Neoproterozoic glacial strata in Oman are key to the ongoing Snowball Earth discussion, providing a great opportunity to test the hypothesis. The Abu Mahara Group (Huqf Supergroup) is well exposed in the core of the Jabal Akhdar of northern Oman. It contains two glaciogenic units, the Ghubrah Formation (723+16/−10 Ma) and the Fiq Formation (currently undated), that are separated by the volcaniclastic Saqlah Formation. An angular unconformity is present between the Ghubrah and Saqlah formations, indicating a significant time gap between the deposition of the Ghubrah and Fiq glacial successions. The localised occurrence of pillow basalts and more widespread volcaniclastics of the Saqlah Formation, suggests the initiation of a rifting phase, which is considered to have continued during deposition of the Fiq Formation. Given the available geochronology, the Ghubrah Formation may correlate with other glaciogenic successions worldwide attributed to the Sturtian glacial epoch, and the Fiq Formation with younger glaciogenic successions attributed to the Marinoan glacial epoch. Neoproterozoic glaciations appear to have taken place at times of tectonically generated accommodation, suggesting a link between geodynamics, basin development and climate change.

The Neoproterozoic Era was characterized by extreme climatic oscillations with extensive glaciations that may have reached intertropical and equatorial latitudes, perhaps covering the whole Earth in ice (‘Snowball Earth’ events, Hoffman et al., 1998). The geological record of these glacial epochs is dominated by glaciogenic diamictites that are typically overlain by carbonate rocks (‘cap carbonates’) bearing a distinctive negative δ13C signature (Hoffman et al., 1998; Hoffman and Schrag, 2002). The exact number of Neoproterozoic glacial epochs is the subject of ongoing discussions. The clustering of geochronological dates and excursions in the composite δ13C reference curve suggest three main periods of glaciation (Table 1; Kaufman and Knoll, 1995; Jacobsen and Kaufman, 1999; Prave, 1999; Knoll, 2000; Hoffman and Schrag, 2002; Corsetti and Kaufman, 2003; Fanning and Link, 2004; Hoffmann et al., 2004; Xiao et al., 2004). These epochs are named ‘Sturtian’ (c. 740–700 Ma), ‘Marinoan’ (c. 650–630 Ma), and ‘Varangerian’ (c. 590–580 Ma) (Halverson et al., in press).

The Neoproterozoic Abu Mahara Group (lower part of the Huqf Supergroup) of Oman (Table 1) contains the glacially influenced Ghubrah and overlying Fiq formations, which are separated by volcanics and volcaniclastics of the Saqlah Formation. The Fiq glacial succession is capped by a regionally trangressive carbonate (Hadash Formation) that has been interpreted as a postglacial ‘cap carbonate’ (Leather et al., 2002 and Allen et al., 2004). Correlation of these glaciogenic successions is only partly resolved. Brasier et al. (2000), Leather (2001), Leather et al. (2002), Allen et al. (2002) and Allen et al. (2004) proposed that the Ghubrah and Fiq formations correlate with the Sturtian and Marinoan glacial epochs, respectively. Published radiometric dates support a Sturtian-equivalent age for the Ghubrah Formation (723+16/-10 Ma; Brasier et al., 2000), but no dates are currently available for the Fiq Formation. As such, the Fiq Formation could potentially represent Sturtian, Marinoan or Varangerian glacial deposits.

In order to resolve this issue, a detailed sedimentological study of the relationships among top Ghubrah, Saqlah and Fiq units was initiated. This paper reports the results of a field-based study, which is particularly focussed on the boundary between the Ghubrah and the overlying Saqlah formations, that in turn helps to resolve whether one or two Neoproterozoic glacial successions occur in Oman. Additionally, the study of the field relationships between the volcanic Saqlah Formation and the glaciogenic Fiq Formation, clarifies the tectono-stratigraphic setting of this part of Oman Neoproterozoic stratigraphy.

