Discovery of a Devonian mafic magmatism on the western border of the Murzuq basin (Saharan metacraton): Paleomagnetic dating and geodynamical implications M.E.M. Derder a, * , S. Maouche a , J.P. Liegeois
b , B. Henry c , M. Amenna a , A. Ouabadi d , H. Bellon e , O. Bruguier f , B. Bayou a , R. Bestandji a , O. Nouar a , H. Bouabdallah a, 1 , M. Ayache a , M. Beddiaf g a CRAAG, B.P. 63, Bouzar
eah, 16340 Alger, Algeria b Geodynamics and Mineral Resources, Royal Museum for Central Africa, B-3080 Tervuren, Belgium c Pal
eomagn
etisme, Institut de Physique du Globe de Paris, Sorbonne Paris Cit
e, Univ. Paris Diderot and UMR 7154 CNRS, 4 avenue de Neptune, 94107 SaintMaur Cedex, France d Laboratoire “G
eodynamique, G
eologie de l'Ing
enieur et Plan
etologie”, FSTGAT / USTHB, BP 32, El-Alia Bab Ezzouar, 16111 Alger, Algeria e Universit
e Europ
eenne de Bretagne, UMR 6538 Domaines Oc
eaniques, IUEM, Universit
e de Bretagne Occidentale, Place Nicolas Copernic, 29280 Plouzan
e, France f G
eosciences Montpellier, Universit
e de Montpellier, UMR-CNRS 5243, 34095 Montpellier, France g Office National du Parc Culturel du Tassili N'ajjer, Djanet, Algeria article info Article history: Received 10 July 2015 Received in revised form 18 November 2015 Accepted 23 November 2015 Available online 2 December 2015 Keywords: Basic magmatism Isotopic (KeAr, UePb) dating Paleomagnetic dating Devonian Murzuq craton Saharan metacraton abstract Intraplate deformation is most often linked to major stress applied on plate margins. When such intraplate events are accompanied by magmatism, the use of several dating methods integrated within a multidisciplinary approach can bring constraints on the age, nature and source mobilized for generating the magma and in turn on the nature of the intraplate deformation. This study focuses on the large gabbro Arrikine sill (35 km in extension) emplaced within the Silurian sediments of the western margin of the Murzuq cratonic basin in southeastern Algeria. Its emplacement is dated during the early Devonian (415-400 Ma) through the determination of a reliable paleomagnetic pole by comparison with the Gondwana Apparent Polar Wander Path (APWP). This age can be correlated with deep phreatic eruptions before Pragian time thought to be at the origin of sand injections and associated circular structures in Algeria and Libya. For the sill, the K eAr age of 325.6 ± 7.7 Ma is related to a K-rich aplitic phase that has K-enriched by more than 20% the Devonian gabbro. Laser-ICP-MS UePb method dates only inherited zircons mostly at c. 2030 Ma with additional ages at c. 2700 Ma and younger ones in the 766-598 Ma age range. The Arrikine sill is a high-Ti alkaline gabbro having the geochemical composition of a hawaiite akin to several intraplate continental and oceanic provinces, including the contemporaneous Aïr ring complexes province in Niger, but also to the Mauna Loa volcano in Hawaii. This peculiar composition akin to that of the contemporaneous Aïr province is in agreement with a lower Devonian age for the Arrikine sill. The lower Devonian Arrikine sill emplacement is related to a “Caledonian” transtensive reactivation of the western metacratonic boundary of the Murzuq craton. This event also generated in the Saharan platform the so-called “Caledonian unconformity” of regional extension, the Aïr ring complexes and magmatic rocks that produced sand injections. It could be related to rifting of the Hun terranes that occurred at the plate margin to the north (Stampfli and Borel, 2002, Blackey, 2008 and references therein). The mid-Carboniferous (c. 326 Ma) reactivation corresponds to Variscan compression on NW Africa generating aplitic fluids, but also to the major “Hercynian unconformity” of regional extension. The * Corresponding author. E-mail addresses: m.e.m.derder@gmail.com (M.E.M. Derder), said.maouche@gmail.com (S. Maouche), jean-paul.liegeois@africamuseum.be (J.P. Liegeois),
henry@ipgp.fr (B. Henry), mohamed20_dz@yahoo.fr (M. Amenna), ouabadi@yahoo.fr (A. Ouabadi), herve.bellon@univ-brest.fr (H. Bellon), bruguier@gm.univ-montp2.fr (O. Bruguier), bbayou57@yahoo.fr (B. Bayou), bestandjirafik@gmail.com (R. Bestandji), obnouar@hotmail.com (O. Nouar), geoexplodz@gmail.com (H. Bouabdallah), mohamedayache62@ yahoo.fr (M. Ayache), mbeddiaf@hotmail.com (M. Beddiaf). 1 Now at geoexplo: www.geoexplo.dz. Contents lists available at ScienceDirect Journal of African Earth Sciences journal homepage: www.elsevier.com/locate/jafrearsci http://dx.doi.org/10.1016/j.jafrearsci.2015.11.019 1464-343X/© 2015 Elsevier Ltd. All rights reserved. Journal of African Earth Sciences 115 (2016) 159e176 generation of the Arrikine magma is attributed to partial melting through adiabatic pressure release of uprising asthenosphere along tectonically reactivated mega-shear zones, here bordering the relictual Murzuq craton enclosed in the Saharan metacraton. © 2015 Elsevier Ltd. All rights reserved. 1. Introduction The Murzuq basin located in central North Africa, in Algeria, Libya and Niger (Fig. 1) is a key area within the Saharan metacraton, being located on a relictual cratonic nucleus, the Murzuq craton (Liegeois et al., 2013
). It is one of the largest cratonic sedimentary basins of the Saharan platform, filled by different Phanerozoic series covered by a very large sandy desert, the Murzuq edeyen or erg. On its western border, we discovered a very large sill of mafic rocks interbedded in the Silurian sedimentary series belonging to the socalled Tassilis series. The sedimentary sequences unconformably rest on the late-Neoproterozoic Pan-African basement of the Tuareg Shield (Fig. 2). Known post-Pan-African magmatism in the InEzzane area is limited to Cenozoic lava flows (2.86 ± 0.07 Ma; Yahiaoui et al., 2014). Around the Murzuq basin, several large Cenozoic volcanic areas are known: Hoggar, Tibesti, Libyan basaltic fields (Liegeois et al., 2005; Fezaa et al., 2010
). On the southwestern border of the Murzuq basin in Niger (between 11
530 E/22
270 N and 12
040 E/21
530 N), Paleozoic dolerites are mapped on the “Carte geologique du Sahara central
” (Lelubre et al., 1962). These dolerites seem to be interbedded between Silurian and Devonian sedimentary sequences and correspond to a very large volume of mafic rocks (the biggest outcrop is indicated to be 90 km long on this map). Several hundred kilometers far to the west and southwest of In-Ezzane, Lower Carboniferous basic intrusions (Djellit et al., 2006; Derder et al., 2006) in the Tin Serririne basin and Devonian ring complexes in Aïr region (Black et al., 1967; Moreau et al., 1994) have been evidenced. In addition, Moreau et al. (2012) suggested that three areas with “sand injections”, which generated circular structures along the Murzuq basin boundaries during lower Devonian, could be related to deep phreatic eruption before Pragian times (Lower Devonian, 407e410 Ma). One of these structures is very close to the sill in the In-Ezzane area (Fig. 1A). We note that the In-Ezzane circular structures proposed to be meteoritic impacts (Bonin et al., 2011; Reimold and Koeberl, 2014) are located more to the West, within the Ordovician sandstones. Such an intraplate magmatism is a key event for understanding the geodynamical evolution of the Saharan platform. The aim of this paper is to date this event and to determine its different implications. 2. Geological setting The Murzuq Basin is the principal feature of Central Sahara. This basin lies across Libya, Algeria and Niger. From a geomorphological point of view, this intracratonic sedimentary basin can be described as an enormous flat hollow of approximately 900 km in diameter from Libya in the north to Niger in the south. This basin is filled by different Phanerozoic series covered by a very large sandy desert (the “Murzuq edeyen or erg””). The main morphological units of this basin consist of plateaus, oueds and sand dunes cover. The Tassili-n-Ajjer (Algeria) and Djado (Niger) uplifted plateaus constitute the southwestern and southern margins of this basin. Primary geological formations dip gently from the flanks towards the basin centre in a concentric pattern. In the In Djerane and Tadrart areas (Figs. 1 and 2), the different Paleozoic series (CambroOrdovician, Silurian, Devonian, Carboniferous and Permian e Henry et al., 2014) outcrop in several trenches of oueds, which produce remarkable rock cliffs (Fig. 2). To the west, the Murzuq craton is bordered by its metacratonic margin, corresponding to the Djanet terrane, made of greenschist Ediacaran sediments (Fezaa et al., 2013) crosscut by 575e545 Ma old high-level plutonic bodies, batholiths, plutons and dyke swarms, (Fezaa et al., 2010). Further west, on the other side of a major NWeSE trending shear zone, the Edembo terrane resulted from a stronger intracontinental metacratonic reactivation that generated amphibolite-facies migmatitic gneisses and aluminous granitoids during the same period (c. 568 Ma; Fezaa et al., 2010). This intracontinental reactivation (575-545 Ma metacratonization) occurred to the east of the Raghane mega-shear zone (western boundary of the Saharan metacraton; Abdelsalam et al., 2002) in a region previously stable, that was covered by sediments deriving from the more western Tuareg Shield in the 595-575 Ma time span (Fezaa et al., 2010). These metacratonic reactivations occurred mostly along pre-existing Eburnean (c. 2 Ga) lithospheric-scale terrane boundaries, along the Murzuq craton, of probable Archean age (Fezaa et al., 2010). These metacratonic zones were reactivated during the Mesozoic, with deposition of several thousand meters of sediments in troughs, and during the late Cenozoic, with the emplacement of the In-Ezzane basaltic field (Liegeois et al., 2013
). The Phanerozoic sediments of the Murzuq basin are located to the East, partly resting on the Murzuq metacratonic margin, the Djanet terrane. The Arrikine area consists of a monoclinal structure dipping 6
toward E to ENE. This structural orientation gradually changes. The Murzuq geological map (Geological Map of Libya (1985)) shows that its borders, including Cretaceous levels, were uplifted and slightly tilted. Very locally, folding has been observed in the InEzzane area (Amenna et al., 2014), probably related to far-field effects of the Hercynian orogeny (Haddoum et al., 2001), again reactivating the metacratonic structures. In this area, we discovered a large sill of mafic rocks (maximum observed NNW-SSE outcrop extension of about 35 km) interbedded in Silurian sediments (Fig. 2). The current description of the local stratigraphy (Freulon, 1964; Bellini and Massa, 1980; Legrand, 2000; Ghienne et al., 2013), together with the published maps of this area, do not mention any occurrence of mafic rocks. The sill intruded the upper member of the Oued In Djerane Formation, mainly made of shales. This upper member displays signs of a more marine environment than the underlying levels, but in its top part, presents facies that are more and more regressive (Legrand, 2000). It is overlain by the detritic Acacus Formation. The host-rocks constituting the roof of the sill are affected by a contact metamorphism. 3. Petrography and mineralogy The Arrikine sill (Fig. 2) has a maximum thickness of c. 250 m and is dipping gently to the ENE similarly to the host-rocks series. The Silurian host-rocks, both at the floor and roof of the sill, show local folding due to the intrusion. Farther from the intrusion, in the Ordovician series, deformations become brittle. The sill is 160 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 Fig. 1. Geological map of the southwestern Murzuq basin margin: A) Location in the North African context based on Lelubre et al., 1962; Liegeois et al., 2003, 2005; Fezaa et al., 2010
. White rectangles indicate the circular structures discussed in the text: Mt T. ¼ Mount Telout; I.E. ¼ In-Ezzane; El Meh. ¼ El Meherschema. B) Geological map of the studied area (modified after Le Caignec et al., 1957; Fezaa et al., 2010 and Geological Map of Libya, 1985, SW sheet). M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 161 constituted of a pyroxene-rich gabbro. On its borders, it is light gray in color due to a strong weathering, while in the other parts, it is dark in appearance and not or slightly altered. These dark-facies rocks are medium-grained grayegreen gabbros. The main mineral is plagioclase appearing as large lengthened intertwined laths. The mafic mineral assemblage is dominated by granular clinopyroxene. Minor constituents are thin laths of biotite associated with oxides, few small quartz grains and relatively large granules of opaque minerals. Weathering renders plagioclase brownish in color and is at the origin of the incomplete replacement of some pyroxene by chlorite. Microprobe analyses (Table 1) confirm that the magmatic plagioclase is a labrador (An55), the pyroxene a diopside, the presence of biotite and apatite and that the oxides are ilmenite and magnetite. The biotite is particularly rich in TiO2 (c. 7%), which is also a characteristic of the biotite from the Iskou ring complex in Aïr (Leger, 1985
