TECTONOPHYSICS ELSEVIER Tectonophysics 278 (1997) 329-352 Styles of continental rifting: crust-mantle detachment and mantle plumes a* H e r m a n n Z e y e n , , F r a n k Volker b, Veronika Wehrle a, Karl Fuchs a, Stephan V. Sobolev a,l Rainer Altherr c a Geophysical Institute, University ofKarlsruhe, Hertzstrafie 16, D-76187 Karlsruhe, Germany b Institutfiir Geowissenschaften und Lithosphiirenforschung, Justus-Liebig-Universitdt Giessen, Senckenbergstrafle 3, D-35390 Giessen, Germany c Mineralogisch.-Petrographisches Institut, Universitiit Heidelberg, Im Neuenheimer Feld 236, D-69120 Heidelberg, Germany Accepted 1 May 1997 Abstract Observations made in different continental rift systems (European, Red Sea-Gulf of Aden, and East African Rift Systems) were investigated in terms of the influence of different parameters on the style of tiffing. Apart from the lithospheric thermal regime at the time of rift initiation, the process of rifting seems to be mainly controlled by the far-field stress regime and the presence or absence of a mantle plume. In a hot lithosphere the low viscosity of the lower crust enables the upper crust to be detached from the mantle and be deformed independently under far-field stresses. Therefore, in western Europe the main rifts could open obliquely to the direction of mantle movement in crustal levels without appreciable extension in the lithospheric mantle. In contrast, the colder lithosphere of Arabia did not allow detachment of crust and mantle. Therefore, despite being in a similar tectonic situation as in western Europe, i.e. rifting in front of an orogen, the whole lithosphere deformed congruently. Rift opening occurred parallel to mantle movement, i.e. parallel to the direction of extensional stress in the lithospheric mantle induced by the pull of the subducting slab at the orogenic front. The forces needed to extend the whole relatively cool Arabian lithosphere could, however, not be produced by slab pull alone. Additional forces and weakening of the lithosphere were produced by the Afar mantle plume. Mantle plumes are generally not able to break very thick cratonic lithosphere but they deflect sidewards when hitting this kind of lithosphere. Wanner (but still relatively cool) lithosphere like in the surroundings of the East African Tanzania craton or in Arabia can, by the buoyancy of a plume, be bent strongly enough to break. As a consequence, long linear rift structures develop with generally high shoulders. The presence of a plume explains thus the position of the East African and Red Sea-Gulf of Aden rifts. Under far-field compression, rifts will open only a small amount, whereas under far-field extension continental break-up may occur. A plume hitting a hot lithosphere may penetrate it without producing long linear rifts. Instead, crustal deformation will be distributed in parallel basins over a wide area with only minor amounts of rift shoulder uplift as has happened in northern Kenya and the French Massif Central. Keywords: plate tectonics; yield strength; stress; lithosphere; lower crust; Great Rift Valley; European Cenozoic Rift System * Corresponding author. Present address: Institutionenf6r Geofysik, Uppsala Universitet,Villav/igen16, S-752 36 Uppsala, Sweden. Tel: +46 (18) 4713781; fax: +48 (18) 501110; e-mail: hz@geofys.uu.se l Present address: GeoForschungsZentrumPotsdam, Telegrafenberg,Potsdam, Germany. 0040-1951/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. Pll S 0 0 4 0 - 1 9 5 1 ( 9 7 ) 0 0 1 1 1 - X 330 H. Zeyen et al./ Tectonophysics 278 (1997) 329-352 1. Introduction Continental crust, in contrast to oceanic crust, is easily deformed (Steckler and ten Brink, 1986). The most spectacular deformations take place in orogenic belts and continental rifts, giving deeply eroded terranes, thick sedimentary basins or new oceans. Generally, the processes which build these structures are due to or lead to a reorganisation of lithospheric plates and plate movements. The last major plate reorganisation took place at the end of the Cretaceous and the beginning of the Tertiary and was accompanied by widespread orogeny and rifting (Ziegler, 1992a). Large rift systems evolved in the European and African realms during the Tertiary, namely the continental European Cenozoic Rift System (ECRIS; Ziegler, 1992b), the Red Sea-Gulf of Aden Rift System (RSGA), and the East African Rift System (EARS) (Fig. 1). These rifts evolved in tectonically very different regimes, controlled by young or old, hot or cold lithosphere, far-field compression or extension and the existence of plumes of different sizes. The youth of these rifts, which are still evolving, their different tectonic background and surface expressions make them an extremely valuable target for comparative studies of the reaction of different kinds of lithosphere to different kinds of forces. During the last 15 years, parts of these rifts have been investigated in detail by a joint effort of geoscientists of the University of Karlsruhe in co-operation with many other geoscientific institutes around the world. Geophysical, petrological and geological studies shed new light on the history and actual state of the lithosphere in these areas. The aim of this paper is to summarize the findings under the focus of influences of the thermal regime (lithospheric thickness) and forces acting on the rift zones. We will start our presentation in the East African Rift System (EARS) which represents extension of cold to cool lithosphere under far-field compression but under the influence of a major mantle plume. Strong, cold and thick cratonic lithosphere in the area (Tanzania craton) forms an essentially undeformed core. Rifting is concentrated in its weaker vicinity, resulting in only minor opening (10-15 km; Baker and Wohlenberg, 1971) which, however, involves the whole lithosphere. We will continue fur- ther north and show how, in the area of the Red Sea-Gulf of Aden System (RSGA), the combination of far-field extension and a major mantle plume gave rise to strong rifting of a cool lithosphere, affecting the whole lithosphere (crustal separation of up to more than 400 km; Joffe and Garfunkel, 1987). We will finally turn to the European Cenozoic Rift System (ECRIS), where two further types of tectonic settings are encountered. Rifting in a warm and weak lithosphere under far-field extension and without plume influence led to minor rifting in the crust (10-15 km extension) without affecting the mantle, since crust and mantle could decouple due to the weakness of the lower crust. In the French Massif Central, a small plume led to the formation of several subparallel rifts above the plume joining into a single rift outside of the area of plume influence. The deeper lithosphere, in contrast to the Afro-Arabian Rift System (EARS and RSGA), is deformed in an essentially isometric area corresponding to the plume, without any correspondence to the linear features in the crust. The observed effects of different combinations of far-field stresses and lithospheric thermal state are summarized in Table 1 which will be referred to during the detailed description of each of the investigated rifts. 2. The East African Rift System 2.1. D a t a The EARS extends in two branches in a general N-S direction from 20°S to 5°N (Fig. 2). Its western branch, which is the youngest rift presented in this study, started to develop with the onset of volcanism near its northern end (Virunga Basin at 2°S) around 12-13 Ma ago (Ebinger, 1989). Basin formation was generally preceded by volcanism which in turn followed regional updoming (Ebinger, 1989). Volcanism and basin formation proceeded towards the south, bordering the Archaean Tanzania craton and cross-cutting along its southern end (Malawi rift basins), the Late Proterozoic/early Palaeozoic PanAfrican belt and the Permo-Jurassic Karoo Province. The shallow structure is dominated by up to 4 km deep and 60-70 km wide sedimentary basins, bordered by steep, essentially normal faults which, as indicated by seismicity, seem to penetrate to a H. Zeyen et al. / Tectonophysics 278 (1997) 329-352 40 331 ° _ 20 ° v¢= km 0 500 1000 Heat flow < 60 0o M .0 ° Heat flow: 60-90 1~ Heatflow > 90 :;i!;!!ilil iiiiiii¸, ~--~ No plume influence Transition Plumeinfluence f-'~ .~0~ -20 °~ , ° Fig. 1. L o c a t i o n o f the investigated rifts. Ri-G = R h i n e graben; Ro-G = R h 6 n e graben; L-G = L i m a g n e g r a b e n ; MC = M a s s i f Central; DS-J = D e a d S e a - J o r d a n T r a n s f o r m Fault; RS = R e d Sea; GA = G u l f o f A d e n ; ER = E t h i o p i a n rift; TR = T u r k a n a rift; GR = G r e g o r y rift; VT = Valencia T r o u g h ; WR = W e s t e r n rift. depth of at least 30 kin. Little is known about the deeper crustal structure and possible crustal thinning. Balanced cross-sections and shallow reflection seismic profiles indicate an extension of 10-15% which could be somewhat underestimated if inter- nal block deformation is considered (Morley, 1988; Ebinger, 1989). Also information about lithospheric thickness and thermal regime is sparse. From longwavelength gravimetric data Girdler (1983) claims a lithosphere-asthenosphere boundary at about 50 km 332 H. Zeyen et al./ Tectonophysics 278 (1997) 329-352 Table 1 Observed effects of different combinations of far-field stresses, plume presence and lithospheric thermal state on the style of rifting Surf. heat flow; Lith. thickness; Lith. strength Far field extension Far field compression plume no plume plume no plume (3) Crust-mantle decoupling; diffuse deformation; difficult (no) break-up (Massif Central). (4) Crust-mantle decoupling; linear, crustal rifting; no break-up; little extension (Rhine graben). Hot >70 m W/m2; _<100 km; _<40 TN/m (1) Diffuse deformation; (2) Not applicable. difficult (no) break-up (Turkana). Intermediate 50-70 mW/m2; 100-150 km; 40-80 TN/m (5) Whole lithosphere breaks; little extension (Kenya). (6) Not applicable. Cold <50 mW/m2; > 150 km; >80 T N / m (9) Bending but little inelastic deformation (Tanzaniacraton). (10) Not applicable. (1 l) Cracks may open and allow magma ascent; no rifting (Nubian shield). (8) Once initiated somewhere (7) Whole lithosphere breaks; continental break-up (Red Sea). (plume), a crack propagates through the whole lithosphere (break-up: Red Sea or transform fault: Dead Sea). It should be noted that extensional deformation under far-field compression in absence