This article was published in an Elsevier journal. The attached copy
is furnished to the author for non-commercial research and
education use, including for instruction at the author’s institution,
sharing with colleagues and providing to institution administration.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Author's personal copy
Rift propagation at craton margin. Distribution of faulting and volcanism in
the North Tanzanian Divergence (East Africa) during Neogene times
B. Le Galla,⁎, P. Nonnottea, J. Roleta, M. Benoita, H. Guilloub,
M. Mousseau-Nonnottea, J. Albarica, J. Deverchèrea
aInstitut Universitaire Européen de la Mer, UMR 6538 UBO/CNRS, Place Nicolas Copernic, 29280, Plouzané, France
bCEA–CNRS, UMR 1572, LSCE, Domaine du CNRS Bat. 12, Avenue de la Terrasse, 91198 Gif sur Yvette Cedex, France
Received 4 June 2007; received in revised form 30 October 2007; accepted 5 November 2007
Available online 17 November 2007
A revised kinematic model is proposed for the Neogene tectono-magmatic development of the North Tanzanian Divergence where the axial
valley in S Kenya splits southwards into a wide diverging pattern of block faulting in association with the disappearance of volcanism.
Propagation of rifting along the S Kenya proto-rift during the last 8 Ma is first assumed to have operated by linkage of discrete magmatic cells as
far S as the Ngorongoro–Kilimanjaro transverse volcanic belt that follows the margin of cratonic blocks in N Tanzania. Strain is believed to have
nucleated throughout the thermally-weakened lithosphere in the transverse volcanic belt that might have later linked the S Kenya and N Tanzania
rift segments with marked structural changes along-strike. The North Tanzanian Divergence is now regarded as a two-armed rift pattern involving:
(1) a wide domain of tilted fault blocks to the W (Mbulu) that encompasses the Eyasi and Manyara fault systems, in direct continuation with the
Natron northern trough. The reactivation of basement fabrics in the cold and intact Precambrian lithosphere in the Mbulu domain resulted in an
oblique rift pattern that contrasts with the orthogonal extension that prevailed in the Magadi–Natron trough above a more attenuated lithosphere.
(2) To the E, the Pangani horst-like range is thought to be a younger (b1 Ma) structure that formed in response to the relocation of extension S of
the Kilimanjaro magmatic center. A significant contrast in the mechanical behaviour of the stretched lithosphere in the North Tanzanian diverging
rift is assumed to have occurred on both sides of the Masai cratonic block with a mid-crustal decoupling level to the W where asymmetrical fault-
basin patterns are dominant (Magadi–Natron and Mbulu), whereas a component of dynamical uplift is suspected to have caused the topographic
elevation of the Pangani range in relation with possible far-travelled mantle melts produced at depth further N.
© 2007 Elsevier B.V. All rights reserved.
Keywords: East Africa rift; North Tanzania; Neogene; Strain localization; Magmatism; Craton
Dynamics of strain propagation is influenced by numerous
physical parameters which, according to their supposed
respective role, lead to varying kinematic models. In continental
or oceanic extensional settings, rift propagation is usually
considered to operate (1) almost simultaneously along the entire
length of the extended domain (Le Pichon and Gaulier, 1988;
Fournier et al., 2004), (2) as a V-shaped lithospheric crack
propagating and widening throughout the undeformed litho-
sphere (Cochran, 1981; Courtillot, 1982), (3) along regular rift
segments bounded either by transform faults (Manighetti et al.,
1997) or accommodation zones (Bosworth, 1985), and (4) by
lateral connection of initial isolated zones (Bonatti, 1985;
Nicolas et al., 1994). In the reference East African Rift System
that connects northwards via the Afar triple junction with the
above-mentioned Red Sea and Gulf of Aden oceanic spreading
systems (Fig. 1), strain propagation is generally thought to have
migrated continuously southwards with time from the Afar
Triangle at around 30 Ma to the S Kenya rift, 2000 km further S,
at about 5–8 Ma (Cerling and Powers, 1977; Crossley and
Available online at www.sciencedirect.com
Tectonophysics 448 (2008) 1–19
⁎Corresponding author. Tel.: +33 2 98 49 87 56; fax: +33 2 98 49 87 60.
E-mail address: email@example.com (B. Le Gall).
0040-1951/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
Author's personal copy
Knight, 1981). However, a more complex rift kinematics is
suggested by the presence of discrete initial depocenters, as old
as 35 Ma, along the Ethiopian and N Kenyan rifts (Morley et al.,
1992), and probably Lower Miocene in age in Central Kenya
(Mugisha et al., 1997; Hautot et al., 2000), that might have
subsequently linked at a more mature rift stage (Vétel and Le
Gall, 2006). For any rift system, the more appropriate area to
fully discuss kinematics of strain propagation is the frontal part
where young structures are generally still preserved with their
original arrangement, and are thus prone to constrain the spatial/
temporal evolution of incipient rift structures throughout the
intact crust/lithosphere. The present paper is focused on the
frontal structures in the Eastern branch of the East African Rift
System, at 2–3° S close to the Kenya-Tanzania border (Fig. 1).
There, rift structures are expressed at the surface by abrupt
changes in crustal extension that result in (1) the widening of the
extended zone from 50–60 km in the S Kenya axial valley to
200 km further S throughout the diverging arms of the North
Tanzanian Divergence (NTD in the text) (Dawson, 1992), and
(2) the rapid disappearance of plume-related synrift magmatism
S of the 200 km-long Ngorongoro–Kilimanjaro transverse
volcanic belt (NKVB hereafter) (Fig. 1). Since its initiation at c.
8 Ma (Dawson, 1992), the NTD Neogene volcanic province has
recorded several episodes of faulting, but no clear time
framework has been so far established for its tectonic and
magmatic development. It is the main goal of our study to
investigate the spatial and timing relationships between tectonic
and magmatic processes during the Neogene (b8 Ma) rift
evolution of the NTD in order to apply a model of rift propa-
gation at the southern extremity of the eastern branch of the East
African Rift System. Our work is primarily based on (1) com-
pilation of existing radiometric dataset, completed by a few K/
Ar age dating results obtained here on selected volcanic rocks,
and (2) the structural analysis of major rift fault network
Fig. 1. Main structural and magmatic features in the S Kenya and N Tanzania rift system. Structures are mainly extracted from the digital elevation model obtained by
Shuttle Radar Topographic Mission (SRTM) data (see Fig. 4A). ASWA SZ., ASWA shear zone; volcanoes: B. Burko; Em., Embagai; Es., Essimingor; G., Gelai; H.,
Hanang; K., Kerimasi; Ke., Ketumbeine; Ki., Kibo; Kw., Kwaraha; L., Lemagrut; M. Monduli; Ma., Mawenzi; Me., Meru; Ng., Ngorongoro; OS., Ol Donyo Sambu;
S., Shira; Sh., Shompole; T. Tarosero.
2 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
extracted from remote sensing data. The resulting 2D-map
arrangement of both tectonic and dated volcanic structures in
the diverging rift pattern leads us to unravel the chronology of
its structural development as approaching the northern margin
of the Tanzanian craton. Emphasis is put on the key-role played
by along-axis jumps and lateral shift of magmatism on the
distribution and propagation of strain. From the surface rift
pattern, a number of assumptions are also made about deeper
rifting processes in relation with structural inheritance and the
various types of crust–lithosphere present in the NTD.
2. General rift setting in the NTD
2.1. Geological context
The S Kenya–N Tanzania rifted area under study shows
pronounced changes in the surface expression of rifting in
coincidence with fundamental variations of the crust/lithosphere
structural pattern at depth. Three distinct domains are
distinguished from N to S on the sketch structural map of Fig. 1.
(1) The Magadi–Natron rift system is a NS-oriented depres-
sion, 50–80 km-wide, occupied by Late Miocene–
Present volcanics directly overlying metamorphic base-
ment rocks of the Proterozoic Belt (Baker et al., 1971;
Fairhead et al., 1972). Its overall structure evolves
southwards from an asymmetrical graben basin to a E-
facing half-graben bounding to the W by a double system
of normal faults (Ol Donyo Ogol and Natron master
faults). The S Kenya rift valleyshows an important micro-
seismic activity associated with geothermal processes
(Maguire and Long, 1976). From geophysical records,
mantle structure beneath the trough is considered as a
narrow and linear (NS) steep-sided low-velocity channel
from 80 to 200 km depth (Green et al., 1991; Achauer
et al., 1992), suggesting local lithospheric thinning. Moho
depth is known to shallow from c. 40 km beneath the
unrifted domain to 35 km beneath the S Kenya rift axis
(Prodehl et al., 1994; Last et al., 1997).
(2) The 200×50 km transverse volcanic belt extending at
N80°E from the Ngorongoro crater to the Kilimanjaro
includes numerous (b20) volcanic edifices, and their
extensively distributed effusive and air-fall material, that
were emplaced during the time interval 8 Ma–Present.
The NKVB is very little deformed and shows an
inhomogeneous distribution of extensional faulting.
