Introduction

The breakup and dispersal of Pangaea and Gondwana as its counterpart began in the late Paleozoic (Muttoni et al. 2009) and continued throughout the Mesozoic with repeated events of rifting and tectonic re-organizations. The sequence and the timing of the related geodynamic processes, as well as the nomenclature of Gondwana-derived blocks, have been analyzed and discussed in numerous studies (Stampfli and Borel 2002; Jokat et al. 2003; Golonka 2007; Blakey 2008; Torsvik and Cocks 2013; Vérard et al. 2015; Jordan and Becker 2018; Müller et al. 2019; Thompson et al. 2019; Gatti et al. 2021; Kocsis and Scotese 2021; Scotese 2021). Africa remained at the very core of the fragmented supercontinent, and the Mesozoic sedimentary archives of its African–Arabian periphery provide important clue to the understanding of this supercontinent’s evolution punctuated by breakup and dispersal events (Guiraud et al. 2005).

The Mesozoic development of northeastern Africa is known relatively well (Guiraud et al. 2005; Issawi et al. 2009; Said 2017; Hamimi et al. 2020), but many details remain uncertain. For instance, it is well-known that ancient rivers drained the continental interiors (the territory of present Egypt) and flowed toward the nearby Tethys Ocean (Bhattacharyya and Dunn 1986; Allam 1989; Beauchamp et al. 1990; Hendriks et al. 1990; El-Azabi and El-Araby 2005; Abd-Elshafy and El-Azeam 2010; Ruban and Sallam 2018, 2019; Sallam and Wanas 2019), but it is yet to be fully understood whether the related depositional systems changed fundamentally through time and what were the source rocks. It would be an oversimplification to assume that the rivers remained constant and they eroded only crystalline (largely Pan-African) rocks in the continent’s interiors. New stratigraphical, petrographic, and geochemical studies, as well as innovative provenance interpretations, are required to shed light on such geological evolution.

Field investigations undertaken in the northwestern periphery of the Gulf of Suez in Egypt have permitted to collect new information about the Mesozoic fluvial sandstones. The purpose of the present paper is to shed light on their composition and provenance, to extend the knowledge about Mesozoic fluvial systems near the edge of Gondwana. The examined sandstones are linked to the different stages of the evolution of this continental domain, from the Pangaea breakup in the beginning of the era to the significant fragmentation and final dispersal in the Middle-Late Jurassic and Early Cretaceous. The comparison of the related episodes of fluvial sedimentation and a detailed illustration of similarities and differences among several sandstone packages is of crucial importance for understanding the changes of the geological environment in this part of Gondwana.

Geological setting

The study area is located in the northwestern part of the Gulf of Suez. Geographically, it comprises the Northern Galala Plateau and the Gulf of Suez coastal plain. This area represents the northeastern part of the African continent where it is bounded by the relatively young Red Sea rift system including the Gulf of Suez failed rift arm. Before the Oligocene rifting (Garfunkel 1988; Khalil and McClay 2001; Boone et al. 2021; Saada et al. 2021), this was the northern periphery of a larger African–Arabian continental block of Gondwanan origin (Guiraud et al. 2005; Jagger et al. 2020).

In the study area, Paleozoic, Mesozoic, and Cenozoic sedimentary sequences crop out (Fig. 1). The Rod El-Hamal, Abu Darag, and Aheimer formations belong to the Carboniferous (Allam 1989; Abd-Elhameed et al. 2020). These are mixed siliciclastic–carbonate deposits with the total thickness of > 300 m. The Mesozoic units include the Qiseib, Rieina, Ras El-Abd, and Malha formations dominated by fluvial sandstones.

Fig. 1
figure 1

Geological map and composite stratigraphic section in the northwest Gulf of Suez (modified after CONOCO 1987, scale 1: 500,000)

