Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

The Red Sea Rift is a unique spot on Earth to observe the transition from continental rifting to seafloor spreading. Along the 2,000 km long axis of the Red Sea, all stages from rifting to drifting can be examined (see Fig. 1). Because direct basement observations are prevented by area-wide evaporite sequences (e.g., Bosworth et al. 2005; Whitmarsh et al. 1974) that overly the crystalline basement, near surface structures above the evaporite layers are investigated in order to get information about the inherent basement dynamics. Ocean deeps, that is, isolated pools with a size of some tens of square kilometers and depths of few hundreds of meters that penetrate the seafloor within the Red Sea are considered to be an accompanying phenomenon of the process from continental rifting to seafloor spreading (Bonatti 1985; Cochran and Martinez 1988; Martinez and Cochran 1988) which makes the meaning of the ocean deeps important in order to understand the plate tectonic setting of this region. Many Red Sea ocean deeps are associated with magmatism, either directly observed as volcanoes or indirectly as magnetic dipole anomalies. These deeps were suggested to represent initial seafloor spreading cells (Bonatti 1985). Taking into account the thick sedimentary cover that generally characterizes the entire Red Sea basin and especially the massive Miocene evaporite layers in the main trough (Cochran et al. 1991), the analysis of surficial structures such as the Red Sea deeps represents one way to extract information about the status of the rifting and the nature of the buried and masked northern Red Sea crust.

Table 1 Main parameters for northern Red Sea deeps derived mainly from high-resolution multi-beam bathymetric, magnetic, and MCS data sets
Full size table
Fig. 1
figure 1

The Red Sea can be separated along its strike into three provinces. The southern province is characterized by organized seafloor spreading, the central part is in a transition stage, and the northern province of the Red Sea is most probably in the late stage of continental rifting (Cochran et al. 1986). Red dots mark the position of prominent Red Sea deeps. Old tectonic lineaments such as the Neoproterozoic Pan-African Najd fault system terminate against the Red Sea rift (Bosworth et al. 2005)

Full size image

In this paper, we compare area-wide geophysical data such as magnetic data and satellite altimetry data with regionally acquired multi-channel seismic and hydro-acoustic data from the northern Red Sea oceanic deeps (Fig. 1). First, we review what we know about the Conrad Deep (CD) published by Ehrhardt et al. (2005). The CD as well as the Shaban Deep (SD) is located in the NW–SE segment of the northern Red Sea. The “magnetic quiet” N–S segment hosts the Kebrit Deep (KD). From these individual studies, we infer the genesis of the deeps, associated with salt tectonics and fluid flow.

Geological Setting

For the geological setting of the Red Sea, we refer to Bosworth (this volume). In this chapter, we focus our attention on the analysis of the Red Sea deeps. Since the 1960s, several ocean deeps were discovered along the axis of the Red Sea (Bäcker and Schoell 1972; Degens and Ross 1969; Miller et al. 1966). Because of size, age, and formation history, we distinguish between central and northern Red Sea deeps. The larger central Red Sea deeps are located in the axial trough and are separated by inter-trough zones (Cochran and Martinez 1988; SearIe and Ross 1975). They are floored by young basaltic crust and exhibit magnetic anomalies not older than 1.7 Ma. The northern Red Sea deeps are smaller and form only isolated deeps within the axial depression. Some of them are accompanied by volcanic activity. Many Red Sea deeps contain bottom-water brines and metalliferous sediments pointing to hydrothermal circulation of seawater (Botz et al. 2011; Hartmann 1980; Scholten et al. 1991; Winckler et al. 2000). The largest and most prominent deep is the Atlantis II Deep, located in the central part of the Red Sea in the vicinity of other large deeps such as the Chain Deep and Discovery Deep. Other prominent deeps are the Tethys and Nereus Deeps further north, but still in the central part of the Red Sea.

Transition from Rifting to Seafloor Spreading

According to its development, the Red Sea can be separated into three parts from south to north (Cochran et al. 1986). In the southern part, seafloor spreading since ca. 5 Ma forms a continuous mid-ocean ridge system (MOR) (Röser 1975). The inception of spreading becomes progressively younger to the north. In the central Red Sea, the MOR becomes discontinuous forming single deeps (oceanized cells) that are separated by inter-trough zones (SearIe and Ross 1975) where the spreading commenced not earlier than 1.7 Ma (Cochran et al. 1986). In the northern Red Sea, only isolated smaller deeps are distributed along an axial depression (Table 1 and Fig. 1). The axial depression is the northward prolongation of the axial trough of the central Red Sea. It is a characteristic feature of the northern Red Sea and bisects the main trough with water depths of about 1,200 m. The main trough is covered in the most areas of the Red Sea with huge amounts of Miocene evaporites with an average thickness of about 3,000 m (Girdler and Styles 1974; McKenzie et al. 1970; Mitchell et al. 1992; SearIe and Ross 1975) that mask the underlying basement on seismic data. The top of the evaporites is seismically imaged as a prominent reflector that was named S-Reflector (SearIe and Ross 1975; Whitmarsh et al. 1974) which can be identified easily in most areas of the Red Sea. The alignment of the coastal shelves, the main trough, and the axial depression is shown in Fig. 2.

Fig. 2
figure 2

3D block diagram showing northern Red Sea bathymetry (data from GEBCO) and characterization of seafloor morphology along an east–west cross section. The positions of the surveyed deeps are marked with white circles

Full size image

Opening of the Red Sea became increasingly focused to the axial depression that represents the tectonically most active part of the rift (Cochran et al. 1991).

Segmentation of the Northern Red Sea

Although the Red Sea is still in the transition from continental rifting to seafloor spreading, it is already possible to observe segmentation along the rift. The bathymetric trend of the axial depression separates the northern Red Sea into segments; one NW–SE segment from 25°N to 28°N and one N–S-trending segment from 24°N to 25°N (Fig. 3). The N–S segment is also known as the Zabargad Fracture Zone (Bonatti et al. 1984; Marshak et al. 1992). The N–S segment links the NW–SE segment in the north with another NW–SE segment in the south so that the axial depression forms a kind of en-echelon pattern (Figs. 1 and 3). Segmentation along passive continental margins is also observed in Baffin Bay (Skaarup et al. 2006) where a sequence of volcanic and non-volcanic segments was identified. Similar segmentation is also known from the South Atlantic Ocean (Franke et al. 2007).

