SPE418 16atakor

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/281133279
Tertiary alkaline volcanism of the Atakor Massif (Hoggar, Algeria): ﬁeld,

petrological, mineralogical, geochemical, and geodynamic features
Article · January 2007
CITATIONS READS
4 404
5 authors, including:
Bernard Bonin Amel Benhallou

Université Paris-Sud 11 Center for Research in Astronomy and Astrophysics Geophysics
195 PUBLICATIONS 9,036 CITATIONS 49 PUBLICATIONS 340 CITATIONS
SEE PROFILE SEE PROFILE
R. Yahiaoui J.-P. Liégeois

École Normale Supérieure de Kouba Royal Museum for Central Africa
10 PUBLICATIONS 260 CITATIONS 229 PUBLICATIONS 11,590 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Mo-W-Bi Deposits in Anorogenic granites of Northcentral Nigeria: Insights from isotopic data, fluid inclusions and Re-Os geochrons View project
is the silet terrane a puzzle View project
All content following this page was uploaded by Bernard Bonin on 09 January 2018.
The user has requested enhancement of the downloaded file.

Geological Society of America
Special Paper 418
2007
Tertiary alkaline volcanism of the Atakor massif, Hoggar, Algeria

Abla Azzouni-Sekkal*
Faculté des Sciences de la Terre, de l’Aménagement du Territoire et de la Géographie, Université des
Sciences et Techniques Houari Boumediene, B.P. 32, El Alia, 16111 Bab Ezzouar, Algiers, Algeria
Bernard Bonin*
UMR 8148 “IDES,” Département des Sciences de la Terre, Université de Paris-Sud, F-91405 Orsay Cedex, France
Amel Benhallou*
Centre de Recherches en Astronomie, Astrophysique, et Géophysique, Bouzaréah, Algeria
Rachid Yahiaoui
École Normale Supérieure, Kouba, Algiers, Algeria
Jean-Paul Liégeois*
Isotope Geology, Africa Museum, B-3080 Tervuren, Belgium
ABSTRACT
The Atakor massif is a part of the Hoggar volcanic province, which was emplaced
on top of a basement swell initiated during the Cretaceous. There have been three
main episodes of volcanic activity since the Miocene, separated by long periods of qui-
escence. The lava flows and domes were emitted along lithosphere-scale fault zones.
With its famous scenery, the Atakor massif is one of the largest (2150 km2) volcanic
districts of the province. Mafic volcanic rocks are abundant in the center of the mas-
sif, but become scarce to the south, where only few scarps are observed. Phonolites
occur only in the Assekrem area, whereas trachytes occur everywhere, with a marked
enrichment in quartz to the south and the southeast (Tahifet area), where rhyolites are
also exposed. Two magmatic groups have been identified based on field and petrologi-
cal observations. The mafic group has a basanite-phonotephrite association, forming
uplifted plateaus, scoria cones, and valley-filling lava flows. The presence of mantle-
derived amphibole ± biotite megacrysts and peridotite mantle xenoliths together with
the nonprimary chemical compositions of the magmatic rocks suggest that magmatic
differentiation may have occurred within the upper mantle. The felsic group is com-
posed of two diverging trends, a silica-saturated benmoreite-trachyte-rhyolite trend
and a silica-undersaturated trachyte-phonolite trend. The primary magmas are con-
sidered to have been produced as a consequence of lithospheric mantle delamination
along linear megashear zones inducing low degrees of decompression partial melting
at variable depths (110–40 km) in the upwelling asthenosphere. The discrete volcanic
episodes correspond to periods of reactivation of the major fault zones in response to
discrete Neogene extensional tectonic events associated with Alpine orogenesis in the
Western Mediterranean region induced by Africa-Eurasia collision.
Keywords: Hoggar (Algeria), volcanism, alkali basalt, phonolite-trachyte, upper mantle.
*E-mails: Azzouni-Sekkal: azzouniabla@yahoo.fr; Bonin (corresponding author): bbonin@geol.u-psud.fr; Benhallou: zoulema@yahoo.com; Liégeois: jean-paul.
liegeois@africamuseum.be.
Azzouni-Sekkal, A., Bonin, B., Benhallou, A., Yahiaoui, R., and Liégeois, J.-P., 2007, Tertiary alkaline volcanism of the Atakor massif, Hoggar, Algeria, in Bec-
caluva, L., Bianchini, G., and Wilson, M., eds., Cenozoic Volcanism in the Mediterranean Area: Geological Society of America Special Paper 418, p. 321–340, doi:
10.1130/2007.2418(16). For permission to copy, contact editing@geosociety.org. ©2007 Geological Society of America. All rights reserved.
321
322 Azzouni-Sekkal et al.
INTRODUCTION in the central part of the massif (Yahiaoui, 2003) and the volcanic
edifices near Tamanrasset in the southwest were previously studied
The Mediterranean Sea is the product of a complex series of by Girod (1971). Representative samples, collected on the basis of
plate tectonic processes. It corresponds broadly to an extensional apparent lack of alteration in the field were analyzed for their bulk-
tectonic system overlying the convergence zone between the rock and mineral chemistry. Bulk-rock sample powders were pre-
European continental plate to the north and the African continen- pared at Centre de Recherches Pétrographiques et Géochimiques
tal plate and related microplates to the south (e.g., Jolivet and Fac- (CRPG), Nancy (France), followed by fusion with LiBO2 and
cenna, 2000; Rosenbaum et al., 2002; Frizon de Lamotte, 2006). HNO3 dissolution, and analysis by inductively coupled plasma–
Oceanic domains emplaced in between continental blocks are electron spectrometry (ICP-ES; major and minor elements) and
exposed as small ophiolite slices, or have been consumed through inductively coupled plasma–mass spectrometry (ICP-MS; trace
subduction beneath the Italian peninsula and the Aegean Sea, as elements, with analytical uncertainties ranging from 5% to 20%,
well as along the Rif-Tell belts in North Africa. The present-day according to the element and the concentration), with quality con-
tectonic settings include both continent-continent collisional and trol using international geostandards. Representative whole-rock
postcollisional regimes (Jolivet et al., 1999). Numerous mag- compositions are listed in Tables 1 and 2. Minerals were analyzed
matic provinces developed along seismically active tectonic lin- using the CAMEBAX electron microprobe at the Université de
eaments within the Alps and Apennine fold belts (Lustrino, 2000; Paris–Sud, Orsay, using routine techniques, i.e., natural minerals
Savelli, 2002 and references therein), as well as further inland as standards, 15 kV acceleration voltage, 12 nA beam intensity,
within the adjacent continents of Africa (Black and Girod, 1970) and 10 s counting times.
and Europe (Wilson and Downes, 1991).
The Cenozoic volcanic province of Hoggar (also named GEOLOGICAL SETTING OF THE MAGMATISM
Ahaggar in the literature) is located within the African plate
~1200 km south of the Mediterranean coast (Fig. 1). It is com- The volcanic formations of the Atakor massif are exposed
posed of several massifs of contrasting ages and eruptive styles, at high altitudes, often more than 2000 m above sea level (asl).
covering a total area of ~11,700 km2, with an eruptive volume The highest peak is the 2918-m-high Tahat phonolite dome. The
of ~1650–2500 km3. The volcanic formations rest upon uplifted, basement constitutes a large SW-NE–trending dome structure,
weathered, and eroded Archean to Neoproterozoic basement. has altitudes consistently above 1400 m asl, and culminates in a
Magmatic activity began at the Eocene-Oligocene boundary small, 2570-m-high, Precambrian basement inlier exposed south
ca. 34 Ma (Aït Hamou et al., 2000) and has continued until the of Assekrem and a 2710-m-high phonolite lava dome (Girod,
present (Girod, 1971). Associated with a crustal swell, 1000 km in 1971). Contrasting morphologies, which result from recent tec-
long axis, the Hoggar volcanic province is classically considered tonic uplift, erosion of the basement, and contrasting eruptive
to be the product of a mantle plume (Sleep, 1990; Burke, 1996), styles, dominate the landscape (Fig. 3). Outpouring of mafic
although no significant thermal anomaly has been detected, nor magmas produced extensive lava flows that form large, dissected
is uplift limited to the central Hoggar area. Clear links are evident plateaus, or have filled valleys associated with scoria cones and
between Cenozoic volcanism in West Africa and Pan-African necks. More viscous, felsic magmas were extruded as domes and
tectonic lineaments, and also between the timing of volcanism spines associated with and intruding pyroclastic deposits and
and Alpine orogenic events (for a review and discussion, see rare thick lava flows. Detailed field descriptions can be found in
Liégeois et al., 2005). Rognon (1967) and Girod (1971).
The Atakor massif is located right in the center of the Hoggar
swell (Fig. 2). Its volcanic formations cover an area of 2150 km2 Age of Magmatism
with a total volume estimated at ~250 km3; this corresponds to
~10%–15% of the entire Hoggar volcanic province. Because of Radiometric dates for the magmatic rocks of the Atakor mas-
its exceptional position within the swell and the excellent expo- sif are scarce. Coupled with sedimentological constraints, such as
sure, most attention has been focused on the Atakor massif (e.g., the presence of lacustrine travertine and diatomite, geomorpho-
Allègre et al., 1981). New whole-rock major-, trace-element, and logical observations, such as alluvial terraces (Rognon, 1967), and
mineral chemical data are presented here that allow us to recon- paleontological data, such as silicified wood (Rognon et al., 1983),
struct the evolution in space and time of the Atakor Neogene available K/Ar isotopic and 14C data indicate three discrete epi-
volcanic episodes and their geodynamic relationships with the sodes of eruptions, separated by two periods of quiescence:
evolving Alpine orogenic system. 1. The first, and the most voluminous, magmatic episode
occurred during the Miocene from 19.9 ± 1.9 Ma to 12.4 ± 2 Ma,
Sampling and Analytical Techniques i.e., from the Burdigalian to the Serravallian (according to the
most recent geological time scale, Gradstein et al., 2004). Flood
Three districts of the Atakor massif were investigated in this basalts, intercalated with trachytic tuffs (redefined as benmoreite,
study. The Tahifet area in the southeast (Benhallou, 2000) had never see following), form a large plateau. The episode ended with the
been studied before. The Assekrem plateau and nearby formations emplacement of the 2918-m-high Tahat phonolite dome. Other
1000 km
Peri a
dr iat
Az= Azrou ic lin
Ta= Taourirt France e
GB= G rande Kabylie

KH= Kef Hahouner
Ita
H-A= High Atlas
ly
M-A= MiddleAtlas
S-A= SaharanAtlas
A-A= Anti-Atlas
Tam=Tamanrasset
in
Spa
Alger KB
cc o Oran Tunis
oro Tell
Tunisia
MRabat Rif KH Nefza
Az M- A Ta Oranie
Oulmès Rekkame S-A
Tamazert Laghouat
H-A
Tripoli
Sirwa Sahro
A- A Gharyan
Algeria N
Libya
30°
As Sawda
West Al Hasawinah
Hoggar ?
African Egéré Al
In Teria
Haruj
craton n’Ajjer
Djanet
Murzuq
Manzaz
Taoudeni Atakor Anahef
Tam In Ezzane
Tahalra
Djado
Mali 20°
Tin
Zaouatene Aïr Tibesti
Tin Taralle
Todra Chad
Timbuctu Gao Iforas Niger
0° Agadès 10°
Phanerozoic sedimentary Archaean, Palaeoproterozoic
cover and Neoproterozoic terranes
Cenozoic volcanism affected by the Pan-Africanorogeny
Archaean and Palaeoproterozoic
Atlas mountains West Africancraton and its
sedimentary cover
Major Pan-Africanshear zones Inferred major Pan-African
shear zones
Uplifts structures Mediterranean
shear zones Thrusts
in Libya
Figure 1. Cenozoic volcanism in NW Africa and the Western Mediterranean. Map is

