Tectonophysics 607 (2013) 80–97

Contents lists available at ScienceDirect

Tectonophysics journal homepage: www.elsevier.com/locate/tecto

The effect of thermal weakening and buoyancy forces on rift localization: Field evidences from the Gulf of Aden oblique rifting N. Bellahsen a,b,⁎, L. Husson c,d, J. Autin e, S. Leroy a,b, E. d'Acremont a,b a

UPMC Univ. Paris 06, UMR 7193, ISTeP, F-75005 Paris, France CNRS, UMR 7193, ISTeP, F-75005 Paris, France CNRS-UMR 6118, Géosciences Rennes, Univ. Rennes 1, France d CNRS-UMR 6112, LPG Nantes, Univ. Nantes, France e CGS-EOST, CNRS-Univ. Louis Pasteur, 1 rue Blessig, F-67084 Strasbourg, France b c

a r t i c l e

i n f o

Article history: Received 30 September 2012 Received in revised form 11 May 2013 Accepted 30 May 2013 Available online 10 June 2013 Keywords: Rifting Gulf of Aden Buoyancy forces Analog models

a b s t r a c t On the basis of field and geophysical data, analog and numerical models, we here discuss the role of buoyancy forces arising from thickness variations in the lithosphere during rifting. In the Gulf of Aden, an oceanized Tertiary oblique rift, several successive directions of extension and associated normal faults suggest that transient stress rotations occurred during rifting. Especially, rift-parallel faults (070°E) overprinted the early divergenceperpendicular normal faults (110°E). Moreover, some first-order differences are noticeable between the western part of the Gulf, which deformed under the Afar hot spot influence, and the eastern part. In the western Gulf of Aden, the ocean–continent transition (OCT) and the oceanic ridge have cut obliquely through the inherited and reactivated Mesozoic basins (100°E to 140°E). The OCT trend is parallel to the overall Gulf trend (070°E). In the eastern part, the oceanization occurred within few syn-rift 110°E-trending basins and the OCT trends mostly perpendicular to the divergence direction. Here, we propose that this contrast is strongly controlled by the Afar hot spot: during rifting times, the hot spot likely induced a hot thermal anomaly in the western asthenosphere. This may have triggered both thermal buoyancy forces and thermal weakening of the lithosphere that helped localizing the rift obliquely. In such localized rift, rift-perpendicular trending crustal buoyancy forces (i.e. around 160°E) have enhanced rift-parallel normal faults (070°E) during final rift localization into a narrow zone strongly oblique to the early syn-rift basins. As a consequence of the Afar hot spot, in the west, the ridge is long and straight; in the east, the ridge segments are rather long too (although less than in the west) as the ridge initiated parallel to the OCT; in between, the ridge is more segmented as both the hot spot influence gradually decreases eastward and the ridge initiated obliquely to the OCT. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Rift localization occurs in the lithosphere when extensional strain focuses in a relatively narrow zone, possibly leading to continental breakup and oceanization. Such localization is obviously witnessed by rifts that evolved into oceanic basins wherein strain localization occurred in the distal passive margin, leading to mantle exhumation at some margins (Beslier et al., 1996, 2004; Boillot et al., 1987; Bonatti et al., 1990) and subsequent oceanic spreading. Strain localization is also documented in active rifts where old faults are located on the rift shoulders while new and active ones are observed in the rift center, along with seismicity and volcanism, as in the East African Rift System (EARS; Corti, 2009; Ebinger and Casey, 2001; Ebinger et al., 1993; and references therein). On the contrary, the localization did not occur in some wide rifts (sensu Buck, 1991; Brun, 1999), like in the basin and range ⁎ Corresponding author at: UPMC Univ. Paris 06, UMR 7193 ISTeP, 4 place Jussieu, 75252 Paris cedex 05, France. Tel.: +33 1 44 27 74 64. E-mail address: [email protected] (N. Bellahsen). 0040-1951/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tecto.2013.05.042

(e.g. Dickinson, 2002; and references therein): in this context, extension is distributed over wide areas, mainly because of the thickness and the thermal state of the lithosphere, without any further oceanization. Less clear however is why and when such localization occurs, and what processes control it. Strain localization may be due to several processes. (1) It may occur because of significant strain softening (e.g. Huismans and Beaumont, 2007; Lavier and Manatschal, 2006), due to the feldspar–mica reaction (Bos and Spiers, 2002), pore pressure (Sibson, 1990), grain-size reduction, and/or mantle serpentinization (Pérez-Gussinyé and Reston, 2001). (2) Specific rheological stratification of the lithosphere may also have a localizing effect: analog as well as numerical models have shown that 4 layer brittle–ductile models promote a strongly localized rifting while 2-layer models promote a rather distributed one (e.g. Brun, 1999; Gueydan et al., 2008; and references therein). (3) In volcanic margins, the break up is accompanied by high amount of magmatism, underplated, intruded in the crust, and/or extruded at surface in Seaward Dipping Reflectors (SDR, e.g. Berndt et al., 2001; Planke et al., 2000), suggesting temperature anomaly in the asthenosphere

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

(Holbrook et al., 2001). Such hot asthenosphere may weaken the lithosphere by injections of melts (Saunders et al., 1992; and references therein) and by dyking that allows stretching at low stress (Buck, 2004). It may also decrease the strength by heat conduction and thermal erosion at the base of the lithosphere (Saunders et al., 1992; White, 1992; and references therein). (4) Finally, buoyancy forces arising from density variations in the lithosphere (Artyushkov, 1973; Fleitout and Froidevaux, 1982) may also provide driving forces for rift localization (Burov, 2007; Davis and Kusznir, 2002; Huismans et al., 2001). Crustal thickness variations are primordial, but the mantle lithosphere may also have slightly negative buoyancy (Griffin et al., 2009; Watremez et al., in press). In such case, a mantle lithosphere denser than the asthenosphere may help the localization. Such local stresses arising from density and thickness variations have been suggested to be active in the EARS (Zoback, 1992) and in the Gulf of Aden (Autin et al., 2010b) and may have produced stress rotations, because of oblique rifting context, that are actually observed in the field (Bellahsen et al., 2006). Buck (1991) and Buck et al. (1999) proposed that the mode of continental rifting was partly controlled by the ratio between crustal buoyancy and lithospheric necking effect and that the temperature controls this ratio. In the mantle lithosphere, it has been suggested that Rayleigh–Taylor instabilities can significantly thin the lithosphere (Conrad and Molnar, 1997; Gemmer and Houseman, 2007; Molnar and Houseman, 2004; Molnar et al., 1998). Similarly, Pascal and Cloetingh (2009) showed that gravitational potential stresses deriving from lateral variations in lithosphere structure at continental margins (South-Norway shelf) may explain features such as the seismicity pattern and the present-day stress orientations. In orthogonal rifts, such interpretations may be supported by numerical models (Burov, 2007; Huismans et al., 2001). However, the force that causes faulting, whether originating from far-field tectonics or local buoyancy forces, cannot unambiguously be disentangled from the geological record as all faults have the same trend and their ages are uncertain due to the lack of precise absolute and relative dating. Conversely, oblique rifts provide a unique opportunity to decipher the mechanisms at play and to establish their timing during rift localization as many fault populations develop, leading to complex fault patterns (see Autin et al., 2010b; Corti, 2009 and references therein) that includes both rift-parallel and rift-oblique faults. As suggested by Bellahsen et al. (2006), rift-parallel faults may witness rift localization processes. The mechanisms that promote oblique rifting are multiple and do not univocally include lithospheric structural inheritance or intrinsic mechanical behavior of the lithosphere (Autin et al., 2010b). Among the youngest and best-documented oblique continental rifted margins are the eastern margins of the Gulf of Aden, whose ridge is the Arabia/Somalia plate boundary (Fig. 1). Based on the study of the Gulf opening and its structural evolution from rifting to active spreading, the aim of this contribution is to show that buoyancy forces have a feedback effect on rift dynamics. Moreover, we discuss the differences between the eastern and the western parts of the Gulf and the role of the Afar hot spot on the rifting geometry and evolution. We synthesize published and new onshore and offshore structural data and compare them to a synthesis of analog models. This analysis advocates for a structural evolution that is discussed in the light of numerical considerations of stresses arising from buoyancy forces.

81

to the rift itself (Chang and Van der Lee, 2011; Sicilia et al., 2008). The low velocity body is not documented further east because of a very poor resolution due to the lack of receiver in eastern Yemen and southern Oman. This low seismic velocity anomaly certainly mirrors the source of the Afar hot spot, but imaged by Basuyau et al. (2010) and potentially laterally feeds the Aden spreading ridge up to the eastern Gulf of Aden (Leroy et al., 2010b). The hot spot has been active since 45 Ma (George et al, 1998), with pulses ca. 30 Ma (Hofmann et al, 1997) and is still active as attested by volcanic activity in the Afar area, as well as Quaternary volcanics along the Yemeni and Somali margins (Leroy et al., 2010b). This corresponds to a hotter asthenosphere that is readily inferred from the shallower bathymetry and higher heat flow in the western Gulf of Aden (Fig. 1) (e.g. Lucazeau et al., 2010; Rolandone et al., 2013 and references therein). Moreover, the sedimentary record shows that the western Gulf has been topographically high compared to the eastern part since at least Eocene times: siliciclastic deposits indicate that the western part was emerged and submitted to erosion while the eastern part was mainly the location of a carbonate-rich sedimentation (Leroy et al., 2012). 2.2. Continental rifting Rifting started at about 34 Ma (Leroy et al., 2012; Roger et al., 1989; Watchorn et al., 1998) all along the Gulf. At this time, the subduction of Tethyan slabs underneath the Eurasian plates induced extension in the Afro-Arabian plate (Bellahsen et al., 2003). Such extensional strain field is due to the collision in the northern Arabian plate while the northeastern part was still subducting, this along-strike variation of the convergence mode being most probably the relevant boundary condition for intraplate extension. The extension was located in the Afro-Arabian rifts (Red Sea, Gulf of Aden and East African rifts) as the Afar hot spot activity localized the rifts. The combination of intraplate stresses with such a weakness produced oblique rifts, without any oblique preexisting lithospheric weakness (Autin et al., 2010b; Bellahsen et al., 2003, 2006). The inherited Precambrian structures likely did not have a strong impact on the rift development (Fig. 1): none of the N–S, NE–SW, or NW–SE structures were reactivated in the Arabian plate. On the contrary, the structural pattern is strongly controlled by basins inherited from Cretaceous intraplate extensive event (see Birse et al., 1997; Brannan et al., 1997; Fantozzi, 1996; Leroy et al., 2012 for a synthesis; and Fig. 1). Those basins were partly reactivated during Tertiary rifting and trend from 90°E to 140°E (Fig. 1, Balhaf, Masila, Jiza-Qamar, Berbera, Nogal, Darror, and Gardafui basins). As a result of rift obliquity, the reactivated basins (and newly formed ones) are arranged en échelon along the continental margins (Fig. 2). Finally, the Indian Ocean and its ridge (Carlsberg, Fig. 1) most likely influenced the location of extensive deformation in the eastern most part of the Gulf of Aden (Manighetti et al., 1997). At the Gulf scale, many recent studies showed that there were successive directions of extension during rifting ranging from 020°E to 160°E (Bellahsen et al., 2006; Fournier et al., 2004; Huchon and Khanbari, 2003; Lepvrier et al., 2002) (Fig. 2). It appears that extensional stresses most probably rotated counter-clockwise from 020° to 160°E, although it might be more complicated especially in distal parts of the margins (Autin et al., 2010b). 2.3. Ocean–continent transition

2. Geological setting 2.1. The Afar hot spot Seismic tomography reveals the presence of a body of low seismic velocities, in the west, underneath the Afar (e.g. Bastow et al., 2005; Benoit et al., 2006; Chang and Van der Lee, 2011; Hansen et al., 2012; Montelli et al., 2006). It ramifies to the Gulf of Aden, but does not go further east than, approximately, the Alula Fartak fault zone (Fig. 1) and, at shallow depths underneath the lithosphere, remains confined

