Source	Image	Description
Ben-Avraham et al (2012)	Fig. 17.1 Tectonic setting of the Dead Sea fault — a transform plate boundary that connects the opening of the Red Sea with the collision in Turkey. Boxes indicate basins along the fault valley discussed in the text. Gulf of Elat Arava valley Dead Sea basin Sea of Galilee basin Hula basin Ghab basin Ben-Avraham et al (2012)	Fig. 17.1 - DST Tectonics

The Dead Sea fault, which spans around 1000 km (Fig. 17.1), is an active left lateral, strike-slip (transform) plate boundary. This feature forms the northern part of the Syrian-African rift system, which extends for >6000 km from Turkey to Mozambique. The fault formed as a consequence of late-Cenozoic breakup of the Afro-Arabian continent and the subsequent drift of Arabia away from Africa. Older structures resulting from mild compression during the Senonian to Mio¬cene are evident throughout the region. These east-northeast to northeast trending and east to west trending folds and faults compose the Syrian-Arc system. The Dead Sea fault obliquely crosses and offsets laterally these structures and is hence, younger. Recent continental breakup was accompanied by a marked change in the behaviour of the region - the Syrian-Arc deformation virtually ceased, the region was uplifted, especially near the newly formed plate boundaries and widespread igneous activity took place. In particular, this resulted in the uplift of the margins of the Dead Sea fault system as well as subsidence along much of its length forming the pronounced physiography seen today (a profound valley, bordered by high mountainous scarps and occupied by deep basins). The fault valley is for the most part, bordered by high escarpments commonly known as the Western and Eastern Boundary faults.

In terms of plate tectonics, the Dead Sea fault is considered to be a left-lateral transform plate boundary with an extensional component, separating the Arabian Plate and the Sinai sub-plate. Activity along the fault is thought to have started in the middle Miocene, when a transition of motion took place from opening in the Gulf of Suez to transcurrent displacement along the Dead Sea fault. The total amount of left lateral slip is estimated at about 105 km. History of motion is not well constrained because young markers cannot be reliably matched across the Fault.

Calculation of regional plate kinematics provides an instantaneous rate of motion along the Fault boundary of about 5-6 mm/yr for the South Sinai triple Junction. This average rate is similar to the average slip computed from the ratio between the entire 105 km slip to ca. 17-20 Ma of tectonic activity (presuming that the overall slip rate has remained approximately uniform during the history of the Fault). More recent GPS measurements provide a minimum slip rate of 3.7 ± 1.1 mm/yr for the three years between 1996 and 1999.

The fault is divided into two sections — the southern section from the Gulf of Elat to south of Lebanon, and a section continuing northward from Lebanon to the Taurus mountain range in Turkey (Garfunkel, 1981). The fault valley is character¬ized by large-scale topographic and structural asymmetries (Wdowinski and Zilberman, 1996). Along the southern section, particularly in the northern Arava valley, the fault's eastern shoulder is flexed upwards towards the axis, resembling an uplifted shoulder, while the western side is flexed down towards the axis, resembling a wide arch. In addition, the eastern side is for the most part topographically higher than the western side. The two sections of the Dead Sea fault differ by a reversal in the large-scale asymmetry; in the southern section (Gulf of Elat to south of Lebanon), the eastern side is usually higher than the western side, while in the northern section (north of Lebanon), the western side is mostly higher than the eastern one.

Along strike variations in topography and structure subdivide the southern Dead Sea fault into five segments. From south to north, these are the Gulf of Elat, Arava, Dead Sea and the Sea of Galilee and the Hula Valley (Fig. 17.1) (Garfunkel, 1981). Deep pull-apart basins, formed between left-stepping fault segments of the Dead Sea fault, can be found in topographically lower areas. Lengths of the individual pull apart basins range from 15 to 50 km and are 5-20 km wide, while their depths range between 5 and 15 km. Marginal normal faults usually border these depressions, while transverse or oblique normal faults define their southern and northern limits. Structural saddles typically occur between the basins and there, transpression often takes place. The crystalline basement underlying the larger pull-aparts is thinner than normal, which accounts for basin subsidence. The basins tend to become shorter and younger in age north¬wards along the fault. Asymmetry occurs in most of the basins, interpreted by Ben-Avraham and Zobak (1992) as indicating that they are bounded by transform segments on one side and sub-parallel normal faults on the other. In these cases, the simple pull-apart model cannot be applied in a straightforward way. Compressional features can be found along many sub-basins, suggesting that the tectonic regime is more complex than specified in the simple models. Both asymmetry and compressional features will be examined in the basins along the southern Dead Sea fault, from the south to the north.