Fig. 3
figure 3

a Bathymetric map (from satellite altimetry; Sandwell and Smith 1997). b magnetic map of the northern Red Sea (after Coutelle et al. 1991). Isolated magmatic dipole anomalies are interpreted as igneous intrusions into the basement. Strong magnetic anomalies are found at the Shaban and Conrad Deeps as well as Oceanographer, Mabahiss and unnamed deeps (red dots). Other deeps are not associated with strong magnetic anomalies such as Kebrit and Vema Deeps (black dots)

Full size image

Two of the examined northern Red Sea deeps (Conrad and Shaban Deeps, CD and SD) are associated with magnetic anomalies. This indicates the presence of highly magnetic rocks, which are interpreted as magmatic intrusions and extrusions (Fig. 3 and Table 1). The Kebrit Deep (KD) is located in the N–S-oriented segment and is not correlated with distinct magnetic anomalies (Table 1, Figs. 1 and 3). This segment represents a magnetic quiet zone in contrast to the more magnetic segment to the north. The southern adjoining segment comprises deeps with pronounced rift structures and initial spreading activity, for example the Nereus Deep (Antonini et al. 1998). As the northern part of the Red Sea rift is close to the stage of seafloor spreading and the extension already is focused to the axial depression, the heat flow is anomalously high. Heat flow between 250 and 350 mW/m2 was measured in the axial depression (Martinez and Cochran 1988), which is about 6–10 times higher than the world mean (Makris and Rihm 1991). This high heat flow is further evidence of recent igneous activity which probably was also responsible for hydrothermal activity. These hydrothermal systems cause the deposition of the abundant metalliferous sediments in and near the deeps and the hyper-saline brine bodies within the deeps (Cocherie et al. 1994).

The Northernmost Red Sea Deeps

The CD and SD are located in the NW–SE-trending segment of the northern Red Sea (Figs. 1 and 3) and were discovered in 1986 and 1984 (Cochran et al. 1986; Pautot et al. 1984), respectively. The CD is one of the northernmost Red Sea deeps. Only one deep is reported north of the CD (at 27°19′N, 35°23′E), but without any significant magnetic anomaly (Guennoc et al. 1988).

The CD was discovered during a multi-component survey, including seismic, hydro-acoustic, magnetic, gravity, and heat flow measurements, by the R/V Conrad in 1984 (Cochran et al. 1986). The CD is conspicuous because of its elongated N20°E-trending shape and the fact that it is located between two magnetic dipole-type anomalies NE and SW of it. The strike of the deep is almost parallel to the strike of the Gulf of Aqaba and the Dead Sea Transform, although the location of deep is a southward prolongation of the Gulf of Aqaba (Figs. 1, 2, and 3). Cochran et al. (1986) modeled two intrusion bodies in order to explain the anomalies and dated the emplacement of the intrusions to be only 40,000 years ago. Based on the high-resolution multi-channel seismic and multi-beam bathymetric data, Ehrhardt et al. (2005) linked the trend and shape of the CD to deep-rooted transtensional tectonics. They further suggested that the enhanced heat transfer between the two magmatic intrusions lowered the viscosity of the Miocene evaporites, thus supporting the evolution of the deep. Since heat flow values change rapidly and maximum values of 605 mWm−2 were reported from CD area (Cochran et al. 1986). Ehrhardt et al. (2005) suggested hydrothermal circulation and salt dissolution as the source of the brine.

The southeastern part of the SD was discovered in the early 1980s by the Preussag cruises (see e.g., Bäcker et al. 1975), but the SD in its entire extent was mapped in 1984. It is named after the Arab month of Shaban and is sometimes also termed the Jean Charcot Deep. The SD is more or less rhombic in shape and associated with a dipole-type magnetic anomaly that is caused by a magmatic extrusion (Pautot et al. 1984). This volcanic edifice forms an elongated NW–SE-trending ridge in the center of the deep. The SD is split up into four sub-basins that are filled with hyper-saline brine. Geochemical analyses of the brine water and the surrounding sediments result in contradictory interpretations of its formation. While Blum and Puchelt (1991) suggest that hydrothermal circulation was responsible for brine formation, Cocherie et al. (1994) analyzed REE signatures of SD metalliferous sediments and concluded that the SD brine water is seawater dominated, established by mainly lateral dissolution of Miocene evaporites.

The KD, located in the N–S segment of the northern Red Sea (Figs. 1, 2 and 3), was discovered in 1974 (Bäcker et al. 1975) and was revisited several times in order to provide long-term datasets of the brine body (Blum and Puchelt 1991; Hartmann et al. 1998; Winckler et al. 2001). Its name (Kebrit is Arabic for sulfur) points to hydrothermal activity. Blum and Puchelt (1991) discovered black smokers at the flanks of the deep and sampled them. Some KD chimneys consisted of pure sulfur. The KD is oval shaped and includes a hyper-saline brine body (e.g., Hartmann et al. 1998). Smooth slopes of the deeps are intersected by graben- or half-graben-like structures. The origin of the KD hyper-saline brine body is also controversial, as Blum and Puchelt (1991) propose hydrothermal circulation, but Winckler et al. (2001) prefer that the brines formed by the lateral solution of Miocene evaporites. At the link between the NW-SE segment and the N–S segment the largest northern Red Sea deep, the Mabahiss Deep developed. It is associated with a high amount of magmatic activity (Guennoc et al. 1988).

Methods and Observations

Our study is based on high-resolution multi-channel seismic, multi-beam and parametric echosounder data collected during R/V Meteor Cruises M44/3 and M52/3 in 1999 and 2002. Three northern Red Sea deeps (CD, SD, and KD) were investigated (Ehrhardt and Hübscher 2003; Hübscher et al. 2000). The close spacing of the seismic lines led to complete bathymetric mapping of the seafloor by the swath echosounder and enabled interpolation between the 2D-seismic lines (Fig. 4). These data are discussed together with previously published magnetic data (Coutelle et al. 1991) and gravity inferred from satellite altimetry data (Sandwell and Smith 1997).

