ABSTRACT

Improved methods of analysis and quantification of heavy mineral assemblages in Cambrian to Early Cretaceous sandstones of southwest Sinai have revealed successive changes in provenance that reflect both rejuvenation of the Arabian Shield and changes in the topographic configuration of the source area. Three mineral units have been identified in the Cambrian succession, at least three in the Carboniferous and three in the Cretaceous. It is predicted that the genetic units defined by these successive changes in mineralogy will be of regional extent and thus assist in elucidating the history of uplift of the Arabian-Nubian Shield and provide a better means of correlating sandstone units into adjacent areas. Variation in the abundance of apatite in the Cambrian succession is independent of provenance signature and is interpreted as reflecting alternating dry and humid climatic conditions.

INTRODUCTION

Previous studies on the heavy mineral stratigraphy of Palaeozoic and Mesozoic sandstones of the Sinai Peninsula and adjacent areas include those of Shukri and Said (1946), Shukri and El Ayouti (1953), Bender (1968), Lillich (1969), Weissbrod and Nachmias (1986), Nasir and Saddedin (1989), Amireh (1991, 1992, 1994) and Weissbrod and Bogoch (2007). The most comprehensive of these studies is that of Weissbrod and Nachmias (1986), with data presented from the Palaeozoic and Mesozoic successions of Sinai, Jordan and the Levant. In that study and in the previous studies, data were presented in the form of whole-assemblage percentage data, which, while providing an adequate qualitative measure of provenance character, also reflect the effects of hydraulic sorting, weathering-related dissolution, and diagenesis. As a consequence, the data acquired in all of these studies proved to be of limited use in stratigraphic applications (Weissbrod and Bogoch, 2007, p. 648).

The limitations inherent in percentage-based analysis can be largely overcome by restricting analysis to a narrow grain-size range and by comparing the relative abundances of minerals with similar hydraulic properties (Morton, 1985; Morton and Hallsworth, 1994). By using these alternative methods, the present study takes a fresh look at the evolution of sand provenance in southwestern Sinai. The samples used in this study were collected in 2001, 2005 and 2006 by two of the authors (M.F.S. and M.A.E.) from Wadi Nasib and Wadi Abu Thora, both in the Um Bogma area, from Gebel Musaba Salama about 30 km to the northwest of Um Bogma, and from a section 10 km NNW of Abu Durba (Figure 1).

GEOLOGICAL SETTING

Outcrops of Palaeozoic and Mesozoic sandstones in Sinai are largely restricted to a narrow zone that runs east–west across southern central Sinai, more or less along the line of latitude 29°N (Figure 1). To the south of these outcrops lies a largely mountainous terrain made up of Precambrian basement rocks, which constitutes the northernmost part of what was, in pre-Oligocene times, a unified Nubian–Arabian Shield. To the immediate north of the sandstone outcrop is an extensive plateau composed mainly of Upper Cretaceous and Eocene limestones. This outcrop pattern reflects the fundamental underlying structure of the Sinai Peninsula, in which the Phanerozoic sedimentary succession rests on a northward-dipping surface of Precambrian shield rocks (Jenkins, 1990). The overall northward thickening of successive Phanerozoic formations indicates that northern Sinai underwent relative subsidence through much of Phanerozoic time, while the shield area to the south underwent a progressive uplift. The shield area, including the basement and its Precambrian sedimentary cover, is thought by most workers to have supplied much, if not all, of the clastic fill of the northern Sinai basin.

Although the fundamental setting of sedimentation appears to have changed little during the Palaeozoic and Mesozoic, the sedimentary succession is discontinuous as a result of successive phases of regional tectonic activity. The Ordovician is probably missing and there is no evidence of Silurian or Devonian strata in the Sinai Peninsula (Klitzsch, 1990a, p. 396). The Early Palaeozoic is thus believed to be represented by the Cambrian alone. Overlying the Cambrian is a succession ranging in age from Carboniferous to Palaeogene, with substantial hiatuses in the Late Carboniferous (Hercynian structural event) and in the Late Jurassic to Early Cretaceous (eustatic sea-level fall) (Jenkins, 1990). Recent reviews of the stratigraphic succession are presented by Jenkins (1990), Klitzsch (1990), Kerdany and Cherif (1990), El-Hawat (1997), Issawi et al. (1999) and Tawadros (2001).

Since no single system of nomenclature has yet been established for the Sinai succession and since a regional review of the nomenclature is beyond the scope of this paper, discussion of the lithostratigraphic terminology is restricted to the individual sections that are the focus of this study. These are at Wadi Nasib (Lower Palaeozoic), Wadi Abu Thora (Carboniferous), Abu Durba (Carboniferous, Triassic and Cretaceous) and Gebel Musaba Salama (Cretaceous). The location of the four sections is shown in Figure 1. The lithostratigraphic nomenclature and ages applied to the units examined in this study are shown in Figures 2 and 3. These assignments are discussed below in relation to the individual sections. In the case of the Cambrian, we apply Late/Upper, Mid/Middle and Early/Lower as applied by previous authors, although it should be noted that these are not precisely related to the latest stages and series as applied by the ICS.

HEAVY MINERAL ANALYSIS

Although it is widely accepted that heavy minerals are valuable indicators of provenance, it is also apparent that the composition of an individual heavy mineral assemblage is dependent on many other factors (Morton and Hallsworth, 1999; Mange and Maurer, 1992; papers in Mange and Wright, 2007). These include chemical weathering (leading to removal of minerals susceptible to oxidation or acid dissolution), grain size availability (leading to the underrepresentation in coarse sands of minerals that typically occur as small crystals), transport (leading to hydraulic separation of minerals with differing densities and shape) and diagenesis (leading to removal of minerals susceptible to dissolution by the circulation of warm pore waters).

These processes can lead to detrital assemblages that are very different in character to those of their host rocks, with some minerals being absent and others present in very different proportions. Thus the processes of meteoric and diagenetic dissolution of unstable minerals can result in what were originally very different assemblages being reduced to an assemblage composed of relatively few ultrastable minerals (typically zircon, rutile and tourmaline and monazite). Conversely, assemblages derived from a single source rock can display substantial variation in composition because of the effects of grain size availability and hydraulic sorting. Unless highly distinctive minerals are present that can be related to a specific source rocks, this modification in the range and proportion can lead to considerable difficulty in provenance interpretation.

Because of these effects, comparison of heavy mineral assemblages using percentage data alone can be highly unreliable. An alternative approach was proposed by Morton (1985) who recommended the use of two-component indices, which reflect the relative abundance of minerals with similar hydraulic behaviour (i.e. similar density and shape characteristics). These two-component indices may relate to two different minerals or two varieties of the same minerals. A further constraint proposed by Morton (1985) was that heavy mineral analysis should be carried out on a single size fraction, thus eliminating the effects of grain size availability. The size fraction proposed was the very fine sand fraction (64–125 microns) because heavy minerals of this size can be obtained from sandstones of all grain sizes.

The four most reliable heavy mineral indices routinely used in provenance studies are the rutile:zircon, monazite:zircon, chrome-spinel: zircon, and apatite:tourmaline indices (Morton, 1985 – see Table 1 for formulae). These are favoured because of the resistance of all the component minerals to diagenetic dissolution. They are thus equally applicable to deeply buried sands as to sands that have undergone minimal burial. The scarcity of chrome spinel in the Sinai sandstones means that the chrome-spinel:tourmaline index is not used in this study.

For sands that have experienced little or no burial dissolution (as indicated by a lack of significant surface etching on unstable minerals), it is possible to use additional indices, such as the staurolite:tourmaline index (see Table 1). Though not subject to the testing undertaken by Morton (1985), this index is considered to provide a broad indication of the relative proportion of the component minerals in the source rocks. Additional indices used in this study are the pink (purple) zircon index (pZi), defined by the percentage of pink (purple) zircons in the total zircon assemblage, and the euhedral zircon index (eZi), defined by the percentage of euhedral zircons within the total zircon assemblage.

The fundamental premise of these mineral indices is that because they are more or less unaffected by the effects of hydraulic sorting, and thus provide a reasonably accurate reflection of the relative abundance of minerals in the source rock. The mineral indices thus provide the best available means of obtaining a numerical relationship between detrital assemblages and specific source rocks and also of making numerical comparison between the provenance character of different sandstone bodies.

The mineral indices presented in Table 1 are considered to be largely independent of the effects of hydraulic sorting and shallow burial-related dissolution, but one index, the apatite: tourmaline index (ATi), is strongly affected by weathering. This is because apatite readily dissolves in acidic meteoric groundwater, in either the source area or the depositional area. For this reason, low ATi values do not necessarily reflect the apatite:tourmaline ratio within the source rocks. ATi values are most reliable as a provenance indicator where transport and deposition has taken place under arid and/or cold climatic conditions. They can, however, provide useful information on sediment maturity, since the degree of weathering is dependent not only on climate, but also on relief and rates of erosion, transportation and deposition. Variation in ATi values in particular can reflect sea-level history (Morton and Hallsworth, 1999) and thus be of value in correlating cycles of sedimentation between sections.