After the development of syn-rift grabens and horsts, the deformation localized where the crust was the thinnest, i.e. in distal margin grabens close to the future ocean–continent transition (OCT, Fig. 2, yellow). In the eastern Gulf of Aden (i.e. east of Alula Fartak F.Z.), the OCT ridge may represent exhumed serpentinized mantle locally intruded by post-rift magmatic material, which modified the OCT after its emplacement (Autin et al., 2010a; d'Acremont et al., 2006, 2010; Leroy et al., 2010a; Watremez et al., 2011). The OCT segments are about 50 km long and oriented 110°E (Fig. 2). It is noteworthy that

82

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

Fig. 1. Topography, bathymetry, heat flux, and inherited structures in the Gulf of Aden and the Arabian plate. a) Heat-flow measurements in the Arabian Shield and Platform, the Red Sea and the Gulf of Aden (Rolandone et al., 2013). Small squares are published measurements in the Red Sea and the Gulf of Aden (Pollack et al., 1992). Large squares are from Gettings (1982), stars from Galanis et al. (1986) and triangles from Förster et al. (2007). Small dots are from marine measurements (Lucazeau et al., 2010). Large dots are from ANR Yocmal experiments (Rolandone et al., 2013). The topography and the heat flow increase westwards; the bathymetry depths increase eastwards. The white line represents the extent of the low velocity body at 400 km depth after Hansen et al. (2012). The dashed white line represents the extent of lava flows during Tertiary times after Leroy et al. (2012). b) Inherited structures. In orange are the Mesozoic basins from Birse et al. (1997), Brannan et al. (1997), and Leroy et al. (2012). Other structures are from Stern and Johnson (2010).

the spreading ridge initiated within a large syn-rift basin (Fig. 2, to the north of the inherited Gardafui basin, see Autin et al., 2013-this volume; d'Acremont et al., 2005). The trend of the oceanic spreading ridge segments is similar to the trend of the syn-rift basins, whose geometry is detailed in the next section. Between the Balhaf basin and Alula Fartak F. Z., margins are rather magma-poor and an OCT with an undulating shape (Leroy et al., 2012) (Fig. 2) emplaced slightly obliquely to the inherited basins. The Afar hot spot may have influenced the break up after or during emplacement of the OCT (Leroy et al., 2010a,b; Lucazeau et al., 2008). In the westernmost part (west of Shukra El Sheik F.Z., Figs. 1 and 2), the Afar hot spot influence is syn-rift. SDR emplaced in the future distal

margins (Ahmed et al., 2013; Tard et al., 1991) and there is a sharp transition between the rifted continental crust and the oceanic crust (Leroy et al., 2012). There, as well as east of Shukra El Sheik F.Z. and up to Balhaf graben, the breakup occurred along a trend strongly oblique to the continental syn-rift basins. 2.4. Aden–Sheba spreading ridge The Gulf of Aden oceanic spreading ridge (Fig. 1) is the active divergent boundary between the Arabian and Somalian plates. It opens at a rate that varies from 18 mm/yr (azimuth 025°E) to 13 mm/yr (azimuth 035°E) from east to west, respectively (Jestin et al., 1994; Vigny et al.,

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

83

Fig. 2. Structural map of the Gulf of Aden modified from Leroy et al. (2012). Sterodiagram of fault slip data inversion is from Lepvrier et al. (2002), Huchon and Khanbari (2003), Fournier et al. (2004), Bellahsen et al. (2006), Leroy et al. (2012) and Bellahsen et al. (2013-in this volume). Note the directions of extension between 020°E and 160°E. If the intermediate trend is represented, the optima are much more numerous. The syn-rift basins are arranged en échelon along the margin, with trend around 140°E in the west to 110°E in the east. The Gulf is divided in three zones: west of Shukra El Sheik Fracture Zone (F.Z.), the young (and straight) ridge propagates into the Gulf of Tadjoura; between Shukra El Sheik F.Z. and Alula Fartak F.Z., the Aden ridge is active since 17.6 Ma (A 5d) and is segmented between Balhaf and Bosaso grabens; the Sheba ridge east of Alula Fartak F.Z. also initiated at 17.6 Ma (d'Acremont et al., 2006; Leroy et al., 2012).

2006). The 075°E trending Gulf is thus oblique to the main direction of divergence. In the westernmost part, the Aden ridge propagates into the African continent in the Gulf of Tadjoura and the Asal rift (Fig. 1). West of the Shukra El Sheik F.Z. (Fig. 1), oceanic spreading started at about 6 Ma (Audin et al., 2004; Leroy et al., 2012). From Shukra El Sheik to east of Alula Fartak F.Z., oceanic spreading started at 17.6 Ma (d'Acremont et al., 2006, 2010; Leroy et al., 2012) and probably at 20 Ma in the extreme eastern part where the rifting was emplaced in an old oceanic lithosphere (Fournier et al., 2010). The oceanic spreading ridge, as far as the eastern part of the Gulf, was influenced by melting anomaly (d'Acremont et al., 2010) implying coeval ridge jumps. This has been explained by ridge–hot spot interaction with channelization of the hot spot material along the spreading ridge from west to east (Leroy et al., 2010b). 3. Structural analysis The Dhofar is the southernmost region of Oman (Fig. 2), where we carried out several surveys that are synthesized here. Fig. 3 is a new structural map modified from previous studies (Autin et al., 2010a; Bache et al., 2011; Bellahsen et al., 2006; d'Acremont et al., 2005; Leroy et al., 2010a, 2012). Geophysical data have been carefully accounted for in order to provide the most precise information such

as the displacement on main faults. Several fault populations can be discerned, from 070°E to 110°E. Onshore, four grabens accommodate the extension and are disposed en échelon along the margin (Fig. 3: Rakhyut, Ashawq–Salalah, Haluf, and Sala'Afan grabens). Syn-rift sedimentary rocks crop out only in very localized area, such as the Rakhyut basin, except in the main onshore Ashawq–Salalah basin (Fig. 3), where they are widespread underneath the Quaternary rocks in the Salalah plain (Platel et al., 1992). This basin is controlled by normal faults whose trend is alternatively 070°E and 110°E. Bellahsen et al. (2006) showed that the 110°E segments predated the 070°E ones, consistently with fault slip analysis and analog models (Bellahsen and Daniel, 2005). The rift parallel segments (070°E, north of Salalah) have significant throw and accommodated much extension. Such a geometry suggests that the 020°E trending extension predates the 160°E trending one. The thick depocenter located in this basin is due to the specific geometry caused by the two different fault trends. The proposed chronology is confirmed by detailed geometries, such as a horsetail that attests for left-lateral reactivation of 110°E normal faults (see in the west of the Ashawq graben on Fig. 3). Moreover the global tilting of strata in this basin is southward (Platel et al., 1992) and most probably controlled by north dipping 070°E normal faults; this suggests that those faults are late and confirms the relative chronology proposed above. Similar geometries are observed in western Dhofar (near Rakhyut, Fig. 3), where faults have various trends, from 070° to

84

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

Fig. 3. Detailed onshore–offshore structural map of the Dhofar area, southern Sultanate of Oman, eastern Gulf of Aden modified is from d'Acremont et al. (2005), Bellahsen et al. (2006), Autin et al. (2010a), Leroy et al. (2010a) and Bache et al. (2011). Offshore, the normal fault horizontal throw is represented by variable thickness of the fault line. This throw has been estimated from the horizontal offset observed on seismic data (Encens-Sheba and Encens cruises, Leroy et al., 2004, 2010a). Note the curved proximal boundary of the OCT and the offset of the first oceanic magnetic anomaly (A5d).

110°E. The two other onshore grabens (Haluf and Sala'Afan grabens, Fig. 3) also display similar trends, yet have more linear geometries. Offshore, the fault pattern is imaged with various seismic lines from several datasets (see Autin et al., 2010a; Bache et al., 2011; d'Acremont et al., 2005; Leroy et al., 2010b). Fig. 3 is a structural map from a new compilation of several published and unpublished seismic lines. Near the shore (central part, south of Salalah), much of the mapped faults are rift-parallel and strike 070–080°E. Those faults trend parallel to the faults nearshore Mughsayl (Fig. 3), with alternating senses of dip. In the west and the east, we mapped 110°E trending normal faults in the continuity of onshore faults: southeast of Rakhyut and east of Salalah. Besides those faults, the main nearshore faults are rift-parallel. The 110°E trending faults, i.e. extension perpendicular, are more numerous just north or within the OCT. The occurrence of such faults within the OCT shows that a 020°E direction of extension was active during the OCT formation, i.e. during a late phase as suggested by Autin et al. (2010b). Thus, the high resolution analysis of the fault network of the Dhofar margin corroborates that temporal variations in stress regime controlled the structural evolution of the Gulf: it is most likely that the extension rotated from 020°E to 160°E during early rifting phases and rotated back to 020°E during a late phase, most probably during the OCT formation. Transfer zones strike approximately 020°E and separate three segments, that themselves control the initiation of the OCT and oceanic spreading (d'Acremont et al., 2005). Other transfer zones are deduced from seismic data and may have been active throughout: seismically blurred zones trending around 160°E seem to separate domains in the near shore margin where most of the faults trend 070°E. Because of the lack of 3D grip of seismic reflection profiles, those transfer zones are

not very well imaged, but are nevertheless consistent with the 160°E extension active during rifting as attested by the 070°E faults and fault– slip data. 4. Insights from published analog models Several studies used analog models to constrain the evolution of oblique rifts (see Corti, 2012 for a review). Among them, models of an extending crust above an oblique velocity discontinuity (Clifton et al., 2000; McClay and White, 1995; Tron and Brun, 1991; Withjack and Jamison, 1986; Fig. 4) or above a mantle weakness zone (Agostini et al., 2009; Corti, 2004, 2012; Corti et al., 2001, 2003; Mart and Dauteuil, 2000; Sokoutis et al., 2007) showed that, for moderate obliquity (from 30° to 60°), the main fault population strikes at an intermediate angle between the trend of the rift and the direction perpendicular to the plate motion (Fig. 4). In the case of transtension, opening direction and direction of extension are indeed not collinear (Withjack and Jamison, 1986). These authors explain the intermediate strike by a rotation of the minimum principal stress in the applied zone of obliquity, induced by the early transtensional regime of oblique rifts. These intermediate faults characterize the early stage of deformation of oblique rift as observed in the evolution of moderate obliquity models (Autin et al., 2010b) (Fig. 5), whose simplified structural evolution of the models is presented in Fig. 6. These authors presented two models: one with four-layer (brittle–ductile) uniform rheology (Figs. 5a and 6a, b, c, d) and one with a pre-existing oblique lithospheric weakness (Figs. 5b and 6e, f, g, h). In models with a uniform rheological structure, during a first stage, en échelon normal fault initiate (Fig. 6a) as in the above-cited models. In models with an oblique weakness, en

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

échelon normal fault also initiate, but with more rift-parallel faults than in the previous model (Fig. 6e). A second population of faults can be observed with strikes parallel to the oblique rift, especially in lithospheric-scale models (Agostini et al., 2009; Autin et al., 2010b; Corti, in press). When the pre-existing

