The Dead Sea Transform - an introduction

The following short essay (edited from Butler et al. 1998) outlines general aspects of the Dead Sea Transform, its finite displacement and possible modern activity. You can use this as a broader introduction before visiting the sites in this web resource. The reference cited below can be found in the amalgamated reference list.

The plate boundary

The Dead Sea Fault System forms part of the plate boundary network of the Mediterranean-Arabian region linking sea-floor spreading in the Red Sea with the destructive southern margin of the Tethyan collision belt (see map). The structure of the southern part of this fault system is relatively simple, with deformation focussed along a narrow depression running up the Gulf of Aqaba, along the Dead Sea and Jordan valley (Quennell 1958; Freund et al. 1970).

Finite displacements

A variety of pre-existing markers have been correlated across the fault system, including the continental margin across the Gulf of Aqaba (e.g. Girdler 1990), to establish bulk displacements (reviewed by Zak & Freund 1981, Quennell 1984, Hempton 1987, Joffe & Garfunkel 1987). Zak & Freund (1981) amongst others estimate the total left lateral displacement across the Dead Sea Fault System is generally recognised to be at 105 km, based on correlations across its southern segments. A swarm of diabase dykes dated at 22-18 Ma show the full offset (Eyal et al. 1981) and provide a maximum age for the onset of transform activity. The displacement apparently accumulated in two distinct periods (Quennell 1984; Hempton 1987, Joffe & Garfunkel 1987). The first of these occurred up to the late Miocene and accounted for 65km displacement, at about 0.5 cm/yr. The remaining 40km of displacement has accumulated since the earliest Pliocene (> c. 4.5 Ma), at about 1 cm/yr.

The northern transform

The structure and kinematics of the Dead Sea Transform are more complex and obscure to the north. Ron & Eyal (1985), amongst many others (e.g. Quennell 1984) were only able to recognise about 25km of left lateral displacement in Lebanon and western Syria. Cenozoic deformation is distributed across a range of faults and folds from the Levant coast across to the Palmyrides of Syria. The timing and significance of these structures is highly controversial (e.g. Chaimov et al. 1990). Within Lebanon, the Dead Sea Transform is represented by some or all of an array of fault strands (e.g. Beydoun 1977, Arthaud et al. 1978, Walley 1988). Most large-scale reviews of plate boundary continuity (e.g. Hempton 1987; Kempler & Garfunkel 1994) consider one of these structures, the Yammouneh Fault, to be the main strand. This fault maps along a NNE-SSW trend (Dubertret 1955), implying a general restraining bend geometry to the transform. Morphological characteristics are documented by Garfunkel et al. (1981) who interpret these in terms of an active tectonic landscape. They link the Yammouneh Fault northwards on the Ghab Fault in NW Syria to complete the connection with the southern margin of the Tethyan collision belt. Ophiolites in southern Turkey are offset by about 75km across the Ghab Fault system (Freund et al. 1970), although Chaimov et al. (1990) suggest that the offset is just 20-30 km. A consequence of a composite Dead Sea-Yammouneh-Ghab system acting as the overall transform plate boundary is a transform-transform-transform triple junction with the left-lateral East Anatolian Fault in SE Turkey (e.g. Westaway 1994).

The active strand goes offshore?

An alternative view of the modern plate boundary geometry is that the Dead Sea Fault System diverts offshore onto the Levantine continental margin. Girdler (1990) takes the active transform out through SW Lebanon. Nur & Ben-Avraham (1978) suggest that some displacements pass offshore along the Carmel Fault to link out to the Levantine abyssal plain and hence to the Cypriot arc subduction zone. The coastal area north of the Carmel Fault is seismically active, with focal mechanisms suggesting left-lateral strike-slip (NW-SE) and broadly N-S extension (Hofstetter et al. 1996). In this case a broad zone of transcurrent shearing represents the active Dead Sea Transform, linking into the Cypriot arc at a transform-transform-trench triple junction. Recent studies by Butler et al. (1997, 1998) seem to confirm Girdler's view and are described in greater detail in this web site.

Tectonic structure of Lebanon and environs

This short essay (after Butler et al. 1998) summarises the structure of Lebanon. There are links embedded within the text to locations within these web pages. The references cited are listed elsewhere.

Northern Israel

Cenozoic deformation in Lebanon is characterised by combinations of folding and dominantly left-lateral strike-slip faulting. Immediately to the south, the Dead Sea Transform lies within the Hula Valley. Here the deformation zone is less than 10km across and characterised by generally low-lying topography flanked by uplifts. Detailed mapping by Heimann & Ron (1987, 1993) shows the structure to consist of alternating transform overlaps with linking panels of deformation. These 'away-from-fault' deformations are best displayed by pressure ridges, rotated and uplifted blocks with subsidiary extensional faults. The transform is therefore represented by segmented active faults that alternate back and forth across the topographic depression. Given the large bulk displacements across the transform as a whole, Heimann & Ron (1993) suggest that fault segmentation is transitory .

