Endogenous and nonimpact origin of the Arkenu circular structures (al-Kufrah basin—SE Libya)

Authors

  • Corrado CIGOLINI,

    Corresponding author
    1. Dipartimento di Scienze della Terra, Università degli Studi di Torino, Via Valperga Caluso 35, 10125 Torino, Italy
    2. NatRisk, Centro Interdipartimentale sui Rischi Naturali in Ambiente Montano e Collinare, Università degli Studi di Torino, Torino, Italy
      Corresponding author. E-mail: corrado.cigolini@unito.it
    Search for more papers by this author
  • Mario DI MARTINO,

    1. INAF–Osservatorio Astronomico di Torino, 10025 Pino Torinese, Italy
    Search for more papers by this author
  • Marco LAIOLO,

    1. Dipartimento di Scienze della Terra, Università degli Studi di Torino, Via Valperga Caluso 35, 10125 Torino, Italy
    Search for more papers by this author
  • Diego COPPOLA,

    1. Dipartimento di Scienze della Terra, Università degli Studi di Torino, Via Valperga Caluso 35, 10125 Torino, Italy
    Search for more papers by this author
  • Piergiorgio ROSSETTI,

    1. Dipartimento di Scienze della Terra, Università degli Studi di Torino, Via Valperga Caluso 35, 10125 Torino, Italy
    Search for more papers by this author
  • Marco MORELLI

    1. Fondazione Prato Ricerche, Museo di Scienze Planetarie, 59100 Prato, Italy
    Search for more papers by this author

Corresponding author. E-mail: corrado.cigolini@unito.it

Abstract

Abstract– The twin Arkenu circular structures (ACS), located in the al-Kufrah basin in southeastern Libya, were previously considered as double impact craters (the “Arkenu craters”). The ACS consist of a NE (Arkenu 1) and a SW structure (Arkenu 2), with approximate diameters of about 10 km. They are characterized by two shallow depressions surrounded by concentric circular ridges and silica-impregnated sedimentary dikes cut by local faults. Our field, petrographic, and textural observations exclude that the ACS have an impact origin. In fact, we did not observe any evidence of shock metamorphism, such as planar deformation features in the quartz grains of the collected samples, and the previously reported “shatter cones” are wind-erosion features in sandstones (ventifacts). Conversely, the ACS should be regarded as a “paired” intrusion of porphyritic stocks of syenitic composition that inject the Nubia Formation and form a rather simple and eroded ring dike complex. Stock emplacement was followed by hydrothermal activity that involved the deposition of massive magnetite–hematite horizons (typical of iron oxide copper-gold deposits). Their origin was nearly coeval with the development of silicified dikes in the surroundings. Plugs of tephritic-phonolitic rocks and lamprophyres (monchiquites) inject the Nubian sandstone along conjugate fracture zones, trending NNW–SSE and NE–SW, that crosscut the structural axis of the basin.

Introduction

The twin Arkenu circular structures (ACS) are located in southeastern Libya, on the eastern margin of the al-Kufrah Basin (Bellini et al. 1991) about 90 km southwest of Jebel al-Arkenu (Fig. 1). The NE (Arkenu 1) and the SW (Arkenu 2) structures have approximate diameters of about 10 km and consist of two shallow depressions with inner low hills surrounded by curved ridges and silica-impregnated cone sheets cut by local faults.

Figure 1.

 Location of the Arkenu circular structures and schematic sketch of the geology of the surrounding region (modified after Schlüter 2006).

Paillou et al. (2003) suggested that these structures could be the result of meteorite impacts. They claimed that this hypothesis was supported by the occurrence of impact breccias at the craters’ bottom, the existence of shatter cones (pointing toward the center of the craters), and the presence of microscopic planar deformation features (PDFs) found in some quartz grains of outcropping sandstones. In contrast, Di Martino et al. (2008) proposed that the origin of these “pseudo-craters” could be the result of intrusion of paired, nearly cylindrical, subvolcanic stocks accompanied by dike injections and hydrofracturing, followed by hydrothermal alteration. For the latter authors the vicinity and the similar geological setting of subvolcanic complexes of Jebel al-Arkenu and Jebel al-Uweinat strengthened this hypothesis. The neighboring regions of this part of the Sahara (e.g., in Egypt, Sudan, Chad, and even in Libya) show the presence of several circular features with possible distinct origins.

The region surrounding the Jebel al-Arkenu, Jebel al-Uweinat, and the Gilf el Kebir plateau in far southwestern Egypt, as well as the top of the plateau itself, are covered by an impressive number of crater-like or circular forms. Some of these are associated with Tertiary basalts (Klitzsch et al. 1987), others have a sandstone or breccia rim and are filled with volcanic rocks exhibiting a typical caldera-like structure (e.g., the Clayton craters, located about 50 km northeast from Jebel al-Uweinat; Clayton 1933; Peel 1939; El-Baz 1982), whereas others consist only of volcanic rocks (e.g., Jebel Peter and Paul; Dardir 1982). Some of these crater-like features, not associated with volcanic rocks, were interpreted as cryptoexplosion structures (El-Baz and Issawi 1982). The volcanic stocks of the area generally consisting of trachyte and olivine basalts, locally associated with phonolites, rhyolites, and microsyenites, are believed to be related to the Hercynian orogeny (El-Baz and Issawi 1982).

El-Baz crater (Egypt), 320 km east of Kufrah, is another typical circular structure delimited by basaltic dikes intruded into a quartz-arenitic sequence (El-Baz 1982; Barakat 1994).

Also in northern Sudan a group of circular features of unknown origin, located about 130 km east/south-east of Jebel al-Uweinat, have been described (El-Baz and Issawi 1982).

In this wider region several other crater-like structures have been identified

In a recent article, Schmieder et al. (2009) interpreted from remote sensing data the Jebel Hadid structure, in the southern Kufrah Basin, as an eroded complex impact structure, somehow similar to the Tin Bider impact structure (Algeria).

In this article, we describe the geology of the Arkenu structures and report the petrographic and petrochemical features of the igneous rocks that crop out in this area and its surroundings, thus indicating that the Arkenu structures are the result of complex endogenous processes, so that the idea that they were of impact origin should be abandoned. Some preliminary data on this issue were reported by Di Martino et al. (2008). This work will further contribute to refine the Earth Impact Database (e.g., Reimold 2007).

