Gondwana Breakup Via Double-Saloon-Door Rifting and Seafloor Spreading in a Backarc Basin During Subduction Rollback

Home , Agulhas Plateau, Gastre Fault, Gibraltar Arc, Sea of Crete

Available online at www.sciencedirect.com

Tectonophysics 445 (2007) 245–272 www.elsevier.com/locate/tecto

Gondwana breakup via double-saloon-door rifting and seafloor spreading in a backarc basin during subduction rollback

A.K. Martin

Repsol YPF Exploración, Al Fattan Plaza, PO Box 35700, Dubai, United Arab Emirates Received 2 February 2007; received in revised form 4 July 2007; accepted 21 August 2007 Available online 28 August 2007

Abstract

A model has been developed where two arc-parallel rifts propagate in opposite directions from an initial central location during backarc seafloor spreading and subduction rollback. The resultant geometry causes pairs of terranes to simultaneously rotate clockwise and counterclockwise like the motion of double-saloon-doors about their hinges. As movement proceeds and the two terranes rotate, a gap begins to extend between them, where a third rift initiates and propagates in the opposite direction to subduction rollback. Observations from the Oligocene to Recent Western Mediterranean, the Miocene to Recent Carpathians, the Miocene to Recent Aegean and the Oligocene to Recent Caribbean point to a two-stage process. Initially, pairs of terranes comprising a pre-existing retro-arc fold thrust belt and magmatic arc rotate about poles and accrete to adjacent continents. Terrane docking reduces the width of the subduction zone, leading to a second phase during which subduction to strike-slip transitions initiate. The clockwise rotated terrane is caught up in a dextral strike-slip zone, whereas the counterclockwise rotated terrane is entrained in a sinistral strike-slip fault system. The likely driving force is a pair of rotational torques caused by slab sinking and rollback of a curved subduction hingeline. By analogy with the above model, a revised five-stage Early Jurassic to Early Cretaceous Gondwana dispersal model is proposed in which three plates always separate about a single triple rift or triple junction in the Weddell Sea area. Seven features are considered diagnostic of double-saloon-door rifting and seafloor spreading:

i) earliest movement involves clockwise and counterclockwise rotations of the Falkland Islands Block and the Ellsworth Whitmore Terrane respectively; ii) terranes comprise areas of a pre-existing retro-arc fold thrust belt (the Permo-Triassic Gondwanide Orogeny) attached to an accretionary wedge/magmatic arc; the Falklands Islands Block is initially attached to Southern Patagonia/West Antarctic Peninsula, while the Ellsworth Whitmore Terrane is combined with the Thurston Island Block; iii) paleogeographies demonstrate rifting and extension in a backarc environment relative to a Pacific margin subduction zone/ accretionary wedge where simultaneous crustal shortening occurs; iv) a ridge jump towards the subduction zone from east of the Falkland Islands to the Rocas Verdes Basin evinces subduction rollback; v) this ridge jump combined with backarc extension isolated an area of thicker continental crust — The Falkland Islands Block; vi) well-documented EW oriented seafloor spreading anomalies in the Weddell Sea are perpendicular to the subduction zone and propagate in the opposite direction to rollback;

E-mail address: [email protected].

vii) the dextral strike-slip Gastre and sub-parallel faults form one boundary of the Gondwana subduction rollback, whereas the other boundary may be formed by inferred sinistral strike-slip motion between a combined Thurston Island/Ellsworth Whitmore Terrane and Marie Byrd Land/East Antarctica. © 2007 Elsevier B.V. All rights reserved.

Keywords: Gondwana breakup; Double-saloon-door seafloor spreading; Plate tectonics; Backarc basin; Subduction rollback; Opposite rotations of terranes

1. Introduction Fourthly, it is shown that these symptomatic features match the characteristics of Gondwana breakup. Although it is known that Gondwana breakup Fifthly, previously proposed driving mechanisms are initiated in a backarc inboard of a subduction zone on compared to the double-saloon-door model. The its Pacific margin (Storey et al., 1996; 1999), and was geodynamics of double-saloon-door rifting and seafloor contemporaneous with eruption of the Karoo, Ferrar and spreading constitute a new driving mechanism for Chon Aike Large Igneous Provinces (Encarnacion et al., continental breakup. 1996; Duncan et al., 1997; Pankhurst et al., 2000), Finally, some issues are discussed and possible future proposed driving mechanisms (Dalziel and Grunow, tests of the hypothesis are outlined. 1992; Rapela and Pankhurst, 1992; Storey, 1995) explain neither the breakup geometry nor the details of 2. Double-saloon-door rifting and seafloor spreading plate dispersal. Similarly, Weddell Sea magnetic sea- in the Chattian – Langhian Valencia Trough, floor spreading anomalies (e.g. Livermore and Hunter, Liguro-Provencal and Algerian Basins, Western 1996; Ghidella et al., 2002) which are perpendicular to Mediterranean the reconstructed Pacific margin subduction zone do not match models which envisage arc-parallel seafloor The double-saloon-door rifting and seafloor spread- spreading. Furthermore, no coherent geodynamic ing model was developed in the Chattian to Langhian cause has emerged for the opposite rotations of the Western Mediterranean where two arc-parallel rifts Ellsworth Whitmore Terrane and the Falkland Islands propagate in opposite directions from an initial central (Watts and Bramall, 1981; Mitchell et al., 1986). location during backarc seafloor spreading and subduc- Here it is proposed that initial Gondwana breakup tion rollback (Martin, 2006). The resultant geometry (Fig. 1) was driven by double-saloon-door rifting and causes pairs of terranes to simultaneously rotate seafloor spreading and that this process successfully clockwise and counterclockwise like double-saloon- explains breakup geodynamics thereby resolving the doors revolving about their hinges. In Fig. 2, plate West above issues. The purpose of this contribution is six-fold. 2 rotates clockwise about pole P1, whereas plate East 2 First, the double-saloon-door rifting and seafloor rotates counterclockwise about pole P2. Movement is spreading scenario, which was developed for the driven by a centrally located force pulling in the Chattian to Langhian development of the Western direction of subduction rollback (b towards q in Fig. 2b). Mediterranean (Martin, 2006) is summarised. Between a and c as well as d and f, movement is accom- Secondly, it is proposed that double-saloon-door modated by extension of thinned continental crust (dark rifting and seafloor spreading also occurred in the shading) whereas from c to b and from f to e, oceanic Serravalian Gibraltar Arc, the Tortonian — Recent accretion occurs (light shading). With further rotation, Tyrrhenian Basin, the Burdigalian — Recent Pannonian (Fig. 2b and c) rifts propagate towards the rotation poles, as Basin, the Serravalian — Recent Aegean, and the demonstrated by seafloor spreading isochrons (dark lines Oligocene — Recent Caribbean. within light shaded oceanic crust) which successively abut Thirdly, a review of these areas (Lonergan and thinned continental crust. A third rift, orthogonal to the White, 1997; Pindell et al., 1999; Duermeijer et al., subduction zone, opens between the rotating terranes and 2000; Mann et al., 2002; Rosenbaum et al., 2002; propagates in the opposite direction to subduction Wenzel et al., 2002; Van Hinsbergen et al., 2006) rollback. suggests a generalised two-stage model, incorporating a In its nascent stage when it is floored by thinned number of symptomatic features. These characteristics continental crust before oceanic crustal breakthrough, constrain a proposed driving mechanism. the spherical triangular area b i j is akin to a rift formed A.K. Martin / Tectonophysics 445 (2007) 245–272 247

Fig. 1. Reconstruction of pre-breakup (pre-190 Ma) Gondwana which aligns Permo-Triassic fold thrust belts (thick dark dashed lines) of the Sierra de la Ventana in South America (SV), Cape Fold Belt in South Africa (CFB), via the Falkland Islands (FI) in their rotated position (Mitchell et al., 1986), the Ellsworth Whitmore Terrane (EWT) in its rotated position (Watts and Bramall, 1981), and the TransAntarctic Mountains (TAM). West Antarctic Peninsula (WAP) and Thurston Island (TI) positions after Grunow et al (1991) and Grunow (1993a,b). Eastern displacement of southern Patagonia (PAT) along the Gastre Fault (GF) to its pre-breakup position (after Rapela and Pankhurst, 1992; Ben-Avraham et al., 1993; Marshall, 1994). East Antarctica close to the Lebombo volcanic lineament (LE) and the southeastern face of the Maurice Ewing Bank (MEB) (after Martin and Hartnady, 1986; Livermore and Hunter, 1996; Lawver et al., 1999, 160 Ma fit; Jokat et al., 2003). Paleopositions of the Agulhas Plateau (AP) and southern Mozambique Ridge (MR) are not constrained. Accretionary wedges and subduction zones of WAP and PAT overlap (cf Lawver et al., 1999). MBL=Marie Byrd Land. NZ=New Zealand. Dot dash line outlines the Karoo Large Igneous Province in Africa, extending to Dronning Maud Land (DML) in East Antarctica, and the Ferrar Province in East Antarctica (dates after Encarnacion et al., 1996; Duncan et al., 1997). Dot dash lines dated 175 Ma and 160 Ma in southwest Gondwana separate the three divisions of the Chon Aike Large Igneous Province (188–178 Ma, 172–162 Ma, and 157–153 Ma) and demonstrate Pacificward migration of volcanism over time (Pankhurst et al., 2000). at a triple junction (Fig. 2a). When oceanic crustal Provencal Basin opened as the combined Balearic breakthrough occurs, and the rift has propagated Peninsula/Kabylies/Alboran terrane rotated clockwise, northwards (Fig. 2c), the three oceanic rifts form an and the Sardinia/Corsica/Calabria terrane rotated counter RRR triple junction. The scenario is similar to the North clockwise (Fig. 3). In the following Burdigalian/Langhian Fiji Basin where a spreading ridge perpendicular to a stage, seafloor spreading initiated in the Algerian and curved subduction slab is associated with clockwise and Liguro-Provencal Basins, and the pair of rotating terranes counter clockwise rotations of the New Hebrides Arc comprised the Kabylies/Alboran and the Sardinia/Cor- and the Fiji Platform respectively (Pelletier et al., 1998; sica/Calabria Block (Pares et al., 1992; Lonergan and Ruellan and Lagabrielle, 2005). White, 1997; Gueguen et al., 1998; Rosenbaum et al., In the Chattian – Burdigalian episode of Western 2002; Ferrandini et al., 2003). These terranes comprise Mediterranean development, the Valencia Trough, floored Hercynian basement with Mesozoic cover (Bouillin et al., entirely by thinned continental crust, and the Liguro- 1986; Roca et al., 1999) deformed in the preceding 248 A.K. Martin / Tectonophysics 445 (2007) 245–272

whereas a rift propagating northeast is shown by successive magnetic anomalies abutting thinned continental crust towards the apex of the Liguro-Provencal Basin (Fig. 3). The central location of the initial rift, and the axial position of greatest extension as shown by refraction, reflection, gravity and heatflow data (Maerten and Séranne, 1995; Maillard and Mauffret, 1999; Negredo et al., 1999; Rollet et al., 2002) agrees with the model. The boundary between clockwise and counter clockwise rotations is marked by the North Balearic Fracture Zone, which in turn is oriented towards the maximum curvature of the reconstructed subduction/forearc zone (Fig. 3). In a triangular area southwest of Sardinia, magnetic anomalies are aligned almost perpendicular to the reconstructed subduction zone, with the central (presumed youngest) anomaly extending farthest to the northwest in agreement with the model-predicted northwest propagation of a rift in this area (compare Figs. 2 and 3). During the Valencia Trough/Liguro-Provencal and the Liguro-Provencal/Algerian Basin opening phases, there are no major fracture zones at the extremities of the reconstructed subduction zone, although there are several fracture zones within the backarc basin areas (Fig. 3). The implication is that, in these cases, the pairs of clockwise and counterclockwise rotated terranes pivoted about rotation poles rather than resulting from interaction with sinistral and dextral strike-slip zones respectively.

3. Additional examples of double-saloon-door rifting and seafloor spreading

3.1. Gibraltar Arc and the Alboran Basin

Fig. 2. Double-saloon-door rifting and seafloor spreading model (after Rollback of a subducting slab (Royden, 1993; Martin, 2006). Crosses depict continental crust. Dark shaded areas are extended continental crust, whereas light shading signifies oceanic Lonergan and White, 1997; Rosenbaum et al., 2002) crust, with solid lines portraying seafloor spreading isochrons. Fig. 2a, initially towards the south, and then towards the west b and c are further clarified in Section 2. southwest is consistent with overall NW oriented convergence and clockwise rotation in Spain (Fig. 3) Cretaceous and Paleogene phases of the Alpine orogeny and SW oriented convergence and counterclockwise 65 – 25 Ma (Brunet et al., 2000; Rossetti et al., 2001; rotation in Morocco. Although individual clockwise Rosenbaum and Lister, 2004a). HPLT rocks from this rotations reach 200° (Platt et al., 2003 and references orogen are exhumed via relaxation of former low-angle therein), it is thought that the Betic sector of the arc as a thrusts as extensional detachments (Monié et al., 1992; whole rotated 25° clockwise. Similarly individual Crespo-Blanc and Campos, 2001; Rossetti et al., 2001; rotations in the Rif of up to 100° occur, whereas the Martínez-Martínez and Azañón, 2002), accompanied by entire Rif sector rotated 55° counterclockwise (Platt Late Oligocene Miocene rift fill. Calc-alkaline magmatic et al., 2003). No single through-going strike-slip fault arc volcanics are generally contemporaneous with rift occurs as originally thought (LeBlanc and Olivier, sediments followed later by alkaline volcanics (Marti 1984). Rather, oblique dextral transpressive movement et al., 1992; Maury et al., 2000; Duggen et al., 2004). in Spain was taken up along the Socovos and Tiscar A rift propagating southwest in the Valencia Trough is Faults, and the Zegri and Malaga-Teba shear zones, demonstrated by well and seismic data (Alvarez-de- whereas oblique sinistral motion in Morocco was Buergo and Melendez-Hevia, 1994; Roca et al., 1999), accommodated by the Jebha and Nekor Faults (Platt ..Mri etnpyis45(07 245 (2007) 445 Tectonophysics / Martin A.K. – 272

