Episodic sea-floor spreading in the Southern Red Sea

Tectonophysics 617 (2014) 140–149

Contents lists available at ScienceDirect

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

Episodic sea-floor spreading in the Southern Red Sea Khalid A. Almalki a,b,⁎, Peter G. Betts a,1, Laurent Ailleres a,2 a b

School of Geosciences Monash University, PO Box 28E, Wellington Road, Clayton, VIC 3800, Australia King Abdulaziz City for Science and Technology, PO Box 6086, Riyadh 11442, Saudi Arabia

a r t i c l e

i n f o

Article history: Received 16 July 2013 Received in revised form 19 January 2014 Accepted 24 January 2014 Available online 31 January 2014 Keywords: Aeromagnetic Magnetic stripe Transitional crust Spreading

a b s t r a c t The Red Sea represents the most spectacular example of a juvenile ocean basin on the modern Earth. Synthesis of regional aeromagnetic data, gravity data, seismic refraction data coupled with structural mapping from the Farasan Islands suggest that the opening of the Red Sea is complex and episodic. Modeling of magnetic and gravity data constrained by seismic refraction data reveals the Arabian Shelf is underlain by oceanic and transitional crust and that mafic diking and intrusions are focused at the continental–transitional crust boundary. This relationship is interpreted to indicate that early Miocene diking along the Arabian Escarpment heralded termination of oceanic basin formation and a shift in the locus of extension focused from a central mid-ocean ridge spreading center to the continental–transitional crust zone. Uplift along the Arabian Escarpment caused erosion and Middle to Late Miocene sedimentation of the Farasan Bank onto existing oceanic crust, suggesting that the extensive sedimentary banks of the southern Red Sea are not passive margins. Re-initiation of spreading occurred at ca 5 Ma. Pliocene to Pleistocene Shelf reef systems (Farasan Islands), developed on the flanks of the spreading ridge, are extensively overprinted by normal faults, suggesting that not all crustal extension is accommodated by active spreading. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The Red Sea represents the modern example of a juvenile crust that has recently undergone the transition from continental rifting, characterized by crustal attenuation, to active mid-ocean ridge spreading (Bosworth et al., 2005; Lazar et al., 2012) (Fig. 1). There is a consensus that the axial trough of the Red Sea formed by mid-ocean ridge began at ca 5 Ma (e.g., Axen et al., 2001; Chu and Gordon, 1998; Pallister et al., 2010). However, an intriguing problem concerning the evolution of the Red Sea is whether it formed during one-stage (e.g. Bosworth et al., 2005) or two-stage spreading as proposed by several early researchers (e.g., Brown and Girdler, 1982; Girdler and Styles, 1974; Hall, 1989). Tectonic models based on geological data are dominated by single stage rift models involving protracted stretching of continental crust followed by sea floor spreading at ca 5 Ma, and include both asymmetric (e.g., Dixon et al., 1989; Voggenreiter et al., 1988; Wernicke, 1985) and symmetrical extension models (e.g., Berhe, 1986; Bohannon and Eittreim, 1991; Martinez and Cochran, 1988). These models suggest that the Miocene sedimentary shelves (known as the Farasan and ⁎ Corresponding author at: School of Geosciences Monash University, PO Box 28E, Wellington Road, Clayton, VIC 3800, Australia. Tel.: +61 3 9905 4886; fax: +61 3 9905 4903. E-mail addresses: [email protected] (K.A. Almalki), [email protected] (P.G. Betts), [email protected] (L. Ailleres). 1 Tel.: +61 3 9905 4150; fax: +61 3 9905 4903. 2 Tel.: +61 3 9905 1526; fax: +61 3 9905 4093. 0040-1951/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tecto.2014.01.030

Dahlak Banks) (Fig. 1) of the southern Red Sea represent passive margins underlain by attenuated continental lithosphere. However, these models are not supported by reconstructions of the Red Sea by Sultan et al. (1992), which suggest that much of the Red Sea substrate may be entirely oceanic crust. Further, the composition and architecture of the crust that lies on either side of the spreading ridge beneath the Miocene marine sedimentary shelves show geophysical affinities to oceanic crust (e.g., Brown and Girdler, 1982; Gettings et al., 1986; Hall, 1989; Mechie et al., 1986). Magnetic profiles in the southern Red Sea were used by Hall (1989) to interpret the presence of Oligocene oceanic crust (Fig. 2). However, the validity of magnetic stripes to interpret the onset of ocean floor spreading has been called into question in recent literature by the recognition that magnetic stripes during the transition from rift to drift can form as a result of dense magma intrusion into stretched transitional crust (e.g., Bronner et al., 2011). That said, documented examples of magnetic striping in transitional crust show evidence for one or two magnetic stripes associated transitional crust, suggesting that such striping must occur immediately before the onset of ocean spreading (Bridges et al., 2012). These relatively recent observations could be used to support assertions by Voggenreiter et al. (1988) that many of the magnetic anomalies interpreted by Hall (1989) where sourced from linear mafic intrusions parallel to the rift. We combine high resolution aeromagnetic data with regional Bouguer gravity, seismic refraction data from the Farasan Bank, and field structural mapping from the Farasan Islands to constrain the evolution of the southern Red Sea since the Oligocene. In particular, using

K.A. Almalki et al. / Tectonophysics 617 (2014) 140–149

141

Fig. 1. Bathymetry and topography image after Becker et al. (2009) showing major Arabian plate boundaries and extension direction with residual velocities of the region after ArRajehi et al. (2010) and Bellahsen et al. (2003), wells, and seismic profiles locations at the Red Sea.

reprocessed and filtered magnetic data we address the nature of the basement under the Farasan Bank. The results show remarkable complexity including multiple episodes of mid-ocean ridge spreading, intervened by an episode of failed rifting and Miocene shelf sedimentation.

