjord and Sandertop Formations, respectively, which form the basal part of the Lyell Land Group (Smith and Robertson, 1999a). This correlation gives critical three-dimensionality to interpretations of stage 1. Sequence 1b thins from 3.55 km in the Alpefjord–Forsblad Fjord area to around 2300 m in the Petermann Bjerg Group (Fig. 23), and it becomes correspondingly dominated by the quartzarenite shoreface association—in Alpefjord, this accounts for 28% of sequence 1b, whereas in the Petermann Bjerg Group, it amounts to around 44%. Similarly, sequence 1c thins from 2 km in the type area of the Nathorst Land Group to a little over 1 km in the Petermann Group, and quartzarenites again dominate the sequence in the latter case—the quartzarenites and associated sediments of the shoreface facies association comprise 20% of sequence 1c in the type area of the Nathorst Land Group, whereas they constitute 79% of the same sequence in the Petermann Bjerg Group (Fig. 23).
Stage 2: Stable Siliciclastic Shelf
Stage 2 is composed of three second-order sequences (Figs. 20 and 24), all of which are characterized by thick retrogradational to progradational successions that show a similar trend in sedimentary evolution through time. In these sequences, the sequence boundary is placed where shoreface or coastal plain deposits sharply overlie outer-shelf or storm-dominated inner-shelf deposits. The vertical development of facies associations indicates that a major regional translation of facies is associated with the abrupt transition from shelf mudstone to shoreface or coastal plain sandstone. This translation is also interpreted to have been associated with regional erosion and appears to have formed in relation to a large-scale forced regression.
Sequence 2a. The lower sequence is represented by the Kempe Fjord and Sandertop Formations, which have a total thickness that reaches 600–1000 m in the central fjord zone (Fig. 24). It overlies a major angular unconformity that cuts out 2.5–3 km of the top of the Nathorst Land Group in Andrée Land (Fig. 20). The regional extent of this unconformity is unknown, but both 50 km to the south in Lyell Land (along apparent basinal strike) and 50 km to the WSW (approximately perpendicular to basinal strike), the Lyell Land Group rests without angular discordance upon the uppermost formation of the Nathorst Land Group. The lower boundary is, however, marked by an abrupt shallowing from outer-shelf mudstone to shoreface and coastal plain sandstone at other localities. The initial shallowing is followed by a more gradual shallowing that is reflected by the lower 200 m of the Kempe Fjord Formation, where deposits of the storm- and wave-dominated shoreface association are overlain by subtidal sandstone. Toward the top, these grade into a 300-m-thick succession of predominantly intertidal deposits. The intertidal channel deposits are succeeded by ∼100 m of sandstone of mainly subtidal origin, implying a subtle rise in relative sea level. This weak deepening trend observed in the upper part of the Kempe Fjord Formation becomes more pronounced in the Sandertop Formation (Fig. 24). The deepening of the shelf is reflected in a change in the lower 80 m of the Sandertop Formation, where storm-dominated inner-shelf deposits gradually give way to outer-shelf deposits that represent the maximum flooding of the basin. Above this unit, storm- and wave-dominated inner-shelf deposits gradually become more abundant, signifying renewed progradation (Fig. 20).
A poorly developed overall shallowing is signified by the overlying 320 m of sediment, which constitute the rest of the Sandertop Formation (Fig. 24). Sharp-based shoreface deposits occur within the storm-dominated inner-shelf deposits and reflect minor progradational events, possibly caused by forced regressions.
The sequence boundary is located at the contact between the Sandertop Formation and the Berzelius Bjerg Formation (Fig. 24), and it marks an abrupt shift from heterolithic mudstone deposited in a storm- and wave-dominated inner-shelf environment to tidally dominated coastal plain sandstone. In contrast to sequence 2a, shoreface deposits are only a few meters thick, and coastal plain deposits are present almost directly above the sequence boundary. Similar to the development seen in the lower part of sequence 2a, the lower 400 m of sequence 2b reflect a stillstand in relative sea level, following the abrupt fall inferred at the base. A relative sea-level rise is indicated by the deepening trend recorded by the top of the Berzelius Bjerg Formation, where the coastal plain deposits pass up into 80 m of fine-grained shoreface sandstone. This trend continues into the Kap Alfred Formation, where the shoreface deposits pass into storm- and wave-dominated inner-shelf deposits that form a relatively uniform 130-m-thick succession (Fig. 24). On top of this, there is a 400-m-thick succession that consists of interbedded tidally influenced shoreface and storm- and wave-dominated inner-shelf deposits, which mark a shift toward a more tidally dominated shelf. The uppermost 100 m of the sequence consists entirely of outer-shelf mudstone, representing the culmination of the overall deepening (Figs. 20 and 24).
Sequence 2c. Sequence 2c is made up of the sediments of the Vibeke Sø and the Skjoldungebræ Formations (Figs. 17 and 24); it reaches a thickness of 500–550 m in the central fjord zone and in Canning Land.
Sequence 2c starts with abrupt shift from outer-shelf mud-stone to mature, structureless sandstone deposited in a storm- and wave-dominated shoreface environment. Numerous mudstone intraclasts and a highly irregular base indicate that the sequence boundary was erosive. Following the initial pronounced sea-level fall, the sequence records a sea-level stillstand followed by a gradual deepening. The lower 150 m of the sequence consist primarily of shoreface deposits. However, in the central fjord zone, 10–20 m of heterolithic storm- and wave-dominated inner-shelf deposits occur 20–25 m above the base, reflecting minor, high-frequency variations in relative sea level.