A key element in the interpretation of all heavy mineral assemblages is the degree to which the mineral composition may have been inherited from pre-existing sediments. In particular, a high degree of physical and mineralogical maturity does not necessarily reflect intense weathering and prolonged transport: it can equally well reflect repeated recycling of the component sands. Fuller accounts of the concept and interpretation of heavy mineral indices is provided by Morton and Hallsworth (1994, 1999) and references therein.

Each of the mineral indices used in this study is briefly discussed below. Heavy mineral data are presented in Tables 2 and 3.

The rutile:zircon index (RZi) is a highly reliable provenance indicator as both rutile and zircon are highly stable. However, this stability also favours recycling, such that much of the signature may be inherited from earlier sediments. Strong variation in RZi values may reflect variation in the proportion of metamorphic to plutonic source rocks or variation in the type of metamorphic source rock.

The monazite:zircon index (MZi) is a reliable provenance indicator as both minerals are highly stable. However, in these sandstones monazite occurs in relatively low proportions, such that variation is limited and accurate ratio figures demand relatively high counts. Monazite is a component of both igneous and metamorphic rocks.

The staurolite:tourmaline index (STi) is a measure of the relative abundance of staurolite, which is an index mineral for medium-grade metasedimentary rocks. Because of the difference in density of the two minerals (3.7 and ca. 3.1, respectively), this index may be subject to a degree of density sorting effects. Also, staurolite is susceptible to dissolution at burial depths greater than ca. 2,000 m. It appears to be relatively resistant to weathering-related dissolution.

The apatite:tourmaline index (ATi) is a potential provenance indicator, but is also strongly influenced by weathering, especially under conditions of low relief and humid weathering. Apatite is a ubiquitous and abundant component of heavy mineral assemblages in both igneous and metamorphic source rocks. Low or very low ATi values are therefore indicative of a significant degree of weathering of the primary heavy mineral assemblages.

The euhedral zircon index (eZi) indicates the percentage of zircons with euhedral crystal form. For this purpose, euhedral zircons are defined as those with well-preserved, unabraded prism faces, including those with abraded interfacial angles. Most of the zircons thus defined are likely to be of first-cycle origin. While the euhedral zircon index can act as a reliable provenance indicator (e.g. distinguishing between sands derived from igneous or mature sedimentary sources) variation in eZi values may also reflect variation in the degree of abrasion that has taken place during transport and deposition. An increase in eZi values is often related to rejuvenation of crystalline source areas, and hence regression, but can also be the result of reworking of weathered basement during transgression. These different mechanisms can usually be inferred from the associated lithofacies succession.

Stratigraphic plots of mineral indices for each of the sections studied are presented in Figures 5, 9, 10 and 15.

CAMBRIAN SANDSTONES

Lithostratigraphy

In his summary of the Palaeozoic of Egypt, Klitzsch (1990a) adopted the scheme of Said (1971) and divided the Cambrian succession at Um Bogma into a lower part with abundant trace fossils (Araba Formation) and an upper unfossiliferous part (Naqus Formation). However, although the term ‘Araba Formation’ is widely used, Weissbrod (2003, p. 103) has argued that its definition is ambiguous. The ambiguity arises because the strata included within the Araba Formation in its type area (Wadi Araba, southwestern Sinai: Said, 1971) are in fact equivalent to the ‘Naqus Formation’ of the Um Bogma area. The status of the ‘Naqus Formation’ is also uncertain because of its differing application in different parts of the Gulf of Suez area (Weissbrod, 2004). Thus the ‘Naqus Formation’ as applied to the Um Bogma area by Klitzsch (1990a) equates with the Adedia Formation of Soliman and Fetouh (1969) (see also El-Shahat and Kora, 1986) and with the Netafim Formation of Weissbrod and Nachmias (1986) (Figure 4). It underlies the Lower Carboniferous Um Bogma Formation and is of Cambrian age. According to Weissbrod (2004), however, sandstones assigned by previous authors to the Naqus Formation in the Abu Durba area are of Early Carboniferous age and equivalent to the Abu Thora Formation, as discussed below. These sandstones thus post-date the Um Bogma Formation, whereas the ‘Naqus Formation’ of the Um Bogma area pre-dates the Um Bogma Formation.

Because of these ambiguities, and to provide a more detailed lithostratigraphic framework for the current heavy mineral study, we here use the lithostratigraphic subdivisions of Kora (1984) and El-Shahat and Kora (1986). These authors identified four formations, Sarabit El Khadim, Abu Hamata, Nasib, and Adedia, within the Lower Palaeozoic succession. These four units are almost identical to the four regional formations identified by Weissbrod (1969) (see Figure 4). The first three units equate with the Araba Formation as interpreted by Klitzsch (1990a) and we follow El Kelani and Bakry (2000) in regarding them as members of the Araba Formation. The Nasib Formation was divided by El-Shahat and Kora (1986) into the White Member, below, and the Variegated Member, above. These are here referred to as the ‘White Unit’ and the ‘Variegated Unit’.

The age assignments shown in Figure 4 for the four Cambrian lithostratigraphic units are taken from Weissbrod (2005), i.e. a late Early Cambrian age for the Sarabit El Khadim Member (Amudei Shelomo Formation), a Middle Cambrian age for the Abu Hamata and Nasib members (Timna and Shehoret formations) and a Late Cambrian age for the Adedia (Netafim) Formation. In the section at Wadi Nasib (Figure 5) and in the Um Bogma area to the west, the Cambrian succession is overlain by dolomitic shales and carbonates of the Lower Carboniferous Um Bogma Formation. The characteristic features of the four Cambrian units are summarised below.

The basal Cambrian Sarabit El Khadim Member, rests unconformably on Precambrian basement rocks. It consists largely of coarse-grained, variably conglomeratic, arkosic sandstones with thin beds of siltstone and a locally developed basal conglomerate. Its variable thickness (11–23 m) may in part reflect a buried topography on the surface of the underlying Precambrian basement. Weissbrod (2003) attributed these sediments to deposition in braided river channel systems.

The overlying Abu Hamata Member is dominated by fine-grained sandstones but also includes beds of medium to coarse grained sandstone and siltstone, along with partings of micaceous shale. Sedimentary facies, trace fossils (including Cruziana and Skolithos) indicate deposition in subtidal to intertidal environments (El-Shahat and Kora, 1986; Weissbrod, 2003). This interpretation is compatible with the presence of glauconite near the base of the unit (21 m: Figure 5) and at higher levels (27.5, 29, 30, 40 m: Figure 5). The upper few metres of the section here assigned to the Abu Hamata Member were assigned by Weissbrod (1969) and Weissbrod and Nachmias (1986) to the ‘Multicoloured Member’ and included in the Shehoret Formation (see Figure 4).

The Nasib Member consists mainly of fine grained sandstones in the lower part (White Unit) and of alternating sandstone, siltstone and mudstone in the upper part (Variegated Unit). Tabular cross-bedding in the sandstones of the White Unit is directed towards the northwest (El-Shahat and Kora, 1986). Khalifa et al. (2006, p. 879) also recorded cross-bedding directed towards the northwest in their ‘Middle Unit’ (of the Araba Formation), which is here regarded as broadly equivalent to the Nasib Member. The presence of flaser bedding and common channel structures in the White Unit indicates that deposition took place in a tidal-flat environment (El-Shahat and Kora, 1986; Weissbrod, 2003). The Variegated Unit is characterised by a higher proportion of siltstone and micaceous shale and was interpreted by El-Shahat and Kora as marking a rise in sea level, leading to the establishment of a subtidal shelf environment. This interpretation is supported by the presence of the trace-fossil Cruziana. The absence of glauconite in the sandstones of the Nasib Member may reflect higher rates of sedimentation than in the Abu Hamata Member.

The Adedia Formation (Netafim Formation) consists of fine grained violet and brown, laminated, cross-bedded sandstones with minor coarse sandstone and pebbly sandstone. The sandstones contrast with those of the underlying formations in being quartzarenitic rather than arkosic or subarkosic (Weissbrod and Perath, 1990; Weissbrod, 2003). The sandstones include the trace-fossil Skolithos. The Adedia (Netafim) Formation was considered by Weissbrod and Perath (1990, p. 265) to represent a new sedimentary cycle: the ‘second Palaeozoic cycle’. Sediments of this formation overstep the underlying Cambrian strata (‘Araba Formation’) to rest directly on basement rocks in many parts of northeastern Egypt (see Weissbrod, 2004, figure 2). A possible beach environment was suggested for this unit (‘Naqus Formation’) by Klitzsch (1990a, p. 395). Weissbrod (2005) also envisaged deposition in a shoreline belt, but with short-lived periods of fluvial progradation. The formation is probably Late Cambrian in age (Weissbrod, 2003, p. 102).