(3) The Eyasi, Manyara and Pangani fault systems form the
main diverging rift structures of the 300 km-wide NTD
sensu stricto which is underlain by Precambrian basement
rocks of the Mozambique Belt (E) and the Archaean
Tanzanian craton (W). Major fault-bounded half-graben
basins, ∼3 km-deep, are only documented along the
Eyasi and Manyara structures (Ebinger et al., 1997). Most
of fault structures are difficult to be directly dated by
radiometric methods because of the lack of volcanic
association, but the large number of earthquakes recorded
in the Eyasi–Manyara tilted fault blocks area (Fig. 4)
indicates that many faults are still active in the western
part of the NTD (Shudofsky et al., 1987; Nyblade and
Langston, 1995; Ibs von Seth et al., 2001). It is worth
noting that the rift structures in the NTD do not represent
the tip zone of the Kenya rift system, since the
topographic expression of Cenozoic extension still exists
600 km further S throughout Tanzania where it is
superimposed to older (Karoo) basinal trends in the
Kilombero region, and might link into the western branch
of the East African Rift System in the N Malawi region
(Le Gall et al., 2004).
2.2. Volcano-tectonic axes
One of the main characteristics of volcanoes in the NTD is to
be spatially arranged along a number of linear axes, showing
four dominant orientations, and locally outlined by faulted
sedimentary basins or parallel fault arrays (Fig. 2):
(1) The NS axial trend corresponds to the axis of the 50 km-
wide Magadi–Natron inner trough that extends as far S as
the Ketumbeine edifice. It continues S of the NKVB with
a slight deviation at N20°E along the Manyara fault
system and its associated hangingwall basin.
(2) The NE–SW trend is marked by various composite
structures. The main axis is the c. 200 km-long Eyasi
structure that comprises to the NE a N40°E volcanic
segment, enclosing ten major shield volcanoes, that
passes laterally to the SW into the N100 km-long Eyasi
fault and its associated N60°E-trending hangingwall half-
graben basin. The Balangida fault lies further SE with a
quite parallel direction. The Tarosero, Essimingor and
Burko volcanoes also strike NE–SW along the Tarosero
lineament across the NKVB, whereas further N, in the
Magadi inner trough, the trace of the eastern uplifted rift
flank shows a pronounced map inflection in the Kajiado
(3) The N80°E trend is exclusively expressed by three
aligned volcanoes (Monduli, Meru and Kilimanjaro)
forming the Meru lineament in the eastern part of the
(4) Lastly, the NW–SE trend is outlined by (a) the curved
southern extremity of the Natron inner trough in the
Engaruka depression, (b) the fissure volcanism associated
with the parasitic activity on the SE flank of Kilimanjaro
and on the Chyulu Hills that both represent the most
external off-axis magmatism in the NKVB, and (c) the
four en echelon uplifted blocks in the 200 km-long
Pangani branch, S of Kilimanjaro.
A number of volcanoes belong to intersecting belts, as
exemplified by Mount Kilimanjaro which occurs at the
intersection of two fundamental basement discontinuities
trending at N150°E (ASWA shear zone) and N80°E (the
Meru lineament along the northern margin of the Masai block).
It is therefore suggested that a strong structural control on
volcanism exists in the NTD, as already reported for many other
3B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
magmatic provinces in the East African Rift (e.g. Ebinger et al.,
In order to address both structural and magmatic aspects of
the rift evolution of the NTD, two complementary approaches
have been carried out in the present work. The overall
structural arrangement of the NTD has been precise from
SPOT4 satellite images (×16, 60×60 km, 20 m lateral
resolution) and SRTM (Shuttle Radar Topographic Mission)
3D dataset (16 m vertical resolution) that also allow us to
extract and analyze fault populations. Topographic profiles
across the major extensional fault structures are drawn from the
3D-digital elevation models. For young and little eroded fault
structures, displacement and fault dip are directly estimated
from measuring the corresponding topographic scarps, hence
providing minimal throw values related to the youngest
faulting event (Mousseau, 2004). For older and more eroded
structures, these values are deduced from the restored geometry
of the fault surfaces. For a few major basin-bounding faults
(Eyasi and Manyara structures), total displacement has been
calculated by Ebinger et al. (1997) from aeromagnetic records
of basement top. Field measurements in selected key-zones
supply additional morphological and kinematic information
about the master fault structures. Fruitful observations have
also been realized in the less well-known Pangani basement
range to the E where active thermal (volcanic) processes are
now documented S of the Kilimanjaro province. Some
uncertainties exist about timing of fault emplacement over
the NTD since syn- and post-faulting dated rocks are missing in
many key-areas where master rift structures are present (Ol
Donyo Ogol, Manyara, Eyasi and Pangani). In order to better
constrain the tectonic and volcanic frameworks in the NTD,
volcanic rocks samples have been collected during the field
investigations. Six new K/Ar age determinations (listed in
Table 1) from long-lived volcanic edifices occurring mainly on
the western side of the NTD (Manyara scarp and Ngorongoro
volcano) have been performed at the LSCE Laboratoire, CEA–
CNRS–UVSQ (Gif/Yvette, France), using the unspiked K/Ar
method on lava groundmass (Guillou et al., 1998; Charbit et al.,
1998; Guillou et al., 2004). Seventeen ages obtained on the
three centers of Mount Kilimanjaro, using the same dating
method, are discussed elsewhere (Nonnotte et al., submitted).
These newly-acquired ages are compiled to existing radio-
metric ages selected as a function of a number of criteria
dealing chiefly with reasonable errors bars, precise sample
location and volcano-stratigraphic attribution.
4. Magmatism in the S Kenya–N Tanzania rifted domain
We attempt here to discuss the c. 8 My magmatic history of
the NTD with regards to the contemporaneous part of the Kenya
rift system lying immediately to the N. This leads us to consider
a propagating rifted domain extending 200 km N of the NTD,
up to Central Kenya. The corresponding Upper Cenozoic
volcanic formations consist of (1) extensive, dominantly basalt/
phonolites fissure-type sequences showing little lateral varia-
tions and forming the main volcanic plateaus (uplifted footwall
terranes) and plains (rift floor), and (2) more localized series
erupted from central shield volcanoes. Particular attention is
paid to the NTD magmatic province where the great number of
Fig. 2. Major rift trends and basement fabrics in the North Tanzanian Divergence. Rift axes are deduced from the spatial distribution of both fault structures and
volcanic eruptive centers. ASWA SZ., ASWA shear zone; BT., CT., ET., EnT., MT., MeT., MNT., PT., SKT., TT., Balangida, Chyulu Hills, Eyasi, Engaruka, Manyara,
Meru, Magadi–Natron, Pangani, South Kilimanjaro, Tarosero, trend, respectively.
4 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
dated volcanic features (list in Table 1), further arranged into
tectono-magmatic lineaments, supplies additional insights on its
structural history during the onset of magmatism. Five
successive stages, typified by specific spatial distribution of
synrift volcanism, are distinguished on the evolutionary sketch
model in Fig. 3:
Dated synrift volcanics in the S Kenya–N Tanzania rift
Relative chronologyTiming (Ma) Magmatic structures Volcano-tectonic trendsReferences
Ol Donyo Sambu
North Chyulu Hills
Magadi grid faulting
South Chyulu Hills
6, 7, 12
Magmatic structures are listed with decreasing ages of initial activity. The magmatic history of long-lived volcanic edifices is assumed to have resulted from either a
single and continuous eruptive episode (time boundaries are indicated), or a multi-phased activity with quiescence periods, as it is the case for the Ngorongoro,
Essimingoror Lemagrut volcanoesthat arethus listed manytimes. The corresponding volcano-tectonic trendis drawnonthe mapof Fig.2. Numbersin theleft column
refer to their relative chronology(they areshownon the fivesuccessivemagmaticstages of the evolutionary modelin Fig. 3). References:1) Bagdasaryan et al. (1973);
2) Evernden and Curtiss (1965); 3) Evans et al. (1971); 4) Foster et al. (1997); 5) Grommé et al. (1970); 6) Wilkinson et al. (1986); 7) our work; 8) MacIntyre et al.
(1974); 9) Isaacs and Curtis (1974); 10) Manega (1993); 11) Haug and Strecker (1995); 12) Hay (1976); 13) Fairhead et al. (1972).
5 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
4.1. Stage 1: ∼8.0–6.0 Ma
At this stage, the southernmost extent of the developing
magmatic rift floor is located in Central Kenya where Lower-Mid
Pliocene phonolites/trachytes (plateaus) series were emplaced for
the first time in the Tinderet and Bahati regions, at the latitude of
S as discrete occurrences in the NS prolongation of the S Kenya
proto-rift, in the Ol Esayeti area where 6–7 Ma-old volcanics are
region where nephelinites, 8.1–7.3 Ma in age (Bagdasaryan et al.,
1973), are the precursor volcanics in the NTD.
4.2. Stage 2: ∼6.0–5.0 Ma
In this time interval (Fig. 3B), the still unrifted domain
extending between Central Kenya and the Essimingor area was
progressively involved into magmatic processes thatexpressed at
the surface by laterally extensive basalt/phonolite sequences,
dated at 5.1 Ma in the rift floor of the Nairobi area (Williams,
1967), and inferred to be Pliocene in age in the Narok region
(Williams, 1964). Quite contemporaneous basaltic series were
erupted in the Magadi area currently forming marginal step-fault
platforms on both sides of the inner trough (Baker, 1958;
Matheson, 1966; Wright, 1967). They consist in the ∼5 Ma-old
Kirikiti Fm. that interdigitated locally with volcanics from the
Olorgesaillie axial volcano of presumed Pliocene age (∼5.8 Ma)
(Baker, 1958; Matheson, 1966). Building up of the S Kenya rift
floor is also accompanied by synchronous eruption of axial
volcanoes (Baker et al., 1971).