The Neoproterozoic Huqf Supergroup of Oman crops out mainly in northern (Jabal Akhdar; Figure 1), central (Huqf) and southern (Mirbat) areas, and is subdivided into three groups; from base up: (1) Abu Mahara, (2) Nafun and (3) Ara (Table 1; Glennie et al., 1974; Gorin et al., 1982; Hughes-Clarke, 1988; Rabu et al., 1993; Loosveld et al., 1996). This study concentrates on the Jabal Akhdar region where the most complete and geochronologically constrained successions are preserved. Here, the top of the Huqf Supergroup is calibrated by radiogenic dates from the middle of the Fara Formation (Ara Group) at 544.5 ± 3.3 Ma (Brasier et al., 2000). This age is consistent with the occurrence of the skeletal fossils Cloudina and Namacalathus in correlative subsurface rocks (Amthor et al., 2003). Recent radiogenic dates obtained from core material in the South Oman Salt Basin have been used to constrain the age of the Cambrian/Precambrian boundary at 542 ± 0.3 Ma, when Cloudina and Namacalathus become extinct (Amthor et al., 2003). This age is now used as the reference for the Cambrian/Precambrian boundary (Table 1).

The Abu Mahara Group contains the two glaciogenic Ghubrah and Fiq formations, separated by the volcanics and volcaniclastics of the Saqlah Formation. In the original nomenclature (Beurrier, 1987; Rabu, 1988), the Fiq and Saqlah formations were grouped in the Ghadir Manqil Formation. This study assigns more importance to the Saqlah and Fiq in Oman’s Neoproterozoic stratigraphy, and recommends that the Fiq and Saqlah be raised to the status of formations (Table 1). Date for the Ghubrah Formation includes 723+16/−10 Ma (Brasier et al., 2000). The Ghubrah Formation is thus considered a representative of the Sturtian glacial epoch (Brasier et al., 2000).

At the base of the Nafun Group, the Hadash Formation (Leather, 2001; Cozzi and Al-Siyabi, 2004; Allen et al., 2004) represents a trangressive cap carbonate that directly overlies and oversteps the Fiq Formation of the Abu Mahara Group. The Hadash Formation is overlain by two siliciclastic-carbonate depositional cycles; from base up: (1) Masirah Bay-Khufai formations, and (2) Shuram-Buah formations.

In the Huqf region, the Abu Mahara Group is thin (20 m) and does not contain glaciogenic deposits. The volcaniclastic Halfayn Formation, dated at c. 802 Ma (Leather, 2001), overlies granodioritic basement dated at c. 822–825 Ma (Leather, 2001). The base of the overlying Nafun Group is marked by the occurrence of a laterally correlative carbonate unit with the Hadash Formation, representing a regional transgressive event (Leather, 2001; Cozzi and Al-Siyabi, 2004).

Neoproterozoic glaciogenic successions also crop out in the Dhofar region of south Oman (Figure 1), where the Mirbat Sandstone Formation (Platel et al., 1992) overlies granodioritic bodies dated at c. 706 ± 40 Ma (Rb-Sr, Gass et al., 1990) and the small Leger Granite pluton (723 ± 2 Ma; U-Pb, Kellerhals and Matter, 2003) intruded into older-than 800 Ma gneissic basement. The Mirbat Sandstone Formation comprises three members; from base up: (1) glaciogenic Ayn (Kellerhals and Matter, 2003); (2) deep-marine Arkahawl; and (3) shallow- to deep-marine Marsham. Correlation of the glaciogenic rocks of the Mirbat Sandstone Formation with those in the Jabal Akhdar is tentative (Cozzi and Al-Siyabi, 2004) and the subject of ongoing studies.

The present study focuses on the Jabal Akhdar region, where the best and most complete exposures of the Abu Mahara Group occur. Table 2 shows a summary of the major lithofacies present within each formation. Particular attention is paid to the nature of the contacts between the Ghubrah, Saqlah and Fiq formations.