Above the shoreface deposits, there is ∼150 m of sediments that show a recurrent interbedding of 1–5-m-thick shoreface deposits and 5–40-m-thick storm-dominated inner-shelf deposits, reflecting repeated episodes of regression followed by transgression. An overall deepening is manifested by a gradual upward thinning of the sandstone beds and a gradual thickening of the heterolithic deposits. In the uppermost 200 m of the sequence, storm-dominated inner-shelf deposits give way to outer-shelf mudstone interbedded with thin, sharp-based shoreface sandstone. A weak tendency toward upward coarsening is present in the uppermost 50 m, where a return to interbedded inner-shelf and shoreface deposits is seen (Fig. 24).
The repeated episodes of forced regression seen throughout most of this sequence reflect the superimposition of high-frequency relative-sea-level variations upon the overall relative sea-level rise. The high-frequency sea-level variations give rise to 20–50-m-thick regressive-transgressive cycles, similar to those observed in the sequences below. These cycles cannot convincingly be correlated between outcrops, and their lateral extent is unknown. Cyclic regressive-transgressive events on a scale of 70–120 m are also visible within the sequence and appear to be correlatable throughout the fjord zone and out into Canning Land. The cyclic pattern is produced by stacked successions of upward-thickening sandstone units.
Stage 3: Carbonate-Platform Development
Stage 3 is represented by the uppermost formation of the Lyell Land Group and the Ymer Ø and Andrée Land Groups, and it includes four second-order sequences with a total thickness of ∼2700 m in the central fjord zone (Figs. 20 and 25). The lower sequence marks the change from siliciclastic to dominantly carbonate deposition and is characterized by a progradational evolution of the carbonate platform (sensu Read, 1982). The second sequence reflects a retrogradational to progradational and aggradational carbonate-platform succession. A shift toward a more humid climate in the later part is suggested by a marked increase in red siliciclastic mud that intermittently caused carbonate deposition to cease, at least on the inner parts of the platform. The third sequence shows an overall retrogradational pattern terminated by an abrupt fall in sea level that resulted in exposure of most of the carbonate platform (Figs. 20 and 25). The upper sequence, heralding the Marinoan glaciation represented by the deposits of the Tillite Group, consists of a strongly retrogradational succession that indicates drowning of the carbonate platform of the previous stage and deposition of deep-marine deposits, followed by a short period of carbonate-platform progradation before glacigenic deposition took over (Fig. 20).
Sequence 3a. Sequence 3a is made up of sediments of the Teufelsschloss, Kap Peterséns, and Antarctic Sund Formations, and it reaches a thickness of ∼500 m in the central fjord zone (Figs. 17, 25, and 26). The sequence was induced by a major transgressive event, possibly accompanied by a shift to a more arid climate, and it records a gradual change from siliciclastic to carbonate deposition (Fig. 20).
Overlying the sequence boundary, initial lowstand forced regression deposits are represented by shoreface sandstone of the Teufelsschloss Formation (Fig. 26), which attains a thickness of around 130 m in most of the outcrop area. However, rapid thinning occurs in the southeastern part of the fjord zone and in Canning Land, where thicknesses of only 15–60 m are reported (Tirsgaard and Sønderholm, 1997). The shoreface deposits are succeeded by outer-shelf mudstone of the Kap Peterséns Formation (∼280 m), and local large-scale channelized slump structures and progradational patterns become conspicuous upward. However, the shift to a more arid climate and ensuing carbonate deposition resulted in declining sedimentation rates and renewed generation of accommodation space. Initial progradation of a carbonate platform is indicated by a gradual increase in carbonate content in the mudstone of the Kap Peterséns Formation (Fig. 25). These are succeeded by cherty limestone of the Antarctic Sund Formation (115 m), in which common resedimented and slumped horizons reflect continued progradation of carbonate and slope and shelf sediments (Fig. 27). The top of the sequence is marked by a dolomitized, erosional unconformity.
Sequence 3b. Sequence 3b is made up of sediments of the Tågefjeld, Rytterknægten, Skildvagten, and Elisabeth Bjerg Formations of the Ymer Ø Group and formations AL1–AL3 of the Andrée Land Group (Figs. 17, 20, 25, and 27; cf. Sønderholm and Tirsgaard, 1993). The sequence reaches a thickness of ∼1050 m in the central fjord zone. The formations of the Ymer Ø Group are relatively constant in thickness throughout the area of exposure, whereas the formations of the Andrée Land Group show marked local thickness variations in both S-N and W-E directions.
Although the facies development in the Andrée Land Group seems to record a deepening trend toward the north or northeast, there are no signs of a major tectonic reorganization of the overall approximate N–S basinal trend during this time. The change from a N–S coastline trend during the early part of the sequence to a more E–W trend reported by Frederiksen (2000a) therefore probably reflects the development of a more indented coastline resulting from varying growth rates of individual ramp segments along the general N–S coastline trend.
The basal part of sequence 3b is represented by the Tågefjeld Formation (∼190 m) and is characterized by a retrogradational succession initiated by widespread succession of evaporitic, shallow-water carbonates and lagoonal mudstone (Fig. 27) grading upward into deeper-water limestone. These are locally associated with complex algal mound structures up to 80 m thick and 250 m wide and possible large-scale slumped units, which suggest platform-margin deposition. This major transgressive phase resulted in drowning of the platform and eventually led to a change to basinal and slope deposition recorded by finely laminated, locally shaly limestone of the 140-m-thick Rytterknægten Formation.