The extensional regime combined with the dominant lateral motion along the Dead Sea fault resulted in the formation of a series of pull-apart basins (or rhomb-shaped grabens). Basins are separated by structural saddles, which formed as a result of local stress fields. In many places, transverse faults cross the saddle, thus separating the basins tectonically as well as structurally.

Source	Image	Description
Ben-Avraham et al (2012)	Fig. 17.2 Profiles and location map from the Gulf of Elat. Location map showing general bathymetry and the location of sub-basins and deeps. Heavy lines and arrows indicate segments of the Dead Sea fault; thin lines indicate other faults in the gulf (dashed where inferred). Dotted lines show the location of seismic profiles. Reverse in asymmetry is evident from the southern sub-basin, where the main strike-slip fault is on the west, to the northern sub-basin, where it is on the east. The central basin is unique in the sense that this is the only area where strike-slip faults overlap, producing an area of extreme deformation and compression. Seismic profile from the Amona Deep in the central sub-basin (Ben-Avraham, 1985). The Arnona Deep is bordered on the east by a large compressional feature. Seismic profile from the Elat Deep in the northern sub-basin (Ben-Avraham, 1985). Basin asymmetry is evident in both profiles. Ben-Avraham et al (2012)	Fig. 17.2 - Gulf of Elat

The Gulf of Elat is the largest depression along the Dead Sea fault (~180 km long and up to 20 km wide) and occupies its southernmost part. The interior of the gulf comprises three elongated coalescing en-echelon pull apart basins, which strike N20°-25°E. Undulations in the floors of the basins produce several distinct deeps: Tiran and Dakar in the southern depression, Amona and Aragonese in the central depression and the Elat Deep in the northern depression. The Hume Deep, south of the Straights of Tiran, links the Gulf of Elat with the axial depression of the Red Sea. The northern part of the gulf has a relatively simple bathym-etry and is dominated by a single flat-bottomed basin with one deep, the Elat Deep, which is the largest and shallowest in the gulf. Only this part of the gulf is symmetrical in cross section with respect to the bathymetry. Although the bathymetry of the Elat Deep is relatively flat and simple, its underlying structure suggests that it was originally composed of at least two separate deeps. The central part of the gulf is the narrowest and deepest and is occupied by a basin, which crosses the axis of the gulf diagonally from northeast to southwest. The southern part of the gulf is the widest and occupies the eastern side of the gulf. The internal structure of the gulf of Elat is dominated by intra-basinal longitudinal faults and by oblique faults that define these sub-basins. In the central part of the gulf, the older sediments in the deep depressions and on much of the marginal blocks are considerably distorted into arches and open folds, raising the possibility of local transpression (Fig. 17.2A). However, some of these features may also be the expression of underlying salt (or shale) diapirs. It is interesting to note that this is the only area where there seems to be evi¬dence for strike slip motion on both sides of the basin (Ben-Avraham, 1985). The sub-basins are separated by structural saddles, which also internally exhibit folding, flexure and uplift suggesting transpression (Fig. 17.2A).

At the southern extremity of the Gulf of Elat, local transpression is evident by the uplift of the Tiran Island block on which Quaternary reefs are found at elevations exceeding 500 m above sea level. An additional feature observed in Figs. 17.2A and 17.2B is the existence of asymmetry within the sub-basins of the Gulf of Elat. The transition from the Arnona Deep (Fig. 17.2A) to the Elat Deep (Fig. 17.2B) is accompanied by a change in lateral asymmetry. A possible explanation may be that the basins are bounded on one side by a strike-slip master fault and on the other by a predominantly normal boundary fault. Where motion along en-echelon segments of the main fault switches from one segment to another (i.e., from one side of a basin to the other), asymmetry in the basin structure is also expected to switch sides. According to this interpretation, asymmetry within the basins suggests that in the southern and central portion of the gulf, active strike-slip motion is taking place primarily on the western boundary faults, while in the northern portion it is taking place primarily on the eastern boundary fault (Ben-Avraham, 1985). In addition, it can be seen from Fig. 17.2A that the Arnona Deep is bounded on the east by a large compressional feature.