Fig. 4
figure 4

Base maps of the multi-channel seismic surveys in the areas of the Conrad Deep (a), Shaban Deep (b) and Kebrit Deep (c). The contour maps show where deeps have been imaged by multi-channel seismic reflection data (MCS). Note the dense MCS line spacing with an average distance of less than 2 km

Full size image

The three northern Red Sea deeps show major differences in size and shape and in the accompanying geophysical anomalies. The bathymetry of the three northern Red Sea deeps is displayed in Figs. 4, 5, and 6. Selected 2D lines will illustrate the subsurface structure of the deeps and their surroundings. The main parameters regarding the deeps are listed in Table 1.

Fig. 5
figure 5

Bathymetry map of the Conrad Deep (CD) area. The CD is located between the slopes of the axial depression. The southwestern slope is characterized by volcanic extrusions (see V/E). Volcanic extrusions on the northeastern slope are not imaged but the magnetic data give strong evidence for the presence of a volcanic/magmatic body (see Fig. 3). Note the second deep (Klauke Deep) west of the CD. Red lines and corresponding annotations show the location of the following seismic lines. The right-hand figure shows a sketch of the bathymetry with all structures quoted before (CD Conrad Deep; KD Klauke Deep; SW Salt wall; SR Salt rise; V/E Volcanic extrusion; SL Streamlines)

Full size image
Fig. 6
figure 6

Left MCS reflection line GeoB99_072, NW–SE cross section through the Conrad Deep (see Fig. 5). Right interpretation of MCS profile, including pre-kinematic (yellow) and syn-kinematic (green) units, was identified SE of the CD. The pre-unit reveals concordant reflection patterns; the syn-unit (green) is characterized by onlap structures and divergent reflection patterns. Both units are cut by a vertical transparent unit (salt wall). Basement-induced salt tectonics are a good explanation for the elongated salt rise

Full size image

Conrad Deep

The CD is the northernmost Red Sea deep and is located within the axial depression of the northern Red Sea (Figs. 3 and 5). The elongated deep is 10.5 km long and 2 km wide and has an overall strike of N20°E, which is sub-parallel to the trend of the Gulf of Aqaba and thus the Dead Sea Transform (cf. Figs. 3 and 5). It comprises two sub-basins showing a small left stepping offset. Its maximum depth is 1,520 m which is approx. 200 m deeper than the surrounding seafloor. Its slopes are steepest on the longitudinal sides and distributed in an en-echelon pattern. The Klauke Deep to the west is approximately 150 m deeper than the inter-deep area. Circular and elongated volcanic edifices SW of the CD are grouped along the western flank of the axial depression. The appearance of the seafloor in the vicinity of the CD is different from NW to SE. That to the NW shows a gentle stepwise decrease of water depth from south to north and is obviously affected by the presence of the Klauke Deep. That to the SE shows two parallel to sub-parallel ridge-like structures, called the salt rise and salt wall (Fig. 5).

The seismic data resolve the uppermost Miocene evaporites and the Pliocene–Quaternary (PQ) overburden (Figs. 6, 7, 8 and 9). PQ sediments are as thick as 150–250 m which is typical for the Red Sea (Guennoc et al. 1988). The S-reflection corresponding to the top of the evaporites is clearly visible in all seismic lines. The S-reflector is not continuous across the CD. It terminates against the CD and the salt wall, respectively. On its southeastern side, the S-reflector is about 200 ms “deeper” compared to the northwestern side. The longitudinal margins of the CD are bounded by normal faults with inclination angles of 16° and 11°, respectively.

Fig. 7
figure 7

Left MCS reflection line GeoB99_073, NW–SE MCS profile across the Conrad Deep (see Fig. 5). Right Interpretation of pre-kinematic (yellow) and syn-kinematic (green) units occurs SE of the deep. The salt wall builds the SE slope of the deep and forms a backstop for Plio-Quaternary sediments

Full size image
Fig. 8
figure 8

Left MCS reflection line GeoB99_075, NW–SE MCS reflection profile across the Conrad Deep (see Fig. 5). Right Interpretation of MCS profile. Pre-kinematic (yellow) and syn-kinematic (green) units were identified SE and NW of the deep. SE slope collapsed without the salt/mud wall as backstop. On the NW side, a high-amplitude reflector (HA) is located next to the layered Plio-Quaternary sediments

Full size image
Fig. 9
figure 9

Top MCS reflection line GeoB99_079, SW–NE cross section (see Fig. 5). Bottom interpretation of MCS profile, including pre-kinematic (yellow) and syn-kinematic (green) units on the SE area of the deep. The seismic line is parallel to the Red Sea extension

Full size image

We identified the following characteristics of the CD:

  • The Pliocene–Quaternary supra-salt succession is subdivided into parallel layered lower and an upper succession which shows wavy and divergent internal reflections.

  • Elongated salt rises—both straight and arcuate—create topographic ridges.

  • The salt wall emerges from the evaporite sequence. It pierces the seafloor and is not covered by PQ sediments.

  • High-amplitude reflector is located close to the CD (Fig. 8).

  • At the SW tip of the CD, close to the slope of the axial depression, volcanic extrusions are observed in both seismic records and the bathymetry (Figs. 5 and 9).

  • The reflection characteristics from the evaporites beneath the S-reflection change laterally from well-stratified to reflection-free (Fig. 8).

Pliocene–Quaternary succession: The appearance of Plio-Quaternary sediments varies from well-layered (Fig. 8, southeastern side) to chaotic (Fig. 6, northwestern side). Based on this seismic character, a subdivision into two seismic units is possible, pre-kinematic, and syn-kinematic. On the SE side of the deep, over the elongated salt rise, the change from a divergent to a concordant reflection pattern marks the onset of the salt tectonic deformation that led to the formation of the salt rise. The distinction between both units corresponds probably to the Units I and II sampled during DSDP Leg XXIII sites 225, 227, and 228 (Stoffers and Ross 1974).

Salt Rise: The salt rise on the SE side of the CD (Figs. 5, 6, 7, 8 and 9) strikes SW–NE and parallel to the strike of the CD. This rise structure could be the result of lateral salt flowage over a step in the underlying basement. Although the seismic data do not show evidence of the basement, we interpret these salt rises as the consequence of salt flow above basement faults in analogy to numerous instances from the Permian Central European Basin System where Zechstein salt ridges are frequently underlain by basement faults (see Warren 2006 and references therein).