85

weakness zone is located in the lower crust (Agostini et al., 2009; Corti, in press) external faults (faults that are located on rift shoulders) are rift-parallel and predate the intermediate to extension-perpendicular internal faults. When the weakness zone is in the lithospheric mantle (Autin et al., 2010b), external faults (red faults on Fig. 6b, f) tend to be late and rift-parallel. The latter evolution suggests that the local extension direction rotates from the regional divergence trend to rift-perpendicular extension trend. Another type of faults can accompany the rift-parallel faults: in models with uniform rheology (4-layer brittle–ductile lithosphere but without oblique weakness), riftperpendicular transfer faults, i.e. parallel to the local rift-perpendicular extension, initiate (Fig. 6b). These transfer faults may be related to the initial en échelon pattern, to accommodate the high amount of extension in the offset basins: when extension localizes, transfer faults initiates between these basins. As a consequence of the stress rotation, early intermediate faults are obliquely reactivated and rift-perpendicular transfer faults must initiate to accommodate the transfer of deformation from a basin to another. Such transfer faults are not observed in models that include an oblique weakness zone: during the second stage, most of the newly formed faults are rift-parallel (red faults on Fig. 6f) and the rift localizes early along its oblique trend; in the analog models, it localizes along the pre-existing weakness. In this case, no transfer faults are necessary as the basins are aligned and not significantly offset and en échelon after this stage. The third observed fault population is composed of displacementnormal faults. In crustal-scale models, this population is not observed (e.g. Clifton et al., 2000; Tron and Brun, 1991) and faults always form at a slight angle to the displacement normal direction. In lithosphericscale models, extension-perpendicular faults form during the late stages of deformation and in internal parts of the rift (with an obliquity of 30°, as in Agostini et al., 2009). In particular, in the models that are controlled by moderate obliquity (40° in Autin et al., 2010b), these faults form especially during the late stages of deformation. In rheologically uniform models, the deformation localizes into narrow divergenceperpendicular zones in the rift center (red faults on Fig. 6c). This evolution suggests that the extension direction is mostly controlled by the regional divergence direction. After the late rifting stage, further divergence would lead to the formation of the OCT and future distal margins, which are not observed in the models. However, looking at the late stages of oblique rift, we can propose a pattern of evolution for the OCT and spreading centers in Fig. 6. In lithosphere with no weakness, the OCT would then probably be characterized by a very segmented pattern (Fig. 6c). Finally, the oceanic ridges would most likely initiate in these distal basins, perpendicular to the far-field divergence, separated by few large transform faults (Fig. 6d). Transform faults initiate because the rift is segmented. Such geometry may thus be controlled by the segmentation of the rift, but one should note that transfer faults are not parallel to transform ones. This is due to the fact that active extensions during these two distinct periods were not striking in the same direction: rift-perpendicular extension during rifting that controls the trend of the transfer faults, and divergence-parallel movement during oceanic spreading. In the model with an oblique weakness, the deformation localizes into a long and continuous narrow zone in the rift center (Fig. 6g), and may lead to the formation of the OCT and thus the future distal margins. In a fourth stage (Fig. 6h), the oceanic ridges form in this distal basin, with an overall direction that is strongly oblique to the far-field divergence, thus separated by very numerous but small transform faults. 5. Interpretation and working hypothesis

Fig. 4. Line drawing of published analog models of oblique rifting is from McClay and White (1995), Tron and Brun (1991), Clifton et al. (2000), from a) to c), respectively. In these models, the oblique rifting is modeled with a sand layer over a basal velocity discontinuity in a), with a sand layer over a basal discontinuity and a silicone putty layer in b), and with wet clay layer over a basal rubber sheet. In all models, the main fault population has a trend intermediate between the extension-perpendicular and the rift axis.

The fault analyses along the Gulf of Aden margins, combined with analog model results, show an overall structural evolution with extension trends varying from 020°E to 160°E and back to 020°. It can be interpreted as a consequence of the competing effects of far-field and buoyancy forces. In the following, we refer to buoyancy forces those

86

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

Fig. 5. Photo of analog models of oblique rifting from Autin et al. (2010b) and line drawing. These models were performed with 4-layer lithosphere (sand-silicone brittle–ductile) over a low viscosity glucose syrup, representing the asthenosphere. The photos are final state picture. Note the differences between the “homogeneous” model on the left and the “heterogeneous” model on the right. In the heterogeneous model, the oblique rifting is influenced by an oblique weakness located in the lithosphere that joins the two lateral velocity discontinuities. In the homogeneous one, there is no weakness in the lithosphere and the rift obliquity results from a spontaneous obliquity only due to the offset of the lateral velocity discontinuities. The homogeneous model is characterized by a segmented rift with extension perpendicular normal faults and many rift-perpendicular transfer faults. The heterogeneous model is characterized by a long and oblique rift composed of rift parallel faults and very few transfer faults.

that are due to lateral density variations. Lateral buoyancy forces during rifting chiefly arise because of crustal thinning: the uneven distribution of crustal thickness corresponds to a laterally uneven distribution of

density that is the source of those buoyancy forces (see e.g. Husson and Ricard, 2004; Ricard and Husson, 2005, and many others, following Artyushkov, 1973, or Fleitout and Froidevaux, 1982).

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

87

Fig. 6. Model-inspired conceptual model of rift evolution in the heterogeneous and homogeneous cases. See text for comparison with the eastern and western Gulf of Aden. See text for detailed comments. The stages c), d), g), and h) (OCT formation and onset of accretion) are extrapolated from the analog model results. It is considered that the OCT and oceanic segments will initiate within the stretched-most basins. The position of transfer faults is observed while the position of transform is conceptual, considering that the oceanic segments will be oriented perpendicular to the imposed divergence. Thus, the segments size and number of transform is deduced from the geometry of the OCT. In red, the structures that initiate for each stage.

88

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

The lithosphere is first exposed to 020°E trending far field forces: en-échelon basins controlled by E–W to 110°E normal faults delineate the proto-Gulf of Aden. Thinning causes buoyancy forces to gradually build up and eventually overcome far field forces; this leads to the development of 070°E rift-parallel normal faults. Since rift-parallel faults are mainly observed in lithospheric-scale models, their development may be linked to the main lateral density variations that produce buoyancy forces, which act perpendicularly to the oblique rift, as suggested in Autin et al. (2010b). The crustal buoyancy forces have a delocalizing effect, as they tend to remove crustal thickness variations (see e.g. Davis and Kusznir, 2002). Thus, they may produce some rift-parallel faults without any significant localizing effect. On the contrary, thermal buoyancy forces (Burov, 2007; Davis and Kusznir, 2002; Huismans et al., 2001) may have a localizing effect if the density of the asthenosphere is lower than that of the mantle lithosphere, as a consequence of temperature difference especially. This process was put aside in the models of Autin et al. (2010b), where the asthenosphere is denser than the mantle lithosphere. In these models, the asthenosphere/mantle lithosphere density contrast may thus help to initiate rift-parallel faults because of this buoyancy instability but without any localizing effect. In these models, crustal buoyancy instabilities (due to the crustal thickness variations) may also have helped to initiate rift-parallel faults. As thinning proceeds, buoyancy forces decrease in the rift center and far field tractions dominate back, exciting 110°E trending normal faults. This suggests that the far-field extension governs the final stages of crustal thinning in the center of the rift. In the analog models, in the thinnest part of the rift, i.e. in the center, lithospheric thickness variations are not significant anymore. Similarly, in Davis and Kusznir (2002), pressure gradients calculated from crustal and lithospheric thicknesses and associated buoyancy forces are the highest at transition zones between stretched and unstretched part of the continental lithosphere. In the rift center, thickness variations are negligible; only the far field regional extension applies, producing purely divergence-perpendicular faults. Moreover, the different structural evolution between the eastern and western Gulf of Aden yields complementary information that are useful to corroborate this hypothesis. In the eastern Gulf of Aden, Bellahsen et al. (2006) and Autin et al. (2010b) proposed that buoyancy forces were active during rifting. This is witnessed by faults trending parallel to the rift axis. In the western Gulf of Aden, the normal faults as well as the OCT and the oceanic ridge cross cut the syn-rift basins through structures running parallel to the rift axis, while in the eastern part, the OCT and ridge initiate within syn-rift basins (Fig. 2). This difference resembles the structural evolution of oceanic spreading ridges that we can extrapolate from the two above-described analog models (Fig. 6): in models with a uniform rheological structure, the ridge presents few divergence perpendicular segments separated by few long transform faults, while in the model with a preexisting oblique structure, the ridge is very segmented and its overall trend is strongly oblique to the divergence. Thus, the evolution of the eastern Gulf of Aden may have been close to that observed in the oblique rift model of Autin et al. (2010b), wherein the lithosphere has a uniform rheological structure. On the contrary, the western part may resemble the oblique rift model that embeds a pre-existing, rift-parallel density heterogeneity, and thus where the buoyancy forces are stronger. Of course, we do not suggest that there has been a preexisting lithospheric weakness that localized the rifting in the west, as there is no such evidence. We therefore regard this model as representative of an oblique rift with enhanced buoyancy forces. 6. Buoyancy dynamics As suggested in the previous section, the two groups of faults may mirror the directions of the two principal forces at play, namely the far-field tectonic forces, chiefly related to the subduction of the Arabian plate underneath Eurasia (e.g. Bellahsen et al., 2003), and the buoyancy

forces that are due to lateral variations in the density/thickness of the lithosphere. The first one approximately strikes 020°E and corresponds to the motion of Arabia towards the subduction zone. Some additional – yet difficult to quantify – shear traction may apply at the base of the lithosphere. The buoyancy forces are due to the gravitational potential energy and vary as rifting evolves. It is a direct consequence of lithospheric thinning. The time evolution of those two forces may control the distribution of stresses within the crust, and as such, the tectonic evolution of the rift (e.g. Burov, 2007; Davis and Kusznir, 2002; Huismans et al., 2001). In order to investigate the interplay between the two contributors, we use a simplified, semi-analytical model, wherein the lithosphere is treated as a thin viscous sheet (e.g. England and McKenzie, 1982; Houseman and England, 1993; Husson and Ricard, 2004). This approach is based on the vertical integration of the Navier–Stokes equations coupled with mass (and optionally heat) conservation. We use the same formalism as Ricard and Husson (2005), who not only consider the compositional buoyancy of the crust, whose role has been recognized since Argand (1924), but also the thermal buoyancy of the lithosphere, viewed as a thermal boundary layer. Deformation in the lithosphere is þ∞

excited by the far-field force Σ and the moment M ¼ ∫ δρ g z dz that is 0

exerted by the lithospheric density anomalies within the lithosphere, such that 4L

∂ ∂u ∂M η ¼ −Σ ∂x ∂x ∂x

where L is the thickness of the lithosphere and η is the viscosity. The total moment M further splits into the compositional Mc and thermal Mθ components of the density anomalies, with Mc ¼


 ρc g ρ 2 1− c S 2 ρm

and þ∞

Mθ ¼ ∫ ρm gzαθdz; 0

where ρc and ρm are the crust and sub-lithospheric mantle densities (set to 2800 and 3200 kg/m3, respectively), g the gravitational acceleration, S the crustal thickness. z is vertical direction, α the coefficient of thermal expansion (set to 2.4 · 10−5 K−1), θ the temperature. Upon this condition, the two terms are opposite, and as such their contributions to the tectonic evolution greatly vary. Last, mass and heat conservation yield the transport equations of Mc and Mθ: ∂Mc ∂M c ∂u ∂Mθ ∂M θ ∂u ∂2 Mθ ; þu þ 2M c þu þ 2Mθ ¼ 0 and ¼K ∂t ∂x ∂t ∂x ∂x ∂x ∂x2 where t is time, u the horizontal velocity and K the thermal diffusivity (10−6 m2/s). Note that the transport of the thermal moment differs from that of the compositional moment because of thermal diffusion (see Ricard and Husson, 2005, for details). Our setup (Fig. 7a) accounts for the contributions of both Mc and Mθ, and allows for basal shear to be modulated. Lateral boundary conditions are dynamic (setting a value for the normal force at the right and left ends of the lithosphere) and lateral viscosity variations can be included. Rifting initiates in the middle of the model thanks to a small initial weakness. We define this weakness by a locally lower thermal moment, but a viscosity drop would do the same. We compute the crustal buoyancy force as the difference between the minimum and maximum values of Mc across the model. The crustal and thermal moments are due to the density contrasts between the mantle and the crust in the one hand, between the mantle and

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

the thermal lithosphere, respectively. As such, they are of opposite signs, and thinning or thickening or both cause competing effects (e.g. Ricard and Husson, 2005). The lateral variations of the crustal moment cause more extension in the edges than in the center of the rift. Conversely, the thermal moment induces more extension in the center of the rift, having a localizing effect. In the following, we explore whether the contribution of the buoyancy forces may overcome that of the far-field force and modify the tectonic framework. There is no obvious reason for far-field forces to have changed throughout the time of rifting, and we therefore assume them constant. Thus, any observed change in the structural evolution, as recorded by the fault pattern, is interpreted as resulting from a change in the crustal buoyancy force, that we measure as the difference in the crustal moments from the rift shoulder to the center. Thermal buoyancy forces do play a role in the crustal structural evolution, but to an unknown extent, mainly because the coupling between the lithospheric mantle and the crust is arguable. This makes us drop their contributions to the crustal structural development. It nevertheless remains accounted for in the computation of the rifting localization process. Fig. 7b depicts the time evolution of the crustal buoyancy force through time upon various conditions. We vary the stress condition

89

at the lateral ends of the lithosphere from stress free (0 TN/m, solid curves) to 2 TN/m (dashed), and 4 TN/m (dotted). The viscosity of the lithosphere is either uniform (1022 Pa s, green) or locally reduced tenfold (sine function over 200 km, reaching a minimum value of 1021 Pa s in its center, red and blue). Red curves depict a situation where the base of the lithosphere is stress free, while blue curves correspond to a situation where a basal shear traction T resists the motion of the lithosphere. In that case, shear stresses T scale linearly with velocity u, such that T = 0.5 MPa/(cm/yr). When the tectonic forces are null, only the buoyancy component drives the system. The crustal buoyancy force gradually grows and reaches its maximum value when breakup occurs (crustal moment decreases to zero in the center of the rift). Introducing tension at the lateral ends speeds up the process but after some delay, it decreases the crustal buoyancy force by stretching the entire lithosphere (Fig. 7c). When lateral viscosity variations are introduced, stretching focuses in the center of the rift and the lateral gradient of the moment increases, as well as the crustal buoyancy force that may reach larger values (Fig. 7c and f). Last, resistance to displacement (basal shear) improves this effect. A Rayleigh–Taylor instability appears and destabilize the system, provided that edge tension is low enough (blue, Fig. 7e).