based on Fabre (2005); Black et al. (1967, 1994); Saadi (1982); Savelli (2002); Vai
(2001); Bigi et al. (1989). The volcanic fields in the Tuareg shield were drawn from a
satellite photograph; orthorectified Landsat Thematic Mapper mosaics as compressed
color imagery in MrSIDTM file format are from Lizardtech.
Manzaz
zone
Int
er-
Atakor
shear
ter
ran
eP
frican
Assekrem
an
-Af
Tahifet
Pan-A
rica
ns
he
rrane
ar
zon
Inter-te
e
Cambrian
topaz granite
one
ar z
she
Tamanrasset
ican
20 km
-Afr
Tahalra
Pan
Figure 2. Satellite photograph of the Atakor massif (orthorectified Landsat Thematic Mapper mosaics
as compressed color imagery in MrSIDTM file are from Lizardtech). The Atakor massif is located in
between the Manzaz district to the northeast, and the Tahalra district to the southwest. The lavas appear
dark blue and the scoria cones red or brown due to several factors, including iron oxidation. The
linear eastern boundary corresponds to a Pan-African shear zone within the LATEA metacraton.
high-altitude domes and flows emplaced onto the basaltic pla- the massif. The older scoria cones were again destroyed by ero-
teau, such as the undated Assekrem phonolite dome and related sion and can be clearly distinguished from the morphologically
Zermezoug lava flow, could belong to the same stage. Morpho- younger scoria cones of the most recent volcanic episode. The
logical observations (Girod, 1971) suggest that the igneous epi- original topography was inverted, so that the plateau-forming
sode was not coeval with any large tectonic displacements. and older valley-filling basaltic lava flows are now exposed as
A long period of quiescence of ~6 m.y. took place between dissected cliffs and scarps. The felsic domes, spines, and flows
12.4 ± 2 Ma and 6.7 ± 2 Ma, i.e., during the Tortonian. The older occur on top of hills composed of Precambrian basement capped
scoria cones were eroded, and their remnants, such as the basal- by a thick blanket of pyroclastic deposits. On the basis of the
tic tuffs and ash deposits that are exposed under the Akarakar variations in basement altitudes from 1390 m in the city of Taman-
trachyte complex, are preserved under the deposits of the sub- rasset to 2570 m in the Assekrem inlier, the total amount of uplift
sequent magmatic episodes. The more resistant lava plateau was is estimated at ~1200 m, which corresponds to an average crustal
dissected by water-driven erosion. uplift rate of 0.5 ± 0.1 mm yr–1 during a period of 2.25 m.y.
2. The second magmatic episode began after the Tortonian 3. The final magmatic episode started after the Pliocene
period of quiescence. It spanned the Miocene-Pliocene bound- period of quiescence, from 1.95 ± 0.2 Ma to the present, i.e.,
ary from 6.7 ± 2 Ma to 4.2 ± 0.2 Ma, i.e., from the Messinian from the Gelasian to the Holocene. The erupted products are
to the Zanclean. The erupted products are compositionally more exclusively basanites-tephrites that form scoria cones, necks,
variable, ranging from basalt to phonolite (Girod, 1971; Rognon and valley-filling lava flows, which constitute less than 3% of the
et al., 1983). Available data suggest that the episode started with total area of the Atakor massif (Rognon et al., 1981, 1983). Some
the emplacement of phonolite and trachyte domes, spines, and lava flows, emitted from the freshest-looking scoria cones, cover
flows, and ended with the outpouring in western Atakor of mafic Holocene lacustrine formations, dated by 14C at ca. 10 k.y. B.P.
plateau-forming lava flows. (cited by Girod, 1971). To the north, small basanite lava flows
An ~2.25 m.y. period of magmatic quiescence, from 4.2 ± cover the youngest alluvial terrace, in which Neolithic artifacts
0.2 Ma to 1.95 ± 0.2 Ma, i.e., during most of the Pliocene, was have been found in situ (Rognon, 1967). Moreover, the Tuareg
marked by active normal faulting and complete resurfacing of people relate ancestral stories suggesting that the volcanoes were
TABLE 1. REPRESENTATIVE WHOLE-ROCK (MAJOR AND TRACE ELEMENTS) COMPOSITIONS OF MAFIC ROCKS
Sample number: IA8 IJ1 IJ2 SA06 ISK7 SA03 PAS06 TAH01 BS01 Y6A
SiO2 42.73 43.24 43.93 44.05 45.25 50.44 43.64 49.66 46.77 46.88
TiO2 3.76 3.79 3.64 3.49 3.14 2.04 3.29 2.25 2.23 3.02
Al2O3 14.12 14.44 15.17 15.95 15.34 17.96 15.37 16.62 16.17 14.71
t
Fe2O3 * 14.44 14.41 14.44 13.77 12.86 9.25 13.54 10.28 10.67 11.95
MnO 0.18 0.19 0.19 0.20 0.17 0.25 0.19 0.18 0.20 0.18
MgO 7.57 7.03 6.19 5.79 5.66 2.67 7.39 4.93 5.65 7.73
CaO 10.06 9.57 9.07 9.20 9.48 6.52 10.86 6.88 9.38 8.79
Na2O 2.79 3.55 3.54 3.68 3.17 5.99 3.34 4.22 5.04 4.52
K2O 1.54 1.40 1.90 1.08 1.57 3.11 1.47 2.98 1.96 1.64
P2O5 0.87 0.93 1.12 0.70 0.81 0.85 0.88 0.65 0.84 0.54
LOI † 1.94 1.66 0.81 1.69 2.54 1.02 0.49 1.37 1.04 0.39
Total 100.00 100.21 100.00 99.60 99.99 100.10 100.46 100.02 99.95 100.35
Rb § 32.6 38.7 32.8 40.7 29.5 91.34 31.96 66.7 57.15 30.0
Cs 0.19 0.22 0.20 0.25 B.D.L. 1.26 0.363 0.36 0.708 0.64
Be 1.23 1.52 1.22 1.99 1.30 3.57 1.42 2.16 3.18 2.02
Sr 882 832 911 873 821 1302 830 618 1458 973
Ba 425 451 440 597 401 1085 506 737 960 663
Y 31.6 32.4 33.2 28.5 29.0 50.7 27.8 22.6 36.0 27.2
La 52.2 54.7 56.1 51.6 52.6 154.5 47.25 57.6 114.7 65.0
Ce 108 118 123 107 110 296.5 96.38 108 204.1 135
Pr 13.2 14.0 14.7 13.2 12.4 32.77 11.24 11.2 22.29 14.9
Nd 56.0 59.1 60.9 51.9 49.6 116.3 45.78 42.2 76.46 57.1
Sm 10.8 11.8 12.2 10.4 9.89 17.94 8.81 7.73 13.07 10.1
Eu 3.51 3.56 3.65 3.17 3.09 5.15 2.6 2.49 3.62 3.32
Gd 9.14 9.34 9.50 8.31 7.75 14.04 6.91 5.89 10.41 8.26
Tb 1.24 1.28 1.33 1.11 1.06 1.88 1.02 0.827 1.36 1.09
Dy 6.27 6.68 6.89 6.17 5.74 10.12 5.46 4.22 6.58 5.82
Ho 1.17 1.24 1.22 1.14 1.03 1.68 0.91 0.795 1.11 1.10
Er 2.73 2.90 2.87 2.57 2.49 4.91 2.34 2.10 2.85 2.42
Tm 0.324 0.333 0.373 0.339 0.363 0.66 0.3 0.267 0.35 0.304
Yb 2.15 2.34 2.28 2.28 2.15 3.67 1.91 1.83 2.32 1.93
Lu 0.308 0.303 0.320 0.312 0.265 0.59 0.236 0.265 0.31 0.273
Th 5.80 6.20 6.15 6.91 5.82 19.38 6.75 9.36 15.33 7.65
U 1.44 1.65 1.7 1.48 1.60 3.84 1.58 2.48 4.0 2.06
Zr 289 308 324 273 310 591 262 336 453 412
Hf 6.72 7.07 7.02 6.24 6.33 10.8 5.62 6.58 8.19 8.68
V 273 263 222 267 191 121 297 137 213 207
Nb 73.8 77.2 80.0 86.9 72.3 247.2 76.06 86.9 143.8 93.4
Ta 5.47 5.81 5.88 6.31 5.14 21.14 5.22 5.91 10.72 6.49
Cr 217 180 79.8 61.9 121 6.56 102 93.7 215 284
Mo 1.37 2.11 2.70 2.35 2.31 6.45 2.81 3.94 7.51 4.57
W 0.40 0.50 0.41 0.76 0.71 1.69 0.74 0.79 2.7 0.86
Co 49.2 47.1 40.1 49.2 37.9 17.1 49.7 26.7 32.7 47.1
Ni 106 85.9 54.2 59.1 73.3 4.18 76.1 63.4 80.6 168
Cu 54.5 49.0 40.1 49.7 38.2 14.0 65.4 27.6 31.7 51.5
Zn 154 159 157 148 140 154 145 119 156 143
Cd 0.26 0.21 0.31 0.23 B.D.L. 0.38 0.85 0.36 1.18 0.57
Ga 22.2 24.0 21.5 22.0 21.4 25.4 23.1 21.5 23.8 22.5
In 0.10 B.D.L. 0.10 B.D.L. B.D.L. 0.08 0.16 B.D.L. 0.17 B.D.L.
Ge 1.03 1.07 1.04 1.11 1.04 1.4 1.32 0.84 1.31 1.02
Sn 1.97 2.31 2.01 2.07 1.79 2.76 2.61 1.79 2.43 1.85
Pb 3.38 3.20 3.04 5.27 3.35 7.76 3.75 5.63 12.4 3.93
As 0.71 0.81 1.06 0.88 0.70 1.81 0.8 1.48 3.12 0.97
Sb B.D.L. # B.D.L. 0.10 B.D.L. B.D.L. 0.25 0.11 B.D.L. 0.32 0.12
Bi Traces ** Traces B.D.L. Traces Traces 0.04 0.17 Traces 0.19 Traces
* Total iron determined as wt% Fe2O3.
†
LOI = Loss on Ignition (wt%).
§
Trace elements (in ppm) are arranged in the increasing order of columns of the periodic table of elements and of atomic numbers.
#
B.D.L. = Below detection limit.
** Traces: No detection
IA8 Miocene basanite lava flow, South Tahifet area.
IJ1 Miocene basanite lava flow, South Tahifet area.
IJ2 Miocene basanite lava flow, South Tahifet area.
SA06 Miocene basanite lava flow, Northwest Assekrem area.
ISK7 Miocene alkali basalt lava flow, Isekram, Tamanrasset area.
SA03 Miocene phonotephrite lava flow, Northwest Assekrem area.
PAS06 Miocene—Pliocene basanite neck, Assekrem plateau.
TAH01 Miocene—Pliocene mugearite neck, Tahaleft, Assekrem area.
BS01 Pliocene tephrite lava flow, Oued Segueïka, Assekrem area.
Y6A Pliocene—Quaternary hawaiite neck, North Tahifet area.
TABLE 2. REPRESENTATIVE WHOLE-ROCK (MAJOR AND TRACE ELEMENTS) COMPOSITIONS OF FELSIC ROCKS
Sample TIL8 TIN3 OTAB01 SKR6 TEZ06 ADRN4 Y1C1 IM2 TEZ01 Y1F1 N1B LHA6 Y2A4 PAS01 DEB2'
number:
SiO2 59.52 61.38 61.77 62.34 62.73 62.81 64.73 65.80 66.81 66.82 67.47 67.61 70.59 60.19 59.81
TiO2 0.71 0.4 0.32 0.26 0.16 0.33 0.32 0.20 0.09 0.27 0.21 0.16 0.11 0.30 0.04
Al2O3 18.03 17.42 17.09 17.43 17.94 17.79 17.48 15.53 16.10 16.01 16.32 14.84 15.31 19.59 18.81
Fe2O3t 5.68 5.22 5.42 5.17 3.34 4.50 3.55 4.30 2.99 3.61 3.25 4.20 1.05 3.40 5.07
MnO 0.24 0.29 0.27 0.34 0.13 0.67 Traces 0.15 0.12 0.05 0.06 0.28 Traces 0.21 0.23
MgO 1.35 0.81 0.36 0.52 0.22 0.23 Traces 0.17 0.15 0.17 0.08 0.08 0.16 0.24 0.26
CaO 1.39 1.38 1.39 0.80 0.70 0.68 0.45 0.71 0.64 0.75 0.46 0.55 0.35 0.98 1.26
Na2O 6.09 7.13 7.14 7.00 6.21 7.04 5.99 7.07 6.83 5.43 5.57 6.80 4.91 8.02 8.59
K2O 4.84 5.01 5.08 5.15 4.74 4.92 5.61 4.75 4.55 5.31 5.31 4.53 4.40 5.57 5.13
P2O5 0.38 0.27 0.2 0.16 0.11 0.23 0.12 0.08 0.11 0.11 0.17 0.02 0.05 0.04 0.14
LOI 1.76 0.64 0.91 0.75 3.43 0.73 2.04 0.80 1.17 1.82 1.37 0.89 2.57 1.43 0.60
Total 99.99 99.95 99.95 99.92 99.71 99.93 100.30 99.56 99.56 100.35 100.27 99.97 99.50 99.97 99.94
Rb 113 132.7 172.5 127 353 115 191.2 313.4 269 165.5 204.8 308 345 268 227
Cs 0.47 1.63 1.06 0.54 18.65 0.29 0.916 1.36 0.785 0.443 0.922 0.68 1.24 2.18 1.47
Be 2.90 4.5 7.33 4.53 15.6 4.63 5.97 12.4 3.4 4.73 3.27 12.1 7.57 9.09 9.74
Sr 313 70.7 41.5 31.1 39.6 60.2 23.4 21.7 35.7 16.0 15.3 17.4 36.8 38.0 64.4
Ba 972 311 58 189 41 527 34 30 34 36 25 22.8 13 49.7 21
Y 21.7 38.3 60.3 37.6 62.0 31.7 44.5 91.9 74.9 55.0 46.7 62.1 62.6 34.6 78.9
La 71.6 109.4 163.1 103 142 95.6 95.38 263.5 62.16 112 62.03 120 48.26 156 228.6
Ce 149 195.5 283.6 182 253.3 182 132 440 129.4 221.6 208.1 229 79.71 226 420.4
Pr 13.9 19.78 29.94 17.7 27.07 18.3 18.9 47.8 14.63 22.33 14.16 21.1 8.8 19.0 44.6
Nd 46.4 65.16 95.12 56.7 84.85 60.7 64.23 156.5 48.68 72.28 48.56 62.4 26.75 49.6 148
Sm 7.20 10.68 15.38 9.08 13.71 9.38 10.96 26.49 9.56 12.68 9.03 10.5 6.07 6.61 23.17
Eu 2.52 2.23 1.51 1.57 1.28 1.96 0.59 0.77 0.544 0.61 0.526 0.401 0.08 0.731 0.77
Gd 4.72 7.46 11.31 6.69 10.48 6.85 8.96 20.01 8.12 10.62 7.54 8.64 6.37 5.11 16.86
Tb 0.774 1.17 1.67 1.05 1.7 1.04 1.34 2.96 1.55 1.73 1.34 1.57 1.31 0.827 2.70
Dy 3.91 6.92 10.79 6.05 10.38 5.77 7.86 16.52 10.92 10.24 8.15 9.50 9.38 5.11 15.53
Ho 0.786 1.36 2.01 1.27 1.99 1.11 1.51 2.92 2.27 2.01 1.67 2.02 2.18 1.11 2.75
Er 2.05 3.46 5.7 3.46 5.86 3.04 4.04 8.44 6.93 5.51 4.68 5.71 5.96 3.10 8.02
Tm 0.321 0.58 0.91 0.585 1.02 0.481 0.66 1.22 1.16 0.88 0.75 1.03 1.02 0.534 1.20
Yb 1.97 4.02 5.98 4.17 6.53 3.19 4.33 7.66 7.49 5.64 5.02 7.49 6.82 3.82 7.37
Lu 0.282 0.63 1.04 0.602 1.03 0.503 0.68 1.13 1.16 0.81 0.72 1.01 0.91 0.628 1.20
Th 14.3 18.79 29.47 26.7 86.35 19.3 34.07 48.72 41.64 29.54 33.09 58.1 57.12 51.8 43.11
U 2.96 5.92 5.35 4.43 20.44 4.56 5.09 11.87 11.26 6.42 7.53 10.4 12.56 13.6 9.87
Zr 528 718 1001 893 1217 827 954 986 643 778 657 1173 630 1191 1118
Hf 11.1 16.5 24.4 17.9 37.6 16.4 22.3 29.1 23.0 19.8 16.8 34.0 24.4 24.4 32.1
V 7.5 2.53 0.6 B.D.L. 1.91 B.D.L. 2.17 0.99 0.78 3.64 0.25 Traces Traces 2.60 Traces
Nb 127 171.5 247.8 193 508 163 204.4 379.1 320.8 184.2 189.6 450 294.5 288 365.2
Ta 9.49 13.22 17.65 14.5 52.04 12.7 13.9 29.3 33.18 14.52 14.93 38.5 31.28 19.4 29.42
Cr B.D.L. 3.03 3.65 Traces 2.82 Traces 4.01 1.79 1.52 3.01 3.56 Traces 6.57 B.D.L. 1.24
Mo 1.46 7.4 4.78 2.83 3.25 3.32 2.08 13.7 1.01 7.4 12.7 5.86 0.55 22.4 2.88
W 0.65 0.71 1.78 1.40 3.28 2.30 3.22 5.54 4.79 4.66 3.41 3.49 3.96 1.45 2.18
Co 3.12 0.75 0.61 0.75 0.86 0.37 0.59 0.57 2.72 0.44 0.50 B.D.L. 0.25 1.22 0.09
Ni B.D.L. 0.92 1.43 B.D.L. 4.57 B.D.L. 1.15 0.93 0.83 3.17 0.81 B.D.L. 1.10 B.D.L. Traces
Cu 5.7 2.6 5.2 B.D.L. 6.4 3.8 4.7 4.9 3.9 5.3 5.30 B.D.L. 1.3 7.2 5.6
Zn 121 143 192 143 211 119 67.8 264 230 92.8 151 238 29.8 142 326
Cd 0.35 0.27 0.71 0.25 0.66 0.39 0.55 0.97 0.47 0.43 0.39 0.87 0.35 0.84 0.35
Ga 22.2 24.5 32,5 24.2 48.8 24.1 32.2 47.3 41.6 34.4 35.1 33.9 38.5 34.3 52.9
In 0.12 0.14 0.2 0.12 0.06 B.D.L. 0.1 0.2 0.09 0.17 0.15 0.13 0.12 B.D.L. 0.25
Ge 1.19 1.29 1.85 1.30 2.33 1.18 1.89 2.1 2.12 1.79 1.76 1.61 1.79 1.30 2.26
Sn 2.59 2.98 4.65 3.12 10.2 2.21 5.65 11.2 11.4 5.71 6.74 8.53 13.7 4.75 9.47
Pb 5.90 9.69 16.4 23.9 29.9 9.01 9.79 24.4 29.7 18.9 19.0 26.8 38.6 21.3 29.7
As 0.48 2.02 2.83 0.76 3.49 0.57 1.66 6.39 0.58 2.82 1.30 2.20 1.51 2.07 1.82
Sb 0.13 0.26 0.3 0.22 1.12 0.17 0.28 0.68 0.76 0.6 0.45 0.38 0.39 0.46 0.33
Bi Traces 0.02 0.04 Traces 0.21 Traces 0.02 0.1 0.01 0.04 0.04 Traces 0.11 B.D.L. Traces
Note: Same mode of presentation as in Table 1.
TIL 8 Trachyte neck, Tilaouis, near Tamanrasset.
TIN3 Trachyte dome, Tindi, near Tamanrasset.
OTAB01 Trachyte dome, Oued Tabbezet, Assekrem area.
SKR6 Trachyte plateau, Isekram, near Tamanrasset.
TEZ06 Trachyte dome, Tezouiadje, Assekrem area.
ADRN4 Trachyte dome-lava, Adriane, near Tamanrasset.
Y1C1 Trachyte dome, North Tahifet area.
IM2 Aegirine trachyte neck, Imross, South Tahifet area.
TEZ01 Trachyte dome, Tezouiadje, Assekrem area.
YIF1 Trachyte dome, North Tahifet area.
N1B Trachyte neck, Northeast Tahifet area.
LHA6 Trachyte neck, Iharen, near Tamanrasset.
Y2A4 Rhyolite dome, North Tahifet area.
PAS 01 Miaskitic phonolite dome-lava, Assekrem plateau.
DEB2' Agpaitic phonolite dome-lava, Debnat, near Tamanrasset.
Tertiary alkaline volcanism of the Atakor massif 327
Figure 3. Atakor landscape viewed from north of Tamanrasset showing flood lava plateau in the background, dome lavas
and necks, and Proterozoic basement in the foreground.
actually seen erupting. Sporadic fumaroles, observed by Girod 2. The felsic group (17% of the total area) includes pyro-
in 1961, persistent, yet local, seismic activity of small amplitude clastic formations, lava flows, and the impressive domes, necks,
(Grandjean, 1962), and fault-controlled mineral springs with and spines that dominate the landscape. It is composed of ben-
mantle-derived gas emissions provide evidence that the Atakor moreites, miaskitic and agpaitic phonolites, metaluminous and
massif should not be considered as extinct. peralkaline trachytes to quartz trachytes, and peraluminous
rhyolites (Table 2). Two diverging evolutionary trends that origi-
Classification of the Volcanic Rocks nate in a common ensemble of intermediate rocks (benmoreite
and metaluminous trachyte) can be traced in the TAS diagram
In the Atakor literature, the volcanic rocks have typically (Fig. 4): the first trend, from trachyte to phonolite, is character-
been named according to a simplified “field” classification. ized by intermediate silica and increasing alkali contents, while
“Basalt” is used for any dark-colored lava flow, and includes both the second trend, from trachyte to rhyolite, is marked by increas-
basanite and tephrite, whereas “hawaiite” and “mugearite” cor- ing silica contents.
respond to gray lava flows and domes, and include phonoteph- The distribution of the various rock types is not random.
rites. Likewise, “phonolite” is used for any felsic rocks bearing Mafic rocks are abundant in the center, north, and west of the
feldspathoid(s). Its correct use, however, should be restricted to massif, where they constitute large plateaus, but are scarce in the
felsic rocks with either more than 10% modal foid in the alkali south, where they occur as scattered outcrops preserved by ero-
feldspar-plagioclase-feldspathoid (APF) diagram (Le Maitre, sion, and as valley-filling lava flows. Benmoreites, previously
1989), or located above the silica undersaturated-saturated (U-S) classified as trachytes by Girod (1971), occur as tuffs intercalated
boundary in the total alkalis versus silica (TAS) variation dia- within the Miocene basalt plateau and also as domes. Phono-
gram, i.e., having more than 10% normative foid (Le Bas et al., lites, in the IUGS sense, are widespread in the Assekrem plateau
1986, 1992). According to recommendations of the Subcommis- (Yahiaoui, 2003) and Debnat complex, east of Tamanrasset, but
sion on the Systematics of Igneous Rocks of the International notably lacking in the Tahifet area. Trachyte, in the IUGS sense,
Union of Geological Sciences (IUGS) (Le Maitre, 2002), most i.e., including “phonolite” that contains less than 8 wt% CIPW-
“phonolites” of the Atakor massif are not sufficiently silica- normative nepheline, occurs everywhere as domes and spines,
undersaturated, and actually plot in the trachyte field. with related lava flows. A single occurrence of peraluminous
Based on field evidence, two groups of magmatic rocks, one rhyolite was discovered in the Tahifet area (Benhallou, 2000).
mafic and one felsic, separated by a “Daly gap” from 51 to 57
wt% SiO2, have been identified (Tables 1 and 2): MAFIC GROUP OF ROCKS
1. The mafic group (83% of the total area) is composed of
dark lava flows and scoria cones, which are made up of basalt and The mafic group is characterized by CIPW nepheline-
basanite, and of gray lava flows, which were previously determined normative compositions. Although it can be subdivided into com-
(Girod, 1971) to be hawaiite and mugearite, but which include positionally dispersed subtypes, i.e., basanite, basalt, hawaiite,
more strongly silica-undersaturated types, such as phonoteph- mugearite, and phonotephrite, the range of compositions plots
rite (Table 1). Whatever their age of eruption, they constitute a within a rather small area of the TAS diagram (Fig. 4). The ultra-
coherent group in the TAS diagram straddling the U-S boundary basic to basic (43–47 wt% SiO2) magmatic rocks have pheno-
from basanite-basalt to mugearite-phonotephrite (Fig. 4). crysts of olivine + clinopyroxene ± plagioclase. The more differ-
Atakor volcanic massif
18 Na2O + K2O
silica-undersaturated
16 felsic trend
14
silica-saturated
felsic trend
12
U
10
S benmoreite tuff
mafic trend
8
Tahifet area (Benhallou, 2000)
Assekrem area (Yahiaoui, 2002)
6
Tamanrasset area (unpublished data)
U-S boundary (Le Bas et al., 1986, 1992)
4
phonolites (Girod, 1971)
trachytes (Girod, 1971)
2
SiO2
0
35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77
Figure 4. Selected volcanic rock compositions, recalculated on a H2O- and CO2-free basis, plotted in the total alkali–silica
(TAS) diagram (Le Bas et al., 1986, 1992). The U-S boundary is indicated. Diamonds: Tahifet samples (Benhallou,
2000); squares: Assekrem samples (Yahiaoui, 2003); triangles: Tamanrasset samples (this paper); open symbols: selected
samples of benmoreite, phonolite, and trachyte from Girod (1971).
entiated mafic rocks (47–51 wt% SiO2) have a similar phenocryst ulvöspinel-magnetite and ilmenite-hematite solid solutions, scarce
mineralogy but with more abundant plagioclase. Widespread sulfides, such as chalcopyrite and pyrite, abundant apatite, and late-
secondary “sunburn” analcime occurs in the groundmass. stage carbonates complete the accessory mineral assemblage.
Mineral Chemistry Megacrysts
Olivine occurs as phenocrysts and microcrysts and contains Megacrysts of Al-Ti diopside and kaersutite, associated
inclusions of Cr-spinel and Fe-Ti oxides. Iddingsitized rims pro- with euhedral crystals of zircon and oxide minerals, including
vide evidence of incipient alteration before and during lava flow Mg-rich ilmenite, occur together with peridotite xenoliths and
eruption. The phenocrysts are normally zoned from Fo87–72 cores amphibole-rich enclaves (Dautria et al., 1987). The enclaves
to Fo74–50 rims; the microcrysts have the same compositions as the have Ti-rich, Si-poor chemistries, and are genetically unrelated
phenocryst rims. Concentrations of Mn (up to 3.6 wt% MnO) and to their host rocks, implying that they crystallized under upper-
Ca (up to 0.51 wt% CaO) correlate positively with Fe content. mantle conditions during earlier, undated, igneous episodes.
Clinopyroxene is highly calcic, varying from Al-Ti diopside The presence of kaersutite suggests that metasomatic fluids may
to diopside to hedenbergite, with phenocryst cores containing have modified the subcontinental lithospheric mantle. Based on
more Al and Ti than rims and microcrysts in the groundmass. the occurrence of Mg-rich ilmenite (with 31%–37% geikielite
Plagioclase is normally zoned from bytownite to labradorite and 10%–25% hematite components) and the abundances of
to andesine (An71 to An39) in the ultrabasic to basic rocks, and from incompatible trace elements in the peridotite xenoliths, Dautria
labradorite to oligoclase (An69 to An17) in the more differentiated et al. (1988) identified the metasomatizing agent as a kimber-
mafic rocks, in which it coexists with anorthoclase microcrysts. lite magma that interacted in small and variable amounts with
In addition to the olivine + clinopyroxene + plagioclase micro- upper-mantle peridotite.
cryst assemblage, the groundmass contains interstitial to poikilitic Kaersutite occurs as unaltered, centimeter-size, euhedral
crystals of analcime and Ti-rich (4.9–8.5 wt% TiO2) biotite (Girod, megacrysts in volcanic ash and scoria deposits. It has been par-
1971; Benhallou, 2000), indicating that fluids were available in fairly tially to totally converted into a plagioclase + clinopyroxene +
large amounts at temperatures close to the solidus. Discrete grains of rhönite + oxides ± olivine assemblage in the basanite-tephrite lava
Common Assemblage of Major Rock-Forming Minerals
Well-aligned alkali feldspar phenocrysts and microcrysts