The onset of the formation of the salt rise is marked by the dividing unconformity in the overlying Plio-Quaternary deposits. It is important to note that the age of the unconformity only refers to the development of the rise and not to the opening of the CD, other than previously explained by Ehrhardt et al. (2005). If the seismic units divided by the unconformity can be correlated to Units I and II of DSDP Leg XXIII sites 225, 227, and 228 (Stoffers and Ross 1974), then the unconformity is of Pliocene age. Thus, the deformation significantly predates the emplacement of magmatic intrusions 40,000 years ago (Cochran et al. 1986) and therefore the opening of the CD.

Salt Wall: The salt wall (Figs. 6 and 7) is conspicuous as it trends oblique to the CD and to the salt rise and it cuts through the entire Pliocene–Quaternary sediments. The seismic lines and the bathymetry (Fig. 5) reveal that the salt wall strikes sub-parallel to the strike of the CD; thus, it intersects the deep, building the SE slope of the deep in its southern sub-basin. The salt wall acts obviously as backstop for the Plio-Quaternary sediments. Where this backstop is absent, the slope is not stabilized and mass wasting into the deep occurred along listric fault planes (Fig. 8).

An evolutionary model which could explain such a feature by hydrothermally driven ascent of hot brines is presented by Hovland et al. (this volume). Hydrothermal circulation above basement faults started and the salt wall was created due to salt precipitation out of the expelling fluids. The general possibility of hydrothermal circulation of salty fluids had been suggested before by Hovland et al. (2006) and was confirmed by laboratory- and molecular-scale dynamic simulations (Hovland et al. 2006). The occurrence of hydrothermal liquids is supported by the report of unusual high heat flow values (605 mW m−2) at the northern end of the deep (Cochran et al. 1986).

Arcuate salt rises: The arcuate shape of salt rises and topographic ridges that terminate obliquely against the CD slopes are not related to deep-rooted faults. Mitchell et al. (2010) presented multi-beam bathymetric data from the central Red Sea which showed that salt flows downslope similar to ice-glaciers. Talbot and Pohjola (2009) documented the similarities between ice streams and salt glacier dynamics. Consequently, we see these salt rises and the corresponding bathymetric ridges as streamlines which emerge sub-parallel to the flowing direction.

High-amplitude Reflector: Consistent to this model is the observation of a high-amplitude reflector (HA, see Fig. 8b).

The high-amplitude reflector is not phase-reversed implying a strong positive acoustic impedance contrast caused by precipitated salt. Similar processes were observed by Netzeband et al. (2006) in the Mediterranean. Hydrothermal circulation caused salt precipitation near the seafloor (because of cooling) and the development of salt bodies or salt filled pore volumes that increased the bulk density and interval velocity.

Volcanic Extrusions: Southwest of the CD at the slope of the axial depression several volcanic extrusions were mapped by seismic (Fig. 9), bathymetric (Fig. 5) and magnetic methods (Fig. 3). Probably the volcanic extrusions ascended along the normal faulted slope of the axial depression. Similar volcanic bodies are likely at the northeastern end of the CD, also at the slope of the axial depression. However, they are not imaged by seismic data and only magnetic data are available as evidence (Fig. 3).

Lessons learned from the CD: The CD is located within the axial depression of the northern Red Sea. The deep disrupts the top of the Miocene evaporites, but it does not reach crystalline basement. In comparison with the central Red Sea deeps, it is in a very juvenile stage. However, the position of the CD is not arbitrary; it developed above the junction of two tectonic lineaments, the NNW–SSE Red Sea Rift and an oblique, almost perpendicular (parallel to the Aqaba trend) transform fault. Magmas used this zone of weakness to ascend. In Late Pleistocene time, magma ascended at the margins of the axial depression to form both intrusions and volcanoes. Temperature radiation decreased the salt viscosity between the intrusions. As a result of the ongoing rifting of the Red Sea, a transtensional regime was established between both magmatic intrusions/extrusions causing the opening of the CD (Ehrhardt et al. 2005). Salt flow above the basement faults formed the salt rises. Salt flow toward the axial depressions formed streamlines analogous to glaciers.

Shaban Deep

The 6 by 5 km wide SD is located in the central axial depression (Fig. 10). In the center of the SD, a volcano built an elongated NW–SE ridge, rising to a level of about 900 m bsl. The SD is subdivided into four sub-basins. The two northern sub-basins are shallower, with maximum water depths of about 1,400 m, and the southern sub-basins are deeper with maximum water depths of more than 1,600 m. The northwestern escarpment consists of a prominent, about 100-m-high rim anticline. The seafloor north of the SD is smoother than that to the south. The eastern side is also affected by two major grabens, striking NE–SW. The transition from the smooth northern domain of the SD to the southern deeper situated domain can be traced across the SD where the difference in depth is reflected by the depth of the sub-basins.

Fig. 10
figure 10

a High-resolution bathymetric map of the Shaban Deep (SD). Red lines indicate seismic reflection profiles in Figs. 11, 12, 13 and 14. The SD is characterized by the volcano in the center of the deep. Note the difference in depth between the northern and southern SD. A WSW–ENE-trending graben seems to cross the SD. The northern part of the SD is bounded by an anticline. The south is influenced by SW–NE-trending graben- or valley-like structures that coalesce with the WSW–ENE graben. b The sketch illustrates the mentioned structures. Note the dashed lines showing the two trends normal to each other. SL Streamline

Full size image

The bathymetry reflects two main tectonic trends. The NW–SE trend, parallel to the axial depression of the Red Sea and the NE–SW trend that crosscuts the SD. This trend builds a smaller ridge that is perpendicular to the volcanic ridge and several elongated NE–SW ridges and valleys in the east of the SD.