Fig. 3. Summary of three episodes of opposite rotations (marked by circular arrows) in the Western Mediterranean (after Pares et al., 1992; Gueguen et al., 1998; Ferrandini et al., 2003; Martin, 2006). In the Chattian to Burdigalian rotation (poles and arrows marked cb) of the Balearic Peninsula (BAL) and the Sardinia (SAR)/Corsica (CO) Block the Valencia Trough (VT) opens and the Liguro- Provencal Basin (LPB) initiates. In the Burdigalian to Langhian (bl), oceanic crust (light shading) is emplaced (pole P1bl now near the Gibraltar Arc, while pole P2bl remains at the northeast tip of the Liguro-Provencal Basin). AB=Algerian Basin. Black lines with crosses=positive magnetic anomalies. Black lines=Fracture zones (after Rehault et al., 1984; Rollet et al., 2002). Note anomalies orthogonal to the subduction zone southwest of Sardinia. NBFZ=North Balearic Fracture Zone. At the end of this phase Sardinia has docked with the Galite Bank off Tunisia while Grande and Petite Kabylies (GK and PK) have accreted to North Africa at the reconstructed subduction zone (line with dark triangles). Pel=Peloritani. Ap=Appenines. Serravalian poles P1s and P2s are speculatively positioned based on clockwise rotations in the Betics of southern Spain, counterclockwise rotations in the Morrocan Rif and west southwest extensional detachments (Martínez-Martínez and Azañón, 2002). CA=Cadiz Alicante Fault (originally thought to extend from Cadiz on the west coast to Alicante on the east coast). Dextral faults in the Betics include CRF=Crevillente Fault. MS=Malaga – Teba Shear Zone, SF=Socovos Fault, TF=Tiscar Fault and ZS=Zegri Shear Zone (Platt et al., 2003). JB and NF are the sinistral Jebha and Nekor Faults in the Morrocan Rif. 249 250 A.K. Martin / Tectonophysics 445 (2007) 245–272 et al., 2003). Thrusting was coeval with a NNW–SSE slab detachment, asthenospheric upwelling and bimodal oriented Burdigalian–Langhian extension phase fol- volcanism (Maury et al., 2000; Duggen et al., 2004). lowed by a WSW directed Serravalian phase (Comas et al., 1999; Crespo-Blanc and Campos, 2001; Martínez- 3.2. Tyrrhenian Basin Martínez and Azañón, 2002), with extension on low- angle detachment faults. Teleseismic tomography has The Tyrrhenian Basin developed in two episodes imaged a narrow high-velocity east-dipping subducting (Fig. 4). The first is post-Langhian, with intra-Tortonian – slab under the Gibraltar Arc (Wortel and Spakman, Messinian rifting (9 – 5 Ma) leading to emplacement of 2000; Faccenna et al., 2004) which has been linked to oceanic crust in the Vavilov Basin from 4.3 Ma to 2.5 Ma

Fig. 4. P1t and P2t=schematic rotation poles for pre-Late Tortonian clockwise rotation of Sicily and contemporaneous (post-Langhian) counter clockwise rotation of the Appenines respectively. Based on the form of rifted crust off Sardinia, (areas deeper than 2500 m shown by inclined parallel lines) and paleogeographic reconstructions (Gattacceca and Speranza, 2002; Rosenbaum and Lister, 2004b). Shaded area=oceanic crust of the Vavilov (V) and Marsili (M) Basins, separated by the Issel Bridge (IB). Br and O=interpreted Brunhes and Olduvai magnetic seafloor spreading anomalies after Nicolosi et al. (2006). Co=Corsica. Sar=Sardinia. LA and CL=Latium-Abruzzi and Campania–Lucania carbonate platforms of peninsular Italy whereas Pa=Panormide platform in west Sicily. P1pp and P2pp=possible rotation poles for Plio-Pleistocene clockwise rotation of Sicily and counter clockwise rotation of Southern Appenines coeval with seafloor spreading in the Vavilov Basin 4.6 – 2.6 Ma and post-1.9 Ma in the Marsili Basin. Note sinistral faults north of the oceanic crust in the Tyrrhenian Sea. 41P=forty first parallel fault. SL=SanGineto Line. Dextral faults occur to the south. NSF=North Sicily Fault, TL=Taormina Line (after Brunet et al., 2000; Sartori et al., 2001; Rosenbaum et al., 2002; Sartori, 2003; Faccenna et al., 2004). A.K. Martin / Tectonophysics 445 (2007) 245–272 251

(Feraud, 1990; Sartori et al., 2001). After a ridge jump to emplacement of thrust sheets in the fold thrust belts, and the southeast which isolated the thicker crust of the Issel contemporaneous with rifting and emplacement of Bridge, the Marsili Basin developed from 3 Ma, with MORB oceanic crust in the Tyrrhenian backarc basin. oceanic crust from 1.9 Ma. Magnetic anomalies are Rollback was restricted to a narrow zone 250 km related to individual volcanos (Tontini et al., 2004)orto wide by the arrival of buoyant continental crust at the linear volcanic centres exploiting faults (De Ritis et al., subductionzonefrom6to5Ma(Sartori, 2003; 2005), but spectral analysis and orthogonal filtering Rosenbaum and Lister, 2004b). This comprised the suggest linear seafloor spreading anomalies (Nicolosi Campania–Lucania, Latium-Abruzzi, and Apulian car- et al., 2006). bonate platforms to the northeast of the Calabrian Arc, This two-stage development is reflected in paleo- while the Panormide, Trapanese, and Saccanese plat- magnetic data. A post-Langhian 60° counter clockwise forms docked in Sicily. Late-stage rotations likely rotation in the Apennines is mirrored by a pre-Late occurred about poles near these collision points Tortonian 70° clockwise rotation in Sicily, whereas a (Fig. 4). This geometry is consistent with the pattern further 20° Plio-Pleistocene counter clockwise rotation of rifting in the Tyrrhenian Basin where sub-basins in southern Italy is matched by a contemporaneous 30° narrow both to the southwest and northeast, and with clockwise rotation in Sicily (Gattacceca and Speranza, sinistral strike-slip along the SanGineto Line and related 2002; Speranza et al., 2003). Rotations are coeval with faults, and dextral movement on the North Sicily Fault

Fig. 5. Carpathians and Pannonian Basin (After Patrascu et al., 1994; Panaiotu, 1998; Chalot-Prat and Girbacea, 2000; Konecny et al., 2002; Wenzel et al., 2002). Lines with triangles=fold thrust belt/accretionary wedge (AW) of the Carpathians. AB=Alcapa or North Pannonian Block. TDB=Tisza–Dacia Block. MHL =mid-Hungarian Line. TCB=Trans Carpathian Basin. TB=Transylvanian Basin. NTF and TF=North Transylvanian and Trotus sinistral strike-slip faults whereas STF, IMF and SF=South Transylvanian, Intra-Moesian and Sinaia dextral faults. V=Vrancea Zone of earthquake epicentres. Subscript b refers to Burdigalian counter clockwise paleomagnetic rotations in the Alcapa Block which may have occurred about rotation pole P1b, whereas P1t is a suggested pole for Tortonian rotations in the TransCarpathian Basin shown by subscript t. P2 is a suggested pole for the clockwise rotations of the Tisza–Dacia Block. 252 A.K. Martin / Tectonophysics 445 (2007) 245–272 and the Taormina Line. Tears in the subducting slab the European Platform is accommodated by eastward have been imaged close to these fault systems in the rollback and conjugate sinistral and dextral strike-slip southern Apennines and the Sicily Channel (Wortel and fault systems (Royden, 1993; Patrascu et al., 1994; Spakman, 2000; Faccenna et al., 2004). Panaiotu, 1998; Wortel and Spakman, 2000; Wenzel et al., Development of the Western Mediterranean (Sec- 2002). The separation of the two blocks by the Mid tions 2–3.2) was episodic. Successive rollback phases Hungarian strike-slip fault is similar to the model shown were closed by docking of terranes with buoyant in Fig. 2a. Counter clockwise rotations had migrated to continental crust both on the over-riding and subducting the TransCarpathian Basin by the Tortonian implying plate (Gueguen et al., 1998; Rosenbaum et al., 2002; early and late-stage rotation poles (Fig. 5). Terrane Martin, 2006). These punctuations caused ridge jumps docking and progressive detachment led to the present towards the subduction zone, isolating slivers of narrow subducting slab which coincides with a restricted continental crust (termed “boudins” by Gueguen et al., area of seismicity (Wenzel et al., 2002¸Martin and Ritter, 1997) such as the Balearic Peninsula, the Corsica/ 2005). Associated volcanism evolved from extension- Sardinia Block and the Issel Bridge separating the related silicic volcanism to subduction-related calc- Vavilov and Marsili Basins. alkaline magmatics and thence to alkaline volcanism (Chalot-Prat and Girbacea, 2000; Konecny et al., 2002). 3.3. The Carpathians and the Pannonian Basin 3.4. The Aegean A similar geodynamic scenario is displayed by the Carpathians (Fig. 5), where opposite rotations of the A pair of opposite rotations (Fig. 6) also occurs during Alcapa and Tisza–Dacia Blocks and their docking with southwestward rollback of the Hellenic subduction zone

Fig. 6. Aegean Sea (after Hetzel et al., 1995; Duermeijer et al., 2000; Gurer and Yilmaz, 2002; Ten Veen and Kleinspehn, 2003; Benetatos et al., 2004; Van Hinsbergen et al., 2006). In Greece, clockwise rotations have been recorded in Ev=Evia, M=Mykonos, Pe=Peleponessos, S=Skyros and T=Tinos. A, C and L refer to the Argolis, Corinth and Lucania Gulfs. KFZ=dextral Kefallonia Fault. Counter clockwise rotations have been registered in the central and southeastern Aegean in Crete, Kar=Karpathos, Kas=Kassos, N=Naxos, Rh=Rhodos, and in the Be=Bey Daglari area in southwestern Turkey. B, G, BM and O=Bergama, Gediz, Buyuk Menderes and Oren Graben. PT=sinistral strike-slip faults in the Pliny Trench southeast of Karpathos and Rhodos. P1 and P2 are suggested poles of rotation for clockwise and counter clockwise rotated blocks. ..Mri etnpyis45(07 245 (2007) 445 Tectonophysics / Martin A.K. – 272

Fig. 7. Caribbean Sea (after Muller et al., 1999; Pindell et al., 1999; Pindell and Kennan, 2001; Mann et al., 2002). On the northeastern Caribbean, EF and SF=Enriquillo and Septentrional sinistral strike-slip faults. MuF and NHF=Muertos and North Hispaniola thrust faults. PR=Puerto Rico and PRT=Puerto Rican Trench. On the southeastern flank, CF, EPF, MF, NCF and OF=respectively Coche, El Pinar, Moron, North Coast and Oca dextral faults. Si=Serranía Interior. T=Tobago Tr=Trinidad. Rotations after Van Fossen et al. (1989), Reid et al. (1991), and Burmester et al. (1996). 253 254 A.K. Martin / Tectonophysics 445 (2007) 245–272 and backarc extension in the Aegean (Angelier et al., Interior has been interpreted as a retro-arc fold/thrust 1982; Kissel and Laj, 1988). The clockwise rotated block belt (Pindell et al., 1999). The collision zone is also comprises the Peleponessos and the islands of Evia, marked by a slab detachment (VanDecar et al., 2003). Skyros, Mykonos and Tinos, whereas the islands of Rhodos, Karpathos, Kassos, Crete and Naxos along with 4. Generalised two-stage model for double-saloon-door the Bey Daglari area in southwestern Turkey register rifting and seafloor spreading during subduction counter clockwise rotations (Duermeijer et al., 2000; Van rollback, and proposed driving mechanism Hinsbergen et al., 2006). Clockwise rotation of the Peleponessos is consistent with NW – SE oriented The examples given in Sections 2–3.5 above (Figs. 3 extensional basins of the Gulf of Corinth (Armijo et al., 4 5 6 and 7) suggest a two-stage generalised model 1996) and the neighbouring Gulfs of Argolis and Lucania. (Figs. 8 and 9). The early stage (Fig. 8a) is based on the Counterclockwise rotation of Rhodos and Crete accords Valencia Trough, Liguro-Provencal and Algerian Basins with the E–Woriented Bergama, Gediz, Buyuk Menderes (Fig. 3). In these cases, two microplates or terranes and Oren graben with Miocene to Recent fill (Gurer and simultaneously rotate clockwise and counter clockwise Yilmaz, 2002) where basin boundary detachments expose in a backarc. Terranes comprise a pre-existing retro-arc the Central Menderes metamorphic core complex (Hetzel fold/thrust belt inboard of a forearc/subduction zone/ et al., 1995; Gessner et al., 2001). Dextral strain is taken magmatic arc. In the Valencia Trough, opening was up along the transpressive Kefallonia, Kastaniotikos and achieved through extension of continental crust (compare related faults (Van Hinsbergen et al., 2006), whereas field Fig. 2a) whereas oceanic breakthrough occurred in the work and earthquake focal mechanisms demonstrate Liguro-Provencal and Algerian Basins. The extreme sinistral transtensive strike-slip faults in Crete and in edges of the rhomboid or triangular shaped backarc the Pliny Trench southeast of Rhodos (Ten Veen and basin are not defined by significant strike-slip fault Kleinspehn, 2003; Benetatos et al., 2004). systems. Rather, the terranes rotate clockwise and anticlockwise about rotation poles (compare (Figs. 8 and 9 3.5. The Caribbean with Fig. 2)). Rifts propagate towards both rotation poles. A third rift develops with anomalies orthogonal to the Oblique collision of Hispaniola Island with buoyant reconstructed subduction zone. This is consistent with crust of the Bahamas Platform from the Oligocene– arc-parallel extension observed in many arcs (McCaffrey, Miocene (Fig. 7) was accompanied by thrusting, nappe- 1996). stacking and counter clockwise rotations in Puerto Rico The late stage (Figs. 8b 9a and b) is based on the (Van Fossen et al., 1989; Reid et al., 1991; Muller et al., Serravalian Betic Rif Arc (Fig. 3), the Tortonian–Recent 1999; Pindell et al., 1999; Pindell and Kennan, 2001). Tyrrhenian Basin (Fig. 4), the Late Burdigalian to GPS studies show that Hispaniola acts as a separate Recent Pannonian Basin (Fig. 5), the Serravalian to microplate as it slows and accretes, whereas Puerto Rico Recent Aegean (Fig. 6), and the Late Oligocene–Recent is carried towards the trench on the Caribbean Plate Caribbean (Fig. 7). In these cases, the subduction zone (Mann et al., 2002). Strain is partitioned between the becomes partially choked by the arrival of buoyant outer low-angle North Hispaniola and Muertos thrust continental crust which resists subduction. Such colli- faults, and the sub-vertical sinistral Septentrional and sions restrict active subduction of oceanic crust to a Enriquillo strike-slip faults. The subducting slab is narrower zone where continued pull by the slab sinking detached under the Hispaniola collision zone, whereas it rollback force breaks rotated terranes into sub-terranes. is continuous under the Puerto Rico trench. These rotate about late-stage poles P1L and P2L. The complementary oblique collision of the Carib- Further subduction rollback (Fig. 9a and b) leads to bean Plate with South America is marked by the Lara renewed collision with continental crust on the southern and Villa de Cura nappes and clockwise rotations in the flanks of the sub-terranes, which probably induces strike- Guajira Peninsula, the Caribbean Mountains and the slip deformation. In reality a combination of rotation islands of Aruba, Bonnaire and Tobago (Burmester (Fig. 8b) and strike-slip (Fig. 9a and b) would be likely, et al., 1996). The Margarita Island and Tobago terranes resulting in dextral strike-slip faults rotated to the were accreted along the dextral Moron, Coche, North southwest about pole P1L, and sinistral faults rotated Coast and El Pilar strike-slip faults, while backarc southeast about pole P2L. This geometry is shown by the extension exhumed HPLT core complexes (Ave Lalle- Gibraltar, Calabrian and Carpathian Arcs (Figs. 3, 4 and ment and Guth, 1990; Ave Lallement, 1997). Accreted 5). A second backarc basin, which opens north of the terranes comprise island arcs whereas the Serranía rotating sub-terranes, has a complex form, being a A.K. Martin / Tectonophysics 445 (2007) 245–272 255 combination of E–W extension between the original rotating terranes, and further opening between the rotating sub-terranes (compare Fig. 2 with Figs. 8band9a and b).