Source	Image	Description
Ben-Avraham et al (2012)	Fig. 17.3 Profiles from the Zofar basin in the Arava valley. Location map showing the division of the valley into two main segments comprised of a number of smaller sub-basins (after Frieslander, 2000). The en-echelon arrangement of the segments, which are separated by a large structural saddle, is evident. Main faults appear as bold lines and location of the seismic profiles, as dotted lines. Example of classic strike-slip fault, the Zofar fault, where reflectors dip towards a main fault (after Frieslander, 2000). Reverse in dip of sediments in the Zofar basin may indicate an additional structural subdivision (after Frieslander, 2000). Ben-Avraham et al (2012)	Fig. 17.3 - Arava

The Arava valley is a morphotectonic depression connecting the Gulf of Elat and the Dead Sea (Fig. 17.3A). The transition from the Gulf of Elat to the Arava Valley is marked by a structural saddle (Fig. 17.1), where mild compression is apparent by upwarping of Neogene and younger sediments. During the Pliocene, this saddle was high and wide enough to allow a major drainage line to flow west¬ward across the Dead Sea fault into the Negev. In the Arava Valley itself, a few small transpressional uplifts developed. The small size of these features may indicate that they are young and formed as a result of recent rearrangements of the active fault strand.

This segment of the Dead Sea fault is characterized by a series of elongated, en-echelon tectonic sub-basins. The left stepping arrangement of the basins (Evrona, Ya'alon, Zofar and Shizaf; Frieslander, 2000; Fig. 17.3) coincides with the junction of E-W dextral faults in the central Sinai-Negev area, which predate the left lateral motion on the Dead Sea fault. The depths of these sub-basins, which are filled with clastic sediments, vary from several hundred meters to a few kilometres. The basins are bounded by sub-parallel, longitudinal boundary faults. In general, the boundary faults are high-angle strike-slip faults with normal displacement. In places, due to the left lateral strike slip motion, reverse faults, collapse antithetic features and compressional 'flower structures' were formed (Fig. 17.3). The pattern of these basins subdivides the Arava valley into two low structural segments, separated by an elevated saddle (Frieslander, 2000).

North of this saddle, the Zofar and Shezaf basins are separated by two listric faults with large displacements. The nature of features of the fault on the east side of this segment of the Arava indicate that it is the main strike slip fault (Garfunkel et al., 1981). The asymmetric fill in these sub-basins may well be related to the fact that strike-slip faulting took place mainly along the eastern boundary fault. In the shallowest (Zofar) basin, the seismic data indicate a regional dip toward the north (Frieslander, 2000).

As in the Gulf of Elat, reversal in asymmetry is also observed in the Zofar sub-basin. Figure 17.3B shows a classic example of a strike-slip fault, where horizons on both sides of the fault, dip towards it. Figure 17.3C, located around 15 km to the north, shows a distinctly different picture. Instead of reflectors of the Zofar basin dipping westward towards the fault, they dip to the east and away from the fault. This reversal in dip direction is the only known example along the Arava valley where reversal in asymmetry occurs within a sub-basin (instead of between sub-basins), and may indicate that the Zofar basin can be structurally sub-divided.

Source	Image	Description
Ben-Avraham et al (2012)	Fig. 17.5 Profiles from the Sea of Galilee. A Location map showing main tectonic elements in the Sea of Galilee area. Bold lines indicated major faults, while dotted lines show the position of seismic profiles. Grey area shows the extent of the Yehudiyya transtensional one (after Shulman et al., 2004). Ben-Avraham et al (2012)	Fig. 17.5A - Sea of Galilee
Ben-Avraham et al (2012)	Fig. 17.5 B and C Profiles from the Sea of Galilee. B Profile from south of the lake showing both compressional and asymmetrical features (Zurieli, 2002). C Profile from north of the lake showing the Yehudiyya transtensional zone. Within the zone lies a compressional feature, the Qazrin structure, which is bordered on the west by the main fault zone (Shulman et al., 2004). Ben-Avraham et al (2012)	Fig. 17.5 B and C - Sea of Galilee

Situated at 200 m below mean sea level, the Sea of Galilee (Lake Kinneret) is around 5 km wide and ca. 20 km long (Fig. 17.5A). The basin is characterized by a negative gravity anomaly (Ben-Avraham et al., 1996), indicating a ~8 km thick sedimentary fill under the central part of the lake (thinning to ~5 km towards the southern extension of the basin). Around the lake, Plio-Pleistocene outcrops of volcanic rocks are present. The Sea of Galilee was divided by Ben-Avraham et al. (1996) into two structural units. The southern half is a pull-apart that extends southwards from the lake, whereas the northern half is thought to have been formed by rotational opening and transverse normal faulting. A later seismic reflection survey (Hurwitz et al., 2002) has shown that the southern basin, delimited by two longitudinal faults, extends under much of the lake. Seismic sections show that the deepest part of the basin is located well south of the deepest bathymetric depression implying that the later is an actively subsiding young feature. The eastern boundary fault extends to the northern part of the lake, while at least on the surface the western fault does not cross the northern section.