Geological Setting

General

The geology of Libya has been summarized in the Geological Atlas of Africa published by Schlüter (2006). Here, we report the events that describe the evolution of this region (e.g., Guiraud et al. [2005] and references therein). The Pan-African orogeny (730–550 Ma) in Saharan Africa involved the amalgamation of the West African Craton (stable since 2 Ga) with mobile belts (island arc and cordilleran-type units) and associated terranes. Mobile belts, preserved in sectors of the Western African craton and along the Red Sea, were associated with magmatic events coupled with tectonic reactivation and basin development (during the Phanerozoic). Infracambrian extensional structures opened up in sectors of the northern Gondwana region, including the Murzuq and Kufrah basins (in South-Central Libya, Niger, and Chad). These units were juxtaposed with the Precambrian basement rocks (consisting of low to medium and locally high-grade metamorphic rocks) during the subsequent regional closures.

According to Black and Liégeois (1993) the 5000 km wide Pan-African domain of Saharan Africa suffered regional continental lithospheric mantle delamination during the early stages of the orogeny, and southern Libya was affected by moderate intraplate tectonic activity during the Paleozoic. This was characterized by phases of transtensional and transpressional stresses, leading to the formation of some structures that later became oil field traps. Recycling into the asthenosphere of large amounts of lithospheric mantle likely provided the magma source for the origin of igneous suites (from Early Permian to Jurassic) with an isotope and trace element signature typical of ocean island basalts (e.g., Black and Liégeois 1993). However, the formation of rift systems across North Africa from the Triassic to the Early Cretaceous was accompanied by the deposition of thick continental strata in both the Murzuq and Kufrah basins of southern Libya (e.g., Nubian sandstones). This phase is coeval with the generation of oil reservoirs in the region. Alpine continental collision of Africa and Europe uplifted the southern Mediterranean shore and partly affected the northern margins and parts of the basins’ interiors. Magmatic activity occurred during the Tertiary until the Pliocene with the emplacement of calc-alkaline and alkaline lavas, dikes, and stocks (e.g., Darfur, Hoggar, and Tibesti volcanic provinces where lavas are generally undersaturated; Franz et al. 1999).

Geology of South-Eastern Libya

The investigated area is located on the eastern edge of the al-Kufrah Basin forming an NE–SW elongated depression between the Tibesti Massif in the West, Jebel Uweinat in the East, the Ennedi and Barkou in the South, and Jebel Dalma in the North, where the metamorphic basement rocks crop out locally, together with Paleozoic sediments (IRC, 1985; Lüning et al. 1999). The central part of the basin is filled with clastic units of Paleozoic and Mesozoic age, which show several unconformities. The thickness of the entire sedimentary sequence exceeds 4000 m at the center of the basin.

The sedimentary succession in the surveyed area is quite different from the typical succession of the al-Kufrah Basin, being limited to some Paleozoic units. We can summarize the geology of the region along a transect from Jebel al-Arkenu to the study area where the Arkenu structures are located from the base to the top.

  • 1 The complex Precambrian basement (Archean to early Proterozoic) mainly consists of strongly folded metamorphic rocks including amphibolites, serpentinites, and metagabbros alternating with quartzite, quartz-feldspar gneisses, biotite gneisses, migmatites, marbles, and banded iron formation. Subordinate diorite, granodiorite, and granite gneisses crop out locally as well.

The metamorphic assemblages of the basement rocks are comprised within the garnet-granulite facies and amphibolite facies, although some sectors may show green-schist facies assemblages (the Saharan Metacraton, according to Abdelsalam et al. [2002]).

  • 2 A Cambrian to Carboniferous clastic succession composed of fluvial to shallow marine-laminated sandstones (Hassaouna Formation) and gravels (Memouniat Formation), deltaic to epicontinental marine sediments (Acacus and Tanezzuft Formations), shallow marine shales and sandstones (Binem Formation), and continental to locally marine sandstones (Dalma Formation).
  • 3 Permian (Continental post-Tassilian) and Triassic to Cretaceous (Nubia Formation) sandstones of continental, fluvial, and shallow marine origin, with irregular thickness. Basement rocks and the sedimentary units are associated with the following igneous rocks: Paleozoic to Tertiary volcanics consisting of basalts, phonolites, and trachytes; Tertiary ring complexes of granite, syenite, and nepheline-syenite that inject older rocks; Pliocene and more recent (2.9–1.8 Ma) calc-alkaline to alkaline plutons, dikes, and sills intruding all the above units.
  • 4 Quaternary deposits include sand dunes, sand sheets, and gravel. Debris plains are widespread.

The overall structural evolution of Libya has recently been discussed by Guiraud et al. (2005; cf. their fig. 7), who described the tectonic style of the al-Kufrah Basin and its surroundings. The basin is an open synclinal fold locally crosscut by NE–SW trending graben-like normal faults. However, the local tectonics are more complex and represented by nearly N–S trending dextral strike–slip faults and two sets of conjugate faults: NW–SE sinistral and NE–SW dextral. Finally, an approximately E–W system of faults is also known. These systems have been subject to crustal adjustment and stress changes that periodically reactivated the faulting network.

In the northern al-Kufrah Basin, Ghanush and Abubaker (2007) reported low gravity anomalies ranging from −70 to −75 mGal probably due to the thickness of the sedimentary cover and the presence of crustal rocks. They also report large positive magnetic anomalies approaching 60 nT, probably due to the presence of high Fe-mineral contents in clastic rocks. Basement rocks of the Archaean metamorphic units, found in the surrounding areas, may reach 160 nT.

Geology of the ACS Site

Remote Sensing

In Fig. 2, we show an ASTER image of the Arkenu circular structures and surrounding area. The image of Fig. 3a is the original digital elevation model, whereas the one in Fig. 3c is an interpretative sketch and reports the major morpho-structural lineaments.

Figure 2.

 Ortho-rectified false-color ASTER image (bands 3-2-1), acquired on 21 January 2005 over the Arkenu structures. The ASTER data were provided by LP DAAC (Land Processes Distributed Active Archive Center; https://lpdaac.usgs.gov/).

Figure 3.