Fig. 8. a) Early stage model for double-saloon-door seafloor spreading during subduction rollback, based on Chattian–Burdigalian and Burdigalian–Langhian (Fig. 3) phases of West Mediterranean backarc extension and subduction rollback. Legend as in Fig. 2. Black line with dark triangles=subduction rollback. Short lines with smaller triangles=- Fig. 9. a) Late stage model for double-saloon-door seafloor spreading fold thrust belt/accretionary wedge. Open arrow=force exerted by slab with strike-slip faults. Rotation of the sub-terranes about late-stage poles P L and P L leads to further collision with continental crust as sinking and rollback. With rotation about early poles P1E and P2E, 1 2 terranes collide with buoyant continental crust. Note that during this shown by open triangles on the southern flanks of the sub-terranes. stage, the extremities of the backarc basin are not bounded by major This foments deformation of the sub-terranes by strike-slip faults. It strike-slip faults. b) Late stage of double-saloon-door seafloor is likely that a combination of rotation (Fig. 8b) and strike-slip spreading model based on Serravalian Gibraltar Arc, Tortonian — deformation (Fig. 9a) occurs simultaneously. This would produce Recent Tyrrhenian Basin, Burdigalian — Recent Pannonian Basin, dextral strike-slip faults oriented more south southwesterly related to Serravalian — Recent Aegean, and Oligocene — Recent Caribbean Sea late-stage pole P1L, and more south southeasterly sinistral faults (Figs.3456and7). Open triangles=backthrusts where rotated terranes associated with late-stage pole P2L (compare Figs. 3, 4 and 5). have collided with continental crust to the south choking subduction in STEP=subduction transform edge propagator. See Section 4 for that zone. Continued application of the slab sinking rollback force exerts further explanation. b) Further subduction rollback may induce ad- a torque which breaks the rotated terranes into sub-terranes, which rotate ditional strike-slip deformation. Note that the STEP not only propagates in the direction of subduction rollback, but propagates laterally about late-stage poles P1L and P2L. A second basin opens in a backarc position relative to the rotating sub-terranes. The model (Fig. 2)predictsa too (compare Fig. 9a and b). third basin which opens close to the accretionary wedge and propagates away from the subduction zone. Further explained in Section 4, whereas the late-stage double-saloon-door model is examined further in Fig. 9. 256 A.K. Martin / Tectonophysics 445 (2007) 245–272

The lateral extremes of the basin which is backarc to the plan view of convection within the mantle wedge (Martin, sub-terranes are marked by pairs of sinistral and dextral 2006). strike-slip fault systems. Associated slab detachments indicate that subduction transform edge propagators 5. Gondwana breakup (STEPS — Bilich et al., 2001; Govers and Wortel, 2005) develop at the boundary between active and blocked The reconstruction used here (Fig. 1) juxtaposes the subduction (Fig. 9). With time, the STEPS not only Dronning Maud Land margin of East Antarctica with propagate in the direction of subduction rollback, but also the Lebombo Line in southeastern Africa and the eastern propagate laterally as the subduction zone is progressively face of the Falkland Plateau (Martin and Hartnady, choked by further collisions of rotated sub-terranes 1986; Livermore and Hunter, 1996; Lawver et al., 1999 (compare Fig. 9a and b). Associated volcanism evolves their 160 Ma reconstruction; Jokat et al., 2003). This from extension-related silicic magmatism, to subduction- aligns fracture zones off Africa with their counterparts related calc-alkaline volcanics, and then to alkaline off Antarctica (Haxby, 1987; Sandwell and Smith, volcanism related to asthenospheric upwelling. 1997). Fitting East Antarctica and Mozambique in this The driving force for subduction rollback is the way causes overlap of West Antarctica with Patagonia negative buoyancy of a sinking slab (Karig, 1974; and the Falkland Plateau. Solutions to this problem are Malinverno and Ryan, 1986; Royden, 1993; Rosenbaum provided by the recognition of microplates in Antarctica et al., 2002). Induced mantle flow in the mantle wedge and southern South America (Section 5.1 below). above the subducting slab is seen as the mechanism driving backarc seafloor spreading. In earlier two- 5.1. Microplates in Antarctica and South America, and dimensional models, such induced mantle flow was constraints on rotation timing related to flux under the tip of the subducting slab (poloidal flow) (e.g. Garfunkel et al., 1986). Three- Exotic geology (Schopf, 1969) and paleomagnetic dimensional studies have emphasised flow around the evidence (Watts and Bramall, 1981; Grunow et al., 1987) lateral edges of slabs (toroidal flow) (Kincaid and led to the recognition of the Ellsworth Whitmore Terrane Griffiths, 2003; Schellart, 2004), although with different as a displaced block within West Antarctica (Fig. 1). A 90° relative importances (Funiciello et al., 2003). Both counterclockwise rotation occurred after Permo-Triassic analogue and numerical models have also demonstrated deformation, and before 175 Ma granitoids were that as slabs sink and rollback occurs, the hingeline emplaced, with the possibility that up to 15° of rotation adopts a curved shape, convex in the rollback direction occurred after their emplacement (Grunow et al., 1987). In (Schellart, 2004; Stegman et al., 2006; Schellart et al., the TransAntarctic Mountains' equivalent in South Africa, 2007). Curvature of the hingeline is produced by rollback- the Cape Fold Belt, isotopic ages as late as 215 Ma have induced toroidal mantle flow, particularly within 800 km been returned (Halbich et al., 1983), possibly further of slab edges. With a curved hingeline, the slab sinking constraining the timing of microplate rotation. force would no longer be planar, but would rather produce Similarly, distinctive geology (Adie, 1952)and a pair of opposite rotational torques. In concert with paleomagnetic evidence (Mitchell et al., 1986; Taylor opposite toroidal flow cells, this is the likely driving and Shaw, 1989) point to a 180° clockwise rotation of the mechanism for double-saloon-door rifting and seafloor Falkland Island Block to a paleoposition aligned with the spreading. Cape Fold Belt off the Transkei coast of South Africa This scheme is similar to the piercing point concept (Fig. 1). In fact 160° of rotation is sufficient to achieve this (Marshak, 1988; Mann et al., 2002; Wallace et al., 2005), paleoposition (Storey et al., 1999), of which 54.6° is whereby buoyant material impedes subduction at one accounted for by the well constrained opening of the point on the arc/subduction zone (equivalent of the South Atlantic Ocean. The earliest undisturbed oceanic rotation pole in Figs. 2, 8 and 9), while rollback continues crust abutting rifted continental crust in the South Atlantic elsewhere, and the combination exerts a rotational torque. is M10 (Martin, 1984) or Early Hauterivian — 134 Ma on Where the two toroidal flow cells meet near the centre the timescale of Gradstein et al. (2004). of the mantle wedge above the subducting slab, an upward Assuming the Falkland Island Block's pre-breakup flow occurs orthogonal to the curved subduction hinge paleoposition (pre-190 Ma — Mussett and Taylor, (Schellart, 2004; Stegman et al., 2006). This convection 1994) lies adjacent to the Maurice Ewing Bank (Ben- cell may be related to the seafloor spreading centre which Avraham et al., 1993; Marshall, 1994; Storey et al., is perpendicular to the subducting slab (Fig. 2). The 1999), then DSPD wells there further constrain rotation backarc spreading ridges may therefore be providing a timing. Earliest sediments are Callovian with an A.K. Martin / Tectonophysics 445 (2007) 245–272 257 extensive pre-Callovian sequence visible downdip on 5.2. Geological setting: retro-arc fold thrust belt, seismic (Lorenzo and Mutter, 1988), implying that magmatic arc, and subduction/accretion complex earliest movement of the Falkland Islands Block is at least pre-Callovian (pre-165 Ma). The South African Cape Fold Belt is a multiple phase If we assume that the Falkland Islands Block and Permo-Triassic deformation (Halbich et al., 1983) Ellsworth Whitmore Terrane rotate simultaneously described as a retro-arc fold/thrust belt (Cole, 1992). (applying the models of Figs. 2, 8 and 9), then the Direct evidence of a magmatic arc does not exist, but combined timing restraints noted above point to initial Permian tuffs and volcanoclastics attest a relatively Gondwana breakup between 190 and 175 Ma. nearby volcanic arc (Johnson, 1991). Ordovician – In addition, paleomagnetic work demonstates that Permian rocks on the Falkland Islands correlate well West Antarctica also comprises the West Antarctic with those in South Africa (Marshall, 1994; Macdonald Peninsula, Thurston Island, and Marie Byrd Land blocks et al., 2003) and the equivalent retro-arc foreland basin (Grunow et al., 1991; Grunow, 1993a, b). Similarly it rocks on the Falklands also contain calc-alkaline has been suggested that southern Patagonia moved volcanoclastics (Scasso and Mendia, 1985; Curtis and independently of northern South America along the Hyam, 1998). dextral strike-slip Gastre Fault (Rapela and Pankhurst, The Ellsworth Whitmore Terrane's paleoposition 1992; Ben-Avraham et al., 1993; Marshall, 1994). (Fig. 1) is justified because its geology is considered

Fig. 10. Phase I Gondwana breakup (190 – 175 Ma)=Pliensbachian – Toarcian. Constraints on rotation timing in Section 5.1. Abbreviations of microplates as in Fig. 1. The combined Falkland Islands/Southern Patagonia/West Antarctic Peninsula Block has rotated clockwise about notional rotation pole P1. The combined Ellsworth Whitmore Terrane/Thurston Island Block is rotated counter clockwise about P2. Subduction rollback towards the Pacific accommodates the rotations which occur in a backarc location. Propagating rifts are shown between the Falkland Islands and Africa/Maurice Ewing Bank, and between the Ellsworth Whitmore Terrane and East Antarctica. The third rift is not well constrained but has been tentatively linked to the Evans Ice Shelf Graben between West Antarctic Peninsula and Thurston Island and the Central Graben in the Weddell Sea (Section 5.3.1). During Phases I and II (Figs. 10 and 11), these rifts probably did not reach the oceanic accretion phase (compare Fig. 2a and b). However, if they did, they would have formed an RRR triple junction. 258 A.K. Martin / Tectonophysics 445 (2007) 245–272 intermediate between the Western Cape of South Africa Antarctica provides the most likely source for the above- and East Antarctica (Curtis and Storey, 1996; Storey et al., mentioned tuffs and volcanoclastics extending from the 1996; Curtis, 2001). Late Carboniferous to Permian Cape Fold Belt to East Antarctica (Pankhurst et al., sequences correlate well with their equivalents in 1998). Reconstructions of the type shown in Fig. 1 bring Antarctica, and both are deformed by a Permo-Triassic the magmatic arc of southern Patagonia and West orogeny. The Permian Polarstar Formation of the Ells- Antarctica to within 500 km of the Cape Fold Belt worth Whitmore Terrane also includes calc-alkaline (Martin, 1986; Dalziel and Grunow, 1992), rather volcanoclastics derived from a volcanic arc (Collinson than being over 1200 km distant (Lock, 1980; Cole, et al., 1992). 1992). U–Pb zircon dating (Pankhurst et al., 1998; Millar In West Antarctica, three domains have been et al., 2002; Hervé and Fanning, 2001, 2003) shows that recognised. The Easterly Domain includes continental Patagonia and West Antarctica have been the sites of rift to backarc basin sequences associated with 183 – intermittent magmatic arc volcanism from the Silurian 156 Ma calc-alkaline silicic volcanic rocks (Hathway, to the Recent. Permian calc-alkaline magmatism in West 2000; Vaughan and Storey, 2000; Riley et al., 2001;

Fig. 11. Phase II Gondwana breakup (175 – 165 Ma)=Aalenian–Bathonian. The combined Falkland Island/Southern Patagonia/West Antarctic Peninsula Block has rotated further clockwise. Rapid rotation of Ellsworth Whitmore Terrane is schematically shown, whereas its final position docked against East Antarctica is shown in Fig. 12. Docking was achieved at the end of Phase I or early in Phase II. Thurston Island also rotates counter clockwise but continues rotating after Ellsworth Whitmore Terrane docking. Rifted basins are shown in the combined Falkland Plateau Basin (FPB) and Outeniqua Basin (OB) of South Africa. FG and CG=Filchner and Central Graben in the Weddell Sea between East Antarctica and Ellsworth Whitmore Terrane. Additional shaded areas represent graben between the Filchner Graben and the Ellsworth Whitmore Terrane demonstrated by gravity and magnetic data (Studinger and Miller, 1999; Ferris et al., 2000; Golynsky et al., 2001). SJB, RVB and MB=San Jorge, Rocas Verdes and Malvinas Basins in Patagonia. Rifted basins in West Antarctic Peninsula relate to the Latady and Botany Bay Formations onshore and the Black Coast Basin offshore. Aligning West Antarctic Peninsula with Southern Patagonia (cf Storey et al., 1999) rather than overlapping them, would obviate the broad gap between the West Antarctic Peninsula and Thurston Island (as discussed in Section 7.2). A.K. Martin / Tectonophysics 445 (2007) 245–272 259

Hunter et al., 2005). A Central Domain consists of a for Early Jurassic to Early Cretaceous Gondwana magmatic arc showing typical island arc geochemistry, breakup is proposed in which three plates always whereas the Western Domain comprises a forearc and separate about a single triple junction in the Weddell subduction accretion complex (MacDonald and Butter- Sea area (Figs. 10 11 12 13 and 14). worth, 1990; Vaughan and Storey, 2000). These three domains have been correlated with similar tripartite 5.3.1. Phase I: initial breakup 190–175 Ma zonations all along the Pacific Gondwana margin from Initial clockwise and counterclockwise movement of New Zealand to Southern Patagonia (Vaughan and the Falkland Islands Block and the Ellsworth Whitmore Storey, 2000). Terrane began between 190 and 175 Ma (Section 5.1). Paleomagnetic work (Grunow, 1993a, b) shows that the 5.3. Reconstruction of backarc rifting and seafloor West Antarctic Peninsula rotated clockwise between spreading during Gondwana breakup 175 and 155 Ma. In detail, the 175 Ma pole is based on rocks dated 178 and 174 Ma, and here it is assumed that Guided by the double-saloon-door model, while this motion began in the 190 – 175 Ma time period. honouring available data, an amended dispersal model Further, southern Patagonia moved right laterally and