Ghubrah Formation

The rocks comprising the Ghubrah Formation were first recognized and defined by Kapp and Llewellyn (1965) as the ‘Mistal conglomerates’. The stratotype is located in Wadi Hajir (Jabal Akhdar; Figure 1). The Ghubrah Formation was defined following the work of Rabu (1988) and Beurrier (1987), who subdivided the Abu Mahara Group into the Ghubrah Formation and Saqlah, Fiq and Amq members of the Ghadir Manqil Formation. The Amq Member has subsequently been reclassified as the Hadash Formation (McCarron, 2000) and incorporated into the Nafun Group (Table 1, Leather, 2001; Cozzi and Al-Siyabi, 2004). Rabu (1988) estimated the Ghubrah thickness to be in the order of 300 m; however, the true thickness is unknown since the base of the Formation is not exposed in outcrop and the unit is tectonically deformed. The true thickness is likely to be considerably greater.

In Wadi Mistal (Figure 1), the Ghubrah Formation is dominated by diamictite lithofacies (Gd, Table 2; Brasier et al., 2000) characterized by: (1) poor stratification; (2) unsorted clasts of diverse size and lithology (crystalline and metamorphic rocks, acid and basic volcanites and sedimentary rocks) some of which bear striations and are interpreted as lonestones and dropstones; and (3) unsorted silty-shaly or sandy matrix (Rabu, 1988; Leather, 2001). Within the diamictites, massive and graded light-coloured siltstone (Gs) occurs as units up to 10 m thick. Rare carbonate beds also occur (Leather, 2001). A glaciogenic origin was first suggested by Tschopp (1967) and Glennie et al. (1974), and later confirmed by Brasier et al. (2000). The diamictite lithofacies (Gd) is interpreted as ice-rafted in origin, deposited together with suspension-fallout in a subaqueous environment relatively distal from glacier margins (Rabu, 1988; Leather, 2001). Siltstone units (Gs) were most likely deposited in a marine environment during a time of reduced influence from ice-rafting.

Saqlah Formation

The overlying Saqlah Formation is less than 100 m thick, and occurs throughout the Jabal Akhdar region. The Saqlah Formation comprises basalt (Sb, Table 2) up to 60 m thick (Rabu, 1988), which is associated with volcaniclastics and lithic sandstones (Sv) and carbonates (Sc) that occur above and below the basaltic unit. Leather (2001) attributed only the basalt to the Saqlah Formation.

Geochemically, the basalt of the Saqlah Formation has a composition between alkali basalts and trachy-andesites (Winchester and Floyd, 1977; Rabu, 1988). The magmatic activity was principally effusive and occasionally hypovolcanic. The basalt pillows suggest subaqueous eruptions. The basaltic unit is laterally discontinuous over tens to hundreds of metres and its thickness increases toward the south (Saiq Plateau; Figure 1b) where a few effusive basalts up to 60 m thick are present. In the north (Wadi Mistal), only one effusive deposit is locally present and it is less than 25 m thick (Rabu, 1988; Leather, 2001). The Saqlah Formation is also found in the Saih Hatat ‘window’ to the east of the Jabal Akhdar (Figure 1b), where a tholeiitic magma series substitutes for the alkaline basalt, attesting to a higher amount of crustal stretching (Le Métour, 1988). The Saqlah Formation basalt is characteristic of intra-continental extension (Pearce and Cann, 1973; Meschede, 1986; Rabu, 1988).

Rabu (1988) interpreted the Saqlah to conformably overlie the Ghubrah Formation; or where the Saqlah is absent, that the Ghubrah passes transitionally into the overlying Fiq Formation. Brasier et al. (2000) and Leather (2001), however, considered the possibility of an unconformity between the Ghubrah and the Saqlah formations, with an upward transition into the Fiq Formation. Leather (2001) incorporated the volcaniclastic unit overlying the Saqlah basalts within the basal Fiq Formation; we propose that the volcaniclastics be assigned to the Saqlah Formation. Throughout the Saqlah succession, the dominance of volcanic and paravolcanic products is evident, although some of the volcanic material is reworked.