). These analyses also show that FeeMg minerals are often altered but above all, that an aplitic component is present and is represented by small spots of K-feldspar, with diverse K compositions ranging from 2 wt % to 15 wt %. This K enrichment is accompanied by a Ba-enrichment (Table 1) developed within acid plagioclase (An13-17), two minerals that cannot be in equilibrium with the gabbroic magma. The development of this phase enriched in K and Ba, has consequences for the whole rock geochemistry and its KeAr dating (see below). 4. Isotopic dating 4.1. K/Ar geochronology Analyses were done by one of us (H.B) at the KeAr dating platform of UMR 6538 “Domaines Oceaniques
” CNRS, UBO at Brest. 4.1.1. Analytical procedure This sample of gabbro (AF-ALG/Dj-Site58 from site 58 e Fig. 2) was also considered for dating on a whole-rock fraction made of grains 0.3 to 0.15 mm in size, prepared after hand-crushing and sieving of the solid rock, then cleaned using ultrapure water. One aliquot of grains was powdered in an agate grinder. Then two independent dissolutions of 80 mg of powdered sample were realized in two well-sealed synthetic bottles using 4 cm3 of hydrofluoric acid at a temperature of 80
C for 8 h. K content of the two HF solutions was determined by Atomic Absorption Spectrometry after dilution (1/2500). A second aliquot of grains was reserved for under vacuum extraction of all gases by heating in a molybdenum crucible, using a high frequency generator. Cleaning of active gases from the whole extracted gases was realized during and after their extraction by gettering on hot titanium sponge contained in three quartz traps successively activated when their temperature was decreasing from 800
C to the ambient one during 10 min. The final step was Fig. 2. Detailed map of the Arrikine sill, from image analysis through Landsat 8 satellite images and from field observations and measurements: a) Names on the map concern geographical location of the two main “oueds” in this area. A color composite analysis, using TM6, TM3 and TM1 bands with contrast enhancement, highlights a NW-SE dark band corresponding to magmatism observed and sampled on the field. Contrasting spectral responses allow differentiating the Silurian clays in purple and the Devonian sandstones in brown. b) Names on the map concern geological formations. The relative homogeneity of responses within these different areas, leads us to try to specify the extension and edges of the sill. Four classes have been defined (green clays, marls and limestones; red dark rocks; brown sandstones; white alluvial material). The result is a classified image by maximum likelihood method. The Cambro-Ordovician levels are not differentiated by this classification and the magmatism in red is well underlined with its NW-SE orientation. However, the northern part of the red area shows a surprising lateral eastward extension. Comparison with Fig. 2a rather suggests drainage of dark material by the Oued Arrikine. The paleomagnetic sampling sites are indicated with numbers by yellow squares (gabbro) and black dots (sedimentary formations). Cambro-Ordovician (ks), Silurian (s) with its two formations (Oued In Djerane and Acacus) and Devonian (d). Yellow circles correspond to injectites. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 162 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 the ultra-purification using an Al-Zr SAES getter. Isotopic composition of argon and concentration of 40ArR were measured in a stainless steel mass spectrometer with a 180
geometry. Isotopic dilution was realized during the fusion step, using precise concentrations of 38Ar buried as ions in aluminum targets (Bellon et al., 1981). A huge degassing of grains was observed and measured for the two first extractions from 0.4 to 0.3 g of sample and led to a long and difficult purification of argon; consequently it was less easy to reduce the effects of signal evolution through time. Two complementary extractions of gases from significantly lighter weights (0.07 and 0.06 g) have permitted to restore better conditions for purification and for isotopic analysis. 4.1.2. Data Resulting data of the four extractions and isotopic analyses are presented in Table 2. Ages are calculated using Steiger and J€ ager's (1977) constants and errors following the equation of Cox and Dalrymple (1967). 4.1.3. Interpretation A mean age of 325.6 ± 7.7 Ma results from the analyses, i.e. a Serpukhovian age (from uppermost Visean to upper Bashkirian, if taking error margins into account). The Silurian age of the host sedimentary series questions the significance of the KeAr age particularly when considering the huge degassing of samples under vacuum that may indicate the effects of secondary processes that affected the Arrikine sill. Such a result points to an event occurring during the Variscan period. Therefore, it is legitimate to raise the question whether a geochemical remobilization of the sill, leading to a potassium and barium enrichment in the plagioclases (Table 1), has occurred. This alteration could have been promoted by cryptocirculations of fluids (supposed hot but not exactly known) that may have reset the KeAr system by complete external diffusion of radiogenic 40Ar. Following such a hypothesis, the 325.6 Ma age marks the time to which the gabbro should have been rejuvenated during the Variscan evolution. 4.2. UePb geochronology 4.2.1. Analytical procedure Zircons were separated from 4 kg of rocks and were subsequently processed according to standard technique (e.g. Bosch et al., 1996), including magnetic separation and careful handpicking under a binocular microscope. UePb dating on zircon was performed by laser ablation sector field ICP-MS at Geosciences Montpellier (France). The ICP-MS (ThermoFinnigan Element XR high resolution) was coupled to a Lambda Physik CompEx 102 excimer laser generating 15 ns duration pulses of radiation at a wavelength of 193 nm. Data were acquired at low resolution (DM/ M ¼ 300) in the fast electrostatic scan mode. Signals were measured in pulse counting mode using 3 points per peaks and a 20% mass window. A pre-ablation step consisting of 7 pulses was applied to clean the sample surface. The acquisition step began with 15 s of gas blank measurement followed by 45 s of laser ablation at a frequency of 4 Hz with a spot size of 26 mm and an energy density of 10 J/cm2 . Measured isotopic ratios were monitored with reference to the G91500 zircon standard with a 206Pb/238U ratio of 0.17917 and a 207Pb/206Pb ratio of 0.07488 equivalent to ages of 1062 Ma and 1065 Ma, respectively (Wiedenbeck et al., 1995). The standard was measured four times each five unknowns in a sequence of 2 standards, 5 unknowns and 2 standards (see Bosch et al., 2011 for details). Analytical data are plotted and ages calculated using the IsoplotEx program (Ludwig 2000). Analyses in Table 3 and in the Concordia plots are ±1s errors and uncertainties in ages quoted in the text are given at ±2s. 4.2.2. Results Only 17 zircons were recovered, which is not so surprising in a basic rock. They display a large range of shapes and colors, indicating they do not form a homogeneous population. Most grains are Table 1 Microprobe analysis. Mineral SiO2 TiO2 Al2O3 FeO MnO MgO CaO Na2O K2O P2O5 Cr2O3 NiO BaO Total Magmatic minerals Plagioclase 58 A-8 54.43 0.150 28.51 0.316 0.000 0.056 11.33 4.933 0.328 0.060 0.000 0.000 0.081 100.19 Pyroxene 58 B-2 50.76 1.027 1.252 12.818 0.246 11.16 21.99 0.506 0.014 0.000 0.002 0.000 0.025 99.81 Biotite 58 B-1 34.86 6.895 12.86 24.52 0.162 7.943 0.000 0.441 9.051 0.011 0.000 0.000 0.000 96.74 Biotite 58 A-5 34.99 7.247 13.53 21.42 0.087 9.491 0.040 0.688 9.068 0.000 0.000 0.000 0.000 96.57 Apatite 58 A-4 0.473 0.000 0.068 0.496 0.048 0.112 54.80 0.003 0.000 42.49 0.000 0.000 0.043 98.53 Opaque 58 A-6 0.005 52.51 0.069 45.22 0.684 1.636 0.051 0.000 0.000 0.000 0.000 0.002 0.000 100.17 Opaque 58 A-7 0.070 26.62 1.269 67.61 1.398 0.014 0.049 0.008 0.006 0.010 0.000 0.001 0.000 97.06 Aplitic secondary minerals Hyper-K spot 58 B-1 64.33 0.000 18.42 0.039 0.003 0.051 0.097 0.837 15.57 0.005 0.000 0.008 0.260 99.61 Hyper -K-Ba spot 58 B-1 61.97 0.073 20.09 0.201 0.000 0.003 0.832 2.837 9.821 0.000 0.000 0.000 3.481 99.30 High-Na spot 58 B-1 65.10 0.125 21.12 0.172 0.015 0.003 2.206 8.490 2.801 0.000 0.000 0.000 0.098 100.13 Altered FeeMg minerals Altered spot 58 A-1 37.97 0.146 2.455 23.85 0.195 6.631 10.80 0.139 0.206 3.912 0.000 0.000 0.000 86.30 Altered spot (brown) 58 A-2 31.03 0.025 12.26 29.37 0.179 14.40 0.340 0.029 0.029 0.000 0.000 0.013 0.000 87.67 Altered spot (green) 58 A-3 31.86 0.057 11.26 30.80 0.129 13.52 0.447 0.105 0.045 0.000 0.000 0.000 0.034 88.25 Table 2 KeAr dating of the Arrikine sill. Sample code Lab. anal. ref. Age (Ma) ± Error at ± 1 s K2O wt. % 40Ar R (105 cm3 g1 ) 40Ar R (%) Fused weight (g) 36Arexp (1010cm3 ) AF-ALG/Dj-Ste58 B 7398-1 327.2 ± 7.6 1.40 1.619 85.3 0.4042 38.2 B 7399-2 314.2 ± 7.3 1.549 86.8 0.3008 23.4 B 7437-8 333.5 ± 7.7 1.654 86.7 0.0580 5.00 B 7459-11 327.9 ± 7.6 1.623 83.4 0.0729 7.96 Mean age 325.6 ± 7.7 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 163 subhedral to rounded, translucent and colorless to pink in color. Although some grains have euhedral shapes, the variety of shapes and the occurrence of rounded grains indicate that most grains may have been snatched from the country rocks, which possibly include a detrital component. Reported in the Concordia diagram (Fig. 3), analyzed grains display a wide range of ages, from Neoproterozoic to Archean, ranging from, 598 ± 14 Ma to 2707 ± 38 Ma (2s, see Table 3). The main population is Paleoproterozoic with a mean age of around 2030 Ma (n ¼ 7). The Neoproterozoic population is dominated by a coherent group of four grains plotting concordantly at c. 682 Ma, while other grains are concordant at 766, 632 and 598 Ma. Lastly the recovered zircon population includes a homogeneous Archean component dated at c. 2700 Ma. 4.2.3. Interpretation The Arrikine sill cannot be older than the intruded Silurian sediments The dated zircons are therefore xenocrysts and not new minerals crystallized within the sill. This agrees with its low-silica geochemical composition and high temperature of crystallization, conditions during which the zircon saturation is very rarely attained (Watson and Harrison, 1983; Boehnke et al., 2013). The inherited zircons can have been incorporated in the Arrikine magma by incorporating (1) unmelted zircons from the source, (2) zircons from the basement during its ascent or (3) zircons from the Ediacaran or Paleozoic sediments during its final emplacement, these possibilities being not mutually exclusive. Incorporating unmelted zircon crystals from the source is unlikely, again because such magmas are far from zircon saturation and are thus efficient to dissolve zircon (Watson and Harrison, 1983). Moreover, the source of the Arrikine gabbro (the mantle or the lower mafic crust) itself most probably bears very few zircons for the same reason. The ages of the inherited zircons in the Arrikine gabbro (0.6, 0.63, 0.68, 2.03 and 2.7 Ga) match those of the detrital zircons from the Ediacaran sediments of the Djanet terrane, just to the west and probably present at depth below the Arrikine area (Fezaa et al., 2010). However, 2.7 Ga zircons are, strictly speaking, not present in these Ediacaran sediments (only 2.65 and 2.8 Ga old zircons are present). In addition, it is noteworthy that they bear other age peaks, especially at 0.7, 1.9 and 2.5 Ga, which are not present in the Arrikine sill. The source of these sediments is Central Hoggar, especially its closest eastern part, explaining that some samples have a proximal immature source with few detrital zircon ages Table 3 UeThePb laser ablation ICP-MS analyses for zircons extracted from the Arrikine gabbro. Sample Pb* (ppm) Th (ppm) U (ppm) Th/U 208Pb/206Pb 207Pb/206Pb ± (1s) 207Pb/235U ± (1s) 206Pb/238U ± (1s) Rho Apparent age (Ma) Conc (%) 206Pb/238U ±(1s) 207Pb/206Pb ±(1s) Arrikine gabbro #1-1 87 82 142 0.58 0.16 0.18473 0.00207 13.247 0.112 0.5197 0.0041 0.94 2698 18 2696 18 100.1 #2-1 44 17 421 0.04 0.02 0.06024 0.00094 0.856 0.011 0.1030 0.0009 0.64 632 5 612 34 103.2 #2-2 25 7 247 0.03 0.02 0.12609 0.00177 6.777 0.078 0.3895 0.0037 0.82 2121 17 2044 25 103.7 #3-1 219 112 523 0.21 0.07 0.18420 0.00221 13.218 0.124 0.5201 0.0045 0.92 2700 19 2691 20 100.3 #4-1 134 151 213 0.71 0.19 0.12365 0.00147 6.635 0.061 0.3889 0.0032 0.88 2118 15 2010 21 105.4 #5-1 664 113 1037 0.11 0.02 0.18601 0.00217 13.412 0.121 0.5227 0.0044 0.93 2711 19 2707 19 100.1 #6-1 96 118 155 0.77 0.22 0.12506 0.00137 6.572 0.054 0.3809 0.0029 0.93 2081 13 2030 19 102.5 #7-1 308 84 804 0.10 0.04 0.06124 0.00189 1.066 0.030 0.1262 0.0017 0.46 766 10 648 65 118.3 #8-1 8 19 62 0.31 0.10 0.12485 0.00138 6.397 0.053 0.3715 0.0028 0.92 2036 13 2027 20 100.5 #9-1 12 56 103 0.55 0.16 0.06410 0.00108 0.946 0.014 0.1070 0.0010 0.63 656 6 745 35 88.0 #10-1 16 102 130 0.78 0.25 0.06221 0.00149 0.945 0.021 0.1102 0.0012 0.50 674 7 681 50 98.9 #11-1 10 38 91 0.41 0.14 0.06207 0.00129 0.955 0.018 0.1116 0.0011 0.54 682 6 676 44 100.9 #12-1 22 143 177 0.81 0.25 0.12459 0.00135 6.088 0.049 0.3543 0.0027 0.94 1955 13 2023 19 96.7 #13-1 648 127 1628 0.08 0.02 0.06548 0.00159 1.004 0.022 0.1112 0.0012 0.50 680 7 790 50 86.1 #13-2 338 72 1007 0.07 0.02 0.06236 0.00116 0.970 0.016 0.1129 0.0011 0.57 690 6 686 39 100.5 #14-1 7 47 68 0.69 0.21 0.05945 0.00195 0.796 0.024 0.0972 0.0012 0.42 598 7 584 70 102.4 #15-1 106 58 327 0.18 0.06 0.13819 0.00152 6.015 0.049 0.3158 0.0024 0.93 1769 12 2205 19 80.2 #16-1 466 204 1162 0.18 0.05 0.12585 0.00135 6.459 0.051 0.3724 0.0028 0.94 2041 13 2041 19 100.0 #17-1 567 210 1533 0.14 0.05 0.12481 0.00132 6.358 0.049 0.3696 0.0027 0.95 2027 13 2026 19 100.1 Fig. 3. UePb concordia diagram. 