The Zemah anticline, a large compressional feature south of the Sea of Galilee in the Kinarot valley (Fig. 17.5B), is at least 4 km long, 2500 m wide and reaches depths of at least 4 km. Figure 17.5B (Zurieli, 2002) clearly hows that the eastern limb of this structure is thicker than the western one (by almost 50%).The anticline is bounded by two faults; the one to the east reach the surface as opposed to the one in the west, which exhibits reverse fault geometry. Other faults can be found at the base of the feature. The Zemah-1 well was drilled on top of the structure and penetrated 4300 m of graben fill consisting of clas-tics, evaporites, intrusive and igneous rocks, with only relatively thin layers of salt present. An additional compressional structure the Ubeidiya structure) is evi¬dent in Fig. 17.5B around 3.5 km west of the Zemah anticline. This element is much smaller (around 800 m wide) and oriented to the northwest. Both elements are thought to have been formed sometime between 0.8 and 1.4 Myr. Rotstein et al. (1992) have suggested that reverse faulting in the Zemah anticline indicates that it is a compressional feature unrelated to igneous activity or salt diapirsm. Between these two compressional structures, an asymmetrical basin, which is the continuation of the Sea of Galilee syncline, is present. Rotstein et al. (1992) have also suggested that the basin in the southern Sea of Galilee and the Kinarot Valley was formed by two overlapping strike-slip faults. Zurieli (2002) proposed that the compressional features can be explained by assuming that the basin is a hybrid pull-apart, that is bordered on both sides by faults with different rates of lateral motion. The Golan Heights lie to the east of the Sea of Galilee and their internal structure is governed by the adjacent northern Dead Sea fault and the Palmyride folding. The Golan Heights can be divided into two separate sections by the Yehudiyya block, or the Yehudiyya transtensional zone (Shulman et al., 2004). The zone is part of a distinct extensional region, or sub-basin, which is confined by:

1. a branch of the Jordan Fault (Fig. 17.5C) in the west (the Meshoshim fault)
2. the southwest-northeast trending Sheikh-Ali fault, in the east and northeast
3. by the Sea of Galilee in the southwest. The Neogene activation of the Sheikh-Ali fault, which branches to the northeast off the north-south trending Dead Sea fault, was the most significant tectonic phase that shaped the structural framework of the Golan Heights.

Source	Image	Description
Ben-Avraham et al (2012)	Fig. 17.6 Location map of the Hula basin and main tectonic elements in the northern Dead Sea fault valley. Schematic map of the northeastern Hula valley and the associated push-up structures (after Zilberman et al., 2000). Ben-Avraham et al (2012)	Fig. 17.6 - Hula Basin

The Hula Valley is the northernmost pull-apart basin along the southern section of the Dead Sea fault. The basin developed between left-stepping segments of the Dead Sea fault: the Jordan and the Rachaya fault in the east and the Yammu-neh fault in the west (Heimann, 1990), which extends northward to the Zagros-Taurus zone of plate convergence (Fig. 17.6).

The eastern structural boundary of the Hula Valley forms a complex fault pattern, reflecting a westward migration of tectonic activity during the Pleistocene (Heimann, 1990). The youngest branch is the Azaz fault, separates the western margins of the basalt plateau of the Golan Heights from the sedimentary fill of the valley (Fig. 17.6).

The Azaz fault is a sinistral strike—slip fault consisting of several segments arranged in a right-stepping en-echelon pattern striking NNE (Heimann, 1990). Small push-up structures, expressed as small hills, occur between neighboring segments (Fig. 17.6). Both the Rachaya fault and the Azaz fault are considered to be currently active.

Source	Image	Description
Ben-Avraham et al (2012)	Fig. 17.7 Fault map in the Ghab Basin and immediate surroundings. Faults have been mapped either from surface observations and geomorphology (gray lines) or from seismic reflection and other interpretations (black lines) (after Brew, 2001). Interpreted migrated seismic profile across the asymmetrical Ghab Basin. Quaternary age deposits are the surfical strata along the whole line. Most of the faults shown have components of both normal and strike-slip fault movement. The fault marked F is the main strand of the Dead Sea fault and the major eastern bounding fault of the Ghab Basin (after Brew, 2001). Ben-Avraham et al (2012)	Fig. 17.7 - Ghab Basin

The northern segment of the Dead Sea fault

The structure and kinematics of the Dead Sea Transform are more complex and obscure to the north (Walley, 1998). While about 25 km of left lateral displacement in Lebanon and western Syria is generally accepted, there is some debate and claims of an offset of 80 km or more near the Syrian-Turkish border. A lateral offset of ca. 100 km across Lebanon has been suggested on the basis of structural correlations across the Dead Sea fault. Resolution of this question must await a comprehensive comparison of the structure and stratigraphy on both sides of this segment of the Dead Sea fault, similar to what was done along its southern segment (e.g., Freund et al., 1970). 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. Within Lebanon, the Dead Sea Transform is represented by some or all of an array of fault strands. Most large-scale reviews of plate boundary continuity consider one of these structures, the Yammouneh Fault, to be the main strand. This fault maps along a NNE-SSW trend, implying a general restraining bend geometry to the transform.