Fig. 7. a) Sketch showing the rifting of a lithosphere submitted to boundary forces (including far-field Σ and shear traction Τ) and buoyancy forces arising from the crustal moment Mc and thermal moment Mθ. b) Maximum force exerted by the compositional moment through time for a uniformly viscous lithosphere (green), and lithosphere with a tenfold viscosity drop in the center of the rift (red and blue). Lateral boundaries are either stress free (solid) or submitted to a tension of 2 TN/m (dashed) or 4 TN/m (dotted). The base of the lithosphere is either stress free (green and red) or undergoing a resisting shear traction (that scales with horizontal motion, see text, blue). c, d, e) Lateral variations of the compositional moment at 10 Myr and 20 Myr for the three setups. f) Comparison between the lateral variations of the compositional moment for comparable stretching.

90

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

In summary, our analysis of the competition between the buoyancy forces and the far-field forces indicate that rift-parallel faults, which are assumed to be caused by crustal buoyancy forces, would be promoted by local softening of the lithosphere (thermal weakening) but hampered by far-field forces that tend to moderate their influence by thinning the lithosphere at a larger scale. Because crustal buoyancy forces are assumed to control the structural evolution, some relative timing of the tectonic framework can be proposed from our experiments. Note that, although plausible in our results, the absolute chronology is not relevant in the sense that time in our model mostly scales with the viscosity of the lithosphere, here set to 1022 Pa s, but for which the uncertainty is higher than one order of magnitude (moreover, time evolution also depends in our models, but to a minor extent, on the thermal diffusivity). We therefore emphasize that only the relative timings are relevant. These results suggest that during a first stage of extension, far field forces dominate the system, before buoyancy forces do apply, and possibly overwhelm the far field forces. Purely active rifts (only excited by their own buoyancies, solid curves) more efficiently grow buoyancy forces. But above all, lateral viscosity variations foster the process (Fig. 7). Not only does the magnitude of the force increases, but also it happens in a shorter time. If rifting occurs more efficiently than wholesale stretching of the rest of the lithosphere due to far field tension (Σ > 0 TN/m), the tectonic framework may evolve from a far-field dominated regime (i.e. ~020°E, for the Gulf of Aden) to a buoyancy dominated regime (i.e. riftperpendicular, ~160°E), and then back to a far-field regime. Another aspect of our tectonic analysis is the along strike variation of the structural style along the Gulf of Aden. In particular, the western Gulf seems (i) more dissected by 070°E striking faults than the eastern Gulf of Aden and (ii) to have a sharpest, more localized distal deformation field. We thus here explore the effect of a hotter asthenosphere and associated lower density due to the Afar hot spot activity. First, density is temperature-dependent. As such, one may expect that it changes the buoyancy forces. However, (i) mass conservation implies that the total load in the lithosphere remains identical and (ii) density changes are small (~10 kg/m3 for a 200 K change in rock temperature). Density variations play a localizing role upon certain conditions: for instance, when the hot asthenosphere rises below a not-yet-warmed lithosphere (Burov, 2007; Huismans et al., 2001), but most important is the temperature dependence of viscosity. This metric is certainly poorly known, as it depends on many factors that are themselves poorly accessible. However, the thermal dependence is most likely exponential and minor changes in the temperature convert into large viscosity variations. In the Gulf of Aden, heat flow measurements show an increase by some 50 mW/m2 from the eastern Gulf to the western Gulf (Fig. 1), which converts into a mean lithospheric temperature change of several 100s of Kelvin, consistently with the model predictions of Schmeling and h i Ea Ea Wallner (2012). Temperature dependence writes η ¼ η0 exp RT , − RT 0 where Ea is the activation energy, R is the gas constant, T the temperature and η0 the viscosity at reference temperature T0 which indicates that the asthenosphere viscosity drop in the western Gulf of Aden likely exceeds an order of magnitude. Our results (Fig. 7) show that lateral viscosity changes drastically foster the dynamics of rifting. In particular, buoyancy forces reach high values early in the history of rifting. This implies that, for a given amount of extension, the buoyancy forces will dominate the structural framework when the viscosity drops in the rift.

7. Discussion In this contribution, we aim at documenting and discussing both the structural evolution of the Gulf Aden and the effect of the Afar hot spot. We have shown that the effect of the hot spot seems significant in the late stage of the rifting and during break up processes. Indeed, during the early syn-rift times, which can be documented in the onshore proximal margins, the evolution in the western part is similar to the one in the eastern part: the early geometry and structural evolution is mainly controlled by the reactivation of inherited Mesozoic basins (Fig. 1). The differences between the eastern and the western parts are significant during the late rifting evolution, OCT emplacement, and spreading ridge evolution. Late syn-rift influence of magmatism is consistent with (i) numerical modeling suggesting that late syn-rift melting occurs as a consequence of lithospheric deformation and not only related to the hot spot timing (van Wijk et al., 2001) and (ii) the hot spot paroxysm timing at 30 Ma, thus later than the rifting onset (34 Ma) both in the eastern (Leroy et al., 2012; Robinet et al., 2013; Roger et al., 1989) and the western Gulf of Aden (Watchorn et al., 1998), even if the hot spot has probably been active since 45 Ma (George et al, 1998). We proposed above that the Afar hot spot is a main parameter influencing the rifting evolution, since it may have heated the mantle. Two main questions arise: what triggers and what fosters the localization of the oblique rift, especially in the western Gulf of Aden? The first question is beyond the scope of the present contribution and we assume that the rift obliquity results from the interplay between Afro-Arabian intraplate stresses during the Tertiary and the influence of the Afar hot spot related weakness, as in Bellahsen et al. (2003). Such oblique rifting dynamics, in the absence of any preexisting oblique lithospheric weakness, was studied in Autin et al. (2010b, 2013-this volume). In the next sections, we discuss the mechanisms that may enhance rift localization and the possible role of buoyancy forces, temperature, and magmatism on the rifting, in this case of east–west asthenosphere temperature variation. In the last section, we discuss the consequences on the oceanic spreading ridge geometry. 7.1. Effects of a hotter-than-normal asthenosphere Near the Afar hot spot, the asthenosphere was and is still most likely hotter than further east: there is a clear low velocity body beneath the Afar (e.g. Hansen et al., 2012, Fig. 1) and the ridge is much higher in the western Gulf of Aden than in the eastern part for oceanic lithosphere of similar ages (White and McKensie, 1989). Moreover, the hot spot is considered to have influenced the rifting until approximately 50°E (Fig. 1) on the basis of the geographical extent of Tertiary lava flows mostly mapped offshore (Leroy et al., 2012). A hot mantle should be less dense than a “normal” one because of thermal expansion and possible depletion during partial melting (as melting first remove dense aluminous silicate phases) (e.g. Oxburgh and Parmentier, 1977; see Buck, 2004). What may have been the effect of a hotter asthenosphere on the lithosphere thinning dynamics? 7.1.1. Plume-derived melts Successive injections of plume-derived melts may reduce its viscosity (Callot et al., 2001; Saunders et al., 1992). This would clearly have a

Fig. 8. Gulf of Aden evolution. On the left, represents a simplified structural map of the Gulf at four times. In the center is a detailed structural map of the Dhofar area at three times. On the right, there are two cross-sections for each time, one in the western part, one in the eastern part. a) Rifting initiation. The basins are arranged en échelon, striking around 120°E; the stretching is rather distributed; the fault network and associated stretching is dominated by far-field forces (020°E extension). b) The rift localizes along the future Gulf axis (075°E). Such processes trigger buoyancy forces due to the lithosphere thinning and its associated spatial thickness variations. Such local forces create a buoyancy-dominated fault network (in red). Such faults are most probably very numerous in the western part of the Gulf. Unfortunately, very few accurate structural data are available. c) OCT formation. The rift localizes even more in the future distal part and an OCT initiates possibly with mantle exhumation. In the west, the OCT cuts across the syn-rift basins while it is rather parallel to syn-rift basins in the east. This suggests that buoyancy forces were stronger and active longer in the west. This is explained by the Afar hot spot influence that heated up the asthenosphere (and lithosphere). d) Present day time kinematics.

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

91

92

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

localizing effect, if the melt intrusions are localized (see Callot et al., 2001). Melt intrusions may be witnessed by lava flows extrusions, such as SDR that are indeed observed in the western Gulf of Aden (Tard et al., 1991) almost exclusively offshore, yet imaged onshore by Ahmed et al. (2013) as several localized small areas. Thus, they most likely trend obliquely to the early-reactivated inherited Mesozoic basins onshore, even though this still has to be better documented. This suggests that the melting processes participated to rift localization. However, they probably did not trigger the localization as the rift was already oblique when the lava flows emplaced, but they enhanced it. 7.1.2. Lithospheric dyking The melt intrusions may directly accommodate the stretching at significantly low stress (Buck, 2004). Buck (2004) underlined that the available force for rifting may not overcome the lithosphere resistance of a “normally” cold, thick and strong lithosphere (Kusznir and Park, 1987). Buck (2004) proposed that lithosphere dyking might occur under extensional forces an order of magnitude less than forces necessary for lithosphere thinning. Moreover, he showed that dyking during few Myr weakens the lithosphere enough to ensure its thinning. Buck (2006) argued that the rift straightness over large distance (larger than the thickness of brittle layer involved) might be an argument for dyke injection, as long dykes (longer than the brittle layer thickness) have to be relatively straight to remain connected to their source. The western Gulf of Aden may indeed be considered as relatively straight (Fig. 2). The normal faults on the onshore proximal margins are definitely not straight neither parallel to the Gulf axis but the distal normal faults, OCT and coastline are much more parallel between each other. Moreover, these features are oblique to the divergence. As the dyke is usually perpendicular to the least compressive stress, this may suggest that the least compressive stress during late rift localization and eventual dyking, in the western Gulf, was rift-perpendicular (as suggested above and below). This shows that the oblique rift localization in the western Gulf cannot initially be due to dyking processes only but that the latter may greatly enhance the localization. 7.1.3. Thermal weakening The asthenosphere temperature may, by conduction, increase the lithosphere thermal gradient and decrease the strength of the lithosphere. Saunders et al. (1992) suggested that conduction is unlikely to efficiently raise the isothermal surfaces in the lithosphere: only the lowest parts of the lithosphere may be affected. However, adding the effect of intruded melts and dykes (Buck, 2004) that may also heat up the lithosphere by conduction, it is possible that all magmatic processes along with the high asthenosphere temperature significantly heated the lithosphere, decreasing its strength. Assuming that such heating was efficient, the lithosphere rheology should change. In relatively cold settings, it may be assimilated to a 4-layer brittle–ductile sandwich where the upper mantle lithosphere is brittle and represents the most resistant part (e.g. Burov, 2010; Burov and Watts, 2006; Handy and Brun, 2004; Watts and Burov, 2003). Such model was challenged by Maggi et al. (2000) and Jackson (2002), but Precigout et al. (2007) reconciled the concept of a resistant mantle lithosphere with the lack of earthquakes in the mantle in various settings by introducing a localizing ductile uppermost mantle. In a hot setting (such as close to or above a hot spot), the strength of the lithosphere is reduced, especially the ductile parts. It is most likely that, in such setting, there is no (brittle) resistant mantle lithosphere. A strong temperature increase would probably lead to a 2-layer rheology where the most resistant part is the crust: this would lead to a relatively de-localized mode of rifting (Brun, 1999; Buck, 1991; Buck et al., 1999). It is obvious that in the Gulf of Aden, the deformation is not more distributed in the western part than in the eastern part (Fig. 2). Thus, hot spot activity during Oligocene times did not result in a distributed rifting in the western Gulf. Then, it seems that the main influence of the asthenosphere temperature was not a significant change of the