Fig. 12. Phase III Gondwana breakup (165 – 150 Ma)=Callovian – Kimmeridgian, magnetic anomalies M39 – M22. East Antarctica (with India and Madagascar attached) has begun to move south southeast relative to Africa as the earliest oceanic crust is emplaced (see Section 5.3.3 and references therein). This spreading ridge, shown by double lines offset by fracture zones, replaces the Filchner Graben which is shown as a stalled rift, commensurate with docking of EWT. The arrow denoting the rift in the Weddell Sea propagating north northeast (in the African reference frame of Fig. 12) is superimposed on the earliest oceanic crust identified – anomaly M29. Key magnetic anomalies are shown in Fig. 14, and in the Weddell Sea, their configuration demonstrates a propagating rift. Paleomagnetic data (Grunow, 1993a,b) show that WAP and TI rotate counter clockwise relative to East Antarctica, whereas the combined PAT/Falkland Islands Block continues to rotate clockwise. This implies the separation of WAP and PAT, but it is not known whether this occurred at a spreading ridge (double lines with question marks). Continued extension in the combined FPB and OB is shown as a propagating rift. If this rift reached the oceanic accretion stage, then spreading ridges would have met in the Weddell Sea at an RRT triple junction. 260 A.K. Martin / Tectonophysics 445 (2007) 245–272 northwards relative to northern Patagonia along the In a complementary scheme, it is here suggested that Gastre and sub-parallel dextral transpressional faults part of the 230 Ma – 130 Ma counter clockwise rotation of (Forsythe et al., 1987; Rapela and Pankhurst, 1992; Thurston Island (Grunow et al., 1991)occurredwhileit Ben-Avraham et al., 1993; Marshall, 1994) which were formed a combined block with the Ellsworth Whitmore active in Late Triassic to Late Jurassic times (Martin Terrane. Any shortening required to accommodate the et al., 1999). The above points suggest that rotation rotation would then take place at the accretionary wedge/ involved a composite block comprising southern subduction zone associated with the magmatic arc on the Patagonia, West Antarctic Peninsula and the Falkland Pacific margin of Thurston Island. Islands Block. Any shortening required to accommodate Sinistral strike-slip movement was invoked to the rotation occurred at the accretionary wedge/subduc- explain the final paleoposition of the Ellsworth tion zone on the western flank of southern Patagonia and Whitmore Terrane relative to East Antarctica (Grunow the West Antarctic Peninsula. This scenario obviates the et al., 1991; Grunow, 1993a,b), although it is not clearly perceived lack of evidence (Lawrence and Johnson, 1995; evident in geophysical data (Studinger and Miller, Richards and Fannin, 1997) for docking of the Falkland 1999). The model shown here predicts rifting and Islands Block. extension in the Filchner Graben area, whereas strike-

Fig. 13. Phase IV Gondwana breakup (150 – 134 Ma)=Tithonian – Early Hauterivian, magnetic anomalies M22 – M10. The reconstruction is for M22 times. The East Antarctic/Madagascar/Indian plate is shown continuing its south southeastward motion relative to Africa, with anomaly M24 (Jokat et al., 2003) slightly landward of the spreading ridge in the Mozambique Basin. The rift opening the combined FPB/OB is shown as a stalled rift, replaced by the RVB oceanic rift which propagates northwestwards in southern Patagonia from 150 to 138 Ma (references in Section 5.3.4). In the Weddell Sea the propagating rift (implied by the geometry of spreading anomalies — Fig. 14) is superimposed on the interpolated position of M22 (after Ghidella et al., 2002). It is uncertain how this rift connects to the RVB rift, and here they are speculatively connected by a transform fault, suggesting an RTT triple junction in the Weddell Sea. A.K. Martin / Tectonophysics 445 (2007) 245–272 261

Fig. 14. Phase V Gondwana breakup (post-134 Ma), or post-magnetic anomaly M10=post-Early Hauterivian. M10 anomalies from the Argentine Basin and off the tip of the Maurice Ewing Bank (which now forms part of the Falkland Plateau — see references in Section 5.3.5) are superimposed on their equivalents in the Cape Basin in the South Atlantic and off southeast Africa. Similarly M10 anomalies off East Antarctica and Mozambique are matched. Key anomalies in the Weddell Sea after Ghidella et al. (2002) and Jokat et al. (2003). Their configuration demonstrates a rift propagating east (in an Antarctic reference frame) whereas rifts propagating north northwest in the South Atlantic and off the Maurice Ewing Bank have already been described (Martin, 1984, 1987). These rifts are offset by the Agulhas Falkland Fracture Zone, and together they replace the Rocas Verdes Basin rift which is shown stalled. The Weddell Sea propagating rift has extended close to the southeastern tip of the Maurice Ewing Bank, where it connects to an RRR triple junction. Subduction rollback was terminated by the Palmer Land event in West Antarctica 103 – 113 Ma and by the inversion of the Rocas Verdes/Magallanes Basin in Patagonia 94 Ma (Vaughan et al., 2002; Fildani and Hessler, 2005). slip is more likely either between Thurston Island and (Garrett et al., 1987; Grunow, 1993a). Extending to the East Antarctica or between Thurston Island and Marie northeast is a bouguer and free air gravity low and Byrd Land in the Pine Island area (Fig. 10). magnetic anomaly (Studinger and Miller, 1999; The third rift in the double-saloon-door model Golynsky et al., 2001), interpreted as the Central (Fig. 2) should be orthogonal to the subduction zone. Graben. Given its orientation relative to the Black In a Gondwana context, this implies a rift between West Coast Basin and Filchner Graben, this may have formed Antarctica and Thurston Island extending towards the a failed rift at a triple junction dated tentatively as Falkland Islands Block and the Ellsworth Whitmore 170 Ma (Ferris et al., 2000). The combined Evans Ice Terrane, thereby providing a mechanism for their Shelf/Central Graben may be a failed precursor of the separation (Fig. 10). Although poorly exposed, the Weddell Sea spreading ridge which did develop West Antarctic Peninsula and Thurston Island are perpendicular to the subduction zone during Phases III separated by a graben under the Evans Ice Stream to V (Sections 5.3.3–5.3.5, Figs. 12 13 and 14). 262 A.K. Martin / Tectonophysics 445 (2007) 245–272

In the pre-breakup reconstruction (Fig. 1), the 16 km — Ludwig, 1983; Lorenzo and Mutter, 1988), southern Mozambique Ridge is provisionally left in its thickens to 20 – 25 km under the offshore Outeniqua present position relative to Africa, and the Agulhas Basin (Scrutton, 1976) and reaches 30 – 40 km under its Plateau is juxtaposed against it (Martin and Hartnady, intermontane onland extensions within the Cape Fold Belt 1986). Both features have yielded continental rocks (Harvey et al., 2001). Combined with basin geometry (Tucholke et al., 1981; Mougenot et al., 1991), but also narrowing to the NW, and possibly earlier deposition in have oceanic affinities (Maia et al., 1990; Gohl and the Falkland Plateau Basin, this connotes a rift propagat- Uenzelmann-Neben, 2001). Although the Cretaceous ing towards the northwest. movement of the Falkland Plateau relative to Africa is The complementary rift propagating to the southeast is well constrained, no identified seafloor anomalies less obvious, but the Filchner Graben is a candidate constrain the paleopositions of the southern Mozambi- (Fig. 11). Here, 29 – 27 km thick continental crust with an que Ridge and the Agulhas Plateau. An alternative estimated thermal age of 230 – 165 Ma indicates a rifting Mozambique Ridge paleoposition farther north between and crustal thinning event at that time (Studinger and Mozambique and East Antarctica is implied by the Miller, 1999). Continental crust of the Filchner Graben identified extinct spreading centre in the northern Natal may obviate a tight fit of the Ellsworth Whitmore Terrane Valley from M11 to M2 time (Tikku et al., 2002). against East Antarctica (Storey et al., 1996). On the other Note that the time period of Phase I encompasses the hand, a loose fit does not align the Ellsworth Mountains eruption of the Karoo, Ferrar and the first phase of the with their equivalents in the TransAntarctic Mountains. Chon Aike large igneous provinces-LIPS (respectively The solution is a tight pre-breakup fit thereby aligning the 184 – 179 Ma, 184 – 180 Ma and 188 – 178 Ma — Permo-Triassic fold belts (Fig. 1), but with rifting in the Encarnacion et al., 1996; Duncan et al., 1997; Pankhurst Filchner Graben as the Ellsworth Whitmore Terrane et al., 2000). Backarc sedimentation had also initiated in separates from East Antarctica in the Early and Middle the southern West Antarctic Peninsula post-183 Ma Jurassic (Figs. 10 and 11). (Hunter et al., 2004, 2005). Rifting and extension was widespread during Phase II (Fig. 11). Extension and silicic magmatism continued in 5.3.2. Phase II: docking of the Ellsworth Whitmore Southern Patagonia and West Antarctica with the 172 – Terrane to earliest seafloor spreading between Africa 162 Ma second phase of the Chon Aike LIP (Pankhurst and East Antarctica: 175–165 Ma et al., 2000). Phase II times also encompass the early part Phase II extends from docking of the Ellsworth of the volcanoclastic and sometimes organic-rich Tobifera Whitmore Terrane at 175 Ma or soon thereafter FormationriftfillinPatagonia,withU–Pb ages of 178, (Grunow et al., 1987) until initial seafloor spreading 171 and 149 Ma (Pankhurst et al., 2000; Fildani and between Africa and East Antarctica approximately Hessler, 2005). Backarc rift sequences of the Latady and 165 Ma (Section 5.3.3 below). Botany Bay Formations extended throughout West As the Falkland Islands Block moved from its pre- Antarctica (Hunter et al., 2004, 2005). In this context of breakup position (Fig. 1) to its pre-South Atlantic widespread extension, the third rift in the double-saloon- opening position (Fig. 14), the Falkland Plateau Basin door scenario is schematically shown between the West formed, with its extensive pre-Callovian sequence (pre- Antarctic Peninsula and the Ellsworth Whitmore Terrane. 165 Ma — Section 5.1). The Outeniqua Basin of South There is no evidence to suggest that any of the three Africa has been correlated with the Falkland Plateau propagating rifts shown (Fig. 11) evolved beyond the Basin on the basis of their similar stratigraphies and continental crustal extension phase (compare Fig. 2a). If matching paleopositions (Martin et al., 1981). Earliest they did however, they may harbour older oceanic crust Outeniqua Basin sediments are Oxfordian (161 – 156 Ma) than the oldest seafloor spreading anomaly yet recognised and are related to the 162 Ma Zuurberg volcanics (Dingle in the Pacific (M41, 167 Ma — Sager et al., 1998). et al., 1983; Petroleum Agency SA, 2003). Note that the gap shown in Fig. 11 between the West Ben-Avraham et al. (1993) proposed a rotation pole for Antarctic Peninsula and the Ellsworth Whitmore Terrane the Falkland Islands Block on the southern margin of is broader than the present-day Central or Evans Ice South Africa. If the rotation pole was situated farther Stream Graben. This is an artefact of the reconstruction of northwest (Fig. 10), then the combined Outeniqua/ Patagonia and the West Antarctic Peninsula (cf Lawver Falkland Plateau Basin may have formed as a backarc et al., 1999) and a possible solution is given in Section 7.2 basin as the Falkland Islands Block rotated to the below. southwest (Figs.10111213and14). Crustal thickness In Fig. 11, the Ellsworth Whitmore Terrane is is least under the central Falkland Plateau Basin (12 – schematically shown rotating as it docks whereas its A.K. Martin / Tectonophysics 445 (2007) 245–272 263 final position relative to East Antarctica is shown in three plates involved were Gondwana, the combined Fig. 12. Thurston Island continued to rotate after the southern Patagonia/West Antarctic Peninsula/Falkland Ellsworth Whitmore Terrane docked (Grunow et al., Islands Block, and thirdly the combined Ellsworth 1987, 1991). The Caribbean (Fig. 7)providesan Whitmore Terrane/Thurston Island Block. In late phase analogous situation. There, Hispaniola is accreting to II and phase III, with docking of the Ellsworth Whitmore the Bahamas Platform while Puerto Rico continues to Terrane, the three plates became West Gondwana, East move with the Caribbean plate in a zone of sinistral strike- Gondwana, and the combined southern Patagonia/Falk- slip and counter clockwise rotations. Continued slab land Islands Block which continues to rotate clockwise. rollback adjacent to a blocked area caused by terrane Paleomagnetic work (Grunow, 1993a)demonstrating accretion exerts a rotational torque, thereby explaining the similar counterclockwise rotations of the West Antarctic rapid rotation of the Ellsworth Whitmore Terrane. Peninsula and Thurston Island relative to East Antarctica between 155 and 130 Ma suggests that West Antarctic 5.3.3. Phase III: early separation of East Antarctica Peninsula/Thurston Island began to separate from south- and Africa 165–150 Ma ern Patagonia/Falkland Islands in late Stage III (Fig. 12). With the docking of the Ellsworth Whitmore Terrane at The latter constitutes a fourth plate possibly separating at a the end of Phase I or during Phase II, the Filchner Graben spreading ridge. If this is so, then this ridge would have had stalled (Fig. 12) and was replaced by separation extended towards a second triple junction to the southwest of Africa and East Antarctica. Earliest magnetic anoma- in the Pacific. Alternatively, the separation of the clock- lies between Mozambique and East Antarctica (M21 – wise and counter clockwise rotating blocks may have been M24 — Bergh, 1977; Segoufin, 1978; Simpson et al., achieved via arc-parallel extension and transtensional 1979; Jokat et al., 2003) date from 147 to 152 Ma. strike-slip faults without reaching the seafloor spreading Extrapolating to the continental margin using spreading stage, as appears to have occurred in the Gibraltar, rates where there is a good model/profile match (Jokat Carpathian and Hellenic arcs (Sections 3.1, 3.3 and 3.4). et al., 2003 their Fig. 6), the earliest oceanic crust off East Note that phase III times encompass the third phase Antarctica is 163 – 168 Ma. Similar extrapolation shows of Chon Aike volcanism from 157 to 153 Ma. that earliest oceanic crust between Africa and Madagas- car/East Antarctica in the Somali Basin dates from 162 to 5.3.4. Phase IV: backarc seafloor spreading in the 168 Ma (Martin and Hartnady, 1986). Rocas Verdes Basin 150–134 Ma — a ridge jump Apart from the interpretation of Kovacs et al. (2002), towards the subduction zone E–W oriented seafloor spreading anomalies have been Phase IVextends from the earliest MORB magmatics identified in the Weddell Sea herringbone of magnetic in the Rocas Verdes Basin to the opening of the South data (LaBrecque and Barker, 1981; Livermore and Atlantic. Eruption of MORB-like pillow lavas dykes Hunter, 1996; McAdoo and Laxon, 1996) and date from and gabbros (De Wit and Stern, 1981) imply that the M25 to M29 (154–157 Ma) (Ghidella et al., 2002). This Rocas Verdes backarc basin reached the oceanic shows that the Weddell Sea rift evolved to the oceanic accretion stage (compare Fig. 2). Northward propaga- accretion stage, with anomalies orthogonal to the tion of oceanic accretion is demonstrated by younging reconstructed subduction zone (Figs.1213and14). of zircon U–Pb dates from 150 to 138 Ma (Stern and De Furthermore, anomalies M19n – M13n successively abut Wit, 2003). the continental margin off East Antarctica, demonstrating A second inference is that the rift on the east side of a propagating rift 146 – 138 Ma (Jokat et al., 2003). This the Falkland Islands Block associated with its clockwise geometry aligns exactly with the model shown in Fig. 2. rotation during phases I–III jumped to the west flank of In criticising interpretations which extend Weddell Sea the Falkland Islands Block (Fig. 13). In phase IV magnetic anomalies back to M29, and proposing that the therefore, the three plates involved are East Gondwana, earliest anomaly was M10, Martin and Hartnady (1986) West Gondwana, including Africa, South America and assumed that the third arm of the triple junction was the Falkland Islands Block, and finally the southern related to opening of the South Atlantic at M10 times. Patagonia plate west of the Rocas Verdes Basin. As However, Figs. 10 11 and 12 demonstrate that a three plate discussed in Section 5.3.3 above, continued separation system existed from initial Gondwana breakup with the of West Antarctica and southern Patagonia may have third rift separating the Falkland Islands Block from West occurred at a seafloor spreading ridge or via block Gondwana. Such a three plate scheme is compatible with rotations within the accretionary wedge/magmatic arc. the earliest Weddell Sea anomalies dating from M29 or Thirdly, as occurred in the Western Mediterranean even earlier. In phase I and the early part of phase II the (Sections 2 and 3.1 – 3.2), the seafloor spreading ridge 264 A.K. Martin / Tectonophysics 445 (2007) 245–272 jump to the Rocas Verdes Basin from east of the Bank from M10n to M8 times (Martin, 1987). South Falkland Islands Block is towards the accretionary America and the newly constituted Falkland Plateau (with wedge/subduction zone. By analogy to the Western the attached Maurice Ewing Bank) acted as a single rigid Mediterranean, this provides strong circumstantial plate. This is shown by the good match of magnetic evidence for subduction rollback towards the Pacific seafloor spreading anomalies north and south of the during Gondwana breakup. Agulhas Fracture Zone (Martin et al., 1982), which also The change in plate interactions from Phase III to negates the suggestion that the Falkland Plateau Basin Phase IV was probably driven by the docking of the continued to open in M10 – M0 times (Tucholke et al., rotating microplate. Western and northern relative 1981; Ludwig, 1983; Thomson, 1998). The configuration movement of southern Patagonia (Forsythe et al., of magnetic anomalies (Fig. 14) shows that by M10 time, 1987) and dextral transpression on the Gastre Fault the Weddell Sea rift had propagated 1650 km to the tip of from Late Triassic to Late Jurassic time (Martin et al., the Maurice Ewing Bank where it joined an RRR triple 1999) constitutes docking and collision of southern junction. This implies an overall propagation rate from Patagonia with northern Patagonia. The Gastre Fault cuts M29 to M10 of 7.17 cm/year whereas between M19 and Patagonia between 38° S and 44° S, whereas the Rocas M17 it was 6.3 cm/year (Jokat et al., 2003). For com- Verdes Basin extends only to 51° S (Stern and De Wit, parison, the rate of Pacificward migration of magmatism 2003; Fildani and Hessler, 2005). This implies that in the demonstrated by Chon Aike volcanism between 175 and docking process, the extent of the subduction rollback 160 Ma (Pankhurst et al., 2000,theirFig. 10)isbetween zone was restricted to areas south of the northern tip of 2.7 and 5.3 cm/year. If, as suggested here, this is the Rocas Verdes Basin, in alignment with the model contemporaneous with subduction rollback, then these discussed in Section 4 above (Figs. 8 and 9). rates of migration are reasonable for a subduction zone Rotating the Falkland Islands Block about the pole close to 4000 km long (Schellart et al., 2007). The shown in Fig. 10 requires about 80° rotation to reach its reconstructed subduction zone may have been shorter pre-South Atlantic opening position (Fig. 14) whereas than this given the points raised in Section 7.2 below. other data suggest a rotation of 105° (Section 5.1). The solution is provided by the model presented in 6. Driving mechanism for Gondwana breakup Figs. 8 and 9. As terranes dock and accrete to adjacent continents, terranes are broken into sub-terranes and An extensional regime backarc to the Early Jurassic late-stage poles develop. The southern Patagonia/Falk- Patagonia/West Antarctica Peninsula subduction zone land Islands Block began to break up, and was likely has been previously noted (De Wit and Stern, 1981; rotated about a late-stage pole. This process led to the Storey et al., 1992; Pankhurst et al., 2000). Subduction final rotation of the Falklands Islands Block, and rollback was proposed to accommodate rotation of the culminated in the creation of the sub-terrane of southern Falkland Islands Block (Rapela and Pankhurst, 1992; Patagonia west of the Rocas Verdes Basin (Fig. 13). Ben-Avraham et al., 1993), while Pacificward migration of 188 – 153 Ma Chon Aike magmatism has been 5.3.5. Phase V: South Atlantic opening from 134 Ma documented (Pankhurst et al., 2000: Riley et al., 2001). onwards An “indenter” has been invoked as the cause of Initial seafloor spreading in the South Atlantic at Gondwana breakup (Dalziel and Grunow, 1992), M10 (134 Ma) closely follows final emplacement of whereas others have emphasised the role of LIPS MORB magmatism in the Rocas Verdes Basin at (White and McKenzie, 1989; Cox, 1992; Storey, 1995; 138 Ma. Rollback ceased in Patagonia as the backarc Dalziel et al., 2000). None of these theories provide Rocas Verdes/Magallanes Basin was inverted and mechanisms for opposite rotations of the Falkland converted to a retro-arc foreland basin between 138 Islands Block and the Ellsworth Whitmore Terrane, and 94 Ma (Stern and De Wit, 2003; Fildani and Hessler, nor explain their separation in a northwest southeast 2005). Similarly, in the West Antarctic Peninsula, any direction, nor reveal why they accreted to different potential backarc extension was brought to an end in the continents. Palmer Land Event dated at 113 – 103 Ma (Vaughan Although initial Gondwana breakup occurred when et al., 2002). major LIP-related volcanics erupted, plume-related During initial separation of Africa and South America mechanisms do not explain the details of the dispersal (Fig. 14), a rift propagated north northwest in the South process. In contrast, the double-saloon-door postulate Atlantic from M10 times (Martin, 1984). Similarly, a rift affords an agent which drives opposite rotations of the propagated northwest along the tip of the Maurice Ewing Falkland Islands Block and the Ellsworth Whitmore A.K. Martin / Tectonophysics 445 (2007) 245–272 265