Full size image

The Qiseib Formation was proposed by Abdallah and Adindani (1963) and later studied by Allam (1989), El-Barkooky (1986), Alsharhan and Salah (1997), Kora (1998), Guiraud et al. (2005), and Wanas and Soliman (2018). This formation is mainly composed of medium to coarse, lenticular, oblique-laminated sandstones, locally ferruginous, intercalating with siltstones and claystones, and punctuated by several paleosol horizons that display characteristic pedogenic features of color-mottling, scours, desiccation cracks, and iron-crusts (Fig. 2A). The total thickness is ~ 60 m. The age of this formation remains debated, but the available lines of evidence (see literature above) permit to assign it to the Early Triassic. Plant remains, lenticular and oblique lamination, and scour surfaces suggest channel deposition in a fluvial depositional system (Issawi et al. 2009). The sandstones accumulated in channels and point bars of braided and meander streams, respectively (Soliman and Wanas 2006). Siltstones and claystones indicate overbank (floodplain) sedimentation. The occurrence of paleosol horizons implies subaerial exposure and paleopedogenic modifications in arid and semi-arid environments (e.g., Sallam et al. 2015a, 2015b; Wanas et al. 2015). The Qiseib Formation is unconformably overlain by Jurassic and/or Lower Cretaceous deposits (Fig. 2B).

Fig. 2
figure 2

General views of the studied rock units: A, B ferruginous, oblique-laminated sandstones of the Qisieb Formation overlain by the Malha Formation at Wadi Qiseib; C outcrop photograph of the Rieina and Ras El-Abd formations at Khashm El-Galala; D oblique-laminated sandstones of the Malha Formation at Wadi Malha; E, F alternating claystones, siltstones, and sandstones of the Malha Formation at the southern cliffs of the Northern Galala Plateau; G paleosol bed displaying colour mottling within the Malha Formation at Wadi Malha

Full size image

The Rieina and Ras El-Abd formations are well-exposed in the study area (Abd-Elshafy 1981; Hassanein 1985) (Fig. 2C). The Rieina Formation consists mainly of oblique-laminated sandstones with a total thickness of ~ 70 m. These sandstones are capped by a 1.5-m-thick claystone bed containing abundant well-preserved imprints of macrofloral remains and by a 0.5-m-thick bed with Thalassinoides sp. The Ras El-Abd Fm. is composed of alternating sandstones, siltstones, and claystones, with several limestones rich in brachiopods, gastropods, and bivalves. The total thickness is ~ 150 m. The Rieina and Ras El-Abd formations are Middle–Late Jurassic in age as suggested by rhynchonellids and foraminifers (Abd El-Shafy, 1981; Hassanein 1985). The Rieina and Ras El-Abd formations were deposited in fluvial (lower part) and subtidal-lagoonal shallow-marine (upper part) environments of a platform-interior setting influenced by strong tidal currents and storm waves (El-Younsy 2001; Ruban and Sallam 2016; Ruban et al. 2019). The Rieina and Ras El-Abd formations unconformably overlie the Qiseib Formation, and they are overlain also unconformably by the Malha Formation.

The Malha Formation, introduced by Abdallah and Adindani (1963) and studied by El Beialy et al. (2010) and Wanas et al. (2015), consists of kaolinitic pebbly sandstones interbedded with blackish-grey claystones and siltstones in the lower part. Its middle part is dominated by medium to coarse sandstones exhibiting characteristic planar and trough oblique lamination (Fig. 2D). The upper part of the formation is composed of massive sandstones alternating with thin-bedded, laminated siltstones, and claystones (Fig. 2E, F) and punctuated by several paleosol horizons (Fig. 2G), in addition to the presence of slightly convolute lamination, hematite-crusts, and load casts. The Malha Formation was deposited in fluvial/deltaic environments, has a thickness of ~ 80 m, and is Aptian–Albian in age (Abdallah and Adindani 1963). Oblique laminated sandstones in the middle part of the formation indicate downstream accretion of bedforms in fluvial channels (Miall 1996). Some interbeds of floodplain and deltaic siltstones and claystones were exposed subaerially with paleosol formation in arid and semi-arid conditions (e.g., Wanas et al. 2015; Zobaa et al. 2015). The Malha Formation underlies unconformably the Galala Formation (Fig. 3).

Fig. 3
figure 3

The Malha Formation overlain by the Cenomanian Galala Formation at the southern cliffs of the Northern Galala Plateau

Full size image

The Upper Cretaceous–Eocene carbonate deposits of the Galala, Wata, and Thebes formations have a total thickness exceeding 200 m (Saber 2012; Abdel-Fattah et al. 2018; Ahmed and Afife 2018; Issawi et al. 2018). The youngest rocks are wadi deposits widely distributed in the southern part of the study area (Fig. 1). The contacts of the majority of the above-mentioned rock packages are disconformable, and they reflect deformation phases, uplifts, and regression episodes (Guiraud et al. 2005; Issawi et al. 2016; Issawi and Sallam 2018).