The Ghab basin

The Ghab basin is located on the active, yet poorly understood, northern segment of the Dead Sea fault system. It formed during the Plio-Quaternary as a result of a complex step-over zone along the fault. Subsidence occurred along cross-basin and transform-parallel faults in two asymmetric depocenters. The larger depocenter in the south of the basin is asymmetric towards the east, the margin along which most active transform displacement apparently occurs. This suggests that, at the latitude of the Ghab Basin, most of the lateral movement on the Dead Sea fault is accommodated on the eastern bounding fault of the basin.In contrast, the second, smaller depocenter to the north is somewhat asymmetric toward the western bounding fault (Brew, 2001)(Fig. 17.7).

Motion along the Dead Sea fault is not pure strike-slip and the direction of the plate boundary changes several times resulting in areas of transtension and transpression. This is evident by the variable morphology and structure, which is characterized by extensional, compressional and asymmetrical features. These features vary in size, from the large-scale, which define the general structure of the valley, to the small-scale, which define the internal structure. The Dead Sea fault valley itself is a large-scale transtensional feature, which formed as a result of oblique strike-slip motion along the fault. The extensional component of motion is largely responsible for the formation of pull-apart basins that occupy its length. Compression is caused by a step to the right of the left lateral strike-slip master fault and results in structural saddles, which separate the main basins along the valley. Smaller compressional uplifts divide several of the major pull-aparts into sub-basins and also occur along the fault valley. However, it is often hard to determine whether these intra-basinal features are salt-related. Asymmetry is evident in the structure and topography of the highlands, which border the valley and in the basins that lie within. The large-scale asymmetry across the Dead Sea fault can perhaps be explained by the combined effect of normal faulting and isostatic uplift on existing pre-rift topography (Wdowinski and Zilberman, 1996).

Numerous examples of basin asymmetry and compression features have been presented from locations worldwide, indicating that these are not local phenom¬ena limited to the Dead Sea fault valley. The Cariaco basin in Venezuela, which crosses the El Pilar fault, shows many similarities to the Dead Sea basin, displaying a clear asymmetry towards the main fault (Ben-Avraham and Zoback, 1992). Additional examples exist along other continental transform faults and rift valleys, such as the Motagua fault system in Guatemala, Lake Baikal, Lake Tanganyika in the East African rift system and the north Anatolian fault in Turkey. Transform-normal extension may help explain the small-scale asymmetries observed within the basins themselves. This seemingly contradicts the basic assumption that pull-aparts are formed as a result of overlapping strike-slip faults. However, it is the extensional component that arises from such fault arrangements that may well be responsible for resolving this 'conflict' ('leaky transform', Garfunkel, 1981). It is interesting to note that very little or no mantle uplift exists under the Dead Sea fault in the Arava valley. This may explain the large negative gravity anomalies associated with the fault that reflect primarily the thick sedimentary sequences including salt layers, which are abundant along the length of the valley.

• This study focuses on the southern segment based mainlyon the wealth of geophysical data. Owing to transtention caused by oblique-slip and the overlapping of en-echelon fault strands, a series of pull-apart basins were formed along the fault’s length. These basins are long and deep-reaching in places more than 10 km deep.They are characterized by extensional, compressional, and asymmetrical structures varying in size from large-scale (defining the general structure of the Dead Sea fault valley) to small-scale (defining the internal structure). This study examines the internal structure of thesebasins from south to north and summarizes the state of knowledge to date.


• Today there is a marked difference between the geology of areas facing each other across the Dead Sea transform — a result of lateral motion that juxtaposed areas that were originally far from each other. Matching of numerous markers across the transform indicates a left-lateral offset of approximately 105 km (Quennell 1959, Bartov 1974).
  


• ... these studies indicate an average slip rate of 5–7 mm year⁻¹ in the past 5 Ma.


• Igneous activity — overwhelmingly basaltic volcanism — occurred on a regional scale, mostly east of the transform (Giannerini et al. 1988, Garfunkel 1989, Steinitz and Bartov 1991, Mouti et al. 1992, Ilani et al. 2001). The volcanics occur in a wide belt that broadly follows the transform, but a series of relatively small occurrences exist within the transform. In places the transform appears to have been a preferred site of magma upwelling along the northern two-thirds of its length. Volcanism began locally 18–20 Ma ago near Tiberias and in the Damascus region; however, most volcanics were extruded since 5-12 Ma.
  