164 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 (Fezaa et al., 2010, 2013). So, the Ediacaran sediments could have provided some of the inherited zircons in the Arrikine sill, but other sources should be found. The detrital zircons from the Tassilis Ordovician sandstones just to the West (Linnemann et al., 2011) display much more numerous age peaks. These sediments are considered as deriving from sources located all over West Africa and even Amazonia (Linnemann et al., 2011). However, the 2.7 Ga age is systematically absent. Incorporation of detrital zircons from these Ordovician mature sediments is thus not favored. The same conclusion can be reached for the Silurian series in which the Arrikine sill intruded. Its source is also distal and it is made of shales whose fine granulometry does not favor the presence of zircons. The basement is not outcropping in the Easternmost Hoggar. The oldest known rocks are the Ediacaran sediments, metamorphosed in the greenschist facies (Djanet terrane) or migmatized in the amphibolite facies (Edembo terrane), intrusive and subvolcanic rocks being younger (Fezaa et al., 2010). Ediacaran magmatic rocks could have had extracted material but few old ages have been measured in their zircons. The Tin Amali dyke swarm for example provided a 2090 ± 40 Ma age (Fezaa et al., 2010) for an inherited component. This age is not present in the detrital zircon age spectrum of the Ediacaran sediments and is likely to be related to the age of one basement component. More information can be obtained from the Nd isotopes. The mean TDM model ages for the various magmatic bodies (batholith, pluton, and dyke swarm in the 575-545 Ma age range) fall in the 1.6e1.8 Ga age interval (Fezaa et al., 2010). This points to a Paleoproterozoic and/or Archean basement in the area. More to the west, but still to the east of the Raghane shear zone that constitutes the western boundary of the Saharan metacraton, the basement of the Aouzegueur terrane (the third Eastern Hoggar terrane; Fig. 1A) has been dated at 1918 ± 5 Ma (LA-ICP-MS UePb zircon age; Nouar et al., 2011). This basement is intruded by various granitoids close to the Raghane shear-zone at c. 790 Ma, c. 590 and c. 550 Ma (Henry et al., 2009; Nouar et al., 2011). All these ages do not correspond to the inherited zircons dated in the Arrikine sill. So few informations are available on the basement below the Arrikine sill. Nd isotopes indicate that the basement is Paleoproterozoic and/or Archean in age and an inherited zircon point to an age of 2090 ± 40 Ma. On the other side, the Arrikine inherited zircons have two well-defined old ages at 2698 ± 22 Ma (3 zircons) and 2028 ± 15 Ma (7 zircons) with no other zircons older than 1 Ga. These zircons can be considered as having been snatched from the underlying basement, but in very small quantities due to the low viscosity of this magma, and not from the superficial sediments that bear many ages but these ones. On the other hand, the younger inherited zircons, at 682 ± 7 Ma (4 zircons), 632 ± 10 Ma (1 zircon) and 598 ± 15 Ma (1 zircon) could be sourced from the Ediacaran sediments that derived from the erosion of Neoproterozoic units close to the Raghane shear zone. The Arrikine sill, through its inherited zircons, is thus currently the best probe for getting information on the basement of the Murzuq craton, at least of its western boundary, indicating the existence of Archean (c 2.7 Ga) and Paleoproterozoic (c. 2.03 Ga) lithologies at depth. On the other hand, due to the absence of magmatic zircons from the sill itself, the zircon dating was not suitable to date emplacement of the Arrikine sill. 5. Paleomagnetic dating 5.1. Sampling and analysis procedure The sampling sites for paleomagnetism are located near the oued Arrikine (Fig. 2). 110 standard paleomagnetic samples, taken using a portable gasoline-powered drill and oriented with magnetic and sun compasses, were collected from the 12 sites (sites 50 to 61 on Fig. 2) in the gabbro (including one site in the weathered gray facies on the sill border and another one within a late dyke). 9 samples were chosen in one site (site 49 on Fig. 2) of Silurian hostrocks affected by contact metamorphism close to the roof of the sill. One to three specimens were cut from each core in order to have additional specimens for pilot studies and rock magnetic analysis. Prior to any demagnetization analysis, the specimens were stored in a zero field for at least one month, in order to reduce possible viscous components. The Natural Remanent Magnetization (NRM) of the specimens was measured using a JR5 spinner magnetometer (AGICO, Brno). Several pilot specimens from each site were subjected to a stepwise alternating field (AF) demagnetization up to 100 mT or thermal (TH) demagnetization up to 670
C in order to characterize the components of magnetization. Following combined demagnetization process, some other specimens were subjected to combined AF and TH treatments. The directions of the magnetization components were plotted on orthogonal vector plots (As and Zijderveld, 1958; Wilson and Everitt, 1963; Zijderveld, 1967) and the remaining vectors and vectorial differences of the magnetization on equal-area projections. The direction of the different components was computed using principal component analysis (Kirschvink, 1980). Fisher (1953) statistics was used to determine the mean directions. 5.2. Rock magnetism The NRM has a relatively high intensity in dark facies (mean 3.2 A/m) compared to the weathered site close to the sill border (mean 2.7 102 A/m) and the late dyke (mean 5.3 102 A/m). AF demagnetization yields a relatively regular decrease of magnetization intensity. For most samples, magnetization decrease during thermal demagnetization (Fig. 4) is progressive (IZ647, IZ668) or steep above 400e450
C treatment up to about 530e560
C (IZ654). These maximum blocking temperature values suggest the presence of Ti-poor titanomagnetite as the principal mineral carrier. In few samples (Fig. 4), a strong increase of the magnetization for about 400
C is followed by a sharp decrease around 450
C up to about 530e560
C (IZ718, IZ729). No increase appears during AF demagnetization of these samples, for which complementary studies of the magnetic properties have to be performed and will be the subject of an independent paper. Thermomagnetic curves (low-field magnetic susceptibility as a function of temperature) have been measured for representative samples by heating in air using CS2e3 oven and KLY3 Kappabridge (AGICO, Brno). For the gray weathered facies, during heating, the 1 0.8 0.6 0.4 0.2 0 0 100 200 300 400 500 600 IZ 739 IZ 668 IZ 654 IZ 647 IZ 718 IZ 729 J/J0 T (°C) Fig. 4. Thermal demagnetization curves for representative samples of the dark facies. M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 165 partial cooling curves are often reversible with the heating ones (Fig. 5 - IZ696). In such cases, irreversibility of the thermomagnetic curves appears only after heating at the highest temperatures. During heating at temperatures higher than 400
C, a decrease of susceptibility is observed, giving Curie temperature of about 550
C. This confirms that the principal magnetic carrier is a Ti-poor titanomagnetite. However for some samples, a slight decrease for temperatures higher than 580
C points out the presence of a small amount of hematite besides the titanomagnetite (Fig. 5 - IZ696). In the dark facies, the sample thermomagnetic curves are not reversible since 400e500
C heating, highlighting a relatively complicated mineralogical alteration during heating at higher temperature (Fig. 6 IZ716). Hysteresis loops were obtained for small cores (cylindrical samples of about 3 cm3 ) using a laboratory-made translation inductometer within an electromagnet capable of reaching 1.6 T. The loops show that the remanent coercive force determined from several representative samples is low (around 30 mT) and that a partial saturation of the magnetization is at least reached between approximately 0.2 to 0.4 T (IZ678, IZ725), confirming the presence of a low coercivity component (like Ti-poor titanomagnetite). For weathered gray facies, susceptibility is very low and almost linear hysteresis loops have been obtained, indicating largely dominating effects of the paramagnetic minerals. For dark facies (Fig. 6), a much higher susceptibility is associated with a clear dominance of ferrimagnetic minerals. For these sites, Day plot (Day et al., 1977) indicates pseudo single-domain (PSD) grain size for the titanomagnetite (Fig. 6). In conclusion, Ti-poor titanomagnetite is the main magnetic carrier in the non-weathered gabbro. It is also present in the weathered gray facies, but in much lower amount and associated with few hematite. 5.3. Paleomagnetic analyses TH or AF or mixed AF-TH demagnetizations (Fig. 7) were performed on all samples. Very similar Characteristic Remanent Magnetization (ChRM B) data have been obtained from all methods, in addition to a viscous component A. During the demagnetization process, three different main kinds of evolution were observed for gabbro sites: octobre 2011 08:35 0 0.5 1 1.5 2 2.5 0 100 200 300 400 500 600 700 K/Ko T (°C) IZ 716 0 0.5 1 1.5 0 100 200 300 400 500 600 700 K/Ko T (°C) IZ 696 Fig. 5. Typical thermomagnetic curves (normalized low field susceptibility K/Ko as a function of the temperature T - in
C) for representative samples of the gray (IZ716) and dark (IZ696) facies of the gabbro. Fig. 6. a, b): Hysteresis loops (a: sample IZ725 and b: sample IZ678, both for dark facies); Hcr is the remanent coercive force, Field H in Tesla. c) Day plot (Day et al., 1977) of the hysteresis ratios; PSD (pseudo-single domain) and MD (multidomain) areas for magnetite. 166 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 Fig. 7. Orthogonal vector plots (filled circles: horizontal plane, crosses: vertical plane), in stratigraphic coordinates for AF (a, c and e), thermal (b, d, f) and mixed (g) treatments for samples IZ659A (a), IZ664A (b), IZ709A (c), IZ739A (d), IZ711A and B (e, f) and IZ660A (g). M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 167 - The first type, obtained on half the samples, is illustrated by a single stable magnetic direction. Fig. 7a, b and g show examples of this behavior obtained from an AF method (IZ659), thermal process (IZ 644) or mixed AF-TH (IZ660A). In this case, after elimination of the viscous component A, the analysis of the magnetization shows that the linear segments towards the origin on Zijderveld (1967) plots, define the ChRM B. - The second type of behavior shows magnetic directions evolving first along a great circle (remagnetization circle), which indicates superimposition of the unblocking temperature (or unblocking fields) spectra for at least two remanent magnetization components. This evolution always reaches a ‘‘stable end point’’ direction yielding the ChRM B. This case is illustrated on Fig. 7c and d for the AF method (IZ 709) and the thermal method (IZ 739). Here also, similar results are obtained by the two method of demagnetization. The secondary component at the origin of the variation along the great circles cannot be properly identified. It could correspond to a Cenozoic overprint (Henry et al., 2004; see also for example Hargraves et al., 1987), but other origins cannot be rejected by lack of constraints. - The third case (Fig. 7e, f) displays similar variation than the first case, except that, during the thermal treatment, a component, with the same direction as the main magnetization but with opposite polarity, appears at c. 400