After an apparent time gap of ∼3 Ma, a new volcanic
activity, involving three magmatic stages, started in the proto-
NTD, and lasted until recent times.
4.3. Stage 3: 5.0–2.5 Ma
From 5.0 to 2.5 Ma (Fig. 3C), magmatism was still persistent
in the Essimingor area, but shifted for the first time to the W, at
the southern extremity of the incipient Eyasi oblique belt where
the Lemagrut, Sadiman and Manyara edifices were erupted
intermittently in the time range 5.5–4.5 Ma (Bagdasaryan et al.,
1973; Foster et al., 1997). Radiometric records indicate an
interval of reduced magmatic activity from 4.5 to 3.7 Ma in the
Fig. 3. A five-staged evolutionary magmatic model for the S Kenya–N Tanzania rift system during the last 8 Ma. Numbers refer to the relative chronology of dated
volcanics(seelist inTable1). The fiveselectedtime periodscorrespondto majorchangesin the spatialdistribution of synriftmagmatism.Trace of national bordersand
outlines of the Eyasi, Manyara and Natron lakes are plotted to provide geographical coordinates. Ba., Bahati; Es., Essimingor; Le., Lemagrut; Mag., Magadi; OE.,
Esayeti; Ol., Olorgesailie; N., Nairobi; Na., Narok; Ti., Tinderet.
6 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
overall proto-NTD. In the time period 3.7–2.5 Ma, renewal of
magmatism was marked by ongoing activity in older volcanic
areas (Lemagrut, Sadiman), and its lateral spreading over intact
domains such as (1) the central (Ngorongoro) and NE
(Mosonik) parts of the Eyasi volcanic belt, and (2) the Engaruka
oblique depressed zone in the SSE prolongation of the Magadi–
Natron axis (Bagdasaryan et al., 1973).
4.4. Stage 4: 2.5–1.5 Ma
Dramatic changes in magmatism distribution took place at
around 2.5 Ma in the developing NTD (Fig. 3D). To the W,
magmatism waned markedly at the SW extremity of the Eyasi
belt. Massive explosive phases occurred in the Ngorongoro
area, synchronously with emplacement of trachydacite lava
flows at ∼1.9 Ma (this work), whereas new eruptive centers
were emplaced further NE (Olmoti) up to the intersection with
the Magadi NS axis (Humbu Fm) (Manega, 1993; Foster et al.,
1997). In the Essimingor central domain, ongoing volcanism
shifted eastwards along the NE–SW-trending Tarosero axis, as
well as in the Engaruka and Ketumbeine areas (Bagdasaryan
et al., 1973; MacIntyre et al., 1974). Lateral widespreading of
the volcanic province to the E also expressed by the building up
of a new magmatic segment, about 100 km-long, enclosing the
Monduli, Arusha and Shira edifices (Wilkinson et al., 1986; this
4.5. 1.5 Ma–Present
The general tendency that prevailed at earlier stages in the
spatial distribution of magmatism still persisted during recent
times, but with some noticeable modifications about the mode
of emplacement of volcanics. These latter evolved from mainly
explosive and massive extrusion associated to basaltic axial
volcanoes to give rise to cones dominated by pyroclastic
material (Dawson, 1992) (Fig. 3E). There was a nearly total
abandonment of magmatism throughout the Eyasi transverse
belt to the W. Only sporadic volcanic activity occurred within
the Ngorongoro crater by local fissure-type basaltic eruptions
dated at 1.16±0.02 Ma (this work). Along the Manyara–
Balangida N20–N60°E axis, the Kwaraha and Hanang isolated
eruptive centers were synchronously emplaced at ∼1.5 Ma
(Bagdasaryan et al., 1973), hence forming the southernmost
surface expression of synrift magmatism in the eastern branch
of the East Africa Rift. In the central domain, magmatism
spread over a larger province including the Embagai, Gelai and
Burko volcanoes (Manega, 1993). The eastward shift of
magmatism persisted along the NKVB by (1) the eruption of
492–209 ka-old phonolite lavas in the eastern (Mawenzi) and
central (Kibo) centers of the Kilimanjaro polyphased edifice
(Nonnotte et al., submitted), and (2) the intrusion of a swarm of
volcanic cones along the NW–SE-trending Chyulu Hills chain
(Saggerson, 1963; Haug and Strecker, 1995). The latest
evidence of volcanism in this eastern area are fissure-type
parasitic tuff cones aligned along NW–SE lineaments, and
dated at 195–165 ka in the Kilimanjaro sector (Nonnotte et al.,
submitted) and at c. 79 ka in the S Arusha sector (this work).
5. Synrift faulting
It is argued below that, in contrast with previous structural
models of the S Kenya–N Tanzania rift system, the splay faults
in the NTD result from (1) the deflection (to the W) of the
Magadi axial trough into a wider domain of more diffuse strain,
the so-called Mbulu faulted domain, and (2) the development of
a separate and younger oblique arm to the E, the Pangani range
(Fig. 4). The structure of the two diverging branches, assumed
here to form the NTD, is fully discussed below, with a special
attention to the Mbulu faulted domain. Comparing its structure
with those of the Magadi and Pangani extensional patterns
supplies new insights into kinematics of rifting in the NTD. On
an other hand, the Magadi–Natron axial trough is only briefly
presented since it has been extensively investigated by previous
5.1. The Magadi–Natron fault system
The general structure of the Magadi and Natron axial troughs
is illustrated on the simplified cross-sections of Fig. 5A and B.
The Magadi trough is a 60 km-wide asymmetric graben
showing a maximum elevation of 1500 m from the rift floor
to the plateau forming its western shoulder. At the latitude of
Lake Natron, the rift valley widens into a 100 km-wide E-facing
half-graben, bounded to the W by a double system of
extensional master faults. The older Ol Donyo Ogol master
fault is post-dated by c. 3.5 Ma-old basalts in its hangingwall
(MacIntyre et al., 1974), whilst the Natron inner fault, that limits
the axial trough to the W, cuts through basalts, 3.5–1.8 Ma in
age (Isaacs and Curtis, 1974). Further E, younger (1.4–0.7 Ma-
old) lava flows of the Plateau Trachyte series resting over a W-
dipping basement flexural margin are intensely disrupted by a
dense grid of rift-parallel minor faults (Baker, 1958; Fairhead
et al., 1972), commonly regarded as resulting from the focus of
recent strain above a thermally-soften lithosphere. The
statistical and geometrical analysis of this dominantly exten-
sional fault pattern indicates a mean fault length of 5–7 km with
higher values up to 20 km (Fig. 5C). Mean throw values are
b100 m and average fault spacing is 0.5–1.0 km. Total
extension across the Natron half-graben is estimated at 10–
15 km (Gloaguen, 2000).
5.2. The Ngorongoro–Kilimanjaro volcanic belt
The structural map of Fig. 1 shows the significant decrease of
faulting throughout the NKVB where fault structures are
furthermore inhomogeneously distributed. They preferentially
occur to the Wacross a 50 km-wide zone connecting the Natron
and Eyasi–Manyara fault/basin patterns. There, the Sadiman
and Ngorongoro volcanoes are seen to be sharply cut by
extensional faults that start to deviate from the NS axial trend,
and dip dominantly towards the E. Conversely, the shield
volcanoes forming the central and eastern parts of the NKVB
are nearly unfaulted, except the Monduli and Meru edifices
which are transected by minor transverse fault structures. The
N140E-striking extensional fault set bounding the Engaruka
7B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
depression, in the SE continuation of the Natron fault system,
cuts through extensively distributed pyroclastic material on the
northern flank of the NKVB. Most of faults die out south-
eastwards before reaching the N80E axis of the volcanic chain.
Further E, the only evidence of faulting is the NW–SE
topographic ridges within the transverse parasitic cone belt on
the SE flank of Kilimanjaro.
5.3. The Mbulu faulted domain
The plateau region extending between the Eyasi and
Manyara–Balangida fault system is regarded as a specific
structural unit, the Mbulu faulted domain, which forms in map-
view a c. 100×130 km rectangle-shaped zone bounded by four
main fault structures trending at N60°E (Eyasi and Balangida),
N20°E (Manyara) and N160°E (Iramba) (Fig. 5D).
5.4. The Eyasi bounding fault
The ∼100 km-long rectilinear Eyasi fault parallels the
N60°E trend of a Proterozoic mafic dyke swarm intruding
Archaean metamorphic country-rocks in its hangingwall (Vail,
1970). The Eyasi fault cuts through both Archaean (craton) and
Proterozoic (Mozambique belt) crystalline terranes on both
sides of a sub-meridian suture zone (Ebinger et al., 1997).
∼3.1 Ma-old lavas from the Lemagrut volcano are also
Fig. 4. Topography of the NTD from SRTM remote sensing dataset. A. 2D-map arrangement of the main tectonic and volcanic structures. The trace of six topographic
profiles discussed in the text is drawn as well as the epicentral locations of earthquakes recorded by the International Seismological Center (http://www.isc.ac.uk, ISC.,
Thatcham, UK) from 1964 to 2004. See Fig. 1 for the toponymy of main structures. B. E–W topographic cross-section of the DNT from the Eyasi to the Pangani fault
systems. Vertical exaggeration∼25.