 Shaded relief map (a) and major morpho-structural features (b) retrieved from the SRTM-DEM of the investigated area. The shaded image is combined with a color scale to show variations in elevation (the abbreviation m a.s.l., indicates meters above sea level). Synthetic illumination (from the south) enhances the main morphological structure and lineaments which are sketched on the right. Thick and thin black lines represent major and minor morpho-structural lineaments, respectively, while the red circles show the location of the concentric rings associated with the Arkenu structures (see the text for explanation).

The topography of the investigated area has been reconstructed from Shuttle Radar Topography Mission (SRTM) data (Farr et al. 2007). In particular we used digital elevation data provided by the Consortium of Spatial Information (http://www.cgiar-csi.org) which released a new version of the original SRTM DEMs, with filled data voids (Jarvis et al. 2008). Reprocessed digital elevation models are available in mosaic 5 deg × 5 deg tiles with a spatial resolution of 90 m at the equator. It should be noted that to gather topographic (elevation) data of Earth’s surface, SRTM used the technique of interferometry based on reflected “radar” signals (on bands C at wavelengths of 5.6; Farr et al. 2007). As a result, one of the most interesting features of the SRTM data collected on desert regions is the capability of the radar signals to penetrate the dry sand surface (to a depth of about 0.5 m on band C; Schaber et al. 1997), so that the resulting DEM is able to image several sand-covered morphological structures such as subsurface drainage systems (Ghoneim et al. 2007). Accordingly, the elevation model retrieved using SRTM DEMs does not necessarily reflect surface topography of an investigated area but more likely it represents the subsurface morphology produced by any kind of surface or structure that is rough enough to generate backscattered radar signals (Elachi et al. 1984).

A shaded relief map of an area of approximately 50 × 50 km around the Arkenu circular structures is illustrated in Fig. 4. This map has been generated using a synthetic illumination (from the south) over the virtual 3-D landscape obtained by the SRTM DEM, and combining the shaded image with a color scale to show the elevation. From this image it appears that despite the apparent gentle morphology of the area, at a larger spatial scale the subsurface topography provide a complex morpho-structural pattern with several major lineaments essentially represented by scarps, straight valley, and dunes (Fig. 3b).

Figure 4.

 a) Shaded relief map of the Arkenu structures with illumination from the east. b) A detailed view of the topography approximately 10 km NE of Arkenu 1 where a third circular concentric structure (about 5 km in diameter) has been observed. Actually this structure is completely covered by sand.

As discussed in more detail in the following section, the most striking features are the circular structures that coaxially surround the two alleged “craters.” There are two types: concentric ridges caused by variations in topography, and silica-impregnated sedimentary dikes found in the Nubian sandstones (that are more resistant to erosion and result in minor topographic discontinuities). Local structures are represented by two major “conjugate” systems of faults: a NNW–SSE and a NE–SW system. Another morpho-structural lineament is trending E–W. This structural arrangement is in agreement with the previously described regional setting. It must be emphasized that a NNW–SSE fault (parallel to the major one cited above) crosscuts the eastern sector of the Arkenu 1 SE rim, and shows a dextral strike–slip motion (with an offset of about 60 m).

Intriguing questions arise if we analyze the SRTM-DEM that allows a better detection of subsurface structures (i.e., up to about 50 cm below the sand cover). In Figs. 4a and 4b, we report an additional image of the paired Arkenu structures (with a synthetic illumination from the east). Both Arkenu 1 and Arkenu 2 are visible and the latter is clearly crosscut, along its northern external ridges, by the NE–SW fault. However, if we move along this structure, just at the NNE of Arkenu 1 we can notice another circular structure that is very likely the surface expression of a subvolcanic stock related to the paired Arkenu structures. Further field and geophysical investigations are needed to test this possibility.

Field Observations

Here, we present the data from our field observations essentially related to the geological features that outcrop in the ACS area. Additional field and petrographic evidences for the nonimpact origin of the Arkenu structures is given in a later section.

As mentioned above, the sedimentary rocks in the ACS area are essentially Nubian sandstones (designating a thick series of quartzose sandstones overlying the Precambrian and Paleozoic rocks). Nubian sandstones crop out at the floors and rims of both structures (Fig. 5). In the study area, the Nubia Formation includes several stratigraphic units represented by shallow marine near shore to coastal successions, alluvial to flood-plain sediments with interbedded channel deposits, and paleosoil horizons (such as those described by Klitzsch et al. 1987; Said 1990).

Figure 5.

 a) Vacuolar coarse-grained sandstone in the central part of Arkenu 2. b) Below vacuolar sandstones, hard quartz-arenites are present. c) Thin-bedded siltstones underlie quartz-arenites (central and NW sector of Arkenu 2).

Main lithologies consist of brownish, reddish, and purple, coarse to fine-grained quartz sandstones (locally cross-bedded), generally slightly cemented, with interbedded varicolored siltstones and shales, of breccias and conglomerates, and rare, very hard, purple-brown and ocher ironstone beds (e.g., Bellini et al. 1991).

In the Arkenu 1 structure, Nubian sandstones are locally overlain by massive magnetite–hematite bodies, whereas in the Arkenu 2 area such bodies are sparse and limited to some sectors in the southern area. Rocks are essentially brownish porous medium to coarse-grained (Fig. 5a), occasionally fine-grained (Figs. 5b and 5c), moderately to poorly sorted sandstones with subrounded to subangular quartz and rarely sedimentary lithic fragments. Thin bedded siltsones underlie quartz-arenites (Fig. 5b).

No specific data about the thickness of the Nubia Formation could be obtained in the study area, but the regional geologic setting allows us to suppose a thickness of only a few hundred meters.

In the north-central part of the Arkenu 1 structure, we found a decameter-sized body of syenite porphyry locally surrounded by meter-thick veins of similar composition (Fig. 6a). In some sectors of the structure’s floor these rocks may show a shell of argillic alteration. At the contact with syenite intrusive veins, we found a brownish horizon of muscovite-rich fragmental feldspathic material (containing magnetite, apatite, barite, and Mn and Zn oxides) that is overlain by a massive magnetite–hematite unit with subhorizontal attitude and a thickness of several meters. This unit is mainly composed of massive magnetite (partially transformed to hematite and/or limonitic products), with abundant centimeter-sized cavities often lined with octahedral magnetite crystals that are vertically oriented. In some places, in the eastern sector the contact between the sandstones and the overlying massive magnetite–hematite horizon is marked by a few-meters thick hydrothermal breccia (Fig. 6b). This breccia in places shows silicified syenite clasts embedded in apatite-rich veinlets. Pervasive hydrothermal alteration was observed at the bottom of the Arkenu 2 structure (Fig. 6c), in the central part of the southwestern sector, where a fine-grained sandstone is strongly affected by argillic alteration and shows typically bleached white to gray color. Together with the findings at Arkenu 1, this strongly supports the idea that “paired” syenitic bodies may be present at depth.