Fiq Formation

The Fiq Formation crops out ubiquitously in the Jabal Akhdar region with considerable lithofacies variation. The Fiq Formation lies conformably on the volcaniclastic and siliciclastic deposits of the Saqlah Formation and it is truncated at the top by an angular unconformity at the base of the Permian Saiq Formation. The total thickness of the Fiq Formation is estimated at 1.5 km. Radiometric dates are currently unavailable for the Fiq Formation, though Brasier et al. (2000), Leather (2001), Allen et al. (2002) and Allen et al. (2004) suggested it is most likely to be a Marinoan-equivalent (c. 650–630 Ma). Rabu (1988) also assigned a Vendian age based on the presence of the cyanophycae palaeocryptidium cayeuxi (Chauvel and Mansuy, 1983).

Rabu (1988) described the Fiq as immature greywacke sandstones with crystalline and/or metamorphic, volcanic and sedimentary clasts. Subsequent studies by Leather (2001), Leather et al. (2002) and Allen et al. (2004) divided the Fiq Formation into 2 facies associations: (1) proximal and distal glaciomarine (Fd); and (2) non-glacial sediment gravity flow (Fg) and shallow marine (Fm). The characteristic lithofacies (Fd) is a massive diamictite, forming bodies up to 30 m thick and laterally persistent over hundreds of metres. Clast abundance varies from less than 5% to less than 40%, depending on the proximality of the facies. Clasts range in size from 10 mm to 2 m, and include faceted and striated samples, locally forming dump structures.

Rabu (1988) proposed a glacial origin for the Fiq suggesting that it was deposited in a rift basin tectonic setting. The glacial origin was later confirmed by Leather (2001), Leather et al. (2002) and Allen et al. (2004) who interpreted the massive diamictite lithofacies as ice-rafted rain-out facies and suspension-fallout deposits in a deep-water setting. Leather (2001) interpreted the Fiq Formation as consisting of seven stratigraphic cycles originating by growth and recession of ice sheets. The youngest diamictite is abruptly overlain by the transgressive Hadash Formation. The nature of the volcanics of the Saqlah Formation, and the large thickness differences observed in the sedimentary rocks of the Fiq Formation, suggest that they were both deposited in a large graben or half-graben structure c. 80 km across (in the E-W direction) (Brasier et al., 2000; Leather, 2001; Leather et al., 2002 and Allen et al., 2004). This interpretation is supported by similar-scale graben structures observed in the subsurface (Loosveld et al., 1996).

Continuous Abu Mahara Group sections crop out in two areas of the Jabal Akhdar (Figure 1b): (1) the central part of the Jabal Akhdar (Wadi Mistal); and (2) along the southern flank of the Jabal Akhdar (Wadi Mu’aydin of the Saiq ‘window’). Tracing lateral continuity of stratigraphy at Wadi Mu’aydin is difficult due to tectonic deformation; accordingly, the present study concentrates on Wadi Mistal where the Abu Mahara Group crops out in an erosive ‘bowl’ (Figure 1c). On the western side of Wadi Mistal, the Abu Mahara Group is locally deformed by thrusts, with strong deformation fabrics and repetitive stacking of depositional units. However, the eastern side of Wadi Mistal presents sections that manifest lateral facies changes over several kilometres distance. The Ghubrah Formation crops out mostly in the centre of the wadi, while the Saqlah and Fiq formations occupy the wadi flanks (Figure 1c). Two key localities are considered here (Figure 1c): Locality 1, in the northeast of the Mistal ‘bowl’; and Locality 2 on the southeast side of the ‘bowl’, where the Saqlah Formation attains its maximum thickness. Particular attention was paid to the boundary relationships between the formations. Further details regarding the sedimentology of the Abu Mahara Group in Wadi Mistal can be found in Leather (2001) and Allen et al. (2004).

Locality 1: Northeast Wadi Mistal

At this locality, the top Ghubrah, Saqlah and lower Fiq formations crop out continuously over several kilometres distance, allowing lateral changes in lithofacies and key stratigraphic surfaces to be traced. Seven stratigraphic sections along a NW/SE profile were measured (sections 14, 1, 2, 3, 4, 13 and 12; Figure 2).