  The structure along the Dead Sea Fault varies considerably, and this is expressed in its physiography. Structurally, it can be divided into two segments: south and northof lat. 33° 10' N. The southern segment is marked by an almost continuous valley, most of which is underlain by a series of deep basins that are separated by less pronounced saddles (Garfunkel 1981, Garfunkel and Ben-Avraham 2001). These structures are controlled by longitudinal en-echelon faults on which lateral motion takes place. These are flanked by normal faults that produced the morphologic boundaries of the valley. Transverse faults are also present, but less conspicuous. In map view, the segment has an arcuate shape, and the strike of the major longitudinal faults varies gradually from approximately 25° NE in the south to nearly N-S in the north, which is compatible with the kinematics of Arabia-Africa plate separation. The margins of this segment were hardly deformed during the development of the Dead Sea Fault, except in the area to the west of lat. 32° N (Samaria and Galilee).
  
  The northern segment is quite different. It comprises well-expressed longitudinal faults with variable strikes. Valleys are developed only along parts of this segment. In Lebanon, the main fault extends along a high-standing area—the topographically highest area along the transform. The flanks of this segment were considerably deformed during the development of the transform. Given the changes in strike of theDead Sea Fault, continuing lateral motion had to deform its flanks. In particular, the fault bends some 30° clockwise in Lebanon. This would be expected to produce substantial transpression, which is considered to be the cause of the great uplifting of its flanks (Quennell 1959), and their intense faulting, which involved great rotations of individual fault blocks about vertical axes (Ron et al. 1984, Ron 1987). East of the transform the compression led to folding.

To the north of the gulf's head lies the Arava Valley (Supplemental Figure 2), a 160-km-long morphotectonic depression connecting the Gulf of Elat and the Dead Sea (Bentor et al. 1965, Ben-Avraham et al. 1979). The southern part of the Arava Valley is marked by a structural saddle, resulting from mild compression (Garfunkel 1981, Garfunkel et al. 1981, Garfunkel & Ben-Avraham 2001, Basson et al. 2002). This compression can clearly be seen in upwarping Neogene and younger sediments. According to Ginat & Avni (1994) the saddle, located 50-70 km north of the Gulf of Elat, was high and wide enough to allow a major drainage line to flow westward across the Dead Sea Fault into the Negev during the Pliocene. In the southern section of the Arava, just north of the Gulf of Elat, compressional features developed in the sedimentary fill (Basson et al. 2002). These structures are young and formed as a result of recent changes in the geometry of the longitudinal master faults.

The Arava Valley is characterized by a series of elongated, en-echelon (left-stepping), tectonic subbasins filled with clastic sediments, the Evrona, Ya'alon, and Zofar basins, and the Shizaf Basin north of the Zofar Basin (Supplemental Figure 2a) (Bartov et al. 1998, Frieslander 2000). These basins are bound by subparallel high-angle strike-slip longitudinal faults with normal displacement and range in depths from several hundred meters to a few kilometers and are filled with elastic sediments. The southern ends of the Yaalon and Zofar basins seem to be aligned with the inter¬section of E-W Themed and Paran dextral faults in the central Sinai-Negev and the Dead Sea Fault. These E-W faults predate left-lateral motion along the Dead Sea Fault, as they themselves are displaced laterally. As such, they may have influenced the younger structure along the Arava.

To the north, the Zofar and Shezaf basins are delimited by two NW-SE striking listric faults with large displacements. In this area, the main strike-slip fault, the Arava Fault, is located on the eastern side of the valley (Garfunkel et al. 1981, Garfunkel 1981, Atallah 1992). Predominance of strike-slip motion along the eastern boundary fault may explain the asymmetry in the sedimentary fill of these subbasins. The Zofar Basin is bounded on the west and east by the Zofar and Arava faults, respectively. Whereas motion along the Arava Fault is left-lateral strike-slip, the young motion along the Zofar Fault is predominantly normal (Bartov et al. 1998, Frieslander 2000).

Seismic data indicates that the shallow Zofar Basin dips toward the north (Frieslander 2000). In addition, a unique reversal in asymmetry occurs (Supplemental Figure 2b,c), where horizons within the southern part of the basin dip toward the west, whereas in the northern section, horizons dip to the east. This is the only known example along the Arava Valley where reversal in asymmetry occurs within a subbasin (instead of between subbasins), and may indicate that the Zofar Basin can be structurally subdivided.