Mineral stratigraphy

A stratigraphic plot of mineral indices for the Wadi Nasib section is shown in Figure 5. Provenance signatures in the Sarabit El Khadim and Abu Hamata members (Amudei Shelomo, Timna and Multicoloured units of Weissbrod and Nachmias, 1986) show marked variation in all but the MZi values. RZi values fall into two groups, one with low to extremely low values and the other with moderate to high values. The latter group is characterised by extremely low STi values. Variation in RZi values does not correspond to the lithostratigraphic boundaries, with a marked upward decrease in values occurring in the upper part of the Sarabit El Khadim Member and a sharp upward increase in values occurring in the middle of the Abu Hamata Member. A sharp upward increase in eZi values at ca. 25 m (associated with a sharp upward increase in ATi values) occurs above a thin conglomerate bed and indicates an influx of less mature sand. The uppermost two samples from the Abu Hamata Member are characterised by extremely high ATi values.

The base of the Nasib Member is associated with a sharp upward increase in MZi values, indicating a significant change in provenance. The extremely high ATi values associated with the uppermost Abu Hamata sandstones persist into the lower part of the Nasib Member. Provenance signatures within the Nasib Formation show an overall upward decrease in both RZi and MZi values, with a slight increase in STi values near the top of the section. ATi values are extremely high in the lower part of the Nasib Member, and extremely low in the remainder. eZi values are lower than those of the underlying formations, ranging from low in the lower part to extremely low in the upper part.

The variability in heavy mineral signatures in the Araba Formation is also apparent in the results obtained by Weissbrod and Nachmias (1986, figure 8), although monazite was not identified as a significant component of the ‘White’ and ‘Variegated’ units of the Nasib Member (Shehoret Formation).

Sandstones of the Adedia Formation display moderate RZi values and are distinguished from those of the underlying Nasib Member by possessing extremely low MZi values. They possess extremely low STi and eZi values, together with zero ATi values.

A cross-plot of RZi and MZi values (Figure 6a) shows that the Cambrian sandstones fall into three provenance groups. A group with very low MZi and RZi values, including sandstones of the Sarabit El Khadim and Abu Hamata members, a group with very low MZi values and high RZi values, including sandstones of the Sarabit El Khadim and Abu Hamata members and the Adedia Formation, and a group with relatively high but variable MZi values and variable RZi values, restricted to sandstones of the Nasib Member. A cross-plot of RZi and STi values (Figure 6b) shows that sandstones of the first group mostly possess STi values higher than those of the second group.

A striking feature of these two cross-plots is that the sandstones of the Sarabit El Khadim and Abu Hamata members have closely similar provenance signatures, with samples of both formations falling within the low-RZi and the high-RZi groups. This indicates that there was no long-term change in heavy-mineral provenance over this interval and that the two formations may be regarded as constituting a single genetic unit, in which sand was supplied from two separate sources. It may also be noted that no fundamental change in heavy mineral provenance signature is associated with the base of the Timna Formation of Weissbrod and Nachmias (1986), even though the composition of the feldspars changes from mixed potassic/sodic to dominantly potassic at this level (Weissbrod and Perath, 1990).

Two distinct cycles of upward-decreasing eZi values are apparent within the Sarabit El Khadim and Abu Hamata succession (Figure 5), but are related to different facies. The relatively high eZi values at the base of the succession are associated with proximal fluviomarine facies, whereas those in the Abu Hamata Member are associated with transgressive marine sediments. The former can be confidently ascribed to relatively high erosion rates at the commencement of Cambrian sedimentation, whereas the latter may reflect increased coastal erosion or the temporary exposure of source rocks rich in euhedral zircon. Sandstones of the Nasib Member and Adedia Formation possess significantly lower eZi values (Figure 7).

The upward increase in MZi values at the base of the Nasib Member (White Unit) clearly marks a significant change in provenance. This level equates with the base of the White Member (Shehoret Formation) of Weissbrod and Nachmias (1986) and Weissbrod and Perath (1990) and is believed to correlate with level at which cross-bedded sandstones first appear in the sections described by Khalifa et al. (2006). Cross-bedding directions in these sandstones indicate transport to the north and west (El Shahat and Kora, 1986, p. 159; Khalifa et al., 2006, p. 879), away from the land area inferred for the Cambrian by Klitzsch and Squyres (1990, figure 1) and Guiraud (1999, figure 2). Since the northern margin of this land area was presumably the source of the more proximal fluvial sandstones of the Sarabit El Khadim Member, which lack monazite, it is likely that the Nasib sands were derived from a source further to the south. Transport to the depositional area may have been by longshore drift, with prolonged residence time in the beach zone accounting for the relatively low eZi values in the Nasib sands.

In RZi/MZi and RZi/STi cross-plots (Figure 6a, b) the sandstones of the Adedia Formation plot close to the field for the high-RZi sandstones of the Sarabit El Khadim and Abu Hamata members, although their eZi values are much lower (Figure 7). Since neither the low-RZi component of the Sarabit El Khadim and Abu Hamata sands nor the high MZi sands of the Nasib Member are represented, the Adedia Formation must have been derived from a new source. The upward change from arkosic to quartz arenitic sand composition at the boundary between the Araba and Adedia formations indicates an increase in the intensity of chemical weathering before or during deposition of the Adedia sands.

CARBONIFEROUS SANDSTONES

Lithostratigraphy

In the Um Bogma area (Figure 1), the oldest Carboniferous sediments are the dolomitic shales and carbonates of the Early Carboniferous (Visean) Um Bogma Formation (Figure 8). These are overlain by sandstones of the Lower to Upper Carboniferous Abu Thora Formation. The carbonates of the Um Bogma Formation are believed to have been deposited under open marine shelf conditions whereas the sandstones of the Abu Thora Formation are interpreted as representing a range of environments from fully marine to fluvial (Kora 1995). In its type section at Wadi Abu Thora, near Um Bogma, the Abu Thora Formation is about 190 m thick (Kora, 1998), but only the lowest 50 m were sampled in this study (Figure 9).

Some 50 km south of Wadi Abu Thora, at Abu Durba (Figure 1), the Um Bogma Formation is missing and the thickness of the Abu Thora Formation reduced to around 50–80 m thick. The overlying section includes Carboniferous and Triassic sediments overlain unconformably by the Lower Cretaceous Malha Formation (Figure 10). Published accounts of the section at Abu Durba show differing interpretations of the lithostratigraphic assignments in this section, as discussed below.

A detailed description of the pre-Cretaceous succession in the Wadi Feiran – El Tor area (which includes the Abu Durba section - see Figure 1) is given by Allam (1989). He identified a succession that consisted of the Araba Formation (Cambrian), Naqus Formation (Cambrian), Abu Durba Formation (Early Carboniferous), Aheimer Formation (Late Carboniferous) and Qiseib Formation (Late Carboniferous to Permian). The Abu Durba Formation was described as consisting of 110 m of interbedded yellowish sandstones and shales with a reddish conglomerate at the base. The Aheimer Formation was described as 90 m thick and consisting of a lower unit of interbedded shales, sandstones and dolomite layers, a middle unit of massive sandstones and a top unit of highly fossiliferous shales with hard, sandy limestones. The Qiseib Formation is now known to be of Permian to Triassic age (see below).

In interpreting lithostratigraphic assignment of Allam (1989), it should be noted that he followed Issawi and Jux (1982) in applying the name ‘Abu Durba Formation’ to a sandstone unit, whereas the term should correctly be restricted to shale-dominated facies, following the original definition by Said (1971) as ‘Durba Shale Formation’ or ‘Durba Black Shales’. Klitzsch (1990a, p. 402-403) described this same sandstone unit as underlying the ‘Abu Durba shale’ and consisting of a 50 to 80 m thick unit of marine siltstone and sandstone interbedded with fluvial sandstone. This sandstone unit had been assigned to the Naqus Formation by several previous authors (e.g., Weissbrod, 1969; Issawi and Jux, 1982), but Klitzsch (1990a, p. 402) considered that it “most likely correlates with the lower part of the Ataqa or Abu Thora Formation in the Um Bogma area”. Kora (1998) reached a similar conclusion, applying the term Abu Thora Formation to the sandstone unit and the term Abu Durba Formation to the overlying shale-dominated section.

Klitzsch’s (1990a) reinterpretation of the section at Abu Durba and Wadi Feiran was discussed by Weissbrod (2004), referring to Klitzsch’s reassignment of the ‘upper part of the Naqus Formation’ to the Early Carboniferous. Weissbrod (2004) evidently extended the Early Carboniferous assignment to the ‘lower part of the Naqus Formation’, since in his conclusions he states that “The names Naqus, Gilf, and Somr El Qaa given to the white sandstone formations that overlie the Cambrian Araba Formation (=Netafim Formation) are synonymous, designating the same rock unit. They are all time-equivalents, and are correlative to the Abu Thora Formation (Early to Late Carboniferous) in the Um Bogma area …”. He referred to these sandstones as ‘undifferentiated Carboniferous sandstones’ and later as Abu Thora / Naqus Formation (Weissbrod, 1995).

Not all of the sandstones beneath the Abu Durba Formation at Abu Durba are white sandstones of ‘Naqus type’, however. In our section, the conglomerate that occurs at the base of the Abu Thora Formation (sensuKlitzsch, 1990a) is underlain by brown sandstones, of which 20 m are exposed. These may equate with the “dark brown to red sandstone and clay unit of 15 m thickness” reported by Issawi and Jux (1982) as resting unconformably on the Naqus Formation and assigned to the ‘Wadi Malik Formation’. We provisionally include these sandstones in the Abu Thora Formation.