C and progressively disappears at higher temperature (from c 500
C), yielding a welldefined ChRM magnetization. During AF demagnetization, this intermediate component of opposite polarity is not present. The final ChRM directions obtained for both demagnetization processes are the same. This additional component is therefore related to a mineralogical alteration due to heating. These samples will be the subject of a complementary analysis about magnetic properties. The ChRM B obtained from samples with these three behaviors was isolated between 200 and 500
C for the lowest temperatures and 400e560
C for the highest ones (thermal process), and between 10 and 70 mT for the lowest blocking field and 60 and 102 mT for the highest ones (AF demagnetization). Results from the sites sampled in weathered facies on the sill border (14 samples) and within a late dyke (12 samples) are mostly not reliable. The obtained results from all the other samples are homogeneous (Table 4). The mean direction of this ChRM B computed from 84 specimens (from 12 sites) has the following directions: D ¼ 355.4
, I ¼ 38.4
, k ¼ 38, a95 ¼ 2.5
and D ¼ 350.7
, I ¼ 38.6
, k ¼ 38, a95 ¼ 2.5
, before and after dip correction, respectively (Table 4 e Fig. 8). This result is similar to that obtained from the mean direction of the 10 sites (two sites with less than 3 ChRMs being not considered): D ¼ 355.5
, I ¼ 38.8
, k ¼ 86, a95 ¼ 5.2
and D ¼ 350.6
, I ¼ 38.9
, k ¼ 86, a95 ¼ 5.2
, before and after dip correction, respectively (Fig. 8). The paleomagnetic data obtained from the site of Silurian hostrocks affected by contact metamorphism have similar characteristics with neighboring ChRM orientation (N ¼ 9 samples, D ¼ 358.6
, I ¼ 42.8
, k ¼ 402, a95 ¼ 2.6
and D ¼ 349.5
, I ¼ 41.7
, k ¼ 402, a95 ¼ 2.6
, before and after dip correction, respectively) but with lower magnetization intensity. Other sites, taken in the Ordovician and Silurian rocks far from the sill (sites 43 to 48 on Fig. 2), yield only weak Cenozoic remagnetization of normal and reversed polarity (see Henry et al., 2004), totally overprinting any possible previous earlier magnetization. That is then not the case for the site affected by contact metamorphism. The samples from this site carried a much more stable magnetization than the other sites. That is attributed to an effect of the contact metamorphism generated by the sill. This kind of positive contact test strongly argues for a primary magnetization of the gabbro. The corresponding paleomagnetic pole determined from the sill data is defined by l ¼ 43.6
S, F ¼ 17.0
E, K ¼ 88 and A95 ¼ 4.7
and l ¼ 42.8
S, F ¼ 22.9
E, K ¼ 88 and A95 ¼ 4.7
, before and after bedding correction, respectively. 5.4. Paleomagnetic dating of the Arrikine sill A gabbro magnetization age younger than the post-Cretaceous tilting of the formations is unrealistic when comparing the Arrikine poles with the known Apparent Polar Wander Path - APWP - (Fig. 9) and data can only be used after applying a bedding correction. The so corrected Arrikine pole (Fig. 9) corresponds to the Gondwana APWP segment for the 400 Ma period of Derder et al., 2006, Amenna et al., 2014 and Henry et al., 2016 and to that of Torsvik et al., 2012 (overlapping of confidence zones) for the 410 Ma mean pole. In the latter, the uncertainty on the paleomagnetic mean pole age is not taken into account while it corresponds to a single pole (Aïr intrusives, Hargraves et al., 1987), which can explain the difference existing with the former one. Indeed, integration of such uncertainties in the first APWP adds a weighted effect for poles having a large age uncertainty and thus covering several windows in time, smoothing the curve. This is particularly critical for periods affected by strong changes of the APWP drift such as the 400e410 Ma period (Fig. 9). Geologically, the Arrikine sill is post-Silurian and pre-Cretaceous. The Arrikine pole is very far from all mean paleopoles from Middle Devonian to Present times, especially for Carboniferous times. It is also far from Silurian mean paleopoles (Fig. 9). By contrast, the Arrikine pole is consistent with Early Devonian poles of Gondwana: Mt. Leyshon Devonian dykes in Australia (Clark, 1996), Snowy River Volcanics in Australia (Schmidt et al., 1987), Herrada Member, Sierra Grande in Argentina (Rapalini and Vilas, 1991) and Aïr alkaline ring-complexes in Niger (Hargraves et al., 1987). In particular, the coincidence of the Arrikine pole with this last pole (l ¼ 43.4
S, F ¼ 8.6
E, K ¼ 50 and A95 ¼ 6.2
) is remarkable. The angular difference between these paleomagnetic poles (10.7
) is of the order of what can be ascribed to the secular variation. In both the Arrikine sill and the Aïr alkaline ring-complexes, the paleomagnetic directions have the same single polarity. So the Aïr ring-complex age of 407 ± 8 Ma (Moreau et al., 1994) can be considered as a good approximation of the age of the Arrikine sill, which can be fixed to the 400e415 Ma age interval, i.e. at the base of the Emsian, from mid-Lochkovian to mid-Emsian (Cohen et al., 2013), thus globally Lower Devonian. The loop of the Gondwana APWP between the Late Ordovician (pole 450 Ma) and Late Devonian (370 Ma) was not generally deemed as very reliable. The Aïr and Arrikine coherent mean is “unique” and supports the existence of this loop that has to be considered for the paleocontinental reconstructions for this period. 6. Geochemistry Only one analysis (sample AF-ALG/Dj-Site58 from site 58 e Fig. 2) is available and listed in Table 5. However, its geochemical signature (Fig. 10) is sufficiently diagnostic to allow a proper geochemical characterization of the Arrikine sill and to deduce constrained conclusions about its origin. All oxides are considered in column a* as deduced from the geochemical analysis by ICP and in column b* are recalculated to 100% on an anhydrous basis. Oxide % are always wt. %, unless otherwise stated. The Arrikine sill is a fine-grained gabbro but (see below) its chemistry is similar to intraplate alkali basaltic rocks in continental flood basalt (CFB) and ocean island basalt (OIB) provinces. So, in order to make the comparison more efficient, geochemical diagrams dedicated to volcanic rocks and volcanic nomenclature are applied here, which is justi- fied by the subvolcanic character of the sill and by the difficulty of 168 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 counting a mode on this kind of rock. The Arrikine gabbro having 45.99% SiO2 and 5.79% Na2O þ K2O, on an anhydrous basis, belongs to the alkali series, being an alkali gabbro. It has the composition of a trachybasalt (Le Bas et al., 1986), more precisely of an hawaiite (wt%Na2O-2 > wt%K2O), i.e. a sodic trachybasalt (Le Maitre et al., 2002). The depleted MgO content (4.1%) corresponds to typical hawaiite (3e5.5% MgO; MacDonald and Katsura, 1964) as is the absence of chromite (Cr: 3 ppm; Ni: 39 ppm), a characteristic of hawaiite (Cr: 5e150 ppm; Ni: 5e80 ppm; Nelson and Carmichael, 1984 and references therein), in agreement with the differentiated character of hawaiites. Immobile trace elements confirm that the Arrikine gabbro has the composition of an alkali basalt with Nb/Y ¼ 0.85 and Zr/ TiO2*0.0001 ¼ 0.0064 (diagram not shown; Winchester and Floyd, 1977). Low concentration of Ni, Cr and absence of positive Eu anomaly do not point to a significative cumulative component. Even the very low Cr content in Arrikine gabbro points to a differentiated character. In addition, with TiO2 ¼ 5.01%, the Arrikine gabbro belongs to the high-Ti series (TiO2 >2%); it has the composition a high-Ti alkali basalt. CIPW norm points to 60% plagioclase (An30), 11% K-feldspar, 9% pyroxene, 6% olivine, 2% apatite and 11% FeeTi oxides (Table 5) in rough agreement with the petrography except that the plagioclase is acid (actually Labrador) and the absence of biotite, not computed in the norm. This norm is in agreement with a composition of a high-Ti hawaiite. The important difference in the An content of the plagioclase can be ascribed to the late “aplitic” phase determined by the microprobe measurements to be acid oligoclase (An13-17). The An30 value determined by the norm suggests that this aplitic phase was quite pervasive at least in the analyzed sample. This implies that concerned elements (i.e. K, Ba but also Rb, Sr) must be used with caution when the original magma is concerned as well as the KeAr age for which we have calculated (method based on an initial age of 410 Ma given by paleomagnetic considerations and the rejuvenated result at 326 Ma) a K enrichment of up to 20%. The contemporaneous (407 ± 8 Ma; Moreau et al., 1994) large province of alkaline ring-complexes of Aïr bears also high-Ti gabbros, especially the large Ofoud ring-complex where they are associated with numerous layers and/or lenses of titanomagnetite up to 10 m thick (Karche and Moreau, 1977) and the smaller Abontorok ring-complex, associated with massive anorthosites (Brown et al., 1989). These gabbros are of uniform grain size (2e7 mm) and massive in texture and titanomagnetite and ilmenite can make up as much as 10 percent of the rock (Husch and Moreau, 1982). The Aïr ring-complex province is intrusive along the western boundary of the Saharan metacraton (Liegeois et al., 1994
, 2013) as a consequence of its reactivation during early Devonian (Moreau et al., 1994). The Arrikine gabbro is compared to the available analysis from Abontorok ring complex (Brown et al., 1989) and to several intraplate continental high-Ti alkali basalt or gabbro provinces: the Trias rift-related from SW China, similar to the Permian high-Ti basalts from the Emeishan LIP (Zhang et al., 2013), the Ordovician Suordakh and Cambrian Kharaulakh mafic rocks from Siberia (Khudoley et al., 2013) or the 1.5 Ga Yangtse mafic province Table 4 Mean direction: Site, number of ChRM N, Declination D and Inclination I before and after dip correction, Fisher (1953) parameters k and a95, and VGP or paleomagnetic pole: Lat. (
S), Long (
E), with for the latter corresponding Fisher parameters K and A95. Before and after dip correction VGPs and Paleomagnetic pole Site N D(
) I(
) D(
) I(
) k a95(
) Lat(
S) Long(
E) K A95(
) 50 6 357.4 24.7 354.6 25.2 40 10.8 52.4 19.1 51 3 342.6 47.4 336.3 46.3 416 6.1 33.5 35.8 52 6 10.7 45.7 4.6 47.4 27 13.1 37.3 5.4 53 9 357.8 37.1 353.2 37.5 109 4.9 44.5 19.4 54 9 349.8 36.2 345.4 35.8 118 4.8 43.9 29.7 55 2 e e e e ee - - 56 1 e e e e ee - - 57 10 355.8 36.1 352.9 36.5 66 6.0 45.2 20.0 58 6 347.9 35.2 343.7 34.7 72 7.9 44.1 32.2 59 7 2.7 45.5 356.6 46.4 44 9.2 38.2 14.3 60 12 353.7 36.0 349.3 36.0 115 4.1 44.8 24.7 61 13 351.1 41.6 345.8 41.3 21 9.4 40.3 27.6 Mean (specimens) 84 355.4 38.4 350.7 38.6 38 2.5 Mean (Sites) 10 355.5 38.8 350.6 38.9 86 5.2 42.8 22.9 88 4.7 Fig. 8. Equal-area plot after dip correction (a) of ChRMs C and (b) of the mean-site paleomagnetic directions with associated 95% confidence zone (open circles: negative inclinations). M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 169 Fig. 9. In North-West Africa coordinates, comparison of the Arrikine pole with Lower Devonian poles (see text), with the 560-210 Ma Gondwana APWP (in red 560-250 Ma - Amenna et al., 2014; Henry et al., 2016 e in orange 240-210 Ma - Domeier et al., 2012), the Africa APWP (in green 200-0 Ma - Besse and Courtillot, 2002) and in blue the 430-380 Ma Gondwana APWP of Torsvik et al. (2012). Paleomagnetic poles with associated confidence zone of Mt. Leyshon Devonian dykes (Mt. LDD) in Australia (Clark, 1996), Snowy River Volcanics (SRV) in Australia (Schmidt et al., 1987), Herrada Member, Sierra Grande (HM) in Argentina (Rapalini and Vilas, 1991), Aïr intrusives (Aïr) in Niger (Hargraves et al., 1987), Tin Serririne (TS e Derder et al., 2006) and Arrikine (ARK) in Algeria (this study). Confidence zones are also presented for the 250 or 500 Ma poles of the APWPs. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Table 5 Major and trace elements for the Arrikine sill measured on a Horiba-Jobin-Yvon Ultima 2 ICP-AES at the IUEM, Pole de Spectrom ^ etrie Oc
ean, Brest, France, by C
eline Liorzou.