The study area is located at the northeastern periphery of the African continent affected by the development of the Red Sea rift system. Although there was a single African–Arabian block before the mid-Cenozoic, it was affected by several phases of tectonic and magmatic activity in the Mesozoic–early Cenozoic (Guiraud et al. 2005). Rifting started in the Oligocene (Garfunkel 1988; Khalil and McClay 2001; Sallam et al. 2019, 2022; Boone et al. 2021), initiating a new phase of palaeogeographic reorganization (Guiraud et al. 2005; Segev et al. 2017) as documented by extensive faulting and block structure (Hussein and Abd-Allah 2001).

Material and methods

Sampling and petrography

The present study focuses on Mesozoic fluvial deposits represented in the study area by the Lower Triassic, Middle Jurassic, and Lower Cretaceous Qiseib, Rieina, Ras El-Abd, and Malha formations, respectively. These units were examined in three representative sections, namely Khashm El-Galala, Wadi Qiseib, and Wadi Malha (Fig. 4), where a total of 120 sandstone and claystone samples were collected.

Fig. 4
figure 4

Correlation of the three measured stratigraphic sections. Numbers refer to the collected samples

Full size image

Sandstone petrography analyses were carried out on 28 thin sections under the polarizing microscope and their description followed the classification schemes of Garzanti (2016, 2019) and Garzanti et al. (2018a). Approximately 200–300 grains were counted on each sample according to the Gazzi–Dickinson point-counting method (formalized by Ingersoll et al. 1984). Lithic fragments were classified according to protolith composition and metamorphic rank (Garzanti and Vezzoli 2003).

X-ray diffraction (XRD) analysis was performed on 10 claystone and clayey sandstone samples (Fig. 4). Analytical X-ray diffraction equipment model X′Pert PRO with secondary monochromator, Cu-Kα radiation (λ = 1.542 Å) operating at 45 K.V., 35 M.A., and scanning speed 0.04°/s were used. The diffraction peaks between 2θ = 2° and 60°, corresponding spacing (d, Å), and relative intensities (I/I°) were obtained. The diffraction charts and relative intensities were compared with ICDD files. The samples were processed using zero background holder.

Geochemistry

A Thermo Scientific ARL 9900 X-ray fluorescence (XRF) spectrometer was used for measuring whole-rock major elements. This analysis was performed with the facility of the State Key Laboratory for Mineral Deposits Research of the Nanjing University (China). All samples were crushed into gravel-size chips. The latter were then powdered to 200 mesh. As suggested by the reference materials BHVO-2 and BCR-2, the uncertainties are less than ± 3% for Si, Ti, Al, Fe, Mn, Mg, Ca, K, and P and less than ± 6% for Na.

Trace and rare earth elements (REEs) were analyzed by inductively coupled plasma emission spectrometry (ICP-MS) at the same laboratory. The analytical ICP-MS procedures started with obtaining clean dissolved sample jars by vigorously wiping and heat soaking the PTFE dissolved sample jars in 50% HNO3 and 10% HNO3 solutions. Sample testing was performed on a Thermo Scientific Element XR mass spectrometer. Rh was used as an internal standard to correct changes in mass spectrometry sensitivity. An unknown sample of similar lithology was tested at intervals of about 10 samples, and the results were corrected for the unknown sample before and after the test by the ratio of the test results to the reference value to reduce the bias caused by the basal effect of different lithology samples. Blanks and replicates were tested to monitor data quality. All major, trace, and REEs whole-rock geochemical data are compared with the average values in the upper continental crust (UCC; Rudnick and Gao 2003) and post-Archean Australian Shale (PAAS; McLennan 1989).