initial lithosphere rheology. However, as the rift started to localize (obliquely), thermal weakening was most likely more important inside the rift than outside the rift. Our buoyancy analysis presented above suggests that both crustal (see Section 7.2) and thermal buoyancy forces (see below) are enhanced in case of thermal weakening of the lithosphere. Moreover, for comparable stretching amounts, buoyancy forces are more focused in the rift center when lateral viscosity variations occur (Fig. 7f). 7.1.4. Thermal buoyancy forces Heat is transferred to the lithosphere first conductively but also eventually convectively and the thermal boundary layer, at the base of the lithosphere, may be thinned by thermal erosion from below (d'Acremont et al., 2003; Olson et al., 1988; Saunders et al., 1992; Spohn and Schubert, 1982, 1983; Storey et al., 1989; White, 1992; Withjack, 1979). The higher this contrast between the densities of the lithosphere and asthenosphere is, the stronger the forces are. Such forces may be active as soon as the asthenosphere is heated. Thus, once rifting initiates thermal buoyancy forces may enhance any lithosphere thickness variations and actively thin the lithosphere. However, temperature only causes mild density variations from one setting to the other, and the overall effect shall not be considered as a primordial explanation to variable kinematics. It may be the initial destabilizing parameter that is subsequently relayed by other ones. 7.2. Effect of crustal buoyancy forces Crustal buoyancy forces are not localizing forces (Buck, 1991; Buck et al., 1999; Davis and Kusznir, 2002; Newman and White, 1999). They tend to remove the thickness variations created in the crust by the far-field extension. As a consequence, in the Gulf of Aden as well as in other passive continental margins, crustal buoyancy forces were probably dominated by thermal buoyancy forces (Davis and Kusznir, 2002) and/or weakening of the lithosphere. On the contrary, wide rifts (sensu Buck, 1991) may be dominated by crustal buoyancy forces that act as de-localizing effect and result in the development of several hundred of meters wide rifts: this “local isostatic stress effect” or “crustal buoyancy” described in Buck et al. (1999) dominates over the lithospheric necking effect. The main structural observation discussed in this contribution is the presence of rift-parallel faults all along the Gulf of Aden. These faults witness that extensional stresses have been, during at least transient periods, perpendicular to the rift axis. In the previous section, we showed that thermal buoyancy forces, thermal weakening, and magmatism (dyking) might result in a (rapid) localization of the rifting obliquely to the divergence, especially in the western part of the Gulf. Such an oblique localization in the mantle lithosphere would also localize the crustal extensional deformation obliquely. Then, acting as a feedback mechanism, crustal buoyancy forces would be oriented perpendicular to the rift axis and would produce rift-parallel normal faults. 7.3. Scenario for the Gulf of Aden rifting evolution Assuming that buoyancy forces and oblique weakening may have occurred in the Gulf of Aden lithosphere, we can infer the following structural evolution for the Gulf of Aden. During an early rifting phase (early Oligocene), the extension is rather distributed yet disposed en échelon along the rift axis that displays a large oblique zone (Bellahsen et al., 2003) of rather distributed basins (inherited or not). The lithosphere starts thinning. In the Dhofar area, as well as in the western part (although it is much less documented), this step is attested by divergence-perpendicular faults, trending around 110°E, as well as intermediate E–W faults (Fig. 8a). During a second phase (Late Oligocene–early Miocene), the rift localizes obliquely. This is attested by rift-parallel normal faults, clearly

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

93

Fig. 9. Three types of opening in the Gulf of Aden. In the east, a Sheba-type ridge strikes parallel to long segments of the continental margins and OCT. Few (rather) large transform faults are active. In such domain, the asthenosphere is rather cold and buoyancy forces were rather weak (but active) during rifting. In the central part, an East Aden-type ridge is very segmented, with six short segments and seven transform faults or accommodation zones that shift the ridge northwards by successive offsets ranging from 14 to 17 km. The ridge overall trend is parallel to the OCT (parallel to the Gulf axis) but oblique to the syn-rift basins. This suggests active buoyancy forces due a hot asthenosphere, less dense than in the east, as a consequence of the Afar hot spot activity. However, the segmented ridge suggests that the hot spot influence either was not too strong in this area to avoid long magmatic segments or lower at spreading ridge onset than during rifting time. In the western most part of the Gulf, a West Aden-type ridge is long and straight (weakly segmented) even if the breakup occurred strongly oblique to the syn-rift basins. This is most likely due to the hot spot influence that implied a high melt supply resulting in a long ridge.

mapped onshore and offshore Dhofar (eastern Gulf of Aden, red faults on Fig. 8b). It is rather difficult to decipher which one of the abovecited localizing parameters is the most efficient during early rift localization. Melt intrusions, thermal buoyancy forces, and thermal weakening may altogether be active processes. However, one may think that thermal buoyancy forces may be active first (especially in the west) as melt intrusion and heat conduction may take some time to occur and be efficient. Yet, at some point, it is most likely that the two latter processes get stronger and drive the rift localization. At this time, as the rift localizes, the weakening of the lithosphere and thermal buoyancy forces help more efficiently the localization of the rift in the west than in the east (Fig. 8b). This is most likely due to the Afar hot spot that has been active since at least 45 Ma (George et al, 1998), with a paroxysmal activity at 30 Ma (Hofmann et al., 1997). Le Garzic et al. (2011) showed that, in the Mukalla region (Southern Yemen), the main fault population is trending around 070°E. Conversely, in the eastern Gulf of Aden, far-field forces may prevail longer and impose that 110°E trending faults may dominate throughout rifting, as observed on the structural map (Fig. 2).

This step may correspond to the “thinning phase” proposed in Manatschal (2004) and Lavier and Manatschal (2006). During this step, the mantle is significantly thinned, and the crust and mantle lithosphere are getting coupled. When the crust and mantle are coupled (likely by removal and thinning of the lower or middle crust), the localizing effects in the mantle lithosphere, discussed above, gain a stronger influence on rift localization in the crust. This step may also correspond to the active rift step in Huismans et al. (2001), although these authors showed, from numerical models, that it was only active since late syn-rift to early post-rift times. In our models, this phase is clearly syn-rift and the extension is “buoyancy-dominated” (Fig. 8b). During a third phase, the rifting localizes even more into the distal parts. This phase may correspond to the exhumation phase of Manatschal (2004) and Lavier and Manatschal (2006), as mantle exhumation as been proposed for the eastern Gulf (Autin et al., 2010a; d'Acremont et al., 2006; Leroy et al., 2010a; Watremez et al., 2011). In the eastern Gulf of Aden, the faults in the distal parts or within the OCT are 110°E (perpendicular to divergence, red faults on Fig. 8c), showing that stresses are “far-field stresses dominated”. There, in the

94

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

rift center, no significant lithosphere variations occur and no buoyancy forces are active. Differences between east and west are probably enhanced: in the west, the same (and strong) localizing processes (than during the second phase) still control the formation of the OCT that strikes parallel to the rift axis (075°E, Figs. 1 and 8). There, the continental margins are very narrow and, as a consequence, buoyancy forces are still active. 7.4. Present-day ridge segmentation The Fig. 2 shows the present-day ridge segmentation. In the eastern part of the Gulf of Aden, i.e. east of Alula Fartak F.Z., the Sheba ridge presents few segments. Between Alula Fartak and Socotra–Hadbeen F.Z., there are three segments. Eastward propagation of these ridge segments occurred since the oceanic spreading onset and until at least 9 Ma with a ridge jump at 11 Ma. This is due to a melting anomaly east of the Alula Fartak F.Z. (d'Acremont et al., 2010) most likely itself related to eastward channelization of hot material in the Aden ridge away from the Afar hot spot (Leroy et al., 2010b). The lengthening of the ridge segments in this area, from about 50 km long to more than 130 km long is thus due to the mantle temperature, enhanced and focused magmatic activity being related to a melting anomaly imaged within and below the oceanic lithosphere (Leroy et al., 2010b). West of Alula Fartak F.Z., two areas are defined: eastward, between longitude 46°E and 50°E (Fig. 2), the Aden oceanic spreading ridge is strongly segmented without large offset between segments; the segment length ranges between 10 and 40 km. Westward, a long E–W trending oceanic spreading segment (Fig. 2, more than 150 km long) presents only one transform F.Z. (Shukra El Sheik F.Z., Fig. 2) inactive since 9 Ma (Leroy et al., 2012). The kinematics of this long segment is thus oblique (Dauteuil et al., 2001): there is an angle of about 55° between the ridge and the divergence (035°E). To briefly sum up, there are zones where the segments are long (west of longitude 46°E and between longitude 51°E and 53°E) and zones where the ridge displays short segments (between longitude 46°E and 50°E, and between longitude 53°E and 54°E). Sauter et al. (2001) proposed that, in slow-spreading ridges, the spacing of spreading cells (and then of the oceanic spreading segments) is controlled by the spreading rate with larger spacing in ultraslow-spreading ridges than on slow-spreading ridges. However, in the Gulf of Aden, variations in spreading rate from west to east are too small (from 13 to 18 mm/yr, respectively) to explain the ridge segmentation variations. Are these variations in accordance with the results proposed above for rifting times? Does the temperature variation in the mantle from west to east also explain these ridge geometry and kinematics changes? In the western most part (West Aden-type ridge, Figs. 2 and 9), in the axial valley, the faults and basins are disposed en échelon and are overlapping each other (Dauteuil et al., 2001) forming non-transform discontinuities associated with oblique topography such as in the Mid Atlantic Ridge (e.g. Dauteuil and Brun, 1993, 1996; Grindlay et al., 1991). The oceanic ridge is long (about 150 km long) and close to the Afar hot spot. Thus, its length may be linked to the high mantle temperature due to the hot spot influence, as suggested by the extent of the present-day low velocity body in the mantle at 400 km beneath Afar (white line on Fig. 1a). Worldwide, ridge interaction with a hot spot shows a quite similar configuration: for example, south of the Iceland, the present-day Reykjanes ridge is long and unsegmented while it was strongly segmented between anomalies 13 and 19 (Vogt and Avery, 1974). This is interpreted as due to long‐term variations in mantle temperature, with periods of ∼30–50 °C warmer mantle flowing away from hot spot producing unsegmented ridges (e.g., Jones et al., 2002; Smallwood and White, 2002; White, 1997; White et al., 1995). Similarly, south of the Azores hot spot, oceanic spreading segments propagated southward and got longer during a high temperature/high melting period that implied high amount of magmatism (Cannat et al., 1999). Further away from hot spots, oceanic spreading center lengthening also