8 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
involved in the northern extremity of its uplifted footwall block
(Fig. 5D) (Foster et al., 1997). The pre-existing large
wavelength volcanic topography in the Ngorongoro–Lemagrut
eruptive area might account for the moderate WSW tilt of the
footwall surface of the Eyasi fault. From aeromagnetic records,
the Eyasi hangingwall lake-basin is known to be a 20×70 km
half-graben, b2 km in depth (Ebinger et al., 1997). The sketch
map of Fig. 6A illustrates the regular trace of the Eyasi
bounding fault that comprises three collinear fault segments
linked by a triangle-shaped extensional faulted block, and a
Fig. 5. Geometrical features of rift fault patterns in the Magadi–Natron and Mbulu extended zones. A. Topographic cross-section in the Magadi trough (vertical
exaggeration∼20). B. Topographic cross-section in the Natron axial valley (similar V.E.). C. Histogram of fault length in the Natron grid fault (data extracted from
SRTM images). D. Simplified 2D-map distribution of rift faults in the Mbulu plateau between the Eyasi and Manyara master extensional faults. E. Structural cross-
section in the Mbulu plateau (vertical exaggeration∼10). Same captions as for Fig. 4B.
9 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
N90°E jog-like transverse structure. The overall asymmetric
morphology of its footwall block, with altitudes ranging from
1800 down to 1000 m, shows major disturbances in relation
with the transverse linkage features. Fault segmentation also
results in the development of a narrow synthetic tilted block
(Seketeti structure) in the central hangingwall segment
(Fig. 6A). The exposed fault scarp provides only limited
information about the youngest fault movement with footwall
elevations decreasing gradually southwards from c. 750 m in its
central part to zero at a western tip point. Extension is probably
transferred further SWalong the parallel Kitangiri underlapping
fault that forms, together with the Eyasi fault, a convergent
approaching fault pattern (Fig. 5D). The northern tip zone of
the Eyasi fault is not so clearly defined because of the above-
mentioned pre-existing dome-like topography in the Ngor-
ongoro volcanic area. The parallel topographic cross-sections
of Fig. 6B show the smooth and highly degraded morphology
of its fault scarps. Maximum total (dip-slip) fault movement
can be roughly estimated by summing (1) the ∼3 km fault
displacement deduced from the c. 2.0 km vertical basement
offset in the Eyasi half-graben (Ebinger et al., 1997) by
assuming an average standard fault dip of 60° at depth, and (2)
the maximum height of exposed fault scarps (740 m) that might
coincide with the maximum hangingwall subsidence. This
estimate yields a throw/length ratio of 0.04 that differs of one
order of magnitude from the 0.003 global ratio of Schlische
et al. (1996). Depending on the method used to evaluate fault
heave (Gibbs, 1984), the total horizontal extension recorded by
the Eyasi fault is in the range 1.5–3.0 km. Since lateral
distribution of total slip along the fault trace is not yet known, it
is difficult to apply a fault growth model to the Eyasi master
5.5. The Manyara–Balangida bounding fault system
The Manyara and Balangida faults are part of a complex
regional-scale extensional linked fault system bounding the
Mbulu faulted domain to the E and S (Fig. 5D). The trace of the
Manyara fault extends over ∼120 km at NS–N20°E from its
northern intersection with the Engaruka N150°E fault (that post-
Fig. 6. Geometry of the Eyasi fault network (EF). A. Linear map trace of the Eyasi master fault. B. Topographic cross-sections of the southern (B1) and central (B2)
segments of the Eyasi fault (location on map A).
10 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
dates the Manyara structure) up to a relay structure with the
Balangida fault to the S (Figs. 5D and 7). The footwall block of
the Manyara fault structure is primarily composed of Proter-
ozoic basement rocks overlain to the N by basaltic lava flows
ranging in age from c. 4.9 Ma (Foster et al., 1997) to c. 1.3 Ma
(this work). The hangingwall is occupied by the Manyara lake-
basin which forms a 20×50 km half-graben, 3 km-deep
(Ebinger et al., 1997). The master fault scarp is continuous and
consists of arrays of distinct segments, linked by transverse jogs
that define a two-order segmentation pattern (Fig. 7A). The
N20°E 1st-order segment to the N is 70 km-long from the
Engaruka fault to the Endabash river. Its zig-zag geometry
comprises six en echelon 2nd-order fault segments, ∼10–
15 km-long in average at N20°E, consistently offset in a left-
lateral manner by short (b2 km) cross-faults with a dominant
N150°E azimuth. Maximum scarp elevations are b600 m.
Measurements of striated fault surfaces in basalts to the N
indicate (1) a slight component of sinistral movement along
N20°E-trending oblique normal fault segments, and (2) dextral
shearing along N150°E fault structures. S of Endabash river, the
overall strike of the range front deviates at NS over more than
40 km along the second 1st-order segment that displays a right-
stepping en echelon pattern with surface offsets along N50°-
trending structures. The marked increase of cross-fault length
southwards suggests the greater influence of reactivated base-
ment structures in the rift fault geometry to the S. Fault scarps
Fig. 7. Geometrical characteristics of the Manyara segmented fault (MF). A. 2D-map geometry of the Manyara extensional range front. Tnrefers to transverse fault
structures defining the two-order segmentation of the Manyara fault. B. Topographic cross-sections showing the smooth and degraded morphology of the Manyar fault
scarp (location on map A). C. Abrupt along-strike variations of fault scarp dimensions in relation with the segmented origin of the master fault.
11 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
12 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
are as high as 800 m, and commonly display an eroded mor-
phology on topographic cross-sections (Fig. 7B). The overall
segmented 2D-geometry of the Manyara fault is also detected
by rapid along-strike variations in footwall elevations (Fig. 7C).
Applying the same method as above for the Eyasi structure
indicates a maximum extension value in the range 2.0–4.0 km
for the Manyara master fault.
5.6. The Balangida fault
Fault trace geometry changes markedly along the Balangida
structure that comprises four linear segments, ∼20 km-long
each, arranged in a left-stepping pattern, and linked by two
overlapping fault zones in the Hanang volcanic area (Fig. 5D).
The strike of fault segments deviates gradually southwards from
N30°E (N) to N50°E (Central) and N60°E (S) where it parallels
basement structures in Archaean terrains. The fault scarp height
ranges from 600 m to zero westwards.
5.7. The Mbulu internal fault network
Fault pattern within the Mbulu domain is dominated by a
regular array of 40–50 km-long isolated extensional structures
striking at N40–50°E, at a moderate angle to the N60°E trend of
the Eyasi and Balangida bounding faults, and cutting both
brittle (N60°E) and folded ductile basement fabrics (Fig. 5D).
They preferentially occur in the elevated plateau to the E where
most of them dip to the SE and limit a series of westerly-tilted
fault blocks. The faulted plateau gives way westwards, via the
Yaida asymmetrical graben, to the subsided and less deformed
Eyasi lake-basin. A second graben structure extends, with a
quite similar width of 20 km, in the immediate footwall of the
master fault. The main individual rift structures in the Mbulu
domain are unrestricted faults, with a mean spacing of c. 20 km,
and an average length of 30 km. Topographic fault scarps are
systematically modest, indicating maximum vertical
displacementb100 m (except the fault bounding the Yaida
graben to the E), so that the internal fault network in the Mbulu
domain is assumed not to add a significant contribution to the
3.5–7.0 km (3.5–7%) total extension chiefly accommodated by
the Eyasi and Manyara master faults. A few sub-meridian faults
with minor extensional displacement also occur in two specific
areas, (1) in association with the Iramba boundary fault to the
W, and (2) within a 20 km-wide faulted corridor extending to
the E from the Oldeani (N) to the Hanang (S) volcanoes. They
show very little interaction with the dominant N40–50°E fault
set, and the two subpopulations are likely to accommodate
partitioning of strain.
5.8. The Pangani rift range
The Pangani rift arm is a prominent 250 km-long extensional
range front, striking at N150–160°E, SE of Mount Kilimanjaro
(Fig. 8A). It involves four separate en echelon horst-like blocks,
exclusively composed of Proterozoic gneissic material. Their
respective geometry evolves southwards from the N Pare block
to the S Usambara block, with increasing surface dimensions
(from 20×50 km to 50×50 km), and fault complexity, whereas
maximum topographic elevations are nearly constant at 2300 m,
2700 m and 2500 m in the N, Central and S blocks, respectively.
Each fault block is an asymmetrical horst structure bounded to
the W by the Pangani W-facing extensional fault system that
locally dips N60° (Fig. 8B). The exposed fault scarp elevations
indicate that minimum vertical movement along the Pangani
fault increases significantly southwards from N1000 m (N Pare)
to 1500 m (central block) and N1500 m (Usambara). However,
major uncertainty in total displacement estimate is mainly due
to the lack of hangingwall subsurface records. The hangingwall
depressed zone to the W appears to be occupied to the N by a
∼20×30 km asymmetrical graben-like structure (Fig. 8B1) that
passes laterally southwards into a basement flexural warp on the
eastern side of the Masai cratonic domain. The trace of the main
bounding fault segments parallels the N150–160°E trend of
ductile fabrics in basement rocks (McConnell, 1972), whilst
internal normal faults within each block display oblique trends
at N0–20°E. The series of minor westerly-facing normal faults
observed in the Pare–Usambara basement range, cutting
through shallowly-dipping (20° to the E) foliated gneisses, are
also assumed to be the field expression of the Cenozoic
extensional deformation (Le Gall et al., 2004).