Terrane. It furnishes a cogent argument for their dominant driving mechanism for Gondwana breakup. separation in a northwest – southeast direction, and The implication is that in addition to being the agent their eventual accretion respectively to West Gondwana which forms continental crust, subduction, arc volcanism, (eventually South America) and East Gondwana and the related double-saloon-door mechanism is the (eventually East Antarctica). A number of features of engine which breaks up continents too. Early Jurassic to Early Cretaceous Gondwana dispersal duplicate processes demonstrated in other basins where 7. Discussion double-saloon-door rifting and seafloor spreading occurs (Sections 2 and 3). These include: 7.1. Gastre Fault

i) a pair of oppositely rotating microplates — the The Gastre Fault (Rapela and Pankhurst, 1992) and clockwise rotating Falkland Islands Block, and faults on the north flank of the North Patagonia Massif the counterclockwise rotating Ellsworth Whit- (Ben-Avraham et al., 1993) have been proposed as the more Terrane; main dextral strike-slip zone accommodating western ii) microplates comprising parts of a pre-existing rotation of southern Patagonia and the Falkland Islands retro-arc fold thrust belt, attached to an accretion- Block. By analogy to the Gibraltar and Calabrian arcs ary wedge/magmatic arc (the combined southern (Sections 3.1 and 3.2), no single through-going fault Patagonia/West Antarctic Peninsula/Falkland need exist. Rather, strain is likely partitioned between Islands Block and the combined Ellsworth strike-slip and offset thrust faults as occurs on Whitmore Terrane/Thurston Island Block); Hispaniola Island in the Caribbean (Section 3.5). iii) paleogeographies evincing rifting and widespread Furthermore, there is no need for the Late Triassic – extension in a backarc environment relative to an Late Jurassic Gastre Fault to be directly related to the accretionary wedge/subduction zone; Agulhas Falkland Fracture Zone which was active later iv) subduction rollback implied by a ridge jump during South Atlantic opening. Observed strike-slip towards the subduction zone, from east of the related features from the Valanginian in the Outeniqua Falkland Islands to the Rocas Verdes Basin; Basin are therefore better related to movement on the Pacificward migration of magmatic activity has Agulhas Fracture Zone during initial opening of the already been documented (Pankhurst et al., 2000; South Atlantic (Malan et al., 1990), rather than invoking Riley et al., 2001); rotation of the Falkland Islands Block microplate in v) isolation of the smaller Falkland Islands microplate post-Valanginian times (Thomson, 1998). via backarc extension and the ridge-jump noted in iv) above; this process is analogous to the separation 7.2. Extent of Patagonia and West Antarctica in of “boudins” (sensu Gueguen et al., 1997)orareas Gondwana reconstructions of thicker continental crust demonstrated by the multi-stage development of the Western Mediterra- Extension of the Falkland Plateau Basin, as well as the nean (Sections 2 and 3.1 –3.2); Magallanes and Malvinas Basins justifies the eastern pre- vi) well-documented E–W oriented seafloor spreading breakup paleoposition of southern Patagonia (Fig. 1). anomalies in the Weddell Sea (LaBrecque and However, no such adjustment in a north – south direction Barker, 1981; Livermore and Hunter, 1996; has been considered. North–south extension occurred in Ghidella et al., 2002; Jokat et al., 2003) which are the San Jorge Basin (Figari et al., 1999; Sylwan, 2001) perpendicular to the subduction zone and which and extension in the Rocas Verdes/Magallanes and propagate in the opposite direction to subduction Malvinas Basins had a north – south component. Arc- rollback; parallel extension has been documented in the Hellenic vii) the documented involvement of the Gastre Fault Arc (Armijo et al., 1992), and appears to be a basic feature and sub-parallel dextral strike-slip faults and the of a majority of the world's arcs (McCaffrey, 1996). inferred (although not proven) sinistral strike-slip Southern Patagonia and the West Antarctic Peninsula movement between the combined Thurston Is- therefore likely suffered north–south oriented extension land/Ellsworth Whitmore Terrane and Marie Byrd during their convex to the west Triassic to Early Land/East Antarctica. Cretaceous arc history (Figs.1,1,10111213and14). Similarly, during the convex to the east Tertiary history of These characteristics present a strong case for double- the Scotia Arc, southern Patagonia and the northern West saloon-door rifting and seafloor spreading being the Antarctic Peninsula suffered extension which included a 266 A.K. Martin / Tectonophysics 445 (2007) 245–272 north–south component (Barker et al., 2003). If southern thereby address the geometry of Jurassic/Cretaceous Patagonia and the West Antarctic Peninsula were smaller subduction in southwest Gondwana? during the Early Jurassic, then they could be reconstructed Is there any evidence to suggest that subduction with their magmatic arcs aligned (cf Storey et al., 1999) transform edge propagators (STEPS) formed close rather than overlapping (e.g. Lawver et al., 1999). Such a to significant strike-slip zones such as the Gastre Fault? scenario would simplify the pre-breakup and subsequent Would a re-appraisal of paleomagnetic data (Grunow, reconstructions and obviate the broad gap between West 1993a), calibrated against a different reconstruction of Antarctic Peninsula and Ellsworth Whitmore Terrane in Gondwana, influence the conclusions regarding microplate Figs. 10 and 11. rotations in southern South America and West Antarctica? Can new paleomagnetic studies, particularly in 7.3. Potential tests for the proposed process Patagonia and West Antarctica, delineate all the terranes and sub-terranes involved in early and late-stage rotations? The Western Mediterranean, Gibraltar, Tyrrhenian, If the Filchner Graben reached the oceanic accretion Aegean, Pannonian and Caribbean examples cited in stage at its northern extent, would E–W magnetic Sections 2 and 3 provide analogies which suggest a profiles in the southern Weddell Sea reveal any series of tests to the theory. associated N – S oriented seafloor spreading anomalies? Does extension within backarc basins occur on low- Similarly, can any Mesozoic oceanic crust be identified angle extensional detachments which may have acted as where the Falkland Plateau Basin collides with the thrust faults in a preceding fold/thrust belt episode? Scotia Arc? Such detachments have been interpreted in the Out- Can analogue or numerical models throw more light eniqua Basin (Malan et al., 1990; Thomson, 1999). on the double-saloon-door process? Analogue models of Multiple sub-graben of the Outeniqua Basin with curved backarc rifting include an asymmetric mode where a boundary faults may be analogous to the Gulfs of single rotation with an associated strike-slip fault is Corinth Argolis and Laconia in the Aegean where modelled (Schellart et al., 2002). The geometry of curved faults have been related to the effect of block double-saloon-door rifting could be modelled by two rotations on earlier formed faults (Van Hinsbergen et al., adjacent mirror image asymmetric rotations with a 2006). centrally located strike-slip fault (compare Fig. 2 here Do extensional detachments expose HPLT rocks with Schellart et al., 2002 their Fig. 5c). which formed in an earlier compressional orogen? To date no HPLT rocks are associated with the Gondwanide 8. Conclusions Orogeny where metamorphism reached only greenschist grade (Marshall, 1994; Curtis, 2001; Frimmel et al., Double-saloon-door rifting and seafloor spreading 2001). However detachment footwalls have not yet been successfully describes backarc extension and spreading in sampled to test this. the Western Mediterranean during subduction rollback Is there a progression from silicic to calc-alkaline and (Martin, 2006). A review of backarc basins in the thence to alkaline volcanism? Such progressions in the Mediterranean, Carpathians, Aegean and Caribbean inti- Western Mediterranean and Pannonian Basins are mates that this process is widespread, and points to a two- related to sub-crustal heating and extension, subduction, stage model. The first stage is based on the Chattian– eventual slab detachment and asthenospheric upwelling Burdigalian development of the Valencia Trough and the (e.g. Konecny et al., 2002). The sequence of events Liguro-Provencal Basin, as well as the Burdigalian– outlined in Sections 5.3.1–5.3.5 suggests bimodal Langhian opening of the Liguro-Provencal and Algerian volcanism in southern West Antarctica may date from Basins. The second stage of the model is based on the 175 Ma, if docking of the Ellsworth Whitmore Terrane Alboran, Tyrrhenian, Pannonian, Aegean and Caribbean involved slab detachment. Any bimodal volcanism examples. In the first stage, two microplates comprising would be younger further north and in Patagonia. retro-arc fold thrust belts and magmatic arcs simultaneous- Further study of Zuurberg volcanics in South Africa, ly rotate clockwise and counterclockwise about rotation Sweeney Formation basalts, Hjort Formation mafics and poles. As they accrete to adjacent buoyant continental crust, ultramafics in West Antarctica, and alkali basalts on the extent of the subduction rollback zone is restricted, Alexander Island may be revealing. thereby initiating strike-slip to subduction transitions. The Can teleseismic tomography image subducted oce- restricted rollback zone is bounded by a pair of dextral and anic crust well over 100 Ma old, which by now may sinistral strike-slip zones which respectively foment further have pierced or be ponded at the 660 km transition, and clockwise and counterclockwise rotations about late-stage A.K. Martin / Tectonophysics 445 (2007) 245–272 267 rotation poles. The probable geodynamic mechanism is a References pair of rotational torques induced by slab sinking and rollback of a curved subduction hingeline. This scenario is Adie, R.J., 1952. The position of the Falkland Islands in a – used to develop a model for Early Jurassic to Early reconstruction of Gondwanaland. Geol. Mag. 89, 401 410. Alvarez-de-Buergo, E., Melendez-Hevia, F., 1994. Características Cretaceous Gondwana breakup, comprising five stages in generales de las subcuencas del margen peninsular mediterráneo which dispersal always involves three plates separating (Rift del Surco de Valencia). Acta Geol. Hisp. 29 (1), 67–79. about a single triple rift or triple junction within the Weddell Angelier, J., Lyberis, N., Le Pichon, X., Barrier, E., Huchon, P., 1982. Sea. Seven criteria similar to those observed in other The tectonic development of the Hellenic arc and the Sea of Crete: – backarc basins are considered characteristic of double- a synthesis. Tectonophysics 86, 159 196. Armijo, R., Lyon-Caen, H., Papanastassiou, D., 1992. East–west saloon-door rifting and seafloor spreading: extension and Holocene normal-fault scarps in the Hellenic arc. Geology 20, 491–494. i) initial breakup involves a pair of oppositely Armijo, R., Meyer, B., King, G.C.P., Rigo, A., Papanastassiou, D., rotated microplates — the Falkland Islands 1996. Quarternary evolution of the Corinth Rift and its implica- Block and the Ellsworth Whitmore Terrane; tions for the later Cenozoic evolution of the Aegean. Geophys. J. Int. 126, 11–53. ii) the microplates comprise parts of a pre-existing Ave Lallement, H.G., 1997. Transpression, displacement partitioning, retro-arc fold thrust belt (the Permo-Triassic and exhumation in the eastern Caribbean/South American plate Gondwanide Orogeny), attached to a nearby boundary zone. Tectonics 16, 272–289. magmatic arc; the Falkland Island Block was Ave Lallement, H.G., Guth, L.R., 1990. Role of extensional tec- initially attached to Southern Patagonia and the tonics in exhumation of eclogites and blueschists in an oblique subduction setting: Northeastern Venezuela. Geology 18, West Antarctic Peninsula, whereas The Ellsworth 950–953. Whitmore Terrane was combined with the Thur- Barker, D.H.N., Christeson, G.L., Austin, J.A., Dalziel, I.W.D., 2003. ston Island Block; Backarc basin evolution and cordilleran orogenesis: insights from iii) these rotations are associated with widespread new ocean-bottom seismograph refraction profiling in Bransfield – extension in a backarc environment and are Strait, Antarctica. Geology 31 (2), 107 110. Ben-Avraham, Z., Hartnady, C.J.H., Malan, J.A., 1993. Early tectonic accommodated by rollback of a contemporary extension between the Agulhas Bank and the Falkland Plateau due to subduction zone, and shortening at the associated the rotation of the Lafonia microplate. Earth Planet. Sci. Lett. 117, accretionary wedge; 43–58. iv) a ridge jump towards the subduction zone from Benetatos, C., Kiratzi, A., Papazachos, C., Karakaisis, G., 2004. Focal east of the Falkland Islands to the Rocas Verdes mechanisms of shallow and intermediate depth earthquakes along the Hellenic Arc. J. Geodyn. 37, 253–296. Basin demonstrates subduction rollback; Bergh, H.W., 1977. Mesozoic sea floor off Dronning Maud land, v) this ridge jump and/or backarc extension separat- Antarctica. Nature 269, 686–687. ed off the Falkland Islands Block from southern Bilich, A., Frohlich, C., Mann, P., 2001. Global seismicity characteristics Patagonia; this process is similar to that occurring of subduction-to-strike-slip transitions. J. Geophys. Res. (Solid – in the Western Mediterranean where areas of Earth) 106, 19443 19452. Bouillin, J.P., Durand-Delga, M., Olivier, P., 1986. Betic, Rifian and continental crust were isolated during episodic Tyrrhenian Arc: Distinctive features, genesis and development subduction rollback; stages. In: Wezel, F.C. (Ed.), The Origin of Arcs. Elsevier, New vi) well-documented EW oriented seafloor spreading York, pp. 281–304. anomalies in the Weddell Sea are perpendicular to Brunet, C., Monié, P., Jolivet, L., Cadet, J.P., 2000. Migration of the subduction zone and propagate in the opposite compression and extension in the Tyrrhenian Sea, insights from 40Ar/39Ar ages on micas along a transect from Corsica to Tuscany. direction to rollback; Tectonophysics 321, 127–155. vii) the rollback zone is bordered by the documented Burmester, R.F., Beck, M.E., Speed, R.C., Snoke, C.W., 1996. A dextral Gastre and sub-parallel faults in Patagonia, preliminary paleomagnetic pole for mid-cretaceous rocks from and by inferred sinistral strike-slip between the Tobago: further evidence for large clockwise rotations in the – combined Thurston Island/Ellsworth Whitmore Caribbean South American plate boundary zone. Earth Planet. Sci. Lett. 139, 79–90. Terrane and Marie Byrd Land/East Antarctica in Chalot-Prat, F., Girbacea, R., 2000. Partial delamination of continental the Pine Island Bay area. mantle lithosphere, uplift-related crust–mantle decoupling, volcanism and basin formation: a new model for the Pliocene–Quater- Acknowledgements nary evolution of the southern East-Carpathians, Romania. Tectonophysics 327, 83–107. Cole, D.J., 1992. Evolution and development of the Karoo Basin. In: I thank Angel Fernandez Lozano for his help with the De Wit, M.J., Ransome, I.G.D. (Eds.), Inversion Tectonics of the artwork. Two anonymous reviews helped improve the Cape Fold Belt, Karoo and Cretaceous Basins of Southern Africa. manuscript. Balkema, Rotterdam, pp. 109–115. 268 A.K. Martin / Tectonophysics 445 (2007) 245–272