Renewed platform progradation is indicated by an upward increase in abundance of slumped horizons in the Rytterknægten Formation (Fig. 20). These deposits are followed by a distinctive dolomite unit (Skildvagten Formation; 180 m) in which irregular stromatolite growth patterns reflect progradation of high-energy platform-margin environments. A shift toward a more humid climate is suggested by a marked increase in red siliciclastic mud, which eventually caused carbonate deposition to cease, at least on the inner parts of the platform, where a mixed siliciclastic-carbonate succession consisting of stacked inner- and outer-shelf deposits was laid down (Figs. 20 and 27; Elisabeth Bjerg Formation; 290 m). These are traceable throughout the entire outcrop area and probably reflect a response to third-order eustatic sea-level changes (Tirsgaard, 1996). An overall decrease in siliciclastic sediment supply within the Elisabeth Bjerg Formation heralds the return to inner-shelf, high-energy carbonate deposition and evaporite formation, recorded by the lower part of the Andrée Land Group.
The lower three formations of the Andrée Land Group reflect a complex relationship between storm-dominated, shallow-marine microbial and pisolitic ramp environments and mid- to outer-ramp environments (Fig. 25; Frederiksen et al., 1999; Frederiksen, 2000a, 2000b). There is no overall trend in stacking patterns, suggesting a general aggradation of the carbonate ramps during times when carbonate production, sea-level change, and subsidence were more or less in balance. The general tendency for very shallow environments and the exposures of parts of the inner ramps with subsequent evaporation, dissolution, and collapse suggest, however, that deposition occurred during an overall fall in relative sea level (Fig. 20).
Formation AL2 consists of dolomites that show a high proportion of polymict, stromatoclast, and dissolution collapse breccias. The polymict breccias, which show a high diversity of clast types, suggesting that the channel fills were probably derived from a large source area, are interpreted to have formed in an extensive, laterally migrating channel belt (Frederiksen, 2000b) that may have formed a lowstand, incised valley system.
Sequence 3c. Sequence 3c includes sediments of formations AL4 and AL5 of the Andrée Land Group (sensu Fränkl, 1953; Sønderholm and Tirsgaard, 1993). The thickness of this sequence is ∼700 m in the central fjord zone (Figs. 25 and 27).
The lower sequence boundary is placed at the base of dolomitic polymict, stromatoclast, and dissolution collapse breccias that occur close to the base of formation AL4 (Figs. 20 and 25). The formation is 30–130 m thick and shows considerable thickness variation. During sea-level lowstand, the polymict and stromatoclast breccias were probably formed in channels and around storm-dominated microbial reef tracts that occurred between subaerially exposed highs experiencing dissolution collapse (Frederiksen, 2000b).
The subsequent overall transgression is documented by the lower ∼500 m of formation AL5, in which inner- to mid-ramp environments such as subaerially exposed flats, lagoons, microbial reefs, pisoid shoals and flats, and channel flats are followed by outer-ramp environments.
A major sea-level fall resulted in exposure of most of the carbonate platform and ensuing extensive dolomitization and dissolution collapse. This is shown by a 20–130-m-thick dolomitic unit, which forms the top of formation AL5, that can be traced throughout most of the region (Frederiksen, 2000b). The upper sequence boundary is placed on top of this unit, since it records dissolution rather than erosion and redeposition (Fig. 25).
Sequence 3d. The uppermost sequence is a retrogradational to progradational succession made up of formations AL6 and AL7 (Figs. 20, 25, and 28). The combined thickness of this succession is ∼250 m, but individual units show large regional variation in thickness (Figs. 25 and 28).
A major relative sea-level rise resulted in drowning of the partly exposed carbonate platform of sequence 3c (Fig. 20; Frederiksen, 2000b). A thin succession of possible shallow-water carbonates overlies the dissolution collapse breccias, followed by ∼100 m of deep-water shaly carbonate turbidites and slump breccias of formation AL6.
In the northern part of the central fjord zone, shallow-water carbonate-platform sedimentation resumed (formation AL7) before glacigenic deposition abruptly took over as recorded by the sharp but transitional contact to the overlying Tillite Group (Fig. 28). Carbonate-platform conditions were not established further to the south, and it is therefore possible that the upper part of formation AL6 is time equivalent to formation AL7 (Fig. 20; Herrington and Fairchild, 1989).
The abrupt increase in accommodation space, recorded by the cessation of stable shallow-water carbonate-platform environments and documented by the change from sequences 3a–3c to the mixed siliciclastic–carbonate deep-water sediments of sequence 3d, could reflect a change toward a more humid climate combined with a major relative sea-level rise. It has been suggested that the sea-level rise may have been driven by tectonically induced subsidence and tilting related to an initial rifting phase associated with the opening of the Iapetus Ocean (Herrington and Fairchild, 1989; Fairchild and Hambrey, 1995; Frederiksen, 2000b). However, direct evidence of synsedimentary faulting, apart from large variations in depositional environment and thickness of individual units over short distances, has not been observed.