2. Data and analysis Aeromagnetic data and Bouguer gravity data were collected by the Deputy Ministry of Minerals Resources of Saudi Arabia. Aeromagnetic

Fig. 2. Observed magnetic profile across the Arabian Shelf correlation with synthetic sea-floor spreading profile after Hall (1989). Location of the profile is showing in Fig. 4A.

142

K.A. Almalki et al. / Tectonophysics 617 (2014) 140–149

data were collected along NE–SW flight lines with a flight elevation of 300 m and line spacing of 1250 m. The gravity data comprised of 145 stations covering an area of 150 km × 120 km and were tied to the gravity base station (KGN-47). Stations in the Farasan Islands were collected at or near sea level. The survey was completed on nominal 10 km station

spacing on a square grid (Gettings, 1977). The distribution of the gravity stations are shown in Fig. 3 with the density of measurements slightly greater on the Arabian landmass compared with the Red Sea. The reference density used to calculate the Bouguer anomaly is 2.67 g/cm3 (for land) and 1.03 g/cm3 (for water) (Gettings, 1977). The complete

Fig. 3. (A) Map showing the location of the gravity stations after Gettings (1977, 1983). (B) Contour map of the Bouguer gravity data. (C) Elevation image of the gravity stations.

K.A. Almalki et al. / Tectonophysics 617 (2014) 140–149

Bouguer anomaly values include corrections for latitude (theoretical gravity), elevation (free-air), Bouguer slab (simple Bouguer), curvature, and terrain effects (Gettings, 1983). Detailed information about data procedures are outlined in Directorate General of Mineral Resources Bulletin 22 and the U.S. Geological Survey open-file report 83-789 (Gettings, 1977; Gettings, 1983). We present the tilt-derivative (Verduzco et al., 2004) of the total magnetic intensity aeromagnetic data (Fig. 4A,B). This type of filtering or data enhancement normalizes magnetic anomaly amplitudes and is effective for imaging subtle anomalies within shallow basement structure (Verduzco et al., 2004). The tilt-derivative is also effective because it highlights subdued amplitudes due to the burial of the magnetic basement, such as the basement buried beneath the 4 km of stratigraphy of the Farasan Bank. A complete Bouguer gravity dataset was gridded with a grid cell size of 500 m (Fig. 4C). The gravity data was effective at illustrating density variations between different types of crust as it effectively discriminates zones of intense mafic magmatism from the continental crust for example. The seismic reflection data were collected by the USGS and traverse the Farasan Bank and the Arabian Shield (Fig. 1). Data was collected at six shot points covering a 1000 km long north east trending traverse from the Red Sea trough to Riyadh (Gettings et al., 1986) (Figs. 1, 4D). We combine different geophysical datasets (Fig. 4) along the southern coast of Saudi Arabia to test the existence of oceanic crust under the Arabian Shelf and present an assessment of the crustal structure relationship between the Arabian Shelf and the Arabian Shield. We have also forward

143

modeled the magnetic and Bouguer gravity data to illustrate the architecture of the Farasan Bank and the underlying basement. Forward models are constrained by seismic refraction data. Four separate crustal zones have been identified based on geophysical criteria. Fieldwork throughout the Farasan Islands involved stratigraphic logging and structural mapping and analysis that provide evidence for the tectonic processes interpreted in this part of the Red Sea basin. 2.1. Mapping results The Farasan Islands contain a lower package of Middle to Late Miocene shale and evaporite successions up to 4 km thick overlain by Pliocene and Pleistocene shallow shelf limestone (Dabbagh et al., 1984) (Fig. 5). The fault architecture of the Farasan Islands (Fig. 5) indicates two patterns of extensional fault activity. Extensional faults and fractures parallel to the rift axis of the Red Sea overprint Pliocene reefs and bound half graben (Fig. 5). Elliptical normal fault systems bound salt domes and were formed as a result of diapirism of Miocene evaporite successions. The salt domes display elongate geometries parallel or orthogonal to the Red Sea axis, suggesting that salt deposition and diapirism were controlled by existing basement structures. 2.2. Aeromagnetic interpretation The aeromagnetic data reveal evidence for seven magnetic stripes (Fig. 4B). Four prominent linear magnetic anomalies that underlie the Miocene Farasan Bank and an additional three stripes lie beneath the

Fig. 4. (A) Aeromagnetic data: Reduce to the Pole, the Farasan Islands and the coast line position, profile location of Figs. 2 and 7A, seismic line and shot point 6 location of panel D. (B) Aeromagnetic data: Tilt Derivative, the Farasan Islands and the coast line position, seismic line and shot point 6 location of panel D. (C) Complete Bouguer gravity data, the Farasan Islands and the coast line position, profile location of Fig. 7B, seismic line and shot point 6 location of panel D. (D) Seismic refraction crustal model modified after Gettings et al. (1986).

144

K.A. Almalki et al. / Tectonophysics 617 (2014) 140–149

Fig. 5. Geological map of the Farasan Island showing the location of salt domes and NW–SE trending normal faults parallel with the Red Sea rift axis. Inset show detailed maps of extensional structures and stratigraphic log of the Farasan Bank.

coastal plain of SW Saudi Arabia (Fig. 4B). All of these anomalies are parallel with the Red Sea rift axis. The stripes under the bank exhibit normal polarity surrounded by wider zones of dominantly reverse polarity.

Magnetic stripes are transected and offset by NE–SW faults which are interpreted to be transform faults (Fig. 6). The faults offset stripes with apparent dextral movement. Smaller scale faults are interpreted

Fig. 6. Geological interpretation map of aeromagnetic and gravity data constrained by seismic, and outcrop information.