CIPW norm and other parameters have been calculated using Norm4 Excel spreadsheet from Kurt Hollocher: (http://minerva.union.edu/hollochk/c_petrology/norms.htm). Sample AF-ALG/Dj 58 Oxide a % b % Element ppm Normative minerals Vol % SiO2 43.80 45.99 Rb 27.2 Plagioclase 60.47 TiO2 4.77 5.01 Sr 442 Orthoclase 11.05 Al2O3 13.97 14.66 Ba 1247 Diopside 8.16 Fe2O3 15.94 16.73 Sc 19.3 Hypersthene 1.13 MnO 0.19 0.20 V 171 Olivine 5.64 MgO 3.90 4.10 Cr 3 Ilmenite 6.13 CaO 6.28 6.60 Co 46 Magnetite 5.36 Na2O 4.12 4.32 Ni 39 Apatite 2.02 K2O 1.40 1.47 Y 38.6 Zircon 0.04 P2O5 0.87 0.91 Zr 320 Plagioclase An content 29.15 Nb 32.7 LOI 1.84 1.84 La 37.7 Differentiation Index 62.31 Total 97.11 97.11 Ce 84.0 Nd 48.9 Sm 9.9 Calculated density, g/cc 3.03 Eu 3.3 Calculated liquid density, g/cc 2.70 Gd 9.6 Calculated viscosity, dry, Pas 0.17 Dy 7.7 Calculated viscosity, wet, Pas 0.16 Er 3.8 Estimated liquidus temp,
C 1280 Yb 3.3 Estimated H2O content, wt, % 0.20 Th 2.3 a Analysis. b Analysis recalculated to 100% on an anhydrous basis. Calculated FeO/Fe2O3¼ 0.62 170 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 (Fan et al., 2013). We added also for comparison the intraplate highTi potassic basalts from the Virunga (Rogers et al., 1998). When compared to these high-Ti alkali basalts from different settings and different ages, the Arrikine gabbro has globally a similar pattern (Fig. 10). Compared to Mauna Loa volcano in Hawaii (Wanless et al., 2006), the Arrikine gabbro displays a subparallel REE pattern, being slightly more enriched, especially in LREE (Fig. 3a). Similarly, its pattern in the spidergram (Fig. 10b) is parallel, only very slightly enriched in HFSE, more enriched in LILE especially in Ba. High-Ti basalts have a largely preponderant mantle origin by contrast to low-Ti basalts (e.g. Deckart et al., 2005), explaining similar compositions in the oceans and in the continents. This characteristic renders, however, this kind of magmas relatively rare within the continents, which is an important fingerprint of the Arrikine sill. The spatially close Aïr high-Ti gabbros (Abontorok ring complex; Brown et al., 1989) display REE patterns close to that of Arrikine (Fig. 10a) except for Eu, probably due to a plagioclase cumulative pattern (Aïr ring-complexes comprise massive anorthosites - Brown et al., 1989; Demaiffe et al., 1991). Abontorok spidergrams are also close to that of Arrikine, the latter being on the most enriched side. This can be related to its low Cr content suggesting a differentiated character, the differentiation being able to concentrate incompatible elements. However, in the Arrikine gabbro Rb, Ba and K appear particularly enriched; these enrichments can be related to the presence of the late metasomatic aplitic phase (See KeAr section; Table 5) probably favored by the sedimentary environment. This is particularly visible when compared with the Suordakh high-Ti basalt from Siberia (Khudoley et al., 2013) in a primitive mantle-normalized graph (Fig. 11): the Arrikine basalt is very similar to the Suordakh basalts but with concentrations in K as high as the richer Suordakh basalts and with Ba contents much higher than any Suordakh basalts. This indicates that the K, Rb and Ba concentrations in the pristine Arrikine gabbro were probably lower than that currently observed. Fig. 10. Arrikine sill representative sample AF-ALG/Dj-Site58 compared with rocks having also a high-Ti basaltic composition (Aïr, Niger (Abontorok ring-complex): Brown et al., 1989; Siberia: Khudoley et al., 2013, Virunga: Rogers et al., 1998, Hawaii: Wanless et al., 2006, Yangtse: Fan et al., 2013, SW China: Zhang et al., 2013). (A) REE normalized to chondrites (Taylor and McLennan, 1985), (B) Dj 58 spectrum for incompatible elements normalized to MORB (Sun, 1980, Pearce, 1982) compared with high-Ti basaltic composition of rocks of Siberia, Virunga, Hawaii, Yangtse and SW China (see references above) M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 171 By opposition, the other elements (HFSE, Th) can be considered as pristine, being not mobilized by an aplitic fluid. If the Arrikine gabbro composition is similar to the intraplate alkaline continental high-Ti alkali basalt or gabbro provinces, it is distinct from the intraplate high-Ti potassic basalts from the Virunga (Rogers et al., 1998), except for Ba. By opposition to the latter, the sources of the formers are considered to be a more or less enriched asthenospheric source with very low lithospheric participation, similarly to Hawaii volcanoes but in a continental setting. The similarity of the geochemistry of the Arrikine sill with the lower Devonian Aïr magmatism concurs to support the lower Devonian age determined in the paleomagnetic section. Based on the silica content, a simple calculation leads to an estimated liquidus temperature of 1280
C. When considering the experimental work of Botcharnikov et al. (2008) determining a liquid line of descent for ferro-basalts, the Arrikine concentrations in SiO2, MgO, TiO2, K2O, Na2O, Al2O3, FeO, CaO point all to a temperature of c. 1080
C, which is the temperature where the maximum concentration in FeO and TiO2 is attained. This is true when considering nearly anhydrous magmas, which is the case of Arrikine (0.2% H2O, Table 5). Considering the temperature of the mantle partial melting at 1280
C and the crystallization of the Arrikine sill at 1080
C seems reasonable. This is indeed the temperature range observed in experimentation concerning Fe-rich basalts (Chevrel et al., 2014). The calculated viscosity at liquidus temperature is very low (0.17 Pas), as it is the case for Fe-rich basalts whose viscosity is less than 1 Pas (Chevrel et al., 2014). In a 3 mwidth dyke, a magma rise speed higher than 10 m/s can be proposed, a speed that limits fractional crystallization and interaction with country-rocks during ascent (Chevrel et al., 2014). This implies that the differentiation that affected the Arrikine magma (low Cr contents) occurred at depth. Most of these provinces are considered to be or having been fed by a mantle plume. Notwithstanding the fact that the existence of mantle plumes is currently challenged (e.g. Foulger and Hamilton, 2014), there is no one argument for a plume origin in the case of the Arrikine sill: very small magmatic volume, no swell, no rift, no magmatic track, even if a similar asthenospheric mantle source can be inferred for the Arrikine sill. The proposed alternative will be discussed below in the geodynamic section. 7. Emplacement and reactivation of the Arrikine sill: why there and why at these times? 7.1. Regional considerations 7.1.1. The time of emplacement, the lower Devonian During the Pan-African orogeny and the Phanerozoic, the Murzuq craton played an important role as a rheologically rigid body. Magmatism and compressive (Pan-African) or extensional (Phanerozoic) tectonics occurred around it, and the Murzuq basin developed as an intracontinental sag basin in a similar way as Al Kufra and Chad basins (Heine et al., 2008). Formation of the Murzuq basin did not require accommodation of tectonic subsidence through faulting and rifting but resulted from the negative buoyancy caused by thick cratonic lithosphere (Ritzmann and Faleide, 2009). As a consequence, the Murzuq craton is devoid of magmatism except along its metacratonic boundaries where Cenozoic volcanic fields are particularly present all around it (Fezaa et al., 2010; Liegeois et al., 2013
). Indeed, the West African Cenozoic volcanism is attributed to tectonic reactivations induced by the EuropeeAfrica convergence (Liegeois et al., 2005; Beccaluva et al.,
2007; Lustrino et al., 2012; Radivojevi
c et al., 2015). These reactivations, even induced by the rather low stress propagating in remote intracontinental regions, must affect the whole lithospheric thickness allowing for small mantle melt volumes to reach the surface. This is the case of the mega-shear zones within metacratonic areas that are rigid but fractured (Liegeois et al., 2003, 2005,
2013; Bardintzeff et al., 2011) through vertical planar delamination as recently evidenced by magnetotelluric studies in Hoggar Fig. 11. Arrikine gabbro AF-ALG/Dj-Site58 spectrum for incompatible elements normalized to the Primitive Mantle (Sun and McDonough, 1989) compared with Siberian high-Ti basalts (Suordakh event, Eastern margin of the Siberian craton; Khudoley et al., 2013). 172 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 (Bouzid et al., 2015). The Arrikine sill is lower Devonian in age and intruded Silurian sediments. It can be correlated with Paleozoic dolerites in Niger (Fig. 1B), which are also intrusive within Silurian sediments, suggesting a regional event along the southwestern border of the Murzuq basin, even if limited in volume. By opposition, no direct evidence of magmatism, even as detrital pebbles, were described within the sedimentary Devonian formations of the Murzuq basin (Bellini and Massa, 1980). This fact is tentatively related to the so-called Caledonian event (base of the Devonian) that prevented the deposition of lower Devonian sediments (Ghienne et al., 2013). This “Caledonian” event” was marked in the Murzuq basin by large scale uplifts and secondary highs roughly oriented SSEeNNW in SW Libya, i.e. parallel to the PanAfrican major shear zones in the Djanet and Edembo terranes (Davidson et al., 2000; Ghienne et al., 2013) constituting the western metacratonic boundary of the Murzuq craton (Fezaa et al., 2010). This means that the Arrikine sill intruded during a period of reactivations of the mega-shear zones of the region, known to be former terrane boundaries and as such of lithospheric scale (Fezaa et al., 2010). The Aïr ring complex province also resulted from the reactivation of a major shear zone, the Raghane shear zone (Moreau et al., 1994), which corresponds to the western boundary of the Saharan metacraton (Liegeois et al., 1994
; Abdelsalam et al., 2002). The shift from the rapid vertical ascent along the mega-shear zone of the low-viscosity Arrikine magma to the widespread horizontal sill development can be tentatively attributed to the characteristics of the Silurian Acacus formation. The latter is a massive level (about 100 m thick) overlying lagunal deposits that include gypsum-rich beds (Freulon, 1964). We thus suggest that the Arrikine magma intruded within evaporitic-rich facies, which, at the contact with the high temperature magma, became much more easily deformable and moreover released water due to gypsum transformation, largely blocking the magma ascent. However, looking at the age of the magmatism and at the stratigraphic series, the Arrikine sill should have intruded the cover at shallow depth (of the order of less than 300 m), in agreement with the moderate size (PSD) of the magnetite grains. These renders the presence of Arrikine aerial products likely but in that case it is likely that they were eroded and transported far away by the large fluvial system that developed at this period in the Murzuq basin (Ghienne et al., 2013) as it was the case for most of the Lower Devonian deposits in the Murzuq basin (Davidson et al., 2000). This kind of emplacement would be favorable to suggest a link between the Arrikine magmatism and the magmatic rocks proposed to be at the origin of sand injections in the Murzuq area (Fig. 1A; Moreau et al., 2012). Indeed, these peculiar structures occur in the Silurian sediments around the Murzuq basin in a similar manner as the Arrikine sill and are considered as Devonian in age. These sand injections would be the result of the combined effects of tectonic uplift, igneous intrusions in the sand-rich Cambrian-Ordovician sediments and overpressures generated by the paleo-relief existing below the Silurian strata (Moreau et al., 2012). The Arrikine sill would correspond to the magma that intruded the Silurian sediments, lower Devonian sediments having not deposited in the area due to uplift. Let us remark that the three known occurrences of sand injections (Mt Telout, In-Ezzane to the west and El Meherschema to the east; Fig. 1A) are also located around the Murzuq craton, within its metacratonic margin, the Tibesti area being the equivalent of the Djanet and Edembo terranes from Eastern Hoggar (Fezaa et al., 2010). 7.1.2. The time of rejuvenation, the mid-Carboniferous The KeAr method has given an age of 325.6 ± 7.7 Ma that we consider as close to that of a rejuvenation event during the Serpukhovian, i.e. from uppermost Visean to upper Bashkirian when taking into account the error bracket-attributed to aplitic fluid percolation. Such an age is also that of the erasure of the primary magnetization in the older sedimentary rocks of the Saharan platform (Henry et al., 2004). It corresponds thus to the second main tectonic event in the area. This Mid-Carboniferous compression event led to the major so-called Hercynian unconformity that affected the upper part of the Paleozoic sequence in the Murzuq Basin but whose initial phases started during middle Carboniferous (Bellini and Massa, 1980), or at the Visean-Serpukhovian boundary (Frohlich et al., 2010 € and references therein). This event was responsible for the uplift of the Tiririne High (NW Murzuq basin) at the end of the Visean, removing the Assedjefar-Marar (Lower Carboniferous) sequence (Echikh and Sola, 2000). It began with transpressional movements along a series of NNE-SSW trending wrench faults (parallel to the Murzuq craton boundary), producing a NEeSW en echelon pattern of folds on the Tiririne High (Echikh and Sola, 2000). At the scale of the Murzuq basin, the last marine deposits are of Moscovian age (e.g. Massa et al., 1988; Amenna et al., 2014) and the upper Carboniferous is nearly entirely lacking (due to uplift or to erosion) and only some red lacustrine mudstone (e.g. upper Carboniferous Tiguentourine Formation) can be rarely found (Davidson et al., 2000). Such a major compressive event, with transpressional movements and various uplifts at the regional scale is favorable for fluid movements, able to reset the KeAr geochronometer that in turn is able to date it. A Serpukhovian reactivation (“Hercynian”) occurred at the same place as the early Devonian (“Caledonian”) reactivation. This indicates that emplacement of the Arrikine sill was related to the lithospheric character of the mega-shear zones of the region (Fezaa et al., 2010). This implies that these intraplate reactivations are quite important and must be considered. 7.2. Global considerations For a long time, the use of the “Caledonian” and “Hercynian” expressions was suggested for the intraplate intense tectonic reactivations that occurred around the Murzuq craton. Causes must be found in events that occurred at the plate margins at these periods. In addition to tectonic reactivations, volcanism may be generated by adiabatic pressure release of uprising asthenosphere in response to stress resulting from far field collisions such as, for the West African Cenozoic volcanism, the AfricaeEurope collision (Liegeois et al., 2005