Figure 6.

 a) Syenite apophysis (white) with fragmental brownish gangue (upper contact) overlain by massive magnetite and vacuolar sandstone (left upper corner). b) Hydrothermally altered breccia in altered syenite at the contact with magnetite bodies. c) Argillic alteration at the bottom of Arkenu 2 (western sector) crosscut by a distinct NNW–SSE fracture.

Thus, the most important feature of the Arkenu 1 structure is the occurrence of a massive magnetite–hematite horizon. This mineralization is somehow similar to the ores described by Grez et al. (1991) and Nyström and Henriquez (1994) for Magnetita Pedernales and Kiruna-type deposits, respectively, whose origin is highly debated. We will later discuss the geoeconomic implications of such deposits.

At the inner rim of Arkenu 1, the magnetite–hematite horizons are covered by the above described porous sandstones, but this sedimentary unit is missing at the top of the mesas in the interior of the structures (Fig. 7). Thus, we regard this setting as a result of an erosive process that acted differentially and likely mobilized the most altered horizons that were concentrated on top of the inner part of the circular structures.

Figure 7.

 Massive magnetite horizons in the central part of Arkenu 1 (looking south from the northern rim): the bottom of the structure is essentially covered by a sand sheet. Vacuolar sandstones (on top of these horizons at the structure rim) are missing, which is indicative of a higher degree of erosion in the area of the structure.

In the surroundings of both Arkenu 1 and Arkenu 2 we found silicified dikes that transect the Nubian sandstones (Fig. 8). These dikes are essentially vertical at the surface but seem to follow a cone sheets geometry, at depth, that extend coaxially (for approximately 5 km) from the central axis of the circular structures. Within these circular features we found the presence of hypabyssal tephrite to phonolite and lamprophyric plugs (the latters may also outcrop outside the circular cone sheet area) (Fig. 9). This supports the idea that a simple tectonic style, consistent with the reactivation of an al-Kufrah graben-like structure, controlled their emplacement. In fact, these intrusive bodies inject the Nubian sandstone along conjugate fracture zones trending NE–SW (parallel with the orientation of the circular structures) and NNW–SSE that crosscut the structural axis of the graben.

Figure 8.

 a) Perspective view of the silicified dike complex, with cone sheet geometry in Nubian sandstone approaching the Arkenu 1 circular structure (from the NE). b) Detail of a silicified dike in the southern sector of the Arkenu 2 circular structure.

Figure 9.

 a) Phonolitic plug about 2 km from the SSE rim of Arkenu 1. b) Effusive plug of lamprophyre cropping out in the outer southern rim area of Arkenu 2 (about 2 km from the center).

Petrography and Petrochemistry

Petrography of Igneous Rocks

The petrographic analysis of Nubian sandstones and breccias outcropping in the areas of the Arkenu structures did not reveal any characteristic shock deformation features. Here, we present a brief description of the igneous rocks that can be subdivided into: syenite porphyries, tephrites and phonolites, and lamprophyres (monchiquites).

There are two types of syenite porphyries. They outcrop within the northeastern sector of the Arkenu 1 structure. Syenite of the first type is more peripheral and is proximal to the intrastructural massive magnetite bodies and shows a porphyritic texture consisting of large K-feldspar (sanidine with subordinated microcline) grains, up to 5 mm across and very rare albite at the rims of feldspar laths in a fine-grained feldspathic matrix. The rock has been subject to strong potassic alteration. Secondary phases are sericite, carbonate, ankerite, and Fe-Mn oxides. Sparse zircon, apatite, and rutile are accessory phases.

The second type is a quartz-bearing syenite porphyry with crystals of sanidine-microcline, albite, abundant microphertite, rare biotite flakes, and patches of acicular greenish amphibole (likely of the richterite-arfedsonite series) locally altered to chlorite. Matrix minerals are of the same mineral phases with patches of granophyric intergrowths consisting of K-feldspar and quartz. Secondary minerals are sericite, Mg-chlorite, carbonate, ankerite, and Fe-Mn oxides (in veinlets). Zircon, apatite, rutile, allanite, opaques (Ti-magnetite and ilmenite) are accessory phases.

Tephrites-phonotephrite-tephriphonolites and phonolites are well represented in the Arkenu area. These rocks are found as necks and dikes that crosscut the Nubian sandstones in the areas surrounding both Arkenu structures.

Tephrites and related lavas (phonotephrite-tephriphonolites) show a porphyritic texture with laths (up to 2–3 mm across) of euhedral plagioclase (albite-oligoclase), subhedral amphibole (of the richterite-arfedsonite series), microphenocrysts of nepheline and sanidine in a microgranular texture consisting of these same mineral phases. Additional matrix minerals are calcite and ankerite, apatite, Fe-Mn oxides. In more altered samples the amphibole is altered to chlorite; secondary K-feldspar replaces sodalite; secondary carbonate and ankerite are common and coexist with zeolites. Accessory minerals are zircon and apatite.

Only a single neck of phonolite has been observed in the area SE of Arkenu 1. It shows a porphyritic texture with sanidine macrophenocrysts (up to 5 mm across), phenocrysts of aegirine, and laths of subhedral phlogopite in a microfelsitic matrix of the same phases, nepheline and secondary zeolites, carbonates and opaques. Accessory phases are sparse zircon and apatite.

Lamprophyres are also well represented as plugs and subordinate dikes. They show a holocrystalline porphyritic texture of euhedral olivine (with 2 mm on average size) locally altered to iddingsite, euhedral to subhedral Ti-augite, kaersutitic amphibole, microphenocryts of nepheline, and late poikiolitic phlogopite (which locally overgrows iddingsitic alteration haloes around olivine), in a pilotaxitic to cryptofelsitic matrix consisting of these mineral phases plus abundant rutile, sparse apatite, and microgranular zircon. Secondary phases are zeolites in patches and analcime. According to the IUGS nomenclature (Le Maitre et al. 1989), these rocks belong to the monchiquite family.