K.A. Almalki et al. / Tectonophysics 617 (2014) 140–149

from the magnetic data based on steep gradients in tilt derivative data (Figs. 4B, 6). These faults trend NW–SE and N–S. Their importance is difficult to interpret, however, because they offset magnetic anomalies they are interpreted as basement faults. Faults at high angles to the rift axis and exhibiting large offsets (N2 km) may represent transverse faults accommodating the extension and rifting over the last 30 Ma. The magnetic data also show NW–SE striking magnetic dikes that trend parallel with the Red Sea rift along the edge of the Arabian Shield (Figs. 4B, 6). These dikes characterized by irregular short wavelength magnetic anomalies (2–10 km) of smaller amplitude between 20 and 50 nT over the coastal plain. At least thirteen individual dikes, forming a NW–SE trending and 20 km wide swarm, have been identified from the aeromagnetic data (Figs. 4B, 6). This narrow zone of diking correlates with the outcropping of the Tihama Asir Magmatic Complex, which is part of a larger dike swarm that extends along the Arabian margin (e.g., Bosworth et al., 2005). The dikes have been dated at ca 22–24 Ma (Ar40–Ar39 cooling ages: Sebai et al., 1991). Magnetic interpretation suggests that normal faulting (or fault zones) along the coastal plain (Fig. 6), overprint the Tihama Asir Magmatic Complex dikes. The Arabian Escarpment represents a first order normal fault (Fig. 6). This fault is characterized by a curvilinear strike that rotates towards the west in the southern part of the study area. To the north the escarpment trends NW–SE, striking (Fig. 6) parallel with the Red Sea. 2.3. Bouguer gravity interpretation The Bouguer gravity data show negative (− 25 to − 133 mGal) anomalies over the Arabian shield, whereas the Arabian Shelf has a long wavelength response of approximately zero mGal (Fig. 4C). We attribute this response to the influence of 4 km thick low density sedimentary successions of the Farasan Bank overlying denser crust. Between the Arabian Shield and the Farasan Bank, a 30 km wide, positive anomaly (+ 25 to + 35 mGal) exists parallel to the Red Sea axis (Fig. 4C). The long wavelength component of the anomaly (Fig. 4C) suggests that the source is comprised of mid-crustal mafic intrusions emplaced into extended transitional crust. This anomaly is comparable with gravity anomalies above transitional crust in the southern Atlantic ocean (Blaich et al., 2011), where the greater volume of mafic intrusions occur (Direen et al., 2012). This interpreted zone of transitional crust has subsequently been intruded by the Miocene and Quaternary dikes, which coincide geographically with the anomaly peaks. The long wavelength component of the anomaly is coincident with three magnetic stripes (Fig. 4B) suggesting that they correspond to the development of a transitional crust as previously proposed in the Afar Rift (e.g., Bastow and Keir, 2011) and the southern Atlantic (Blaich et al., 2011). 2.4. Seismic refraction interpretation Seismic refraction data show 9 km of mafic crust under the Farasan Bank (e.g., Mooney et al., 1985) (Fig. 4D). Gettings et al. (1986) interpreted a lower layer with a velocity of 6.8 km/s and the upper layer with a velocity of 6.4 km/s beneath the Farasan Bank. Farasan Bank sediments have a modeled seismic velocity of 4.2 km/s. The high velocity layers beneath the bank have similar velocities to the present day axial trough of the Red Sea (6.8–7.3 km/s) (Gettings et al., 1986), and to other oceanic crust (e.g. offshore from Wilkes Land (Antarctica): Close et al., 2009), suggesting that the crust beneath the Farasan Bank is likely oceanic in origin. Seismic refraction crustal models show a rapid decrease in crustal thickness at the Arabian Escarpment from 40 km below the Arabian Shield to 11 km or 9 km beneath the Farasan Bank (e.g., Mechie et al., 1986; Milkereit and Fluh, 1985; Mooney et al., 1985) (Fig. 4D). This step in the Moho is coincident with the positive Bouguer anomaly (Fig. 4C) as well as the three magnetic stripes (Figs. 4B, 6). These relationships suggest a rapid transition from Arabian Shield continental crust to Oligocene ocean crust.