In most of these localities, the basalt of the Saqlah Formation is pillowed (pillows less than 1.5 m in size, Figure 3c) and is composed of microlithic lava with a spilitised andesitic basaltic composition (Rabu, 1988). The basalt is vesicular, with 10-mm-scale vesicles. A few brownish dolomitic carbonate beds (less than 0.3 m thick) with planar lamination are present above and below the basalt (Sc). These are usually weakly folded and laterally discontinuous over tens to hundreds of metres. Carbon isotope measurements for the carbonate bands show widespread values ranging from +4.5 to −10‰ δ13C VPDB (Table 3). Although occupying the same stratigraphic position, the beds cannot be laterally correlated using δ13C values. Such large lateral variations in δ13C may have resulted from the large chemical variations in the carbon pool in proximity of volcanic centres.

In section 14, the top of the Ghubrah Formation comprises massive glacial diamictite (Gd) sharply overlain by the tuffaceous siltstone and brownish dolomitic carbonates of the Saqlah Formation (Sv, Sc). Locally, a pillowed basalt (Sb) showing large lateral thickness variations (5–10 m), cuts down into the Ghubrah and base of the Saqlah formations. The basalt is overlain by thin tuffaceous siltstones and channelised volcaniclastic sandstones (Sv, Figure 3b) that pass upwards into massive diamictite (Fd) of the Fiq Formation. The Fiq diamictite contains abundant dropstones, some of which exceed one metre in size.

Section 1 does not show a direct Saqlah/Ghubrah contact. However, the Ghubrah Formation outcrops show metre-scale beds of massive diamictite interbedded with centimetre-scale yellowish beds of sandstone and siltstone. The Saqlah/Ghubrah boundary is an angular unconformity (20–30°), also observed in section 2. Both sections 2 and 3 show the same Saqlah Formation facies. The basal few metres comprise tuffaceous siltstones (Sv) with few basaltic pillows and one 0.4-m-thick dolomite bed (section 1 only). The basalt (Sv) is continuous between these locations and thickens towards the southeast, while the overlying massive volcaniclastic sandstones (Sv) become thinner. The basalt is overlain by graded m-scale sandstone beds with white angular volcanic clasts, interbedded with thin tuffaceous siltstones and carbonate bands (Figure 3a). Above the lower 20 m, the volcanic influence diminishes considerably, the strata becoming dominated by thinner sandstone beds alternating with m-scale beds of siltstone. This is defined as the Fiq/Saqlah boundary (Figure 4), although the first diamictite appears 150 m higher in the succession. Consequently, in the north, the boundary appears above a cliff-forming unit of thickly-bedded lithic (volcaniclastic) sandstones (Figure 4). Localities 1 to 4 are characterised by the same gradational Fiq/Saqlah boundary (Figures 2, 5 and 6). The Fiq/Saqlah contacts are conformable, precluding the presence of an unconformity at this stratigraphic level.

A few kilometres to the southeast, pillowed basalt is absent in section 4 (Figure 2), and the Saqlah Formation is very similar to the upper part of sections 1 to 3. The best evidence for an angular unconformity between the Ghubrah and Saqlah occurs in section 4, where the angular discordance reaches 40° (Figure 6).

The final two sections (12 and 13, Figure 2) contain characteristic reddish tuffaceous siltstone layers interbedded with lithic sandstones dominated by volcanic fragments. In section 13, a laterally discontinuous 15-m-thick pillowed basalt is present. In sections 12 and 13, the top of the Ghubrah Formation is not well exposed, but regional bedding measurements show a 10–20° angular unconformity between the Ghubrah and Saqlah formations. The Fiq/Saqlah boundary is conformable and relatively sharp, and placed between the volcaniclastic sandstones and reddish siltstones of the Saqlah Formation and the greenish-yellowish Fiq diamictites.