of a plume is also possible if high topography is present as in active orogenic areas but is not the subject of this study. The different fields are numbered for referring to them in the text. depth compared to about 80 km at a short distance further west and 60 km beneath the East African Plateau (Tanzania craton). Braile et al. (1996) model a lithospheric thinning from 90 km beneath the East African Plateau to 65 km beneath the western rift. Lithospheric thinning is also indicated by reduction of P-wave velocities in the upper mantle (Fairhead and Girdler, 1971). The stress regime during rifting is described as dominantly E - W (Ebinger, 1989) to N W - S E extensional (Rosendahl et al., 1992). The eastern branch of the EARS consists of the Turkana rift in the north and the Gregory rift in the south. This rift branch started to develop near its northern end with extensive flood basalt eruptions some 33 Ma ago, preceding half-graben formation by a few million years (Morley et al., 1992). These flood basalts are estimated to amount to approximately 45,000 k m 3, i.e. some 20% of the total volcanism erupted along the eastern branch (Williams, 1972; Baker, 1986, 1987; Morley et al., 1992). Apart from the initial N - S trend a second, more recent, W - E trend of volcanism is indicated by the evolution of eruption centres east of the rift, beyond its shoulders. The most important of these volcanic areas are (from north to south) the Huri Hills, Marsabit, Nyambeni Range Mt. Kenya, and the Chyulu Hills (Henjes- Kunst and Altherr, 1992; Class et al., 1994). It is worth noting that, both in the eastern and western branch, regional updoming and volcanism generally preceded basin formation and the rift evolved from north to south (Baker, 1987; Ebinger, 1989). The overall extension and topographic expression of the rift varies significantly from north to south. In northern Kenya the area of deformation is very wide (some 150-250 km in E - W direction; Morley et al., 1992; Smith and Mosley, 1993; Hendrie et al., 1994). A series of subparallel basins evolved with a total extension of about 40 km since Miocene times (Morley et al., 1992). A /3-factor of up to 2 has been reported by Ebinger and Ibrahim (1994), probably accumulated during three distinct rift phases since Late Jurassic times (Jurassic-Cretaceous: Anza rift and Blue and White Nile rifts in southern Sudan; Palaeogene: Anza rift and western and central Turkana rift; Miocene-Recent: Ethiopian rift, central and eastern Turkana rift, Kenya rift; for more details see Ebinger and Ibrahim, 1994). The topographic expression, however, is relatively poor, the basin flanks overlying the background topography by only a few hundred metres. Towards the south the rift narrows to 6 0 - 7 0 k m and the amount of extension decreases to 10-15 km (Morley et al., 1992). However, the rift H. Zeyen et aL / Tectonophysics 278 (1997) 329-352 b g co c~ 333 oa o 20" 20" JurassicCretaceous rift Craton TertiaryRecent rift 10" 10" m. km 0 500 7~ 0o _10 ° .20 ° 0o -10" ; -20 ° Fig. 2. The East African Rift System and important rift-controlling tectonic units: the Tanzania and Nyasa cratons as hard, undeformed cores and the Jurassic-Cretaceous Anza rift as weak zone. shoulders rise up to 2500 m above the rift bottom and the surrounding undisturbed areas. Since 1985 several refraction and teleseismic experiments as well as geological and petrological studies have been carried out in the Kenya rift and its surroundings, yielding a fairly detailed image of the crust and upper mantle (KRISP Working Group, 1987; Prodehl et al., 1994b; Ritter et al., 1995; No- vak et al., 1997). Crustal thickness varies strongly along the rift, from 34 km in the area of the Kenya Dome south of the Equator to only 20 km north of Lokori (2°N) and beneath Lake Turkana (Mechie et al., 1994). Nearly 10 km of the thick crust in the south are made up of a layer with velocities above 6.8 km/s which, taking into account the expected high temperatures, has been interpreted as due to 334 H. Zeyen et al./ Tectonophysics 278 (1997) 329-352 basaltic intrusions (Mechie et al., 1994), or underplating (Mooney and Christensen, 1994). In the area of the Kenya Dome the crust is thickened to about 40 km beneath the shoulders, reducing again to about 36 km towards the east and west (Braile et al., 1994; Maguire et al., 1994). Seismic velocities in the upper mantle are strongly reduced beneath the whole length of the Kenya rift to great depths. Refraction seismic experiments yielded velocities of 7.5-7.6 km/s just beneath the crustmantle boundary (Moho discontinuity) underneath the whole rift (Mechie et al., 1994). In contrast, velocities are in the range of 8.0-8.2 km/s beneath the shoulders and the outer areas (Prodehl et al., 1994a; Braile et al., 1994; Byrne et al., 1997; Novak et al., 1997). Teleseismic tomography shows a strong velocity reduction of 6-12% in a nearly vertical channel which is well correlated to the position of the rift down to a depth of at least 160 k m (Achauer et al., 1994; Slack et al., 1994). This velocity reduction can be explained by the joint effects of small amounts of melt and an increase of potential temperature by about 300 K beneath the rift (Sobolev et al., 1996), indicating a reduction of thermal lithospheric thickness to about 60 km. In contrast to lithospheric models determined gravimetrically, the teleseismic model indicates a lithospheric thickness of at least 130 km to the east of the Gregory rift and 160 km to the west. Few stress data exist for East Africa. Zoback (1992) explained the observed N40-50°E orientation of SH in the EARS as superposition of E - W far-field compression with local deviatoric E - W extension in the rift area. Palaeo-stress indicators reveal a rotation of SH from N - S to the actual N E - S W direction during the Holocene (Bosworth et al., 1992; Haug and Strecker, 1995). Comparing the western and eastern branches of the EARS, striking similarities are observed between the entire western branch and the southern part of the eastern branch (Gregory rift). Both rifts are very narrow with small stretching factors and high shoulders, both are situated near the border between a cratonic block (Tanzania craton) and the overthrusted Late Proterozoic/early Palaeozoic Pan-African belt but die out where a subbasin enters the craton itself (e.g. the Nyanza rift in Kenya). Differences reside mainly in the timing and the amount of volcanism: in the southern part of the eastern rift volcanism started a few Ma earlier and was much more productive than in the western rift. In contrast, the northern part of the eastern rift (Turkana rift) has a very different appearance. The rift splits apart towards the north, forming a wide zone of diffuse extension without appreciable shoulders and, compared to the southern part of the eastern branch, the crust is much more thinned and consequently also the overall topography is much lower. The volcanism of the EARS started here with a vast amount of magma, erupted within a relatively short time span. 2.2. Model of rift formation These and other observations can be best explained by a model in which a large mantle plume is the principal source for the forces which lead to rift formation. Many other models (starting with Vening Meinesz, 1960, until recently Hendrie et al., 1994) are able to explain structural features in the crust and upper mantle, but do not take into account the source of the extensional forces in an overall compressional regime and the regional timing of rifting and volcanism. Evidence for the existence of a plume beneath the East African Plateau, which comprises the region of roughly 1000 x 1000 km 2 between the two rift branches (Ebinger, 1989), is manifold. The average topographic elevation of the plateau of about 1000 m above most of its surroundings is, as is to be expected for this long wavelength, in isostatic equilibrium (Baker and Wohlenberg, 197 l). An important mass deficit in the mantle of at least 15 kg/m 3 (equivalent to a temperature increase of about 100-150 K) over a depth range of 200 km is needed to support this topographic load in the case of static equilibrium. Petrological evidence for hot uprising asthenospheric material as source for the volcanism in the area of the eastern branch has been reported by several authors (e.g. Macdonald, 1994; Class et al., 1994; Hay et al., 1995). Smith (1994) argues for the rise of an asthenospheric plume as cause for volcanism and rifting of the eastern branch. Also the initiation of volcanism in the northern part of the eastern branch in the form of flood basalts is an indicator for the rise of a mantle plume (Campbell and Griffiths, 1990). Why then, if the whole East African Plateau is uplifted, did the rifting start at 335 H. Zeyen et al./Tectonophysics 278 (1997) 329-352 200 - I 2001 .cr,.,st { L i t h o s p h e r i c mantle . "~" -200 - 2 0 0 ~ v -400 - -400 4 Forces -600 q d3 -600 Movements -8oo4 -800 [Before C. 30 Ma] -1000-1000 i0001 i 1 -500 0 500 1000 -1000 -500 0 500 1000 -500 0 Distance (km) 500 1000 -1000 -500 0 Distance (krn) 500 1000 I 20 E -2o -40 a -60 ,8O -100 -1000 Fig. 3. Sketch of plume and rift evolution in East Africa. (A) Plume rises off-centre beneath the Tanzania craton. Stresses in the lithosphere are mainly compressional. (B) When the plume approaches the lithosphere, the eastern part of the Tanzania craton and surrounding areas start to be uplifted. Compressional far-field stresses are superposed by extensional stresses due to bending and topography. (C) The plume, extending laterally beneath the craton, reaches the thinner lithosphere of the Pan-African realm. Extensional stresses increase and volcanism and subsequently rifting starts. (D) The plume extends further to both sides. It has reached the western edge of the craton, forcing extension in the area of the western rift. East of the craton the rift matures and new volcanic centres evolve east of the rift. its northeastern edge and propagate southward along the eastern side until it started also at the western side of the plateau? Fig. 3 summarises our concept of the evolution of the mantle plume and its effects on the lithosphere. This diagram is based on the assumption of a lithosphere which is stationary above the plume (e.g. Ebinger et al., 1989). With slight modifications this model may, however, also be applied to a scenario in which Africa is moving northeastwards with respect to the plume (Bonavia et al., 1995). In the Early Oligocene a rising plume (Fig. 3A) hit the cratonic lithosphere of about 200 km thick some distance east of the centre of the craton (or arrived from the northeast) (Fig. 3B). Due to the asymmetrically acting buoyancy force the cratonic block tilted on a large scale, being uplifted mainly on its eastern side. At this time the whole area was still submitted to compressional far-field forces (Fig. 4) which arose from ridge push around most of the African continent and high topography of the area overlying the Afar plume (see below). The uplift of the East African Plateau caused regionally strong bending forces which were able to invert the compressional stress regime into extension but were not strong enough to break the craton (Table 1, field no. 9). The plume then spread beneath the cratonic lithosphere (Griffiths and Campbell, 1990), but it was not able to penetrate the lithosphere, partly due to the low temperatures and therefore high strength of the craton already at relatively great depth, partly due to the increased buoyancy of H. Zeyen et al./Tectonophysics 278 (1997) 329-352 336 4~ --~I 20 ° C;n , ~ 1 20° ~6 "_ 10"4 -- -__ . 