Results

Petrographic characteristics and detrital modes

The studied Mesozoic sandstones from the northwestern Gulf of Suez are mostly quartzose, ranging from quartz-rich litho-quartzose to pure quartzose, with average quartz content ~ 86% of the detrital components (Table 1, Figs. 5 and 6). Monocrystalline quartz prevails (average ~ 71%), exhibiting unit (72%) or undulatory (28%) extinctions (Fig. 5). Polycrystalline quartz grains (> 2 crystals per grain) are also frequent and form in average ~ 29% of total quartz (Fig. 5). Other detrital constituents are represented by lithic fragments (average ~ 5%), feldspars (average ~ 1.3%), and heavy minerals (average ~ 7%). Detrital grains have angular, sub-angular, sub-rounded, and rounded shapes. Grain size and sorting vary widely from fine to very coarse and from poor to fair. The matrix consists mostly of quartzose silt. Cements are represented chiefly by authigenic kaolinite, quartz overgrowths and ferruginous material (Fig. 5). The grain-to-grain contacts are point, long, concavo-convex, and stylolite (sutured) contacts (Fig. 5). Contact type is largely depend on the relevant packing variety in which loosely-packed sandstone is generally dominated by no and/or point contacts, whereas closely-packed varieties display more concavo-convex and stylolitic contacts. Intercrystal boundaries within polycrystalline quartz are crenulated to sutured and long to slightly curved.

Table 1 Detrital modes (%) of the studied sandstones
Full size table
Fig. 5
figure 5

Photomicrographs of the studied sandstones: A–C poorly-sorted quartz-rich litho-quartzose sandstone dominated by monocrystalline (Qm) and polycrystalline (Qp) quartz grains and heavy minerals (hvy) embedded in phyllosilicate cement; some quartz grains display undulose extinction (U), Qiseib Formation. The red arrow points to a sutured intercrystal boundary within polycrystalline quartz, whereas the black arrow shows a long contact; D poorly-sorted litho-quartzose sandstone consisting predominantly of monocrystalline quartz (Qm), lithics (Lf), and plagioclase (P), Qiseib Formation; E, F quartzose sandstone of the Rieina Formation dominated by ill-sorted monocrystalline and polycrystalline quartz grains, some of them displaying undulose extinction (U). The red arrows points at stylolitic contact between quartz grains; G litho-quartzose sandstone consisting mainly of monocrystalline (Qm), polycrystalline (Qp) quartz, and lithic fragments (Lf) embedded in phyllosilicate cement, Malha Formation. The black arrow points at a long contact between quartz grains; H pebbly sandstone of the Malha Formation composed mainly of coarse-grained polycrystalline (Qp) quartz and lithic fragments (Lf) cemented by phyllosilicate. Some quartz grains show undulose extinction (U). The red arrows indicate curved and sutured intercrystal boundaries within polycrystalline quartz

Full size image
Fig. 6
figure 6

Overall compositional homogeneity of the studied Mesozoic sandstones. Q quartz, F feldspars, L lithic fragments, ZTR + zircon + tourmaline + rutile, qFQ and qLQ quartz-rich feldspatho-quartzose and litho-quartzose sandstones (classification after Garzanti 2019). The biplot (centre; Gabriel 1971) displays multivariate observations (points) and variables (rays). The length of each ray is proportional to the variance of the corresponding variable; if the angle between two rays is 0° or 180°, then the corresponding variables are perfectly correlated or anticorrelated

Full size image

K-feldspar with cross-hatched microcline predominates in the Lower Cretaceous Malha Formation, whereas plagioclase prevails and microcline was not observed in the Lower Triassic Qiseib Formation and Middle-Late Jurassic Rieina and Ras El-Abd formations. Lithic fragments are dominantly siltstone and sandy siltstone, with a few slate and metasiltstone, limestone, dolostone, and rare quartz-muscovite schist and gneiss (Table 1). Carbonate grains are more common in older units. One dubious serpentinite grain was observed in one sandstone sample from the Qiseib Formation. Heavy minerals are mainly zircon, rutile, tourmaline, epidote, garnet; monazite; opaque Fe-Ti-Cr oxides are also frequently observed. Intersample mineralogical variability is significant, but of similar magnitude within each formation (Fig. 7).

Fig. 7
figure 7

Overview of compositional and textural features of the studied sandstone: A Percentage of quartz, feldspars, and lithic fragments based on point counting; B Sorting (1—poor, 2—moderate, 3—good, 4—very good) and roundness (1—angular, 2—subangular, 3—subrounded, 4—rounded grains); C Average grain size

Full size image

The XRD analysis of claystone and clayey sandstones show dominant quartz (15–75%, average ~ 54%) and kaolinite (12–55%, average ~ 29%) (Fig. 8, Table 2). Alunite and talc were recorded in samples 5c and 8c, respectively, from the Ras El-Abd Formation (Table 2). No other clay minerals were detected. The XRD peaks also document occurrence of chabazite and goethite in clayey sandstones of the Malha Formation. Mica (most commonly muscovite) is found in a few samples from the Qiseib and Ras El-Abd formations. Hematite, calcium carbonates, and halite were also recorded (Table 2).