The sub-surface structure of the SD area is best seen on the NNW–SSE MCS reflection lines (Figs. 11, 12, 13 and 14). The typical thickness of Plio-Quaternary sediments, around 150–250 m, is equivalent to 200–300 ms TWT and covers the evaporites. Similar to the CD region, the supra-salt succession reveals two sequences with different reflection characteristics. The S-reflection was identified on all lines as marking the top of the Miocene evaporites. It is significantly deeper in the south compared to the northern margin of the SD. In the north, it terminates against the slope approximately at 1.7 s TWT. A brine reflector is detected within all four SD sub-basins and in similar water depths. East of the SD, a low frequency reflection probably images the base of the Miocene evaporites (Figs. 13 and 14). The two-way travel time is comparable to base reflections known from old industry data (Preussag data from the 1980s; Wiedicke-Hombach, pers. comm.). Similar to the CD area further north, two units of the Pliocene–Quaternary supra-salt succession and the lateral variation of the evaporites, well-stratified or transparent, are observed. The rim anticline is a peculiar structure that looks similar to the CD salt wall.

Fig. 11
figure 11

Top MCS reflection line GeoB99_085, NNW–SSE section crossing the western extension of the SD. The area experiences N–S extension. Bottom Interpretation of MCS profile. Extensional faulting is identified in the Plio-Quaternary sediments. The anticline structure at the northern rim of the SD is probably caused by hydrothermal circulation similar to that seen in the CD

Full size image
Fig. 12
figure 12

Top MCS reflection line GeoB99_088, NNW–SSE section crossing the western part of the SD (see Fig. 10). Bottom Interpretation of MCS profile. N–S extension caused block rotation along normal faults in the area of the SD. The Plio-Quaternary sediments are rather thin with two identified units

Full size image
Fig. 13
figure 13

Top MCS reflection line HH02_45, NNW–SSE section crossing the SD and the volcanic cone. Bottom Interpretation of MCS profile. The ascent of the magmatic body is associated with N–S extension (see Fig. 10). The brine level is equal between the different sub-basins

Full size image
Fig. 14
figure 14

Top MCS reflection line HH02_47, NNW–SSE section crossing the eastern part of the SD (see Fig. 10). Bottom Interpretation of MCS profile. Extensional normal faulting is complemented by strike-slip faulting in the southern part

Full size image

The following features characterize the SD:

  • A central volcanic ridge that strikes along the axial rift.

  • A perpendicular SSW–NNE striking ridge across the SD.

  • The northern sub-basins reveal a rim anticline.

  • Straight and arcuate grabens south and southeast of the SD.

  • Topographic ridges along the rift axis.

  • The reflection characteristics from evaporites beneath the S-reflection change laterally from well-stratified to transparent (Fig. 11).

Volcanic Ridge: The NW–SE ridge comprises a volcanic extrusion (Haase et al. 2000; Pautot et al. 1984) that causes a magnetic dipole anomaly (Fig. 3). The ridge crosses the SD from NW to SE and rises in the center to a water depth of 900 m (Fig. 10). Seismic lines only touch the volcano (Fig. 13) but show clearly the western slope of the volcanic cone. The shape of the volcano is narrow and elongated. No lava flows could be observed; thus, it is unclear how long it was active. Basaltic rock fragments found within the SD point to a young volcanic edifice. As the volcano is limited to the area of the SD and according to the magnetic map (Fig. 3), no other volcanic or magmatic activity is present in the vicinity, an interrelation between the emplacement of the volcano and the development of the SD is obvious.

Perpendicular Ridge: A smaller SW–NE-trending ridge crosses the SD (Fig. 10). The volcano and the perpendicular ridge subdivide the SD into four sub-basins. The nature of the perpendicular ridge is unclear. Interpretation of the MCS reflection data suggests that the ridge is an elevated edge of a rotated block (Fig. 12) or may be a step from the northern shallow sub-basins to the southern deep sub-basins. The extensional normal faulting and the stepwise descent of the S-reflection (i.e., the top of the Miocene evaporites) gives evidence of a basement fault perpendicular to the Red Sea rift axis, and it is likely that this junction of tectonic lineaments localized the volcanic activity.

Rim anticline: The northern margin of the SD is a prominent elongated structure forming a rim anticline (Fig. 10). This diapir-like or wall-like structure is sometimes very pronounced (Figs. 11 and 12) and is sometimes subdued and part of the slope. In any case, it serves as backstop similar to the salt wall of the CD and prevents Plio-Quaternary sediment and the Miocene S-reflector from terminating into the deep. Two scenarios explain the conspicuous rim: (a) Similar to the discussion of salt flow in the CD area, we take these observations as evidence for southward-directed salt creep (Fig. 15). The elongated ridges represent streamlines, and salt thickening due to shortening creates the rim anticline. South of the deep, the salt flows away from the SD causing salt withdrawal and therefore a decreased slope angle; (b) A diapir or salt wall similar to the CD is formed by hydrothermal circulation and salt re-precipitation (Figs. 11, 12, 13 and 14).

Fig. 15
figure 15

Sketch of a N–S section across the SD explaining the bulge-like salt wall and the decreased slope angle in the south by lateral salt flow from north to south along the axial depression

Full size image

Straight and arcuate graben: The eastern and southeastern margin of the SD is transected by several SW–NE-trending graben structures (Fig. 10). Seismic lines across these grabens (Fig. 14) show thick and faulted Plio-Quaternary sediments. The arcuate graben (Figs. 10 and 14) is probably the result of an underlying strike-slip fault that compensates SW–NE transform motion.

Lessons learned from Shaban Deep: Similar to the CD, the SD is located at a junction of the rift axis and a perpendicular fault. While the CD emerged between two volcanoes, the SD developed around a central volcanic ridge. We suggest that hydrothermal circulation associated with the ascent of magmas dissolved the salt and opened the SD. Additionally, the preexisting SW–NE-trending fault plays a vital part which is depicted in a sketch (Fig. 16). The left side shows the situation before the onset of magmatic activity. The SW–NE-orientated fault zone was most probably a strike-slip fault with an additional normal component of faulting (in order to explain the different depths between the northern and southern parts). The right side of Fig. 16 illustrates the development of the SD in the successive stages of the ascent of the magmatic body. The intrusion used the preexisting fault for its ascent (Fig. 16a). Because of the lateral replacement of the sediments by the magmatic cone, an extensional regime perpendicular to the strike of the fault, that is, in NW–SE direction was initiated (Fig. 16b). Heat from the magma and associated hydrothermal circulation lowered the viscosity of the surrounding evaporites and remobilized these. The strike-slip displacement along the preexisting fault was blocked by the volcano; thus, it side-stepped to the south around the barrier (Fig. 16d). This scenario could explain the SW–NE striking fault that is separated from the volcano, the circular shape of the deep, and the faulted SE-part of the survey area. Similar to the CD, rift-related extension, intersecting fault systems, and the emplacement of magmatic intrusion/extrusion are the main causes forming the deep.