145

3. Data modeling The magnetic profile across the stripes (Fig. 7A) beneath the bank demonstrates an asymmetric pattern in which the crust is dominated by a positive total field magnetic anomaly flanked by negative total field anomalies with amplitudes similar to those observed over oceanic spreading centers. The stripes under the shelf are characterized by long wavelength anomalies with large amplitude between 60 and 90 nT over the Arabian Shelf (Fig. 7A). In the coincident gravity profile (Fig. 7B), an oceanic crust is represented by a relatively flat response between the Saudi Arabian coast and the axial trough (0 to −25 mGal). The oceanic sources of the gravity anomalies have been modeled using a tripartite density profile to represent different oceanic crustal levels (Fig. 7C). The upper oceanic crust is modeled with a density of 2.6 g/cm3 and is less than 4 km thick. This part of the crust is interpreted to consist of altered basalts. The underlying sheeted dikes have been modeled with a density of 2.7 g/cm3. This layer is up to 2 km thick and has an undulating base. The lower oceanic crust is interpreted to consist of gabbro and is modeled with a density of 2.9 g/cm3. The depth to Moho varies from 8 km near the axial trough to 15 km to the coastal plain (Fig. 7C) based on seismic refraction data (Fig. 4D). We have modeled a transitional crust with a homogeneous density of 3.0 g/cm3 that is characterized by short wavelength anomalies (10–20 km) with amplitude of ~40 nT. This region of intense diking associated with the Tihama Asir Magmatic Complex has a wedge-shaped geometry in cross-section with a modeled density of 2.85 g/cm3 (Fig. 7C). The continental crust is characterized by a regional low Bouguer gravity anomaly (−25 to −84 mGal) (Fig. 7B). The boundary between transitional and continental crust is located ~35 km to the east from the Saudi Arabian coastline (Figs. 4B, 6) and is coincident with the Arabian Escarpment. The magnetic response of the Arabian shield is characterized by short wavelength anomalies (5–15 km) with variable amplitude between 40 and 70 nT (Fig. 7A). Typical anomalies within this zone have short wavelengths (e.g. 10 km) probably because of increasing thickness of the onlapping sedimentary cover and the heterogeneous distribution of magnetite in the Precambrian rocks. The Arabian Shield continental crust is modeled with a density of 2.76 g/cm3 and shows the depth to Moho stepping from 18 km in the transitional crust to ~40 km (Fig. 7C) over a distance of 20 km consistent with the seismic refraction profile. 4. Discussion Previous interpretations have attempted to account for the three stripes under the coastal region by proposing extended oceanic crust between 28 and 26 Ma (e.g., Hall, 1989) (Fig. 2). However, these models are not supported by the seismic and gravity data which show crustal thicknesses of up to 15 km (Fig. 4D) (e.g., Bohannon, 1989) and an increase in gravity response related to mid-crustal mafic intrusions emplaced into extended transitional crust (Fig. 4C) (e.g. Blaich et al., 2011; Direen et al., 2012). The three oldest magnetic stripes identified in aeromagnetic data (Figs. 4B, 6) are interpreted to have formed within transitional crust at the margin of the Arabian plate. If this interpretation is correct then this represents the most magnetic striping recorded in transitional crust at a continental margin. The boundary between transitional crust and the continental crust is defined by a relatively steep gravity gradient (Figs. 4C, 7B) (e.g., Gettings et al., 1986), which coincides with a step in the Moho in seismic refraction data (e.g., Mooney et al., 1985) (Fig. 4D) and a major normal fault with a down throw to the west which is expressed as the Arabian Escarpment (Fig. 6). Four additional stripes identified in aeromagnetic data lie beneath the Farasan Bank (Fig. 4B). These magnetic stripes represent ~ 75 km of oceanic crust that formed before the deposition of Middle to Late Miocene (15–5 Ma) sedimentary succession of the bank. If our interpretations are correct then similar ocean crust may exist beneath the

146

K.A. Almalki et al. / Tectonophysics 617 (2014) 140–149

Fig. 7. (A) Observed and calculated magnetic profile showing different types of crust, refer to Fig. 4A for location. (B) Observed and calculated bouguer gravity profile, refer to Fig. 4C for location. (C) Geophysical cross-section based on forward modeling of the geophysical datasets illustrates the architecture of the Arabian Shelf and part of the Arabian Shield underlying basement.

Dahlak Bank on the conjugate margin (Fig. 1). However, quality of magnetic data on the Dahlak Bank is not sufficient to test this hypothesis and therefore we cannot distinguish if the stripes on the African side “Dahlak Bank” are correlated to Farasan Bank stripes or simply represent transitional stripes or a combination of both. Our modeling suggests that the crust beneath the Farasan Bank is ~ 9 km thick. This crust may be thicker than typical oceanic crust, but not uncommon for oceanic crust formed during initiation of spreading due to elevated melt production induced by steep gradients on the lithosphere–asthenosphere boundary (e.g., Ligi et al., 2011). Oceanic crust is interpreted to be located beneath the Arabian Shelf. This interpretation is poorly resolved in the gravity data and corresponds to a shallow gravity gradient from − 5 to − 25 mGal (Fig. 7B). However, the interpreted boundary between oceanic and transitional crust coincides with a decrease in the long wavelength (N60 km) response of magnetic data (Fig. 7A). This interpretation is also supported by seismic refraction velocity (6.8 km/s), which are comparable to seismic velocities at the active mid-ocean ridge (Gettings et al., 1986) but this is only the case for the lowermost crust shown in Fig. 4D. The rest of the crust below the Farasan Bank has a seismic velocity of 6.4 km/s, which is lower than the 6.8 km/s crust at the current oceanic ridge axis. We attribute this difference to the thick (up to 4 km) sedimentary cover above the ocean crust, which is absent in the axial trough (e.g., Mooney et al., 1985). This interpretation is consistent with more evolved ocean basins where crustal velocity are lower where sedimentary cover is thicker (e.g., western Atlantic seafloor) compared to regions of thinner sedimentary cover (e.g., eastern Pacific seafloor) (Chulick and Mooney, 2002; Chulick et al., 2013). The presence of ocean crust beneath the Miocene Arabian Shelf supports interpretation of the crustal structure beneath the southern Red Sea margin in Yemen (Ahmed et al., 2013) suggesting a two-stage spreading evolution for the Red Sea, and not a single stage of MOR spreading since the Pliocene (e.g., Bosworth et al., 2005). Initial interpretations of the youngest stripe age at 34 Ma (Girdler and Styles, 1974) and 32 Ma (Girdler and Underwood, 1985) suggests that this stripe would pre-date the arrival of the Afar Plume at 30 Ma (e.g., Bosworth et al., 2005) and therefore are considered incorrect