10" km 0 500 o - g • +4- , -20 . . o . . . . . + . . ,¢ . . . o . . ) . . -20 ° ,¢ Fig. 4. Forces acting in East Africa at the onset of rifting. White arrows represent compressiona] forces due to elevated topography of the Ethiopian Plateau above the Afar plume towards the northeast and due to ridge push from the Indian Ocean, and grey arrows the e×tensJona] forces due to lithosphere bending and increased topography above a plume. The thick numbered line delimits areas above ]000 m topographica] elevation. the cratonic root (e.g. Forte et al., 1995). Fig. 5 shows the surface stress distribution in elastic plates of different thickness, resulting from a distributed force from below which is able to lift a 1000 km wide region by 1000 m (Turcotte and Schubert, 1982). The 90-km-thick plate represents a craton (as indicated by examples presented by McNutt et al., 1988 or Ebinger et al., 1989), whereas the 40-km-thick plate represents the younger and weaker surrounding area affected by the Pan-African orogeny. These calculations suggest that, as the plume head approached the edge of the craton, the extensional forces con- H. Zeyen et al,/ Tectonophysics 278 (1997) 329-352 40 ~J L. . . . . . . . J Y . m~ 20 5 I0 jl . . L EET=9O kr~) . i z ' "= EEl"=40 k m I [ F I 0 • - @ -io I I/ - ! '~ -20 H J -30 ! d -40 500 Distance from plume 1000 centre 1500 (km} Fig. 5. Extensional stresses produced by the bending of an elastic plate due to buoyancy forces of an underlying plume. Dashed line indicates that the plume is located underneath a craton with an equivalent elastic thickness (EET) of 90 km, and continuous line, the plume bends thinner lithosphere (EET = 40 km). centrated in a narrower zone and became larger by a factor of about 1.5, eventually causing rupture of the weaker Pan-African belt (Fig. 3C; Table 1, field no. 5). Spreading further, the plume head reached the western end of the craton causing there failure for the same reason, whereas towards the east new volcanic centres formed east of the rift shoulders (Fig. 3D). Further support for local driving forces as opposed to far-field extensional forces for the opening of the northern Kenya rift comes from the observation of inversion structures in the Anza rift and its surroundings east of the Turkana rift. Bosworth and Morley (1994) report basin inversion in the northwestern Anza rift during the Late Eocene and Oligocene which they relate mainly to a thermal event, probably the arrival of the plume in our model. However, Tertiary basin inversion in the Anza graben observed further to the southeast has been interpreted as compressional activity related to the formation of the EARS (Reeves et al., 1987; Bosworth, 1992). Partly, this activity may also be connected to the SSE-directed movement of Somalia with respect to Africa (Jestin et al., 1994). This intraplate accommodation of the EARS extension would not be expected if the rift were formed by far-field extensional forces. It is a strong indicator for local stress sources arising from buoyancy forces exerted by the low-density 337 plume material. It must be emphasised that this scenario implies brittle fracture of the crust and at least the upper part of the lithospheric mantle. The model of velocity anomalies derived from teleseismic delay times (e.g. Achauer et al., 1994) with its steep and narrow low-velocity channel below the Kenya rift seems to indicate this mechanism as the most probable for the initiation of the Gregory rift. However, crack opening (due to brittle fracture) alone cannot explain the observed 50-km-wide low-velocity zone in the lithospheric mantle since the total amount of crustal extension is estimated to be not more than 15 km. An explanation of this discrepancy might be that the crack served as channel for the ascent of hot plume material to shallow depths where it triggered more widespread deformation and melting of fertile, during the Pan-African orogeny subducted material in this area (Smith, 1994). Another possibility for the formation of the channel might be that the crack allowed the metastable, dense lithospheric mantle to become unstable and sink into the buoyant plume material. This model is able to explain the position and evolution of the EARS. But it does not explain the obvious differences between the Turkana rift and the Gregory rift. In order to understand the evolution of the Turkana rift we have to take into account its geological history which consists of three distinct rifting events (Ebinger and Ibrahim, 1994). In Late Jurassic/Cretaceous times, related to the opening of the southern Atlantic and the separation of Madagascar from Africa (Hankel, 1994), a large rift system evolved in central and northern Africa (Lambiase, 1989; Daly et al., 1989; Bosworth, 1994). Part of this system was the Anza graben (Fig. 2) which runs from the coast into southern Kenya in a NW direction until the Turkana area where its evidence is lost beneath Miocene rift sediments (Morley et al., 1992). This rifting event produced only little, if any, volcanism but up to 8-km-deep sedimentary basins in the Anza graben (Dindi, 1994; Bosworth and Morley, 1994), which points to strong crustal thinning. Both Tertiary rifting phases in the Turkana area started with flood basalt eruptions, not in the Anza graben but near its borders. At first glance, this seems surprising since heating during rifting weakens strongly the lithosphere beneath the graben. Fig. 6, however, reveals that the borders of a rift may 338 E H. Zeyen et al./ Tectonophysics 278 (1997) 329-352 45 ! Z 40 f f dfting 35 ~ ' ~ - - ~ J ; , , ~ / ' - ~ ' ~/¢/~p "~ "O CO ~ 5Ma " 10Ma / / / J / 15 Ma 3o - j - - h 20M, 25 Ma "E 25 ~ 0 T 100 200 300 Distance from rift centre (km) Fig. 6. Post-rift strength evolution in a rift basin and its shoulders. The model consists initially of a 40-km-thick crust and a 150-km-thick lithosphere which is stretched in a 100-km-wide area with an extension velocity of 0.5 mm/year and a rifting period which lasted during the entire Cretaceous. The strength is calculated by integrating vertically yield strength curves (for more information see Negredo et al., 1995) and gives an estimate of the force needed to deform the entire lithospheric column in extension. Only half of the model is represented, the rift axis at 0 km being treated as symmetry axis. become the weakest parts within 2 0 - 2 5 Ma after the end of rifting due to lateral heat conduction and relaxation (see also Negredo et al., 1995). Therefore, the onset of new rifting and volcanism after a period of ca. 30 Ma, as expected for the Turkana rift, should be localised not inside the abandoned rift but in its vicinity (Table 1, field no. 1). In addition, the formation of a wide rift and migration of deformation as observed in northern Kenya (Morley et al., 1992) is expected for a preheated and weak lithosphere (Negredo et al., 1995). 3. The Red S e a - G u l f of Aden rift system 3.1. D a t a The RSGA consists of the Gulf of Suez, the Red Sea, the Gulf of Aden, the Afar area, and the Ethiopian rift and is connected to the Dead S e a Jordan transfault zone (Fig. 7). It started to develop a few Ma earlier than the EARS in the E o c e n e Oligocene. First volcanism is reported from Ethiopia at about 45 Ma and is generally agreed to be connected to the ascent of a large mantle plume, the Afar plume (e.g. WoldeGabriel et al., 1991; Ebinger et al., 1993). An enormous amount of volcanic material has been erupted in a short time, especially in the area of the Ethiopian Plateau (Davidson and Rex, 1980; Mohr and Zanettin, 1988, and references therein). Since approximately 30 Ma the centre of volcanism moved further to the northeast into the Afar area (Vellutini, 1990). Topographically the R S G A is characterised by a high plateau at 2 - 3 0 0 0 m elevation (Ethiopia) and depressions which evolved into new ocean basins (Red Sea, Gulf of Aden) with high shoulders. The Afar depression has an intermediate position with elevations between slightly below sea level and 1500 m. Geophysical measurements indicate a thick crust (40 km) with high velocities in the lower crust underneath Ethiopia (Makris and Ginzburg, 1987). This may be explained similarly to the relatively thick crust of the Kenya Dome by underplating or intrusions in relation to the strong volcanism of the area. Comparison with areas of similar crustal thickness like the Arabian Platform (Mechie et al., 1986) shows that the high topographic position cannot be explained by crustal thickening but a strong mass deficit in the mantle must be invoked, related to the plume and lithospheric heating. The elevation anomaly caused by the plume can be traced on a large-scale topographic map and crosses the Red Sea north of the Farasan Islands between 18° and 20°N into Arabia. Its influence can be observed petrologically in the composition of the volcanic rocks which show clear plume traces still in the area of the Farasan Islands but no more north of 20°N where sea-floor spreading is connected to N-MORB basaltic volcanism (Volker et al., 1993, 1997). The crust thins rapidly to 20 km in the Afar area where it has seismic characteristics which lie in between typical oceanic and continental crust (Makris and Ginzburg, 1987). Further north, it thins to only 10-15 km in the central Red Sea (Makris et al., 1991), indicating here clearly newly formed oceanic crust. Magnetic stripes confirm the existence of oceanic crust since about 5 Ma in the southern Red Sea, extending now between 16° and 22°N (Makris and Rihm, 1991). In the Gulf of Aden little is known about the deep crustal structure but magnetic lineaments indicate that sea-floor spreading started approximately 10 Ma ago (Joffe and Garfunkel, 1987) and evolved towards the Afar area. Although the Afar forms the link between the two basins with true oceanic crust, H. Zeyen et al./ Tectonophysics 278 (1997) 329-352 339 40 ° 3 o ° -I- . . . . . . 