Fig. 8
figure 8

X-ray diffraction pattern of the minerals identified in the studied claystone and clayey sandstone samples. A Malha Formation; B Ras El-Abd Formation; C Qiseib Formation

Full size image
Table 2 Semi-quantitative X-ray diffraction (XRD) data of the analyzed claystone and clayey sandstone samples
Full size table

Diagenetic features

The studied sandstones exhibit a variety of diagenetic features resulting from compaction, cementation, dissolution, and kaolinitization. Overcompaction is indicated by grain fracturing and pressure dissolution along intergranular contacts between quartz grains resulting in concave-convex and suture grain-to-grain contacts. Silica released from pressure solution of detrital quartz is a potential source for syntaxial quartz overgrowths. Authigenic kaolinite is common and occurs as pore-filling cement in well-sorted sandstones. Goethite is the main iron oxide recognized in sandstone samples, with minor hematite.

Whole-rock geochemistry of sandstones

All studied samples show high percentages of SiO2, ranging between 48.2 and 99.9%, with an average of 84.5% and lower concentrations of Fe2O3tot (0.01–12.4, average 2.0%). Compared to both UCC and PAAS, the studied sandstone samples revealed a depletion in TiO2 (0.04–1.9%, average 0.4%), MnO (0.001–0.3%, average 0.02%), P2O5 (0.005–0.06%, average 0.02%), Na2O (average 0.36%), K2O (average 0.2%), CaO (average 1.9%), and MgO (average 0.2%) (Table 3). Lower Triassic and Lower Cretaceous sandstones show higher concentrations of Al2O3 (average 8.9% and 6.4%, respectively) than Middle-Late Jurassic sandstones (average of 2.2%).

Table 3 Major element concentrations and averages (in wt %), CIA, CIW, and ratio values of the studied sandstones
Full size table

Sandstone samples display low Al2O3/SiO2 ratio (average 0.1) and low K2O/Na2O ratio (average 0.6), and relatively high Al2O3/TiO2 ratio (average 16.1) (Table 3). Middle-Late Jurassic sandstones have higher K2O/Na2O ratio (average 1.2) than older and younger sandstones. In addition, Al2O3 shows positive linear correlation with both TiO2 (r2 = 0.89) and Fe2O3tot (r2 = 0.73) (Fig. 9A–E). Loss is positively correlated with CaO (r2 = 0.65) (Fig. 9F).

Fig. 9
figure 9

Correlation diagrams of selected oxides

Full size image

Most trace elements are depleted relative to both UCC and PAAS, excepting Zr (184 ppm on average), Hf (4.2 ppm), and Nb (7.7 ppm) (Table 4, Figs. 10 and 11). Sandstones of the Qiseib and Malha formations (Middle-Late Jurassic samples were not analyzed) exhibit a chondrite-normalized rare earth elements (REEs) pattern (Fig. 12) with high light REEs fractionation, low heavy REEs fractionation, and negative Eu/Eu* anomalies (Table 5). Eu anomalies (Eu/Eu*) were calculated as Eu/Eu* = (Eu)cn/[(Sm)cn × (Gd)cn]0.5 (from McLennan 1989), and Ce anomalies (Ce/Ce*) as Ce/Ce* = (Cecn)/[(Lacn + Prcn)]0.5, where cn is the chondrite-normalized values of the element (Taylor and McLennan 1985; Barrat et al. 2012). The Eu/Eu* anomaly ranges from 0.45 to 0.79 (average 0.56) in Qiseib sandstones and from 0.47 to 0.61 (average 0.58) in Malha sandstones. Normalized (La/Yb)cn, (La/Sm)cn, and (Gd/Yb)cn ratios are provided in Table 5. The total rare earth elements (ΣREE) ranges from 23 to 80 ppm (average 44 ppm) in the Qiseib sandstone samples, and from 16 to 159 ppm (average 53 ppm) in the Malha samples. Values normalized to average upper continental crust (UCC) for selected trace elements, including mobile Ba, Rb, and Sr and high-field strength non-mobile Hf, Nb, and Zr, are shown in Fig. 13.