While the dominant motion along the Dead Sea Fault is left-lateral slip, transtensional and transpressional features are often present. Thus the motion is not purely strike-slip. In detail, the structure probably changed as a result of reorganization of plate boundary. Therefore, the structure is characterized by extensional, compressional, and asymmetrical structures varying in size from large scale (defining the general structure of the Dead Sea Fault Valley) to small scale (defining the internal structure).

Oblique strike-slip motion is largely responsible for the formation of the physiographic valley and of the pull-apart basins along the southern part of the Dead Sea Fault. This oblique plate separation resulted in the creation of collapsed structures and features evident in the many lows, i.e., pull apart basins, along the valley floor. Compressional features are caused by a step to the right of the main left-lateral strike-slip fault strands resulting in the formation of transverse structural saddles that separate the different basins. Within the pull-aparts, smaller compressional uplifts and saddles divide them into distinct subbasins.

Another characteristic feature of the Dead Sea Fault Valley is asymmetry, both in the large scale, as exhibited in the differing topography of the eastern and western-flanking cliffs, and the small scale, i.e., within the basins and subbasins. Large-scale asymmetry could have been influenced by the combined effect of nor-mal faulting and isostatic uplift on existing topography (Wdowinski & Zilberman 1996).

On the smaller scale, asymmetry within the pull-apart basins can be explained by transform-normal extension. Overlapping strike-slip faults, which are respon¬sible for the formation of these basins, together with rearrangement of plate motions, can result in the formation of an extensional component [the "leaky transform," termed by Garfunkel (1981)]. This asymmetry is not a local phenomenon, with other pull-apart basins along continental rift zones clearly showing asymmetry toward the main fault. Examples may be found in the Cariaco Basin in Venezuela (Ben-Avraham & Zoback 1992), the Motagua fault system in Guatemala, Lake Baikal, Lake Tanganyika in the East African rift system, and the North Anatolian Fault in Turkey.

Although it seems that motion along the Dead Sea Fault was initiated some 18 Ma ago (Garfunkel 1981, Garfunkel et al. 1981), the basins started to evolve later owing to changes in the fault pattern, sometimes probably related to slight shifts in the pole of rotation between the Arabian plate and the Sinai subplate. This introduced a component of oblique motion (Joffe & Garfunkel 1987), which led to the formation of the pull-aparts. An interesting question is the spatial evolution of the basins along the Dead Sea Fault, i.e., did the southern basins evolve before the northern ones, or did they all develop more or less simultaneously? Also, how did each basin evolve? For example, did the Dead Sea Basin develop from south to north as some researchers have suggested (e.g., ten-Brink & Ben-Avraham 1989), or from the center first and then to the north and south (Ben-Avraham & Schubert 2006), or simultaneously through the entire length (Lazar et al. 2006)?

One of the characteristics of the Dead Sea Fault is the deep seismicity. As demon¬strated by Aldersons et al. (2003, most microearthquakes nucleate at rather deep depths — some 20-30 km below the surface. In this respect, the Dead Sea Fault is quite different from the San Andreas Fault, where most earthquakes occur within the top 15 km of the subsurface, although some anomalously deep crustal earthquakes have been recorded at depths between 20-30 km (Bryant & Jones 1992). This may have to do with the differences in slip-rates between the two fault systems; the slip along the Dead Sea Fault is only approximately one tenth [4 mm year (Wdowinski et al. 2004)] the overall slip along the San Andreas Fault [-50 mm year-1 (van der Woerd et al. 2006)].

Another interesting characteristic is the low values of heat flow exhibited along the Dead Sea Fault, except for the Sea of Galilee, which is located within a large field of Tertiary volcanism and has a relatively high heat flow, and the southern Gulf of Elat, close to the Red Sea. The Dead Sea Basin exhibits the lowest heat flow values along the fault, 38 mW m-2 (Ben-Avraham et al. 1978). Recently, Forster et al. (2007) reported higher vales in southern Jordan. The deep seismically found by Aldersons et al. (2003) suggests that this may not represent the regional heat flux, but this issue requires further study.

The deep seismicity and low heat flow values suggest that the deformation might be brittle in the lower crust. This also supports recent claims (Ginzburg et al. 2007) that some of the transverse faults within the Dead Sea Basin are deep normal faults extending to the basement and no listric faults as previously assumed.

Elias, A., et al. (2007). "Active thrusting offshore Mount Lebanon: Source of the tsunamigenic A.D. 551 Beirut-Tripoli earthquake." Geology 35(8): 755-758.

Morhange, C., et al. (2006). "Late Holocene relative sea-level changes in Lebanon, Eastern Mediterranean." Marine Geology 230(1): 99-114.