interpretations (Fig. 8). However, Hall (1989) correlated this stripe with synthetic magnetic profiles and interpreted it to be dated at 24 Ma (Fig. 2) consistent with rifting and subsequent ocean floor spreading following the arrival of the Afar Plume. Further, the magnetic stripes distribution suggests that ocean crust production rate increased after ca 26 Ma before termination of spreading (Fig. 6). Independent support for the presence of oceanic crust beneath the shelf comes from our modeling of the magnetic and gravity data (Fig. 7). Although our results are based primarily on the modeling of the gravity and magnetic profile across the shelf, they are consistent with other geophysical and geological data for the region (Fig. 1). For example, four boreholes in the neighborhood of the profile penetrated basalt (Dungunab-1; Durvara-2; Amber-1 and B1; Brown and Girdler, 1982; Coleman, 1977; Girdler and Styles, 1974; Lowell and Genik, 1972) (Fig. 1). Further, there are at least four seismic profiles which constrain the velocity structure to oceanic crust in the main trough of central and northern Red Sea (Davies and Tramontini, 1970; Drake and Girdler, 1964; Egloff et al., 1991; Gaulier et al., 1988; Makris et al., 1991; Rihm et al., 1991) (Fig. 1). A second phase of lithospheric extension, focused at the continent– ocean transition, indicated by intense ca 24–22 Ma dike emplaced along the Arabian Escarpment (Sebai et al., 1991) (Fig. 6). The gravity profile (Fig. 7B) shows high-amplitude gravity maximum over the continental–oceanic transition zone, which suggests that high-density sources in this area were produced by a gradual increase of diking as the lithosphere became extended. We interpret that the high-density source corresponds to the intruded dikes. We cannot distinguish if these dikes represent a distinct extensional event or represents a change in the locus of extension. In either scenario it appears that MOR spreading terminated and was focused at the continental– transitional crust transition. Diking may have led to localized stretching and adiabatic mantle melting right at the transition between oceanic and continental crust (Fig. 8) and may represent a failed attempt at a ridge jump. This episode of magmatism may have initiated thermallydriven uplift along the Arabian continental margin, which we interpret provided the source for the Farasan Bank sediments (Fig. 8). Fission track dates from the Arabian shield indicate that uplift has occurred in

K.A. Almalki et al. / Tectonophysics 617 (2014) 140–149

147

Fig. 8. The development of the southern Red Sea. Inset shows the regional plate tectonic position at ca 5 Ma and approximate cross-section location. (A) Schematic cross-sections of the southern Red Sea illustrating possible sub-crustal relationships at 5–0 Ma. (B) Schematic cross-sections of the southern Red Sea illustrating possible sub-crustal relationships at 22–6 Ma. (C) Schematic cross-sections of the southern Red Sea illustrating possible sub-crustal relationships at 26–24 Ma. (D) Schematic cross-sections of the southern Red Sea illustrating possible sub-crustal relationships at 28 Ma. (E) Schematic cross-sections of the southern Red Sea illustrating possible sub-crustal relationships at 31 Ma.

the last 14 Ma (Bohannon et al., 1989) coincident with the formation of the Farasan Bank (Fig. 8). We suggest that the Farasan Bank represents a prograding shelf deposited on an oceanic substrate and not a passive margin (Fig. 8). The sedimentation during bank formation is interpreted to represent a period of relative tectonic quiescence in which no crustal extension is recorded until the onset of renewed mid-ocean ridge spreading at ca 5 Ma (Fig. 8), which was superimposed on the existing Oligocene ridge and remains active today (e.g., Ligi et al., 2011), creating approximately 75 km of new oceanic crust. Crustal extension is also occurring away from the mid-ocean ridge and is evident by extensional fault development in the Farasan Islands which records present-day normal faulting parallel to the rift axis (Fig. 5), suggesting that separation between the Arabian and African plates is not accommodated by extension

along the mid-ocean ridge alone. The fact that salt diapirs trend parallel to the transform faults and the rift axis suggests a control by preexisting architecture on this part of the basin. 4.1. Geodynamic drivers The geodynamic drivers for opening of the southern Red Sea are complex and possibly include the arrival of the Afar Plume, and forces related to slab pull during closure of the Neo-Tethys during the Oligocene (e.g., Hempton, 1987; Jolivet and Faccenna, 2000), or combinations of both (Bellahsen et al., 2003; Zeyen et al., 1997) (Fig. 8). The magmatic expression of the plume is recorded as far north as 20°N (Zeyen et al., 1997). We interpret this to indicate that the Afar Plume played a major role in triggering ocean spreading at ca