The faults of Lebanon

Notwithstanding the incremental segmented nature of the Dead Sea Fault System in Israel, the structure of Lebanon is far more complex (e.g. Arthaud et al. 1978). Here there are many strike-slip faults, generally oriented N-S or NNE-SSW, together with E-W cross-faults, and folds at the km and 10km scales (e.g. Beydoun 1977; Beydoun & Habib 1995). Walley (1988) termed the map-pattern 'braided' whereby the localised displacements from the south are distributed onto an array of five main transcurrent fault zones and a host of minor ones. These various deformation structures are reviewed in turn.

The Yammouneh Fault

Of the five major faults which outcrop in Lebanon (e.g. Beydoun 1977, Walley 1988), the Yammouneh is the most continuous and clearly evident on satellite images (e.g. Ambraseys & Barazangi 1989). Evidence for its continuity has been described by Garfunkel et al. (1981). They used its morphological characteristics, largely established from topographic maps, to infer that the fault is an active transcurrent structure. The southern portion of the fault coincides with the eastern slopes of the uplifted Jabel Barouk structure, forming the western edge of the southern Bekaa valley. To the north of the town of Chtaura the fault lies within the high ground of Mount Lebanon. Here it is characterised by a series of enclosed karstic sedimentary basins or poljes (Garfunkel et al. 1981). The Yammouneh Fault emerges from the northern edge of the mountains along Wadi Chadra. Here it appears to offset the southern outcrop edge of the Homs Basalt by about 10km. A similar offset is recognised on the north side of the Basalt, in NW Syria, and the Yammouneh is inferred to be continuous with the Ghab Fault (Garfunkel et al. 1981), linking northwards to the East Anatolian Fault (e.g. Joffe & Garfunkel 1987, Hempton 1987).

The Roum fault

Diverging from the southern end of the Yammouneh Fault near the Hula valley, the Roum Fault has been mapped as a discontinuous array of left-stepping segments with associated basins. However, it has been difficult to trace for more than about 30-40km to the north of Hula. There are no through-going structures mapped by Dubertret (1955). However, the region is characterised by flat-bedded Cretaceous carbonates and, as Dubertret's (1955) mapping was largely designed to show the distribution of lithostratigraphy, faults within individual geological units are generally not shown. Probable northward extensions to the Roum have been recognised in the Damour valley (Butler et al. 1997), just south of Beirut (Girdler (1990; his 'Ed Damour Fault') and within Beirut itself (Dubertret 1955). It has a clear geomorphological expression, bounding the coastal Tyre-Nabatiyé plateau to the west and the more dissected and uplifted ranges to the east (Sanlaville 1970). The Roum fault strands are associated with deflections in the courses of the Litani and Zahrani rivers (e.g. Walley 1988).

Other Faults

Splaying to the east of the Yammouneh Fault from the Hula area, the Hasbaya and Rachaiya Faults cut the western flank of the Mount Hermon anticline. The Hasbaya Fault runs broadly along the Jordan (Hasbani) valley north of Hula. A few kilometres to the east lies the Rachaiya Fault. Heimann et al. (1990) describe small step-overs on associated subsiduary fault strands which develop minor basins. They estimate about 1km of Quaternary horizontal movement on the Rachaiya Fault, based on these segment geometries. Dubertret's (1955) mapping shows neither the Rachaiya nor the Hasbaya Faults to be laterally continuous. If this is accurate, we infer that these structures have not accumulated more than a few kilometres total of transcurrent displacement.

The Serghaya fault is the most eastern of the major faults of Lebanon, running along the length of the Mount Hermon-Anti-Lebanon anticline. In the Anti-Lebanon mountains the fault is very difficult to map because of the monotonous, shattered and deeply karstified upper Cretaceous carbonates through which it runs. In the south, where the Serghaya Fault converges with the main part of the Dead Sea Transform near Mount Hermon, there are indications of Plio-Pleistocene tectonic activity.

Apart from the five major faults described above which outcrop in Lebanon, Khair et al. (1997) infer the presence of a buried structure- their Mid-Bekaa Fault. This is inferred on the basis of a step in the buried basement morphology modelled on gravity data. The lack of surface expression make it difficult to evaluate the significance of this structure as a late Cenozoic transcurrent fault.

Summary of Faults

In summary, of the five major faults (six including the hypothesised Mid-Bekaa structure of Khair et al. 1997) , the only structures which can have accommodated a substantial proportion of the 105km total offset of the Dead Sea Transform through Lebanon are the Roum, Serghaya and Yammouneh Faults. Of these, only the Yammouneh Fault appears to have the continuity and landscape characteristics consistent with slip-rates of 6mm/yr operating for several million years. However, to date the structural mapping of the Roum Fault has been inadaequate to establish its continuity and hence its capability to accommodate significant (>10 km) displacement.