Fig. 16
figure 16

Cartoon of possible development of the SD. Left Bird’s-eye view from the south on a preexisting strike-slip normal fault (red) that is crossing the Red Sea rift axis (yellow). Right side Cross section sequence illustrating the development of the SD. See the location of the red line for orientation

Full size image

Kebrit Deep

Compared to the CD and SD, the KD (KD; Fig. 17) has a more oval and asymmetric shape. It is located in the Red Sea axial depression but in contrast to the CD and SD, the KD lies within the N–S-trending segment (see Fig. 3). It has smooth slopes that are intersected by two W–E-trending grabens and half grabens, and a N–S-trending graben that crosses the deep but is slightly offset to the east. All these grabens coalesce in the center of the KD comparable to a star-shaped pattern. It is reasonable that this is the expression of a through-going N–S-trending fault as the surface expression of the rift axis. The morphology of the surrounding seafloor west of the N–S-trending graben is regular and smooth but irregular and rough east of it.

Fig. 17
figure 17

a High-resolution bathymetric map of the Kebrit Deep (KD). Red lines mark the position of the seismic lines shown in Figs. 18, 19, 20, and 21. The KD has an oval shape with grabens emanating from the deep in a star-shaped pattern. b Sketch of the main structures of the Kebrit Deep. Black dashed lines star-shaped pattern. The N–S lines are slightly offset to the east. A circular structure (C) in the northeast overtops the surrounding of the KD

Full size image

The KD differs in many details from the CD and SD. Apart from size, shape, and location, the internal structure also shows major differences (see Table 1).

The following features characterize the KD:

  • Star-shaped pattern.

  • Smooth slopes, no escarpment-like margin.

  • Prominent differences between western and eastern provinces as well as seafloor to the north and south.

  • The reflection characteristics of evaporites beneath the S-reflection changes laterally from well-stratified to transparent.

Star-shaped pattern: There are four main grabens that radiate from the center of the KD. The two W–E-trending grabens developed because of subsidence due to subrosion (salt dissolution) (Fig. 21). The bathymetric map (Fig. 17) shows that graben depth decreases toward the western end of the survey area and the grabens terminate just to the west. The N–S-trending grabens are through going and it is possible that these grabens are the surface expression of a strike-slip fault.

Smooth slopes: The slopes into the KD are smooth and have no escarpment-like margins (Figs. 18 and 19). The entire sedimentary succession bends down into the deep. The prominent brine layer could only be a result of hydrothermal circulation (Fig. 20).

Fig. 18
figure 18

Left MCS reflection line HH02_74, NW–SE section crossing the KD (see Fig. 17). Right Interpretation of MCS profile. Note the thick Plio-Quaternary sediment sequences that bend down into the deep. The Miocene S-reflector does not crop out in the deep. The NW corner of the KD comprises an abundant sequence of concordant Plio-Quaternary sediments (approx. 600 m). The deep represents a collapse structure, most likely due to dissolution of the Miocene salt beneath. In the NW, the evaporites are well-stratified but transparent to the SE

Full size image
Fig. 19
figure 19

Left MCS reflection line HH02_76, W–E section crossing the southern margin of the KD (see Fig. 17). Right Interpretation of MCS profile. Western and eastern sides are separated by a graben. Both sides show an anticline of the Miocene salt, but the Plio-Quaternary is well-layered on the western side and chaotic on the eastern side. It is possible that the graben represents the surface structure of a strike-slip fault. The reflection pattern from the evaporites varies from west to east

Full size image
Fig. 20
figure 20

Left MCS reflection line HH02_63: W–E section crossing the northern margin of the KD (see Fig. 17). Right Interpretation of MCS profile. Note the different appearance between the concordant layered sediments on the western side and the faulted and folded sediments on the eastern side. The intervening graben is a surface expression of a strike-slip fault. The evaporites are layered

Full size image

Prominent differences between western and eastern provinces as well as north and south: The seismic lines clearly show a different appearance of the NW, SW, SE, and NE quadrants of the KD. The NW quadrant shows regular layered sediments on top of the Miocene evaporites. This thick succession bends down into the deep (Fig. 18). The western side is transected by two grabens that compensate the subsidence of the KD (Fig. 21). On the SW quadrant, a salt anticline is responsible for a considerable thinning of the Plio-Quaternary sediments, which are still well-layered. Toward the SE quadrant, the N–S-trending graben separates the well-layered sediments from underlying chaotic structures above a high-amplitude reflector (see Fig. 19) (Bosworth and Strecker 1997). Below this reflector, well-stratified sediments cover the S-reflector. The NW quadrant is dominated by an arcuate anticline with a diameter of 2.5 km caused by vertical salt mobilization. High heat flow and hydrothermal circulation may have remobilized the salt. To the west, separated by the N–S graben, there is a sharp transition to the well-layered, concordant sediments of the NW quadrant.

Fig. 21
figure 21

Left MCS reflection line HH02_75, N–S section crossing the western margin of the KD (see Fig. 17). Right Interpretation of MCS profile. The northern side shows well-layered, slightly divergent sediments. Two graben structures separate the thick Plio-Quaternary sediments from thinner sediments in the south. Vertical salt tectonics influenced the southern side. The evaporites show a southward divergent reflection pattern in the northern segment

Full size image

A clear brine reflection is visible within the KD. There is no contact between the brine layer and the Miocene evaporites; thus, a lateral dissolution of the evaporites as proposed by Winckler et al. (2000) can be ruled out. The accumulation of the hyper-saline brine body is likely due to hydrothermal circulation. No evidence for volcanic activity is seen in the seismic data. The magnetic map of Coutelle et al. (1991) shows only slight magnetic anomalies (Fig. 3).