2o--I , ..... I --.'~ . ~ ~ t 2 o° Afa~Plume ~ o ° ~ - . ~ 3. . . . . . . . . . 12 " _ " . ......... 0 A A - 1 } ~ i} ' initial plate boundary Io ° OO orogenic front ===~ far field forces local forces -[ 0 o ___~_ T----------L-- ~iiiiiiiiiiiii~iiiiiiiiiiii!i!i~i~iiii}iiiiiii R *" q -'I 0 ° b Lo Fig. 7. The Red Sea-Gulf of Aden Rift System. Forces acting at the beginning of rifting. White arrows represent the far-field forces (compression due to ridge push from the Indian Ocean Ridge, extension due to slab pull underneath the Zagros Mountains which are represented by the thrust fault line), and the grey arrows extension due to bending and increased topography above the Afar plume. Dashed line indicates the initial position of the evolving rift. sea-floor spreading is here still in a beginning phase. Vellutini (1990) and Clin (1991) have studied the western continuation of the Gulf of Aden, the Gulf of Tadjura, and argue that faulting and sea-floor spreading are changing from the dominantly N - S direction in the Gulf of Aden to a N E - S W direction in the Afar area. It seems therefore, that a connection is forming between the oceanic crusts of the Gulf of Aden and the Red Sea. From a geometrical point of view, the opening of 340 H. Zeyen et al./ Tectonophysics 278 (1997) 329-352 the Red Sea-Gulf of Aden system increases from the Gulf of Suez in the north towards Afar and the eastern end of the Gulf of Aden. This implies the assumption that the whole Afar depression belongs to the rift floor which is indicated by the abnormal crustal thickness and velocity distribution and by the alignment of the main border faults. In Afar the width of the rift reaches then 350 km, in the Gulf of Aden even 400 km. The Ethiopian rift, in contrast, is much narrower (ca. 50-70 km) and in this respect similar to the Gregory rift. Fig. 8 shows three crustal sections across the Red Sea at different latitudes which will serve to study the form of crustal extension and to explain the apparent paradox that in the area of widest rift opening the crust has remained thickest. The profiles are taken from Makris et al. (1991), slightly simplified in order to distinguish only post-Oligocene sediments or volcanics, upper crust, lower crust and mantle. In the northernmost Red Sea (Fig. 8a), where no plume influence is recognised nor has sea-floor spreading started, the upper crust has thinned by a factor of 2.5 whereas the lower crust has vanished in the centre of the rift. This different thinning of upper and lower crust has also been reported from other rifts (e.g. Valencia Trough in the western Mediterranean; Torn6 et al., 1992) and is explained by mainly brittle extension in the upper crust and faster, ductile deformation in the weaker lower crust (Moretti and Pinet, 1987), or thermo-mechanical/chemical destabilization of the Moho (Kusznir and Ziegler, 1992). If the total amount of crustal stretching (/3 = 4) is distributed over the width of the basin one comes up with an initial width of 30-40 km, similar to that of the African rifts. Further south (Fig. 8b), in the transitional area of plume influence, where also sea-floor spreading has occurred, the upper crustal extension is stronger, but lower crust still remains beneath the flanks of the central trough. If one takes away the 75 km of extension due to sea-floor spreading and calculates the initial width of the rift based on whole crustal thinning (fl = 3.5), a somewhat wider initial rift of about 80 km results. However, it is reasonable to suppose that also here lower crust should have vanished. If, therefore, one uses only the remaining upper crustal thickness for the calculations one obtains a fl-factor of 7 - 8 and again an initial rift width of 30-40 km. The third profile (Fig. 8c) crosses the plume area. Here, the lower crust is again thicker whereas the upper crust is even more thinned. Calculating the initial rift width based on whole crustal thinning (fl = 2.5) one obtains approximately 150 km, whereas ignoring the lower crust (/3 = 9) results again in 30-40 km for the initial rift width. The observed correlation between upper crustal thinning and basin width on one side and lower crustal thickness and plume influence on the other side gives reasonable evidence for two assumptions. (a) The original lower crust must have vanished completely in the centre of the rift due to ductile extension. The actually existing lower crust, in contrast, must be related to underplating. Since the crust beneath the shoulders on both sides of the rift is thickened (Mechie et al., 1986; Makris and Ginzburg, 1987; Voggenreiter et al., 1988) it can be supposed that at least part of this thickening is due to lateral accommodation of lower crustal material which has moved out from beneath the rift (Moretti and Pinet, 1987; Zeyen et al., 1996). (b) Although the Red Sea and Gulf of Aden are now rather wide basins, their initial state was similar to the one of the Western and Gregory rifts and the Ethiopian rift, but different to the Turkana rift. 3.2. Model of rift formation The question then arises why the rifts have had such a different evolution. The answer may be found in the differences in the stress field. At the time of beginning extension in the area of the RSGA, the Tethys ocean was just closing and continentcontinent collision was starting towards the northeast in what was to become the Zagros Mountains (Courtillot et al., 1987). Due to the relatively dense downgoing slab associated with the northeastwardsubducting oceanic lithosphere, the uppermost mantle beneath NE Africa was subjected to extensional forces in the direction of movement (slab pull). The thermal regime beneath the Arabian micro-plate was probably similar to the present-day one (i.e. a relatively low surface heat flow of ca. 50 mW/m2; Eckstein, 1978) resulting in a relatively highly viscous lower crust (see discussion in Section 5). Although the crust was under compression near the collision zone, further away the coupling between crust and H. Zeyen et a l . / Tectonophysics 278 (1997) 329-352 NW . 30" 35 40" 45" 50" 30" 25' 25' 20' 20 t5' 15" Profile 1 341 SE a) 10' 5' 30' 40' 35' 45" Profile 2 sw NE b) sw Profile 3 NE c) (D O4 150 km ~ Mantle Replaced crust Lower crust Oceanic crust Upper crust Fig. 8. Crustal profiles across the Red Sea simplified from M a k r i s et al. (1991~. L -- actual width of the rift measured along the profile; A A -- surface of reduction of crustal material ~- sum of 'new' mantle material and Cenozoic sediments); fluc -- stretching factor of the upper crust: f i t c = stretching factor calculated fron~ whole crustal thickness. (al Profile I inorthern rift, no influence of Afar plume). (b) Profile II ~central rift. transition zone). ~c) Profile III ~Afar area. strong plume influence). Note that the upper crustal stretching factor increases nearly proportionally to the width of the rift, whereas the total crustal stretching factor decreases. 342 H. Zeyen et al./Tectonophysics 278 (1997) 329-352 mantle through the highly viscous lower crust provoked also extension at crustal levels. Under these conditions, favourable for rifting, the rising plume added further extensional forces in the Afar area, bending the lithosphere upwards (Table 1, field no. 7). This triggered the formation of a triple junction with strong extension in the direction of the extensional mantle far-field forces (Red Sea, Gulf of Aden) but little extension perpendicular to it (Ethiopian rift). Due to the coupling between crust and mantle and the additional push from beneath due to buoyancy forces related to the low-density plume material, crust and lithosphere deformed simultaneously, breaking apart in an initially very narrow zone. Indicators for the triggering function of the plume are the timing of the onset of volcanism prior to widespread faulting, and the position of the triple junction. It is situated close to the centre of the uplifted area, and, therefore, most probably in the centre of the plume. The extensional far-field forces, however, have been responsible for the evolution of the Red Sea and the Gulf of Aden towards the break-up and formation of new oceanic crust (Table 1, field no. 8). At the time when the EARS started to develop further to the south, the slab which was responsible for the extensional forces in the Red Sea area was already separated from the area southwest of the Afar plume. As a result, far-field forces were dominated in the EARS region by compression from all sides, enhanced by the topographic high of the Afar Plateau in Ethiopia. 