Folds

A feature of the restraining bend on the Dead Sea Transform are the large-scale fold structures and associated topographic elevations that are the greatest in the Levant. The major anticline of Mount Lebanon and Jabel Barouk ranges is sub-parallel to the trace of the Yammouneh Fault (e.g. Beydoun 1977, Hancock & Atiya 1979). The complementary syncline to the east contains the Bekaa valley, bounded to its east by the major anticlines of the Anti-Lebanon and Mount Hermon ranges. Within the Bekaa valley there are additional folds, lying exclusively to the east of the Yammouneh Fault, generally oriented oblique to its trace with NE-SW -trending hinges. Another oblique fold-belt is found in NW Lebanon, around the city of Tripoli. These complex deformations which occur within Mesozoic-Cenozoic sediments are possibly detached from deep-seated crustal transform faults along Triassic evaporites (e.g. Beydoun & Habib 1995).

Beydoun (1977) and Khair et al. (1997) point out that the crust beneath the uplifted Mount Lebanon ranges is apparently of normal thickness. This observation has been used to infer that the Lebanese crust has not been shortened by Cenozoic tectonics (e.g. Ron 1987). However, the region was a site of substantial Mesozoic subsidence and extension (reviewed by Laws & Wilson 1997) with over 6000m of sedimentary cover (Beydoun & Habib 1995). This implies a stretching factor of about 1.5, a value which estimates the pre-orogenic thickness of the crystalline crust at just 20km. This is the thickness modelled by Khair et al. (1997) for the areas adjacent to the uplifted Mount Lebanon range. However, the present crustal thickness (excluding the cover) beneath Mount Lebanon is 23-25 km (Khair et al. 1997). Assuming homogeneous deformation beneath Mount Lebanon, area balancing implies about 8-10km crustal shortening.

Cross-faults and tectonic rotations

The oblique convergence across the Dead Sea Transform at the Lebanese restraining bend has been considered to be accommodated by block rotations and subsiduary faulting distributed across tens of kilometres off the main transcurrent faults . Following this proposal by Freund & Tarling (1979), Ron and co-workers (Ron 1987, Ron et al. 1990a,b) have identified tectonic rotations, counter-clockwise on the flanks of Mount Hermon and clockwise in Galilee. Declination anomalies are recorded for Cretaceous volcanics, compared with their predicted palaeo-pole positions for the Arabian sub-continent. Ron et al. (1990a) report 69˚13˚ counter-clockwise rotations from Hermon and correlate these with the 55.6˚10.4˚ rotations of the same sense identified by Gregor et al. (1974) for Mount Lebanon. The rotations are apparently related to regenerating NW-SE right-lateral strike-slip faults which rotate to broadly E-W before locking (Ron et al. 1990a). These structures apparently accommodate about 50% E-W elongation, essentially forming zones of non-coaxial plane strain wherein the vertical thickness is conserved.

Arrays of ENE-WSW-trending faults, generally assumed to accommodate right-lateral strike-slip (e.g. Ron 1987), are found on either side of the Roum fault. One array occupies a topographic low between Mount Lebanon and Jabel Barouk, running inland from Beirut into the Chouf hills, dissecting strata which record counter-clockwise rotations (Gregor et al. 1974). Another suite of faults lies on the Tyre-Nabatiyé plateau, between the Roum Fault and the coast of Lebanon. This region has not been studied palaeomagnetically. In contrast with other parts of the restraining bend, the plateau contains few folds. The Cretaceous-Eocene stratigraphy is generally sub-horizontal. Sparse field data confirm the assumption of strike-slip faulting (Nammour, 1992) although map-patterns of the offset of stratigraphy suggest significant vertical slip components (e.g. Hancock & Atiya 1979).

TOTAL SUMMARY

The Lebanese restraining bend of the Dead Sea Transform contains a series of major transcurrent faults of which only the Yammouneh Fault is generally considered to be capable of accommodating many tens of kilometres of left-lateral transcurrent displacement. The restraining bend is characterised by distributed deformation adjacent to the Yammouneh Fault. This is most obviously expressed as long wave-length anticlines, generally aligned sub-parallel to the major fault. The Mount Lebanon anticline is probably the surface expression of crustal thickening, this deformation effecting previously extended Mesozoic sedimentary basins thereby returning the crust to approximately its normal thickness. There are additional minor folds with axes commonly oblique to the trend of the Yammouneh Fault. These presumably record distributed transcurrent shear. Other distributed shears are less obvious than the folds but may be more important as finite deformation mechanisms. Arrays of closely-spaced, generally right-lateral, faults are spatially associated with strata with substantial declination anomalies. Ron et al. (1990a and others) relate these to distributed block rotation within the restraining bend, locally accommodating net elongations of several tens of kilometres.

The above summary is a simplified description of the finite structure of the restraining bend. Several workers (e.g. Westaway 1995) have attempted to explain the relationships between the orientation of the Yammouneh Fault, the total displacement on the Dead Sea Transform, the finite relative plate movement vector and the record of distributed longitudinal strains and tectonic rotations. As discussed earlier however, the regional plate kinematics have varied during the 18 Myr history of the transform. Consequently we suggest that an incremental analysis of tectonic evolution is required.