Table 4 Concentrations of selected trace elements and averages (ppm) and ratio values of the studied sandstones
Full size table
Fig. 10
figure 10

Correlation diagrams of Al2O3 with Co, Ni, Cu, Zn, and V

Full size image
Fig. 11
figure 11

Correlation diagrams of Fe2O3tot with Co, Ni, Cu, Zn, and V

Full size image
Fig. 12
figure 12

Chondrite-normalized REE diagrams for samples from the Qiseib and Malha formations. UCC and PAAS values from Rudnick and Gao (2003) and McLennan (1989), respectively

Full size image
Table 5 Rare earth element concentrations and averages (in ppm) and ratios of the studied sandstones
Full size table
Fig. 13
figure 13

UCC-normalized multi-element diagrams for the studied sandstones. Normalizing values from Rudnick and Gao (2003)

Full size image

Discussion

The compositional signatures of sandstones are controlled by several factors, including the lithology of source rocks as well as physical and chemical processes acting during erosion, transport, deposition, and diagenesis (Johnsson 1993). Information on provenance and paleoenvironmental conditions can be obtained only after careful consideration of the potential effects of diagenetic dissolution.

Diagenesis

Heavy minerals are particularly susceptible to post-depositional leaching, depending on their chemical composition and crystallographic structure as well as on the geochemical environment, thermal gradient, and character of intrastratal fluids. Although these conditions may change from place to place and through time, the durability of rock-forming minerals during diagenesis is generally considered to follow in reverse the order of crystallization from high-temperature melts indicated by the Bowen series (Goldich 1938). A similar sequence of progressive mineral disappearances with burial depth is in fact observed in diverse sedimentary basins (Milliken 2007; Morton and Hallsworth 2007). A typical vertical sequence of mineral facies (“diagenetic minerofacies”) can be defined and used as a standard for comparison when trying to assess diagenetic control in ancient sedimentary successions (Garzanti et al. 2018b).

In the studied sandstones, the lack of chemically labile ferromagnesian minerals such as amphibole, pyroxene, or olivine; the observed presence of epidote in all samples but two; and of garnet in all samples but six (Table 1) can be ascribed to selective diagenetic dissolution rather than to absence in the originally deposited sediment. These features clearly indicate the garnet-epidote minerofacies, generally attained at burial depths of 2–3 km. In the study area, however, the thickness of the sedimentary cover is insufficient to support such a relatively deep burial (Fig. 1). Intense diagenetic dissolution, therefore, suggests a high thermal gradient developed during Miocene rifting (Garfunkel 1988; Khalil and McClay 2001; Boone et al. 2021). Alternatively, diagenetic dissolution may have been inherited by recycling of older, already diagenetically-modified sandstones.

Further clues on the extent of diagenetic dissolution are offered by the relative stability of tectosilicates. Because the removal rate of Na+ and Ca2+ from plagioclase during active leaching generally exceeds the removal rate of K+ from K-feldspar, under most circumstances plagioclase results as the least stable among the feldspar group and the low-temperature-ordered polymorph microcline as the most stable (Blatt 1967; Nesbitt et al. 1997; Garzanti et al. 2021).

In the studied sandstones, the plagioclase/feldspar (P/F) ratio is markedly lower in the Malha Formation (where microcline was commonly observed) than in the older Lower Triassic and Middle-Late Jurassic units (where microcline was not observed). Moreover, two sandstone samples from the Malha Formation resulted to be the only ones lacking both epidote and garnet (Table 1). Such a difference can hardly be considered as a diagenetic effect, because the Malha Formation lies stratigraphically above and, thus, was affected by lower-temperature burial conditions than the underlying units. Nonetheless, these interpretations are done with serious caution due to the low content of feldspars.

Provenance

The quartzose composition of the studied sandstones (Fig. 5), together with the predominance of terrigenous to low-rank meta-terrigenous grains and sporadic occurrence of carbonate grains (Table 1), indicates that these siliciclastic rocks were chiefly derived from erosion of a sedimentary succession. The rare metamorphic lithic fragments and heavy minerals such as epidote and garnet may have also been recycled from older sandstones, although first-cycle provenance from the Pan-African basement cannot be excluded for a minor part of the sediment.