4. The European Cenozoic Rift System The European Cenozoic Rift System (ECRIS) started to evolve ca. 45 Ma ago during the Middle Eocene in the area of the Sa6ne and southern Rhine graben (Fig. 9; Ziegler, 1992b). Within a few Ma it propagated southward across the Pyrenean front into the Valencia Trough, forming the Bresse-Rh6ne graben, and northward to form the Rhine graben. Towards the south, further rifting is indicated by the opening of the Gulf of Lions and the Valencia Trough as well as the volcanism in southern Spain. Northwards, a triple junction formed at the northern end of the Rhine graben leading north-northeastwards into the Leine graben which was abandoned in the Miocene and towards the northwest to the Lower Rhine Embayment (which is now the most active part of the ECRIS). Along this main axis of ECRIS only sporadic volcanic activity occurred except in the Eifel-Westerwald-Vogelsberg region north of the triple junction and in the Valencia Trough. Simultaneously to this more or less continuous line of rift segments two different small rift systems evolved: the grabens of the French Massif Central, the main one being the Limagne graben, and the Eger graben in the Bohemian Massif. In the area of these grabens relatively strong volcanic activity occurred starting in the Oligocene and continuing episodically until Recent times. The crust beneath most of ECRIS is thinned. In the continental part of the rift system, north of the Gulf of Lions, strongest thinning of approximately 30% occurred under the southern end of the Rhine graben and the Limagne graben. The crustal thinning resulted in a Moho uplift of 5-6 km (from 30-32 to 25-26 km; Prodehl and Aichroth, 1992; Zeyen et al., 1997) and a subsidence of sedimentary basins of 2.5-3 km. Much less thinning, however, is observed in the areas of the Lower Rhine Embayment, the Leine graben and the Eger graben (Ziegler, 1990). Thinning underneath the offshore parts, in contrast, is much stronger. In the Valencia Trough crustal thickness is 14 km (Collier et al., 1994) and in the Gulf of Lions oceanic crust has developed (Burrus et al., 1987). Crustal thinning is generally restricted to the area of graben formation. In the Massif Central a regional Moho updoming of about 2 km (from 30-32 km in average to 28-30 km underneath the Massif Central) corresponds well to the estimated amount of regional uplift of the area (Zeyen et al., 1997). Whereas crustal thinning correlates well with the thickness of sedimentary basins, lithospheric thickness is correlated with volcanic activity. A regional average lithospheric thickness of 100-110 km (Babuska and Plomerovfi, 1992) is strongly reduced beneath the Massif Central and the EifelWesterwald area. Teleseismic delay time inversion in the Massif Central (Granet et al., 1995) resulted in clear low-velocity channels beneath the main volcanic eruption centres which join at depths greater than 80 km to a plume-like structure. These results were interpreted by Sobolev et al. (1996) together with petrologically derived thermobarometric data in 343 14. Zeyen et al./Tectonophysics 278 (1997) 329-352 o COo o Fig. 9. The European Cenozoic Rift System and surrounding important tectonic structures. LRE = Lower Rhine Embayment; E = Eifel-Westerwald volcanic area; BM = Bohemian Massif with the Eger graben; RiG = Rhine graben; RoG = Bresse-Rh6ne graben; MC = Massif Central with Limagne graben; Pyr = Pyrenees; Carp = Carpathian Arc; TESZ = Trans-European Suture Zone. The hatching northeast of the TESZ indicates the hard cratonic block of the East European Platform. The shaded area includes onshore areas and the shelf in northwestern Europe. terms of lithospheric thickness. This interpretation yielded a regional, nearly circular thinning to about 80 km with local extrema beneath the volcanic areas where the lithosphere is less than 60 km thick. Potential temperatures in the order of 100-200 K above average are found to depths of more than 250 km (Sobolev et al., 1997). The distribution of anomalously high temperatures does not show, however, any relation to the graben system in the Massif Central. P-wave delay time analyses (Raikes and Bonjer, 1983; Babuska and Plomerovfi, 1992) and petrological evidence support a thinning of the lithosphere beneath the Eifel-Westerwald to about 60 km. A low-velocity anomaly in European P-wave tomographic models down to 300 km (Spakman et al., 1993) indicates a similar plume-like structure in the upper mantle as beneath the Massif Central. Beneath the Rhine graben no clear evidence for substantial lithospheric thinning correlated to the graben has been found. Earlier studies, mainly based on gravimetric and thermal analysis, suggested a mantle plume beneath the southern Rhine graben (Kahle and Werner, 1980), but recent teleseismic studies have not been able to find low-velocity anomalies correlating with the graben beneath 50 km depth (Glahn and Granet, 1992). Surface expressions of the different parts of the rift system are manifold. On the one hand, the main axis from the Mediterranean through the Rh6ne, Bresse and Sa6ne grabens to the Rhine graben forms a single, relatively narrow trough with, in most parts, 344 H. Zeyen et al./Tectonophysics 278 (1997) 329 352 high shoulders which rise up to 1500 m above the graben surface. On the other hand, hardly any uplifted shoulders above the regional uplift are associated with the Limagne graben system and the Lower Rhine Embayment. The high topography west of the Limagne graben is entirely due to syn- and post-rift volcanic edifices. Where volcanic activity fades out towards the north a shoulder is missing although basin subsidence and Moho updoming become maximum. A further distinguishing characteristic of the Massif Central grabens is the splitting apart of a single graben at the northern edge of the Massif Central into several distinct basins towards the centre. This dispersion of deformation happens where the graben enters the area of volcanic activity and regional uplift. This feature is comparable to the splitting up of the Kenya rift at its northern end as it approaches the Anza graben (see above) or the one of the Red Sea towards the southeast above the Afar plume. As in these areas, it may be explained by extension of a weak and relatively hot lithosphere, possibly due to the underlying plume (Table 1, field no. 3). -10 E v' -20 The formation of the ECRIS was contemporaneous with Palaeogene and Neogene phases of Alpine orogeny (Ziegler, 1992b). In contrast to the Red Sea which opened in a similar tectonic regime perpendicular to the orogenic front and parallel to the movement of the lithospheric mantle, in Europe the opening occurred obliquely or perpendicular to the direction of lithospheric mantle movement which is defined by the opening of the Mid-Atlantic ridge and the direction of collision between Africa and Europe. Furthermore, the low level of volcanism and lithospheric thinning underneath the main part of the ECRIS shows that, in contrast to the RSGS, the mantle has hardly been affected by the extensional deformation. This implies that, in contrast to the RSGS, the crust must have moved laterally relative to the underlying mantle in western Europe (Table 1, field no. 4). The reason for this difference must reside in the coupling of the crust to the mantle, i.e. the strength of lower crustal material under the given P - T conditions. The principal possibility and some tectonic consequences of differential movement of f f S a ; / e a -30 / l_l// -40 ' ' ' ' I fast Africa -- -- Africaslow . . . . 0 5. Far-field extension: the lower crust as controlling factor i i ' ' ' ' I ' ' Europe ' ' I 100 200 300 Ext. Strength (MPa) ' ' ' ' 400 Fig. 10. Strength distribution in the lower crust. Comparison between W Europe and NE Africa/Arabia. For Africa strength distributions for deformation rates of 2 x 10-15 s-I (slow) and 2 x 10-14 s I (fast) are shown. the upper crust with respect to the underlying mantle has been investigated, e.g. by Lobkovsky and Kerchman ( 1991 ). We will investigate here the reason for the obvious differences between ECRIS and RSGS. Fig. 10 shows a comparison of lower crustal strength profiles for typical settings in western Europe (last affected by the Hercynian orogeny, ca. 300 Ma ago) and NE Africa (last affected by Pan-African orogenic events, ca. 600 Ma ago), calculated under the assumption of constant strain rate through the lithosphere (constant strain rate model; e.g. Turcotte and Schubert, 1982). A typical western European lithosphere may have a 30-32-km-thick crust, the lithosphere-asthenosphere boundary at a depth of 100 km and a surface heat flow of 80 mW/m 2. If the regions in NE Africa which have not been disturbed by recent rifting are considered as representative for the whole of NE Africa at the beginning of rifting, 345 H. Zeyen et al./Tectonophysics 278 (1997) 329-352 the corresponding parameters in the Red Sea region may have been 40-45 km crustal thickness (Prodehl and Mechie, 1991), 150 km lithospheric thickness and a surface heat flow of 50 mW/m 2. Taking standard rheological parameters for material which is expected to form dominantly the lower crust (Rutter and Brodie, 1992) one sees that the strength at the bottom of the crust is reduced in western Europe by a factor of approximately 2.5 in comparison to NE Africa. This number is based on the assumption of a vertically constant horizontal deformation rate of 2 x 10 -15 s -1 in both areas. The lower crust has, however, deformed considerably faster in the Red Sea area than underneath the ECRIS: in a similar time span it was thinned up to a maximum of 20-30% in Europe whereas it vanished completely underneath the Red Sea. If we assume a ten times higher deformation rate in the latter area, the resistance of the lower crust increases by a factor of two, resulting in a five times larger force to be transmitted through the Moho in the case of NE Africa than underneath Europe. If one does not assume that the crust is deformed as a whole with a constant deformation rate but that the mantle is moving at a constant velocity with respect to the upper crust without deformation in the upper crust, the deformation concentrates in the lowermost part of the lower crust (constant stress model; Mtiller et al., 1997). This concentration results in higher strain rates and therefore larger shear stresses at Moho level. But also in this case the same relations apply for Moho strength as in the constant deformation model. Also the total strength of the crust, calculated as integral of local strength over the whole crust, is much higher for NE Africa than for Europe, by a factor of 2.5 for slow deformation and more than 3.5 for faster deformation. The calculated local ductile strength at Moho level and total crustal strength for the applied theological models are given in Table 2. Where may then the force for convergent crustal deformation in the Alps and Zagros Mountains come from? Ridge push can only account for a part of the required force. Estimates lie generally in the range of 2 - 4 TN/m for ridges of 100 Ma old (e.g. Turcotte and Schubert, 1982; Park, 1988) which was roughly the case in Europe at the time of continent-continent collision. Since generally a continent-continent col- lision is preceded by subduction of oceanic lithosphere, slab pull may also play a major role. The force exerted by a downgoing slab due to its negative buoyancy is calculated as F = p • g - L • H, with p being the mean density difference between slab and surrounding mantle, g