occurs in case of enhanced and focused magmatism (d'Acremont et al., 2010; Leroy et al, 2010b; Rabain et al., 2001). Thus, in the western most Gulf of Aden, the high mantle temperature due to the Afar hot spot activity most likely implies high melting and magmatism amount and, as a consequence, rather long oceanic spreading segments with non-transform discontinuities. In the central and the eastern parts, the ridge is much more segmented with transform faults (east of longitude 46°E). At the first order, this can be explained by the decreasing influence of the hot spot and the associated thermal anomaly and melt supply. However, second order differences can be observed between the central and the eastern parts. In the eastern part, the OCT formed parallel to the syn-rift structures (Figs. 2 and 9). Thus, the oceanic ridge initiated also parallel to the late syn-rift structures that are themselves perpendicular to divergence. As a consequence, there is no need for strong segmentation to accommodate the obliquity: few transform F.Z initiated at late syn-rift transfer faults (d'Acremont et al., 2005; Leroy et al., 2010a). After spreading onset, the segmentation evolved mainly with magma supply variations and kinematics changes (just east of Alula Fartak F. Z., d'Acremont et al., 2010). In the central part (East Aden-type ridge, Figs. 2 and 9), the OCT formed oblique (rift-parallel) to the inherited reactivated basins and to early newly formed syn-rift basins. The oceanic spreading ridge is strongly segmented and is probably the most intriguing part of the ridge in terms of geometry. It is still probably under the Afar hot spot influence as suggested by the eastern extent of syn-rift influence of the hot spot (dashed white line, Fig. 1a). Such limit is at the eastern limit of recognized syn-rift volcanism (Leroy et al., 2012). From that point eastward, the hot spot influence may still been active although decreasing. In such context, the ridge geometry may be explained as follows. The mantle temperature was not anomalously high enough to generate very large amount of melting and thus did not imply long magmatic oceanic spreading centers with non-transform faults: here segments are offset by transform faults. Assuming this, the ridge geometry was probably constrained by the inherited OCT structure: the OCT is, in this area, strongly oblique to the divergence; the ridge had to be very segmented in order not to be a long and oblique-to-divergence spreading center (see similar geometries in Fig. 6h). 8. Conclusions We have presented a synthesis of structural data as well as new structural interpretations of onshore–offshore data from the Gulf of Aden oblique rifting, along with a synthesis of published analog models of oblique rifting. We propose that the fault evolution on the Gulf continental margins have been strongly controlled by buoyancy forces that superimposed on far-field extension. We further suggest that, in the western part of the Gulf, the thermal anomaly due to the impinging of the Afar plume strongly localized the strain and fostered the development of rift-parallel faults. In this part of the Gulf, both buoyancy forces and lithosphere weakening explain the oblique-to-divergence rift localization. Similarly, the OCT is aligned along the Gulf axis. In the western most part, the present-day Afar hot spot influence results in a long oceanic spreading ridge, consistent with high amount of melt supply, while in a more central part, the ridge is strongly segmented. The latter is probably influence by the oblique-to-divergence OCT that promoted short and segmented spreading centers, consistent with decreased mantle temperature. Conversely, in the eastern Gulf of Aden, the development of rift-parallel faults is only a transient process, restricted to late syn-rift times. As a consequence, the OCT and the oceanic ridge segments strike perpendicular to the divergence with only few long transform faults. Acknowledgments We thank F. Roure and two anonymous reviewers for their comments and criticisms that considerably improved the original

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

manuscript. We also thank Gwenn Peron-Pinvidic and Pjer Osmundsen for their great work as invited editors. This work was funded by the YOCMAL ANR project and in the framework of Actions Marges.

References Agostini, A., Corti, G., Zeoli, A., Mulugeta, G., 2009. Evolution, pattern and partitioning of deformation during oblique continental rifting: inferences from lithospheric-scale centrifuge models. Geochemistry, Geophysics, Geosystems 10 (Q11015). http:// dx.doi.org/10.1029/2009GC002676. Ahmed, A., Tiberi, C., Leroy, S., Stuart, G.W., Keir, D., Sholan, J., Khanbari, K., Al-Ganad, I., Basuyau, C., 2013. Crustal structure of the rifted volcanic margins and uplifted plateau of Western Yemen from receiver function analysis. Geophysical Journal International. http://dx.doi.org/10.1093/gji/ggt072. Argand, E., 1924. La tectonique de l'Asie, Congrès Géologique International, Comptes-rendus de la XIII Session (596 pp., Brussels). Artyushkov, E.V., 1973. Stresses in the lithosphere caused by crustal thickness inhomogeneities. Journal of Geophysical Research 78, 7675–7708. Audin, L., Quidelleur, X., Coulié, E., Courtillot, V., Gilder, S., Manighetti, I., Gillot, P., Tapponnier, P., Kidane, T., 2004. Paleomagnetism and K–Ar and 40Ar/39Ar ages in the Ali Sabieh area (Republic of Djibouti and Ethiopia): constraints on the mechanism of Aden ridge propagation into southeastern Afar during the last 10 Myr. Geophysical Journal International 158, 327–345. Autin, J., Leroy, S., Beslier, M.-O., d'Acremont, E., Razin, P., Ribodetti, A., Bellahsen, N., Robin, C., Al Toubi, K., 2010a. Continental break-up history of a deep magma-poor margin based on seismic reflection data (northeastern Gulf of Aden margin, offshore Oman). Geophysical Journal International 180 (2), 501–519. http://dx.doi.org/ 10.1111/j.1365-1246X.2009.04424.x. Autin, J., Bellahsen, N., Husson, L., Beslier, M.-O., Leroy, S., d'Acremont, E., 2010b. Analogue models of oblique rifting in a cold lithosphere. Tectonics 29, 1–23. Autin, J., Bellahsen, N., Leroy, S., Husson, L., Beslier, M.O., d'Acremont, E., 2013. The role of structural inheritance in oblique rifting: insights from analogue models and application to the Gulf of Aden. Tectonophysics 607, 51–64 (in review, this volume). Bache, F., Leroy, S., Baurion, C., Robinet, J., Gorini, C., Lucazeau, F., Razin, P., d'Acremont, E., Al-Toubi, K., 2011. Post-rift uplift of the Dhofar margin (Gulf of Aden). Terra Nova 23, 11–18. Bastow, I.D., Stuart, G.W., Kendall, J.M., Ebinger, C.J., 2005. Upper-mantle seismic structure in a region of incipient continental breakup: northern Ethiopian rift. Geophysical Journal International 162, 479–493. Basuyau, C., Tiberi, C., Leroy, S., Stuart, G., Al-Lazki, A., Al-Toubi, K., Ebinger, C., 2010. Evidence of partial melting beneath a continental margin: case of Dhofar, in the northeast Gulf of Aden (Sultanate of Oman). Geophysical Journal International 180, 520–534. Bellahsen, N., Daniel, J.M., 2005. Fault reactivation control on normal fault growth: an experimental study. Journal of Structural Geology 27, 769–780. Bellahsen, N., Faccenna, C., Funiciello, F., Daniel, J.-M., Jolivet, L., 2003. Why did Arabia separate from Africa? Insights from 3-D laboratory experiments. Earth and Planetary Science Letters 216, 365–381. Bellahsen, N., Fournier, M., d'Acremont, E., Leroy, S., Daniel, J.-M., 2006. Fault reactivation and rift localization: the northeastern Gulf of Aden margin. Tectonics 25. http://dx.doi.org/10.1029/2004TC001626. Bellahsen, N., Leroy, S., Autin, J., Razin, P., d'Acremont, E., Khanbari, K., 2013. Preexisting oblique transfer zones and transfer/transform relationships in continental margins: new insights from the southeastern Gulf of Aden, Socotra Island, Yemen. Tectonophysics 607, 32–50 (this volume, in review). Benoit, M.H., Nyblade, A.A., Owens, T.J., Stuart, G., 2006. Mantle transition zone structure and upper mantle S velocity variations beneath Ethiopia: evidence for a broad, deepseated thermal anomaly. Geochemistry, Geophysics, Geosystems 7, 1–16. Berndt, C., Planke, S., Alvestad, E., Tsikalas, F., Rasmussen, T., 2001. Seismic volcanostratigraphy of the Norwegian Margin: constraints on tectonomagmatic break-up processes. Journal of the Geological Society 158, 413–426. Beslier, M.-O., Cornen, G., Girardeau, J., 1996. Tectono-metamorphic evolution of peridotites from the ocean/continent transition of the Iberia abyssal plain margin. In: Whitmarsh, R.B., et al. (Ed.), Proc. ODP, Sci. Results, pp. 397–412. Beslier, M.-O., Royer, J.-Y., Girardeau, J., Hill, P.J., Boeuf, E., Buchanan, C., Chatin, F., Jacovetti, G., Moreau, A., Munschy, M., Partouch, C., Robert, U., Thomas, S., 2004. A wide ocean– continent transition along the south-west Australian margin: first results of the MARGAU/MD110 cruise. Bulletin de la Société Géologique de France 175, 629–641. Birse, A.C.R., Bott, W.F., Morrison, J., Samuel, M.A., 1997. The Mesozoic and early Tertiary tectonic evolution of the Socotra area, eastern Gulf of Aden, Yemen. Marine and Petroleum Geology 14 (6), 675–684. Boillot, G., Recq, M., Winterer, E.L., Meyer, A.W., Applegate, J., Baltuck, M., Bergen, J.A., Comas, M.C., Davies, T.A., Dunham, K., Evans, C.A., Girardeau, J., Goldberg, G., Haggerty, J., Jansa, L.F., Johnson, J.A., Kasahara, J., Loreau, J.P., Luna-Sierra, E., Moullade, M., Ogg, J., Sarti, M., Thurow, J., Williamson, M., 1987. Tectonic denudation of the upper mantle along passive margins: a model based on drilling results (ODP leg 103, western Galicia margin, Spain). Tectonophysics 132, 335–342. Bonatti, E., Seyler, M., Channell, J., Girardeau, J., Mascle, G., 1990. Peridotites drilled from the Tyrrhenian Sea. In: Cita, M.B., et al. (Ed.), Proc. ODP, Sci. Results, pp. 37–47. Bos, B., Spiers, C.J., 2002. Frictional-viscous flow of phyllosilicate-bearing fault rock: microphysical model and implications for crustal strength profiles. Journal of Geophysical Research 107. http://dx.doi.org/10.1029/2001JB000301.