define a trachytic texture. All phenocrysts and most micro-
crysts are monoclinic (sanidine), and yield a restricted range
of compositions from Or23 to Or45, consistent with hypersolvus
crystallization. Some microcrysts are composed of two discrete
populations of sanidine Or92–95 and albite Or01–10, and therefore
crystallized subsolvus.
In silica-undersaturated felsic rocks, nepheline occurs as
euhedral phenocrysts rimmed by analcime. Analcime occurs also
as late, poikilitic crystals, rarely accompanied by members of the
cancrinite-vishnevite solid solution series. In silica-oversaturated
rocks, occasional β-type quartz crystals have euhedral shapes.
The groundmass is speckled with Fe-Ti oxides of the
ulvöspinel-magnetite solid solution series that are commonly
oxidized to Ti-bearing hematite. The common accessory mineral
Figure 5. Kaersutite phenocryst in Miocene mugearite of the Tahaleft assemblage includes apatite, zircon, titanite, and rutile.
neck (photograph from Yahiaoui, 2003). The size of the crystal is
~0.5 mm.
Rock-Forming Minerals of the Trachyte-Rhyolite Trend
Both peralkaline and peraluminous felsic rock types com-

flows (Girod, 1971), whereas fresh phenocrysts (Fig. 5) occur in prise the trend. The widespread peralkaline suite evolved from
the mugearite of the Tahaleft neck (Yahiaoui, 2003). Megacrysts metaluminous trachyte to aegirine trachyte. The peraluminous
of kaersutite are also found in other districts of the Hoggar vol- group is composed of biotite-trachyte and rhyolite. A third group,
canic province, e.g., Tahalra (Dautria et al. 1987), Manzaz, Adrar well developed in the Akarakar area (Girod, 1971), is composed
N’Ajjer, and Djanet (authors’ observations). of fayalite trachyte, akin to the peralkaline suite.
An altered megacryst of Ti-rich biotite was observed in a Fayalite occurs as phenocrysts as well as microcrysts in the
basanite sample from the Tahifet area (Benhallou, 2000). The rim groundmass, where it coexists with quartz and/or tridymite. Its
of the megacryst is crowded with exsolved rutile needles, and optical characteristics indicate it contains less than 10% forsterite
the core is completely converted into an alkali feldspar + clino- (Girod, 1971).
pyroxene + rhönite + olivine + ilmenite assemblage. Clinopyroxene is the most abundant mafic mineral in the
In addition to mafic silicate megacrysts, euhedral crystals peralkaline suite. Na-bearing (0.15 Na atoms per formula unit
of zircon and octahedral crystals of oxide minerals are found [a.p.f.u.]) hedenbergite is the chief mafic mineral of the Tezouiadje
in the scoria deposits. Crystals of spinel-hercynite solid solu- domes. In the Oued Tabezzet dome, phenocrysts are composed
tion are homogeneous and include early Mg-rich ilmenite, or of large brown Na-bearing (0.07–0.10 Na a.p.f.u.) hedenbergite
exsolved hematite lamellae and corundum grains. Crystals of cores surrounded by narrow green aegirine augite rims (0.24–
ulvöspinel-magnetite solid solution are crowded with exsolved 0.36 Na a.p.f.u.), whereas green microcrysts in the groundmass
ilmenite-hematite and pseudobrookite lamellae. The complex are more sodic (0.43 and 0.71 Na a.p.f.u.) aegirine augite. In the
Fe-Ti oxide assemblage appears to have resulted from fluctuat- Tahifet area, nearly pure (0.89–0.95 Na a.p.f.u.) aegirine occurs
ing oxygen fugacity (fO2) at near-liquidus temperatures (Con- as interstitial microcrysts.
quéré and Girod, 1968). Arfvedsonite occurs only as interstitial, poikilitic crystals in
the groundmass. Its composition is marked by high K2O contents
FELSIC GROUP OF ROCKS (up to 2 wt%), with probable occurrence of Li in the octahedral
sites that are not fully occupied by Ti, Al, Fe, Mn, and Mg.
The less abundant felsic group is compositionally more Aenigmatite forms crystals surrounding sanidine and quartz
varied than the mafic group. Two diverging trends, rooted in the in fayalite trachyte and arfvedsonite trachyte (Girod, 1971). In
benmoreite + trachyte fields, are evident in the TAS diagram a Tezouiadje dome (Yahiaoui, 2003), dark-brown crystals of a
(Fig. 4): these evolve either from trachyte to rhyolite, or from Ti-bearing (0.21–0.35 Ti a.p.f.u.) mineral surround sanidine
trachyte to phonolite. Both trends contain metaluminous types, crystals in the groundmass. The fairly Ti-poor and Fe-rich com-
which evolve to peralkaline and/or peraluminous end mem- position of this mineral, relative to aenigmatite, suggests that it
bers. In peralkaline types, sodic mafic minerals of the agpaitic belongs to the wilkinsonite-aenigmatite (Duggan, 1990) solid
sequence crystallized late. Phenocrysts, when present, are chiefly solution series. This occurrence of a wilkinsonite-like mineral in
composed of alkali feldspar and calcic mafic minerals. a peralkaline trachyte is the third reported so far in the world.
Biotite is rare, occurring only in peraluminous trachytes, contents range from 230 ppm in basanite to 660 ppm in pho-
as brown-red acicular microcrysts in the groundmass and rarely notephrite; chondrite-normalized REE patterns (Figs. 6A–6B)
included within sanidine. Its composition is Ti-rich (~7.00 wt% are strongly fractionated; (La/Yb)N ranges from 20 to 40 without
TiO2). The absence of detectable halogens (F and Cl) suggests Eu anomalies, suggesting a garnet-bearing source and no plagio-
that coexisting fluids were essentially aqueous. clase fractionation. Multi-element diagrams (Figs. 6C–6D)
suggest derivation of the magmas from an enriched mantle
Rock-Forming Minerals of the Trachyte-Phonolite Trend source by fairly low degrees of partial melting. Primitive
mantle–normalized trace-element patterns are strongly frac-
Metaluminous (miaskitic) and peralkaline (agpaitic) varie- tionated, with normalized abundances ranging from 125 to 500
ties comprise the trend, and intermediate types are rare. In addi- (Ta, Nb) to 4–7 (Yb). The more highly incompatible large ion
tion to nepheline and analcime, haüyne was observed in only lithophile elements (LILEs), Rb, Ba, and Th, have normalized
three intrusions (Girod, 1971). The mafic mineralogy resembles values between 50 and 100; K is slightly depleted in some sam-
that of the trachyte-rhyolite trend, but with rare phenocrysts of ples. The less incompatible elements are increasingly depleted
decayed brown amphibole observed in miaskitic trachyte and from 30 to 50 (P) down to 3–7 (Yb).
phonolite. Contrasting with the accessory titanite + apatite +
zircon association of the miaskitic phonolites, the accessory Evolutionary Trends in the Felsic Group of Rocks
mineral assemblage of the agpaitic phonolites (Azambre and
Girod, 1966) includes rare minerals, such as aenigmatite, eudia- The felsic and mafic trends are separated by a Daly gap in the
lyte, lavenite, rinkite (named ‘mosandrite’ in the original pub- 51–57 wt% SiO2 range (Table 2). The felsic group of rocks shows
lication), and possibly rosenbuschite. Other rare minerals have two well-defined diverging trends in the TAS diagram (Fig. 4).
yet to be identified. The silica-oversaturated trachyte-rhyolite trend, marked
by increasing silica and slightly decreasing alkali contents,
GEOCHEMICAL EVOLUTION reflects alkali feldspar fractionation and late precipitation of
quartz. Regarding incompatible trace element behaviors, two
Origin and Evolution of the Mafic Group of Rocks subgroups of rocks have been identified. Tamanrasset trachytes
constitute the first subgroup, and Atakor trachytes-rhyolites, the
Mafic lava flows and necks are easily distinguished in the second one.
field on the basis of their color. Dark-colored lava flows are ultra- The Tamanrasset trachytes are distinguished from the Atakor
basic to basic in composition (43%–47% SiO2), with compat- trachytes by their REE characteristics (Fig. 7A). REE contents
ible trace element concentrations (Ni = 54–168 ppm, Cr = 70– vary from 300 to 500 ppm; REE patterns are strongly fraction-
284 ppm; Table 1) and MgO contents (6–8 wt %) that indicate ated, with (LaN/YbN) of 27–12.5 and light (L) REE enrichment
relatively primitive compositions. The gray lava flows are more relative to heavy (H) REE. Eu anomalies vary from weakly posi-
differentiated (47%–51% SiO2) with lower contents of compat- tive (Eu/Eu* = 1.25 for 60.5 wt% SiO2) to weakly negative (from
ible trace elements (Table 1) and MgO contents < 5 wt%. Eu/Eu* = 0.80 at 61 wt% SiO2 to 0.60 at 63 wt% SiO2), with
The Miocene and early Pliocene basanite plateau lavas one exception of Eu/Eu* = 0.15, which corresponds to a more
and necks are Si-poor (43% SiO2), fairly mafic, with an Mg# evolved (68 wt% SiO2) sample with the highest REE content and
(Mg/[Mg + total Fe] atomic ratio) of ~0.50, and contain 85–110 the least fractionated REE pattern. Based on these data, it can
ppm Ni and 180–220 ppm Cr. Though containing peridotite be inferred that some Tamanrasset trachytes retain a cumulative
xenoliths and mantle-derived megacrysts, the most magne- feldspar component and that feldspar fractionation was delayed
sian compositions are still not consistent with those of primary until SiO2 reached ~61 wt%. In primitive mantle–normalized
magmas, which should have Mg#s in the range 0.70–0.82. The trace element patterns (Fig. 7D), strong fractionation from
magmatic rocks of the basanite-phonotephrite trend, with from LILE (normalized values in the 100–1000 range, except Ba) to
43% to 51% SiO2, are saturated in olivine and clinopyroxene, high field strength element (HFSE) (normalized values around
suggesting that differentiation may have occurred at mantle 10 for Y and Yb) is obvious. In addition, weak to increasingly
depths during the Miocene and the early Pliocene. pronounced negative anomalies in Ba, Sr, P, Eu, and Ti (Figs.
Quaternary basanite-tephrite lava flows and necks are more 7A–7D) provide evidence for incipient feldspar + apatite + Fe-Ti
silicic (47% SiO2), but also more magnesian (Mg# of 0.51–0.56). oxide fractionation.
They contain 80–170 ppm Ni and 215–280 ppm Cr. The mag- Atakor trachytes-rhyolites have high total REE contents, vary-
mas were multiply saturated but no really primitive compositions ing widely from 630 ppm (62 wt% SiO2) to 1000 ppm (66 wt%
have been sampled. No differentiated rocks of the same age have SiO2) to 375 ppm (68 wt% SiO2) to 200 ppm (72 wt% SiO2). REE
been found. are incompatible in trachytic liquids until the trachyte-rhyolite
Incompatible elements are abundant in all rock types, and boundary is reached, where they ultimately become compatible,
their behaviors are roughly similar, implying nearly identical probably because of apatite + zircon fractionation, as evidenced
differentiation processes at depth. Rare earth element (REE) by decreasing amounts of P2O5, Zr, and Hf. LREE patterns are
1000 1000
A C
Normalized to primitive man tle

TAH01 TAH01
SA06 SA06
PAS06 PAS06
SA03 SA03
IJ1 IJ1
Normalized to chondrite
IA8 IA8
IJ2 IJ2
100 ISK7 100 ISK7
10 10
Neogene basanites and phonotephrites
1 1
La Ce Pr Nd SmEu GdTb Dy Ho Er Tm Yb Lu Rb Ba Th U K Nb Ta La Ce Pb Pr Sr P Zr Hf Sm Ti Y Yb V
1000 1000
B D
Normalized to primitive ma ntle
BS01 BS01
Y6A Y6A
100 100
10 10
Quaternary basanites and phonotephrites

1 1
La Ce Pr Nd SmEu Gd Tb DyHo Er Tm Yb Lu Rb Ba Th U K Nb Ta La Ce Pb Pr Sr P Zr Hf Sm Ti Y Yb V
Figure 6. Geochemical characteristics of the mafic group of rocks. (A) Rare earth element (REE) patterns of Neogene rocks (normalization
values of McDonough and Sun, 1995). (B) REE patterns of Quaternary rocks. (C) Primitive mantle–normalized multi-element patterns of Neo-
gene rocks (normalization constants from Sun and McDonough, 1989, updated in McDonough et al., 1992). (D) Primitive mantle–normalized
multi-element patterns of Quaternary rocks.
slightly fractionated (Fig. 7B), with no HREE fractionation, but granite (WPG) field (Pearce et al., 1984). Extremely low (less
with deep negative Eu anomalies. Eu/Eu* values range from 0.5 than 0.3) Y/Nb ratios correspond to the A1 group of A-type gran-
to 0.1 in the trachytes down to 0.04 in the rhyolites, indicating ites (Eby, 1992), which are considered to be representative of rift
extensive feldspar fractionation. Primitive mantle–normalized and/or hotspot tectonic settings.
trace-element patterns (Fig. 7D) show that K is depleted relative The more complex silica-undersaturated trachyte-phonolite
to U and Ta. As a continuation of the incipient fractionation proc- trend involves two steps in the TAS diagram (Fig. 4). The first
esses observed within the Tamanrasset trachytes, deep troughs in step, which has both increasing silica and alkali contents, is
Ba, Sr, P, Ti, and Eu (Figs. 7 B–7D) are conspicuous in the trace- illustrated by the metaluminous (miaskitic) subgroup. The sec-
element patterns of the high-silica trachyte and rhyolite. Other ond step, which has decoupled behaviors of silica (decreas-
incompatible elements can reach extremely high normalized val- ing contents) and alkalis (increasing contents), characterizes
ues, e.g., around 1000 for Th, U, Ta, and Nb in rhyolite. the peralkaline (agpaitic) subgroup, which has a highly sodic,
The silica-oversaturated trachyte-rhyolite trend is character- silica-poor mineralogy. The two-step trend is well expressed in
ized by very high Y + Nb values and plots in the within-plate diagrams of differentiation index (DI), defined as the sum of
1000 1000