Petrochemistry of Igneous Rocks

Previous major and trace elements data for Libyan igneous rocks were reported by Conticelli et al. (1995), who described alkali basalts, basanites, and mela-nephelinites in the Jebel Uweinat area and in the al-Kufrah basin (26.7–28 Ma in age). Bardintzeff et al. (2011) reported data on the recent foidites, alkali basalts, and basanites at Waw an Namous volcano and in the surrounding Al Haruj area that has been also studied by Cvetković et al. (2010). We will describe the petrochemistry of the igneous rocks that outcrop in the ACS area and relate their geochemical signature to similar rock suites, well known in the literature.

Bulk-rock compositions were acquired by inductively coupled plasma mass spectrometry (ICP-MS) analysis by ALS Chemex, Sevilla, Spain. Precision of analyzed major elements are better than ±0.2% for SiO2, and ±0.1% for the other oxides (with the exception of MnO and P2O5, which show errors of about ±0.02%). Analytical precision for trace elements is 2–3% for concentrations up to 100 ppm, 3–6% in the range of 10–100 ppm, and 6–10% below 10 ppm. The analyses for the different magma suites are reported in Table 1.

Table 1. Whole rocks chemical analysis on the materials found at Arkenu 1 and Arkenu 2.
 Syenites Tephritic–Phonolitic SuiteLamprophyresVein
Rock typeSyeniteSyeniteQz-SyeniteQz-SyenitePhonoliteTephri - phonoliteTephri - phonoliteTephri - phonoliteTephriteTephriteMonchiquiteMonchiquiteMonchiquiteMonchiquiteMonchiquiteMonchiquiteApatite vein
Sample NameG2G44233E1E29A16A202228R101414B26R27R25C
  1. *Total iron as Fe2O3; n.d. = not determined; L.O.I = loss on ignition.

SiO264.6065.2068.1067.3053.4054.9051.2051.3047.0048.2036.2036.0038.8038.7039.0037.500.62
TiO20.030.060.320.380.360.480.620.381.000.415.895.954.704.714.854.810.03
Al2O317.5517.4214.2015.3018.4016.5017.5519.5017.5017.959.469.529.299.5410.557.900.24
Fe2O3*0.390.362.652.713.995.444.934.217.134.9515.3015.3513.9513.5013.6514.952.77
MnO0.110.120.160.190.330.660.250.330.290.460.250.240.210.210.190.231.72
MgO0.060.080.430.520.360.620.590.961.130.498.8010.2512.3511.5510.5012.400.12
CaO0.420.501.001.181.692.715.071.988.886.9015.1514.7513.6013.4512.4513.1052.40
Na2O0.780.836.746.838.042.648.068.964.376.363.152.902.762.823.021.940.01
K2O13.6012.802.142.255.698.925.333.753.263.871.741.691.361.420.951.470.01
P2O5*0.080.110.040.090.030.030.350.040.280.031.201.160.780.770.711.0035.40
L.O.I.0.901.203.142.426.386.186.108.337.168.542.392.292.202.693.002.684.08
Total98.5298.6898.9299.1798.6799.08100.0599.7498.0098.1699.53100.10100.0099.3698.8797.9897.40
V985047506945465157412394312350357314159
Cr2220176478977025045050038047010
Co121.51.30.83.62.80.83.31.147.854.25354.557.962.61
Ni2243547532469619218514726210
Zn646168662052631562191842921651651201291301501310
Ga23.22225.324.144.637.735.641.53122.622.522.816.918.720.217.512.4
Rb25525166.565.2160240138.5109123.5133.532.339.929.13525.737.16.6
Sr308302245242382169134051817052410118022501145114564311951670
Y5.95.49.79.435.445.326.636.437.456.236.133.922.525.322.328.2195.5
Zr3233543529284030701685291022203770428327331360318375119
Nb882875108.5104.5399487307388242585108102.585.194.686.2100.5331
Mo8n.d.<2n.d.<2<24<2<2<2<2<27<2<2<266
Sn0.50.3114534453322232
Cs0.310.280.350.3311.892.360.641.352.310.280.280.30.30.770.320.2
Ba4010400270774225002340888189011251780911831623810118524103890
Hf1110.810.348.150.326.749.744.952.41088.59.28.78.72.1
Ta30.529.81122.524.627.421.725.438.18.58.16.26.867.20.6
Pb233229211918231920152635443624
Th5.985.923.373.5236.340.42633.822.440.69.359.76.787.546.118.0812.15
U14.514.10.610.666.764.6214.186.116.32.612.581.742.012.252.19319
Ce24424053.450.6150.5164.5203155199357191195125140.5122.5156.52800
Dy1.581.521.641.587.859.346.168.028.8912.89.278.985.936.695.897.3164.4
Er0.780.70.930.94.535.33.214.444.346.653.993.782.572.832.563.1524
Eu2.041.981.121.063.193.723.263.264.36.34.94.983.193.53.13.9849.6
Gd7.757.333.373.1810.1511.9511.110.513.620.415.215.659.8711.359.7812.5157.5
Ho0.250.230.320.341.551.791.141.521.572.341.551.521.011.130.991.249.79
La13012633.930.373.990.510381.392.818391.393.559.46758.275.41260
Lu0.050.040.130.10.670.760.420.640.510.860.360.320.240.260.240.291.21
Nd92.990.831.130.36272.583.167.190.7144.595.397.262.46959.977.51305
Pr26.926.38.087.9717.6520.623.518.7524.341.124.124.715.817.615.3519.8349
Sm10.9510.84.914.7810.8512.312.5511.2515.22317.4517.5511.512.710.9514.1194
Tb0.560.530.360.311.431.661.321.461.762.581.961.941.251.391.231.5516.55
Tm0.060.050.130.110.640.750.440.610.560.890.480.430.30.340.310.362.03
Yb0.350.330.740.794.315.082.794.13.355.82.72.421.71.871.672.1110.45
La13012633.930.373.990.510381.392.818391.393.559.46758.275.41260
Ce24424053.450.6150.5164.5203155199357191195125140.5122.5156.52800
Pr26.926.38.087.9717.6520.623.518.7524.341.124.124.715.817.615.3519.8349
Nd92.990.831.130.36272.583.167.190.7144.595.397.262.46959.977.51305
Sm10.9510.84.914.7810.8512.312.5511.2515.22317.4517.5511.512.710.9514.1194
Eu2.041.981.121.063.193.723.263.264.36.34.94.983.193.53.13.9849.6
Gd7.757.333.373.1810.1511.9511.110.513.620.415.215.659.8711.359.7812.5157.5
Tb0.560.530.360.311.431.661.321.461.762.581.961.941.251.391.231.5516.55
Dy1.581.521.641.587.859.346.168.028.8912.89.278.985.936.695.897.3164.4
Ho0.250.230.320.341.551.791.141.521.572.341.551.521.011.130.991.249.79
Er0.780.70.930.94.535.33.214.444.346.653.993.782.572.832.563.1524
Tm0.060.050.130.110.640.750.440.610.560.890.480.430.30.340.310.362.03
Yb0.350.330.740.794.315.082.794.13.355.82.72.421.71.871.672.1110.45
Lu0.050.040.130.10.670.760.420.640.510.860.360.320.240.260.240.291.21