). Indeed, these “Caledonian” and “Hercynian” effects are known all over the Saharan platform, to the north of Hoggar (Haddoum et al., 2001; Galeazzi et al., 2010), around the West African craton (Michard et al., 2008) or within the Algerian petroleum basins (Galeazzi et al., 2010) and in the whole Northern Africa (Guiraud et al., 2005). The Alpine reactivation is generally superimposed on the Hercynian reactivated structures (Smith et al., 2006; Derder et al., 2009). These “Caledonian” and “Hercynian” tectonic events are also recorded at the regional scale by detrital zircons in the Murzuq basin (Morton et al., 2011). The Arrikine sill emplacement, as well as that of the Aïr ringcomplex province (Moreau et al., 1994) at c. 410 Ma, occurred during a continental breakup, when the Hun terranes separated from the northern Gondwana margin (Stampfli and Borel, 2002; Blackey, 2008 and references therein). Such a geodynamical environment is favorable to generate intraplate tectonics and magmatism as it was the case, for West Africa, during the continental breakup. This led to the Atlantic ocean opening that generated the Central Atlantic Magmatic Province (Jurassic CAMP; Marzoli et al., 1999), which extends in West Africa (e.g. Chabou et al., 2007, 2010), SW Europe (Callegaro et al., 2014) and even Sicily M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 173 (Cirrincione et al., 2014), and East America, including Bolivia (Bertrand et al., 2014). This province, even tholeiitic, comprises abundant high-Ti basalts related to an asthenospheric mantle source (Deckart et al., 2005). We thus suggest that the Arrikine sill, and associated magmatism, emplaced due to this continental breakup that also generated the Caledonian unconformity. The mid-Carboniferous reactivation with an age at 325.6 ± 7.7 Ma as determined by the KeAr method is easily related to the late Variscan episode that resulted from the collision between Gondwana, especially NW Africa, and southern Europe from 330 Ma to early Permian (Schulmann et al., 2014). The fact that this important event generated no magmas (except the aplitic metasomatism) can be related to the transpressive character of this reactivation, unable to allow magmas to rise if no major horizontal movements occurred, as it is the case in an intraplate setting. To sum up, we propose that the Arrikine sill was emplaced along a major lithospheric weakness zone, the margin of the Murzuq craton during the Gondwana-Hun continental breakup. Similarly, the reactivation of the major Raghane shear zone bordering the Saharan metacraton that allowed the emplacement of the Aïr ring complex province (Moreau et al., 1994), can be attributed to this continental break-up. This lithosphere weakness zone was reactivated during the late Variscan orogeny that involved Gondwana. These two events correspond to the two major unconformities with important sedimentary hiatus known in the region, the so-called Caledonian and Hercynian unconformities. 8. Conclusions The newly discovered Arrikine sill is a large magmatic object (pyroxene-rich gabbro) of Lower Devonian age emplaced in Silurian sediments, rather unexpected in far-east Hoggar. Although emplaced in an intraplate environment, its datation was not straightforward: all zircons are older than the country rocks and are thus inherited (most zircons are at c. 2030 Ma with additional ages at c. 2700 Ma and in the 766-598 Ma age range) and its rejuvenated KeAr age of 325.6 ± 7.7 Ma relates to a remobilization event due to a K-rich aplitic fluid that reset the isotopic system and that is recorded by secondary minerals (K-feldspar, acid plagioclase, quartz). By contrast, the paleomagnetic pole is reliable and allows dating of the Arrikine sill emplacement in the 400e415 Ma age interval (mid-Lochkhovian-mid-Emsian). Geochemically, the Arrikine gabbro has the composition of a high-Ti alkali basalt, actually a high-Ti hawaiite (sodic trachybasalt), with 46% SiO2, 4% MgO, %Na2O-2>%K2O, absence of chromite and 5% TiO2. As a whole, this chemistry is close to that of the Mauna Loa volcano in Hawaii (Wanless et al., 2006), in agreement with an enriched mantle origin needed for such a high-Ti/low-Si composition. This peculiar chemistry is similar to several intraplate continental provinces such as the Trias rift-related from SW China, the high-Ti basalts from the Emeishan LIP (Zhang et al., 2013), the Ordovician Suordakh and Cambrian Kharaulakh mafic rocks from Siberia (Khudoley et al., 2013) or the 1.5 Ga Yangtse mafic province (Fan et al., 2013). It is also similar to some lithologies belonging to the lower Devonian Aïr ring complexes province in Niger, such as the Abontorok ring complex (Brown et al., 1989). The peculiarity of these compositions supports so a lower Devonian age for the Arrikine sill, as is proposed in the paleomagnetic section. The Arrikine gabbro composition points to a liquidus temperature of 1280
C and a temperature of 1080
C for the magma when it crystallized close to the surface. Geodynamically, the Arrikine sill emplaced in Silurian sediments belonging to the western boundary of the Murzuq craton, actually its metacratonic margin (Fezaa et al., 2010; Liegeois et al.,
2013). Its emplacement is related to the reactivation of this metacratonic margin due to a low stress that propagated in this remote intracontinental region. This was a consequence of “Caledonian” events that also generated in the area the so-called Caledonian unconformity, witness of the absence of lower Devonian sediment deposition (e.g. Ghienne et al., 2013). A link with the magmatic rocks at the origin of the sand injections at the Murzuq craton margins (Moreau et al., 2012) is suggested. The Aïr ring complexes were emplaced at the same period, also along a major lithospheric discontinuity, the Raghane shear zone (Moreau et al., 1994), which is the western boundary of the Saharan metacraton (Abdelsalam et al., 2002). The Mid-Carboniferous rejuvenation dated by the KeAr age (325.6 ± 7.7 Ma) and marked by a aplitic fluid phase can be correlated with the erasure of the primary magnetization in the older sedimentary rocks of the Saharan platform (Henry et al., 2004). It corresponds to the second main Phanerozoic tectonic event in the area marked by the so-called Hercynian unconformity, responsible for important transpressive vertical movements such as the uplift of the Tiririne High (NW Murzuq basin). Emplacement (c. 410 Ma) and rejuvenation (c. 326 Ma) of the Arrikine sill are thus attributed to intense tectonic reactivations that occurred around the Murzuq craton, able to generate magmas of mantle origin through adiabatic pressure release of uprising asthenosphere along mega-shear zones. The effects of these two events being known all over the Saharan platform, it is concluded that they resulted from major events that occurred at plate boundaries. We suggest that the extensional event in the Arrikine area at 410 Ma is linked to the continental breakup when the Hun terranes separated from the northern Gondwana margin (Stampfli and Borel, 2002; Blackey, 2008 and references therein) and, that the compressional event at c. 325 Ma reflects the far-field effects of the collision between Africa (belonging to Gondwana) and southern Europe (Schulmann et al., 2014). Acknowledgments This project was supported by the Algerian-French PICS cooperation program “Architecture lithospherique et dynamique du
manteau sous le Hoggar”. We are very grateful to the civil and military authorities at Djanet and to the “Office du Parc National du Tassili” - OPNT (now “Office National du Parc Culturel du Tassili N'ajjer” - ONPCTA) for help during the field work. Whole-rock analyses were measured on the Horiba-Jobin-Yvon Ultima 2 ICP-AES at the IUEM, Pole de Spectrom ^ etrie Oc
ean, Brest, France. C
eline
Liorzou is thanked for her essential geochemical contribution. Microprobe analyses were performed using the Ouest Microprobe located at Brest. Jessica Langlade is thanked for the feldspars analyses. We are very grateful to Rob Van der Voo for his detailed and constructive review. References Abdelsalam, M., Liegeois, J.P., Stern, R.J., 2002. The saharan metacraton. J. Afr. Earth
Sci. 34, 119e136. Amenna, M., Derder, M.E.M., Henry, B., Maouche, S., Bayou, B., Bouabdallah, H., Ayache, M., Bediaf, M., 2014. Improved Moscovian part of the Gondwana APWP for paleocontinental reconstructions, obtained from a first paleomagnetic pole, age-constrained by a fold test, from in Ezzane area in the Murzuq basin (Algeria, stable Africa). J. Afr. Earth Sci. 99, 342e352. http://dx.doi.org/10.1016/ j.jafrearsci.2013.12.006. As, J.A., Zijderveld, J.D.A., 1958. Magnetic cleaning of rocks in paleomagnetic research. Geophys. J. R. Astr. Soc. 1, 308e319. Bardintzeff, J.-M., Deniel, C., Guillou, H., Platevoet, B., Telouk, P., Oun, K.M., 2011. Miocene to recent alkaline volcanism between Al Haruj and Wawan Namous (southern Libya). Int. J. Earth Sci. 101, 1047e1063. Beccaluva, L., Azzouni-Sekkal, A., Benhallou, A., Bianchini, G., Ellam, R.M., Marzola, M., Siena, F., Stuart, F.M., 2007. Intracratonic asthenosperic upwelling and lithosphere rejuvenation beneath the Hoggar swell (Algeria): evidence from HIMU metasomatised lherzolite mantle xenoliths. Earth Planet. Sci. Lett. 174 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 260, 482e494. Bellini, E., Massa, D., 1980. Stratigraphic contribution to the palaeozoic of Southern basins of Libya. In: Salem, M.J., Busrewil, M.T. (Eds.), Geology of Libya (I). Academic Press., London, pp. 2e56. Bellon, H., Quoc Buu, N., Chaumont, J., Philippet, J.C., 1981. Implantation ionique d’argon dans une cible support : application au traçage isotopique de l’argon contenu dans les mineraux et les roches. Comptes Rendus l
’Academie Sci. Paris
II 292, 977e980. Bertrand, H., Fornari, M., Marzoli, A., García-Duarte, R., Sempere, T., 2014. The Central Atlantic magmatic province extends into Bolivia. Lithos 188, 33e43. Black, R., Jaujou, M., Pellaton, C., 1967. Notice explicative sur la carte geologique de
l’Air a l ’echelle du l/500 000
eme. Dir. Mines Geol., Niamey . Blackey, R.C., 2008. Gondwana paleogeography from assembly to breakup d A 500 m.y. odyssey. In: Fielding, C.R., Frank, T.D., Isbell, J.L. (Eds.), Resolving the Late Paleozoic Ice Age in Time and Space, pp. 1e28. Geological Society of America Special Paper 441. Boehnke, P., Watson, E.B., Trail, D., Harrison, T.M., Schmitt, A.K., 2013. Zircon saturation re-revisited. Chem. Geol. 351, 324e334. Bonin, B., Ouabadi, A., Liegeois, J.P., Bruguier, O., Passchier, C., Guillou, H.,
Abdallah, N., Fezaa, N., Laouar, R., Boissonnas, J., 2011. Possible impact craters within the in ezzane Quaternary volcanic province, eastern hoggar, Algeria. In: The Second Arab Impact Cratering and Astrogeology Conference (Aicac II). Morocco, Casablanca. Bosch, D., Bruguier, O., Pidgeon, R.T., 1996. Evolution of an archean Metamorphic Belt: a conventional and SHRIMP U-Pb study of accessory minerals from the Jimperding Metamorphic Belt, Yilgarn craton, west Australia. J. Geol. 104, 695e711. Bosch, D., Garrido, C.J., Bruguier, O., Dhuime, B., Bodinier, J.L., Padron-Navarta, J.A., Galland, B., 2011. Building an island-arc crustal section : time constraints from a LA-ICP-MS zircon study. Earth Planet. Sci. Lett. 309, 268e279. Botcharnikov, R.E., Almeev, R.R., Koepke, J., Holtz, F., 2008. Phase relations and liquid lines of descent in hydrous ferrobasalt - implications for the Skaergaard Intrusion and Columbia River Flood Basalts. J. Petrology 49, 1687e1727. Bouzid, A., Bayou, B., Liegeois, J.-P., Bourouis, S., Bougchiche, S.S., Bendekken, A.,
Abtout, A., Boukhlouf, W., Ouabadi, A., 2015. Lithospheric structure of the Atakor metacratonic volcanic swell (Hoggar, Tuareg Shield, southern Algeria): Electrical constraints from magnetotelluric data. In: Foulger, G.R., Lustrino, M., King, S.D. (Eds.), The Interdisciplinary Earth: a Volume in Honor of Don L. Geological Society of America, Anderson, pp. 239e255. Special Paper 514. Brown, W.L., Moreau, C., Demaiffe, D., 1989. An anorthosite suite in a ring-complex: crystallization and emplacement of an anorogenic type from abontorok, Aïr, Niger. J. Petrol. 30, 1501e1540. Callegaro, S., Rapaille, R., Marzoli, A., Bertrand, H., Chiaradia, M., Reisberg, L., Bellieni, G., Martins, L., Madeira, J., Mata, J., Youbi, N., De Min, A., Azevedo, M.R., Bensalah, M.K., 2014. Enriched mantle source for the Central Atlantic magmatic province: new supporting evidence from Southwestern Europe. Lithos 188, 15e32. Chabou, M.C., Sebai, A., Feraud, G., Bertrand, H., 2007. Datation 40Ar/39Ar de la
province magmatique de l’Atlantique central dans le Sud-Ouest algerien. C. R.