A TIN3
TIL8
D
TIN3 ADRN4
TIL8 SKR6
ADRN4 100 LHA6
SKR6
100 LHA6
10
10
Tamanrasset trachytes
1 0.1
La Ce Pr Nd SmEu Gd Tb Dy Ho Er TmYb Lu Rb Ba Th U K Nb Ta La Ce Pb Pr Sr P Zr Hf Sm Ti Y Yb V
1000 10000
B
Normalized to primitive ma ntle
Y1C1
Y1C1
Y1F1
E
Y1F1 Y2A4
Y2A4 IM2
1000 OTAB01
IM2
OTAB01 TEZ06
TEZ06 TEZ01
100 N1B
TEZ01
N1B 100
10
10
Atakor trachytes and rhyolites
1 0.1
La Ce Pr Nd SmEu Gd Tb Dy Ho Er TmYb Lu
Rb Ba Th U K Nb Ta La Ce Pb Pr Sr P Zr Hf Sm Ti Y Yb V
1000 1000
C
DEB2' DEB2' F
PAS01 PAS01
100
100
10
10
Phonolites
1 0.1
La Ce Pr Nd SmEu Gd Tb Dy Ho Er TmYb Lu Rb Ba Th U K Nb Ta La Ce Pb Pr Sr P Zr Hf Sm Ti Y Yb V
Figure 7. Geochemical characteristics of the felsic group of rocks. (A) Rare earth element (REE) patterns of Tamanrasset trachytes.
(B) REE patterns of Atakor trachytes and rhyolites. (C) REE patterns of phonolites. (D) Primitive mantle–normalized multi-element patterns
of Tamanrasset trachytes. (E) Primitive mantle–normalized multi-element patterns of Atakor trachytes and rhyolites. (F) Primitive mantle–
normalized multi-element patterns of phonolites.
CIPW-normative orthoclase + albite + quartz + nepheline, versus e.g., East African Rift and Ethiopia, Cameroon line (Fitton, 1987),
CIPW-normative nepheline (Fig. 8A). In peralkaline rocks, Na2O Cabo Verde archipelago and Canary Islands, and the Maghreb
in excess of Al2O3 is converted into CIPW-normative acmite (Ac) (Lustrino, 2000; Savelli, 2002). The association of topographic
and sodium disilicate (Ns), which are not considered in the DI swells and Cenozoic volcanism is remarkable, and to account
calculation. The two steps remain well defined in a diagram com- for the relationships among within-plate volcanic activity, topo-
bining DI and (Ac + Ns) (Fig. 8B). Phonolites are poorly repre- graphic elevation, and mantle seismic velocity anomalies, the
sented in our data set, and only two agpaitic types were sampled. involvement of one or several mantle plumes has been suggested
REE contents and patterns resemble those of the trachyte-rhyolite (for contrasting views, see Burke, 1996; Liégeois et al., 2005; Pik
trend, with fractionated LREE, negative Eu anomalies (Eu/Eu* et al., 2006). The popular plume model postulates that hot mantle
of 0.25–0.1), and no HREE fractionation (Fig. 7C). Primitive material upwelling from the lower mantle impinged on the base
mantle–normalized trace-element patterns (Fig. 7E) do not show of the Tuareg shield lithosphere and induced partial melting of
such extreme enrichments as in the rhyolites, but their shapes either the subcontinental lithospheric mantle, or the plume itself,
suggest similar sources and processes. or a mixture of both (e.g., Aït-Hamou et al., 2000).
The Cenozoic evolution of the Tuareg shield has to be con-
DISCUSSION sidered in the context of the motion of the African plate. As noted
by Frizon de Lamotte (2006), Africa is almost entirely (90%) sur-
The Atakor massif is located within the Hoggar province, rounded by passive continental margins, with the Afar triple junc-
which constitutes part of the large Tuareg shield in the north- tion commonly regarded as plume-related. The only exception
central part of West Africa (Black et al., 1994). While some Ceno- is the Mediterranean region (Fig. 1), where Africa-Eurasia con-
zoic volcanic provinces within the African continent (Liégeois vergence has been accommodated since the Cretaceous either by
et al., 2005) are clearly within-plate in character, e.g., Hoggar, subduction, a process that occurs now only in the Eastern Medi-
Libya, Tibesti, and Darfur, others are related to plate boundaries, terranean, or by transcurrent and thrust faults that are currently
active in the Western Mediterranean. The stresses at convergent
plate boundaries are higher, by several orders of magnitude, than
60 65 70 75 80 85 90 95 100 those at passive continental margins, as illustrated by the cur-
0
benmoreite DI rent low levels of seismicity along the coastlines of Africa com-
5
pared with the Atlas belt of Morocco (Sébrier et al., 2006), and
the Maghreb in general. The Tuareg shield is only 1000 km south
10 miaskitic trend of the Mediterranean coast (Fig. 1). In the case of within-plate
igneous activity, the role played by events occurring at nearby
15 phonolite plate margins should be envisaged as an alternative possibility to
mantle plumes of deep origin (Anderson, 2005). The plate tec-
20 agpaitic trend tonics paradigm predicts that the Hoggar volcanic activity and
reactivation of Precambrian fault systems are likely to be more
25
closely related to geodynamic events occurring in the Mediter-
30 CIPW-normative Ne
ranean region than at more remote, passive plate margins.
Atakor volcanic massif In the following discussion, we examine the nature and com-
60 65 70 75 80 85 90 95 100
position of the upper-mantle sources of the magmatism, and the
0 periodicity of the volcanic activity, discuss the various models
D I + CIPW-normative (Ac + Ns) that might explain how magmas can be generated in this within-
5 plate tectonic setting, and the temporal relationships with the
geodynamic evolution of the Western Mediterranean.
10
Upper Mantle Sources of the Cenozoic Magmas