95

Brannan, J., Gerdes, K.D., Newth, I.R., 1997. Tectono-stratigraphic development of the Qamar basin, eastern Yemen. Marine and Petroleum Geology 14, 701–730. 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. Buck, W.R., 1991. Modes of continental lithospheric extension. Journal of Geophysical Research 96, 20161–20178. Buck, W.R., 2004. Consequences of asthenospheric variability on continental rifting. In: Karner, G.D., Taylor, B., Driscoll, N.W., Kohlstedt, D.L. (Eds.), Rheology and Deformation of the Lithosphere at Continental Margins. Columbia University Press, pp. 1–31. Buck, W.R., 2006. The role of magma in the development of the Afro-Arabian rift system. Geological Society, London, Special Publications 259, 43–54. http://dx.doi.org/ 10.1144/GSL.SP.2006.259.01.05. Buck, W.R., Lavier, L., Poliakov, A.N.B., 1999. How to make a rift wide? Philosophical Transactions of the Royal Society of London 357, 671–693. Burov, E., 2007. The role of gravitational instabilities, density structure and extension rate in the evolution of continental margins. In: Karner, G.D., Manatschal, G., Pinheiro, L.M. (Eds.), Imaging, Mapping and Modelling Continental Lithosphere Extension and Breakup: Geological Society, London, Special Publications, 282, pp. 139–156. http://dx.doi.org/10.1144/SP282.7. Burov, E.B., 2010. The equivalent elastic thickness (Te), seismicity and the long-term rheology of continental lithosphere: time to burn-out “crème brûlée”? Insights from large-scale geodynamic modeling. Tectonophysics 484, 4–26. Burov, E., Watts, A.B., 2006. The long-term strength of continental lithosphere “jelly sandwich” or “crème-brûlée”? GSA Today 16, 4–10. Callot, J.P., Grigné, C., Geoffroy, L., Brun, J.P., 2001. Development of volcanic margins: two-dimensional laboratory models. Tectonics 20, 148–159. Cannat, M., Briais, A., Deplus, C., Escartın, J., Georgen, J., Lin, J., Mercouriev, S., Meyzen, C., Muller, M., Pouliquen, G., Rabain, A., da Silva, P., 1999. Mid-Atlantic Ridge– Azores hotspot interactions: along-axis migration of a hotspot-derived event of enhanced magmatism 10 to 4 Ma ago. Earth and Planetary Science Letters 173, 257–269. Chang, S.J., Van der Lee, S., 2011. Mantle plumes and associated flow beneath Arabia and East Africa. Earth and Planetary Science Letters 302, 448–454. Clifton, A.E., Schlische, R.W., Withjack, M.O., Ackermann, R.V., 2000. Influence of rift obliquity on fault-population systematics: results of experimental clay models. Journal of Structural Geology 22, 1491–1509. Conrad, C.P., Molnar, P., 1997. The growth of Rayleigh–Taylor-type instabilities in the lithosphere for various rheological and density structures. Geophysical Journal International 129, 95–112. Corti, G., 2004. Centrifuge modelling of the influence of crustal fabrics on the development of transfer zones: insights into the mechanics of continental rifting architecture. Tectonophysics 384 (1–4), 191–208. Corti, G., 2009. Continental rift evolution: from rift initiation to incipient break-up in the Main Ethiopian Rift, East Africa. Earth-Science Reviews 96, 1–53. Corti, G., 2012. Evolution and characteristics of continental rifting: analog modelinginspired view and comparison with examples from the East African Rift System. Tectonophysics 522, 1–33. Corti, G., 2013. Control of rift obliquity on the evolution and segmentation of the main Ethiopian rift. Nature Geosciences. http://dx.doi.org/10.1038/ngeo160 (in press). Corti, G., Bonini, M., Innocenti, F., Manetti, P., Mulugeta, G., 2001. Centrifuge models simulating magma emplacement during oblique rifting. Journal of Geodynamics 31 (5), 557–576. Corti, G., Bonini, M., Conticelli, S., Innocenti, F., Manetti, P., Sokoutis, D., 2003. Analogue modelling of continental extension: a review focused on the relations between the patterns of deformation and the presence of magma. Earth-Science Reviews 63 (3–4), 169–247. d'Acremont, E., Leroy, S., Burov, E.B., 2003. Numerical modelling of a mantle plume: the plume head–lithosphere interaction in the formation of an oceanic large igneous province. Earth and Planetary Science Letters 206, 379–396. d'Acremont, E., Leroy, S., Beslier, M.-O., Bellahsen, N., Fournier, M., Robin, C., Maia, M., Gente, P., 2005. Structure and evolution of the eastern Gulf of Aden conjugate margins from seismic reflection data. Geophysical Journal International 160, 869–890. d'Acremont, E., Leroy, S., Maia, M., Patriat, P., Beslier Marie, O., Bellahsen, N., Fournier, M., Gente, P., 2006. Structure and evolution of the eastern Gulf of Aden; insights from magnetic and gravity data (Encens-Sheba MD117 cruise). Geophysical Journal International 165 (3), 786–803. d'Acremont, E., Leroy, S., Maia, M., Gente, P., Autin, J., 2010. Volcanism, jump and propagation on the Sheba ridge, eastern Gulf of Aden: segmentation evolution and implications for oceanic accretion processes. Geophysical Journal International 180 (2), 535–551. http://dx.doi.org/10.1111/j.1365-1246X.2009.04448.x. Dauteuil, O., Brun, J.-P., 1993. Oblique rifting in a slow-spreading ridge. Nature 361, 145–148. Dauteuil, O., Brun, J.-P., 1996. Deformation partitioning in a undergoing oblique extension: Mohns Ridge, Norwegian Sea. Tectonics 15, 870–884. Dauteuil, O., Huchon, P., Quemeneur, F., Souriot, T., 2001. Propagation of an oblique spreading centre: the western Gulf of Aden. Tectonophysics 332, 423–442. Davis, M., Kusznir, N., 2002. Are buoyancy forces important during the formation of rifted margins? Geophysical Journal International 149, 524–533. Dickinson, W.R., 2002. The basin and range province as a composite extensional domain. International Geology Review 44, 1–38. Ebinger, C.J., Casey, M., 2001. Continental breakup in magmatic provinces: an Ethiopian example. Geology 29 (6), 527–530. Ebinger, C.J., Yemane, T., Woldegabriel, G., Aronson, J.L., Walter, R.C., 1993. Late Eocene–Recent volcanism and faulting in the southern main Ethiopian rift. Journal of the Geological Society 150 (1), 99–108. England, P., McKenzie, D., 1982. A thin viscous sheet model for continental deformation. Geophysical Journal of the Royal Astronomical Society 70, 295–321.

96

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97

Fantozzi, P.L., 1996. Transition from continental to oceanic rifting in the Gulf of Aden: structural evidence from field mapping in Somalia and Yemen. Tectonophysics 259 (4), 285–311. Fleitout, L., Froidevaux, C., 1982. Tectonics and topography for a lithosphere containing density heterogeneities. Tectonics 1 (1), 21–56. Förster, A., Förster, H.J., Masarweh, R., Masri, A., Tarawneh, K., 2007. The surface heat flow of the Arabian Shield in Jordan. Journal of Asian Earth Sciences 30, 271–284. Fournier, M., Bellahsen, N., Fabbri, O., Gunnell, Y., 2004. Oblique rifting and segmentation of the NE Gulf of Aden passive margin. Geochemistry, Geophysics, Geosystems 5 (11) (24 pp.). Fournier, M., Chamot-Rooke, N., Petit, C., Huchon, P., Al-Kathiri, A., Audin, L., Beslier, M.O., d'Acremont, E., Fabbri, O., Fleury, J.-M., Khanbari, K., Lepvrier, C., Leroy, S., Maillot, B., Merkouriev, S., 2010. Arabia-Somalia plate kinematics, evolution of the Aden-Owen-Carlsberg triple junction, and opening of the Gulf of Aden. Journal of Geophysical Research 115. http://dx.doi.org/10.1029/2008jb006257. Galanis, S.P., Sass, J.H., Munroe, R.J., Abu-Ajamieh, M., 1986. Heat flow at Zerqa Ma'in and Zara and a geothermal reconnaissance of Jordan. US Geological Survey, Open File Report 86–63, Menlo Park. Gemmer, L., Houseman, G.A., 2007. Convergence and extension driven by lithospheric gravitational instability: evolution of the Alpine–Carpathian–Pannonian system. Geophysical Journal International 168, 1276–1290. George, R., Rogers, N., Kelley, S., 1998. Earliest magmatism in Ethiopia: evidence for two mantle plumes in one flood basalt province. Geology 26, 923–926. http:// dx.doi.org/10.1130/0091-7613. Gettings, M.E., 1982. Heat-flow measurements at shot points along the 1978 Saudi Arabian seismic deep-refraction line, Part II. Discussion and interpretation. Saudi Arabia Deputy Min. for Mineral Res., Open File Report USGS-OF-02-40. (40 pp.). Griffin, W., O'Reilly, S.Y., Afonso, J.C., Begg, G., 2009. The composition and evolution of lithospheric mantle: a re-evaluation and its tectonic implications. Journal of Petrology 50, 1185. Grindlay, N.R., Fox, P.J., Macdonald, K.C., 1991. Second-order ridge axis discontinuities in the South Atlantic: morphology, structure, and evolution. Marine Geophysical Researches 13, 21–49. Gueydan, F., Morency, C., Brun, J.P., 2008. Continental rifting as a function of lithospheric mantle strength. Tectonophysics 460, 83–93. Handy, M.R., Brun, J.P., 2004. Seismicity, structure and strength of the continental lithosphere. Earth and Planetary Science Letters 223, 427–441. Hansen, S.E., Nyblade, A.A., Benoit, M.H., 2012. Mantle structure beneath Africa and Arabia from adaptively parameterized P-wave tomography: implications for the origin of Cenozoic Afro-Arabian tectonism. Earth and Planetary Science Letters 319–320, 23–34. Hofmann, C., Courtillot, V., Féraud, G., Rochette, P., Yirgu, G., Ketefo, E., Pik, R., 1997. Timing of the Ethiopian flood basalt event and implications for plume birth and global change. Nature 389, 838–841. Holbrook, W.S., Larsen, H.C., Korenaga, J., Dahl-Jensen, T., Reid, I.D., Kelemen, P.B., Hopper, J.R., Kent, G.M., Izarralde, D., Bernstein, S., Detrick, R.S., 2001. Mantle thermal structure and active upwelling during continental breakup in the North Atlantic. 190, 251–266. Houseman, G., England, P., 1993. Crustal thickening versus lateral expulsion in the Indian–Asian continental collision. Journal of Geophysical Research 98, 12233–12249. Huchon, P., Khanbari, K., 2003. Rotation of the syn-rift stress field of the northern Gulf of Aden margin, Yemen. Tectonophysics 364 (3–4), 147–166. Huismans, R.S., Beaumont, C., 2007. Roles of lithospheric strain softening and heterogeneity in determining the geometry of rifts and continental margins. Geological Society, London, Special Publications 282, 111–138. Huismans, R.S., Podladchikov, Y.Y., Cloethingh, S., 2001. Transition from passive to active rifting: relative importance of asthenospheric doming and passive extension of the lithosphere. Journal of Geophysical Research 106, 11,271–11,291. Husson, L., Ricard, Y., 2004. Stress balance above subduction: application to the Andes. Earth and Planetary Science Letters 222, 1037–1050. Jackson, J., 2002. Strength of the continental lithosphere: time to abandon the jelly sandwich? Geological Society of America Today 12, 4–10. Jestin, F., Huchon, P., Gaulier, J.M., 1994. The Somalia plate and the East African Rift System: present-day kinematics. Geophysical Journal International 116, 637–654. Jones, S.M., White, N., Clarke, B.J., Rowley, E., Gallagher, K., 2002. Present and past influence of the Iceland Plume on sedimentation. In: Dore, A.G., Cartwright, J., Stoker, M.S., Turner, J.P., White, N. (Eds.), Exhumation of the North Atlantic Margin: Timing, Mechanisms and Implications for Petroleum Exploration: Geol. Soc. Spec. Publ., 196, pp. 13–25. Kusznir, N.J., Park, R.G., 1987. The extensional strength of the continental lithosphere: its dependence on geothermal gradient, and crustal composition and thickness. In: Coward, M.P., Dewey, J.F., Hancock, P.L. (Eds.), Continental Extensional Tectonics: Geological Society, London, Special Publication, 39, pp. 35–52. Lavier, L., Manatschal, G., 2006. A mechanism to thin the continental lithosphere at magma-poor margins. Nature 440, 324–328. http://dx.doi.org/10.1038/nature04608. Le Garzic, E., de L'Hamaide, T., Diraison, M., Géraud, Y., Sausse, J., de Urreiztieta, M., Hauville, B., Champanhet, J.M., 2011. Scaling and geometric properties of extensional fracture systems in the Proterozoic basement of Yemen. Tectonic interpretation and fluid flow implications. Journal of Structural Geology 33, 519–536. Lepvrier, C., Fournier, M., Bérard, T., Roger, J., 2002. Cenozoic extension in coastal Dhofar (southern Oman): implications on the oblique rifting of the Gulf of Aden. Tectonophysics 357, 279–293. Leroy, S., et al., 2004. From rifting to spreading in the eastern Gulf of Aden: a geophysical survey of a young oceanic basin from margin to margin. Terra Nova 16 (4), 185–192.