Geosci. 339, 970e978. Chabou, M.C., Bertrand, H., Sebai, A., 2010. Geochemistry of the Central Atlantic magmatic province (CAMP) in south-western Algeria. J. Afr. Earth Sci. 58, 211e219. Chevrel, M.O., Baratoux, D., Hess, K.-U., Dingwell, D.B., 2014. Viscous flow behavior of tholeiitic and alkaline Fe-rich Martian basalts. Geochimica Cosmochimica Acta 124, 348e365. Cirrincione, R., Fiannacca, P., Lustrino, M., Romano, V., Tranchina, A., 2014. Late Triassic tholeiitic magmatisminWestern Sicily: a possible extension of the Central Atlantic Magmatic Province (CAMP) in the Central Mediterranean area? Lithos 188, 60e71. Clark, D.A., 1996. Palaeomagnetism of the Mount Leyshon Intrusive Complex, the Tuckers Igneous Complex and the Ravenwood Batholith. CSIRO Australian Exploration Mining Report, 318R. Cohen, K.M., Finney, S.C., Gibbard, P.L., Fan, J.X., 2013. The ICS international chronostratigraphic chart. Episodes 36, 199e204. Cox, A., Dalrymple, G.B., 1967. Statistical analysis of geomagnetic reversal data and the precision of potassium-argon dating. J. Geophys. Res. 72, 2603e2614. Davidson, L., Beswetherick, S., Craig, J., Eales, M., Fisher, A., Himmali, A., Jho, J., Mejrab, B., Smart, J., 2000. The structure, stratigraphy and petroleum geology of the Murzuq Basin, southwest Libya. In: Sola, M.A., Worsley, D. (Eds.), Geological Exploration in Murzuq Basin. Elsevier, Amsterdam, pp. 295e320. Day, R., Fuller, M., Schmidt, V.A., 1977. Hysteresis properties of titanomagnetites: grain size and compositional dependence. Phys. Earth Planet. Interiors 13, 260e267. Deckart, K., Bertrand, H., Liegeois, J.P., 2005. Geochemistry and Sr, Nd, Pb isotopic
composition of the Central Atlantic magmatic province (CAMP) in Guyana and Guinea. Lithos 82, 282e314. Demaiffe, D., Moreau, C., Brown, W.L., Weis, D., 1991. Geochemical and isotopic (Sr, Nd and Pb) evidence on the origin of the anorthosite-bearing anorogenic complexes of the Aïr Province, Niger. Earth Planet. Sci. Lett. 105, 28e46. Derder, M.E.M., Henry, B., Bayou, B., Ouabadi, A., Bellon, H., Djellit, H., Khaldi, A., Amenna, M., Baziz, K., Hemmi, A., Guemache, M.A., 2006. New African Lower Carboniferous paleomagnetic pole from intrusive rocks of the Tin Serririne basin (Southern border of the Hoggar, Algeria). Tectonophysics 418, 189e203. Derder, M.E.M., Henry, B., Amenna, M., Bayou, B., Djellit, H., Guemache, M.A., Hemmi, A., 2009. New structural implications for central Sahara (Algeria) from revisited Upper Carboniferous “Hassi Bachir” formation: paleomagnetic constraints. Tectonophysics 463, 69e76. http://dx.doi.org/10.1016/ j.tecto.2008.09.012. Djellit, H., Bellon, H., Ouabadi, A., Derder, M.E.M., Henry, B., Bayou, B., Khaldi, A., Baziz, K., Merahi, M.K., 2006. Age 40K/40Ar, Carbonifere inf erieur, du magma-
tisme basique filonien du synclinal paleozoïque de Tin Serririne, sud-est Hoggar
(Algerie). Comptes Rendus Geosci. 338, 624
e631. Domeier, M., Van der Voo, R., Torsvik, T.H., 2012. Paleomagnetism and Pangea: the road to reconciliation. Tectonophysics 514e517, 14e43. Echikh, K., Sola, M.A., 2000. Geology and Hydrocarbon occurrences in the murzuq Basin, SW Libya. In: Sola, M.A., Worsley, D. (Eds.), Geological Exploration in Murzuq Basin. Elsevier, Amsterdam, pp. 175e222. Fan, H.-P., Zhu, W.-G., Li, Z.-X., Zhong, H., Bai, Z.J., He, D.F., Chen, C.-J., Cao, C.-Y., 2013. Ca. 1.5Ga mafic magmatism in South China during the break-up of the supercontinent Nuna/Columbia: the Zhuqing Fe-Ti-V oxide ore-bearing mafic intrusions in western Yangtze Block. Lithos 168e169, 85e98. Fezaa, N., Liegeois, J.P., Abdallah, N., Bruguier, O., Laouar, R., Ouabadi, A., 2013.
Origine du groupe metas
edimentaire de Djanet (Hoggar oriental, Alg
erie).
Geochronologie et g
eochimie. Bull. Du. Serv. G
eologique Natl. 24, 3
e26. Fezaa, N., Liegeois, J.-P., Abdallah, N., Cherfouh, E.H., De Waele, B., Bruguier, O.,
Ouabadi, A., 2010. The Djanet terrane (Eastern Hoggar, Algeria), the Pan-African metacratonic boundary of the Murzuq craton: field, detrital and magmatic U-Pb zircon and Sr-Nd isotopes evidences. Precambrian Res. 180, 299e327. Fisher, R.A., 1953. Dispersion on a sphere. Proc. R. Soc. Lond. A 217, 295e305. Foulger, G., Hamilton, W.R., 2014. Earth science: plume hypothesis challenged. Nature 505, 618. Freulon, J.M., 1964. Etudes geologiques des s
eries primaires du Sahara Central
(Tassili n'Ajjer et Fezzan). In: Publication du Centre de Recherches sur les Zones Arides. CNRS, Paris. Geologie, 3
. Frohlich, S., Petitpierre, L., Redfern, J., Grech, P., Bodin, S., Lang, S., 2010. Sedimen- € tological and sequence stratigraphic analysis of Carboniferous deposits in western Libya: Recording the sedimentary response of the northern Gondwana margin to climate and sea-level changes. J. Afr. Earth Sci. 57, 279e296. Galeazzi, S., Point, O., Haddadi, N., Mather, J., Druesne, D., 2010. Regional geology and petroleum systems of the IllizieBerkine area of the Algerian Saharan Platform: an overview. Mar. Petroleum Geol. 27, 143e178. Geological Map of Libya, 1985. SW Sheet. Compiled by Industrial Research Centre. Geological Research and Min. Department (Tripoli, Libya). Ghienne, J.F., Moreau, J., Degermann, L., Rubino, J.L., 2013. Lower Palaeozoic unconformities in an intracratonic platform setting: glacial erosion versus tectonics in the eastern Murzuq Basin (southern Libya). Int. J. Earth Sci. 102, 455e482. Guiraud, R., Bosworth, W., Thierry, J., Delplanque, J.A., 2005. 2005. Phanerozoic geological evolution of Northern and Central Africa: an overview. J. Afr. Earth Sci. 43, 83e143. Haddoum, H., Guiraud, R., Moussine-Pouchkine, A., 2001. Hercynian compressional deformations of the Ahnet-Mouydir Basin, Algerian Saharan Platform: far-field stress effects of the Late Palaeozoic orogeny. Terra Nova 13, 220e226. Hargraves, R.B., Dawson, E.M., Houten, F.B., 1987. Palaeomagnetism and age of midPalaeozoic ring complexes in Niger, West Africa, and tectonic implications. Geophys. J. R. Astronomical Soc. 90, 705e729. http://dx.doi.org/10.1111/j.1365- 246X.1987.tb00750.x. Heine, C., Müller, R.D., Steinberger, B., Torsvik, T.H., 2008. Subsidence in intracontinental basins due to dynamic topography. Phys. Earth Planet. Interiors 171, 252e264. Henry, B., Merabet, N., Derder, M.E.M., Bayou, B., 2004. Chemical remagnetizations in the Illizi basin (Saharan craton, Algeria). Geophys. J. Int. 156, 200e212. Henry, B., Liegeois, J.P., Nouar, O., Derder, M.E.M., Bayou, B., Bruguier, O., Ouabadi, A.,
Belhai, D., Amenna, M., Hemmi, A., Ayache, M., 2009. Repeated granitoid intrusions during the Neoproterozoic along the western boundary of the Saharan metacraton, Eastern Hoggar, Tuareg shield, Algeria: an AMS and UePb zircon age study. Tectonophysics 474, 417e434. Henry, B., Derder, M.E.M., Amenna, M., Maouche, S., Bayou, B., Ouabadi, A., Bouabdallah, H., Ayache, M., Beddiaf, M., Bestandji, R., 2014. Paleomagnetic dating of continental geological formations: Strong diachronism evidenced in the Saharan platform and geodynamical implications. J. Afr. Earth Sci. 99, 353e362. http://dx.doi.org/10.1016/j.jafrearsci.2014.02.010. Henry, B., Derder, M.E.M., Amenna, M., Maouche, S., Bayou, B., 2016. Better constrained selection of the Paleozoic West Gondwana (South America) paleomagnetic poles for the APWP's determination. J. S. Am. Earth Sci. submitted. Husch, J.M., Moreau, C., 1982. Geology and major element geochemistry of anorthositic rocks associated with paleozoic hypabyssal ring complexes, air massif, Niger, West Africa. J. Volcanol. Geotherm. Res. 14, 47e66. Kirschvink, J.L., 1980. The least-squares line and plane and the analysis of palaeomagnetic data. Geophys. J. R. astronomical Soc. 62, 699e718. Karche, J.P., Moreau, C., 1977. Note preliminaire sur le massif subvolcanique
a structure annulaire d'Abontorok (Aïr-Niger). C. R. Acad. Sci. Paris. 284, 1259e1262. Khudoley, A.K., Prokopiev, A.V., Chamberlain, K.R., Ernst, R.E., Jowitt, S.M., Malyshev, S.V., Zaitsev, A.I., Kropachev, A.P., Koroleva, O.V., 2013. Early Paleozoic mafic magmatic events on the eastern margin of the Siberian Craton. Lithos 174, 44e56. LeBas, M.J., LeMaitre, R.W., Streckeisen, A., Zanettin, B., 1986. A chemical classifi- cation of volcanic rocks based on the total alkaliesilica diagram. J. Petrology 27, M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176 175 745e750. Le Caignec, J., Harel, M., Didier, J., Illy, P., 1957. Reports of the BRMA (Bureau de Recherches Minieres de l'Alg erie)for Djanet area, Mission Illy (Hoggar 1956-
1957). Feuilles au 1/200 000 : Fort Charlet et Tin Alkoum. Legrand, P., 2000. Une region de r
ef
erence pour la limite Ordovicien
eSilurien : l'Oued in Djerane, Sahara algerien. Comptes Rendus de l'Acad
emie des Sci. - Ser.