The shale-dominated section above the sandstones of the Abu Thora Formation at Abu Durba is here assigned to the Abu Durba Formation, following Kora (1998). Earlier assignments to the Aheimer Formation (e.g. Allam, 1989) are deemed invalid, since biostratigraphic analysis indicates that the Aheimer Formation is not present in Sinai (Kora, 1998, p. 714; see also Jenkins, 1990, Figure 19.4). The base of the Abu Durba Formation is marked by a thin conglomerate bed. In the studied section, the Abu Durba Formation is only 54 m thick, whereas Kora (1998, figure 3) indicated a thickness of ca. 105 m for the Abu Durba Formation at Abu Durba. Since the distinctive fossiliferous limestone recorded by Kora from near the top of the Abu Durba Formation is not present in our section, and since the underlying 45 m thick sandstone section is also absent, it appears that our section represents only the lower part of the succession recorded by Kora (1998). The remainder appears to have been removed by erosion beneath the pre-Qiseib unconformity. The Abu Durba Formation is of Late Carboniferous age, ranging from Moscovian to early Kasimovian in complete sections (Kora, 1998).

Mineral stratigraphy

In the section at Wadi Abu Thora (Figure 9), the sandstones of the Abu Thora Formation display very consistent RZi and MZi values, coupled with zero STi and ATi values. Significant variation in eZi values is displayed, however.

By contrast, the Carboniferous sandstones in the section at Abu Durba (Figure 10), show marked variation in RZi and STi values. Sandstones in the lower part of the section are characterised by moderate to high RZi values and zero STi values. The two samples from the ‘brown sandstone unit’ at the base of the section possess slightly higher eZi values than those of the overlying sandstones. RZi values within the Abu Thora Formation are highly variable. By contrast, MZi values and eZi values are consistently low throughout. Staurolite first appears near the top of the formation, reflected in moderate STi values in three samples. ATi values are also extremely low except at the base of the formation, where a moderately high ATi value is recorded from a sandstone sample 6 m above the basal pebbly sandstone bed.

Data for the Abu Durba Formation are very limited because of the scarcity of sandstones and a scarcity of sand of the appropriate grain size. The basal conglomerate yielded valid counts only for the apatite:tourmaline index, which yielded a moderate value. A sandstone sample from the lower part of the formation yielded a moderate RZi value, a rather high STi value, a low ATi value and an extremely low eZi value. Samples from the sandstone unit at the top of the formation yielded moderate RZi and STi values and zero ATi values.

The consistent values for RZi, MZi and STi in the sandstones at Wadi Abu Thora indicate that no significant change in provenance took place during sedimentation of the lower part of the Abu Thora Formation. Two depositional cycles appear to be indicated by the eZi values, with an initial upward decrease in eZi values terminated by a sharp upward increase at ca. 33 m above the base. Since there is no associated change in provenance signature, the variation in eZi values is most likely to have resulted from changes in relative sea level. RZi, MZi and STi values for the Abu Thora Formation are compatible with derivation through reworking of the Cambrian Adedia Formation and Nasib sands (Figure 11a, b). However, their higher eZi values (Figure 12) indicates that there may also have been a contribution from crystalline basement rocks.

In the Abu Durba section (Figure 10), variation in RZi values allows the recognition of three heavy mineral units within the Abu Thora Formation. The lower two units display the strongest contrast in RZi values, while the top unit, with moderate RZi values, is distinguished by the first appearance of staurolite. It is thus clear that significant changes in provenance took place during deposition of the Abu Thora sandstones. A comparison of RZi and MZi values in these sandstones (Figure 11c) and those of the Abu Thora Formation at Wadi Abu Thora (Figure 11a) indicates that they have a different provenance. It is possible, however, that equivalents of the Wadi Abu Thora sandstones are represented beneath the succession measured at Abu Durba (see discussion, above).

Provenance signatures in the Abu Durba Formation are characterised by relatively high STi values (Figure 11d), thus continuing the trend of upward increasing STi values in the upper unit of the Abu Thora Formation.

The relatively high ATi values associated with the basal parts of the Abu Thora and Abu Durba formations at Abu Durba are interpreted as representing periods of reduced chemical weathering at a time of extensive transgression. The rapid reduction in ATi values in the overlying sections is typical of the increased weathering associated with the establishment of extensive alluvial floodplains during the highstand phase (Morton and Hallsworth, 1999).

TRIASSIC SANDSTONES

Lithostratigraphy

The red and variegated sandstones and shales that overlie the Abu Durba Formation at Abu Durba (Figure 11) are here assigned to the Qiseib Formation (equivalent to the Budra Formation of Weissbrod and Nachmias, 1986). A thin conglomerate is present at the base.

On the western side of the Gulf of Suez, the Qiseib Formation consists of a ‘lower clastic part’ and an ‘upper carbonate part’ (Kerdany and Cherif, 1990). The former has yielded an Early Permian flora at Wadi Araba (Lejal-Nicol, 1987; Klitzsch, 1990a) and the latter a Middle Triassic fauna (Weissbrod, 1969; Kerdany and Cherif, 1990, p. 416). In southwestern Sinai, the Qiseib Formation is represented entirely by continental deposits. These deposits were considered by Barakat et al. (1986) to be younger than the basal Qiseib succession of the western side of the Gulf, and we here follow Kerdany and Cherif (1990) in considering them to be entirely of Triassic age. The succession exposed at Abu Durba probably comprises the ‘Lower Member’ and ‘Middle Member’ of El Kelani and Bakry (200), since the pebbly sandstones that characterise the ‘Upper Member’ are not present, and were presumably removed during generation of the base Malha Formation unconformity.

Mineral stratigraphy

Sandstones from the lower part of the Qiseib Formation possess moderate to high RZi values, low STi values and moderate eZi values. MZi values show a slight increase up-section (Figure 10). ATi values are extremely low throughout.

The base of the Qiseib Formation is associated with a sharp upward decrease in STi values. These data appear to be at odds with those of Weissbrod and Nachmias (1986, figure 8) for the Qiseib (Budra) Formation of the Um Bogma area. However, since Weissbrod and Nachmias (1986) did not sample the base of the formation and since the Abu Durba section lacks the ‘Upper Member’ of El Kelani and Bakry (2000), it is possible that the two data sets represent different levels within the Qiseib Formation. The relatively high eZi values in the Qiseib sandstones of Abu Durba indicate that the Triassic sands cannot have been derived by reworking of any of the Carboniferous sands. They probably reflect significant rejuvenation of the source area, perhaps related to the reconfiguration of the regional palaeogeography that took place during the Late Carboniferous (Hercynian) tectonic phase (Klitzsch, 1990a, p. 399-400).

CRETACEOUS SANDSTONES

Lithostratigraphy

The Cretaceous sandstones of southwest Sinai are here assigned to the Malha Formation (of Abdallah et al., 1965). Correlative strata in the Levant were assigned a Barremian to Albian age by Weissbrod et al. (1994, p. 70 and figure 2) and Weissbrod (2002). This age assignment is followed here (Figure 3).

Weissbrod (1969) introduced an alternative lithostratigraphic scheme in which strata assigned to the Malha Formation are included within three lithostratigraphic units. In ascending order, these are the Amir Formation, Avrona Member (Hatira Formation) and ‘unnamed member’ (Hatira Formation) (Figure 13). The last two units have since been given formation status: Avrona Formation and Samar Formation, respectively (Weissbrod, 1993).

Another lithostratigraphic scheme was proposed by Barakat et al. (1986), who also identified three formations within the unit here referred to as the Malha Formation. In ascending order, these are the Raqaba Formation, the Temmariya Formation and the Malha Formation. The practicability of these formations was questioned by Issawi et al. (1998, p. 64) and Kora (1989) on the grounds that their constituent lithologies were not sufficiently distinctive. It is nevertheless possible to correlate between the units of Barakat et al. (1986) and those of Weissbrod (1969) (Figure 13), both of whom included lithological logs of the section at Gebel Raqaba. Barakat et al. (1986, p. 384) indicate that the Raqaba Formation is equivalent to the Amir Formation, while the Temmariya and Malha formations are together equivalent to the Hatira Formation and hence to the Avrona and Samar formations of later usage (see Figure 13).

Of particular significance is the presence of a siltstone unit, several metres thick, in the lower part of the succession. This siltstone was described by Barakat et al. (1986) as red, brown, violet and mottled in colour and referred to as the ‘upper member’ of the Temmariya Formation. Weissbrod (1969) does not indicate the colour of this unit at Gebel Raqaba, but notes (p. 18) the occurrence of a 10–15 m thick red siltstone unit at a similar level at Gebel el Tih (see Figure 1). Beds of kaoliniticclay were identified in the lower part of the “Malha” Formation by Barakat et al. (1986) and at an equivalent level in the upper part of the Avrona Formation (Weissbrod, 1969). The same succession, of red and variegated siltstone and sandstone overlain by interbedded sandstone and kaolinitic clay was recorded in this study at Musaba Salama (see Figure 14), in the workings of the Sinai Manganese Company. This section exposes the upper part of the “Temmariya Member” and the lower part of the “Drab Member”. Two kaolin beds are present in the basal part the “Drab Member”. The lower kaolin bed is believed to equate with the C bed of Kora and El-Beialy (1989) and is overlain by a thin conglomerate with an irregular erosional base.