Leroy, S., et al., 2010a. Contrasted styles of rifting in the eastern Gulf of Aden: a combined wide-angle, MCS and heat flow survey. Geochemistry, Geophysics, Geosystems. http://dx.doi.org/10.1029/2009GC002963. Leroy, S., d'Acremont, E., Tiberi, C., Basuyau, C., Autin, J., Lucazeau, F., Sloan, H., 2010b. Recent off-axis volcanism in the eastern Gulf of Aden: implications for plume–ridge interaction. Earth and Planetary Science Letters 293 (1–2), 140–153. Leroy, S., et al., 2012. From rifting to oceanic spreading in the Gulf of Aden: a synthesis. Arabian Journal of Geosciences 5, 859–901. http://dx.doi.org/10.1007/s12517-0110475-4. Lucazeau, F., Leroy, S., Bonneville, A., Goutorbe, B., Rolandone, F., d'Acremont, E., Watremez, L., Düsünur, D., Tuchais, P., Huchon, P., Bellahsen, N., Al-Toubi, K., 2008. Persistent thermal activity at the eastern Gulf of Aden after continental break-up. Nature Geoscience 854–858. http://dx.doi.org/10.1038/ngeo359. Lucazeau, F., Leroy, S., Rolandone, F., d'Acremont, E., Watremez, L., Bonneville, A., Goutorbe, B., Düsünur, D., 2010. Heat-flow and hydrothermal circulations at the ocean–continent transition of the eastern Gulf of Aden. Earth and Planetary Science Letters 295 (3–4), 554–570. http://dx.doi.org/10.1016/j.epsl.2010.04.039. Maggi, A., Jackson, J.A., Priestley, K., Baker, C., 2000. A reassessment of focal depth distributions in southern Iran, the Tien Shan and northern India: do earthquakes really occur in the continental mantle? Geophysical Journal International 143, 629–661. Manatschal, G., 2004. New models for evolution of magma-poor rifted margins based on a review of data and concepts from West Iberia and the Alps. International Journal of Earth Sciences 93, 432–466. http://dx.doi.org/10.1007/s00531-004-0394-7. 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. Mart, Y., Dauteuil, O., 2000. Analogue experiments of propagation of oblique rifts. Tectonophysics 316, 121–132. http://dx.doi.org/10.1016/S0040-1951(99) 00231-0. McClay, K.R., White, M.J., 1995. Analogue modelling of orthogonal and oblique rifting. Marine and Petroleum Geology 12, 137–151. Molnar, P., Houseman, G.A., 2004. The effects of buoyant crust on the gravitational instability of thickened mantle lithosphere at zones of intracontinental convergence. Geophysical Journal International 158, 1134–1150. Molnar, P., Houseman, G.A., Conrad, C.P., 1998. Rayleigh–Taylor instability and convective thinning of mechanically thickened lithosphere: effects of non-linear viscosity decreasing exponentially with depth and of horizontal shortening of the layer. Geophysical Journal International 133, 568–584. Montelli, R., Nolet, G., Dahlen, F.A., Masters, G., 2006. A catalogue of deep mantle plumes: new results from finite-frequency tomography. Geochemistry, Geophysics, Geosystems 7, 1–69. Newman, R., White, N., 1999. The dynamics of extensional sedimentary basins: constraints from subsidence inversion. Philosophical Transactions of the Royal Society of London A 357, 805–834. Olson, P., Schubert, G., Anderson, C., Goldman, P., 1988. Plume formation and lithosphere erosion: a comparison of laboratory and numerical experiments. Journal of Geophysical Research 93, 15065–15084. Oxburgh, E.R., Parmentier, E.M., 1977. Compositional and density stratification in oceanic lithosphere — causes and consequences. Journal of the Geological Society of London 33, 343–355. Pascal, C., Cloetingh, S.A.P.L., 2009. Gravitational potential stresses and stress field of passive continental margins: insights from the South-Norway shelf. Earth and Planetary Science Letters 277, 464–473. Pérez-Gussinyé, M., Reston, T.J., 2001. Rheological evolution during extension at nonvolcanic rifted margins; onset of serpentinization and development of detachments leading to continental breakup. Journal of Geophysical Research 106, 3961–3975. Planke, S., Symonds, P.A., Alvestad, E., kogseid, J., 2000. Seismic volcanostratigraphy of large volume basaltic extrusive complexes on rifted margins. Journal of Geophysical Research 105, 19335–19351. Platel, J.P., Roger, J., Peters, T., Mercolli, I., Kramers, J.D., Le Métour, J., 1992. Geological Map of Salalah (1/250,000), Sultanate of Oman; sheet NE 40-09. Ministry of Petroleum and Minerals, Directorate General of Minerals, Oman. Pollack, H.N., Hurter, S.J., Johnston, J.R., 1992. Global Heat Flow Data Set, World Data Center A for Solid Earth Geophysics, NOAA E/GCI, 325 Broadway, Boulder, CO 80303, USA. Precigout, J., Gueydan, F., Gapais, D., Garrido, C.J., Essaifi, A., 2007. Strain localisation in the subcontinental mantle — a ductile alternative to the brittle mantle. Tectonophysics 445, 318–336. Rabain, A., Cannat, M., Escartin, J., Pouliquen, G., Deplus, C., Rommevaux-Jestin, C., 2001. Focused volcanism and growth of a slow spreading segment (Mid-Atlantic Ridge, 35°N). Earth and Planetary Science Letters 185, 211–224. Ricard, Y., Husson, L., 2005. Propagation of tectonic waves. Geophysical Research Letters. http://dx.doi.org/10.1029/2005GLO23690. Robinet, J., Razin, P., Serra-Kiel, S., Gallardo Garcia, A., Leroy, S., Roger, J., Grelaud, C., 2013. The Paleogene pre-rift to syn-rift succession in the Dhofar margin (northeastern Gulf of Aden): stratigraphy and depositional environments. Tectonophysics 607, 1–16. Roger, J., Platel, J.P., Cavelier, C., Bourdillon-de-Grisac, C., 1989. Données nouvelles sur la stratigraphie et l'histoire géologique du Dhofar (Sultanat d'Oman). Bulletin de la Société Géologique de France 2, 265–277. Rolandone, F., Lucazeau, F., Leroy, S., Mareschal, J.-C., Jorand, R., Goutorbe, B., Bouquerel, H., 2013. New heat flow measurements in Oman and the thermal state of the Arabian Shield and Platform. Tectonophysics 589, 77–89. Saunders, A.D., Storey, M., Kent, R.W., Norry, M.J., 1992. Consequences of plume–lithosphere interactions. In: Storey, B.C., Alabaster, T., Pankhurst, R.J. (Eds.), Magmatism and the Causes of Continental Break-up: Geological Society Special Publication, 68, pp. 41–60.

N. Bellahsen et al. / Tectonophysics 607 (2013) 80–97 Sauter, D., Patriat, P., Rommevaux-Jestin, C., Cannat, M., Briais, A., Gallien Shipboard Scientific Party, 2001. The Southwest Indian Ridge between 49°15′E and 57°E: focused accretion and magma redistribution. Earth and Planetary Science Letters 192, 303–317. Schmeling, H., Wallner, H., 2012. Magmatic lithospheric heating and weakening during continental rifting: a simple scaling law, a 2-D thermomechanical rifting model and the East African Rift System. Geochemistry, Geophysics, Geosystems 13, Q08001. http://dx.doi.org/10.1029/2012GC004178. Sibson, R.H., 1990. Conditions for fault-valve behaviour. Geological Society, London, Special Publications 54, 15–28. http://dx.doi.org/10.1144/GSL.SP.1990.054.01.02. Sicilia, D., Montagner, J.-P., Cara, M., Stutzmann, E., Debayle, E., Lépine, J.-C., Lévêque, J.-J., Beucler, E., Sebai, A., Roult, G., Ayele, A., Sholan, J., 2008. Upper mantle structure of shear-waves velocities and stratification of anisotropy in the Afar Hotspot region. Tectonophysics 462, 164–177. Smallwood, J.R., White, R.S., 2002. Ridge–plume interaction in the North Atlantic and its influence on the formation of oceanic crust. In: Bell, B., Jolley, D. (Eds.), The North Atlantic Igneous Province: Stratigraphy, Tectonic and Magmatic Processes: Geol. Soc. Spec. Publ., 197, pp. 15–37. http://dx.doi.org/10.1144/GSL.SP.2002. 197.01.02. Sokoutis, D., Corti, G., Bonini, M., Brun, J.P., Cloetingh, S., Mauduit, T., Manetti, P., 2007. Modelling the extension of heterogeneous hot lithosphere. Tectonophysics 444, 63–79. http://dx.doi.org/10.1016/j.tecto.2007.08.012. Spohn, T., Schubert, G., 1982. Convective thinning of the lithosphere: a mechanism for the initiation of continental rifting. Journal of Geophysical Research 87, 4669–4681. Spohn, T., Schubert, G., 1983. Convective thinning of the lithosphere: a mechanism for rifting and midplate volcanism on Earth, Venus and Mars. Tectonophysics 94, 67–90. Stern, R.J., Johnson, P., 2010. Continental lithosphere of the Arabian Plate: a geologic, petrologic, and geophysical synthesis. Earth-Science Reviews 101, 29–67. Storey, M., Saunders, A.D., Tarney, J., Gibson, I.L., Norry, M.J., Tirwall, M.F., Leat, P., Thompson, R.N., Menzies, M.A., 1989. Contamination of Indian Ocean asthenosphere by the Kerguelen–Heard mantle plume. Nature 338, 574–576. Tard, F., Masse, P., Walgenwitz, F., Gruneisen, P., 1991. The volcanic passive margin in the vicinity of Aden, Yemen. Bulletin des Centres de Recherches ExplorationProduction Elf-Aquitaine 15, 1–9. Tron, V., Brun, J.P., 1991. Experiments on oblique rifting in brittle–ductile systems. Tectonophysics 188, 71–84. Van Wijk, J.W., Huismans, R.S., ter Voorde, M., Cloetingh, S.A.P.L., 2001. Melt generation at volcanic continental margins: no need for a mantle plume? Geophysical Research Letters 28, 3995–3998.

97

Vigny, C., Huchon, P., Ruegg, J.C., Khanbari, K., Asfaw, L.M., 2006. Confirmation of Arabia plate slow motion by new GPS data in Yemen. Journal Of Geophysical Research 111. http://dx.doi.org/10.1029/2004JB003229. Vogt, P.R., Avery, O.E., 1974. Detailed magnetic surveys in the Northeast Atlantic and Labrador Sea. Journal of Geophysical Research 79, 363–389. Watchorn, F., Nichols, G.J., Bosence, D.W.J., 1998. Rift-related sedimentation and stratigraphy, southern Yemen (Gulf of Aden). In: Purser, B., Bosence, D. (Eds.), Sedimentation and Tectonics in the Rift Basins Red Sea–Gulf of Aden. Chapman & Hall, London, pp. 165–189. Watremez, L., Leroy, S., Rouzo, S., d'Acremont, E., Unternehr, P., Ebinger, C., Lucazeau, F., Al-Lazki, A., 2011. The crustal structure of the north-eastern Gulf of Aden continental margin: insights from wide-angle seismic data. Geophysical Journal International 184, 575–594. Watremez, L., Burov, E., d'Acremont, E., Leroy, S., Huet, B., Le Pourhiet, L., Bellahsen, N., 2013. Buoyancy and localizing properties of continental mantle lithosphere: Insights from thermomechanical models of the eastern Gulf of Aden. Gcubed. http://dx.doi.org/10.1002/ggge.20179 (in press). Watts, A.B., Burov, E., 2003. Lithospheric strength and its relationship to the elastic and seismogenic layer thickness. Earth and Planetary Science Letters 213, 113–131. White, R.S., 1992. Magmatism during and after continental break-up. In: Storey, B.C., Alabaster, T., Pankhurst, R.J. (Eds.), Magmatism and the Causes of Continental Break-up: Geological Society Special Publication, 68, pp. 1–16. White, R.S., 1997. Rift–plume interaction in the North Atlantic. Philosophical Transactions of the Royal Society of London, Series A: Mathematical, Physical and Engineering Sciences 355, 319–339. White, R.S., McKensie, D., 1989. Magmatism at rift zones: the generation of volcanic continental margins and flood basalts. Journal of Geophysical Research 94, 7685–7729. White, R.S., Bown, J.W., Smallwood, J.R., 1995. The temperature of the Iceland plume and origin of outward-propagating V-shaped ridges. Journal of the Geological Society of London 152, 1039–1045. Withjack, M., 1979. A convective heat transfer model for lithospheric thinning and crustal uplift. Journal of Geophysical Research 84, 3008–3022. Withjack, M.O., Jamison, W.R., 1986. Deformation produced by oblique rifting. Tectonophysics 126 (2–4), 99–124. Zoback, M.L., 1992. First- and second-order patterns of stress in the lithosphere: the World Stress Map Project. Journal of Geophysical Research 97 (B8), 11703–11728.