Locality 2: Southeast Wadi Mistal

In the southeast of Wadi Mistal, laterally continuous outcrops of the Abu Mahara Group are present (section 7, 8 and 10; Figures 7 and 8). The top of the Ghubrah Formation comprises massive diamictite (Gd) with decimetre-scale dropstones, locally interbedded with thin cm-thick yellowish siltstone bands (Gs). The siltstones commonly contain centimetre-scale angular volcanic clasts. Slump and channel structures are common. This volcaniclastic and clastic lithofacies is interpreted as reworked volcanic deposits associated with successive turbidity and sediment gravity flows, probably initiated on local slopes created by nearby volcanoes or on rift basin margins. Along these profiles, the Saqlah/Ghubrah boundary is not well exposed, but the bedding measurements confirm an angular unconformity between the units, reaching a maximum angular discordance of 70°.

The Saqlah Formation is thicker in this part of the Mistal ‘bowl’ (Figure 8). Locally, the base of the Saqlah is marked by a few decimetre-scale carbonate beds (Sc), with few tens of metres of lateral extent (sections 7 and 8). The overlying 15-m-thick pillowed basalt is laterally discontinuous over a 100 m distance (section 7), followed by a white unit of tuffaceous siltstones interbedded with sandstones and conglomerates. This latter unit, mainly turbiditic in origin, shows variable bedding orientations (fan-shaped) and locally a small (less than 10o) angular discordance with the overlying reddish unit consisting of tuffaceous siltstones interbedded with channelised sandstones. The contact with the overlying Fiq Formation is relatively sharp and easily discernible between the uppermost reddish siltstones of the Saqlah Formation and the lowermost greenish diamictites (Fd) of the Fiq Formation (Figure 8). This boundary is conformable and appears above alternations of thin siltstone and sandstone, where the contact is emphasized by a sharp change in colour between the white and reddish Saqlah Formation and grey and brown Fiq Formation (Figure 9).

The Saqlah Formation unconformably overlies the tectonically deformed Ghubrah Formation, with up to 70° of angular discordance. The presence of this unconformity is confirmed by the fact that the Ghubrah, Saqlah and Fiq share a pervasive, late tectonic fabric, whereas the Ghubrah Formation was also deformed prior to the emplacement and deposition of the Saqlah Formation. The possibility of a tectonic contact between the Saqlah and the underlying Ghubrah Formation can be discarded as no fault plane and associated brittle deformation was found in the vicinity of the contact. Moreover, ongoing studies in Wadi Bani Jabir (East Jabal Nakhl; Figure 1b), which are interpreted as having been a rift shoulder during Fiq deposition (Leather, 2001; Allen et al., 2004), show the presence of an extremely thin (5 m) Fiq Formation overlying 400 metres, and possibly more, of non-glaciogenic immature sandstones (named Jabir Formation by Leather, 2001). The stratigraphic position of the Jabir Formation is still uncertain, as it could pre- or post-date the Ghubrah Formation. A possible Jabir Formation equivalent crops out in the Saiq Plateau (southern flank of the Jabal Akhdar anticline), where it is found underneath the glacial diamictites of the Ghubrah Formation. If these interpretations are correct, the unconformity below the Fiq Formation in Wadi Jabir truncates the whole Ghubrah Formation, confirming the presence of a major time gap between the Ghubrah and the Saqlah-Fiq package.

The Fiq/Saqlah boundary is commonly gradual in nature, and is effectively marked by the disappearance of volcaniclastic rocks and/or the appearance of characteristic glaciogenic dropstone lithofacies of the Fiq Formation.

The high lateral variability of bedding within the Saqlah and the presence of localised angular unconformities suggest either syntectonic growth of stratigraphy or volcanogenic tilting. Either interpretation is possible, since the volcanics of the Saqlah Formation indicate the beginning of intraplate extension (Le Métour, 1988; Rabu, 1988; Bell, 1993; Loosveld et al., 1996; Leather, 2001).

The Fiq Formation records large lateral variations in thickness, reflecting deposition during the synrift phase (Rabu, 1988; Leather et al., 2002 and Allen et al., 2004). In this model, the postrift phase would correspond to deposition of the Nafun Group, which is recorded throughout Oman during a time of regional subsidence (Figure 10).