Lessons learned from the KD: Although the three mentioned deeps (CD, SD, and KD) are all northern Red Sea brine deeps, they have little in common. The KD is not associated with magnetic anomalies and thus is probably not affected by magmatic intrusions (Fig. 2 and Table 1). Its internal structure suggests that it is a collapse structure, as argued by Bäcker et al. (1975). The most conspicuous difference, in comparison with the other deeps described above, is the smooth and continuously dipping slopes without scarps. The entire Plio-Quaternary sequence and the top of the evaporites are affected by the subsidence. The cause of the subsidence is linked with dissolution of the Miocene evaporites, because of hydrothermal circulation. A basement fault system such as the N–S graben could have initiated the thermal activity. It is conceivable that this position had an impact on the development of the hydrothermal circulation cell and thus supported the formation of the KD. The circulation of seawater dissolved the evaporites beneath the deep and caused continuous collapse of the overlying sedimentary cover. Hydrothermal circulation and the absence of a magmatic body is also assumed by Blum and Puchelt (1991) who analyzed the deposits of the KD as being of low temperature origin. These facts support the approach that subrosion caused the development of the KD without any significant alteration of the ductile properties of the evaporites.

The KD developed most likely as a collapse structure due to hydrothermal circulation and subrosion (salt dissolution) of the evaporites below the deep. Thus, it developed independently of volcanic extrusions/intrusions and the tectonic activity in the axial depression. Because of the higher heat flow in the center of the Red Sea (axial depression) (Cochran et al. 1986; Martinez and Cochran 1988), it is likely that the necessary hydrothermal circulation cell was established in the axial depression.

Discussion

Using the morphology, internal structure, associated magnetic anomalies and the location of the northern Red Sea deeps, we are able to determine to types of deeps with regard to their development. The first type is based on the CD and SD, representing deeps in the NW–SE segment of the northern Red Sea, and the second is based on the KD, located in the N–S segment of the northern Red Sea (Table 1, Figs. 1, 2 and 3).

Segmentation-Controlled Development

The investigated Red Sea deeps were grouped into intrusion-related deeps (CD and SD) and collapse-type deeps (KD). This classification correlates with the location of the deeps. The intrusion-related deeps are in the NW–SE segment of the northern Red Sea and the collapse-type KD is found in the N–S segment. The magnetic anomaly map (Fig. 3) shows more anomalies in the NW–SE segment. In addition to the already discussed SD and CD, Martinez and Cochran (1988) mapped an unnamed deep at 26° 31′N/35° 00′E associated with a magnetic anomaly which probably developed similarly to CD and SD. The N–S segment does not reveal any significant magnetic anomaly that points to magmatic activity. This is in agreement with the collapse-type development of the KD. Other deeps, such as the Vema Deep at the southern end of the N–S segment (Figs. 3 and 22), exhibit similar patterns to the KD, for example the shallow and smooth slopes and continuously dipping sediments (Fig. 4 in Guennoc et al. 1988). The N–S segment was already mapped as the Zabargad Fracture Zone (Bonatti et al. 1984; Marshak et al. 1992) and may act as a transform zone. Smaller transform zones may exist south of the CD (Ehrhardt et al. 2005) and north of the SD at the Brothers Islands and correlate probably with the accommodation zones identified by Cochran (2005), Cochran et al. (1991), and Cochran and Martinez (1988).

Fig. 22
figure 22

Sketch of the segmentation of the northern Red Sea. The NW–SE segments are oriented normal to the Red Sea extension direction and are termed “volcanic” segments. The N–S segment is oblique to the extension and has a strike-slip component. It correlates with the Zabargad Fracture Zone (Bonatti et al. 1984; Marshak et al. 1992)

Full size image

Further to the south, another NW–SE segment of the central Red Sea adjoins the studied N–S segment (Zabargad Fracture Zone) (Fig. 22). It exhibits well-developed ocean deeps such as the Nereus Deep that forms a 40-km-long trough with a central ridge from which tholeiitic basalts were recovered (Antonini et al. 1998). This segment comprises deeps that are associated with magmatic activity, like in the northern NW–SE segment, but the deeps are further developed, in agreement with the model of a rift that propagated from north to south (Bonatti 1985; Martinez and Cochran 1988). Thus, a correlation between types of the deeps with their location exists and magma-related deeps are limited to the NW–SE segments of the Red Sea.

Following this approach, the development of the northern Red Sea deeps is classified into intrusion-related deeps and collapse-type deeps. This classification correlates also with the geographical allocation of the deeps (see Table 1). Magmatic activity and the following development of intrusion-related type of deeps are restricted to the NW–SE segment, thus to the segment where the Red Sea extension is normal to the strike of the axial depression. The KD and the Vema Deep which are located in the N–S segment are collapse-type deeps. The lack of magmatic intrusions in the N–S segment can be explained by the reduced extension due to the oblique strike of this segment in comparison with the main extension direction. The reduced extension could be insufficient to promote magmatic intrusions. The significant strike-slip component is probably compensated along the apparent strike-slip fault that crosses the KD. The narrow appearance of the N–S fault (Fig. 17) shows no major extension at the seafloor. It is reasonable that the extension is masked by lateral salt flow similar to the model of Augustin et al. (2014) for the central Red Sea. Thus, no intrusion-related deep such as the CD and SD is mapped in the N–S segment. The deeps in the N–S segment such as KD and Vema Deep developed as collapse structures. However, this type of deep is not necessarily limited to the N–S segment, as it is only dependent on the existence of a hydrothermal circulation cell, and this could be anywhere along the rift. Furthermore, the described types of deeps represent end-member models. Deeps of “mixed origin” are possible.

After classifying the deeps as direct and indirect rift-related structures, it is next possible to extract information about the rift. The NW–SE segment is oriented normal to Red Sea extension. With continued extension along the axial depression (Cochran et al. 1991), the NW–SE segment was pierced by magmatic intrusions. Red Sea deeps developed between or around them. This is in agreement with the model of Bonatti (1985) who correlated the conspicuous constant spacing of the magnetic anomalies with Raleigh-Taylor instabilities according to density/viscosity inversions in the upper mantle. Thus, the NW–SE segments are termed “volcanic segments.”