Both Weissbrod (1969) and Barakat et al. (1986) identified an upward change from drab to variegated colours in the upper part of the section at Gebel Raqaba, corresponding to the boundary between the Avrona Member (now Avrona Formation) and the ‘unnamed member’ (now Samar Formation) and between the Lower (drab) Member and the Upper (variegated) Member of the ‘Malha Formation’.

It is thus apparent that it is possible to identify several stratigraphic units within the Lower Cretaceous succession of Gebel Raqaba and surrounding areas, but that no agreement exists on the limits of these units, or on their mappability and appropriate hierarchical status. In this paper, therefore, we have divided the Malha Formation (sensuKerdany and Cherif, 1990) into four informal members, based on the units identified by Barakat et al. (1986). These are the Raqaba, Temmariya, ‘Drab’ and ‘Variegated’ members. Their equivalence to the terminology of Weissbrod (1969) is shown in Figure 11. Correlative strata in the Levant were assigned a Barremian to Albian age by Weissbrod et al. (1994, p. 70 and figure 2) and Weissbrod (2002, 2003). This age assignment is followed here (Figure 3).

The sandstones of the Malha Formation consist dominantly of pale-coloured cross-bedded sandstones, with those of the Temmariya Member being mostly coarser than the remainder. Red and variegated sandstones, siltstones and shales occur in the upper part of the Temmariya Member and constitute the bulk of the sediments in the Variegated Member at the top of the Malha Formation. Concretions rich in iron and manganese oxides as well as iron and manganese crusts are frequent in the upper part of the Temmariya Member. The contact between the Amir Formation and the Avrona Formation was said to be unconformable by Weissbrod and Nachmias (1986) and Weissbrod (1993), but was later regarded as representing only a short depositional hiatus, with minimal erosion (Weissbrod, 2002, p. 63). Barakat et al. (1986) recognised an unconformity at the base of their Temmariya Formation, i.e. at a level above the base of the Avrona Formation (see Figure 12). At Gebel Musaba Salama, however, the full thickness of the upper part of the Temmariya Member is apparently preserved whereas the succession at the base of the ‘Drab Member’ is much reduced, with only two kaolin beds present and the basal 30 m sandstone unit missing. The base of the ‘Drab Member’ may therefore be a surface of non-deposition rather than an erosional unconformity. The base of the ‘Variegated Member’ (Samar Formation) was regarded as marking a regional unconformity by both Barakat et al. (1986) and Weissbrod (2002).

Both of the Malha Formation sections sampled for this study, at Abu Durba (Figure 10) and Gebel Musaba Salama (Figure 15), lie on an uplifted block that influenced Early Cretaceous sedimentation in west-central and southwestern Sinai (see Weissbrod, 2002, p. 73 and figure 5). In the north of this block (including Gebel Musaba Salama) the Raqaba Member (Amir Formation) is missing, and the Temmariya Member (Avrona Formation) rests unconformably on the Qiseib (Budra) Formation (Weissbrod 2002, p. 63). In the south of the block, the Amir and Avrona formations (Raqaba, Temmariya and ‘Drab’ members) are all missing according to Weissbrod (2002, p. 63), with the Lower Cretaceous represented by the Samar Formation (‘Variegated Member’) alone. Barakat et al. (1986, figure 8) show only their ‘Malha Formation’ present at Abu Durba, although they do not indicate which member or members are present. In our section, a 50 cm bed of violet shale separates an underlying unit of greyish green, medium- to coarse-grained friable sandstone from an overlying unit of reddish, pink and violet, medium-grained, cross-bedded sandstone. It is possible that the lower sandstone unit, which is kaolinitic in part, represents the top of the ‘Drab Member’. However, it could also belong to the ‘Variegated Member’, since Barakat et al. (1986, p. 386) state that outside the area of the El Tih escarpment, the Upper (Variegated) Member includes pale-coloured and white sandstones with grey kaolin beds. In the absence of definite evidence for the presence of the ‘Drab Member’ in the Abu Durba succession, we follow Weissbrod (2002) in assigning all of the Lower Cretaceous sandstones to the ‘Variegated Member’.

At Abu Durba, the Malha Formation rests unconformably on the Qiseib Formation (as tentatively proposed by Barakat et al., 1986, figure 8), whereas at nearby Gebel Ekma, they rest on the Abu Durba Formation (Kora, 1998, figure 3). In both localities, the base of the Malha Formation is marked by a thin but prominent bed of indurated goethitic silty sandstone, which at Abu Durba is underlain by a 50 cm thick kaolin bed. An indurated ferruginous bed is also present at the base of the ‘Drab Member’ at Gebel Musaba Salama.

Mineral stratigraphy

The boundary between the Temmariya Member and the ‘Drab Member’ at Gebel Musaba Salama (Figure 15) is marked by a sharp upward increase in STi values and eZi values. STi and eZi values show a progressive upward decrease in the upper part of the section, accompanied by an increase in MZi values and an irregular increase in RZi values. ATi values are extremely low throughout, but are slightly higher in the high-eZi section than in the remainder.

Sandstones in the Malha Formation at Abu Durba (Figure 10) are characterised by moderate to high RZi values, low MZi values, low STi values, extremely low ATi values and moderate to high eZi values, with the highest eZi values occurring at the base of the section.

A cross-plot of RZi and STi values (Figure 16a) shows that the sandstones of the ‘Temmariya Member’ and ‘Drab Member’ at Gebel Musaba Salama and the sandstones of the Malha Formation at Abu Durba plot in largely separate fields. However, they appear to fall within a single compositional trend of decreasing RZi values with increasing STi values. This suggests that the range in composition of the Malha sandstone succession as a whole developed through the mixing of low-STi and high-STi end-member sands. The similarity in composition of the lower and upper sandstone units at Abu Durba indicates that they belong to the same genetic unit, despite their contrasting lithological character. For this reason, the lower, green-grey sandstone unit is provisionally regarded as belonging to the ‘Variegated Member’ rather than to the ‘Drab Member’. Further study of Malha Formation sections will be needed to determine whether sandstones of this composition are indeed restricted to the ‘Variegated Member’.

A progressive increase in RZi and MZi values appears to have taken place over time, with the lowest values in the ‘Temmariya Member’ and the basal part of the ‘Drab Member’ (Figure 15) and the highest values in the ‘Variegated Member’. The influx of high-STi sands appears to have been superimposed on this long term trend.

Since the increase in STi values at the base of the ‘Drab Member’ is associated with an increase in eZi values, it presumably represents the nearby exposure of metamorphic and granitic basement rocks. Comparably high eZi values in the ‘Variegated Member’ (Figure 16b) are not associated with high STi values, however, indicating that these sands were derived from a different area of basement outcrop. Both high-eZi sections probably represent periods of source-area rejuvenation.

The high-RZi, low STi sands of the Malha Formation at Abu Durba may have been derived through reworking of Qiseib sands, which show a comparable range in RZi and STi values (Figure 16a). The analysed Qiseib sands have lower eZi values than the Malha sands (Figure 16b), but higher values may have existed in more proximal sand facies and may be present in the coarse-grained ‘Upper Member’ of Kelani and Bakry (2000), which is presumed to have been removed by erosion at this location.

DISCUSSION

The use of heavy mineral indices has revealed previously unrecognised variations both between and within formations. The following sections demonstrate how such variations can be used to obtain improved regional correlation of individual sandstone units and to better assess the role of recycling and the influence of climate on the composition of the heavy mineral assemblages.

Regional correlation potential

Heavy mineral assemblages are useful in providing a ‘mineral fingerprint’ that can be used in the identification and correlation of sandstones. This approach can be especially helpful in the investigation of subsurface sections, where stratigraphic relationships may not be well established.

The most distinctive units encountered in this study are those of the Araba Formation, in which the mineralogically variable Sarabit El Khadim and Abu Hamata members are overlain by the monazite-bearing Nasib Formation (Figure 5). The unique mineralogy of the Nasib Member, with its relatively high MZi values, indicates that its base is likely to represent a regionally traceable stratigraphic surface and might therefore provide a reliable mapping datum within the Araba Formation. In this respect, it may be noted that the White Unit (White Member) is recognised also within the Shehoret Formation of the Levant and that according to Weissbrod (2003, 2005) its base equates with the boundary between the Burj Dolomite-Shale and Abu Khusheiba Sandstone formations (see Powell, 1989) of Jordan. If a mappable two-fold or three-fold division of the Araba Formation were to prove practicable, this would facilitate the suggestion of Weissbrod (2003, p. 103) to establish an ‘Araba Group’ to include all the Lower Palaeozoic strata of northeast Egypt. It may be noted that the Araba Formation has already been equated with the entire Lower Palaeozoic succession by El-Shahat and Kora (1986, p. 160-161) and that the term Araba Formation was subsequently used in this sense by Abdallah et al. (1992) and Khalifa et al. (2006).