15
20
The primitive mantle–normalized multi-element patterns of
the mafic rocks of the Atakor massif are highly enriched in incom-
25 patible elements (Fig. 6), similar to the HIMU (high μ) end mem-
ber of the oceanic-island basalt (OIB) spectrum (Weaver 1991a,
CIPW-normative Ne
30 1991b). This observation is consistent with Sr-Nd-Pb isotope char-
acteristics of the Atakor basalts (Allègre et al., 1981; Dupuy et al.,
Figure 8. Selected trachyte-phonolite compositions, recalculated on
H2O- and CO2-free basis, using same symbols as in Figure 4. CIPW- 1993), which indicate a HIMU affinity for the mantle source.
normative Ne versus differentiation index (DI) diagram and CIPW- The temperature and depth (or pressure) at which the mafic
normative Ne versus (DI + Ac + Ns) diagram. magmas segregated from their mantle sources can be calcu-
lated based on empirical equations, with statistical uncertain- 8 km (work in progress). A third possibility would be that the mafic
ties of ±40 °C and ±8 km (or ±0.27 GPa) (e.g., Albarède, 1992). magmas originated from different upper-mantle sources. The least
Removal of clinopyroxene can shift the calculated temperature differentiated, noncumulative mafic lavas have Mg# ranging in the
and pressure by ~5–10 °C and 0.1–0.3 GPa. In the following esti- Tahalra massif from 0.57 to 0.69 in basanites, 0.41–0.53 in alkali
mations, the more differentiated mafic magmatic rocks with low basalts (Dautria et al., 1988), and in the Atakor massif from 0.50
Ni and Cr contents have been discarded. The Ni- and Cr-rich, to 0.56 in Neogene basanites. Samples with primitive Mg#, i.e.,
more magnesian, and least siliceous Miocene-Pliocene basanite 0.70 – 0.82, have not been identified anywhere, implying that none
magmas could have segregated from their mantle source at tem- of the mafic magmatic rocks represents unmodified partial melts
peratures of 1260–1270 ± 40 °C and pressures of 1.9–2.2 ± 0.27 of mantle peridotite.
GPa, corresponding to depths of ~55–65 ± 8 km, i.e., within the The Sr-Nd-Pb isotope systematics of the alkali basalts,
subcontinental lithospheric mantle. The temperature and pressure basanites, and nephelinites from the Hoggar alkaline volcanic
of segregation of the more primitive Quaternary basanite-tephrite province (Allègre et al., 1981) are consistent with their generation
magmas are estimated at 1250 ± 40 °C and 1.3 ± 0.27 GPa, cor- from a HIMU-type mantle source; they have 87Sr/86Sr ranging from
responding to a depth of 40 ± 8 km, close to the crust-mantle 0.70310 to 0.70348, εNd(t) varying from +4.5 to +6.9 ± 0.8, and high
boundary, but still within the subcontinental lithospheric mantle. Pb isotopic ratios (206Pb/204Pb up to 20.373 ± 0.014, and 207Pb/204Pb
These values are surprisingly low for silica-deficient mafic up to 15.696 ± 0.017 for a nephelinite). This range of composi-
lavas, which are commonly considered to be generated at con- tions, confirmed subsequently by Dupuy et al. (1993), shows that
siderably greater depths. In the nearby Tahalra massif (Dautria the source of the Neogene magmas is not strongly depleted. The
et al., 1988), low-Ni, low-Cr Miocene basanites are clearly dif- HIMU isotopic signature of the mafic magmatic rocks is consistent
ferentiated, but one primitive Miocene sample yields a depth of with their incompatible trace element patterns (Fig. 6).
105 ± 8 km. High-Ni, high-Cr Quaternary alkali basalts yield
comparable depths of 110 ± 8 km. Because these preliminary Episodic Nature of the Igneous Activity in
calculations indicate that the mafic magmas of the Atakor and the Hoggar Province
Tahalra massifs differ in terms of their calculated segregation
depths, a further 17 mafic rock compositions previously reported The Atakor massif, located on top of a major topographic
by Girod (1971) were also considered. These data are not directly swell, was active only during the more recent magmatic epi-
comparable with our data, because they were obtained by wet sodes of the Hoggar volcanic province. The first, Burdigalian,
chemical analysis and no Ni and Cr abundances were reported. episode in the Atakor massif occurred ca. 20 Ma, while the
For 14 of Girod’s samples, calculated depths vary from 35 to Hoggar province as a whole began to erupt at the Eocene-
80 ± 8 km, which is in agreement with our results. The other three Oligocene boundary, ca. 34 Ma. Though reliable isotopic age
high-Mg, low-Si compositions yield greater depths of 92–100 ± determinations are scarce, they provide a general framework
8 km. Girod (1971) noted that these three compositions corre- for considering the spatial and temporal variations in the char-
spond to Miocene basanites crowded with olivine phenocrysts acteristics of the Cenozoic volcanism.
(limburgitic facies). If the olivine phenocrysts were accumula- The first volcanic episode, located in the Anahef province,
tive, the resultant whole-rock compositions would have lower was marked by outpouring of a thick pile of flood basalt lavas in
SiO2 and higher MgO contents compared with the true parent the Taharaq district (Aït-Hamou, 2000). A highly eroded 300–
magmas, and thus the apparently greater depth of magma segre- 700-m-high plateau is composed of a sequence of 5–10-m-thick
gation could be an artifact. Consequently, we do not consider the lava flows. Its original area probably covered around 1500 km2,
calculated 92–100 km depths to be meaningful. and the total thickness of the eruptive sequence could have been
Thus, the Atakor Neogene mafic magmas indicate fairly as much as 1500 m. The volume of preserved flood basalts is
shallow depths of segregation compared to the Tahalra Quater- ~200 km3, although the original volume has been estimated at
nary basanite magmas (Dautria et al., 1988), which segregated at 1000 km3, and could have been as high as 2000 km3. Preliminary
~110 km. A possible explanation could be that the Atakor magmas K-Ar determinations have suggested a long-lasting volcanic epi-
have experienced some low-pressure differentiation, such that the sode from 44 ± 0.8–24.4 ± 0.5 Ma (Lutetian to Chattian), with a
calculated depths are not meaningful in terms of magma segrega- good correlation between the topographic elevation of the mag-
tion processes. However, the Ni and Cr contents of the samples matic rocks and their isotopic ages. The same samples when dated
selected for the calculations are in the range of 110–170 ppm and by the 40Ar-39Ar method yield identical, though less precise, ages,
220–280 ppm, respectively, which are similar to those of the Tahalra ranging from 34.5 ± 3.5 Ma to 32.8 ± 2.6 Ma, corresponding to
basanites. A second possibility could be that the Atakor magmas the Eocene-Oligocene boundary, i.e., from Priabonian to Rupelian
did actually segregate at shallower depths than the Tahalra mag- (Aït-Hamou et al., 2000). Younger K-Ar ages, from 28.5 ± 0.5 Ma
mas, suggesting that different magma generation processes operate to 24.4 ± 0.5 Ma are regarded, at least partly, as the result of the
within the subcontinental lithospheric upper mantle. In the Tazrouk thermal effects of intrusion of 29–24 Ma alkaline ring complexes
massif, located to the northeast of the Atakor massif, a basanite (Maza et al., 1998). The lavas are mainly high-K, high-P, high-Ti
has also indicated an intermediate depth of segregation of 80 ± (HKPT) tholeiitic basalts with subordinate amounts of alkaline
rocks ranging from alkali basalt to trachyandesite. No low-K, not known precisely, though it is likely to be Miocene and coeval
low-P, low-Ti (LKPT) basalts, characteristic of many continen- with the first volcanic episode in the Atakor massif.
tal flood basalt provinces, have been sampled so far. Using only In summary, the Hoggar volcanic province was the site of
the high-Ni (110–363 ppm), high-Cr (210–698 ppm) unaltered repeated volcanic episodes, separated by periods of quiescence.
mafic rocks, the calculated depths of segregation of the olivine According to currently available data, it is possible to distinguish
tholeiitic basalt magmas range from 73 to 38 ± 8 km. Cumulative six eruptive episodes:
rocks (ankaramites) yield a similar range of depths from 61 to 44 ± 1. The first episode is Oligocene in age, ca. 34–32 Ma, and is
8 km. The factors controlling the calculated segregation depths are characterized by the outpouring in the Taharaq district of a thick
mostly the Mg#, and to a lesser extent the Ni and Cr contents; the sequence of HKPT tholeiitic flood basalts, with a total volume
greatest depths correspond to a Mg# of 69 and the highest Ni and estimated at 1000–2000 km3.
Cr contents, while the shallowest depths correspond to Mg# of 2. The second episode is marked by emplacement at shallow
~50–55, with but varying abundances of Ni and Cr. Sr-Nd-Pb iso- depths (probably as low as 500 m) of tholeiitic to alkaline ring
tope systematics (Aït-Hamou et al., 2000) do not identify a single complexes, ca. 29 Ma, in the same area as the older flood basalts.
source for the magmatism, but rather a heterogeneous source vary- 3. The third episode corresponds to the emplacement, ca.
ing between HIMU and EM1 (enriched mantle) end-member char- 24 Ma, of silicic extrusions (trachyte and peralkaline rhyolite) on
acteristics, with 87Sr/86Sr from 0.703544 to 0.704519, 143Nd/144Nd top of unroofed alkaline ring complexes, indicating uplift rates of
from 0.512522 to 0.512796, 206Pb/204Pb from 18.361 to 19.499, and ~0.4 mm yr–1 within the Taharaq district.
207
Pb/204Pb from 15.580 to 15.624. 4. The different eruptive episodes in the Atakor massif have
The second magmatic episode is located nearly at the same occurred since 20 Ma. The first episode, actually the fourth epi-
place as the Taharaq district, and it consists of plutonic ring com- sode in the Hoggar, lasted from 20 to 12 Ma, with the outpouring
plexes and associated volcanic formations (Rémy, 1967; Maza, of alkaline flood basalts and associated intermediate pyroclastic
1998). Radial dike swarms also occur; these were emplaced at formations, and it ended with the emplacement of large phono-
shallow (~500 m) depths within the 34 Ma tholeiitic flood basalt lite domes and lava flows. The subsequent episode, i.e., the fifth
pile. The magmatic rocks are classified into a restricted suite of episode in the Hoggar, was shorter, between 6.7 and 4.2 Ma,
older tholeiitic olivine gabbros, yielding a K-Ar age of 29 ± 0.6 Ma with eruptions of a compositionally variable bimodal suite of
(late Rupelian), and a younger compositionally more variable alka- basanite to phonotephrite and benmoreite to trachyte, phono-
line suite of essexite, monzonite, and syenite. Extrusions of late lite, and scarce rhyolite. The last episode, i.e., the sixth episode
trachyte and peralkaline rhyolite yield a K-Ar age of 24 ± 0.4 Ma in the Hoggar, began 1.9 Ma and lasted until the present, with
(late Chattian). Field relationships between the unroofed plutonic exclusively basanite lava flows and associated scoria cones. The
ring complexes and the late extrusive volcanic deposits indicate Tahalra massif has also yielded Pliocene–Quaternary ages from
that between 29 and 24 Ma, the Anahef area experienced an uplift 3.5 Ma (unpublished data referred to in Aït-Hamou et al., 2000).
rate of ~0.4 mm yr–1. The Sr-Nd-Pb isotope systematics of the Since the Eocene-Oligocene boundary, igneous activity in
magmatic rocks (Maza et al., 1998) indicate a variety of sources, the Hoggar massif has not been continuous, but periodic. The six
with the tholeiitic plutonic rocks having isotopic compositions periods of igneous activity differ in their duration, ranging from
between those of the EM1 and HIMU mantle end members, while less than 1 m.y. (the oldest episodes) up to 7 m.y. (the Miocene
the alkaline plutonic rocks cluster near the HIMU end member, episode of Atakor). The five intermediate periods of quiescence
and the rhyolite plots in an intermediate position between EM1 and have durations of ~2–5 m.y.
HIMU. The available isotopic data suggest a shift from a hetero-
geneous mantle source in the Oligocene to a more homogeneous Magma-Generating Dynamic Processes
Miocene to Quaternary HIMU-like source.
Following this second magmatic episode, alkaline magma- In the mantle below the Hoggar swell, Neogene HIMU-
tism occurred throughout the Hoggar swell. North of the Taharaq like mantle sources appear to have replaced an Oligocene mixed
flood basalt province, the magmatism of the Serouenout district source, involving both HIMU and EM1 end members (Aït-Hamou
includes nephelinite, basanite, trachyte, and phonolite, with et al., 2000). Sr-Nd-Pb isotopic variations, coupled with the shift
a pronounced Daly gap between 42 and 58 wt% SiO2 (Maza, at 29 Ma from tholeiitic to alkaline igneous activity, are commonly
1998). The most primitive compositions yield magma segrega- explained by the involvement of a deep mantle plume. The popu-
tion depths of 96–100 ± 8 km for basanite and 139–183 ± 8 km lar “plume” model postulates that hot materials coming from the
for nephelinite. While the Mg# is the prime indicator of magma lower mantle impinged on the base of the lithosphere and induced
segregation depth for each magma type, for the same Mg#, partial melting of either the subcontinental lithospheric mantle, or
nephelinite magma was segregated significantly deeper than the interior of the plume itself, or a mixture of both. Allègre et al.
basanite magma, e.g., for a Mg# of 0.58, the calculated depths (1981) suggested that the HIMU source of OIBs originated from
are 139 and 96 km, respectively. These calculated magma segre- the lower mantle, whereas the heterogeneous sources for intra-
gation depths are by far the highest in the entire Hoggar volcanic continental flood basalts were located within the subcontinental
province. The age of the eruptions in the Serouenout district is lithospheric upper mantle. Aït-Hamou et al. (2000) adopted this
model, stating that the isotopic signature of the tholeiites suggests craton). The initiation of igneous activity could have been facili-
that they were derived from a heterogeneous mantle source domi- tated by edge-driven small-scale convection, induced by strong
nated by a lithospheric component near the EM1 end member, and lateral temperature gradients (King and Anderson, 1998).
that the subsequent alkali lavas required the involvement of a near Magma segregation depths along the major axis of the
end-member HIMU component of possible plume origin. Such a Hoggar swell are apparently unrelated to lithosphere thickness.
scenario can be challenged, and these authors admit that the nature In the southwest to northeast transect, the following values have
and origin of the two enriched mantle end members have yet to been calculated from available data for Neogene alkali basalts
be defined, as both components could be lithospheric, or, equally, and basanites: ~110 km in Tahalra, ~65–40 km in Atakor, ~80 km
could have a deeper origin. in Tazrouq, and ~100 km in Serouenout. Taharaq Oligocene
Alternatively, a scenario combining an EM1-like plume tholeiitic basalts yield depths of ~75–40 km, and Serouenout Mio-
component and a HIMU-like subcontinental lithospheric mantle cene (?) nephelinite magmas segregated at depths from 140 up to
source is tenable: the first, short-lived episode would imply fairly 180 km. With the exception of the Serouenout nephelinites, all
high degrees of partial melting (tholeiitic basalt magmas) of both the computed depths are shallower than the LATEA lithospheric