the gravimetric acceleration, L the depth to which the slab reaches and H the slab thickness. This force is counteracted by resistance in the mantle along the down-going slab (LithgowBertelloni and Richards, 1995), and by resistance in the crust which consists in deformational resistance, mostly in the upper crust, along the orogenic front or, in case of detachment between upper crust and mantle, in resistance to movement at Moho level. Density differences are estimated to be in the order of 30-40 kg/m 3, assuming a Moho temperature of 500-700°C. In this case a slab, reaching the bottom of the upper mantle and having thus a vertical extension of about 500-600 km with a thickness of 100 km, pulls with a force of 15-25 TN/m. Due to the high temperatures in the sub-lithospheric mantle the resistance per km slab length is similar or smaller than the resistance at Moho level, even for the relatively hot European lower crust (Fig. 10). Therefore, in order to estimate possible detachment lengths at Moho level one has to sum slab and detachment length when calculating the total resistance. Considering the ductile strengths of the lower crust given in Table 2, the slab pull is able to overcome resistance over a length of 1000-1600 km in Europe, assuming the constant stress model and nearly twice as much, in the constant strain rate model. Even when this amount is reduced by a slab length of 500 km largescale detachment along the European Moho seems to Table 2 Results of strength calculations for different lithospheric thermal regimes (Africa, cool; Europe, hot) and different deformation models (constant strain vs. constant stress) in the lower crust Strength at Moho (MPa) (constant strain/stress) Crustal strength (TN/m) (comp./ext.) Lithospheric strength (TN/m) (comp./ext.) Africa fast Africa slow Europe 45/81 21/38 9/15 26/11 22/9 7/4 94/41 88/33 26/13 346 H. Zeyen et al./Tectonophysics 278 (1997) 329 352 be possible. In Africa, however, not more than 1100 km in the case of slow, homogeneous deformation (i.e. 600 km in the case of constant stress in the lower crust) or 600 km (300 km) in the case of fast deformation could be overcome, which is hardly more than the assumed slab length. However, in Africa as well as in Europe, this force is large enough to deform the whole crust in compression along the orogenic front. Although these calculations are evidently very simple, they do show that it is possible to detach the European crust along the whole length from the Mediterranean to the North Sea from the mantle, enabling it to move independently from the lithospheric mantle. Along a line connecting the rigid blocks of the East-European Platform and the African indenter, i.e. from SW Norway to the W Alps, the crust broke and was sheared towards the southwest, opening the series of rifts which form the ECRIS (Fig. 9). In Africa, in contrast, such a detachment would be much more difficult to be achieved and not on a large enough scale to break the whole Arabian microplate. Therefore, the Arabian crust stuck to the mantle and rifting in the Red Sea could only occur parallel to mantle movement, i.e. parallel to the tensional forces acting in the mantle. For the same reason, and in contrast to Europe, in the Red Sea area not only the crust was affected by rifting but the whole lithosphere broke apart. However, the break-up of the whole lithosphere implies a much stronger force than slab pull alone is able to build up. With the numbers of the above described model 40-50 TN/m would be necessary to break the whole Arabian lithosphere under extension compared to the 15-25 TN/m exerted by slab pull. Therefore, in the Red Sea, an additional force was necessary to induce rifting: the initiation of rifting was here triggered by bending and thermal weakening of the lithosphere by a mantle plume. A consequence arising from the crust-mantle detachment as we propose it for western Europe is compressional crustal tectonics north of the Alps towards the continent-ocean boundary contemporaneous with Alpine compression and rifting. Crust and mantle cannot be detached in oceanic lithosphere of the age observed at the North-Atlantic coast since the lithosphere has already cooled and temperatures are not high enough to produce a ductile regime in the thin oceanic crust. Therefore, if we assume a differential velocity between crust and mantle in central-western Europe, there must be a transition zone towards the oceanic realm in which this relative movement is accommodated by compressional tectonics. This area is located in Europe in southern England, the North Sea and the North German Basin where Jurassic-Cretaceous basins (i.e. areas weakened only a few tens of million years before the onset of rifting in ECRIS) have been strongly inverted (Fig. 9; see various publications in Buchanan and Buchanan, 1995, especially Huyghe and Mugnier, 1995). 6. Plume: the preexisting lithospheric thermal regime as controlling factor In the East African Rift System, the Afar, and the French Massif Central we observe how a plume affects lithosphere of different ages. The very thick lithosphere of the Tanzania craton was hardly affected by the plume. The upward pressure of the low-density plume material was large enough to elevate the block as a whole, but the resulting horizontal bending stresses were not large enough to break the craton. There is also no clear evidence for sublithospheric erosion due to the hot plume. It seems more likely that the plume deviated to the sides when it reached the lithosphere. Numerical (Ribe and Christensen, 1994) and physical (Griffiths and Campbell, 1990) modelling of plume-lithosphere interaction show that this deviation of plume movement beneath thick lithosphere is viable. When the plume hits thinner lithosphere where the flexural rigidity is considerably smaller, like in the Pan-African regions surrounding the East African Plateau and in Ethiopia, the vertical force exerted by the buoyant hot material is able to bend the lithosphere considerably stronger, up to the point to break it. In a far-field compressional stress regime like it may have existed in East Africa, lithospheric failure may induce instabilities in the negatively buoyant lithospheric mantle, which allow plume material to rise massively further up to depths where melting occurs, giving rise for voluminous eruptions like flood basalts. Under far-field extension, however, the combination of plume activity and extension produces rifting and may lead under favourable conditions to 347 H. Zeyen et al./Tectonophysics 278 (1997) 329-352 continental break-up and sea-floor spreading like in the Red Sea and Gulf of Aden. In this scenario the plume adds the force to the far-field forces which is necessary to break the whole lithosphere, which none of the two effects on its own would be able to achieve. Away from the direct influence of the plume a narrow, crack-like rift evolves. The evolution above the plume, in contrast, is controlled at the beginning by widespread formation of new crust due to underplating and ductile deformation of a large part of the heated crust which hinders true sea-floor spreading. When a plume hits thin lithosphere like in W Europe, it is able to penetrate directly into depth levels where melting occurs. In this case it seems that the lithosphere is not able to oppose the rise of the buoyant material even of a small plume by bending, forcing the plume to spread, but allows the melts to penetrate it relatively easily (White and McKenzie, 1995). Therefore, no large-scale rifting occurred above the Massif Central plume and the plume does not show any sign of head flattening. The 80-km isobath of the lithosphere-asthenosphere boundary (Fig. 11) is roughly circular and in velocity and temperature profiles (e.g. Sobolev et al., 1997) no typical flat plume head structure is visible. Therefore it seems that the plume is nearly undeformed beneath that depth. At shallower levels, however, when plume material enters the more brittle part of the lithosphere, the deformation is controlled by the crustal stress field, the influence of which obviously 150 100 50 140 E v 0~ O 120 110 a t00 90 80 -50 70 60 40 -100 -100 -50 0 Distance (km) 50 100 150 Fig. 11. Lithospheric thickness under the Massif Central. Note that the border faults (thick lines) turn from a mainly N-S direction to a NW-SE direction as they enter the area with a Iithospheric thickness smaller than approximately70 km. 348 H. Zeyen et al./Tectonophysics 278 (1997) 329-352 penetrates for some distance into the mantle. This influence is also visible underneath the Rhine graben, where a certain correlation between upper crustal graben and mantle structures can be detected down to a depth of about 50 km (Glahn and Granet, 1992). Rifting in the Massif Central is therefore triggered by the plume, but controlled by the far-field crustal stress field. The crustal stress field is, however, only weakly coupled to the mantle stress field and mainly influenced by the combined effect of continentcontinent collision in the Alps and ridge push from the Mid-Atlantic Ridge. 