Khalifa et al. (2006) recognised a three-fold division within the Araba Formation, in which the ‘Lower Unit’ equates with the Sarabit El Khadim and Abu Hamata members of this study. The equivalence of the ‘Middle’ and ‘Upper’ units is less clear, but the base of the ‘Middle Unit’ clearly equates with the base of the Nasib Member. These authors show the ‘Middle Unit’ (i.e. Nasib Member) resting directly on crystalline basement in two of their six measured sections (Khalifa et al., 2006, figure 2).

The inclusion by Khalifa et al. (2006) of the Sarabit El Khadim and Abu Hamata members within a single lithostratigraphic unit (‘Lower Unit’) is compatible with the mineralogical observations of this study, since the sandstones of the Sarabit el Khadim and Abu Hamata members display a similar range of heavy mineral compositions. This indicates that no significant tectonic restructuring took place during this phase of sedimentation. On the evidence from Sinai alone, therefore, the two members could be interpreted as representing a single genetic sequence. Correlations over the wider Middle East region suggest otherwise, however, since Al-Husseini (2010) has presented evidence for a regionally recognisable sequence boundary at the base of the Burj Formation of Jordan and Saudi Arabia, which is equivalent to the base of the Abu Hamata Member (Weissbrod, 2003, 2005). The arkosic sandstones that underlie the Burj Sequence Boundary were assigned to the ‘Asfar Sequence’ by Al-Husseini (2010). They include the Siq Formation (sensu Al-Husseini, 2010) of Saudi Arabia, the Salib Formation of Jordan and, by correlation (e.g. Weissbrod, 2003, 2005), the Sarabit El Khadim Member of Sinai.

The Cambrian succession in northwestern Saudi Arabia is very similar to that of Sinai (Khalifa et al., 2006, figure 3). Using the lithostratigraphic scheme of Al-Husseini (2010, figure 6), the Abu Hamata Member is there represented by sandstones and mudstones of the Burj Formation, and the Nasib Member by the overlying Saq Formation (Risha Member). The unconformable nature of the base of the Saq Formation supports the conclusion reached in this study that the onset of deposition of the Nasib sands was associated with tectonic restructuring of the Arabian-Nubian Shield.

Heavy mineral signatures also appear to have the potential to establish a stratigraphic subdivision of the Carboniferous sandstone succession of Sinai. In addition to the occurrence of a distinctive assemblage in the Abu Durba Formation, the variation in mineralogy within the Abu Thora sandstones at Abu Durba (Figure 9) is likely to be of stratigraphic significance, as is the contrast in mineralogy between these sandstones and the Abu Thora sandstones at Wadi Abu Thora.

Although stratigraphic coverage of the Malha Formation is incomplete, the marked mineralogical variation recorded from the different units at Gebel Musaba Salama and at Abu Durba indicates that there is a possibility that heavy mineral signatures will help in the recognition of different units and thereby provide a firmer basis for the recognition of formal members.

Sand provenance and recycling

Although this study, like that of Weissbrod and Nachmias (1986), has identified successive changes in sand provenance during both the Palaeozoic and Early Cretaceous, the data are too limited to identify specific source rocks. It is, however, possible to assess the possibility of local recycling of older sands by comparing compositional ranges as defined by index cross-plots.

Regional studies of palaeocurrent indicators have indicated a dominantly northward transport direction for the Palaeozoic and Mesozoic fluvial sands of north Africa, the Sinai Peninsula and Arabia (e.g. McKee, 1962; Karcz and Key, 1966; Beuf et al., 1971; Selley, 1972, 1997; Klitzsch et al., 1979; Klitzsch, 1981; Dabbagh and Rogers, 1983; Van Houten et al., 1984; Klitzsch and Wycisk, 1987; Wycisk, 1990; Stump and van der Eem, 1995; Alsharhan and Nairn, 1997; Burke and Kraus, 2000; Burke et al., 2003; Avigad et al., 2005, figure 1). In central and southern Egypt, this northward transport pattern prevailed in the Cambrian, in the Devonian and Early Carboniferous and again in the Late Jurassic and Cretaceous, as reflected in a number of palaeogeographic reconstructions (Klitzsch, 1990a, b; Klitzsch and Squyres, 1990; Said, 1990; Kuss and Bachmann, 1996; Schandelmeier and Reynolds, 1997; Guiraud, 1999). It was interrupted by two phases of uplift. The first phase involved N-S uplift of central and eastern Egypt during the Silurian and probably also during the Ordovician (Klitzsch and Wycisk, 1987, p. 103 and Figure 3). Sedimentation, at least in the Silurian, was confined to western Egypt and areas east of the Arabian Gulf. Central and eastern parts of Egypt were either exposed or were covered by continental sediments that have since been eroded (Klitzsch and Wycisk, 1987, p. 131). The second phase involved E-W uplift of central Egypt in the mid Carboniferous, leading to the development of a southward drainage pattern in southern Egypt and northern Sudan (Klitzsch and Wycisk, 1987, figure 5). This drainage pattern persisted into the mid Jurassic. In the eastern part of the Arabian Shield, late Palaeozoic uplift of the Riyadh Swell led caused exposure of the basement and acted as a barrier to northward sand transport from the Permian (Weissbrod and Bogoch, 2007) or late Carboniferous (Knox et al., 2007) onwards.

Cambrian

In a review of the provenance of the Early Palaeozoic sands of northeastern Africa and Arabia, Weissbrod and Bogoch (2007) concluded that the main source lay to the south of the Arabian Shield, in interior Gondwana. A source to the south of the main Arabian-Nubian basement outcrop is certainly indicated for the sandstones in the Lower Palaeozoic formations of Yemen and the Wajid area of Saudi Arabia since they show a consistent northerly transport direction (Dabbagh and Rogers, 1983; Stump and Van der Eem, 1986; Hussain et al., 2004). These sands were evidently derived either from the southern fringe of the Arabian–Nubian Shield from an area further to the south (e.g., Dabbagh and Rogers, 1983; Konert et al., 2001; Hussain et al., 2004; Knox et al., 2007). This Lower Palaeozoic succession is believed to have once extended northwards over the crystalline basement of the eastern Arabian-Nubian Shield, as indicated by the continuity of the formations that fringe the present-day basement outcrop (Powers et al., 1966; Vaslet, 1990) and by the presence of outliers of Cambrian sediments on the crystalline basement of northwestern Saudi Arabia (Janjou et al., 1998; Al-Husseini, 2010).

A far southern provenance also may be true for the sands of the Lower Palaeozoic formations of western Egypt and areas further to the west. In central and eastern Egypt, including the Sinai area, however, the basal Cambrian units, are relatively thin and the youngest unit, the Upper Cambrian Adedia (Netafim) Formation shows eastward and southward onlap onto the crystalline basement (Weissbrod, 2005, p. 306). Furthermore, the Lower and Middle Cambrian sandstones possess arkosic and subarkosic compositions, which contrast with the quartz arenitic composition of equivalent sandstones in southwesterrn Saudi Arabia (Stump and Van der Eem, 1995; Hussain et al., 2004; Knox et al., 2007). They also display a wide scatter of compositions when represented on heavy mineral index cross-plots, indicating a lack of homogenisation and hence a short transport distance. The older Cambrian sediments could not, therefore, have extended far to the south of their present outcrop, as pointed out by Weissbrod and Bogoch, 2007, p. 669-670). Significant relief thus appears to have existed on the Nubian segment of the Arabian-Nubian Shield during the Early and Middle Cambrian, perhaps as a result of relative uplift. Significant local topographic relief also appears to have existed in the eastern part of the Arabian-Nubian Shield, since a westerly flow direction has been recorded from basal Cambrian fluvial sandstones in northwestern Saudi Arabia (Janjou et al., 1998; Al-Husseini, 2010, p. 143).

Active differential uplift may explain the change in sand provenance associated with the boundary between the Abu Hamata and Nasib members. A change in relative sea level alone is unlikely to have been responsible for the contrast in sand provenance signatures, since monazite-bearing sands were produced neither during the major transgression represented by Sarabit El Khadim -Abu Hamata transition nor during the subsequent regression represented by the increase in sand content towards the top of the Abu Hamata Member. Weissbrod (2005, p. 306) noted that the time-equivalent White Member of the Shehoret Formation of the Levant constitutes a local facies that pinches out on both the northern and southern margins of the basin. This indicates that the increase in the proportion of sandstone at the base of the Nasib Formation and the associated sharp increase in MZi values is the result of a tectonic reconfiguration of the hinterland rather than a change in relative sea level alone. This interpretation is supported by the lack of close grouping of individual sample compositions (Figure 6), which indicates a lack of sand mixing and hence supply from a source that was not far distant.