In contrast, the N–S segment of the northern Red Sea is not associated with major magnetic anomalies and seems to be unaffected by magmatic intrusions. However, it is possible that intrusive bodies are present with temperatures above the magnetic blocking temperature, but in general the temperature of the top of the crystalline basement should be below the Curie temperature. Although high heat flow values were reported from the northern Red Sea (Martinez and Cochran 1989), the temperature rises gently because of the good thermal conductivity of the evaporites (~5.5 W K−1 m−1) (Cochran and Martinez 1988). The temperature at the sediment-basement interface is calculated to be about 300 °C, using a three-layer model and representative parameters for heat flow, thermal coefficient, and layer thickness (see also Cochran and Martinez 1988). The deeps that were discovered so far in the N–S segment have the topographic characteristics of collapse structures caused by subrosion (salt dissolution) of Miocene evaporites from below. Thus, the N–S segment is termed a “non-volcanic” segment that links the “volcanic” NW–SE segments (Fig. 22).

This sequence of volcanic and non-volcanic passive continental margin segments can be observed also at other oceans such as Baffin Bay between Baffin Island (North America) and Greenland (Skaarup et al. 2006) or the South Atlantic Ocean (Franke et al. 2007).

As the collapse-type Red Sea deeps are only dependent on a hydrothermal circulation spot, it is conceivable that they also developed in the NW–SE “volcanic” segment, independently from magmatic activity and thus not correlated with known magnetic anomalies. This is supported by the existence of the unnamed deep north of the CD (Guennoc et al. 1988) that reveals evidence for a collapse-type deep. In addition, the processing of the bathymetric data of the transits of R/V Meteor Cruise M44/3 points to some more deeps in the NW–SE segment (Fig. 23). This map illustrates the combined swath echosounder data from the cruises M44/3, M52/3 and transits. At 26° 35′N/35° 11′E, the unnamed deep mentioned above could be identified (Cochran et al. 1986), besides other depressions that point to deeps with water depths of more than 1,350–1,400 m.

Fig. 23
figure 23

Bathymetric track of Meteor Cruise M44/3. This line shows the bathymetric link between the two survey areas of the M44/3 cruise. Although the track is slightly offset from the axial depression, several deep-like areas were detected. Grey colors indicate water depths above 1,000 m and white colors depths of more than 1,000 m

Full size image

Seismostratigraphy of Evaporites

The influence of buoyancy on driving salt tectonics has long been overestimated (Hudec and Jackson 2007). Instead, in many settings, evaporite deformation is primarily related to gravitational forces as a consequence of basement tilt (gliding) (Brun and Fort 2011; Hudec and Jackson 2007). Netzeband et al. (2006) stated that salt deformation due to basin tilt may start during deposition. The divergent reflection pattern beneath a flat salt top as revealed in Fig. 21 (KD) corroborates the assumption of syn-depositional salt tectonics.

The lateral variation between layered and transparent evaporites is also intriguing (Figs. 8, 11,18, 19, 20 and 21). Seismic data from the Atlantis II Deep show the same phenomenon (Fig. 5 of Mitchell et al. 2010). Generally, a vertical succession of evaporites of alternating transparent and stratified sequences is well known from the Messinian evaporites of the Levant Basin (Bertoni and Cartwright 2005, 2006; Dümmong and Hübscher 2011; Netzeband et al. 2006). In the Levant Basin, the vertically varying reflection characteristics imply a vertical change of evaporite facies, in which the transparent sequences are typically considered as halite and the stratified sequences are carbonates or sulfates (gypsum or anhydrite). Such alternating successions are well known and reflect the transition from low-to-high salinity or vice versa (e.g., Warren 2006). The lateral presence of transparent as well as stratified evaporites as observed in the Red Sea is uncommon but not unique, for example, it is also known from the Santos Basin off Brazil (Mohriak et al. 2012). In the Santos Basin, the evaporites are generally tabular, but internally seismically transparent diapirs alternate horizontally with folded and stratified evaporites. Similar to the Red Sea, the Santos evaporites developed along a continental rift system. Since the top of the evaporites is just a little deformed compared to the underlying salt, halokinesis must have started during the precipitation phase. In contrast to post-depositional salt diapirism, for example in the Zechstein Basins, the rim synclines were not filled with clastic sediments; instead, they were filled with evaporites other than the transparent halite.

We conclude that the lateral change of evaporite facies is the consequence of rift- or salt load-induced basin tilt, resulting halite gliding and deformation and the infill of evolving accommodation space between halite folds (e.g., diapirs) by evaporite facies other than halite, for example, gypsum or carbonates.

Conclusions

As a result of this work, we conclude that rift-related extension and faulting induced magmatic intrusions and extrusions, respectively. Where rifting and magmatic activity coalesce, the development of the northern Red Sea deeps is preferred. As magmatic-related deeps such as the CD, SD, or Nereus Deep are only mapped in the NW–SE segments of the northern Red Sea, the NW–SE segments were termed “volcanic” segments. The “volcanic” segments are linked by a “non-volcanic” N–S segment that is oblique to the NW–SE segments and acts probably as transform zone. Collapse-type deeps such as the KD are not linked to magmatic activity and are not necessarily limited to the “non-volcanic” segments of the Red Sea. The present multi-beam data suggest that deeps are more abundant than previously known.

The observation of the widespread salt sediments showed, as opposed to most of the worldwide known salt deposits where vertical variations in seismic reflection characteristics reflect deposition of varying salt facies that the northern Red Sea evaporites reveal a lateral variation from stratified to transparent. Accommodation space for stratified evaporites evolved already during salt deposition due to syn-depositional salt tectonics, most likely due to basin floor tilt caused by rifting or salt load. The bathymetry is not only controlled by plate tectonics and volcanism. The sediment covered salt flows as an analogue to ice-glaciers and builds streamlines that are formed sub-parallel to the flow direction. Salt rises are present where the salt flows above basement faults.

In order to broaden the knowledge of the Red Sea deeps and their relation to the rifting, extended detailed bathymetric data sets must be acquired and added to seismic data acquisition at crucial locations, such as, for example, the different deeps. This would be helpful for the understanding of the Red Sea rifting process and to understand the complex behavior of evaporites in tectonically active regions such as the Red Sea.