Collinson, J.W., Vavra, C.L., Zawiskie, J.M., 1992. Sedimentology of Proceedings of the Ocean Drilling Program, Scientific Results, the Polarstar Formation (Permian), Ellsworth Mountains, West College Station TX, vol. 107, pp. 93–97. Antarctica. In: Webers, G.F., Craddock, C., Splettstoesser, J.F. Ferrandini, J., Gattacceca, J., Ferrandini, M., Deino, A., Janin, M.-J., (Eds.), Geology and Paleontology of the Ellsworth Mountains, 2003. Chronostratigraphie et paleomagnetisme des depots oligo- West Antarctica, Geol. Soc. Amer. Mem., vol. 170, pp. 63–79. miocenes de Corse: Implications geodynamiques pour l'ouverture Comas, M.C., Platt, J.P., Soto, J.I., Watts, A.B., 1999. The origin and du bassin liguro-provencal. Bull. Soc. Geol. Fr. 174, 357–371. tectonic history of the Alboran Basin: insights from ODP Leg 161 Ferris, J.K., Vaughan, A.P.M., Storey, B.C., 2000. Relics of a complex results. Proc. Ocean Drill. Program Sci. Results 161, 555–580. triple junction in the Weddell Sea embayment, Antarctica. Earth Cox, K.G., 1992. Karoo igneous activity, and the early stages of the Planet. Sci. Lett. 178, 215–230. break-up of Gondwanaland. In: Storey, B.C., Alabaster, T., Figari, E.G., Strelkov, E., Lafitte, G., Cid de la Paz, M.S., Courtade, S. Pankhurst, R.J. (Eds.), Magmatism and the Causes of Continental F., Celaya, J., Vottero, A., LaFourcade, P., Martinez, R., Villar, H., Break-up, Geol. Soc. London Spec. Publ., vol. 68, pp. 137–148. 1999. Los sistemas petroleros de la Cuenca del Golfo San Jorge: Crespo-Blanc, A., Campos, J., 2001. Structure and kinematics of the Sintesis estructural, estratigrafia y geoquimica, Actas IV Congreso South Iberian paleomargin and its relationship with the Flysch de Exploracion y Desarrollo de hidrocarburos. Mar Del Plata I Trough units: extensional tectonics within the Gibraltar Arc fold 197–237. and thrust belt (western Betics). J. Struct. Geol. 23, 1615–1630. Fildani, A., Hessler, A.M., 2005. Stratigraphic record across a retro-arc Curtis, M.L., 2001. Tectonic history of the Ellsworth Mountains, West basin inversion: Rocas Verdes-Magallanes Basin, Patagonian Antarctica: reconciling a Gondwana enigma. Geol. Soc. Amer. Andes, Chile. Geol. Soc. Amer. Bull. 117, 1596–1614. Bull. 113 (7), 939–958. Forsythe, R.D., Kent, D.V., Mpodozis, C., Davidson, J., 1987. Curtis, M.L., Hyam, D.M., 1998. Late Paleozoic to Mesozoic Paleomagnetism of Permian and Triassic rocks in the central structural evolution of the Falkland Islands: a displaced segment Chilean Andes. In: McKenzie, G.D. (Ed.), Gondwana Six: of the Cape Fold Belt. J. Geol. Soc. ( Lond. ) 155, 115–129. Structure, Tectonics and Geophysics, Geophysical Monograph, Curtis, M.L., Storey, B.C., 1996. A review of the geological vol. 40, pp. 241–252. constraints on the pre-Gondwana break-up position of the Frimmel, H.E., Folling, P.G., Diamond, R., 2001. Metamorphism of Ellsworth Mountains: implications for Weddell Sea evolution. In: the Permo-Triasssic Cape Fold Belt and its basement, South Storey, B.C., King, E.C., Livermore, R.A. (Eds.), Weddell Sea Africa. Mineral. Petrol. 73, 325–346. Tectonics and Gondwana Break-up, Geol. Soc. London Spec. Funiciello, F., Faccenna, C., Giardini, D., Regenauer-Lieb, K., 2003. Publ., vol. 108, pp. 11–30. Dynamics of retreating slabs: 2. Insights from three-dimensional Dalziel, I.W.D., Grunow, A.M., 1992. Late Gondwanide tectonic laboratory experiments. J. Geophys. Res. 108 (B4), 2207. rotations within Gondwanaland. Tectonics 11, 603–606. doi:10.1029/2001JB000896. Dalziel, I.W.D., Lawver, L.A., Murphy, J.B., 2000. Plumes, orogenesis Garfunkel, Z., Anderson, C.A., Schubert, G., 1986. Mantle circulation and supercontinental fragmentation. Earth Planet. Sci. Lett. 178, and the lateral migration of subducted slabs. J. Geophys. Res. 91, 1–11. 7205–7223. De Ritis, R., Blanco-Montenegro, I., Ventura, G., Chiappini, M., 2005. Garrett, S.W., Herrod, L.D.B., Mantripp, D.R., 1987. Crustal structure Aeromagnetic data provide new insights on the volcanism and tectonics of the area around Haag Nunataks, West Antarctica: new of Vulcano Island and offshore areas (southern Tyrrhenian Sea, Italy). areomagnetic and bedrock elevation data. In: McKenzie, G.D. Geophys. Res. Lett. 32, L15305. doi:10.1029/2005GL023465. (Ed.), Gondwana Six: Structure, Tectonics and Geophysics, De Wit, M.J., Stern, C.R., 1981. Variations in the degree of crustal Geophysical Monograph, vol. 40, pp. 109–115. extension during formation of a backarc basin. Tectonophysics 72, Gattacceca, J., Speranza, F., 2002. Paleomagnetism of Jurassic to 229–260. Miocene sediments from the Apenninic carbonate platform Dingle, R.V., Siesser, W.G., Newton, A.R., 1983. Mesozoic and (southern Apennines, Italy): evidence for a 60° counterclockwise Tertiary Geology of Southern Africa. Balkema, Rotterdam. 375 pp. Miocene rotation. Earth Planet. Sci. Lett. 201, 19–34. Duermeijer, C.E., Nyst, M., Meijer, P.T., Langereis, C.G., Spakman, Gessner, K., Ring, U., Johnson, C., Hetzel, R., Passchier, C.W., W., 2000. Neogene evolution of the Aegean arc: Paleomagnetic Gungo, T., 2001. An active bivergent rolling hinge detachment and geodetic evidence for a rapid and young rotation phase. Earth system: Central Menderes metamorphic core complex in Western Planet. Sci. Lett. 17, 509–525. Turkey. Geology 29 (7), 611–614. Duggen, S., Hoernle, K., van den Bogaard, P., Harris, C., 2004. Ghidella, M., Yanez, G., LaBrecque, J.L., 2002. Revised tectonic Magmatic evolution of the Alboran region: the role of subduction implications for the magnetic anomalies of the western Weddell in forming the western Mediterranean and causing the Messinian Sea. Tectonophysics 347, 65–86. Salinity Crisis. Earth Planet. Sci. Lett. 218, 91–108. Gohl, K., Uenzelmann-Neben, G., 2001. The crustal role of the Agulhas Duncan, R.A., Hooper, P.R., Rehacek, J., Marsh, J.S., Duncan, A.R., Plateau, southwest Indian Ocean: evidence from seismic profiling. 1997. The timing and duration of the Karoo igneous event, Geophys. J. Int. 144 (3), 632. doi:10.1046/j.1365-246x.2001.01368.x. southern Gondwana. J. Geophys. Res. 102, 18127–18138. Golynsky, A., Chiappini, M., Damaske, D., Ferraccioli, F., Ferris, J., Encarnacion, J., Flemming, T.H., Elliot, D.H., Eales, H.V., 1996. Finn, C., Ghidella, M., Isihara, T., Johnson, A., Kim, H.R., Synchronous emplacement of Ferrar and Karoo dolerites and the Kovacs, L., LaBrecque, J., Masolov, V., Nogi, Y., Purucker, M., early break up of Gondwana. Geology 24, 535–538. Taylor, P., Torta, M., 2001. ADMAP — Magnetic Anomaly Map Faccenna,C.,Piromalli,C.,Crespo-Blanc,A.,Jolivet,L.,Rossetti,F.,2004. of the Antarctic, 1:10000000 scale map. In: Morris, P., von Frese, Lateral slab deformation and the origin of the western Mediterranean R. (Eds.), BAS (Misc.), vol. 10. British Antarctic Survey, arcs. Tectonics 23, TC1012. doi:10.1029/2002TC001488. Cambridge. Feraud, G., 1990. 39Ar–40Ar analysis on basaltic lava series of Vavilov Govers, R., Wortel, M.J.R., 2005. Lithosphere tearing at STEP faults: Basin, Tyrrhenian Sea (Ocean Drilling Program Leg 107, holes response to edges of subduction zones. Earth Planet. Sci. Lett. 236/ 655B and 651A). In: Kastens, K., Mascle, J., et al. (Eds.), 1–2, 505–523. A.K. Martin / Tectonophysics 445 (2007) 245–272 269