148

K.A. Almalki et al. / Tectonophysics 617 (2014) 140–149

26 Ma by either initiating ocean spreading or weakening the continental lithosphere sufficiently to allow MOR spreading (Fig. 8). Extension may have been driven by slab pull from the down-going Arabian plate during Neo-Tethys closure (Fig. 8). The rate of spreading was established through comparisons between different magnetic reversal time-scales (Girdler and Styles, 1974; Girdler and Underwood, 1985; Hall, 1989) indicating a spreading rate of 1.4 and 2.2 cm/year. This is higher than the current 1.3–1.8 cm/year extension rate measured using magnetic and GPS data at the current spreading center in the same study region (ArRajehi et al., 2010; Chu and Gordon, 1998; McClusky et al., 2010). Plate reconstructions coupled with GPS measurements suggest that Africa plate motion relative to Eurasia has been constant since 13 Ma but is 70% slower than the rate at 30–13 Ma while Arabia–Eurasia motion has been largely constant between 2 and 3 cm/year (McClusky et al., 2010; McQuarrie et al., 2003). Further, the same studies indicate that the rate of Africa–Arabia relative motion increased by 70% at 13 Ma (ArRajehi et al., 2010; McClusky et al., 2010; McQuarrie et al., 2003). This change in plate velocity is caused by independent motion of Arabia due to oblique rotation of the Arabian plate (e.g., Bellahsen et al., 2003; Beydoun, 1999; Bosworth et al., 2005; Hempton, 1987; McQuarrie et al., 2003) and is unlikely to be a reliable indicator of the Oligocene spreading rate between 26 and 24 Ma. We suggest that a faster rate of spreading in the Oligocene reflects greater slab pull forces on the Arabian plate during the final stages of Tethys closure (Bellahsen et al., 2003). This is consistent with the indicated convergence rate between Arabia and Eurasia of 3.2 cm/year at N 25 Ma deduced from GPS data with geological estimates (Floyd et al., 2012; Reilinger and McClusky, 2011). An implication of our findings is that the rate of spreading was approximately 5 cm/year if a conjugate pair exists off the coast of Eritrea, which is twice the rate of the present day spreading rate (e.g. Floyd et al., 2012). The difference may be attributed to the geodynamic drivers operating in the region. Late Oligocene spreading was more likely to be influenced by subduction pull forces related to the closure of the Neotethys (e.g., Bellahsen et al., 2003), whereas in the present-day setting such geodynamic drivers are not present, and therefore spreading rates are slower. Termination of ocean spreading at 24 Ma may have resulted from either slowing of the convergence between Arabia and Eurasia or the onset of the Arabian plate passive margin collision with Eurasia (e.g. Ballato et al., 2011). This is supported by recent Apatite fission track data along the Bitlis–Zagros thrust zone (Fig. 1) that shows the collision of Arabia with Eurasia occurred in the early Miocene (ca 20 Ma; Okay et al., 2010), thus the indicated ~1 cm/year rate of extension for the Miocene (ArRajehi et al., 2010; McClusky et al., 2010; McQuarrie et al., 2003) may reflect either a decrease in slab pull forces or collision between the Arabian and Eurasian Plates. A feature of our interpretation is the stepping of the locus of extension from the spreading ridge to the oceanic–continental transition zone at about the same time as spreading appears to have terminated (Fig. 8). This unusual scenario where crustal extension shifts off-axis (in the form of dikes) and becomes focused at the continent–ocean transitional zone may be driven by external forces related to the closure of the Neo-Tethys and is therefore not typical of juvenile oceanic basins, or possibly heralds a failed attempt for the mid-ocean ridge to jump towards the off-axis region as is currently occurring in the Danakil Depression in the Afar region today (Bridges et al., 2012) (Fig. 1). This off-axis shift in the locus of extension may reflect reduced extensional stress transfer across the Arabian plate which may have happened because the convergence rates decreased and/or there was a decrease in slab pull along the northern margin of the Arabian plate (e.g., Hatzfeld and Molnar, 2010; McQuarrie et al., 2003). Off-axis magmatism near the rift margin is currently observed in Saudi Arabia and further south in Ethiopia (e.g. Pallister et al., 2010; Rooney et al., 2011) where mid-ocean ridge jumping is speculated. Renewed mid-ocean ridge spreading has been related to plate reorganization at ca 5 Ma attributed to the transition from subduction to

oblique collision between Eurasia and Arabia at 6 Ma (Axen et al., 2001) (Fig. 8) and/or active subduction zone at Makran (Fig. 1) (Bellahsen et al., 2003) resulting in a counterclockwise rotation of the Arabian plate with respect to Africa (Westaway, 1994). Further, we attribute the north–south faults interpreted over the old oceanic crust under the bank to be recent structures (b 1 Ma) related to an evident major off-axis shift of this spreading in a north–south direction towards the Arabian plate as interpreted from the presence of anomalously slow S-wave velocities in the mantle beneath the western Arabian plate (Chang et al., 2011).

Acknowledgments We would like to acknowledge the King Abdulaziz City for Science and Technology for financial support to Khalid Almalki. The Geological Survey of Saudia Arabia is thanked for providing the geophysical data. Peter Betts was supported by the Monash Researcher Accelerator Program.