plume and lithosphere materials, and the subsequent, long-lasting thickness of 160 km, implying that the magmas segregated from
episodes would be marked by fairly low degrees of partial melt- mantle sources within the subcontinental lithosphere. Addition-
ing (basanite magmas) of the lithospheric mantle. Whether the ally, those volcanic districts emplaced within major Precambrian
HIMU sources are located in deep peridotitic mantle, or in garnet shear zones yield the lowest values.
pyroxenite-bearing lithospheric mantle, is a highly debated issue Periodic igneous activity in the Hoggar province is incon-
(e.g., Hirschmann et al., 2003; Niu and O’Hara, 2003). sistent with a single event of hot material intrusion at the Eocene-
Any discussion dealing with the plume concept is biased Oligocene boundary. No negative shear-velocity anomalies, diag-
because the term “plume” is frequently defined in a very loose nostic of thermal perturbations, have been detected above 250 km
manner. Volcanic provinces are considered as “hotspots” fuelled depth below the Tuareg shield (e.g., Figure 7E of Ritsema, 2005).
by mantle plumes coming from below, but the origin of the ther- The fairly shallow depths computed for magma segregation need
mal anomalies is a major issue for debate (e.g., Courtillot et al., to be reconciled with the apparently normal heat flow in the
2003; Anderson and Natland, 2005). In the original plume theory, Hoggar province (Lesquer et al., 1989) and seismic tomography
mantle plumes were inferred to come from great depths, possibly data indicating fast wave velocities (Shapiro and Ritzwoller, 2002).
down to the core-mantle boundary, plume heads were large-scale Ayadi et al. (2000) identified only weak velocity contrasts below
features (~1000 km in diameter), inferred to trigger continental Tahalra and Atakor. This paradox would seem to indicate that each
flood basalt magmatism, and smaller wavelength convective thermal event related to magma segregation, ascent, and emplace-
instabilities were not considered as real plumes. The two types of ment was short-lived enough to prevent the development of long-
upwelling materials were expected to sample deep and shallow lasting thermal anomalies in the mantle beneath the Hoggar swell.
mantle regions, respectively. Consequently, we suggest that the periodic igneous activ-
To define a mantle plume requires that several specific ity of the Hoggar is related to episodic fault activity. High fault
features be recognized. The major geochemical argument used stress and low lithospheric stiffness favor large variations of slip
currently to support a deep mantle origin is the occurrence in rates along fault zones (Chéry and Vernant, 2006). In an elas-
the lavas of high 3He/4He ratios, though this argument has been tic lithosphere loaded by plate motion, a low-loading velocity
questioned (Anderson, 2001). A compilation of new helium iso- causes fault-slip rates to cycle between high and low values,
tope results throughout Africa (Pik et al., 2006) indicates that the while a high-loading velocity leads to low stress and constant
magmatism of the Ethiopia-Afar province could originate from slip rates. The African and Eurasian continental plates have been
a large, deep-rooted mantle plume characterized by a high 3He converging since the Cretaceous. Periods of low-velocity plate
signature, possibly originating from the core-mantle boundary motion could, therefore, have promoted reactivation of remote
according to seismic tomography data. All other African intra- fault zones and triggered episodic igneous activity.
plate volcanic provinces, including the Miocene basanites of the The various volcanic districts in the Hoggar province, espe-
central Atakor massif, are more likely linked to a second-order cially the Atakor massif, are closely associated spatially with
type of shallow-mantle upwelling, presumably originating from Pan-African basement lineaments dissecting the Tuareg shield
depths shallower than 400 km, as suggested by seismic wave that were rejuvenated as brittle faults during the Paleozoic and
imaging (Sebai et al., 2006). the Mesozoic. The slip rates along those fault zones reactivated
In the Tuareg shield, the lithospheric thickness varies during the Cenozoic have yet to be determined. Liégeois et al.
significantly from west to east (Liégeois et al., 2005), i.e., (2005) argued that mantle decompression melting and resulting
from ~100 km west of 4°50E (corresponding to juvenile Pan- igneous activity could have taken place during local and short-
African terranes), to ~160 km between 4°50E and 8°30E (older lived episodes of fault reactivation along weakness zones, which
reworked terranes, including Laouni, Azrou-n-Fad, Tefedest, induced near-vertical lithosphere delamination (Fig. 9). Genera-
Egere-Aleksod, grouped into the LATEA microcontinent, Fig. 9), tion of OIB-like magmas is a function of the degree of partial
and to no more than 100 km east of 8°30E (East Saharan meta- melting, the size of the melting zone, and the length scales of
Cenozoic
Cambrian
Africa-Europe
“Taourirt”
granite convergence
province 4°50' E 100 km
LATEA
metacraton
Manzaz
ne
ra
er
en
it
ra
in
r
id
Atakor
te
Te
l
ke
In
Is
23°N Cambrian
Tahalra
topaz
LATEA crust granites
ITC TIC LC LC LC
LC (LC)
LC
35-0 Ma
160 km
50 km
LLM LLM
LATEA
lithospheric
LLM Africa-Europe
mantle
110 km (LLM) within-plate
re-activation
Asthenospheric
mantle
Figure 9. Delamination model for the Hoggar Cenozoic magmatism (after Liégeois et al., 2005).
mantle heterogeneity (Meibom and Anderson, 2004). Linear Eurasia convergence, with the formation of the Pyrenees-Provençe
delamination along shear zones can explain how melt segrega- mountain range, to a period dominated by Apulia-Eurasia con-
tion depths, and possibly even melt generation zones, can be vergence, leading to the formation of the Alpine mountain range,
shallower than expected in a regionally thick lithosphere. during which time the major plate boundary between Africa and
Eurasia was initiated, characterized by a northward-dipping sub-
Relations with the Evolution of the Western duction system. The African continent was still separated from the
Mediterranean Eurasian continent by remnants of subducting Neotethys oceanic
basins, and no high stress fields were applied to the continental pas-
The Hoggar massif, including the Atakor massif, remained sive margin. The period from 35 to 30 Ma was a key one (Jolivet
remarkably stable tectonically for a period of ~500 m.y., i.e., and Faccenna, 2000), because it marked the onset of full collision
since the Cambrian, when the 525 Ma topaz granites (Cheilletz between Africa and Eurasia and the inception of backarc exten-
et al., 1992) of the “Taourirt” granite province (Azzouni-Sekkal sion in the whole region. At 30 Ma, the subduction regime changed
et al., 2003) were emplaced (Fig. 2). The origin of Cenozoic fault from compressional to extensional in the whole Mediterranean
reactivation remains obscure if only local data are considered. region, as a consequence of continental collision and the slow-
Africa-Eurasia convergence since the Cretaceous has resulted ing down of the northward motion of Africa. Slow plate veloci-
in the formation of the Alpine orogenic belt through subduc- ties could have promoted reactivation of a specific Precambrian
tion, shortening, and uplift, and the formation of small oceanic fault zone in the Hoggar that formed the boundary between the
basins in the Western Mediterranean through microplate rotation, Serouenout and LATEA terranes. Periodic linear delamination of
extension, and subsidence. What has happened since the Eocene- the lithospheric mantle in the fault zone triggered, successively, the
Oligocene boundary in the Hoggar province is worth considering development at 34 Ma of a large volume of tholeiitic flood basalts,
in relation to global geodynamic events. the emplacement at 29 Ma of tholeiitic to alkaline ring complexes,
The Paleogene in the Hoggar is marked by three Oligocene and finally the extrusion at 24 Ma of silicic peralkaline volcanic
volcanic episodes at 34, 29, and 24 Ma. No Eocene igneous epi- rocks. The 29 Ma ring complexes illustrate the rapid shift (Maza
sodes are known so far. In the Mediterranean area (Fig. 1), Eocene et al., 1998) from tholeiitic magmas, derived by partial melting
times represent a transition from a period dominated by Iberia- of a heterogeneous mantle source (Aït-Hamou et al., 2000), to
alkaline magmas from a more homogeneous HIMU-like source, Pliocene–Quaternary basanite-tephrite association. Peridotite man-
suggesting that any thermal anomaly associated with the origin of tle xenoliths and megacrysts (kaersutite, mica, diopside, zircon,
the 34 Ma flood basalts must have disappeared progressively in a oxides, etc.) indicate that magmatic differentiation processes must
5 m.y. period of time. have operated within the upper mantle up to the crust-mantle
The Neogene was marked by the closure of the last Neo- boundary. The felsic group, separated by a Daly gap from the mafic
tethys basins between Africa and Eurasia, which was completed lavas, is compositionally more variable with two diverging trends,
by 21 Ma (Rosenbaum et al., 2002). Since then, the Atakor mas- a silica-saturated to oversaturated benmoreite-trachyte-rhyolite
sif experienced three volcanic episodes, separated by long-lasting trend with metaluminous, peralkaline, and scarce peraluminous
periods of quiescence. These episodes reflect the complex evolu- varieties, and a silica-undersaturated trachyte-phonolite trend with
tion of the Western Mediterranean, with the formation of three metaluminous (miaskitic) and peralkaline (agpaitic) varieties. The
backarc basins. The first volcanic episode in the Atakor massif more evolved magma compositions indicate magmatic differentia-
is nearly coeval with extensional events initiating, successively, tion involving feldspars, apatite, zircon, and Fe-Ti oxides.
the 25–23 Ma Valencia trough between Iberia and the Balearic The depths of melt segregation of the most primitive mafic
Islands, the 21–18 Ma oceanic Algero-Provençal Basin, and the magmas in the Hoggar vary from ~110 km in Tahalra to ~65–
18–10 Ma Alboran Sea. The following period of quiescence in 40 km in Atakor, with intermediate values of ~75–40 km (Taha-
the Atakor massif was marked by extensive igneous activity raq), ~80 km (Tazrouq), and ~100 km (Serouenout). These cal-
within the Western Mediterranean area, especially in the Rif-Tell culated depths are shallower than the LATEA (corresponding to
fold belts of the Maghreb (Maury et al., 2000, and references the Laouni, Azrou-n-Fad, Tefedest and Egere-Aleksod terranes)
therein), where discrete igneous episodes correspond to slab lithospheric thickness of 160 km. These apparently contradictory
breakoff and abrupt changes in the kinematics of Africa-Eurasia results can be reconciled if we consider the possibility that mag-
plate convergence (Rosenbaum et al., 2002). It is remarkable that matism was triggered by lithospheric delamination along linear
the major tectonic episodes in the Rif-Tell fold belt were appar- shear zones, allowing the ascent and decompression partial melt-
ently not accompanied by volcanic activity in the Atakor massif. ing of the asthenospheric mantle. The extent of asthenospheric
More isotopic age determinations are necessary to substantiate ascent was more or less significant according to the magmatic
whether this period of time was really quiescent, or not. The sec- period and the location of the igneous activity.
ond volcanic episode in the Atakor massif was coeval with the The episodic nature of the volcanic activity can be related
extensional event that created the northern part of the Tyrrhenian to episodic large-scale tectonic events occurring at the Africa-
Sea, floored by thinned continental crust, and isolated Corsica Eurasia convergence boundary in the Western Mediterranean.
from the central Italian mainland. The following period of qui- The discrete magmatic episodes were likely triggered by a drastic
escence in the Atakor massif ended with the inception of the two slowing down in the motion of the African plate, which induced
deep oceanic basins in the southern part of the Tyrrhenian Sea, major changes in the stress regime and caused reactivation of
isolating Sardinia from southern Italian mainland and Sicily. pre-existing Precambrian faults.
To sum up, episodes of igneous activity in the Hoggar swell
appear to have been nearly coeval with extensional events in the ACKNOWLEDGMENTS
Western Mediterranean, inducing relaxation of boundary stress
fields caused by low-plate-loading velocities, while periods of We are grateful to the editors of this volume, who prompted
quiescence could correspond to compressional events, inducing us to produce this paper based on new results collected on a
a high-stress regime and a high-plate-loading velocity. region apparently remote, but actually related to the formation of
the Mediterranean Sea. Field and laboratory studies were funded
SUMMARY AND CONCLUSIONS partially by the four-year Accord-Programme 94 MDU 282
granted by the “Comité Mixte d’Évaluation et de Prospective de
The Atakor massif is located at the center of the Hoggar la Coopération Inter-Universitaire Franco-Algérienne” (CMEP).
swell within the Tuareg shield. It is characterized by volcanic The Algerian co-authors gratefully acknowledge the logistic sup-
activity volcanic during three discrete Neogene igneous episodes, port provided by the “Centre International des Etudiants et Sta-
separated by fairly long periods of quiescence. The topographic giaires” (CIES) during their stays in France. Many thanks are
highs are composed of Precambrian basement uplifted since the due to the support of the director of the Centre de Recherches en
Pliocene and of overlying lava plateaus, domes and spines, scoria Astronomie Astrophysique et Géophysique in Tamanrasset. J.M.
cones, and valley-filling lava flows. The mantle source of the Dautria kindly provided analytical materials on volcanic districts
magmas has a dominant HIMU-like characteristic, with a K-poor other than the Atakor massif. Discussions with Laurent Jolivet
signature and significant incompatible element enrichment show- and Gideon Rosenbaum were helpful to clarify our ideas on the
ing LILE to HFSE fractionation. Cenozoic evolution of Western Mediterranean. Finally, thanks
The magmatic rocks have been subdivided into two groups are due to Jean-Marie Dautria and two anonymous reviewers for
with contrasting compositions. The mafic group includes a their helpful comments, as well as to the editors for their patience
Miocene–Pliocene basanite-phonotephrite trend and a subsequent and efficient editorial handling of the manuscript.
REFERENCES CITED Chéry, J., and Vernant, P., 2006, Lithospheric elasticity promotes episodic fault
activity: Earth and Planetary Science Letters, v. 243, p. 211–217, doi:
10.1016/j.epsl.2005.12.014.
Aït Hamou, F., 2000. Un exemple de “point chaud” intra-continental en con- Conquéré, F., and Girod, M., 1968, Contribution à l’étude des paragenèses des
texte de plaque quasi-stationnaire: Étude pétrologique et géochimique basaltes alcalins: Les spinelles du volcan de l’Oued Temorte (Massif de
du Djebel Taharaq et évolution du volcanisme Cénozoïque de l’Ahaggar l’Atakor, Sahara Algérien): Contributions to Mineralogy and Petrology,
(Sahara Algérien) [Thèse de Doctorat en Sciences]: Montpellier, Univer- v. 20, p. 1–29, doi: 10.1007/BF00371063.