IIA - Earth Planet. Sci. 330, 61e66. Leger, J.M., 1985. G
eologie et
evolution magmatique du complexe plutonique
d'Iskou (Aïr, Niger). J. Afr. Earth Sci. 3, 89e96. Lelubre, M., Freulon, J.M., Lefranc, P., 1962. Carte geologique du Nord-Ouest de
l'Afrique, feuille Sahara central, 1/2,000,000eme. CNRS, France . Le Maitre, R.W., Streckeisen, A., Zanettin, B., Le Bas, M.J., Bonin, B., Bateman, P., Bellieni, G., Dudek, A., Efremova, S., Keller, J., Lameyre, J., Sabine, P.A., Schmid, R., Sorensen, H., Woolley, A.R., 2002. Igneous Rocks: a Classification and Glossary of Terms: Recommendations of the International Union of Geological Sciences. Subcommission on the Systematics of Igneous Rocks. Cambridge University Press, 236 pp. Linnemann, U., Ouzegane, K., Drareni, A., Hofmann, M., Becker, S., G€ artner, A., Sagawe, A., 2011. Sands of West Gondwana: an archive of secular magmatism and plate interactions. A case study from the Cambro-Ordovician section of the Tassili Ouan Ahaggar (Algerian Sahara) using UePbeLA-ICP-MS detrital zircon ages. Lithos 123, 188e203. Liegeois, J.P., Black, R., Navez, J., Latouche, L., 1994. Early and late Pan-African
orogenies in the Aïr assembly of terranes (Tuareg shield, Niger). Precambrian Res. 67, 59e88. Liegeois, J.P., Latouche, L., Boughrara, M., Navez, J., Guiraud, M., 2003. The LATEA
metacraton (Central Hoggar, Tuareg shield, Algeria): behaviour of an old passive margin during the Pan-African orogeny. J. Afr. Earth Sci. 37, 161e190. Liegeois, J.P., Benhallou, A., Azzouni-Sekkal, A., Yahiaoui, R., Bonin, B., 2005. The
Hoggar swell and volcanism: reactivation of the Precambrian Tuareg shield during Alpine convergence and West African Cenozoic volcanism. In: Foulger, G.R., Natland, J.H., Presnall, D.C., Anderson, D.L. (Eds.), Plates, Plumes and Paradigms: Geological Society of America Special Paper, 388, pp. 379e400. Liegeois, J.P., Abdelsalam, M.G., Ennih, N., Ouabadi, A., 2013. Metacraton: nature,
genesis and behaviour. Gondwana Res. 23, 220e237. Lustrino, M., Cucciniello, C., Melluso, L., Tassinari, C.C.G., De Gennaro, R., Serracino, M., 2012. Petrogenesis of Cenozoic volcanic rocks in the NWsector of the Gharyan volcanic field, Libya. Lithos 155, 218e235. MacDonald, G.A., Katsura, T., 1964. Chemical compositions of Hawaiian lavas. J. Petrology 5, 82e133. Marzoli, A., Renne, P.R., Piccirillo, E.M., Ernesto, M., Bellieni, G., DeMin, A., 1999. Extensive 200-million-year-old continental flood basalts of the Central Atlantic Magmatic Province. Science 284, 616e618. Massa, D., Coquel, R., Doubinger, R., 1988. New Palynological data on the Visean/ Moscovian carboniferous interval in the Ghadames Basin (Libya). Comparison with saharan Basins. Assesment of the Gondwanian and Euramerican In- fluences. Oil & Gas science Technology. Rev. l’Institut Français Petrole 43, 3
e16. Michard, A., Hoepffner, C., Soulaimani, A., Baidder, L., 2008. The variscan Belt. In: Michard, A., et al. (Eds.), Continental Evolution: the Geology of Morocco, Lecture Notes in Earth Sciences, vol. 116, pp. 65e131. Moreau, C., Demaiffe, D., Bellion, Y., Boullier, A.M., 1994. A tectonic model for the location of Paleozoic ring-complexes in Aïr (Niger, West Africa). Tectonophysics 234, 129e146. Moreau, J., Ghienne, J.F., Hurst, A., 2012. Kilometre-scale sand injectites in the intracratonic Murzuq Basin (south-west Libya): an igneous trigger? Sedimentology 59, 1321e1344. Morton, A.C., Meinhold, G., Richard, J.P., Phillips, J., Strogen, D., Abutarruma, Y., Elgadry, M., Thusu, B., Andrew, G., 2011. A heavy mineral study of sandstones from the eastern Murzuq Basin, Libya: constraints on provenance and stratigraphic correlation. J. Afr. Earth Sci. 61, 308e330. Nelson, S.A., Carmichael, I.S.E., 1984. Pleistocene to recent alkalic volcanism in the region of Sanganguey volcano, Nayarit, Mexico. Contributions Mineralogy Petrology 85, 321e335. Nouar, O., Henry, B., Liegeois, J.P., Derder, M.E.M., Bayou, B., Bruguier, O., Ouabadi, A.,
Amenna, M., Hemmi, A., Ayache, M., 2011. Eburnean and Pan-African granitoids and the Raghane mega-shear zone evolution: Image analysis, U-Pb zircon age and AMS study in the Arokam Ten
er
e (Tuareg shield, Algeria). J. Afr. Earth Sci.
60, 133e152. Pearce, J., 1982. Role of the sub-continental lithosphere in magma genesis at active continental margins. In: Hawkesworth, C.J., Norry, M.J. (Eds.), Continental basalts and Mantle Xenoliths. Shiva Geology Series, Nantwich, pp. 230e249. Radivojevi
c, M., Tolji
c, M., Turki, S.M., Boji
c, Z.,  Sari
c, K., Cvetkovi
c, V., 2015. Neogene to Quaternary basalts of the Jabal Eghei (Nuqay) area (south Libya): two distinct volcanic events or continuous volcanism with gradual shift in magma composition? J. Volcanol. Geotherm. Res. 293, 57e74. Rapalini, A.E., Vilas, J.F., 1991. Preliminary paleomagnetic data from the Sierra Grande Formation: tectonic consequences of the first mid-Paleozoic paleopoles from Patagonia. J. S. Am. Earth Sci. 4, 25e41. Reimold, W.E., Koeberl, C., 2014. Impact structures in Africa: a review. J. Afr. Earth Sci. 93, 57e175. Ritzmann, O., Faleide, J.I., 2009. The crust and mantle lithosphere in the Barents Sea/ Kara Sea region. Tectonophysics 470, 89e104. Rogers, N.W., James, D., Kelley, S.P., de Mulder, M., 1998. The generation of potassic lavas from the eastern Virunga Province, Rwanda. J. Petrol. 39, 1223e1247. Schmidt, P.W., Embleton, B.J.J., Palmer, H.C., 1987. Pre- and post-folding magnetizations from the early Devonian Snowy River Volcanics and Buchan Caves Limestone. Vic. Geophys. J. Int. 91, 155e170. Schulmann, K., Martinez Catalan, J.R., Lardeaux, J.M., Janousek, V., Oggiano, G., 2014. The Variscan Orogeny: Extent, Timescale and the Formation of the European Crust. Geological Society, London, pp. 1e6. Special Publications, 405. Smith, B., Derder, M.E.M., Henry, B., Bayou, B., Yelles, A.K., Djellit, H., Amenna, M., Garces, M., Beamud, E., Callot, J.P., Eschard, R., Chambers, A., Aifa, T., Ait Ouali, R., Gandriche, H., 2006. Relative importance of the Hercynian and post-Jurassic tectonic phases in the Saharan platform: a palaeomagnetic study of Jurassic sills in the Reggane Basin (Algeria). Geophys. J. Int. 167, 380e396. Stampfli, G.M., Borel, G.D., 2002. A plate tectonic model for the Paleozoic and Mesozoic constrained by dynamic plate boundaries and restored synthetic oceanic isochrons. Earth Planet. Sci. Lett. 196, 17e33. Steiger, R.H., Jager, E., 1977. Subcommission on geochronology: convention on the € use of decay constants in geo- and cosmo-chronology. Earth Planet. Sci. Lett. 36, 359e362. Sun, S.S., 1980. Lead isotopic study of young volcanic rocks from mid-ocean ridges, ocean islands and islands arcs. Philos. Trans. R. Soc. Lond. A297, 409e445. Sun, S.S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts: implications for mantle compositon and processes. In: Saunders, A.D., Norry, M.J. (Eds.), Magmatism in Ocean Basins. Geological Society of London, pp. 313e345. Special Publication 42. Taylor, S.R., Mc Lennan, S.M., 1985. The Continental Crust: its Composition and Evolution. Blakwell, Oxford, 312 pp. Torsvik, T.H., van der Voo, R., Preeden, U., Mac Niocaill, C., Steinberger, B., Doubrovine, P.V., van Hinsbergen, D.J.J., Domeier, M., Gaina, C., Tohver, E., Meert, J.G., McCausland, P.J.A., Cocks, L.R.M., 2012. Phanerozoic polar wander, paleogeography and dynamics. Earth-Science Rev. 114, 325e368. http:// dx.doi.org/10.1016/j.earscirev.2012.06.007. Wanless, V.D., Garcia, M.O., Rhodes, J.M., Weis, D., Norman, M.D., 2006. Shield-stage alkalic volcanism on mauna loa volcano, Hawaii. J. Volcanol. Geotherm. Res. 151, 141e155. Watson, E.B., Harrison, T.M., 1983. Zircon saturation revisited: temperature and composition effects in a variety of crustal magma types. Earth Planet. Sci. Lett. 64, 295e304. Wiedenbeck, M., Alle, P., Corfu, F., Griffin, W.L., Meier, M., Oberli, F., von Quadt, A., Roddick, J.C., Spiegel, W., 1995. Three natural zircon standards for U-Th-Pb, LuHf, trace element and REE analyses. Geostand. Newsl. 19, 1e23. Wilson, R.L., Everitt, C.W.F., 1963. Thermal demagnetization of some Carboniferous lavas for paleomagnetic purposes. Geophys. J. R. Astronomical Soc. 8, 149e164. Winchester, J.A., Floyd, P.A., 1977. Geochemical discrimination of different magma series and their differentiation products using immobile elements. Chem. Geol. 20, 325e343. Yahiaoui, R., Dautria, J.-M., Allard, O., Bosch, D., Azzouni-Sekkal, A., Bodinier, J.-L., 2014. A volcanic district between the Hoggar uplift and the Tenere Rifts: Volcanology, geochemistry and age of the In-Ezzane lavas (Algerian Sahara). J. Afr. Earth Sci. 92, 14e20. Zhang, J., Huang, Z., Luo, T., Qian, Z., Zhang, Y., 2013. Origin of early Triassic rift related alkaline basalts from Southwest China: age, isotope, and trace-element constraints. Int. Geol. Rev. 55, 1162e1178. Zijderveld, J.D.A., 1967. AC demagnetisation of rocks: analysis of results. In: Collinson, D.W., Creer, K.M., Runcorn, S.K. (Eds.), Method in Paleomagnetism. Elsevier, Amsterdam, pp. 254e286. 176 M.E.M. Derder et al. / Journal of African Earth Sciences 115 (2016) 159e176