The presence of the unconformity between the Ghubrah and Saqlah-Fiq package indicates two separate glacial successions. Since the available U-Pb dates support a Sturtian age for the Ghubrah, a Marinoan or Varangerian age for the Fiq Formation seems likely. This is based upon a threefold (Sturtian-Marinoan-Varangerian) glacial epoch timeframe (Halverson et al., in press). The Shuram Formation of the overlying Nafun Group bears a major negative carbon isotope excursion (Table 1; Burns and Matter, 1993; Le Guerroué et al., 2003; Cozzi and Al-Siyabi, 2004), which may correlate with the Varangerian ice age (Le Guerroué et al., 2003; Halverson et al., in press), implying a Marinoan age for the Fiq Formation.

Importance regarding the ‘snowball Earth’ theory

The two glaciogenic successions of the Ghubrah and Fiq formations can be correlated with Sturtian and Marinoan glacial epochs respectively, whereas the Shuram Formation of the Nafun Group may be a time-equivalent of the Varangerian glacial epoch (Le Guerroué et al., 2003; Cozzi and Al-Siyabi, 2004; Halverson et al., in press).

The rifting event during Saqlah-Fiq times provides a possible explanation for the glacial origin of the Fiq, with glaciers nucleating on rift shoulders and growth and recession of ice sheets and tidewater glaciers driving the observed cyclicity (Leather, 2001; Leather et al., 2002; Allen et al., 2004). A correlation of Neoproterozoic glaciation with rifting has been suggested by Young (1995), who noted a correspondence between the deposition of glacial successions on both sides of Laurentia and the break-up of Neoproterozoic supercontinents. Prave (1999) proposed two separate but superimposed rift events to explain the preservation of the Surprise and Wildrose glacial successions in the Neoproterozoic stratigraphy of the Death Valley region (USA). A similar tectonic driving mechanism has been recently proposed by Eyles and Januszczak (2004) for Neoproterozoic glaciations. They propose that diachronous rifting produced diachronous tectono-stratigraphic cycles rather than the globally synchronous cycles required by glacially-driven eustatic drawdown and flooding. Latitudinally restricted ice centres were located over domal, plume-related continental uplifts and over linear rift flank uplifts. The correspondence of volcanicity, basin development and glaciation at two levels in the Huqf Supergroup therefore supports some aspects of the ‘zipper-rift’ model of Eyles and Januszczak (2004).

However, preliminary estimates of palaeolatitudes from palaeomagnetic studies suggest a tropical position of Oman during the deposition of the Abu Mahara Group. Kempf et al. (2000) considered the Mirbat Formation to be deposited at 32° south. Recently, Kilner (2003) found a paleolatitude of 13° south for both the Mirbat and Fiq formations, which led him to correlate them to the Marinoan ice epoch. This palaeogeographic reconstruction fits into a larger scale plate scenario. Meert and Torsvik (2003) and Meert and Lieberman (2004) predicted a 20° south position of the Congo-Sao Francisco paleoplate (containing the Arabian Shield) at c. 750 Ma, followed by a drift of the Arabian Shield to an equatorial position at c. 580 Ma (Varangerian) and to 60–30° at c. 510 Ma. If these paleolatitudes are correct, they imply that Neoproterozoic glaciers reached intertropical latitudes.

There is, however, a problem regarding evidence for a global Varangerian glaciation. The potential candidate formations for such an event correlate with the largest negative δ13C shift of Earth’s history (Table 1) (Le Guerroué et al., 2003; Cozzi and Al-Siyabi, 2004; Halverson et al., in press); these are: (1) Shuram of Oman (Cozzi et al., 2004); (2) Wonoka of Australia (Calver, 2000); (3) Johnnie of California, USA (Corsetti and Kaufman, 2003); (4) Squantum of Massachusetts, USA (Thompson and Bowring, 2000); (5) Gaskiers of Newfoundland (Krogh et al., 1988); and (6) Hankalchough of Tienshan, China (Xiao et al., 2004). The Shuram Formation of Oman, like the Wonoka Formation of Australia, surprisingly contains neither glaciogenic deposits nor cap carbonate sequences; whereas the Johnnie Formation of California records a major eustatic fall but no glacial strata. Of all of the formations bearing the characteristic δ13C negative shift, only the Squantum, Gaskiers and Hankalchough formations contain glacial rocks (Thompson and Bowring, 2000; Xiao et al., 2004). A ‘Snowball Earth’-type glaciation thus seems unlikely for the Varangerian. Moreover, the observed negative δ13C excursion is greater in amplitude than that predicted by the snowball theory (Le Guerroué et al., 2003; Cozzi and Al-Siyabi, 2004; Halverson et al., 2004), calling for other mechanisms capable of producing such a shift.