6. Kinematic implications
the NTD, as providing new insights into (1) orthogonal versus
oblique rifting, (2) along-axis magmatic segmentation, and
(3) mutual spatio-temporal relationships and propagation of strain
6.1. Timing of faulting
The tectonic setting of dated sedimentary and volcanic
material provides few constraints about the maximum age of
faulting in the NTD. The earliest evidence of extensional faulting
is documented on the western flank of the Natron basin where
3 Ma-old basalts post-dated the Ol Donyo Ogol fault scarp
(Foster et al., 1997). A second episode of younger extension at
1.6 Ma is locally documented in the Olduvai lacustrine area
(Foster et al., 1997). However, the overall rift morphology of the
NTD is usually assigned to a single major faulting event, dated at
1.2 Ma inthe Engaruka depression (MacIntyre et al.,1974). With
regards to this simplified fault timing framework, evidence for
multi-stage extensional faulting are documented in the Engaruka
area where N150°E fault system transects and post-dates the
N20°E extremity of the Manyara fault. Furthermore, there is
Fig. 8. The Pangani extensional faulted range. A. Fault-map organization of four en echelon horst structures. B. Topographic cross-sections of the N Pare (B1), Central
(B2) and Usambara (B3) ranges showing (1) the consistent asymmetry of the horst-like structure, and (2) the increasing structural complexity southwards (location on
map A). The existence of sedimentary depositional areas in the hangingwall of the Pangani bounding fault is not yet documented (vertical exaggeration×6).
13 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
some uncertainty as to the age of possible older movement along
the Eyasi and Manyara basin-bounding structures, prior to their
major activity (b2.0 and 1.2 Ma, respectively) (Foster et al.,
1997; this work). Available radiometric ages, complemented by
regional geology, suggest that most of their ∼3 km-thick fill
materialinitiated atc.4.9 Mainshallow lacustrineenvironments,
with no (or little) extensional faulting (Foster et al., 1997), so that
it might have been rapidly deposited during recent (b2.0 Ma)
fault-induced hangingwall downthrow. This indicates a max-
imum syntectonic sedimentation rate of 1.5 mm y−1, consistent
with the values estimated for other synrift lacustrine basins in the
East African Rift System (Morley and Wescott, 1999). By
contrast, indirect evidence for successive faulting events in the
recent time period is supplied at a greater scale by comparing
master fault scarp morphologies that are either smooth and
subdued along presumably older structures, as it is the case of the
Eyasi and Manyara faults to the W (Figs. 6B and 7B), or sharper
and angular along younger faults such as the Pangani fault to the
confirm the relative faulting chronology applied above to the
6.2. Orthogonal versus oblique rifting
One major assumptionofour structuralevolution schemeisto
re-interpret the diverging fault pattern in the NTD in terms of two
distinct and diachronous extensional arms, i.e. the Mbulu and
Pangani structures, instead of three similar and synchronous
branches, as commonly stated. Indeed, the Eyasi and Manyara
fault networks are regarded as parts of a single deformed unit, the
Mbulu domain that continues with major changes to the SW the
Magadi–Natron axial trough. Though controversy still exists
Neogene times in East Africa (Bosworth, 1992; Delvaux et al.,
1992; Bosworth and Strecker, 1997), the trend of the minimum
principal stress is assumed to have remained nearly constant
during the onset of faulting at around 1 Ma (Strecker et al., 1990;
our work about the Manyara fault, see above). Therefore, most of
the pronounced changes in tectonic style along-strike are rather
assigned to varying mechanical properties of the extended crust,
southwards from nearly 15–20% (Natron) (Gloaguen, 2000) to
4–7% (Mbulu domain, see above). One of the most striking
structural changes depicted on Fig. 9A is, in addition to the
marked widening of the extended domain southwards, the
consistent deviation of the master fault azimuth from a NS
orientation in the Magadi–Natron trough to a N20°E direction
the rift axis strictly coincides with the regional-scale tectono-
magmatic zonation, and is thus believed to reflect the greater
influence of basement fabrics in the cold and amagmatic crust/
lithosphere, S of the NKVB. The NS orientation of rift structures
in the Magadi–Natron magma-rich trough, at high angle to the
EW extension, is typical of an orthogonal rift setting that
developed above a linear mantle upwelling at depth (Achauer
et al., 1992). Conversely, the N60°E azimuth of the probably
inherited master faults bounding the Mbulu domain define an
diffuse strain (estimated at 3.5–7%) within this b130 km-wide
extended zone expresses by a composite set of newly-formed
faults striking NS and N40°E, i.e. at an angle of about 20° to the
boundary structures. Similar 2D-map fault patterns have been
obtained during experimental modeling of oblique continental
2000). The relativelylower densityoffault structuresobservedin
the Mbulu natural case could be due to a difference inthe amount
of applied extension.
To the E, the contribution of deep-seated basement
discontinuities during Cenozoic rifting is better supported by
the parallelism and spatial coincidence of both volcanic and
tectonic rift structures with the SE prolongation of the ASWA
basement shear zone documented further NW in Kenya
(Chorowicz, 1989; Coussement, 1995) (Fig. 9). In response to
∼EW extension (e.g. Brazier et al., 2005; Calais et al., 2006),
steep shear surfaces associated to the ASWA N140°E lineament
might have first been reactivated as oblique fractures that permit
the ascent of magma through the Chyulu Hills, Kibo, Mawenzi
and parasitic belts of Mount Kilimanjaro during the onset of
magmatism (Haug and Strecker, 1995; Ritter and Kaspar,
1997). At a later stage of rifting, the tectonic rejuvenation of
similarly-trending basement structures, still under the inferred
N90–100°E extension, led to sinistral shearing along the
Pangani range which is internally dissected by NS-trending en
echelon horst structures developed nearly orthogonal to
extension. The typical en echelon structures emplaced along
the Pangani range, as well as those associated to the Manyara
bounding fault, show no evidence for significant strike-slip
displacement, in agreement with the results of experimental
modeling of oblique extension (Clifton et al., 2000). In the
debate about the role played (Chorowicz et al., 1987; Hackman
et al., 1990; Smith and Mosley, 1993) or not (Ebinger et al.,
1989) by basement structures on rift geometry, unequivocal
insights for the latter are not supplied by the diverging fault
pattern cutting through the amagmatic part of the NTD. Indeed,
Precambrian ductile fabrics in both the Mbulu and Pangani rift
branches do not seem to exert a significant influence on rift fault
patterns since they either display little lateral continuity in
highly folded terrains (Mbulu), or are shallow-angle (∼20°)
structures (Pangani). The only supportive evidence for base-
ment–rift structural relationships is provided by the parallelism
of Cenozoic mapped fault traces (Eyasi, Balangida and Pangani
structures) and Proterozoic brittle structures related to either
dyke swarms (Mbulu) or transverse shear zone (Pangani).
6.3. Magmatic along-axis segmentation
The evolutionary sketch model of Fig. 3 shows that until
stage 6.0–5.0 Ma igneous activity in the NTD was totally
disconnected from the extensive magmatically active rifted
domain further N in Central Kenya. The occurrence of discrete
volcanic provinces to the S suggests an episodic ascent and
melting of mantle sources through a still normal-thickness
lithosphere in an early stage of the NTD development. The
14 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
Fig. 9. Conceptual model for the tectonic and magmatic development of the S Kenya–N Tanzania rifted domain during Neogene times. A. Two-staged evolutionary model emphasizing the role of magmatism distribution
on the propagation of strain. B1. Interpretation of the Mbulu faulted plateau as an oblique rift zone. B2. Comparisons with fault patterns obtained during experimental modeling of oblique extension (from Clifton et al.,
2000). C. Crustal-scale cross-section of the NTD illustrating the contrasted mechanical behaviour of the stretched crust on both sides of the Masai cratonic block. The location of the decoupling level in the rifted crust to
the W fits with the depth distribution of seismicity (Shudofsky et al., 1987; Nyblade and Langston, 1995; Brazier et al., 2005). Open circle at the bottom of the Pangani uplifted range indicates the northerly (orthogonal to
the section) origin of mantle melts inferred to have caused significant dynamical uplift during recent times.
B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
corresponding melt/magma chamber could have been linked
to the main magmatic body underneath Central Kenya in two
end-members ways. Far-travelled melts might have flowed at
depth from the main mantle perturbation underneath the
Central Kenya rift towards the nascent volcanic province in
the NTD. The southerly lateral flow of plume material along a
distance of c. 200 km might have been channelized by relief at
the base of the lithosphere, in agreement with the model
applied at a greater scale by Ebinger and Sleep (1998) about the
distribution of Cenozoic magmatism in NE Africa in relation
with the Ethiopian megaplume. Conversely, a localized thermal
anomaly could have directly occurred beneath the NTD, the
dimension of which being tentatively deduced from the sub-
radial distribution of nearly all of Neogene eruptive centers
(excepted the Hanang, Kwaraha and Kilimanjaro external
edifices) within a c. 100 km diameter size, centered on the
Burko–Essimingor area (Figs. 1 and 2). Following this
assumption, the rift fault network radiating from the Burko
area (Fig. 1) might have been triggered by additional dome-
induced radial tensile stress, as similarly proposed by Fairhead
and Walker (1979) about the triple-like rift pattern involving
the transverse Kavirondo rift arm further N (Fig. 3A).
Whatever is the local or far-travelled origin of the mantle
perturbation responsible for magmatism in the NTD, the sharp
boundary of surface volcanism along the NKVB suggests the
abrupt arrest of the southerly-migrated melting products at
depth against a mechanical barrier coincident with the step-like
morphology of the lithosphere at the northern craton margin.