7. Conclusions We have shown the interactions of plumes and farfield extensional stress regimes with the lithosphere under different initial thermal conditions (Table 1). The following conclusions rely mainly on the observations in the European Cenozoic Rift System and the Afro-Arabian Rift System. Although not specifically mentioned, comparison with other rifts like Baikal and the North Sea rifts helped in the understanding of the acting processes. Rifting under far-field extension may be produced in two different forms. In a hot lithosphere which contains a weak lower crust, crust and mantle are decoupled and are deformed in different ways. The crust may in this case be affected by rifting without any extensional deformation of the mantle, resulting in relatively small deformation (Rhine-Rh6ne graben). The observed escape tectonics in W Europe results from the combined effects of drag of the subducting lithospheric mantle and ridge push, which both press the European crust against the harder African indenter. In a colder and stronger lithosphere the lower crust is not ductile enough to allow for large-scale crust-mantle detachment. In this case, rifting, if it occurs, must affect the whole lithosphere. However, unrealistically high stresses would be necessary to break the whole lithosphere. Therefore, rifting in cool lithosphere can only occur in connection with plume activity. This may be the reason why in many cases tension in the mantle, produced by a subducting plate, does not result in rifting (e.g. India). In certain occasions a plume may provoke rifting. The strongest rifting events related to a plume seem to occur in not too old and not too young lithosphere. Hill et al. (1992) discussed two major arguments for mantle plumes being incapable to break cratonic lithosphere. On the one hand the plume cannot rise high enough to produce melts, on the other hand the flexural rigidity of cratons is so high that they do not bend strongly enough to break. If a plume hits cratonic lithosphere it expands laterally until it encounters areas of thinner lithosphere which bend stronger and break easier. In this case whole lithosphere rifting may occur even in a far-field compressional stress regime. Horizontal extensional far-field forces cannot be imagined to be the unique source for whole lithosphere rifting as observed in the Afro-Arabian Rift System for two reasons: firstly, horizontal forces necessary to break cold lithosphere even under extension would be unreasonably high; and secondly, one would not expect this good near-vertical correlation of riftrelated structures from the surface down to at least 150 km depth. At least the relatively ductile lower crustal material would deflect any crack towards the horizontal which penetrated from the surface downwards, yielding much less steep detachment zones at depth (Wernicke, 1985). In a far-field extensional regime a rising plume may provide the additional force needed to break even a relatively cold lithosphere as encountered in the Red Sea-Gulf of Aden rifts. In this case the combined action of local force moment above the plume and regional far-field extension induces opening of a crack penetrating into areas not influenced by the plume which may eventually lead to continental break-up. The direction of penetration may be directly controlled by existing zones of weakness like in the Red Sea (e.g. Vail, 1983; Shimron, 1988), but the direction of opening is controlled mainly by the far-field stress direction. Hot lithosphere like in W Europe and N Kenya (Turkana region) is not able to stop a plume building up elastic forces. Since the plume may easily rise to depths where melting occurs its buoyancy increases and it may penetrate the thin lithosphere. Therefore, no elongated single rift develops like in the above-described rifts. Extensional deformation is not concentrated in the original rift centre but migrates outwards, forming a wide area of subparallel basins and distributed deformation. H. Zeyen et al./Tectonophysics 278 (1997) 329 352 Acknowledgements We thank O. Novak and J. Mechie for their remarks on an earlier version of the manuscript and P. Ziegler and an anonymous reviewer for their comments which helped to improve this publication. This study was enabled by grants of the Deutsche Forschungsgemeinschaft (DFG) within the collaborative Research Program 'Stress and Strain in the Lithosphere' (SFB 108). Most figures were elaborated using the Generic Mapping Tool system (GMT; Wessel and Smith, 1995), References Achauer, U., 1994. New ideas on the Kenya rift based on the inversion of the combined dataset of the 1985 and 1989/90 seismic tomography experiments. Tectonophysics 236, 305329. Babuska, V., Plomerovfi, J., 1992. The lithosphere in Central Europe - - seismological and petrological aspects. Tectonophysics 207, 141-163. Baker, B.H., 1986. Tectonics and volcanism of the southern Kenya rift valley and its influence on rift sedimentation. In: Frostick, L.E. et al. (Eds.), Sedimentation in the African rifts. Geol. Soc. London Spec. Publ. 25, 45-57. Baker, B.H., 1987. Outline of the petrology of the Kenya Rift alkaline province. In: Fitton, J.G., Upton, B.G.J. (Eds.), Alkaline Igneous Rocks. Geol. Soc. London Spec. Publ. 30, 293-311. Baker, B.H., Wohlenberg, J., 1971. Structure and evolution of the Kenya rift valley. Nature 229, 538-542. Bonavia, F.F., Chorowicz, J., Knackstedt, M.A., 1995. Have wet and dry Precambrian crust largely governed Cenozoic intraplate magmatism from Arabia to East Africa?. Geophys. Res. Lett. 22, 2337-2340. Bosworth, W., 1992. Mesozoic and early Tertiary rift tectonics in East Africa. Tectonophysics 209, 115-138. Bosworth, W., 1994. A model for the three-dimensional evolution of continental rift basins, north-east Africa. Geol. Rundsch. 83, 671-688. Bosworth, W., Morley, C.K., 1994. Structural and stratigraphic evolution of the Anza rift, Kenya. Tectonophysics 236, 93115. Bosworth, W., Strecker, M.R., Blisniuk, RM., 1992. Integration of East African paleostress and present-day stress data: Implications for continental stress field dynamics. J. Geophys. Res. 97, 11851-11865. Braile, L.W., Wang, B., Dandt, C.R., Keller, G.R., Patel, J.E, 1994. Modeling the 2-D seismic velocity structure across the Kenya rift. Tectonophysics 236, 251-269. Braile, L.W., Keller, G.R., Wendlandt, R.F., Morgan, E, Khan, A.M., 1996. The East African rift system. In: Olsen, K.H. (Ed.), Continental Rifts: Structure, Evolution, Tectonics. Elsevier, Amsterdam, pp. 213-231. 349 Buchanan, J.G., Buchanan, RG., 1995. Basin Inversion. Geol. Soc. London, Spec. Publ. 88, 596 pp. Burrus, J., Bessis, E, Doligez, B., 1987. Heat flow, subsidence and crustal structure of the Gulf of Lions (NW Mediterranean): A quantitative discussion of the classic passive margin model. In: Beaumont, C., Tankard, A.J. (Eds.), Sedimentary Basins and Basin-Forming Mechanisms. Mem. Can. Soc. Pet. Geol. 12, 1-15. Byme, G., Jacob, B., Mechie, J., Dindi, E., 1997. The upper mantle beneath the southwestern Kenya rift. In: Fuchs, K., Altherr, R., Mtiller, B., Prodehl, C. (Eds.), Structure and Dynamic Processes in the Lithosphere of the Afro-Arabian Rift System. Tectonophysics 278, 243-260 (this volume). Campbell, I.H., Griffiths, R.W., 1990. Implications of mantle plume structure for the evolution of flood basalts. Earth Planet. Sci. Lett. 99, 79-93. Class, C., Altherr, R., Volker, F., Eberz, G., McCulloch, M.T., 1994. Geochemistry of Pliocene to Quaternary alkali basalts from the Huri Hills, northern Kenya. Chem. Geol. 113, 1-22. Clin, M., 1991. Evolution of Eastern Afar and the Gulf of Tadjura. Tectonophysics 198, 355-368. Collier, J.S., Buhl, P., Torn, M., Watts, A.B., 1994. Moho and lower crustal reflectivity beneath a young rift basin: results from a two-ship, wide-aperture seismic-reflection experiment in the Valencia Trough (western Mediterranean). Geophys. J. Int. 118, 159-180. Courtillot, V., Armijo, R., Tapponier, E, 1987. The Sinai triple junction revisited. Tectonophysics 141, 181-190. Daly, M.C., Chorowicz, J., Fairbead, J.D., 1989. Rift basin evolution in Africa: the influence of reactivated steep basement shear zones. In: Cooper, M.A., Williams, G.D. (Eds.), Inversion Tectonics. Geol. Soc. Spec. PuN. 44, 309-334. Davidson, A., Rex, D.C., 1980. Age of volcanism and rifting in southern Ethiopia. Nature 283, 657-658. Dindi, E.W., 1994. Crustal structure of the Anza graben from gravity and magnetic investigations. Tectonophysics 236, 359371. Ebinger, C.J., 1989. Tectonic development of the western branch of the East African rift system. Geol. Soc. Am. Bull. 101, 885-903. Ebinger, CJ., Ibrahim, A., 1994. Multiple episodes of rifting in Central and East Africa: A re-evaluation of gravity data. Geol. Rundsch. 83, 689-702. Ebinger, C.J., Bechtel, T.D., Forsyth, D.W., Bowin, C:O., 1989. Effective elastic plate thickness beneath the East African and Afar Plateaus and dynamic compensation of the uplifts. J. Geophys. Res. 94, 2883-2901. Ebinger, C.J., Yemane, T., WoldeGabriel, G., Aronson, J.L., Walter, R.C., 1993. Late Eocene-recent volcanism and faulting in the southern Ethiopian rift. J. Geol. Soc. London 150, 99108. Eckstein, Y., 1978. Review of heat flow data from the eastern Mediterranean region. Pageoph. 117, 150-159. Fairhead, J.D., Girdler, R.W., 1971. The seismicity of Africa. J. R. Astron. Soc. 24, 271-301. Forte, A.M., Dziewonski, A.M., O'Connell, R.J., 1995. 350 H. Zeyen et al./ Tectonophysics 278 (1997) 329-352 Continent-ocean chemical heterogeneity in the mantle based on seismic tomography. Science 268, 286-288. Girdler, R.W., 1983. Processes of planetary rifting as seen in the rifting and break up of Africa. Tectonophysics, 94, 241-252. Glahn, A., Granet, M., 1992. 3-D structure of the lithosphere beneath the southern Rhine graben. Tectonophysics 208, 149158. Granet, M., Stoll, G., Dorel, J., Achauer, U., Poupinet, G., Fuchs, K., 1995. Massif Central (France): new constraints on the geodynamical evolution from teleseismic tomography. Geophys. J. Int. 121, 33-48. Griffiths, R.W., Campbell, I.H., 1990. Stirring and structure in mantle starting plumes. Earth Planet. Sci. Lett. 99, 66-78. Hankel, O., 1994. Early Permian to Middle Jurassic rifting and sedimentation in East Africa and Madagascar. Geol. Rundsch. 83, 703-710. Haug, G.H., Strecker, M.R., 1995. Volcano-tectonic evolution of the Chyulu Hills and implications for the regional stress field in Kenya. Geology 23, 165-168. Hay, D.E., Wendlandt, R.F., Keller, G.R., 1995. Origin of Kenya Rift plateau-type phonolites: Integrated petrological and geophysical constraints on the evolution of the crust and upper mantle beneath the Kenya Rift. J. Geophys. Res. 100, 1054910557. Hendrie, D.B., Kusznir, N.J., Morley, C.K., Ebinger, C.J., 1994. Cenozoic extension in northern Kenya: a quantitative model of rift basin development in the Turkana region. Tectonophysics 236, 409-438. Henjes-Kunst, F., Altherr, R., 1992. Metamorphic petrology of xenoliths from Kenya and northern Tanzania and implications for geotherms and lithospheric structures. J. Petrol. 33, 11251156. Hill, R.I., Campbell, I.H., Davies, G.E, Griffiths, R.W., 1992. Mantle plumes and continental tectonics. Science 256, 186193. Huyghe, P., Mugnier, J.-L., 1995. A comparison of inverted basins of the Southern North Sea and inverted structures of the external Alps. In: Buchanan, J.G., Buchanan, EG. (Eds.), Basin Inversion. Geol. Soc. London, Spec. Publ. 88, 339-353. Jestin, E, Huchon, E, Gaulier, J.M., 1994. The Somalia plate and the East African Rift System: present-day kinematics. Geophys. J. Int. 116, 637-654. Joffe, S., Garfunkel, Z., 1987. Plate kinematics of the circum Red Sea - - A re-evaluation. Tectonophysics 141, 5-22. Kahle, H.