Major Elements

The peripheral syenite porphyry is an alkali-syenite with a SiO2 content of 64.6 wt%, very high K2O (13.6 wt%), and low Na2O contents (0.78 wt%). When plotted on the total alkali versus silica diagram (TAS modified for igneous rocks, cf. Bellieni et al. 1995; Fig. 10a) is an alkaline-rich syenite. The second syenite type is a quartz-bearing syenite comparatively richer in SiO2 (up to 68.1 wt%) and that has a higher content in Na2O (6.74 wt%) due to the abundance of albite and microperthite. These rocks also have moderately higher contents in Fe2O3 (total), CaO, and MgO due to the presence of amphibole and secondary carbonate. In Fig. 10a syenitic rocks are compared with other syenitic suites well known in literatures.

Figure 10.

 Total alkali-silica diagram for the igneous rocks outcropping at ACS and in the surrounding area. a) The alkali syenites have been plotted according to the TAS diagram for the classification of plutonic rocks (e.g., Bellieni et al. 1995). For comparison we included the rocks of the Cameroon Line (Deruelle et al. 1991; Njonfang and Nono 2003), Namibia (Harris 1995; Harris et al. 1999), Monchique Complex and Mount Ormonde (Bernard-Griffiths et al. 1997; Grange et al. 2010). b) Tephritic-phonolitic suites and monchiquite lamprophyres are compared with the Monchique Complex (Cretaceous, Bernard-Griffiths et al. 1997; Grange et al. 2010), the alkali field of Libya-Western Sudan lavas (Franz et al. 1987, 1999; Lustrino and Wilson 2007; Cvetković et al. 2010; Bardintzeff et al. 2011) and the Okenyenya Alkali Igneous Complex of Namibia (Cretaceous in age, Milner and Le Roex 1996; Le Roex and Lanyon 1997), and the monchiquite-camptonite suite of the middle-high Atlas, Morocco (Tertiary, Bouabdli et al. 1988; Wagner et al. 2003);

When plotted on the TAS diagram of Le Bas et al. (1986) subeffusive rocks plot on the field of tephrite, phonotephrite, tephriphonolite, and phonolite (Fig. 10b) with silica ranging from 47 to 54.9 wt% and total alkali element concentrations from 7.90 and 13.8 wt%, respectively. Fe2O3 (tot), CaO, and MgO are higher in tephrites (up to 7.13, 8.9, and 1.13 wt%, respectively). The very high LOI values of some of these rocks are due to marked secondary alteration recorded by the presence of abundant hydrous minerals, secondary calcite, ankerite, and patches of zeolites.

Lamprophyric rocks of the monchiquite family have lower SiO2 contents (from 36.2 to 39 wt%) and considerably higher CaO, Fe2O3 (tot), MgO values (up to 15.2, 15.3, and 12.4 wt%, respectively). However, total alkali element values are lower and range from about 2.7 to 5.9 wt%, so that in the TAS diagram they plot in the low silica field. The very high values for TiO2 (4.7–5.95 wt%) correspond to the high modal content of rutile.

Trace Elements

Trace element data for the syenite porphyries are high in incompatible elements (particularly Ba up to 4010 ppm and Rb 255 ppm) and high field strength elements (HFSE) typical of alkaline subeffusive rocks. The Rb/Sr ratios range from 0.828 to 0.831. Transition metal contents are low, whereas Pb contents are high (up 233 ppm). The overall trace element signature of these rocks is rather peculiar, likely due to the overprint from pervasive potassic alteration. In Fig. 11, we report REE patterns for the Arkenu 1 syenite porphyries (with [La/Lu]N of 2.67–3.24) compared with an apatite vein ([La/Lu]N = 107.3) associated with the overlaying breccias. The degree of enrichment of the veins is quite high but the overall shape of the REE patterns is rather consistent, thus indicating a genetic link between these rock types. In particular, there is a slight negative anomaly in Eu followed by a depletion in HREE.

Figure 11.

 Chondrite-normalized REE patterns (normalization factors from Nakamura 1974) for the syenite porphyry compared with an apatite vein.

The quartz-bearing syenite porphyries have lower contents in incompatible elements (with Ba ranging from 707 to 742 ppm, and Rb from 62 to 67 ppm) than the previously described syenites, and exhibit Rb/Sr ratios around 0.27. These rocks have comparatively higher contents in Zr (up to 534 ppm) consistent with the abundance of zircon as accessory phase. Transition metals, with the exception of Cr (17–20 ppm), are similar to those recorded in the syenites.

These rocks show a lower degree of enrichment in terms of LREE (about 100 times chondritic values, with [La/Lu]N of 26–31) compared with the syenites but have higher HREE contents with a rather constant pattern between Dy and Lu (Fig. 12a). However, these porphyries have a trend that is more similar to those defined by tephritic and phonolitic rocks. This similarity is further emphasized by looking at the spider diagram (Fig. 12b) normalized to primitive mantle values (according to Sun and McDonough 1989). Negative anomalies in U (particularly in the quartz-syenites) and P are likely due to apatite fractionation, whereas the Ti anomaly, typical of orogenic suites, may be also related to magnetite segregation. On the other hand, positive anomalies in Pb and Zr could be inherited from the source region and/or related to the abundance of zircon as accessory phase.