sité Montpellier II–Sciences et Techniques du Languedoc, 200 p. Courtillot, V., Davaille, A., Besse, J., and Stock, J., 2003, Three distinct types
Aït-Hamou, F., Dautria, J.M., Cantagrel, J.M., Dostal, J., and Briqueu, L., 2000, of hotspots in the Earth’s mantle: Earth and Planetary Science Letters,
Nouvelles données géochronologiques et isotopiques sur le volcanisme v. 205, p. 295–308, doi: 10.1016/S0012-821X(02)01048-8.
Cénozoïque de l’Ahaggar (Sahara Algérien): Des arguments en faveur Dautria, J.M., Liotard, J.M., Cabanes, N., Girod, M., and Briqueu, L., 1987,
d’un panache: Comptes Rendus de l’Académie des Sciences de Paris, Amphibole-rich xenoliths and host alkali basalts: Petrogenetic constraints
v. 330, p. 829–836. and implications on the recent evolution of the upper mantle beneath
Albarède, F., 1992, How deep do common basaltic magmas form and differenti- Ahaggar (central Sahara, southern Algeria): Contributions to Mineralogy
ate?: Journal of Geophysical Research, v. 97, p. 10,997–11,009. and Petrology, v. 95, p. 133–144, doi: 10.1007/BF00381263.
Allègre, C.J., Dupré, B., Lambert, B., and Richard, P., 1981, The subcontinental Dautria, J.M., Dostal, J., Dupuy, C., and Liotard, J.M., 1988, Geochemistry and
versus suboceanic debate. Lead-neodymium-strontium isotopes in primary petrogenesis of alkali basalts from Tahalra (Hoggar, northwest Africa):
basalts in a shield area: the Ahaggar volcanic suite: Earth and Planetary Chemical Geology, v. 69, p. 17–35, doi: 10.1016/0009-2541(88)90155-6.
Science Letters, v. 52, p. 85–92, doi: 10.1016/0012-821X(81)90210-7. Duggan, M.B., 1990, Wilkinsonite, Na2Fe2+4Fe3+2Si6O20, a new member of the
Anderson, D.L., 2001, A statistical test of the two reservoir model for helium aenigmatite group from the Warrumbungle volcano, New South Wales,
isotopes: Earth and Planetary Science Letters, v. 193, p. 77–82, doi: Australia: American Mineralogist, v. 75, p. 694–701.
10.1016/S0012-821X(01)00507-6. Dupuy, C., Chikhaoui, M., and Dostal, J., 1993, Trace element and isotopic
Anderson, D.L., 2005, Scoring hotspots: The plume and plate paradigms, in geochemistry of Cenozoic alkali basaltic lavas from the Atakor (central
Foulger, G.R., Natland, J.H., Presnall, D.C., and Anderson, D.L., eds., Sahara): Geochemical Journal, v. 27, p. 131–145.
Plates, Plumes and Paradigms: Geological Society of America Special Eby, G.N., 1992, Chemical subdivision of the A-type granitoids: Petrogenetic
Paper 388, p. 31–54. and tectonic implications: Geology, v. 20, p. 641–644, doi: 10.1130/0091-
Anderson, D.L., and Natland, J.H., 2005. A brief history of the plume hypothesis 7613(1992)020<0641:CSOTAT>2.3.CO;2.
and its competitors: Concepts and controversy, in Foulger, G.R., Natland, Fabre, J., (2005). Géologie du Sahara occidental et central. Série/Reeks: Ter-
J.H., Presnall, D.C., and Anderson, D.L., eds., Plates, Plumes and Para- vuren African Geosciences Collection, MRAC Tervuren, Belgique,
digms: Geological Society of America Special Paper 388, p. 119–145. 572 p.
Ayadi, A., Dorbath, C., Lesquer, A., and Bezzeghoud, M., 2000, Crustal and Fitton, J.G., 1987. The Cameroon line, West Africa: A comparison between
upper mantle velocity structure of the Hoggar swell (central Sahara, oceanic and continental alkaline volcanism, in Fitton, J.G., and Upton,
Algeria): Physics of the Earth and Planetary Interiors, v. 118, p. 111–123, B.G.J., eds., Alkaline Igneous Rocks: Geological Society of London Spe-
doi: 10.1016/S0031-9201(99)00134-X. cial Publication 30, p. 273–291.
Azambre, B., and Girod, M., 1966, Phonolites agpaïtiques: Bulletin de la Société Frizon de Lamotte, D., 2006, Some recent developments on the Maghreb geo-
Française de Minéralogie et de Cristallographie, v. 90, p. 514–520. dynamics: Comptes Rendus Geoscience, v. 338, p. 1–10, doi: 10.1016/
Azzouni-Sekkal, A., Liégeois, J.P., Bechiri-Benmerzoug, F., Belaidi-Zinet, S., j.crte.2005.11.006.
and Bonin, B., 2003, The “Taourirt” magmatic province, a marker of the Girod, M., 1971, Le massif volcanique de l’Atakor (Hoggar, Sahara Algérien).
closing stages of the Pan-African orogeny in the Tuareg shield: Review Étude pétrographique, structurale et volcanologique: Paris, Centre de
of the available data and Sr-Nd isotope evidence: Journal of African Earth Recherches sur les Zones Arides, série Géologie, no. 12, Éditions du Cen-
Sciences, v. 37, p. 331–350, doi: 10.1016/j.jafrearsci.2003.07.001. tre National de la Recherche Scientifique, 158 p.
Benhallou, A.Z., 2000, Étude du volcanisme alcalin Cénozoïque de la région de Gradstein, F.M., Ogg, J.G., Smith, A.G., Bleeker, W., and Lourens, L.J., 2004,
Tahifet (Bloc d’Azrou N’fad, Hoggar central, Algérie) [Magister thesis]: A new geologic time scale, with special reference to Precambrian and
Algiers, USTHB-FSTGAT, 164 p. Neogene: Episodes, v. 27, p. 83–100.
Bigi, G., Castellarin, R., Catalano, R., Coli, M., Cosentino, D., Dal Piaz, G.V., Grandjean, A., 1962, Les séismes au Hoggar: Travaux de l’Institut des Recher-
Lentini, F., Parotto, M., Patacca, E., Praturlon, A., Salvini, F., Sartori, R., ches Sahariennes, v. 21, p. 187.
Scandone, P., and Vai, G.B., 1989, Synthetic structural-kinematic map of Hirschmann, M.M., Kogiso, T., Baker, M.B., and Stolper, E.M., 2003,
Italy: Rome, Progetto Finilizzato Geodinamica, CNR, scale 1:2,000,000. Alkalic magmas generated by partial melting of garnet pyroxenite:
Black, R., and Girod, M., 1970, Late Paleozoic to recent igneous activity in Geology, v. 31, p. 481–484, doi: 10.1130/0091-7613(2003)031<0481:
West Africa and its relationship to basement structure, in Clifford, T.N., AMGBPM>2.0.CO;2.
and Gass, I.G., eds., African Magmatism and Tectonics: Edinburgh, Jolivet, L., and Faccenna, C., 2000, Mediterranean extension and the
Oliver & Boyd, p. 185–210. Africa-Eurasia collision: Tectonics, v. 19, p. 1095–1106, doi: 10.1029/
Black, R., Jaujou, M., and Pellaton, C., 1967, Notice explicative sur la carte 2000TC900018.
géologique de l’Aïr, à l’échelle du 1/500,000: Direction des Mines et de la Jolivet, L., Faccenna, C., D’Agostino, N., Fournier, M., and Worrall, D., 1999,
Géologie, République du Niger; Paris, Editions du BRGM, 57 p. The kinematics of back-arc basins, examples from the Tyrrhenian, Aegean
Black, R., Latouche, L., Liégeois, J.P., Caby, R., and Bertrand, J.M., 1994, and Japan Seas, in Mac Niocaill, C., and Ryan, P.D., eds., Continental Tec-
Pan-African displaced terranes in the Tuareg shield (central Sahara): tonics: Geological Society of London Special Publication 164, p. 21–53.
Geology, v. 22, p. 641–644, doi: 10.1130/0091-7613(1994)022<0641: King, S.D., and Anderson, D.L., 1998, Edge-driven convection: Earth and
PADTIT>2.3.CO;2. Planetary Science Letters, v. 160, p. 289–296, doi: 10.1016/S0012-
Bonin, B., 2004, Do coeval mafic and felsic magmas in post-collisional 821X(98)00089-2.
to within-plate regimes necessarily imply two contrasting, mantle Le Bas, M.J., Le Maitre, R.W., Streckeisen, A., and Zanettin, B., 1986, A
and crustal, sources? A review: Lithos, v. 78, p. 1–24, doi: 10.1016/ chemical classification of volcanic rocks based on the total alkali–silica
j.lithos.2004.04.042. diagram: Journal of Petrology, v. 27, p. 745–750.
Burke, K., 1996, The African plate: South African Journal of Geology, v. 99, Le Bas, M.J., Le Maitre, R.W., and Woolley, A.R., 1992, The construction of
p. 339–410. the total alkali–silica chemical classification of volcanic rocks: Mineral-
Cheilletz, A., Bertrand, J.M.L., Charoy, B., Moulahoum, O., Bouabsa, L., ogy and Petrology, v. 46, p. 1–22, doi: 10.1007/BF01160698.
Farrar, E., Zimmerman, J.L., Dautel, D., Archibald, D.A., and Boul- Le Maitre, R.W., ed., 1989, A Classification of Igneous Rocks and Glossary
lier, A.M., 1992, Géochimie et géochronologie Rb-Sr, K-Ar et 40Ar/39Ar of Terms: Recommendations of the International Union of Geosciences
des complexes granitiques Panafricains de la région de Tamanrasset Subcommission on the Systematics of Igneous Rocks: Oxford, Blackwell
(Algérie): Relations avec les minéralisations Sn-W associées et l’évolution Scientific Publications, 193 p.
tectonique du Hoggar central: Bulletin de la Société Géologique de Le Maitre, R.W., ed., 2002. Igneous Rocks: A Classification and Glossary of
France, v. 163, p. 733–750. Terms (2nd edition): Recommendations of the International Union of
Geosciences Subcommission on the Systematics of Igneous Rocks: Cam- Rognon, P., 1967, Le massif de l’Atakor et ses bordures (Sahara central). Étude
bridge, Cambridge University Press, 236 p. géomorphologique: Paris, Éditions du Centre National de la Recherche
Lesquer, A., Bourmatte, A., Ly, S., and Dautria, J.M., 1989, First heat flow Scientifique, 555 p.
determination from the central Sahara: Relationship with the Pan-African Rognon, P., Gourinard, Y., and Bandet, Y., 1981, Un épisode de climat aride
belt and Hoggar domal uplift: Journal of African Earth Sciences, v. 9, dans l’Atakor (Hoggar) vers 1.5 Ma (datations K/Ar) et sa place dans
p. 41–48, doi: 10.1016/0899-5362(89)90006-7. le contexte paléoclimatique du Plio-Pléistocène Africain: Bulletin de la
Liégeois, J.P., Benhallou, A., Azzouni-Sekkal, A., Yahiaoui, R., and Bonin, B., Société Géologique de France, v. XXIII (7° série), p. 313–318.
2005. The Hoggar swell and volcanism: Reactivation of the Precambrian Rognon, P., Gourinard, Y., Bandet, Y., Koeniguer, J.C., and Delteil-Desneux,
Tuareg shield during Alpine convergence and West African Cenozoic vol- F., 1983, Précisions chronologiques sur l’évolution volcano-tectonique
canism, in Foulger, G.R., Natland, J.H., Presnall, D.C., and Anderson, et géochronologique de l’Atakor (Hoggar): Apports des données radio-
D.L., eds., Plates, Plumes and Paradigms: Geological Society of America métriques (K/Ar) et paléoclimatique du Plio-Pleistocène Africain: Bul-
Special Paper 388, p. 379–400. letin de la Société Géologique de France, v. XXV (7° série), p. 973–980.
Lustrino, M., 2000, Volcanic activity from the Neogene to the present evolution Rosenbaum, G., Lister, G.S., and Duboz, C., 2002, Reconstruction of the tec-
of the Western Mediterranean area: A review: Ofioliti, v. 25, p. 87–101. tonic evolution of the Western Mediterranean since the Oligocene: Jour-
Maury, R.C., Fourcade, S., Coulon, C., El Azzouzi, M’H., Bellon, H., Coutelle, nal of the Virtual Explorer, v. 8, p. 105–124.
A., Ouabadi, A., Semroud, B., Megartsi, M’H., Cotten, J., Belanteur, O., Saadi, M., 1982, Shéma structural du Maroc: Notes et Mémoires du Service
Louni-Hacini, A., Piqué, A., Capdevila, R., Hernandez, J., and Rehault, Géologique du Maroc, v. 278B, attached color map, scale 1:4,000,000.
J.P., 2000, Post-collisional Neogene magmatism in the Mediterranean Savelli, C., 2002, Time-space distribution of magmatic activity in the Western
Maghreb margin: A consequence of slab-breakoff: Comptes Rendus de Mediterranean and peripheral orogens during the past 30 Ma (a stimulus
l’Académie des Sciences, Paris, v. 331, p. 159–173. to geodynamic considerations): Journal of Geodynamics, v. 34, p. 99–
Maza, M., 1998, Transition magmatisme tholéïitique-alcalin en contexte 126, doi: 10.1016/S0264-3707(02)00026-1.
intra-continental: Exemple du “point chaud” de l’Ahaggar. Les com- Sebai, A., Stutzmann, E., Montagner, J.P., Sicilia, D., and Beucler, E., 2006,
plexes volcano-plutoniques annulaires du Sud-Amadghor et le district Anisotropic structure of the African upper mantle from Rayleigh and
de Serouenout (Ahaggar central, Algérie) [Thèse de doctorat en Sci- Love wave tomography: Physics of the Earth and Planetary Interiors,
ences]: Montpellier, Université Montpellier II–Sciences et Techniques du v. 155, p. 48–62, doi: 10.1016/j.pepi.2005.09.009.
Languedoc, 200 p. Sébrier, M., Siame, L., Zouine, M., Winter, T., Leturmy, P., and Missenard,
Maza, M., Briqueu, L., Dautria, J.M., and Bosch, D., 1998, Le complexe annu- Y., 2006, Active tectonics in the Moroccan High Atlas: Comptes Rendus
laire d’âge Oligocène de l’Achkal (Hoggar central, sud Algérie): Témoin Geoscience, v. 338, p. 65–79, doi: 10.1016/j.crte.2005.12.001.
de la transition au Cénozoïque entre les magmatismes tholéitique et Shapiro, N.M., and Ritzwoller, M.H., 2002, Monte-Carlo inversion for a
alcalin. Évidences par les isotopes du Sr, Nd et Pb: Comptes Rendus Geo- global shear velocity model of the crust and upper mantle: Geophysi-
science, v. 327, p. 167–172. cal Journal International, v. 151, p. 88–105, doi: 10.1046/j.1365-
McDonough, W.F., and Sun, S.-s., 1995, The composition of the Earth: Chemi- 246X.2002.01742.x.
cal Geology, v. 120, p. 223–235, doi: 10.1016/0009-2541(94)00140-4. Sleep, N.H., 1990, Hot spots and mantle plumes: Some phenomenology: Jour-
McDonough, W.F., Sun, S.-s., Ringwood, A.E., Jagoutz, E., and Hofmann, nal of Geophysical Research, v. 95, p. 6715–6736.
A.W., 1992, Potassium, rubidium, and cesium in the Earth and Moon and Sun, S.-s., and McDonough, W.F., 1989, Chemical and isotopic systematics of
the evolution of the mantle of the Earth: Geochimica et Cosmochimica oceanic basalts: Implications for mantle composition and processes, in
Acta, v. 56, p. 1001–1012, doi: 10.1016/0016-7037(92)90043-I. Saunders, A.D., and Norry, M.J., eds., Magmatism in the Ocean Basins:
Meibom, A., and Anderson, D.L., 2004, The statistical upper mantle assem- Geological Society of London Special Publication 42, p. 313–345.
blage: Earth and Planetary Science Letters, v. 217, p. 123–139, doi: Vai, G.B., 2001, Structure and stratigraphy: An overview, in Vai, G.B., and
10.1016/S0012-821X(03)00573-9. Martini, I.P., eds., Anatomy of an Orogen: The Apennines and Adjacent
Niu, Y., and O’Hara, M.J., 2003, Origin of ocean island basalts: A new per- Mediterranean Basins: Bodmin, Cornwall, Great Britain, Kluwer Aca-
spective from petrology, geochemistry, and mineral physics consider- demic Publishers, p. 15–32.
ations: Journal of Geophysical Research, v. 108, p. 2209–2228, doi: Weaver, B.L., 1991a, Trace element evidence for the origin of ocean-
10.1029/2002JB002048. island basalts: Geology, v. 19, p. 123–126, doi: 10.1130/0091-7613
Pearce, J.A., Harris, N.B.W., and Tindle, A.G., 1984, Trace element discrimina- (1991)019<0123:TEEFTO>2.3.CO;2.
tion diagrams for the tectonic interpretation of granitic rocks: Journal of Weaver, B.L., 1991b, The origin of ocean island basalt end-member composi-
Petrology, v. 25, p. 956–983. tions: Trace element and isotopic constraints: Earth and Planetary Science
Pik, R., Marty, B., and Hilton, D.R., 2006, How many plumes in Africa? The Letters, v. 104, p. 381–397, doi: 10.1016/0012-821X(91)90217-6.
geochemical point of view: Chemical Geology, v. 226, p. 100–114, doi: Wilson, M., and Downes, H., 1991, Tertiary–Quaternary extension-related
10.1016/j.chemgeo.2005.09.016. alkaline magmatism in western and central Europe: Journal of Petrology,
Rémy, J.M., 1967, Étude géologique et volcanologique du sud-est de l’Amadror v. 32, p. 811–849.
en Ahaggar (Sahara central): Nancy, Sciences de la Terre, Mémoire 10, Yahiaoui, R., 2003, Étude du volcanisme alcalin de la région de l’Assekrem
192 p. (Hoggar central), étude pétrographique, minéralogique, géochimique et
Ritsema, J., 2005, Global seismic structure maps, in Foulger, G.R., Natland, géobarométrique [Magister thesis]: Algiers, USTHB-FSTGAT, 164 p.
J.H., Presnall, D.C., and Anderson, D.L., eds., Plates, Plumes and Para-
digms: Geological Society of America Special Paper 388, p. 11–18. MANUSCRIPT ACCEPTED BY THE SOCIETY 9 JUNE 2006
Printed in the USA
View publication stats