Gradstein, F.M., Ogg, J.G., Smith, A.G., Agterberg, F.P., Bleeker, W., Karig, D.E., 1974. Evolution of arc systems in the Western Pacific. Cooper, R.A., Davydov, V., Gibbard, P., Hinnov, L.A., House, M.R., Ann. Rev. Earth Planet. Sci. 2, 51–78. Lourens, L., Luterbacher, H.P., McArthur, J., Melchin, M.J., Kincaid, C., Griffiths, R.W., 2003. Laboratory models of the thermal Robb, L.J., Shergold, J., Villeneuve, M., Wardlaw, B.R., Ali, J., evolution of the mantle during rollback subduction. Nature 425, Brinkhuis, H., Hilgen, F.J., Hooker, J., Howarth, R.J., Knoll, A.H., 58–62. doi:10.1038/nature01923. Laskar, J., Monechi, S., Plumb, K.A., Powell, J., Raffi, I., Röhl, U., Kissel, C., Laj, C., 1988. The Tertiary geodynamical evolution of the Sadler, P., Sanfilippo, A., Schmitz, B., Shackleton, N.J., Shields, Aegean arc: a paleomagnetic reconstruction. Tectonophysics 146, G.A., Strauss, H., Van Dam, J., van Kolfschoten, T., Veizer, J., 183–201. Wilson, D., 2004. A Geologic Time Scale 2004. Cambridge Konecny, V., Kovac, M., Lexa, J., Sefara, J., 2002. Neogene evolution University Press. 589 pp. of the Carpatho-Pannonian region: an interplay of subduction and Grunow, A.M., 1993a. New Paleomagnetic data from the Antarctic backarc diapiric uprise in the mantle. EGU Stephan Mueller Peninsula and their Tectonic Implications. J. Geophys. Res. 98, Special Publication Series, European Geosciences Union, vol. 1, 13815–13833. pp. 105–123. Grunow, A.M., 1993b. The creation and destruction of Weddell Sea Kovacs, L., Morris, C.P., Brozena, J., Tikku, A., 2002. Seafloor floor in the Jurassic. Geology 21, 647–650. spreading in the Weddell Sea from magnetic and gravity data. Grunow, A.M., Kent, D.V., Dalziel, I.W.D., 1987. Mesozoic evolution Tectonophysics 347, 43–64. of West Antarctica and the Weddell Sea Basin: New paleomagnetic LaBrecque, J.L., Barker, P., 1981. The age of the Weddell Basin. constraints. Earth Planet. Sci. Lett. 86, 16–26. Nature 290, 489–492. Grunow, A.M., Kent, D.V., Dalziel, I.W.D., 1991. New paleomagnetic Lawrence, S., Johnson, M., 1995. Shelf north of the Falklands may be data from Thurston Island: implications for the tectonics of West new South Atlantic petroleum province. Oil & Gas J. 6, 52–55. Antarctica and Weddell Sea opening. J. Geophys. Res. 96 (B11), Lawver, L.A., Gahagan, L.M., Dalziel, I.W.D., 1999. A tight fit — 17935–17954. Early Mesozoic Gondwana. A plate reconstruction perspective. In: Gueguen, E., Doglioni, C., Fernandez, M., 1997. Lithospheric Motoyoshi, Y., Shiraishi, K. (Eds.), Origin and Evolution of boudinage in the western Mediterranean back-arc basin. Terra Continents: Proceedings of the International Symposium “Origin Nova 9, 184–187. and Evolution of Continents” Tokyo, Mem. National Institute Gueguen, E., Doglioni, C., Fernandez, M., 1998. On the post 25 Ma Polar Research, Special Issue, vol. 53, pp. 214–229. geodynamic evolution of the Western Mediterranean. Tectono- LeBlanc, D., Olivier, P., 1984. Role of strike-slip faults in the Betic- physics 298, 259–269. Rifian orogeny. Tectonophysics 101, 345–355. Gurer, O.F., Yilmaz, Y., 2002. Geology of the Oren and surrounding Livermore, R.A., Hunter, R.J., 1996. Mesozoic seafloor spreading in areas, SW Anatolia, Turkish. J. Earth Sci. 11, 1–13. the southern Weddell Sea. In: Storey, B.C., King, E.C., Livermore, Halbich, I.W., Fitch, F.J., Millar, J.A., 1983. Dating the Cape Orogeny. R.A. (Eds.), Weddell Sea Tectonics and Gondwana Break-Up, In: Sohnge, A.P.G., Halbich, I.W. (Eds.), Geodynamics of the Cape Geol. Soc. London Spec. Publ., 108, pp. 227–241. Fold Belt, Geol. Soc. S. Afr. Spec. Publ., vol. 12, pp. 75–100. Lock, B.E., 1980. Flat-plate subduction and the Cape Fold Belt of Harvey, J.D, De Wit, M.J., Stankiewicz, J., Moctar Doucouré, C., 2001. South Africa. Geology 8, 35–39. Structural variations of the crust in the Southwestern Cape, deduced Lorenzo, J.M., Mutter, J.C., 1988. Seismic stratigraphy and tectonic from seismic receiver functions. S. Afr. J. Geol. 104, 231–242. evolution of the Malvinas/Falkland Plateau. Rev. Bras. Geocienc. Hathway, B., 2000. Continental rift to back-arc basin: Jurassic– 18, 191–200. Cretaceous stratigraphical and structural evolution of the Larsen Lonergan, L., White, N., 1997. Origin of the Betic-Rif mountain belt. Basin, Antarctic Peninsula. J. Geol. Soc. ( Lond. ) 157, 417–432. Tectonics 16 (3), 504–522. Haxby, W.F., 1987. Gravity Field of the World's Oceans, National Ludwig, W.J., 1983. Geologic framework of the Falkland Plateau. In: Geophysical Data Centre. NOAA, Boulder, Colorado. Ludwig, W.J., Krashenninikov, V.A., et al. (Eds.), Init. Repts. Hervé, F., Fanning, C.M., 2001. Late Triassic detrital zircons in DSDP 71. U.S. Govt. Printing Office, Washington, pp. 281–293. metaturbidites of the Chonos Metamorphic Complex, southern MacDonald, D.I.M., Butterworth, P.J., 1990. The stratigraphy, setting and Chile. Rev. Geol. Chile 28 (1), 91–104. hydrocarbon potential of the Mesozoic sedimentary basins of the Hervé, F., Fanning, C.M., 2003. Early Cretaceous subduction at the Diego Antarctic Peninsula. In: St. John, B. (Ed.), Antarctica as an Exploration de Almagro archipelago, Southern Chile. Episodes 26 (4), 285–289. Frontier, Am. Assoc. Pet. Geol. Stud. Geol., vol. 31, pp. 101–125. Hetzel, R., Passchier, C.W., Ring, U., Dora, O.O., 1995. Bivergent Macdonald, D., Gomez-Perez, I., Franzese, J., Spalletti, L., Lawver, L., extension in orogenic belts: the Menderes massif (southwestern Gahagan, L., Dalziel, I., Thomas, C., Trewin, N., Hole, M., Paton, Turkey). Geology 23 (5), 455–458. D., 2003. Mesozoic break-up of SW Gondwana: implications for Hunter, M.A., Cantrill, D.J., Flowerdew, M.J., Millar, I.L., 2005. Mid- regional hydrocarbon potential of the southern South Atlantic. Jurassic age for the Botany Bay Group: implications for Weddell Mar. Pet. Geol. 20, 287–308. Sea Basin creation and southern hemisphere biostratigraphy. J. Maerten, L., Séranne, M., 1995. Extensional tectonics in the Oligo- Geol. Soc. ( Lond. ) 162, 745–748. Miocene Hérault Basin (S. France), Gulf of Lion margin. Bull. Hunter, M.A., Riley, T.R., Millar, I.L., 2004. Middle Jurassic air fall Soc. Geol. Fr. 166 (6), 739–749. tuff in the sedimentary Latady Formation, eastern Ellsworth Land. Maia, M., Diament, M., Recq, M., 1990. Isostatic response of the Antarct. Sci. 16, 185–190. lithosphere beneath the Mozambique Ridge (SW Indian Ocean) Johnson, M.R., 1991. Sandstone petrography, provenance and plate and geodynamic implications. Geophys. J. Int. 100, 337–348. tectonic setting in Gondwana context of the southeastern Cape- Maillard, A., Mauffret, A., 1999. Crustal structure and riftogenesis of Karoo Basin. S. Afr. J. Geol. 91 (2/3), 137–154. the Valencia Trough (north-western Mediterranean Sea). Basin Jokat, W., Boebel, T., Konig, M., Meyer, U., 2003. Timing and Res. 11, 357–379. geometry of early Gondwana breakup. J. Geophys. Res. 108 (B9), Malan, J., Martin, A.K., Cartwright, J., 1990. The structural and 2428. doi:10.1029/2002JB001802. stratigraphic development of the Gamtoos and Algoa Basins, 270 A.K. Martin / Tectonophysics 445 (2007) 245–272

offshore South Africa. Abstracts Geocongress '90, Geol. Soc. S. Millar, I.L., Pankhurst, R.J., Fanning, C.M., 2002. Basement chronology Afr. Cape Town 328–331. of the Antarctic Peninsula: recurrent magmatism and anatexis in the Malinverno, A., Ryan, W.B.F., 1986. Extension in the Tyrrhenian Sea Palaeozoic Gondwana Margin. J. Geol. Soc. ( Lond. ) 159 (2), and shortening in the Apennines as result of arc migration driven 145–157. by sinking of the lithosphere. Tectonics 5 (2), 227–245. Mitchell, C., Taylor, G.K., Cox, K.G., Shaw, J., 1986. Are the Falkland Mann, P., Calais, E., Ruegg, J.-C., DeMets, C., Jansma, P., Mattioli, Islands a rotated microplate? Nature 319, 131–134. G., 2002. Oblique collision in the northeastern Caribbean from Monié, P., Montigny, R., Maluski, H., 1992. Age burdigalien de la GPS measurements and geological observations. Tectonics vol. 21 tectonique ductile extensive dans le massif de l'Edough (Kabylies, (no. 6), 1057. doi:10.1029/2001TC001304. Algérie). Donnees radiometriques 39Ar–40Ar. Bull. Soc. Géol. Fr. Marshak, S., 1988. Kinematics of orocline and arc formation in thin- 5, 571–584. skinned orogens. Tectonics 7, 73–86. Mougenot, D., Gennesseaux, M., Hernandez, J., Lepvrier, C., Malod, Marshall, J.E.A., 1994. The Falkland Islands: a key element in J.-A., Raillard, S., Vanney, J.-R., Villeneuve, M., 1991. La ride du Gondwana paleogeography. Tectonics 13, 499–514. Mozambique (Ocean Indien): un fragment continental individual Marti, J., Mitjavila, J., Roca, E., Aparicio, A., 1992. Cenozoic lors du coulissement de l'Amerique et de l'Antarctique le long de magmatism of the Valencia Trough (western Mediterranean): l'Afrique de l'Est? C. R. Acad. Sci., Paris (II) 312, 655–662. relationship between structural evolution and volcanism. Tectono- Muller, R.D., Royer, J.Y., Cande, S.C., Roest, W.R., Maschenkov, S., physics 203 (1–4), 145–165. 1999. New constraints on the Late Cretaceous/Tertiary plate Martin, A.K., 1984. Propagating rifts: crustal extension during continen- tectonic evolution of the Caribbean. In: Mann, P. (Ed.), tal rifting. Tectonics 3, 611–617. Sedimentary Basins of the World, 4, Caribbean Basins. Elsevier, Martin, A.K., 1986. Microplates in Antarctica. Nature 319 (6049), New York, pp. 33–59. 100–101. Mussett, A.E., Taylor, G.K., 1994. 40Ar–39Ar ages for dykes from the Martin, A.K., 1987. Plate reorganisations around Southern Africa, hot- Falkland Islands with implications for the break-up of southern spots and extinctions. Tectonophysics 142, 309–316. Gondwanaland. J. Geol. Soc. ( Lond. ) 151 (1), 79–81. Martin, A.K., 2006. Oppositely directed pairs of propagating rifts in back- Negredo, A., Fernandez, M., Torné, M., Doglioni, C., 1999. Numerical arc basins: double saloon door seafloor spreading during subduction modelling of simultaneous extension and compression: the rollback. Tectonics 25, TC3008. doi:10.1029/2005TC001885. Valencia Trough (Western Mediterranean). Tectonics 18, 361–374. Martin, A.K., Hartnady, C.J.H., 1986. Plate tectonic development of Nicolosi, I., Speranza, F., Chiappini, M., 2006. Ultrafast oceanic the south west Indian Ocean: a revised reconstruction of East spreading of the Marsili Basin, southern Tyrrhenian Sea: evidence Antarctica and Africa. J. Geophys. Res. 91 (B5), 4767–4786. from magnetic anomaly analysis. Geology 34, 717–720. Martin, M., Ritter, J.R.R., 2005. High-resolution teleseismic body- Panaiotu, C., 1998. Paleomagnetic constrains on the geodynamic wave tomography beneath SE Romania-I. Implications for three- history of Romania. In: Sledzinski, J. (Ed.), Monograph of dimensional versus one-dimensional crustal correction strategies Southern Carpathians, Reports on Geodesy, vol. 7, pp. 205–216. with a new crustal velocity model. Geophys. J. Int. 162 (2), 448. Pankhurst, R.J., Weaver, S.D., Bradshaw, J.D., Storey, B.C., Ireland, T.R., doi:10.1111/j.1365-246X.2005.02661.x. 1998. Geochronology and geochemistry of pre-Jurassic superterranes Martin, A.K., Goodlad, S.W., Hartnady, C.J.H., 1981. A revised fit of in Marie Byrd Land, Antarctica. J. Geophys. Res. 103, 2529–2547. South America and South Central Africa. Earth Planet. Sci. Lett. Pankhurst, R.J., Riley, T.R., Fanning, C.M., Kelley, S.P., 2000. Episodic 54 (2), 293–305. silicic volcanism in Patagonia and the Antarctic Peninsula: Martin, A.K., Goodlad, S.W., Hartnady, C.J.H., du Plessis, A., 1982. chronology of magmatism associated with the break-up of Cretaceous palaeopositions of the Falkland Plateau relative to Gondwana. J. Petrol. 41 (5), 605–625. southern Africa using Mesozoic seafloor spreading anomalies. Pares, J.M., Freeman, R., Roca, E., 1992. Neogene structural Geophys. J. R. Astron. Soc. 71, 567–579. development in the Valencia Trough margins from palaeomagnetic Martin,M.,Kato,T.,Rodríguez,C.,Godoy,E.,Duhart,P., data. Tectonophysics 203 (1–4), 111–124. McDonough, M., Campos, A., 1999. Evolution of Late Paleozoic Patrascu, S., Panaiotu, C., Seclaman, M., Panaiotu, C.E., 1994. Timing of accretionary complex and overlying forearc–magmatic arc south- rotational motion of Apuseni Mountains (Romania): paleomagnetic central Chile (38–41°S): constraints for the tectonic setting data from Tertiary magmatic rocks. Tectonophysics 233, 163–176. along the southwestern margin of Gondwana. Tectonics 12 (4), Pelletier, B., Calmant, S., Pillet, R., 1998. Current tectonics of the 582–605. Tonga–New Hebrides region. Earth Planet. Sci. Lett. 164, 263–276. Martínez-Martínez, J.M., Azañón, J.M., 2002. Orthogonal extension Petroleum Agency SA, 2003. South Africa: Petroleum exploration in the hinterland of the Gibraltar Arc (Betic, SE Spain). Journal of opportunities offshore South Africa. www.petroleumagencysa. the Virtual Explorer 8, 1–20. com 2003. Marketing Brochure: 1–32. Maury, R.C., Fourcade, S., Coulon, C., El Azzouzi, M., Bellon, H., Pindell, J.L., Kennan, L.J., 2001. Exploration framework atlas series 4: Coutelle, A., Ouabadi, A., Semroud, B., Megartsi, M., Cotten, J., Mexico and Gulf of Mexico. Tectonic Analyis Ltd., Cokes Barn, Belanteur, O., Louni-Hacini, A., Piqué, A., Capdevila, R., West Sussex, England, Commercial Report. Hernandez, J., Rehault, J.-P., 2000. Post-collisional Neogene Pindell, J., Kennan, L.J., Higgs, R., Barrett, S., Villamil, T., Erikson, J., magmatism of the Mediterranean Maghreb margin: a consequence Algar, S., 1999. Exploration framework atlas series 2: Venezuela– of slab breakoff. C. R. Acad. Sci., Paris 331, 159–173. Trinidad. Tectonic Analysis Ltd., Cokes Barn, West Sussex, McAdoo, D.C., Laxon, S.W., 1996. Marine Gravity from Geosat and England. Commercial Report. ERS-1 Altimetry in the Weddell Sea. In: Storey, B.C., King, E.C., Platt, J.P., Allerton, S., Kirker, A., Mandeville, C., Mayfield, A., Livermore, R.A. (Eds.), Weddell Sea Tectonics and Gondwana Platzman, E.S., Rimi, A., 2003. The ultimate arc: differential Breakup, Geol. Soc. London Spec. Pub., vol. 108, pp. 155–164. displacement, oroclinal bending, and vertical axis rotation in McCaffrey, R., 1996. Estimates of modern arc-parallel strain rates in the External Betic-Rif arc. Tectonics 22 (3), 1017. doi:10.1029/ fore arcs. Geology 24, 27–30. 2001TC001321. A.K. Martin / Tectonophysics 445 (2007) 245–272 271