Stage 4: Glacigenic Deposition
The Tillite Group overlies the Eleonore Bay Supergroup with no apparent major hiatus, and it is divided into five formations (Fig. 29; Hambrey and Spencer, 1987). The oldest unit, the Ulvesø Formation (100–318 m), is dominated by diamictite and is overlain by dolomitic shale, siltstone, and sandstone of the Arena Formation (223–320 m), which is in turn overlain by a second diamictite unit, the Storeelv Formation (60–223 m; Hambrey and Spencer, 1987). The Storeelv Formation is overlain by a cap carbonate, which marks the base of the Ediacaran (Knoll et al., 2004) and is assigned to the Canyon Formation (up to 300 m). The bulk of the Canyon Formation is made up of dolomitic shales, and stromatolitic dolostone is developed toward the top. The youngest unit in the Tillite Group, the Spiral Creek Formation (up to 45 m), is developed only in the central part of the basin and is overlain with slight angular unconformity by quartzarenites of the Lower Cambrian Kløftelv Formation.
The Ulvesø Formation contains a wide range of glacial and related lithofacies, documented by Hambrey and Spencer (1987) and Moncrieff and Hambrey (1988, 1990). These include waterlain tillite, debris-flow diamictite, glaciomarine deposits, and eolian sandstone together with indications of permafrost conditions (contraction wedges and load structures). The environment is interpreted as one in which a floating, debris-bearing ice shelf supplied sediment into a shallow-marine environment (Fig. 30). There are no close analogues of this situation at present—true ice shelves are restricted to Antarctica at the present but are debris-free unless freezeon of saline ice allows sediment to be retained until calving takes place (M. Hambrey, 2007, personal commun.). Current activity was present beneath the floating ice in the Ulvesø Formation, and it produced lenses of sorted material in the waterlain tillites and glaciomarine facies (Moncrieff and Hambrey, 1990).
In most parts of the basin, massive sandstone assigned to the Arena Formation directly overlies the Ulvesø Formation, and there is no cap carbonate (Hambrey and Spencer, 1987). In general, the Arena Formation fines upward to shale or interbedded sandstone and shale, with ripple lamination and interference ripples present. At the snout of Sorteelv Gletscher, in northern Scoresby Land, two thin diamictite horizons similar to those in the Ulvesø Formation are present in the lower two-thirds of the Arena Formation, each of which is up to a few meters thick. The Arena Formation ranges in thickness from 100 to 360 m and was probably deposited in a marine environment (Hambrey and Spencer, 1987).
The younger diamictite unit, the Storeelv Formation (Figs. 29 and 31; 60–223 m), contains a suite of waterlain tillite and glaciomarine lithofacies similar to those in the Ulvesø Formation, but it also contains clast-rich muddy diamictite units with aligned pebble fabrics that have been interpreted as lodgment tills representing deposition beneath grounded ice. In places, this lithofacies is seen to overlie a striated boulder pavement (Moncrieff and Hambrey, 1990). Furthermore, at Kap Weber, a horizon of the lodgment till is seen to pass laterally into a proximal waterlain till (Moncrieff and Hambrey, 1990), which represents the point at which the ice detached from the substrate and began to float. Whereas the Ulvesø Formation is dominated by intrabasinal clasts, derived principally from the Eleonore Bay Supergroup, the Storeelv Formation contains a higher proportion of exotic clasts, including crystalline basement lithologies (Hambrey and Spencer, 1987; Hambrey et al., 1989).
In contrast to the older glacial deposits of the Ulvesø Formation, the Storeelv Formation is overlain by a characteristic 6–13-m-thick cap carbonate (Fig. 29; Hambrey and Spencer, 1987; Fairchild and Hambrey, 1995). The peritidal carbonates of this unit, the Canyon Formation, pass upward into dolomitic shales (25–35+ m), and the deepening trend continues into a 160–190-m-thick mudstone-siltstone unit, which is interpreted as having been deposited at outer-shelf depths. The Canyon Formation shows a marked shallowing trend through the uppermost 50 m, where the shale facies passes upward into peritidal dolostone with storm deposits (Fairchild and Herrington, 1989; Fairchild and Hambrey, 1995). This shallowing trend is capped by the Spiral Creek Formation (24–45 m), which was deposited in playa-lake environments and which is composed of sandstone, siltstone, and mudstone with abundant halite pseudomorphs and gypsum together with mud cracks, ripple lamination, and silicified pebbles (Hambrey and Spencer, 1987; Fairchild and Hambrey, 1995). The formation is present only in the central part of the basin, on Ella Ø and in Andrée Land—to the north and south, the overlying Kløftelv Formation (Lower Cambrian) rests unconformably on the Canyon Formation.
Units of inferred glacial origin have also been identified in the parautochthonous foreland windows of the orogen. Although the Tillite Group, within the Franz Joseph allochthon, is now superimposed upon this foreland, it is separated from the foreland by both the Hagar Bjerg and Niggli Spids thrust sheets (Higgins et al., 2004a). Palinspastic reconstructions of the orogen suggest that the original separation between the foreland tillites and those in the Franz Joseph allochthon was in the order of 200–400 km (Higgins and Leslie, 2000; Higgins et al., 2004a).