-G., Werner, D., 1980. A geophysical study of the Rhinegraben. II. Gravity anomalies and geothermal implications. Geophys. J. R. Astron. Soc. 62, 631-647. KRISP Working Group, 1987. Structure of the Kenya rift from seismic refraction. Nature 325, 239-242. Kusznir, N.J., Ziegler, EA., 1992. The mechanics of continental extension and sedimentary basin formation: A simple-shear/pure shear flexural cantilever model. Tectonophysics 215, 117-131. Lambiase, J.J., 1989. The framework of African rifting during the Phanerozoic. J. Afr. Earth Sci. 8, 183-190. Lithgow-Bertelloni, C., Richards, M.A., 1995. Cenozoic plate driving forces. Geophys. Res. Lett. 22, 1317-1320. Lobkovsky, L.I., Kerchman, V.I., 1991. A two-level plate tectonics: application to geodynamics. Tectonophysics 199, 343374. Macdonald, R., 1994. Petrological evidence regarding the evolution of the Kenya Rift Valley. Tectonophysics 236, 373390. Maguire, P.K.H., Swain, C.J., Masotti, R., Khan, M.A., 1994. A crustal and uppermost mantle cross-sectional model of the Kenya Rift derived from seismic and gravity data. Tectonophysics 236, 217-249. Makris, J., Ginzburg, A., 1987. The Afar Depression: transition between continental rifting and sea-floor spreading. Tectonophysics 141, 199-214. Makris, J., Rihm, R., 1991. Shear-controlled evolution of the Red Sea: pull apart model. Tectonophysics 198, 441-466. Makris, J., Henke, C.H., Egloff, E, Akamaluk, T., 1991. The gravity field of the Red Sea and East Africa. Tectonophysics 198, 369-381. McNutt, M.K., Diament, M., Kogan, M.G., 1988. Variations of elastic plate thickness at continental thrust belts. J. Geophys. Res. 93, 8825-8838. Mechie, J., Prodehl, C., Koptschalitsch, G., 1986. Ray path interpretation of the crustal structure beneath Saudi Arabia. Tectonophysics 131,333-352. Mechie, J., Keller, G.R., Prodehl, C., Gaciri, S., Braile, L.W., Mooney, W.D., Gajewski, D., Sandmeier, K.-J., 1994. Crustal structure beneath the Kenya Rift from axial profile data. Tectonophysics 236, 179-200. Mohr, E, Zanettin, B., 1988. The Ethiopian flood basalt province. In: Macdougall, J.D. (Ed.), Continental Flood Basalt. Kluwer, Dordrecht, pp. 63-110. Mooney, W.D., Christensen, N., 1994. Composition of the crust beneath the Kenya rift. Tectonophysics 236, 391-408. Moretti, I., Pinet, B., 1987. Discrepancy between lower and upper crustal thinning. In: Beaumont, C., Tankard, A.J. (Eds.), Sedimentary Basins and Basin-Forming Mechanisms. Can. Soc. Pet. Geol., Mem. 12, 233-239. Morley, C.K., 1988. Variable extension in Lake Tanganyika. Tectonics 7, 785-801. Morley, C.K., Wescott, W.A., Stone, D.M., Harper, R.M., Wigger, S.T., Karanja, F.M., 1992. Tectonic evolution of the northern Kenyan rift. J. Geol. Soc. London 149, 333-348. Mfiller, B., Wehrle, V., Zeyen, H., Fuchs, K., 1997. Short-scale variations of tectonic regimes in the western European stress province north of the Alps and Pyrenees. In: Fuchs, K., AItheft, R., Mfiller, B., Prodehl, C. (Eds.), Stress and Stress Release in the Lithosphere - - Structure and Dynamic Processes in the Rifts of Western Europe. Tectonophysics 275, 199-210. Negredo, A., Fernandez, M., Zeyen, H., 1995. The lateral synand post-rift evolution of lithospheric yield strength. Constraints on modes of rifting. Earth Planet. Sci. Lett. 134, 8798. Novak, O., Prodehl, C., Jacob, B., Okoth, W., 1997. Crustal structure of the southeastern flank of the Kenya rift deduced from wide-angle P-wave data. In: Fuchs, K., Altherr, R., MUller, B., Prodehl, C. (Eds.), Structure and Dynamic Pro- H. Zeyen et al./ Tectonophysics 278 (1997) 329-352 cesses in the Lithosphere of the Afro-Arabian Rift System. Tectonophysics 278, 171-186 (this volume). Park, R.G., 1988. Geological Structures and Moving Plates. Blackie, London, 337 pp. Prodehl, C., Aichroth, B., 1992. Seismic investigations along the European Geotraverse and its surroundings in Central Europe. Terra Nova 4, 14-24. Prodehl, C., Mechie, J., 1991. Crustal thinning in relationship to the evolution of the Afro-Arabian rift system: a review of seismic-refraction data. Tectonophysics 198, 311-327. Prodehl, C., Jacob, A.W.B., Thybo, H., Dindi, E., Stangl, R., 1994a. Crustal structure on the northeastern flank of the Kenya rift. Tectonophysics 236,271-290. Prodehl, C., Keller, G.R., Khan, M.A. (Eds.), 1994b. Crustal and Upper Mantle Structure of the Kenya Rift. Tectonophysics, 236, 483 pp. Raikes, S.A., Bonjer, K.-E, 1983. Large-scale mantle heterogeneity beneath the Rhenish Massif and its vicinity from teleseismic P-residuals measurements. In: Fuchs, K., von Gehlen, K., M/ilzer, H., Murawski, H., Semmel, A. (Eds.), Plateau Uplift -the Rhenish Shield - - A Case History. Springer, Berlin, pp. 315-33l. Reeves, CN., Karanja, F.M., McLeod, I.N., 1987. Geophysical evidence for a failed Jurassic rift and triple junction in Kenya. Earth Planet. Sci. Lett. 81,299-311. Ribe, N.M., Christensen, U., 1994. Three-dimensional modeling of plume-lithosphere interaction. J. Geophys. Res. 99, 669682. Ritter, J.R.R., Fuchs, K., Kaspar, T., Lange, F.E.I., Nyambok, I.O., Stangl, R.L., 1995. Seismic images illustrate the deep roots of the Chyulu Hills volcanic area, Kenya. EOS 76, 273278. Rosendahl, B.R., Kilembe, E., Kaczmarick, K., 1992. Comparison of the Tanganyika, Malawi, Rukwa and Turkana rift zones from analyses of seismic reflection data. Tectonophysics 213, 235-256. Rutter, E.H., Brodie, K.H., 1992. Rheology of the lower crust. In: Fountain, D.M., Arculus, R., Kay, R.W. (Eds.), Continental Lower Crust. Developments in Geotectonics, 23, Elsevier, Amsterdam, pp. 201-267. Shirnron, A.E., 1988. The Red Sea line; a late Proterozoic transcurrent fault. Proc. Int. Conf. Basement Tectonics, Butte, MT, Aug. 8-12, p. 31. Slack, ED., Davis, P.M., 1994. Attenuation and velocity of P-waves in the mantle beneath the East African Rift, Kenya. Tectonophysics 236, 331-358. Smith, M., 1994. Stratigraphic and structural constraints on mechanisms of active rifting in the Gregory Rift, Kenya. Tectonophysics 236, 3-22. Smith, M., Mosley, R, 1993. Crustal heterogeneity and basement influence on the development of the Kenya rift, East Africa. Tectonics 12, 591-606. Sobolev, S.V., Zeyen, H., Stoll, G., Werling, F., Altherr, R., Fuchs, K., 1996. Upper mantle temperatures from teleseismic tomography of French Massif Central including effects of composition, mineral reactions, anharmonicity, anelasticity and partial melt. Earth Planet. Sci. Lett. 139,147-163. 351 Sobolev, S.V., Zeyen, H., Granet, M., Stoll, G., Achauer, U., Bauer, C., Werling, F., Altherr, R., Fuchs, K., 1997. Upper mantle temperatures and lithosphere-asthenosphere system beneath the French Massif Central constrained by seismic, gravity, petrologic and thermal observations. In; Fuchs, K., Altherr, R., Mtiller, B., Prodehl, C. (Eds.), Stress and Stress Release in the Lithosphere - - Structure and Dynamic Processes in the Rifts of Western Europe. Tectonophysics 275, 143-164. Spakman, W., Van der Lee, S., Van der Hilst, R., 1993. Traveltime tomography of the Eurnpean-Mediterranean mantle down to 1400 kin. Phys. Earth Planet. Inter. 79, 3-74. Steckler, M.S., ten Brink, U.S., 1986. Lithospheric strength variations as a control on new plate boundaries: examples from the northern Red Sea region. Earth Planet. Sci. Lett. 79, 120-132. Torn6, M., Pascal, G., Buhl, P., Watts, A.B., Mauffret, A., 1992. Crustal and velocity structure of the Valencia trough (western Mediterranean), Part I. A combined refraction/wide-angle reflection and near-vertical reflection study. Tectonophysics 203, 1-20. Turcotte, D.L., Schubert, G., 1982. Geodynamics. J. Wiley, New York, 450 pp. Vail, J.R., 1983. Pan-African crustal accretion in northeast Africa. J. Afr. Earth Sci. 1,285-294. Vellutini, R, 1990. The Manda-lnakir Rift, Republic of Djibouti: a comparison with the Asal Rift and its geodynamic interpretation. Tectonophysics 172, 141-153. Vening Meinesz, F.A., 1960. Les grabens Africains. Rdsultat de compression ou de tension dans la crofite terrestre?. Bull. Sdances Inst. R. Colonial Belge 21,539-552. Voggenreiter, W,, H6tzl, H., Mechie, J., 1988. Low-angle detachment origin for the Red Sea Rift System?. Tectonophysics 150, 51-75. Volker, F., McCulloch, M.T., Altherr, R., 1993. Submarine basalts from the Red Sea: New Pb, Sr and Nd isotopic data. Geophys. Res. Lett. 20, 927-930. Volker, F., Altherr, R., Joehum, K.E and McCulloch, M.T., 1997. Quaternary volcanic activity of the southern Red Sea: New data and assessment of models on mantle sources and Afar plume-lithosphere interaction. In: Fuchs, K., Altherr, R., Mtiller, B., Prodehl, C. (Eds.), Structure and Dynamic Processes in the Lithosphere of the Afro-Arabian Rift System. Tectonophysics 278, 15-29 (this volume). Wernicke, B., 1985. Uniform-sense normal simple shear of the continental lithosphere. Can. J. Earth Sci. 22, 108-125. Wessel, R, Smith, W.H.F., 1995. New version of the generic mapping tools released. EOS Trans. AGU 76, 329. White, R.S., McKenzie, D., 1995. Mantle plume and flood basalts. J. Geophys. Res. 100, 17543-17586. Williams, L.A.J., 1972. The Kenya rift volcanics: a note on volumes and chemical compositions. Tectonophysics 15, 8396. WoldeGabriel, G., Jemane, T., Suwa, G., White, T., Asfaw, B., 1991. Age of volcanism and rifting in the Burji-Soyoma area, southern Main Ethiopian rift: Geo- and biochronology data. J. Air. Earth Sci. 13, 437-447. Zeyen, H., Negredo, A. and Fern~mdez, M., 1996. Extension 352 H. Zeyen et al./Tectonophysics 278 (1997) 329-352 with lateral material accommodation - - 'active' vs. 'passive' rifting. Tectonophysics 266, 121-I38. Zeyen, H., Novak, O., Landes, M., Prodehl, C., Driad, L., Hirn, A., 1997. Refraction seismic investigations of the northern Massif Central, France. In: Fuchs, K., Altherr, R., Mtiller, B., Prodehl, C. (Eds.), Stress and Stress Release in the Lithosphere - - Structure and Dynamic Processes in the Rifts of Westyern Europe. Tectonophysics 275, 99-118. Ziegler, EA., 1990. Geological Atlas of Western and Central Europe, 2nd ed. Shell Int. Pet. Mij., distributed by Geol. Soc. Publ. House, Bath, 238 pp. Ziegler, EA., 1992a. Plate tectonics, plate moving mechanisms and rifting. Tectonophysics 215, 9-34. Ziegler, EA., 1992b. European Cenozoic Rift System. Tectonophysics 208, 91-111. Zoback, M.L., 1992. First- and second-order patterns of stress in the lithosphere: The Word Stress Map Project. J. Geophys. Res. 97, 11703-11728.