References Ahmed, A., et al., 2013. Crustal structure of the rifted volcanic margins and uplifted plateau of Western Yemen from receiver function analysis. Geophys. J. Int. 193, 1673–1690. ArRajehi, A., et al., 2010. Geodetic constraints on present-day motion of the Arabian plate: implications for Red Sea and Gulf of Aden rifting. Tectonics 29, 1–10. Axen, G.J., Lam, P.S., Grove, M., Stockli, D.F., Hassanzadeh, J., 2001. Exhumation of the west-central Alborz Mountains, Iran, Caspian subsidence, and collision-related tectonics. Geology 29 (6), 559–562. Ballato, P., Uba, C.E., Landgraf, A., Strecker, M.R., Sudo, M., Stockli, D.F., Friedrich, A., Tabatabaei, S.H., 2011. Arabia–Eurasia continental collision: insights from late Tertiary foreland-basin evolution in the Alborz mountains, northern Iran. Geol. Soc. Am. Bull. 123, 106–131. Bastow, I.D., Keir, D., 2011. The protracted development of the continent–ocean transition in Afar. Nat. Geosci. 4, 248–250. http://dx.doi.org/10.1038/NGEO1095. Becker, J.J., et al., 2009. Global bathymetry and elevation data at 30 arc seconds resolution: SRTM30_PLUS. Mar. Geod. 32, 355–371. Bellahsen, N., Faccenna, C., Funiciello, F., Daniel, J.M., Jolivet, L., 2003. Why did Arabia separate from Africa? Insights from 3-D laboratory experiments. Earth Planet. Sci. Lett. 216, 365–381. Berhe, S.M., 1986. Geologic and geochronologic constraints on the evolution of the Red Sea–Gulf of Aden and Afar Depression. J. Afr. Earth Sci. 5, 101–117. Beydoun, Z.R., 1999. Evolution and development of the Levant (Dead Sea Rift) Transform System: a historical–chronological review of a structural controversy. Geol. Soc. Lond. Spec. Publ. 164, 239–255. Blaich, O.A., Faleide I, J.I., Tsikalas, F., 2011. Crustal breakup and continent‐ocean transition at South Atlantic conjugate margins. J. Geophys. Res. 116, B01402. Bohannon, R.G., 1989. Style of extensional tectonism during rifting, Red Sea and Gulf of Aden. J. Afr. Earth Sci. 8, 589–602. Bohannon, R.G., Eittreim, S.L., 1991. Tectonic development of passive continental margins of the southern and central Red Sea with a comparison to Wilkes Land, Antarctica. Tectonophysics 198, 129–154. Bohannon, R.G., Naeser, C.W., Schmidt, D.G., Zimmwermann, R.G., 1989. The timing of uplift, volcanism and rifting peripheral to the Red Sea, a case for passive rifting. J. Geophys. Res. 94, 1683–1701. Bosworth, W., Huchon, P., McClay, K., 2005. The Red Sea and Gulf of Aden Basins. J. Afr. Earth Sci. 43, 334–378. http://dx.doi.org/10.1016/j.jafrearsci.2005.07.020. Bridges, D.L., Mickus, K., Gao, S.S., Abdelsalam, M.G., Alemu, A., 2012. Magnetic stripes of a transitional continental rift in Afar. Geology 40 (3), 203–206. http://dx.doi.org/ 10.1130/G32697. Bronner, A., Sauter, D., Manatschal, G., Peron-Pinvidic, G., Munschy, M., 2011. Magmatic breakup as an explanation for magnetic anomalies at magma-poor rifted margins. Nat. Geosci. 4, 549–553. http://dx.doi.org/10.1038/NGEO1201. Brown, C., Girdler, R.W., 1982. Structure of the Red Sea at 20° N from gravity data and its implication for continental margins. Nature 298, 51–53. Chang, S.-J., Merino, M., Van der Lee, S., Stein, S., Stein, C.A., 2011. Mantle flow beneath Arabia offset from the opening Red Sea. Geophys. Res. Lett. 38, 1–5. Chu, D., Gordon, R.G., 1998. Current plate motions across the Red Sea. Geophys. J. Int. 135, 313–328. Chulick, G.S., Mooney, W.D., 2002. Seismic structure of the crust and uppermost mantle of North America and adjacent ocean basins: a synthesis. Bull. Seismol. Soc. Am. 92, 2478–2492. Chulick, G.S., Detweiler, S., Mooney, W.D., 2013. Seismic structure of the crust and uppermost mantle of South America and surrounding oceanic basins. J. S. Am. Earth Sci. 42, 260–276. Close, D.I., Watts, A.B., Stagg, H.M.J., 2009. A marine geophysical study of the Wilkes Land rifted continental margin, Antarctica. Geophys. J. Int. 177, 430–450. http://dx.doi.org/ 10.1111/j.1365-246X.2008.04066.x.

K.A. Almalki et al. / Tectonophysics 617 (2014) 140–149 Coleman, R.G., 1977. Geologic background of the Red Sea. In: Hilpert, L.S. (Ed.), Red Sea Research 1970–1975. Bulletin, 22. Saudi Arabian Ministry of Petroleum and Mineral Resources, Directorate General Mineral Resources, Jeddah, pp. G1–G18. Dabbagh, A., Hotzl, H., Schnnier, H., 1984. Farasan Islands. In: Jado, A.R., Zoetl, J.G. (Eds.), Quaternary Period in Saudi Arabia, vol. 2. Springer-Verlag, Vienna, Austria, pp. 212–218. Davies, D., Tramontini, C., 1970. The deep structure of the Red Sea. Philos. Trans. R. Soc. Lond. 267, 181–189. Direen, N.G., Stagg, H.M.J., Symonds, P.A., Norton, I.O., 2012. Variations in rift symmetry: cautionary examples from the Southern Rift System (Australia–Antarctica). Geol. Soc. Lond. Spec. Publ. 369. Dixon, T.H., Ivins, E.R., Franklin, B.J., 1989. Topographic and volcanic asymmetry around the Red Sea: constraints on rift models. Tectonics 8, 1193–1216. Drake, C.L., Girdler, R.W., 1964. A geophysical study of the Red Sea. Geophys. J. R. Astron. Soc. 8, 473–495. Egloff, F., et al., 1991. Contrasting structural styles of the eastern and western margins of the southern Red Sea: the 1988 SONNE experiment. Tectonophysics 198, 329–353. Floyd, M.A., et al., 2012. Active deformation of the Arabian plate and its margins. Abstract G53A-1125 presented at 2012 Fall MeetingAGU, San Francisco, Calif (3–7 Dec.). Gaulier, J.M., et al., 1988. Seismic study of the crust of the northern Red Sea and Gulf of Suez. Tectonophysics 153, 55–88. Gettings, M.E., 1977. Delineation of the continental margin in the southern Red Sea from new gravity evidence. In: Hilpert, L.S. (Ed.), Red Sea Research 1970–1975. Bulletin, 22. Saudi Arabian Ministry of Petroleum and Mineral Resources: DGMR, pp. K1–K11. Gettings, M.E., 1983. A simple Bouguer gravity anomaly map of southwestern Saudi Arabia and an initial interpretation. Saudi Arabia Deputy Ministry for Mineral Recourses: U.S. Geological Survey. Open File Report, USGS-OF-83-789. Gettings, M.E., Blank, H.R., Mooney, W.D., Healey, G.H., 1986. Crustal structure of southwestern Saudi Arabia. J. Geophys. Res. 91, 6491–6512. Girdler, R.W., Styles, P., 1974. Two stage Red Sea floor spreading. Nature 247, 7–11. Girdler, R.W., Underwood, M., 1985. The evolution of early oceanic lithosphere in the southern Red Sea. Tectonophysics 116, 95–108. http://dx.doi.org/10.1016/00401951(85)90223-9. Hall, S.A., 1989. Magnetic evidence for the nature of the crust beneath the Southern Red Sea. J. Geophys. Res. 94, 12, 267–12, 279. http://dx.doi.org/10.1029/JB094iB09p12267. Hatzfeld, D., Molnar, P., 2010. Comparisons of the kinematics and deep structures of the Zagros and Himalaya and of the Iranian and Tibetan plateaus and geodynamic implications. Rev. Geophys. 48, RG2005. http://dx.doi.org/10.1029/2009RG000304. Hempton, M.R., 1987. Constraints on Arabian Plate motion and extensional history of the Red Sea. Tectonics 6 (no. 6), 687–705. http://dx.doi.org/10.1029/TC006i006p00687. Jolivet, L., Faccenna, C., 2000. Mediterranean extension and the Africa–Eurasia collision. Tectonics 19, 1095–1106. Lazar, M., Ben-Avraham, Z., Garfunkel, Z., 2012. The Red Sea — new insights from recent geophysical studies and the connection to the Dead Sea fault. J. Afr. Earth Sci. 68, 96–110. Ligi, M., et al., 2011. Initial burst of oceanic crust accretion in the Red Sea due to edgedriven mantle convection. Geology 39, 1019–1022.