At the scale of the S Kenya–N Tanzania rift system, the age
distribution of Neogene magmatism ahead of the developing rift
undoubtedly demonstrates that rifting propagated discontinu-
ously southwards via discrete magmatic nucleation cells. A
similar punctiform fashion of rift propagation has previously
been applied to the Red Sea and Ethiopian segmented rifts
(Bonatti, 1985; Nicolas et al., 1994; Ebinger and Casey, 2001),
whilst it is considered by other authors as a less energetically
mechanical process thus requiring less total driving force than if
the rift were to be formed simultaneously along its entire length
(Parmentier and Schubert, 1989).
6.4. Strain/magmatism relationships
With regards to the classical issue dealing with timing
relationships between magmatic and faulting activities in
continental rifts, strong evidence exists in the NTD to support
the view that magmatism preceded extension (Baker and
Wohlemberg, 1971; Baker, 1987). Onset of volcanism is thus
too early in the rift evolution of the NTD to have been entirely
caused by lithospheric stretching and it should rather result from
a major mantle thermal anomaly. The model illustrated in Fig. 3
shows that the 1.2 Ma-old major rift episode occurred at stage 5,
i.e. when the surface extent of Neogene volcanism was at its
maximum. From this simple spatial/timing correlation of strain
and magmatism, it can be argued that the inferred melting
material underneath part of the NTD led to considerable thermal
weakening of the lower crustal layer, hence promoting focusing
of higher deviatoric stress in the upper crust which was then
favourable to record brittle tensile failure under applied
extension. That supportsthe view,already recognized elsewhere
by various authors (e.g., Gans et al., 1989; Benes and Davy,
1996; Callot et al., 2002), that localized thermal anomalies and
induced lithospheric magmatism played a key-role in the
focusing and propagation of strain during the Neogene tectonic
development of the NTD. Therefore, recent faulting in the NTD
might have initiated above the magmatic soften area that
roughly coincided with the NKVB before propagating outwards
throughout the adjacent and intact crust (Fig. 9A). The resulting
sub-meridian age migration of strain is further complicated by a
W–E age progression that occurred at a later stage of rifting.
The western part of the NKVB is assumed to have first linked, at
c. 1.2 Ma, the Magadi–Natron axial rift pattern and the Mbulu
less strained domain southwards via the Ngorongoro faulted
area (Fig. 9A). A similar scenario is envisaged for the later
tectono-magmatic evolution of the Kilimanjaro–Pangani east-
ern area where the emplacement of the Mawenzi–Kibo
magmatic centers at b1.0 Ma is likely to have promoted the
nucleation of a discrete cell of strain that subsequently migrated
southeastwards along the ASWA Proterozoic shear zone, giving
rise to the Pangani range (Fig. 9B) and its offshore continuation
in the Kerimbas active fault-basin pattern (Mougenot et al.,
1986). The broadly linear eastward decrease in age of strain is
consistent with the relative youthfulness of master fault scarps
from W to E (Figs. 5 and 8). Similar lateral changes in strain
distribution are reported in the Turkana rift (N Kenya) and have
been assigned by Morley (1994) to lithospheric strengthening
effects. Such a scenario is not envisaged for the NTD for the
following reasons. Firstly, the amount of lithospheric thinning
under the main rift axis in S Kenya (Magadi–Natron trough) is
too small (Achauer et al., 1992) to have been accompanied by
sufficient mineral phase changes (olivine versus Q-Fk) for a
significant strengthening of the lithosphere. Secondly, the
structural development of the eastern rift branch (Pangani) in
the NTD is not accompanied by the abandonment of strain
along the older branch to the W which isstillthe locus of intense
seismic activity (Shudofsky et al., 1987; Nyblade and Langston,
1995; Ibs von Seth et al., 2001). The location of extension in
recent times (b1 Ma) along two diverging branches in the NTD
is believed to be a direct consequence of the eastward shift of
magmatism along the NKVB. The nature of the causal
mechanisms for the W–E migration of magmatism is a long-
standing problem that concerns other off-axis volcanic
provinces along the Kenya rift (Nyambeni and E Turkana,
e.g., Baker et al., 1971; King, 1978; Williams, 1978; Bosworth,
1987). This issue is beyond the scope of the present work, but
one should notice that at a broad scale, the WNW-directed
absolute plate motion of Africa above a fixed mantle anomaly,
as recently refined by Gripp and Gordon (2002), is consistent
with the eastward migration of magmatism along the NKVB.
Although linking the Magadi–Mbulu main rifted domain to
the W and the Pangani discrete range to the E, the NKVB does
not appear to correspond to a large-scale transfer zone. Indeed,
it is deformed very little, except at its western extremity, and
shows no clear evidence for strike-slip faulting. It should rather
correspond to a major change in the rheology of the continental
16 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
lithosphere along the northern boundary of the nearly un-
deformed Masai crustal block.
The spatio-temporal framework of magmatism in Fig. 3 also
reveals clearly the diachronous emplacement of variously-
trending volcanic axes over the NTD. Indeed, older magmatic
axes with dominant NS and NE–SWorientations (referred to as
ET, MT, and TT) were first emplaced to the W, whilst the
intrusion of younger fissure-type volcanics to the E followed
NW–SE-striking axes (EnT, CT, SKT). Given that emplace-
ment of all magmatic axes was likely governed by a steady-state
direction of extension at N90–100°E, their specific distribution
should be rather controlled by the rejuvenation of variously-
oriented basement discontinuities in geographically separate
crustal zones progressively weakened by easterly-shifted
6.5. Surface rift arrangement and lithospheric pattern
In extensional settings, many geometrical parameters of
fault patterns, such as the 2D-map dimension of the deformed
zone, as well as individual fault length distribution, are known
to be partly controlled by mechanical layer thickness (Brun,
1999, and references therein). When comparing fault para-
meters relevant to the Magadi–Natron narrow (60 km) trough
and the Mbulu wider faulted plateau (N120 km), it appears that
mean fault length increases markedly from 5–7 km up to c.
30 km, respectively (Fig. 5C and D). These along-strike
variations in the surface rift fault pattern are consistent with the
southerly increase in crustal thickness deduced from geophy-
sical investigations and earthquakes centroid depths (Last et al.,
1997). More surprisingly, there are no systematic changes in
symmetrical/asymmetrical fault/basin profiles from the
Magadi–Natron to the Mbulu rift portions in relation with
their magmatic versus amagmatic nature. Indeed, both rift
domains are dominated by asymmetrical structures, facing
consistently to the E, with an increasing degree of asymmetry
southwards from a graben (Magadi), to a half-graben (Natron),
and a system of domino-type tilted fault blocks (Mbulu)
(Fig. 5). By analogy with results from rift modeling (Brun et
al., 1985; Faugère and Brun, 1986), it is therefore suggested
that extension along the main (western) rifted zone in the NTD
was accompanied at depth by significant intra-crustal decou-
pling, probably close to the brittle–ductile interface, and with
an eastward sense of shearing, i.e. synthetic to major fault
displacement. The distribution of seismicity indicates a strong
and brittle crust in the depth range 25–34 km (Shudofsky et al.,
1987; Nyblade and Langston, 1995; Brazier et al., 2005) At a
greater scale, it is also noticeable that most (if not all) major
depocenters in the Kenya rift, such as the Lokichar (N Kenya)
or W Kerio (Central Kenya) basins, initiated as asymmetrical
half-grabens, systematically bounded by east-facing normal
faults, probably in response to a common crust/lithosphere-
scale mechanical process operating at the scale of East Africa.
The fact that the inferred mid-crustal decoupling in the NTD
occurred through either Archaean or Proterozoic lithospheres,
that are furthermore either undisturbed (Mbulu) or attenuated
(Magadi–Natron) by mantle perturbations, suggests at first
approximation that mechanisms of crustal extension were not
fully dependent on the thermal structure of the underlying
mantle. That assumption apparently contradicts theoretical and
experimental results of extension modeling that rather predict a
greater degree of coupling, and a resulting upper crustal
symmetrical fault/basin pattern, in cold and thick lithosphere
(Brun et al., 1994), as should have been the case for the
Archaean craton in the NTD. The lithosphere of the Tanzanian
craton is thus believed to have been mechanically disrupted by
Cenozoic extensional tectonism at its margin, in agreement
with the earthquake analysis of Last et al. (1997), whilst it does
not seem to have been significantly modified by the thermal
effects of the adjoining Kenya mantle plume, as documented
from (1) heat flow measurements (Nyblade et al., 1990), and
(2) the spatial distribution of magmatism that avoids craton
nucleus. A more striking difference in the mechanical
behaviour of the stretched lithosphere in the NTD is suspected
to have occurred on both sides of the Masai cratonic block
when considering the structure of the Pangani horst-like range
to the E (Fig. 9C). Its anomalous uplifted geometry, with
maximum fault elevations (∼1800 m) that are twice the values
related to the Mbulu master faults (Eyasi and Manyara
structures), leads us to question the role of a possible
component of dynamical uplift during extension along the
eastern arm of the NTD. A plausible causative mechanism
could be the presence at depth of low-density ponded melt
products, originating from the Kilimanjaro feeding zone and
channelized southeastwards along ASWA-type transverse
discontinuities. Analog off-axis hot mantle material is inferred
to exist further NE underneath the Chyulu Hills area from
lithospheric tomography (Ritter and Kaspar, 1997). Supportive
evidence for mantle thermal anomalies beneath part of the
Pangani range is supplied by field investigations carried out in
July 1998 over the northern part of the Pare Mountains (Ibaya–
Shighatini village, see location on Fig. 8A) where extrusion of
small volume of pyroclastic material (volcanic tuffs mixed with
mud) is likely to be the surface expression of incipient melting
processes at depth. It is also noteworthy that a prominent sub-
meridian gravity low, of unknown origin, is reported by Tesha
et al. (1997) in the Pangani area. Lastly, melting of astheno-
spheric mantle can occur for low lithospheric extension values,
as expected here about the Pangani arm, if abnormally hot
mantle is present at depth (Latin and White, 1990). However,
further gravimetry, geothermal and teleseismic surveys are
required to better define the spatial extent of both the inferred
shallow and deep thermal/melting processes in order to
constrain the ‘dynamical uplift’ hypothesis applied above to
the Pangani rift arm.