That monazite-bearing rocks are present within the Egyptian Arabian-Nubian Shield basement is indicated by the occurrence of a significant monazite component in heavy mineral assemblages of present-day beach sands on the eastern side of the Gulf of Suez (El-Kammar et al., 2007) and of sands that were shed southwards from the Gebel Uweinat area during the Permian, Triassic and Jurassic (Lakia Formation: Klitzsch and Wycisk, 1987, p. 122). The relatively low proportion of euhedral zircons in the Nasib sands may be inherited from the basement source rocks or it may reflect a high degree of abrasion resulting from a long residence time in high-energy marginal marine environments.

Sandstones of the Upper Cambrian Adedia Formation and its equivalent in the Levant, the Netafim Formation, differ from the older Cambrian sandstones of the area in being of quartz arenitic composition. They also show a closer grouping of individual provenance data points (Figure 6), coupled with extremely low eZi values (Figure 7). Since derivation by recycling of the older Cambrian sands appears to be precluded by petrological evidence (Amireh, 1991) and by the incompatibility of the heavy mineral provenance signatures (Figure 6), the greater degree of homogenisation suggests transport from a more distant source. A study of detrital zircons in the equivalent Netafim Formation of the Levant has shown, however, that they show a similar age distribution to those in the underlying sub-arkosic Shehoret Formation (Kolodner et al., 2006, p. 372). This indicates that although the Late Cambrian sands may have been more distantly sourced than the Middle Cambrian sands, they too were derived, in large part at least, from Neoproterozoic basement exposed on the Arabian-Nubian Shield.

Since the heavy mineral data show no evidence for successive recycling of earlier deposited sands, all of the Cambrian sands must be essentially of first-cycle origin. The mechanism for producing first-cycle sands of quartz-rich composition was investigated by Avigad et al. (2005), who concluded climatic conditions in the Early Cambrian were conducive to intense chemical weathering of crystalline basement outcrops. They argued that whereas the earlier arkosic and subarkosic Cambrian sands were derived through local erosion of unweathered basement rocks on the northern fringe of the Arabian-Nubian Shield, the overlying supermature quartz arenites of the Adedia and Netafim formations were generated by southward extension of the catchment area over a large expanse of intensely and deeply leached crystalline basement.

Although detrital zircon ages indicate that the majority of sand was derived from a Neoproterozoic basement source, zircons with pre-Neoproterozoic ages are also present and become increasingly common in the younger Lower Palaeozoic formations. Some of these may have been derived from pre-Neoproterozoic terranes on the southern fringes of the Arabian-Nubian Shield (Avigad et al., 2003; Kolodner et al., 2006). However, the presence of a small population of zircons with ages of 0.9–1.1 Ga (Kibaran) is more difficult to explain, since the nearest known basement of this age lies some 3,000 km to the south, in interior Gondwana. Avigad et al. (2003) and Kolodner et al. (2006) speculated that undetected Kibaran rocks might be present in the Arabian-Nubian Shield. This interpretation is supported by a study of the Neoproterozoic Negash diamictite in the Tigray region of northern Ethiopia (Avigad et al., 2007). The diamictite is clearly sourced from the nearby Arabian-Nubian Shield and while it has a zircon population that is primarily of Neoproterozoic age, a small Kibaran-age component is also present. Since no other pre-Neoproterozoic ages are represented, Avigad et al. (2007, p. 98) argued that the Kibaran zircons cannot have been sourced from outside the Arabian-Nubian Shield and that they must have been derived from Kibaran-age material that had been incorporated into the shield during the accretion process.

Another argument that has been put forward in support of a derivation of these sands from sources south of the Arabian-Nubian Shield concerns the abundance of vein-quartz pebbles (Weissbrod and Bogoch, 2007, p. 670-671). These authors considered that such pebbles have no obvious source on the Arabian-Nubian Shield and noted that they are absent from Shield-derived late Neoproterozoic conglomerates and arkoses. However, if a rapid and intense phase of weathering took place in the earliest Cambrian, as argued by Avigad et al. (2005), this could have led to the destruction of all but the most resistant pebble types and hence to a concentration within the regolith of what was originally a very minor vein-quartz component. In this context it may be noted that the Shield-derived Neoproterozoic diamictite at Negash, northern Ethiopia, contains pebbles of vein quartz, along with others of volcanic and sedimentary composition (Avigad et al., 2007, p. 92).

Carboniferous

According to Klitzsch and Wycisk (1987) and Klitzsch and Squyres (1990), Carboniferous sedimentation in Egypt took place under two tectonic regimes, separated by a period of E-W uplift in mid-Carboniferous times. To the south of this uplift, sedimentation resumed with the deposition of Late Carboniferous tills and glaciofluvial sediments, correlative with those of the Arabian Peninsula (e.g. Alsharhan et al., 1991). The tectonic uplift took place at the same time as the ‘mid-Carboniferous tectonic event’ as identified in the Arabian Peninsula (see Al-Husseini, 2004, figure 4). To the north of the uplift, the Carboniferous is represented by the Lower Carboniferous Um Bogma Formation, the Abu Thora Formation and the Upper Carboniferous Abu Durba Formation.

The Abu Thora Formation is said to range from Late Visean (Serpukovian) to Early Westphalian (late Bashkirian) in age (Kora, 1998), in which case it straddles the mid-Carboniferous hiatus as identified in the Arabian Peninsula. Although no hiatus has been identified within the Abu Thora Formation, it seems likely that some expression of the associated E-W uplift of central Egypt will exist within the Carboniferous sandstone succession. This might well be revealed by heavy mineral analysis, but stratigraphic coverage in the present study is not sufficiently complete to determine whether this is the case. In particular, it will be necessary to obtain data from the complete Abu Thora succession of the Um Bogma area and from the ‘Naqus Formation’ or ‘undifferentiated Carboniferous sandstones’ that underlie the Abu Thora Formation in outcrops to the south.

The only indication of a fundamental change in provenance signature within the Carboniferous sandstones is the appearance near the top of the Abu Thora Formation at Abu Durba (Figure 10) of sands with relatively high STi values. These sands and those of the overlying Abu Durba Formation display a distinctive trend of increasing STi values with increasing RZi values (Figure 11d) that is not seen in the Cambrian sands (Figure 11b). All the other Abu Thora sands have compositions that could possibly have been derived from recycling of the Cambrian sands or from a common source. Another possibility is that these sands could have been generated by reworking of Silurian and perhaps Ordovician sands, which may have been exposed as a result of the uplift of central Egypt. Such reworking could account for the very low eZi values in the Carboniferous sandstones at Abu Durba (Figure 11). This model, of erosion of Early Palaeozoic sandstones exposed on tectonic swells, is the one favoured by Weissbrod and Bogoch (2007).

Triassic and Cretaceous

According to the palaeogeographic reconstruction of Druckman (1974, figure 2) the fluvial sands of the Qiseib (Budra) Formation were derived from the Arabian-Nubian Shield. The heavy mineral assemblages are distinctive in possessing eZi values that are significantly higher than those in the underlying Carboniferous sandstones. This indicates that a significant component of the sands is of first-cycle origin and derived from basement outcrops that were not far distant.

A palaeogeographic reconstruction by Van Houten et al. (1984, figure 6) for the Lower Cretaceous ‘Facies 1’ (Malha Formation) fluvial sands in northern Egypt shows a highly irregular distribution of basement outcrops, with palaeocurrent directions indicating sand dispersal to the north and west of the Arabian-Nubian Shield. This was interpreted as representing infilling of a highly irregular topography, with a wider extension of sedimentation not taking place until the latest stage of ‘Facies 1’ sedimentation. Sediments of this late stage were described as better sorted and more mottled than the earlier deposits and as including common, poorly developed palaeosols. They are here interpreted as equating with the ‘Variegated Member’. Unlike the older sands, they display bimodal palaeocurrent indicators, suggestive of deposition in a marginal marine setting. This palaeogeographic setting implies that the Early Cretaceous sands of northeast Egypt were derived from immediately adjacent parts of the Arabian-Nubian Shield.

Heavy mineral index data are lacking for the ‘Raqaba Member’, but sands in the upper part of the ‘Temmariya Member’ at Gebel Musaba Salama (Figure 15) display a very limited range of RZi, MZi and STi values (Figures 15, 16). They also possess extremely low eZi values. The sands are thus both physically mature and homogeneous, indicating either that they were derived through reworking of pre-Mesozoic older sandstones or that they were derived from a distant source. By contrast, sandstones in the lower part of the overlying ‘Drab Member’ display high STi values and high eZi values (Figures 15, 16). These must include a substantial component derived directly from nearby crystalline basement as most of the index values are higher than any recorded from the pre-Cretaceous sandstones (Figure 17). There is some overlap with sandstones of the Sarabit El Khadim and Abu Hamata members, but the STi values in the ‘Drab Member’ range much higher. Comparable STi values are found in the Abu Durba Formation, but the eZi values are much lower. The uppermost analysed ‘Drab Member’ sample possesses much lower STi and RZi values, indicating a return to compositions similar to those of the Temmariya Member.

Sandstones of the Malha Formation at Abu Durba, assigned here to the ‘Variegated Member’ (see Weissbrod 2002), also possess high eZi values (Figures 10, 16). Unlike the high-eZi sandstones of the ‘Drab Member’ of Gebel Musaba Salama, these sandstones do not display high STi values. As discussed in an earlier section, they may well have been derived by reworking of sands of the Qiseib Formation.