Tillites have been identified in the tectonic windows in Gåseland (Wenk, 1961; Phillips et al., 1973), Charcot Land (Moncrieff, 1989), and at Målebjerg (Fig. 32; Smith and Robertson, 1999b). They differ significantly from those in the Franz Joseph allochthon because they directly overlie crystalline basement, indicating that the entire 13+ km of the Eleonore Bay Supergroup is absent. Moncrieff (1989) documented the tillite of the Charcot Land and Gåseland windows, where tillites sit in hollows on the eroded top of Paleoproterozoic gneisses and supracrustal rocks, beneath thrusts that transport Archean–Paleoproterozoic gneisses overlain by the Krummedal succession. Moncrieff (1989) erected different names for the tillite units in each of the two windows, the Tillit Nunatak Formation in the Charcot Land window and the Støvfanget Formation (misspelled as Stofvanget Formation) in the Gåseland window, but these are clearly lateral equivalents; the former name is selected here for the tillites that occur in this depositional and tectonic context in the Charcot Land, Gåseland, and Målebjerg windows. The lithofacies preserved in the two windows include tillite, sandstone, and laminated mudstone with dropstone. All of the clasts were derived locally and range up to 2 m. The maximum thickness is less than 200 m at Tillit Nunatak, but it is much less at most localities, rarely exceeding 25 m. Moncrieff (1989) interpreted the unit as consisting of lodgment tills near the base of the unit with diamictites deposited from floating ice (ice shelf and icebergs) and debris flows higher in the unit. The finer interbedded sediments were interpreted as being of predominantly turbiditic origin and, in Gåseland, proglacial outwash. Moncrieff (1989) correlated the Tillit Nunatak Formation with the Tillite Group of the Franz Joseph allochthon, and more specifically with the Storeelv Formation.
The glacial sediments assigned to the Tillit Nunatak Formation in the Målebjerg window (Smith and Robertson, 1999b) are directly comparable with those in windows farther to the south, but rather than being truncated by thrusts, they are unconformably overlain by the Lower Cambrian quartzites of the Slottet Formation (Smith et al., 2004b). The unit has a maximum thickness of 31 m and includes two diamictites separated by 21 m of phyllites, platy quartzites, and semipelites. Clasts in the upper diamictite range up to 6 m, and the bed is dominated by clasts of granitic lithology.
As with the glacial deposits of the Franz Joseph allochthon, those of the Tillit Nunatak Formation display a close inter action with adjacent marine environments. There is no evidence that these lay within “paleohighlands” as suggested by Halverson et al. (2004), since there is no evidence of significant relief, but there is evidence of deposition from floating ice (Moncrieff, 1989). These localities were well inboard of the Eleonore Bay Basin (Higgins et al., 2004a). They did not, however, constitute highlands, since they show evidence of marine environments.
Age of the Tillite Group
The age of the Tillite Group has been the subject of significant controversy. Early work correlated the unit with the Varanger tillites of northern Norway (Troelsen, 1956; Haller, 1971; Hambrey and Harland, 1981; Hambrey, 1983) and, more particularly, with the Neoproterozoic tillites of NE Spitsbergen and Nordaustlandet (Svalbard). Detailed work in Svalbard and eastern Greenland has demonstrated unit by unit, and to some extent bed by bed, correlation of the glacial horizons and associated units (cf. Fairchild and Hambrey, 1995), to the extent that it is clear that the two areas must have been part of the same basin prior to Caledonian collision and strike-slip dismemberment of the Laurentian margin. In considering the age of the glacial events in the Eleonore Bay Basin, it is therefore possible to utilize the combined data set from the two regions.
The development of interest in Neoproterozoic glaciation and the “Snowball Earth” hypothesis has led to a considerable focus on developing an understanding of the number, correlation, and dating of the events (for reviews, see Hambrey and Harland, 1985; Knoll, 2000; Hoffman and Schrag, 2002). There is now general consensus that there were two widespread, if not global, Neoproterozoic glaciations—the Sturtian and the Marinoan, which have been recognized and radiometrically dated in the same section of the Flinders Range in South Australia (Knoll et al., 2004)—together with an uncertain number of more localized and/or more poorly known events. There has been uncertainty, however, over whether the events recorded in the Tillite Group and its Svalbard correlatives represent the Marinoan glacial event (Halverson et al., 2004) or the Sturtian (Brasier and Shields, 2000; Robb et al., 2004). The situation has been further confused by uncertainty over whether some of the glacial events contain paired tillites, or whether the two events recorded in Greenland could represent combined Sturtian + Marinoan events (Kennedy et al., 1998) or Marinoan + post-Marinoan (Gaskiers) events (Kaufman et al., 1997), with a condensed succession or hiatus between them. Thus, in the last ten years, virtually every possible permutation of correlation for the glacial events in the Eleonore Bay Basin has been advanced. Detailed work by Halverson et al. (2004, 2005) may point to a resolution of this debate. The detailed correlation of the East Greenland and Svalbard successions proposed by earlier authors was confirmed by Halverson et al. (2004), while Halverson et al. (2005) recognized the presence of the distinctive pre-Marinoan negative δ13C anomaly (the “Trezona anomaly”) beneath the lowest tillite in NE Svalbard and confirmed that the cap carbonate is of typical end-Marinoan type. Recently, this anomaly has also been identified in formation AL7, which underlies the Tillite Group (Kristiansen, 2007). This indicates that both tillite horizons in Svalbard and East Green-land are correlatives of the Marinoan glaciation. Furthermore, this suggests that the Tillite Group does not extend upward as far as the 580 Ma Gaskiers glacial event, which supports the conclusions based on the absence of an Ediacaran fauna, and it may be concluded that the Sturtian glaciation is not represented in the Eleonore Bay Basin. This may be because the Sturtian glaciation did not extend to this part of Laurentia (see Halverson et al.  for discussion) or because the oldest basin infill postdates the Sturtian event.