Research authorization was provided by the Tanzania
Commission for Science and Technology and by Tanzania
National Parks. We are indebted to the authorities of
Kilimanjaro National Park and to Dr. J. Wakibara for their
help to organize field work in the Marangu area. We thank Pr. S.
Muhongo (University of Dar Es Saalam) and the French
17 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy
Embassy for their support. This study was funded by the DyETI
program of CNRS–INSU and by a grant of SUCRI (Western
Brittany University, Brest). René Maury is kindly thanked for
his constructive suggestions of the initial manuscript. The
comments of M. Sandiford (editor) are greatly appreciated. The
paper has also benefited from extensive critical commentsby W.
Bosworth and one anonymous reviewer. Contribution no. 1047
of the IUEM, European Institute for Marine Studies (Brest,
Achauer, U., Maguire, P.K., Mechie, J., Green, W.V., the KRISP Working
Group, 1992. Some remarks on the structure and geodynamics of the Kenya
rift. Tectonophysics 213, 257–268.
Baker, B.H., 1958. Geology of the Magadi area. Geological Survey of Kenya
Report, 42. 81 pp.
Baker, B.H., 1987. Outline of the petrology of the Kenya Rift alkaline province.
In: Fitton, J.G., Upton, B.G. (Eds.), Alkaline Igneous Rocks. Geological
Society of London, Special Publication, 30, pp. 293–311.
Baker, B.H., Wohlenberg, J., 1971. Structure and evolution of the Kenya Rift
Valley. Nature 229, 538–542.
Baker, B.H., Williams, L.A.J., Miller, J.A., Fitch, F.J., 1971. Sequence and
geochronology of the Kenya rift volcanics. Tectonophysics 11, 191–215.
Benes, V., Davy, P., 1996. Modes of continental lithospheric extension:
experimental verification of strain localization processes. Tectonophysics
Bonatti, E., 1985. Punctiform initiation of seafloor spreading in the Red Sea
during transition from a continental to oceanic rift. Nature 316, 33–37.
Bosworth, W., 1985. Geometry of propagating rifts. Nature 316, 625–627.
Bosworth, W., 1987. Off-axis volcanism in th Gregory rift, east Africa:
implications for models of continental rifting. Geology 15, 397–400.
Bosworth, W., 1992. Mesozoic and Tertiary rifting in East Africa. Tectono-
physics 209, 115–137.
Bosworth, W., Strecker, M., 1997. Stress field changes in the Afro-Arabian Rift
System during the Miocene to Recent period. Tectonophysics 278, 47–62.
Brazier, R.A., Nyblade, A., Florentin, J., 2005. Focal mechanisms and the stress
regime in NE and SW Tanzania, East Africa. Geophysical Research Letter
Brun, J.P., 1999. Narrow rifts versus wide rifts: inferences for the mechanics of
rifting from laboratory experiments. Philosophical Transactions of the Royal
Society of London 357, 695–712.
Brun, J.P., Choukroune, P., Faugères, E., 1985. Les discontinuités significatives
de l'amincissement crustal: application aux marges passives. Bulletin de la
Société Géologique de France 8, 139–144.
Brun, J.P., Sokoutis, D., Van den Driessche, J., 1994. Analog modeling of
detachment fault systems and core complexes. Geology 22, 319–322.
Calais, E., Ebinger, C.J., Hartnady, C., Nocquet, J.M., 2006. Kinematics of the
East African Rift from GPS and earthquake slip vector data. In: Yirgu, G.,
Ebinger, C.J., Maguire, P.K. (Eds.), Structure and Evolution of the Rift
Systems Within the Afar Volcanic Province, Northeast Africa. Geological
Society of London, Special Publication, 259, pp. 9–22.
dimensional laboratory models. Tectonics 21. doi:10.1029/2001TC901019.
Cerling, T.E., Powers, D.W., 1977. Paleorifting between the Gregory and
Ethiopian rifts. Geology 5, 441–444.
Charbit, S., Guillou, H., Turpin, L., 1998. Cross calibration of K–Ar standard
minerals using an unspiked Ar measurement technique. Chemical Geology
Chorowicz, J., 1989. Transfer and transform fault zones in continental rifts:
examples in the Afro-Arabian rift system. Implications of crust breaking.
Journal of African Earth Sciences 8, 203–214.
Chorowicz, J., Le Fournier, J., Vidal, G., 1987. A model for rift development in
Eastern Africa. Geological Journal 22, 495–513.
Clifton, A.E., Schlische, R.E., Withjack, M.O., Ackermann, R.V., 2000.
Influence of rift obliquity on fault-population systematics: results of clay
modeling experiments. Journal of Structural Geology 22, 1491–1509.
Cochran, J.R., 1981. The Gulf of Aden: structure and evolution of a young
ocean basin and continental margin. Journal of Geophysical Research 86,
Courtillot, V., 1982. Propagating rifts and continental break-up. Tectonics 3,
Coussement, C., 1995. Structures transverses et extension continentale. Le rôle
des zones de failles d'Assoua et TRM dans la cinématique néogène du
système de rift est-africain. Thèse 3èmecycle, Université de Brest, France.
Crossley, R., Knight, R.M., 1981. Volcanism in the western part of the Rift
valley in Southern Kenya. Bulletin of Volcanology 44, 117–128.
Delvaux, D., Levi, K., Kajara, R., Sarota, J., 1992. Cenozoic paleostress and
kinematic evolution of the Rukwa–North Malawi Rift valley (East African
Rift System). Bulletin des Centres de Recherche Exploration-Production
Elf-Aquitaine 16, 383–406.
Ebinger, C.J., Sleep, N.H., 1998. Cenozoic magmatism throughout east Africa
resulting from impact of a single plume. Nature 395, 788–791.
Ebinger, C.J., Casey, M., 2001. Continental breakup in magmatic provinces; an
Ethiopian example. Geology 29, 527–530.
Ebinger, C.J., Deino, A.L., Drake, R.E., Tesha, A.L., 1989. Chronology of
volcanism and rift propagation: Rungwe volcanic province, East Africa.
Journal of Geophysical Research 94, 15,785–15,803.
Ebinger, C.J., Poudjom Djomani, Y., Mbede, E., Foster, A.N., Dawson, J.B.,
1997. Rifting Archean lithosphere: Eyasi–Manyara–Natron rifts, East
Africa. Journal of the Geological Society of London 154, 947–960.
Evans, A.L., Fairhead, J.D., Mitchell, J.G., 1971. Potassium–argon ages from
the volcanic province of Northern Tanzania. Nature 229, 19–20.
Evernden, J.F., Curtiss, G.H., 1965. The potassium–argon dating of late
Cenozoic rocks in East Africa and Italy. Current Anthropology 6, 343–385.
Fairhead, J.D., Walker, P., 1979. A detailed gravity study of the crustal structure
associated with the Kavirondo rift valley, East Africa. Geodynamical
Evolution of the Afro-Arabian Rift System. Acad. Naz. Lincei, Rome,
Fairhead, J.D., Mitchell, J.G., Williams, L.A., 1972. New K/Ar determinations
on rift volcanics of S Kenya and their bearing on age of rift faulting. Nature
Faugère, E., Brun, J.P., 1986. Modélisation expérimentale de la distension
continentale. Compte Rendus de l'Académie des Sciences, Paris, 299,
Foster, A.N., Ebinger, C.J., Mbede, E., Rex, D., 1997. Tectonic development of
the northern Tanzanian sector of the East African Rift System. Journal of the
Geological Society of London 154, 689–700.
Fournier, M., Bellahsen, N., Fabbri, O., Gunnel, Y., 2004. Oblique rifting and
segmentation of the NE Gulf of Aden passive margin. Geochemistry
Geophysics Geosystems 5, Q11005. doi:10.1029/2004GC000731.
Gans, P.B., Bohrson, W.A., Scherner, E., 1989. Synextensional Magmatism in
The Basin and Range Province: A Case Study from the Eastern Great Basin.
Geological Society of America, Special Paper, 233. 53 pp.
Gibbs, A.D., 1984. Structural evolution of extensional basin margins. Journal of
Structural Geology 141, 609–620.
Gloaguen, R., 2000. Analyse quantitative de l'extension continentale par
imagerie satellitale optique et radar. Application au rift sud-kenyan. Thèse
3èmecycle, Université de Brest, France. 277 pp.
Green, W.V., Achauer, U., Meyer, R.P., 1991. A three-dimensional seismic image
of the crust and upper mantle beneath the Kenya rift. Nature 354, 199–203.