The overall homogeneity of both petrographic and heavy-mineral modes (Fig. 6) does not indicate a major difference in provenance throughout the Mesozoic. The subtle compositional differences characterizing the Malha Formation, including a lower P/F ratio, lesser feldspar content, the presence of microcline, locally ZTR-dominated heavy-mineral assemblages, and rarity of carbonate lithics (Table 1), point at increasing proportion of detritus recycled from older (hence last-eroded) and more strongly diagenized siliciclastic units of Paleozoic age. We envisage that deepening of erosion into stratigraphic successions exposed in source areas may have been promoted by tectonic reactivation affecting north-western Africa in the Early Cretaceous, as originally inferred by Bhattacharyya and Dunn (1986).

Sandstone composition and palaeogeography

The studied Lower Triassic, Middle-Late Jurassic, and Lower Cretaceous sandstones represent three distinct Mesozoic pulses of fluvial sedimentation, even though they are characterized by broadly homogeneous petrographic (Fig. 6) and geochemical signatures (Figs. 9, 10, 11, 12, and 13). Compositional variability is of the same magnitude among and within the three studied formations, a rather unexpected fact if we consider the major gaps occurring between sandstone packages of different age. In all three cases, extensive recycling of older siliciclastic rocks is indicated, thus implying successive stages of erosional exhumation of Paleozoic sedimentary successions (Sallam and Ruban 2021).

In the Early Triassic, the study area was a small intracontinental sedimentary basin fed from uplands located in northeastern Africa and/or northwestern Arabia (Guiraud et al. 2005). In the Middle–Late Jurassic, the study area became a narrow coastal zone between the sea in the north and continental uplands in the south (Guiraud et al. 2005). Palaeogeography changed even more conspicuously in the Lower Cretaceous, when the study area was a low-lying peninsula at the northeastern edge of Africa, surrounded by epeiric seas in the north, west, and east but connected with the Arabian landmass that represented the main sediment source (Guiraud et al. 2005).

Surprisingly, these palaeogeographical changes (Fig. 14) are not reflected by the petrographic and geochemical composition of the studied Mesozoic sandstones. Their rather homogeneous composition, however, may not necessarily indicate unchanged source areas, because the same, mostly siliciclastic lithosome representing their most plausible major detrital source was deposited after the main Pan-African event in a vast area ranging from Oman to Mauritania (e.g., Burke 1999; Avigad et al. 2005). Apparently, during multiple rejuvenation events, thick sedimentary packages accumulated across this part of Gondwana (especially in the Early Paleozoic during and after the last stages of the major Pan-African orogeny) and were subsequently incised by rivers with spatially restricted drainage networks.

Fig. 14
figure 14

Palaeogeographical setting of the study area in the considered time slices. Global reconstructions are simplified from Kocsis and Scotese (2021), and regional reconstructions are based on Guiraud et al. (2005)

Full size image

Further clues on provenance and palaeoclimate are provided by clay mineralogy (Keller 1956; Girard et al. 2000). In the studied samples, kaolinite is dominant and chabazite and goethite occur (Fig. 8). Kaolinite is typically formed by intense leaching of feldspars or other aluminosilicates under hot-wet climate (Girard et al. 2000), but may also be detrital or grown during diagenesis. The occurrence of chabazite suggests derivation from altered basaltic rocks, whereas goethite may indicate extensive weathering of iron-rich sediments and paleosols (Zee et al. 2003). In addition, the presence of alunite in some samples from Jurassic rock units indicates alteration of K-feldspar-rich volcanic rocks by acid-bearing solutions. Fluvial sedimentation in tropical conditions is thus being suggested.

Conclusions

This petrographic and geochemical study of Lower Triassic, Middle-Late Jurassic, and Lower Cretaceous sandstones from the northwestern Gulf of Suez leads to the following major conclusions:

  1. 1.

    The studied sandstones are mainly quartzose, ranging from quartz-rich litho-quartzose to pure quartzose, and their petrographic and geochemical characteristics do not conspicuously change throughout the Mesozoic succession.

  2. 2.

    The lack of major provenance changes suggests extensive recycling of Paleozoic strata widely exposed across northern Gondwana and rule out first-cycle provenance from Pan-African basement rocks exposed in the continental interiors.

  3. 3.

    Progressive deepening of erosion into older siliciclastic units is hypothesized.