Two glaciogenic successions, the Ghubrah and Fiq formations of the Huqf Supergroup of Oman, are separated by an angular unconformity representing a significant period of geological time. The older Ghubrah succession is dated at 712 Ma, and therefore can be correlated to the older than 700 Ma Sturtian glacial epoch. In the Jabal Akhdar, the unconformity is locally overlain by the volcanic and volcaniclastic Saqlah Formation, which in turn passes up transitionally into the glaciogenic Fiq Formation. The Saqlah-Fiq package is interpreted as a rift basin fill initiated by intraplate volcanism, and most likely correlates globally with a cluster of glacial events of Marinoan age.

The correspondence of volcanicity, basin development and glaciation invites a geodynamic-climatic linkage of these processes. We favour the location of ice centres over rift-flank uplifts at the same time as basin subsidence accommodated glaciogenic diamictites, gravity flows and non-glacial marine sediments.

The authors would like to thank the ETH Zürich and Petroleum Development Oman for supporting this work, two anonymous reviewers improved the quality of this paper. We are also grateful to James Etienne who reviewed an earlier version of the manuscript as well as Alex and Benjamin who helped during fieldwork. The design and drafting of the final graphics was by GeoArabia.


Erwan Le Guerroué is currently a PhD student at the Institute of Geology, ETH Zürich, Switzerland. He received a BSc in 2000 in Life, Earth and Planetary Sciences, and an MSc in Geology in 2002 from the University of Rennes 1, France. Erwan’s research interests focus on sedimentology, stratigraphy and chemostratigraphy of Neoproterozoic sections in the Sultanate of Oman.

Philip Allen obtained a Bachelor’s degree in Geology from the University of Wales, Aberystwyth, in 1974. He then worked for British Petroleum Co. prior to completing a PhD at Cambridge University in 1979. Philip spent two years as a post-doctoral scientist at the Geological Institute, University of Berne, before returning to the University of Wales in Cardiff as a Lecturer in 1981. He moved to the Earth Sciences Department at the University of Oxford in 1985 and was elected to the Chair of Geology and Mineralogy at Trinity College Dublin in 1996. He has been Full Professor of ‘Oberflächennahe Geosysteme’ (Sedimentary Systems) at the ETH-Zürich, Switzerland, since September 2001. Philip’s main research interests are in the intersection of physical surface systems, sediment transport and deposition, and tectonics. He has published advanced textbooks entitled Basin Analysis: Principles and Applications (First edition 1990; Second edition 2004), and Earth Surface Processes (1997). His current research projects include an investigation of environmental changes and basin development during the Neoproterozoic, principally based on data from the Sultanate of Oman; landscape evolution and sedimentation in tectonically active terrains such as the Basin and Range province of southwestern USA; the evolution and dynamics of mountain belts and foreland basin systems, particularly the Alps of western Europe.

Andrea Cozzi completed his undergraduate studies in Geology (cum laude) at the University of Trieste, Italy, in 1994. He received an MA and a PhD in Geology from the Johns Hopkins University in 1997 and 1999, respectively. Between 1999 and 2000, Andrea studied Precambrian carbonates in Oman as a post-doctoral fellow at Trinity College Dublin. The following year he worked as a Carbonate Geologist with Petroleum Development Oman. Andrea is currently an Oberassistent at the Institute of Geology, ETH Zürich, Switzerland. His research interests focus on stratigraphic and geochemical techniques for high-resolution correlation, the interplay of tectonics and sedimentation, carbonate sedimentology, cyclostratigraphy and sequence stratigraphy.