COMPARISONS WITH NEOPROTEROZOIC BASINS ALONG THE LAURENTIAN MARGIN OF THE IAPETUS OCEAN
The late Neoproterozoic breakup of Rodinia and the development of the Iapetus margin of Laurentia involved a protracted and complex history. Two pulses of rifting and associated extrusive igneous activity at ∼760–700 Ma and ∼620–590 Ma have been recognized in the Appalachians; however, only the late phase proceeded to continental separation and opening of the Iapetus Ocean (Aleinikoff et al., 1995). In Newfoundland, this final pulse was a two-stage process involving the formation of the Iapetus Ocean between Laurentia and Gondwana at 570 Ma followed by continued rifting that generated one or more terranes at around 535 Ma (Fig. 33; Cawood et al., 2001).
Pre-Iapetus basins that are possibly coeval with those in the Greenland Caledonides have been recognized in Scotland and Ireland (Dalradian Supergroup basin, Moine Supergroup basin, and the Sleat, Stoer, and Torridon Group basins), and Svalbard (the basin or basins filled by the Hekla Hoek succession; Higgins et al., 2001b).
The successions that show the greatest similarity are those of the Eleonore Bay Basin and the Tonian–Cryogenian (early–middle Neoproterozoic) part of the Hekla Hoek succession of northeastern Svalbard. These two successions show much the same sedimentary evolution from siliciclastic to carbonate deposition capped by Marinoan glacigenic deposits, and they seem to lack major regional stratigraphic breaks and evidence of rift-related sedimentation. Furthermore, they are similar in attaining very substantial thicknesses (Harland and Gayer, 1972; Harland, 1985, 1997; Hambrey, 1989; Harland et al., 1992). The close relationship of these two Neoproterozoic successions is also suggested by chemostratigraphic analysis (Knoll et al., 1986) and microfossil assemblages (Vidal, 1985; Green et al., 1989; Swett and Knoll, 1989). It therefore seems reasonable to consider that the two successions formed within one or a series of connected ensialic basins and that eastern Svalbard was positioned very close to central East Greenland during the Neoproterozoic (Hambrey, 1989). It is evident that both successions reflect major, and probably long-lived, extension-related subsidence. Problems with this interpretation arise from the anticipated strain-strengthening of the lithosphere at slow strain rates, and these have been discussed by Soper (1994).
Sandelin et al. (2001) undertook a detailed study of the lower part of the Murchisonfjorden Supergroup in Nord austlandet (NE Svalbard). Significantly, the lowest unit of the Murchison-fjorden Supergroup, the Galtedalen Group, has a well-defined base and well-developed basal conglomerates, and it rests with major unconformity on Mesoproterozoic sediments intruded by 950 Ma granites. The overlying part of the Galtedalen Group and the succeeding Franklinsundet Group make up alternating units of quartzite and shale with subordinate carbonates (Sandelin et al., 2001). This succession, which is equivalent to the lower part of the Lomfjorden Supergroup in Ny Friesland (Harland, 1997; Sandelin et al., 2001), is almost certainly a correlative of the Nathorst Land Group in which the basal relationships with Krummedal equivalents are preserved, although precise correlations are not yet possible.
In North Greenland, although the depositional evolution of the Hekla Sund Basin resembles the Eleonore Bay Basin, the tectonic setting is entirely different, since at least the early history of the Hekla Sund Basin is closely linked to extensional rifting (Rivieradal Group and lower part of Hagen Fjord Group). A correlation between the postrift carbonate-dominated succession (upper part of Hagen Fjord Group) of the Hekla Sund Basin and the upper part of the Eleonore Bay Basin (Ymer Ø and Andrée Land Groups) suggests the presence of a very extensive carbonate platform that testifies to a period of regional tectonic stability prior to the onset of Iapetus rifting during the late Neoproterozoic. It is possible that the extent of this carbonate platform was even larger than suggested by the Arctic localities, since it may have extended into the Caledonides of Scotland and Ireland, where a similar succession is present (Appin Group and Port Askaig tillite; see following).
Within the Scottish Highlands, the siliciclastic sediments of the Moine Supergroup are entirely allochthonous, and they have a depositional age that is well constrained between the youngest detrital zircon ages at 979 ± 77 Ma and 926 ± 68 Ma, and the intrusion of the West Highland granitic gneiss at 870 Ma (Millar, 1999; Cawood et al., 2003). The Moine Supergroup was deposited in a marine environment, but it has often been correlated with the terrestrial, fluvially dominated, sediments of the Torridon Group, which is present largely in the foreland and in large-scale thrust sheets below the Moine thrust. The youngest detrital zircons in the Torridon Group have an age of 1060 ± 18 Ma (Rainbird et al., 2001), which is in agreement with the Rb-Sr whole-rock age of 977 ± 39 Ma that dates diagenetic mineral growth (Turnbull et al., 1996). A.D. Stewart has long favored a rift origin for the Torridon Group (see Stewart  for review), although this has been countered by Prave (1999), who interpreted the Torridon Group as the molasse of the Grenville orogen and the Moine Supergroup as its distal marine equivalent, thus reviving the correlation of Peach et al. (1907). More recent work has noted, however, that although the Torridon Group and Moine Supergroup were deposited at broadly the same time, there are relatively low proportions of Archean and Ketilidian detrital grains in the Moine Supergroup compared to the Torridon Group (Cawood et al., 2003; Friend et al., 2003), and that they differ in other aspects of sediment provenance too, suggesting that the two units were deposited in coeval but spatially discrete basins. The age distribution of detrital zircon grains from the Moine nappe (Friend et al., 2003, their Fig. 3a) is strikingly similar to that of grains from the Krummedal succession (Watt et al., 2000, their Fig. 7a). Although, as noted already, the profile of detrital zircon ages from the Nathorst Land Group is also similar to that from the Krummedal succession and the Moine Supergroup (Watt et al., 2000; Cawood et al., 2001), we do not support a correlation on the grounds of the structural, metamorphic, and sedimentologic context of the Nathorst Land Group.