Although the record is incomplete, therefore, the heavy mineral assemblages indicate that the sandstones of the Malha Formation were derived from both the crystalline basement and its pre-Cretaceous sedimentary cover. Previous authors have considered the Lower Cretaceous sandstones of the Levant and Jordan to have been derived almost wholly by reworking of Palaeozoic sands, with any direct contribution from the crystalline basement having been sporadic and localised. These conclusions were reached both on the basis of mineralogy and detrital zircon dating (Weissbrod and Nachmias, 1986; Amireh, 1992, 1994; Kolodner et al., 2009). These differing conclusions are not mutually exclusive, since it is quite possible that the successions in the Levant and Jordan were derived from a part of the Arabian-Nubian Shield in which basement exposure was relatively limited.

Climate and weathering signatures

The elucidation of climatic change through heavy mineral assemblages is based on the effect of weathering on the proportion of unstable minerals. The latter include ferromagnesian minerals, such as olivine, pyroxene, amphibole and epidote, along with apatite. The ferromagnesian minerals are the most unstable, being susceptible to the combined effects of oxidation and dissolution during under all but the coldest or most arid conditions. Apatite, on the other hand, is unstable only in waters of low pH (< 6) and is thus unstable at the surface when subjected to humid weathering conditions but is stable in groundwaters, which are rarely of such high acidity. For this reason, variation in apatite abundance can be a useful indicator of changing climate.

The climatic significance of variation in apatite abundance in the Sinai sandstones was discussed by Weissbrod et al. (1987). These authors argued that conditions suitable for the dissolution of apatite would not have existed until the late Palaeozoic, when the spread of terrestrial floras allowed the development of humic soils. They thus attributed the high apatite abundances recorded from the Lower Palaeozoic sandstones to a universal absence of acidic weathering profiles. A very different conclusion was reached by Guidry and Mackenzie (2003), however, who concluded on the basis of experimental studies that weathering-related dissolution of apatite in the early Cambrian was responsible for a build-up of phosphorus in the oceans and hence to the appearance of phosphate-shelled organisms. Indeed, the early Cambrian climate is may well have been especially conducive to acid leaching within the weathering profiles in the north Gondwana region (Avigad et al., 2005). These later conclusions are supported by the scarcity of apatite in, for example, the Cambrian sandstones of the Wajid Group in southwest Saudi Arabia (Hussain et al., 2004; Knox et al. (2007), since associated unweathered crystalline basement rocks and unweathered basement-derived sediments possess extremely high ATi values (Knox et al. 2007, table 4, sample W54A; Knox et al., 2010, Figure 31).

Variation in ATi values in the Cambrian sandstones of Sinai must, therefore reflect varying degrees of weathering of the sands prior to their final accumulation. Low ATi values could have resulted from weathering of the source rocks or weathering of the sands themselves prior to their burial below the vadose zone. High ATi values must reflect derivation from unweathered source rocks and a lack of significant weathering during transport and deposition. Stratigraphic variation in ATi values within a single succession, such as the Cambrian sandstones studied here, can be the result of variation in the climate or variation in the rates of erosion and transport. The fluctuating ATi values in the lower part of the Cambrian succession in Wadi Nasib (Figure 5, below 35 m) most probably reflect the latter, since there is broad positive correlation between ATi and eZi values (Figure 18). In contrast, there is no such relationship in the upper part of the succession, where most samples possess very low eZi values and where ATi values are either extremely high or extremely low (Figure 18). The extremely high values indicate sand supply from an unweathered source and occur in sandstones with differing provenance signatures. These high values are here interpreted as representing a period of relatively cool or dry climate during which acid leaching of apatite was greatly reduced. This interpretation is supported by the presence of aeolian dune sandstones at a level equivalent to the uppermost Abu Hamata Member of northwest Saudi Arabia (Janjou et al., 1998: uppermost Siq Formation; Al-Husseini, 2010: uppermost Burj Formation). The upward change to extremely low ATi values above 60 m in the Wadi Nasib section is interpreted as indicating a return to a more humid climate and to more intense chemical weathering.

The virtual absence of apatite from the Late Palaeozoic onwards is almost certainly related to recycling of older apatite-free sands and to a high degree of chemical weathering associated with the development of humic soils (Weissbrod et al., 1987). Two notable exceptions to the general scarcity of apatite are found in the Abu Durba section (Figure 10), one at the base of the Abu Thora Formation and the other at the base of the Abu Durba Formation. In both instances, the high ATi values are associated with relatively high percentages of the unstable minerals amphibole, pyroxene and epidote (Table 2). Both instances are associated with strongly transgressive phases, during which much of the sand may have been generated by coastal erosion and thus not subjected to significant syndepositional weathering. By contrast, the overlying sections represent regressive phases, during which the sands would have experienced substantial leaching during their transit across the alluvial floodplain (see Morton and Hallsworth, 1999).

CONCLUSIONS

This study has shown that the use of heavy mineral indices rather than percentages reveals a much greater degree of stratigraphic mineral variation within the Palaeozoic and Mesozoic sandstones of southwest Sinai. Although a full assessment of the stratigraphic variation is precluded by incomplete stratigraphic coverage, the present study allows the following observations to be made:

  • (1)The Cambrian succession comprises three principal mineral-stratigraphic units, with boundaries at the base of the late Early Cambrian, within the Middle Cambrian and at the base of the Late Cambrian. The successive changes in mineralogy are interpreted as resulting from tectonic restructuring of source areas on the Arabian-Nubian Shield and are thus likely to be represented in other parts of the region, including the offshore succession of the Gulf of Suez.

  • (2)Variation in the relative abundance of detrital apatite indicates that Cambrian sedimentation mostly took place under humid climatic conditions, but that a short period of dry climate occurred during the Middle Cambrian.

  • (3)The Lower to Upper Carboniferous succession, comprising the Abu Thora Formation and the underlying ‘undifferentiated Carboniferous sandstones’, appears to include at least two distinct mineralogical units, which may represent separate depositional sequences. Significant stratigraphic variation in mineralogy, reflecting sharp changes in provenance, also occurs within the Abu Thora Formation and has the potential for regional subdivision and correlation.

  • (4)The sandstones of the Upper Carboniferous Abu Durba Formation have a distinctive mineralogy that indicates sand supply from newly exposed basement rocks. They clearly represent a separate depositional sequence.

  • (5)Sandstones of the Triassic Qiseib Formation are physically immature, reflecting rejuvenation of the source area following the ‘Hercynian’ tectonic phase.

  • (6)Heavy mineral assemblages in the Lower Cretaceous sandstones of the Malha Formation indicate successive phases of rejuvenation of the Arabian-Nubian Shield, accompanied by changes in the topographic configuration of the source area. Because these mineral units appear to be associated with separate cycles of uplift and subsidence, it is likely that they will be represented in all areas adjacent to the Arabian-Nubian Shield, including the offshore sections in the Gulf of Suez.

  • (7)The occurrence of units with highly distinctive provenance signatures throughout the Palaeozoic and Mesozoic sandstone succession will allow improved identification of sandstone units in isolated outcrops or in the subsurface.

  • (8)Future studies should concentrate on two aspects of the mineral stratigraphy: completion of the stratigraphic coverage by sampling complete Carboniferous and Early Cretaceous sandstone sections, and enhancement of palaeogeographic interpretations by assessing lateral variation in provenance signatures within individual lithostratigraphic units.

ACKNOWLEDGEMENTS

The authors are grateful to Dr. John Powell for reviewing an earlier version of this work and to an anonymous reviewer for his very thorough and constructive review. R.W.O’B.K. publishes with the approval of the Executive Director, British Geological Survey (NERC).

ABOUT THE AUTHORS

Robert W. O’B. Knox is an Independent Consultant specializing in the application of heavy mineral analysis to the correlation and delineation of reservoir sandstones in the Paleozoic and Mesozoic successions of the Arabian Peninsula and elsewhere. He was formerly with the British Geological Survey, where he gained over 25 years experience on the stratigraphy and sedimentology of UK and overseas petroleum basins. He is currently an Honorary Research Associate of the BGS.

rwok@btinternet.com

Mamdouh F. Soliman is an Associate Professor of mineralogy and chemostratigraphy at the department of Geological Sciences, Assiut University, Egypt. He received his PhD (1998) at the Ruprecht-Karls-Universitäte Heidelberg, Germany. He directed his investigations towards the stratigraphically significant events of the Upper Cretaceous – Lower Paleogene. His research interests include petrology, mineralogy and chemostratigraphy of the Cretaceous – Tertiary, Danian – Selandian and Paleocene – Eocene sediments in Egypt.

msoliman20032003@yahoo.com

Mahmoud A. Essa is an Associate Professor of mineralogy and geochemistry at the department of Geological Sciences, Assiut University, Egypt. He received his PhD (2000) at the Assiut Universty, Egypt. His research interests include the Upper Cretaceous – Eocene and Neogene mixed siliciclastic-carbonate sediments in Egypt.

maessa2002@yahoo.com