Figure 12.

 a) Chondrite-normalized REE patterns (normalization factors from Nakamura 1974) for the quartz-bearing syenites compared with the rocks of the tephrite-phonolite suite. b) Spider diagram for the trace element abundances of these rocks (normalized to primitive mantle values after Sun and McDonough 1989).

Trace element data for the tephrites-phonotephrite-tephriphonolites and phonolites are plotted in Fig. 12. They show a higher degree of enrichment in REE when compared with the quartz-bearing syenites, but their bulk REE patterns exhibit a similar trend thus suggesting a possible genetic link. They have [La/Lu]N ratios of 14–25. This similarity is further supported by the spider diagram (Fig. 12b) where the main difference is the absence of the U anomaly when compared with that of the quartz-bearing syenites. We observe positive anomalies in Pb and Zr, as well.

Lamprophyres of the monchiquite family are also enriched in incompatible elements (particularly Ba, ranging from 623 to 2410 ppm, the latter values in most altered rocks; Rb ranges in abundance from 25.7 to 39.9 ppm) with Rb/Sr ratios from 0.02 to 0.04. Transition metals are quite abundant with Ni and Cr contents up to 262 and 500 ppm, respectively. HFSE elements contents are slightly higher than those of OIB.

The lamprophyres show a rather coherent trend in terms of REE (Fig. 13) with a uniform pattern (with [La/Lu]N of 25–30), but are slightly less enriched, in terms of LREE when compared with the alkaline lavas. Moreover, they are only slightly more depleted in terms of HREE. In this case the negative Eu anomaly is small or almost negligible. When these samples are compared with OIB, we can notice a generally similar pattern, with moderate enrichments up to about 10 times the mean values of the oceanic island basalts of Sun and McDonough (1989). In particular, they show anomalous enrichments in Ba, Sr, Nd, and Ti (the latter due to the abundance of rutile) and moderate negative anomalies in K.

Figure 13.

 a) Chondrite-normalized REE patterns (normalization factors from Nakamura 1974) for the ACS monchiquites compared with those of b) the Monchique Complex (Bernard-Griffiths et al. 1997; Grange et al. 2010), middle-high Atlas Morocco (Bouabdli et al. 1988; Wagner et al. 2003) and Namibia (Le Roex and Lanyon 1997). c, d) Spider diagram for the above rocks. Trace element abundances were normalized to primitive mantle values after Sun and McDonough (1989); the OIB pattern is included for comparison (average data from Sun and McDonough 1989).

Evidence for the Nonimpact Origin of the ACS

Some rocks outcropping in the surroundings of the alleged crater-like structures are characterized by striations that have been previously interpreted as shatter cones (Paillou et al. 2003).

Shatter cones are conical and have striated surfaces, and they are widely held to be diagnostic of impact structures. They are characteristically found in impact crater floor rocks or in central uplifts (if present). They may also be observed as clasts in impact breccias (e.g., Milton 1977; French 1998; Lugli et al. 2005). Our observations show that the striations are not pervasive but superficial features and are not related to fracturing. In addition, the observed striations are also present on the surfaces of rocks at distal sites from the Arkenu area (Fig. 14). These striations, which may be found throughout the eastern Sahara region, are oriented NS (±20°), thus being in good agreement with the main wind directions (from north-west since early Holocene, and then from north and north-east until present times; cf. Brookes 2003). These structures are definitely not shatter cones but are, in turn, related to the abrasive action of winds on the exposed rock surfaces, representing so-called ventifacts (e.g., Orti et al. 2008).

Figure 14.

 a, b) Ventifacts related to the abrasive action of winds on exposed rock surfaces in the NE sector of Arkenu 1.

In addition, the rocks sampled in the ACS do not show any microscopic effect of shock metamorphism: no planar deformation features (PDFs) in quartz grains, no evidence of melting, or presence of glass. It is well known that, in porous sedimentary targets, shock effects are not always clear and well developed (e.g., Grieve et al. 1996). However, observations of shock effects in Coconino Sandstone from Meteor crater (Arizona) showed a progressive microscopic change from their original textures with increasing shock pressure: at low pressure (<5 GPa) the porosity is reduced to zero and the minerals are fractured; at higher pressures (5–13 GPa) the fractured quartz may coexist with glass and coesite (cf. French [1998]; and references therein). In the Arkenu area, there are no apparent differences between the sedimentary rocks outcropping inside and outside the circular structure. In summary, we did not find evidence for the shocked quartz alleged by Paillou et al. (2003) in sandstones and breccias. Rare deformation textures found in a few of the numerous studied quartz grains do not have any similarities with PDF typical of impactites and may be related to tectonic deformation in the provenance regions and recycling of ancient sedimentary rocks. Moreover, we must recall that even the presence of isolated shocked quartz grains in clastic sedimentary rocks (without additional impact evidence) could represent erosion of distal impact structures and, thus, may not be uncommon in the Sahara region (e.g., Orti et al. 2008).

Discussion

The geological survey conducted at Arkenu circular structures gave us the opportunity to verify in the field the reliability of the suggested hypothesis related to an impact origin. The presence of a fragment of a carapace of a syenitic stock, and related veins, in the Arkenu 1 structure associated with extensive hydrothermal alteration is a key factor for establishing its genesis and evolution. It is quite surprising that the outcrop of this intrusive had never been reported before, as one can easily walk onto it in the northeastern sector of the structure’s bottom. Moreover, the presence of massive magnetite horizons associated with hydrothermal veins and breccias (bearing apatite veinlets) somehow similar to the mineral ore of “Magnetita Pedernales” (northern Chile; cf. Grez et al. 1991) suggests that this structure is related to complex endogenous processes. The origin of such deposits has been a strongly debated issue since their discovery: initially interpreted as the most typical examples of iron oxide lava flows (Park 1961), they have been later reinterpreted as the result of metasomatic-hydrothermal processes (Rhodes et al. 1999; Sillitoe and Burrows 2002). Particularly Sillitoe and Burrows (2002) developed a plausible explanation for the hydrothermal origin for this type of massive magnetite ore, based on observations at the El Laco magnetite deposit in northern Chile. They suggested that these deposits could be the result of metasomatic replacement induced by uprising hydrothermal fluids at shallow depths (within 100–300 m from the paleosurface), i.e., within the framework of fossil geothermal systems. According to their model, ore deposition occurred below a steam-heated zone separated from the paleowater table by a high-temperature alteration halo. The fluid source is, however, still under debate, as clear relationships with underlying magmatic bodies are generally missing and fluid inclusion studies are consistent with both a magmatic (Broman et al. 1999) and a metasomatic-replacement (Sheets et al. 1997) model.