149

Lowell, J.D., Genik, G.J., 1972. Sea-floor spreading and structural evolution of Southern Red Sea. AAPG Bull. 56, 247–259. Makris, J., Henke, C.H., Egloff, F., Akamaluk, T., 1991. The gravity field of the Red Sea and East Africa. Tectonophysics 198, 369–381. Martinez, F., Cochran, J.R., 1988. Structure and tectonics of the northern Red Sea: catching a continental margin between rifting and drifting. Tectonophysics 150, 1–32. McClusky, S., et al., 2010. Kinematics of the southern Red Sea–Afar Triple Junction and implications for plate dynamics. Geophys. Res. Lett. 37, L05301. McQuarrie, N., Stock, J.M., Verdel, C., Wernicke, B.P., 2003. Cenozoic evolution of Neotethys and implications for the causes of plate motions. Geophys. Res. Lett. 30 (n. 20), 2036. http://dx.doi.org/10.1029/2003GL017992. Mechie, J., Prodehl, C., Koptschalitsch, G., 1986. Ray path interpretation of the crustal structure beneath Saudi Arabia. Tectonophysics 131, 333–352. Milkereit, B., Fluh, E.R., 1985. Saudi Arabian refraction profile: crustal structure of the Red Sea–Arabian shield transition. Tectonophysics 111, 283–298. Mooney, W.D., Gettings, M.E., Blank, H.R., Healy, J.H., 1985. Saudi Arabia seismic refraction profile: a travel time interpretation of crustal and upper mantle structure. Tectonophysics 111, 173–246. http://dx.doi.org/10.1016/0040-1951(85)90287-2. Okay, A.I., Zattin, M., Cavazza, W., 2010. Apatite fission-track data for the Miocene Arabia–Eurasia collision. Geology 38, 35–38. Pallister, J.S., et al., 2010. Broad accommodation of rift-related extension recorded by dyke intrusion in Saudi Arabia. Nat. Geosci. 3, 705–712. http://dx.doi.org/10.1038/ NGEO966. Reilinger, R., McClusky, S., 2011. Nubia–Arabia–Eurasia plate motions and the dynamics of Mediterranean and Middle East tectonics. Geophys. J. Int. 186, 971–979. Rihm, R., Makris, J., Moller, L., 1991. Seismic surveys in the Northern Red Sea: asymmetric crustal structure. Tectonophysics 198, 219–295. Rooney, T.O., Bastow, I.D., Keir, D., 2011. Insights into extensional processes during magma assisted rifting: evidence from aligned scoria cones. J. Volcanol. Geotherm. Res. 201, 83–96. http://dx.doi.org/10.1016/j.jvolgeores.2010.07.019. Sebai, A., Zumbo, V., Féraud, G., Bertrand, H., Hussain, A.G., Giannérini, G., Campredon, R., 1991. 40Ar/39Ar dating of alkaline and tholeiitic magmatism of Saudi Arabia related to the early Red Sea rifting. Earth Planet. Sci. Lett. 104, 473–487. Sultan, M., Becker, R., Arvidson, R.E., Sore, P., Stern, R.J., El-Alfy, Z., Guinnes, E.A., 1992. Nature of the Red Sea crust, a controversy revisited. Geology 20, 593–596. Verduzco, B., Fairhead, J.D., Green, C.M., MacKenzie, C., 2004. New insights into magnetic derivatives for structural mapping. Lead. Edge 23, 116–119. Voggenreiter, W., Hotzl, H., Mechie, J., 1988. Low-angle detachment origin for the Red Sea rift system. Tectonophysics 150, 51–75. Wernicke, B., 1985. Uniform-sense normal simple shear of the continental lithosphere. Can. J. Earth Sci. 22, 108–125. Westaway, R., 1994. Present-day kinematics of the Middle East and eastern Mediterranean. J. Geophys. Res. 99 (B6), 12,071–12,090. http://dx.doi.org/10.1029/94JB00335. Zeyen, H., Volker, F., Wehrle, V., Fuchs, K., Sobolev, S.V., Altherr, R., 1997. Styles of continental rifting: crust–mantle detachment and mantle plumes. Tectonophysics 278, 329–352.