SPE418 16atakor

Uploaded by

Copyright:

Available Formats

SPE418 16atakor

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SPE418 16atakor

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Tertiary alkaline volcanism of the Atakor Massif (Hoggar, Algeria): ﬁeld,

Article · January 2007

Bernard Bonin Amel Benhallou

SEE PROFILE SEE PROFILE

R. Yahiaoui J.-P. Liégeois

SEE PROFILE SEE PROFILE

is the silet terrane a puzzle View project

The user has requested enhancement of the downloaded file.

Tertiary alkaline volcanism of the Atakor massif, Hoggar, Algeria

Keywords: Hoggar (Algeria), volcanism, alkali basalt, phonolite-trachyte, upper mantle.

GB= G rande Kabylie

Figure 1. Cenozoic volcanism in NW Africa and the Western Mediterranean. Map is

Atakor volcanic massif

Mineral Chemistry Megacrysts

Common Assemblage of Major Rock-Forming Minerals

Well-aligned alkali feldspar phenocrysts and microcrysts

Both peralkaline and peraluminous felsic rock types com-

Normalized to primitive man tle

Neogene basanites and phonotephrites

Quaternary basanites and phonotephrites

Normalized to primitive man tle

Atakor trachytes and rhyolites

Upper Mantle Sources of the Cenozoic Magmas

Printed in the USA

View publication stats

You might also like