Several coexisting features characterize these massive magnetite deposits, considered as part of the iron oxide copper-gold (IOCG) deposits by some authors (Sillitoe 2003; Williams et al. 2005) or seen in a different group by others (P-rich iron oxide deposits: Groves et al. 2010). They include (1) the abundance of hydrothermal apatite; (2) the occurrence of certain features (vuggy textures; well shaped magnetite crystals) typical of growth at shallow depth; (3) evidence of a composite hydrothermal evolution: magnetite precipitation postdates a high-T (high-Temperature) event and is followed by argillic alteration at lower temperature regimes.

All these features are found in our specific case, where the hydrothermal origin of magnetite ore is strongly supported by its association with apatite-rich hydrothermal veins and breccias. A close relationship between hydrothermal fluids and the syenite body is suggested by the field data and confirmed by the consistent REE patterns of syenite and apatite veins. As described above, syenite is at least locally affected by an early potassic alteration overprint, which is a relatively high-T hydrothermal event. The presence of veins and sectors of argillic alteration at the structure’s bottom are likely related to late hydrothermal stages when, probably due to uplift and erosion, the downward migration of the paleowater table occurred (as is common in epithermal environments), and the steam-heated products overprinted the magnetite horizons. Coeval and subsequent supergene alterations partially transformed the magnetite into hematite and likely produced the leaching of sulfide minerals at the surface. This phenomenon was likely associated with the marked erosion of the softer upper part that was originally steam-heated. Thus, the ACS are a result of complex endogenous processes that involve stock emplacement, hydrofracturing, hydrothermal alteration cycles, and concentrated erosive processes.

The geological setting of the area suggests the nearly coeval emplacement of the paired porphyry stocks into the Nubian sequence. Within Arkenu 1, the porphyry outcrops on the NE “crater” bottom and inner flank (likely due to higher erosive regimes related to the topography), whereas within Arkenu 2 the granite porphyry is likely underlying, at moderate depth, some horizons of argillic alteration that impregnate the quartz-arenite at the structure’s bottom (central southwestern side). Silicified dikes with cone sheets geometry extend coaxially (for approximately 5 km) from the central axis of the circular structures. Within these circular features we found tephrite-phonolite and hypabyssal lamprophyric plugs but these bodies may also be found outside the cone sheets area. This supports the idea that a simple tectonic style, consistent with the reactivation of the al-Kufrah graben-like structure, controlled their emplacement. In fact, these intrusive bodies inject the Nubian sandstone along conjugate fracture zones trending NE–SW (parallel to the orientation of the structures) and NNW–SSE and that moderately crosscut the structural axis of the graben.

On petrogenetic grounds syenites and tephrite-phonolites represent typical alkaline suites of likely Eocene to Oligocene age. We suggest this age as the igneous rocks intrude the Nubian Formation and similar alkaline rock suites outcrop in the nearby Jebel al-Arkenu area (IRC 1985; Lüning et al. 1999) and recent Jebel al-Uweinat (Conticelli et al. 1995). Moreover, similar alkaline intrusives, Upper Cretaceous to Lower Oligocene in age, were reported by Issawi (1982) and Klitzsch et al. (1987) in southwestern Egypt. However, a younger age cannot be excluded (Franz et al. [1999] described similar findings in northwestern Sudan and considered them to be of Pliocene age) although the degree of alteration of these rocks seems to favor our first hypothesis. The lamprophyric rocks (monchiquites) could be somewhat older as their OIB signature suggests recycling into the asthenosphere of large portions of the lithospheric mantle that according to Black and Liégeois (1993) is typical of igneous suites of at least Jurassic age.

Conclusions

In the light of our findings the hypothesis of an impact origin for the Arkenu circular structures should be abandoned. We here provide extensive proof of their endogenous origin related to the emplacement of “paired” subvolcanic porphyries of syenitic composition. These stocks were emplaced along conjugate faults that were then reactivated. We cannot say whether their injection was accompanied by venting at the surface, but at least hydrofracturing has been effective during and/or after their emplacement. Their surroundings were sites of subsequent hydrothermal activity that involved the deposition of massive magnetite horizons. These processes were nearly coeval with the development of silicified dikes in nearby wallrocks that follow a cone sheet geometry. The overall structure of the ACS closely resembles a rather simple ring dike complex consisting of at least two “paired” stocks associated with a single magmatic injection. Plugs of tephritic-phonilitic rocks are nearly coeval and show petrochemical affinities with the unaltered syenite porphyry in terms of their trace element compositions. In turn, lamprophyric dikes of the monchiquite family of definite OIB affinity could be somewhat older.

Hydrothermal alteration at the ACS site has been responsible for the genesis of massive magnetite-hematite horizons similar to “Magnetita Pedernales,” El Laco and Kiruna P-rich iron ore deposits. Metal enrichment was accompanied and followed by fault reactivation, uplift, and erosion that affected the final morphology of this sector of southern Libya. In conclusion, it is not excluded that these circular structures, together with several others in the al-Kufrah basin and its wider surroundings (e.g., Schmieder et al. 2009) could be potential targets for the exploration of iron oxide copper-gold (IOCG) deposits (cf. Williams et al. 2005).

Acknowledgments— This research has been funded by the Italian Ministry for University and Research (MIUR) and by INAF. Additional funds were provided by Fondazione Cassa di Risparmio di Torino. We thank E. Callegari and L. Orti for discussing several aspects of Saharan geology. R. Serra provided logistic support. G.P. Sighinolfi and L. Ferrière reviewed an earlier draft of the paper. We thank W.U. Reimold for his valued advice during the final editing of the manuscript.

Editorial Handling— Dr. Uwe Reimold

Ancillary