Rapela, C.W., Pankhurst, R.J., 1992. The granites of northern Schopf, J., 1969. Ellsworth Mountains: position in west Antarctica due Patagonia and the Gastre Fault system in relation to the break-up to seafloor spreading. Science 164, 63–66. of Gondwana. In: Storey, B.C., Alabaster, T., Pankhurst, R.J. Scrutton, R.A., 1976. Continental break-up and deep crustal structure at the (Eds.), Magmatism and the causes of continental break-up, Geol. margins of southern Africa. Ann. Acad. Bras. Cienc. 48 suppl, Soc. London Spec. Publ., vol. 68, pp. 209–220. 101–120. Rehault, J.-P., Boillot, G., Mauffret, A., 1984. The Western Mediterra- Segoufin, J., 1978. Anomalies magnetiques mesozoiques dans le nean basin geological evolution. Mar. Geol. 55, 447–477. bassin de Mozambique. C. R. Acad. Sci., Paris (D) 287, 109–112. Reid, J.A., Plumley, P.W., Schellekens, J.H., 1991. Paleomagnetic Simpson, E.S.W., Sclater, J.G., Parsons, B., Norton, I., Meinke, L., evidence for late Miocene counterclockwise rotation of north coast 1979. Mesozoic magnetic lineations in the Mozambique Basin. carbonate sequence, Puerto Rico. Geophys. Res. Lett. 18, 565–568. Earth Planet. Sci. Lett. 43, 260–264. Richards, P.C., Fannin, N.G.T., 1997. Geology of the North Falkland Speranza, F., Maniscalco, R., Grasso, M., 2003. Pattern of orogenic rotations Basin. J. Pet. Geol. 20, 165–183. in central-eastern Sicily: implications for the timing of spreading in the Riley, T.R., Leat, P.T., Pankhurst, R.J., Harris, C., 2001. Origins of Tyrrhenian Sea. J. Geol. Soc. ( Lond. ) 160 (2), 183–195. large volume rhyolitic volcanism in the Antarctic Peninsula and Stegman, D.R., Freeman, J., Schellart, W.P., Moresi, L., May, D., Patagonia by crustal melting. J. Petrol. 42 (6), 1043–1065. 2006. Influence of trench width on subduction hinge retreat rates in Roca, E., Sans, M., Cabrera, L., Marzo, M., 1999. Oligocene to Middle 3-D models of slab rollback. Geochem. Geophys. Geosyst. 7, Miocene evolution of the central Catalan margin (northwestern Q03012. doi:10.1029/2005GC001056. Mediterranean). Tectonophysics 315, 209–233. Stern, C., De Wit, M., 2003. Rocas Verdes ophiolites, southernmost Rollet, N., Deverchere, J., Beslier, M.O., Guennoc, P., Rehault, J.-P., South America: remnants of progressive stages of development of Sosson, M., Truffert, C. 2002. Back arc extension, tectonic oceanic-type crust in a continental margin back-arc basin. In: inheritance, and volcanism in the Ligurian Sea, Western Mediterra- Dilek, Y., Robinson, P.T. (Eds.), Ophiolites in Earth History, Geol. nean. Tectonics. 21: 6-1-6-26. Soc. London Spec. Pub., 218, pp. 1–19. Rosenbaum, G., Lister, G.S., 2004a. Formation of arcuate orogenic Storey, B.C., 1995. The role of mantle plumes in continental breakup: belts in the western Mediterranean region. Geol. Soc. Am. Spec. case histories from Gondwanaland. Nature 377, 301–308. Paper 383, 41–56. Storey, B.C., Alabaster, T., Hole, M.J., Pankhurst, R.J., Wever, H.E., Rosenbaum, G., Lister, G.S., 2004b. Neogene and Quaternary rollback 1992. Role of subduction plate boundary forces during the initial evolution of the Tyrrhenian Sea, the Apennines, and the Sicilian stages of Gondwana break-up: evidence from the proto-Pacific Maghrebides. Tectonics 23, TC1013. doi:10.1029/2003TC001518. margin of Antarctica. In: Storey, B.C, Alabaster, T., Pankhurst, R.J. Rosenbaum, G., Lister, G.S., Duboz, C., 2002. Reconstruction of the (Eds.), Magmatism and the Causes of Continental Break-up, Geol. tectonic evolution of the western Mediterranean since the Soc. London Spec. Publ., vol. 68, pp. 149–163. Oligocene. Journal of the Virtual Explorer 8, 107–130. Storey, B.C., Vaughan, A.P.M, Millar, I.L., 1996. Geodynamic Rossetti, F., Faccenna, C., Goffé, B., Monié, P., Argentieri, A., Funiciello, evolution of the Antarctic Peninsula during Mesozoic times and R., Mattei, M., 2001. Alpine structural and metamorphic signature of its bearing on Weddell Sea history. In: Storey, B.C., King, E.C., the Sila Piccola Massif nappe stack (Calabria, Italy): insights for the Livermore, R.A. (Eds.), Weddell Sea Tectonics and Gondwana tectonic evolution of the Calabrian Arc. Tectonics 20 (1), 112–133. Break-up, Geol. Soc. Spec. Publ., vol. 108, pp. 87–103. Royden, L.H., 1993. Evolution of retreating subduction boundaries Storey, B.C., Curtis, M.L., Ferris, J.K., Hunter, M.A., Livermore, R.A., formed during continental collision. Tectonics 12, 629–638. 1999. Reconstruction and break-out model for the Falkland Islands Ruellan, E., Lagabrielle, Y., 2005. Subductions et ouvertures within Gondwana. J. Afr. Earth Sci. 29 (1), 153–163. océaniques dans le Sud-Oeste Pacific. Géomorphologie: Relief, Studinger, M., Miller, H., 1999. Crustal structure of the Filchner-Ronne Processus, Environnement, vol. 2, pp. 121–142. shelf and Coats Land, Antarctica, from gravity and magnetic data: Sager, W.W., Weiss, C.J., Tivey, M.A., Johnson, H.P., 1998. Geomag- implications for the breakup of Gondwana. J. Geophys. Res. 104, netic polarity reversal model of deep-tow profiles from the Pacific 20379–20394. Jurassic Quiet Zone. J. Geophys. Res. 103 (B3), 5269–5286. Sylwan, C.A., 2001. Geology of the Golfo San Jorge Basin, Argentina. Sandwell, D.T., Smith, W.H.F., 1997. Marine gravity anomaly from Geosat J. Iber. Geol. 27, 123–157. and ERS 1 satellite altimetry. J. Geophys. Res. 102, 10039–10054. Taylor, G.K., Shaw, J., 1989. The Falkland Islands: new paleomagnetic Sartori, R., 2003. The Tyrrhenian backarc basin and subduction of the data and their origin as a displaced terrane from southern Africa. Ionian lithosphere. Episodes 26 (3), 217–221. In: Hillhouse, J.W. (Ed.), Deep Structure and Past Kinematics of Sartori, R., Carrara, G., Torelli, L., Zitellini, N., 2001. Neogene Accreted Terranes, Geophysical Monograph, vol. 50, pp. 59–72. evolution of the southwestern Tyrrhenian Sea (Sardinia Basin and Ten Veen, J., Kleinspehn, K.L., 2003. Incipient continental collision and western bathyal plain). Mar. Geol. 175, 47–66. plate-boundary curvature: Late Pliocene–Holocene transtensional Scasso, R.A., Mendia, J.E., 1985. Rasgos estratigraficos y paleoam- Hellenic forearc, Crete, Greece. J. Geol. Soc. ( Lond. ) 60, 161–181. bientales del Paleozoico de las Islas Malvinas. Asoc. Geol. Argent. Thomson, K., 1998. When did the Falklands rotate? Mar. Pet. Geol. 40, 26–50. 15, 723–736. Schellart, W.P., 2004. Kinematics of subduction and subduction- Thomson, K., 1999. Role of continental break-up, mantle plume induced flow in the upper mantle. J. Geophys. Res. 109, B07401. development and fault reactivation in the evolution of the Gamtoos doi:10.1029/2004JB002970. Basin, South Africa. Mar. Pet. Geol. 16, 409–429. Schellart, W.P., Lister, G.S., Jewell, M.W., 2002. Some experimental Tikku, A.A, Marks, K.M., Kovacs, L.C., 2002. An Early Cretaceous insights into arc formation and back-arc extension. Bolletino de extinct spreading centre in the northern Natal Valley. Tectono- Geofisica Teorica ed Applicata 42, 85–91 (N. ½ suppl.). physics 347, 87–108. Schellart, W.P., Freeman, J., Stegman, D.R., Moresi, L., May, D., Tontini, F.C., Stefanelli, P., Giori, I., Fagggioni, O., Carmisciano, C., 2004. 2007. Evolution and diversity of subduction zones controlled by The revised aeromagnetic anomaly map of Italy. Ann. Geophys. 47, slab width. Nature 446, 308–311. doi:10.1038/nature05615. 1547–1555. 272 A.K. Martin / Tectonophysics 445 (2007) 245–272

Tucholke, B.E., Houtz, R.E., Barrett, B.M., 1981. Continental crust Antarctica, and implications for timing of Mesozoic terrane collision. beneath the Agulhas plateau, southwest Indian Ocean. J. Geophys. Geol. Mag. 139 (4), 761–765. Res. 86, 3791–3806. Wallace, L.M., McCaffrey, R., Beavan, J., Ellis, S., 2005. Rapid VanDecar, J.C., Russo, R.M., James, D.E., Ambeh, W.B., Franke, M., microplate rotations and backarc rifting at the transition between 2003. Aseismic continuation of the Lesser Antilles slab beneath collision and subduction. Geology 33 (11), 857–860. continental South America. J. Geophys. Res. 108 (B1), 2043. Watts, D.R., Bramall, A.M., 1981. Paleomagnetic evidence for a doi:10.1029/2001JB000884. displaced terrane in Western Antarctica. Nature 293, 638–641. Van Fossen, M.C., Channell, J.E.T., Schellekens, J.H., 1989. Wenzel, F., Sperner, B., Lorenzo, F., Mocanu, V., 2002. Geodynamics, Paleomagnetic evidence for Tertiary anticlockwise rotation in tomographic images and seismicity of the Vrancea region (SE- southwest Puerto Rico. Geophys. Res. Lett. 16, 819–822. Carpathians, Rumania). EGU Stephan Mueller Special Publication Van Hinsbergen, D.J.J., Van der Meer, D.G., Zachariasse, W.J., Series, European Geosciences Union, vol. 3, pp. 95–104. Meulenkamp, J.E., 2006. Deformation of western Greece during White, R.S., McKenzie, D.P., 1989. Magmatism at rift zones: the Neogene clockwise rotation and collision with Apulia. Int. J. Earth generation of volcanic continental margins and flood basalts. Sci. (Geol Rundsch). doi:10.1007/s00531-005-0047-5. J. Geophys. Res. 94, 7685–7729. Vaughan, A.P.M., Storey, B.C., 2000. The eastern Palmer Land shear Wortel, M.J.R., Spakman, W., 2000. Subduction and slab detachment in zone: a new Terrane accretion model for the Mesozoic development the Mediterranean–Carpathian region. Science 290, 1910–1917. of the Antarctic Peninsula. J. Geol. Soc. ( Lond. ) 157, 1243–1256. Vaughan, A.P.M., Kelley, S.P., Storey, B.C., 2002. Mid-Cretaceous ductile deformation on the Eastern Palmer Land Shear Zone,