Gripp, A.E., Gordon, R.G., 2002. Young tracks of hot spots and current plate
velocities. Geophysical Journal International 150, 321–361.
Grommé, C.S., Reilly, T.A., Mussett, A.E., Hay, R.L., 1970. Paleomagnetism and
potassium–argon ages of volcanic rocks of Ngorongoro Caldera, Tanzania.
Geophysical Journal of the Royal Astronomical Society 22, 101–115.
18B. Le Gall et al. / Tectonophysics 448 (2008) 1–19
Author's personal copy Download full-text
Guillou, H., Carracedo, J.C., Day, S., 1998. Dating the Upper Pleistocene–
Holocene volcanic activity of La Palma using the unspiked K–Ar technique.
Journal of Volcanology and Geothermal Research 86, 137–149.
Guillou, H., Singer, B.S., Laj, C., Kissel, C., Scaillet, S., Jicha, B.R., 2004. On
the age of the Laschamp geomagnetic excursion. Earth and Planetary
Science Letters 227, 331–343.
Hackman, B.D., Charsley, T.J., Key, R.M., Wilkinson, A.F., 1990. The
development of the East African Rift system in north-central Kenya.
Tectonophysics 184, 189–211.
Haug, G.H., Strecker, M.R., 1995. Volcano-tectonic evolution of the Chyulu Hills
and implications for the regional stress field in Kenya. Geology 23, 165–168.
Hautot,S.,Tarits,P.,Wheeler,K.,Le Gall,B.,Tiercelin,J.J., LeTurdu,C.,2000.
Deep structure of the Baringo Rift Basin (central Kenya) from three-
dimensional magnetotelluric imaging: implications for rift evolution.
Journal of Geophysical Research 105, 23,493–23,518.
Hay, R.L., 1976. Geology of the Olduvai Gorge. University of California Press,
Ibs von Seth, M., Blumenstein, S., Wagner, R., Hollnack, D., Wohlenberg, J.,
2001. Seismicity, seismotectonics and crustal structure of the southern
Kenya Rift: new data from the lake Magadi aera. Geophysical Journal
International 146, 439–453.
Isaacs, G.L., Curtis, G.H., 1974. Age of early Acheulian industries from Peninj
Group, Tanzania. Nature 249, 624–627.
King,B.C., 1978.Structural evolution ofthe Gregory rift valley. In: Bishop, W.W.
Last, R.J., Nyblade, A.A., Langston, C.A., Owens, T.J., 1997. Crustal structure
of the East African Plateau from receiver functions and Rayleigh phase
velocities. Journal of Geophysical Research 102, 24,469–24,483.
Latin, D., White, N., 1990. Generating melt during lithospheric extension: pure
shear versus simple shear. Geology 18, 327–331.
Le Gall, B., Gernigon, L., Rolet, J., Ebinger, C., Gloaguen, R., Nilsen, O.,
Dypvik, H., Deffontaines, B., Mruma, A., 2004. Neogene–Holocene rift
propagation in central Tanzania: morphostructural and aeromagnetic
evidence from the Kilombero area. Bulletin of the Geological Society of
America 116, 490–510.
Le Pichon, X., Gaulier, J.M., 1988. The rotation of Arabia and the Levant fault
system. Tectonophysics 153, 271–274.
MacIntyre,R.M., Mitchell, J.G., Dawson, J.B., 1974.Age offault movementsin
Tanzanian sector of East African Rift System. Nature 247, 354–356.
Maguire, P.K., Long, R.E., 1976. The structure of the western flank of the
Gregory Rift (Kenya), part 1, the crust. Geophysical Journal of the Royal
Astronomical Society 44, 661–675.
Manega, P.C. 1993. Geochronology, geochemistry and isotopic study of the
Plio-Pleistocene hominid sites and the Ngorongoro Volcanic Highland in
northern Tanzania. PhD thesis, University of Colorado.
Manighetti, I., Tapponnier, P., Courtillot, V., Gruszow, S., Gillot, P.Y., 1997.
Propagation of rifting along the Arabia–Somalia plate boundary: the Gulfs
of Aden and Tadjoura. Journal of Geophysical Research 102, 2681–2710.
Matheson, F.J., 1966. Geology of the Kajiado area. Geological Survey of Kenya
McCall, G.J., 1967. Geology of the Nakuru–Thomson's Fall-Lake Hannington
area. Geological Survey of Kenya Report, 78. 122 pp.
McConnell, R.B., 1972. Geological development of the rift system of Eastern
Africa. Bulletin of the Geological Society of America 83, 2549–2572.
Morley, C.K., 1994. Interaction of deep and shallow processes in the evolution
of the Kenya rift. Tectonophysics 236, 81–91.
Morley, C.K., Wescott, W.A., 1999. Sedimentary environment and geometry of
sedimentary bodies determined from subsurface studies in East Africa. In:
Morley, C.K. (Ed.), Geoscience of Rift Systems — Evolution of East Africa,
44. American Association of Petroleum Geologists, pp. 211–231.
Morley, C.K., Wescott, W.A., Stone, D.M., Harper, R.M., Wigger, S.T.,
Karanga, F.M., 1992. Tectonic evolution of the northern Kenya rift. Journal
of the Geological Society of London 149, 333–348.
Mougenot, D., Recq, M., Virlogeux, P., Lepvrier, C., 1986. Seaward extension
of the East-African Rift. Nature 321, 599–603.
Mousseau, M., 2004. Champs de failles en contexte de divergence de rift
intracontinental. DEA Report. University of Brest. 50 pp.
Mugisha, F., Ebinger, C.J., Strecker, M., Pope, D., 1997. Two-stage rifting in the
Kenya Rift: Implications for half-graben models. Tectonophysics 278, 61–81.
Nicolas, A., Achauer, U., Daignieres, M., 1994. Rift initiation by lithospheric
rupture. Earth Planetary Science Letter 123, 281–298.
Nyblade, A.A., Langston, C.A., 1995. East African earthquakes below 20km
depth and their implications for crustal structure. Geophysical Journal
International 121, 49–62.
Nyblade, A.A., Pollack, D.L., Jones, F., Mushayandebvu, M., 1990. Terrestrial
heat flow in east and southern Africa. Journal of Geophysical Research 95,
Parmentier, E.M., Schubert, G., 1989. Rift propagation. Geophysical Research
Letters 16, 183–186.
Prodehl, C., Keller, G.R., Khan, M.A., 1994. Crustal and upper mantle structure
of the Kenya rift. Tectonophysics 236, 483.
Ritter, J.R., Kaspar, T., 1997. A tomography study of the Chyulu Hills, Kenya.
Tectonophysics 278, 149–169.
Saggerson, E.P., 1963. Geology of the Simba–Kibwesi area. Geological Survey
of Kenya Report, 58. 70 pp.
Schlische, R.W., Young, S.S., Achermann, R.V., Gupta, A., 1996. Geometry and
scaling relations of a population of very small rift-related normal faults.
Geology 24, 683–686.
Shudofsky, G., Cloetingh, S., Stein, S., Wortel, R., 1987. Unusually deep
earthquakes in East Africa: constraints on the thermo-mechanical structure
of a continental rift system. Geophysical Research Letter 14, 741–744.
Smith, M., Mosley, P., 1993. Crustal heterogeneity and basement influence on
the development of the Kenya Rift, East Africa. Tectonics 12, 591–606.
Strecker, M.R., Blisniuk, P.M., Eisbacher, G.H., 1990. Rotation of extension
direction in the central Kenya Rift. Geology 18, 299–302.
Tesha, A.L., Nyblade, A.A., Keller, G.R., Doser, D.I., 1997. Rift localization in
suture-thickened crust: evidence from Bouguer gravity anomalies in north-
eastern Tanzania, East Africa. Tectonophysics 278, 315–328.
Vail, J.R., 1970. Tectonic control of dykes and related irruptive rocks in eastern
Africa. In: Clifford, T.N., Gass, I.G. (Eds.), African Magmatism and
Tectonics. Oliver and Boyd, Edinburgh, pp. 337–354.
Vétel, W., Le Gall, B., 2006. Dynamics of prolonged continental extension in
magmatic rifts: the Turkana Rift case study (North Kenya). In: Yirgu, G.,
Ebinger, C.J., Maguire, P.K. (Eds.), The Structure and Evolution of the East
African Rift System in the Afar Volcanic Province. Geological Society of
London, Special Publication, 259, pp. 211–235.
Walsh, J., 1969. Geology of the Eldama Ravine–Kabarnet area. Geological
Survey of Kenya Report, 83. 48 pp.
Williams, L.A., 1964. The geology of the Narok district. Proceedings of the East
Africa Academy 1, 37–49.
Williams, L.A., 1967. Geology of the Nairobi region. In: Morgan, W.T. (Ed.),
Nairobi: City and Region. Oxford University Press, Nairobi, pp. 1–13.
Williams, L.A., 1978. The volcanological development of the Kenya rift. In:
Neumann, E.R., Ramberg, I.B. (Eds.), Petrology and Geochemistry of
Continental Rift. Reidel, Dordrecht, pp. 101–122.
Wilkinson, P., Mitchell, J.G., Cattermole, P.J., Downie, C., 1986. Volcanic
chronology of the Meru–Kilimanjaro region, Northern Tanzania. Journal of
the Geological Society of London 143, 601–605.
Wright, J.B., 1967. Geology of the Narok area. Geological Survey of Kenya
Report, vol 80.
19 B. Le Gall et al. / Tectonophysics 448 (2008) 1–19