In the Grampian Highlands of Scotland, to the south of the Great Glen fault, the oldest part of the Dalradian Basin, the Grampian Group, is interpreted as the fill of several NE-SW–trending rift basins (Glover et al., 1995; Robertson and Smith, 1999), and this group rests on basement of enigmatic affinity, referred to as the Dava and Glen Banchor successions. The depositional age of the sub–Grampian Group basement is constrained by the youngest detrital grain at 900 Ma, metamorphism at 840 Ma, and ductile shearing at 806 Ma (Noble at al., 1996; Highton et al., 1999; Cawood et al., 2003). The depositional age of this basement is thus similar to the ages for the Moine and Krummedal successions (Cawood et al., 2003), and the initiation of Dal radian deposition must postdate ca. 800 Ma. Interestingly, however, the Grampian Group and the underlying basement share similar detrital zircon profiles (Cawood et al., 2003; Banks et al., 2007), despite an undoubted depositional break (M. Smith et al., 1999). This is very similar to the situation regarding the zircon profiles of the Krummedal succession and Nathorst Land Group and lends weight to the argument that detrital zircon distributions cannot be used to infer continuous deposition between the two Greenland units (contra Watt and Thrane, 2001).
The presence of the 800 Ma Knoydartian event in the Moine Supergroup (Rogers et al., 1998; Vance et al., 1998; Millar, 1999) and in sub–Grampian Group basement (Highton et al., 1999) and its absence elsewhere on the Laurentian margin introduce the possibility that these rocks represent an exotic terrane, and that Scandian thrusting was a final, minor event in comparison with earlier strike-slip juxtaposition, as suggested by Bluck et al. (1997). Although this hypothesis retains some currency (see Oliver  for review), the detrital zircon data sets of Cawood et al. (2003, 2004), Friend et al. (2003), and Banks et al. (2007) suggest similar provenance for the Moine Supergroup, sub–Grampian Group basement, and Krummedal successions. This indicates that these basin fills were contiguous, if not actually part of the same basin, during the Neoproterozoic. Nevertheless, the Knoydartian event is not represented in East Greenland, and the localized nature of this event remains an enigma.
The Grampian Group of the Central Highlands of Scotland consists of ∼8 km of metasediments, including fluvial fan-delta conglomerate and psammite, which pass upward into submarine-fan psammite-semipelite couplets, followed by shallow-marine arkosic sediments (Glover et al., 1995; Banks and Winchester, 2004). In places, there is a passage up into the Appin Group, but elsewhere, the Appin Group is transgressive across block-tilted Grampian Group strata. It is noteworthy that the transition between the Grampian and Appin Groups has been interpreted as onlap onto an active rift shoulder, such that a conformable relationship pertains away from the margins, but the Appin Group rests unconformably on older Grampian Group sediments on the rift shoulder (Robertson and Smith, 1999, their Fig. 5; M. Smith et al., 1999). This relationship has clear parallels with the observed relationships between the Nathorst Land and the Lyell Land Groups of the Eleonore Bay Supergroup. The Appin Group is a generally upward-shallowing sequence, characterized by phyllites and black shales with wedges of quartzite that pass up into subtidal carbonates and the Port Askaig tillites. As with the Tillite Group in East Greenland, the age of the tillites is clearly critical to determining the upper limit for the depositional age of the Grampian–Appin interval. However, there is currently a lack of consensus over the age and correlation of the Port Askaig tillite; some authors (Brasier and Shields, 2000; Condon and Prave, 2000) favor a Sturtian (∼700 Ma) age, and others (Halverson et al., 2005) prefer a Marinoan correlation.
The thick siliciclastic succession overlying the Port Askaig tillite is assigned to the Argyll and Southern Highland Groups. In contrast to a maximum preserved thickness of ∼340 m between the Ulvesø Formation and the base of the Cambrian Kløftelv Formation in East Greenland (Hambrey and Spencer, 1987), there are up to 9 km of Argyll Group sediments between the Port Askaig tillite and the 600 Ma Tayvallich volcanics, and a further 4 km of Southern Highland Group sediments. Regardless of arguments concerning the correlation of the upper parts of the Southern Highland Group (see Trewin  for reviews), there is general consensus that the group reaches at least into the Lower Cambrian. This correlation points to a significant contrast in subsidence histories for the Dalradian and Eleonore Bay Basins during the later part of the Neoproterozoic.
The Neoproterozoic successions in the Hekla Sund and Eleonore Bay Basins probably formed in response to an early pulse of Iapetan rifting at around 760–700 Ma. The deposits are well exposed in thrust sheets of the East Greenland Caledonides, and both basins attain thicknesses in excess of 10 km. Their lower parts show strong tectonic overprint and metamorphic recrystallization, which impede sedimentologic studies, but sedimentary structures are generally well preserved in the upper parts, making detailed sedimentologic interpretations possible.
The Hekla Sund Basin in eastern North Greenland reflects deposition during active rifting in two half-grabens. The rift phase is represented by the Rivieradal Group, which is only developed in the eastern half-graben, and the much thinner lower part of the Hagen Ford Group in the western half-graben. The lower part of the up to 10-km-thick Rivieradal Group is composed of a thick succession of phyllite, quartzite, and marble that is exposed toward the east. This succession may in part be equivalent to the upper, less-deformed western part that includes conglomerates and thick-bedded sandstone turbidites and mudstones deposited in a fan-delta and deep-marine basin plain setting. The uppermost part of the group reflects siliciclastic deposition in shallow-marine and fluvial environments. Major point sources of sediment input are indicated by the localized presence of thick units of channelized debris-flow deposits. In the western half-graben, the rift-phase deposits are only ∼500 m thick, and they are represented by fluvial to shallow-marine deposits. The postrift, thermal subsidence phase of the Hekla Sund Basin is represented by the upper part of the Hagen Fjord Group (750 m), which is recognized in both half-graben segments of the basin. This phase is characterized by a change from shallow-marine siliciclastic to carbonate-platform deposition. The top of the Hagen Fjord Group is marked by a paleokarst development, reflecting a major hiatus across which much of the later Neoproterozoic is absent.
The deposits of the Eleonore Bay Basin of East Greenland attain a combined thickness of almost 15 km. Major regional stratigraphic breaks seem to be absent, as is evidence of rift-related sedimentation, and deposition within one or a series of connected ensialic basins is thus envisaged. Four stages of basin evolution have been recognized in the more than 14-km-thick Eleonore Bay Supergroup; these stages reflect major paleogeographic reorganizations of the shelf and, in the upper part of the succession, a change from shallow-marine siliciclastic to carbonate-platform deposition prior to the onset of Marinoan glacigenic deposition. Although neither continental deposits nor horizons of fluvial incision have been recognized in the Eleonore Bay Basin succession, at least 10 second-order sequence boundaries have been recognized. The widespread nature of these major lithostratigraphic units and their facies suggest that depositional conditions were uniform and parallel to basin strike for most of the duration of deposition.
Glacigenic deposition in the Eleonore Bay Basin is recorded by the Tillite Group, which is up to 750 m thick and overlies the Eleonore Bay Supergroup. The Tillite Group consist of two formations of diamictite separated by sediments deposited in a marine environment. A conspicuous negative δ13C anomaly (the “Trezona anomaly”) has been recognized beneath the lowest tillite in the Eleonore Bay Supergroup, indicating a Marinoan age for the tillites. The upper tillite is overlain by a cap car bonate that marks the base of the Ediacaran and consists of peritidal carbonate passing into outer-shelf mudstone. These deposits are locally overlain by playa lake sediments that form the top of the Tillite Group. An angular unconformity separates the Eleonore Bay Basin succession from overlying Cambrian deposits.
Pre-Iapetus basins along the Laurentian margin that are possibly coeval with those in the Greenland Caledonides have been recognized in Scotland and Ireland (represented by the Dalradian Supergroup, the Moine Supergroup, and the Sleat, Stoer, and Torridon Groups), and on Svalbard (represented by the Hekla Hoek succession). Of these, the Eleonore Bay Basin and the Hekla Hoek successions show the greatest similarity. Both successions show much the same sedimentary evolution from siliciclastic to carbonate deposition, capped by Marinoan tillite, and they probably reflect long-lived, extension-related subsidence.
The Grampian and Appin Groups of the Central Highlands of Scotland also record a change from siliciclastic to carbonate deposition that in many ways resembles the evolution observed within the Eleonore Bay Supergroup. They are overlain by the Port Askaig tillite, which has been attributed both a Sturtian (∼700 Ma) and a Marinoan (∼635 Ma) age, demonstrating that good dating of the glacigenic deposits is critical in correlating the successions of Scotland and East Greenland.
The authors thank the reviewers Christopher Banks and Michael J. Hambrey for their valuable comments and suggestions, which greatly benefited this manuscript. The paper is dedicated to the memory of Steve Robertson, who died shortly after completing his work on the Nathorst Land Group.
This paper is published with the permission of the Geological Survey of Denmark and Greenland. Field work in North Greenland was partly funded by the Carlsberg Foundation (grant No. 93-0254/10) and Maersk Oil and Gas.
Figures & Tables
The Greenland Caledonides: Evolution of the Northeast Margin of Laurentia
- ancient ice ages
- Arctic region
- carbonate platforms
- clastic rocks
- depositional environment
- East Greenland
- glacial environment
- glacial geology
- marine environment
- plate tectonics
- sedimentary basins
- sedimentary rocks
- sequence stratigraphy
- shelf environment
- upper Precambrian
- Tillite Group
- northeastern Greenland
- Hagen Fjord Group
- Eleonore Bay
- Eleonore Bay Supergroup
- Rivieradal Group
- Hekla Sund