Sivan, D., Schattner, U., Morhange, C., Boaretto, E. (2010). "What can a sessile mollusk tell about neotectonics?" Earth and Planetary Science Letters 296(3): 451-458.

Crustal deformation and seismicity in the Levant region are mainly related to the plate-boundary Dead Sea Fault (DSF) and the intraplate Carmel-Gilboa-Faria Fault System (CGFS). The intersection between these two major fault systems is generally treated as an ~35-km-wide deformation belt stretched between the Faria and Gilboa Faults. Here, we present spatial and temporal analysis of faulting near this intersection. Our analysis is based on new geological mapping, new high-resolution airborne light detection and ranging (LiDAR) data, and seismic reflection profiles and indicates northward migration and localization of the intersection over time since the early Miocene. We discovered and mapped outcrops of Miocene, Pliocene, and Pleistocene rock units as well as faults and reconstructed the evolution of deformation. Three main tectonic phases were identified in this area covering the following periods: the early-middle Miocene, the late Mio¬cene-Pliocene, and the Quaternary. During the first phase, the DSF and the CGFS developed, and the CGFS faulted along a series of subparallel grabens and elongated NW-SE, between the southernmost Faria and the northernmost Gilboa faults, over a belt width of —35 km. During the second phase, deformation along the CGFS migrated northward and concentrated at an ~6-km-wide zone in the northern Faria Anticline. During the third stage, small-scale northward migration and localization of the deformation to a width zone of ~1-2 km at the southern boundary of the Beit She'an Valley occurred. Faults from the third phase reveal both sinistral and nor¬mal faulting. We propose that the currently active intersection between the DSF and the CGFS is located east of this localized deformation zone, near a right step of the DSF and the uplifted area of Tel Al-Qarn in the eastern Jordan Valley. We suggest that the northward migration and localization of this intersection are related to regional tectonic changes, spatial variations in the Sinai-Arabia Euler pole, and the localization of deformation along the DSF.

The CGFS, a major intraplate fault within the Sinai plate (Fig. 1), defines the boundary between two major tectonic blocks within the Sinai plate (e.g., Dembo et al., 2015; Gomez et al., 2020; Hamiel and Piatibratova, 2021). The CGFS is ~80 km long, branches from the DSF at the central part of the Jordan Valley segment, and continues in the ~NW direction to the northern tip of Mt. Carmel and farther northwestward into the Mediterranean continental shelf (e.g., Garfunkel and Almagor, 1984; Ben-Gai and Ben-Avraham, 1995; Hofstetter et al., 1996). The Carmel, Gilboa, and Faria Faults comprise the main fault segments of this system (Fig. 1B). The Carmel Fault, the northwesternmost fault segment of the CGFS, is an oblique, sinistral-normal fault (e.g., Ben-Gai and Ben-Avraham, 1995; Achmon and Ben-Avraham, 1997; Sadeh et al., 2012; Dembo et al., 2015). The Gilboa Fault in the north and the Faria Fault in the south are the two main easternmost segments of this system (Fig. 1). They are normal faults that dip in the NE direction (Hatzor and Reches, 1990; Shaliv eta., 1991). Recent geodetic studies (e.g., Sadeh et al., 2012; Hamiel et al., 2016, 2018; Gomez et al., 2020; Hamiel and Piatibratova, 2021) showed that the current strike-slip rate along the DSF decreases from ~5 mm/yr south of the intersection with the CGFS to ~4 mm/ yr north of it, which suggests that slip is transferred from the DSF to the CGFS and that an active triple junction is located at the intersec¬tion between them. These studies also calculated an oblique strike-slip motion of 0.5-1.6 mm/yr across main segments of the CGFS and proposed that the observed reduction in slip rate along the DSF is due to transfer of slip to the CGFS. Hamiel and Piatibratova (2021) demonstrated a relative motion of 0.8 ± 0.4 mm/yr, sub-parallel to the DSF, between the southern and central Sinai tectonic blocks, i.e., south and north of the CGFS. This result was found to be in a good agreement with the CGFS total slip rate vector and the observed reduction in slip rate along the DSF near the intersection with the CGFS.

The intersection between the DSF and the CGFS is generally treated as an ~35-km-wide deformation zone between the Faria and Gilboa Faults and the central Jordan Valley segment of the DSF (e.g., Shaliv et al., 1991; Segev et al., 2014; Dembo et al., 2015). The northern end of this intersection zone is defined by the Beit She'an Valley, which is located between the Gilboa Fault and the DSF (Fig. 1B). However, it is unclear whether the entire zone between the Gilboa and Faria Faults is active, and if not, where exactly this intersection occurs and how it evolved over time since its formation in the Miocene. In this study, we investigate the spatial and temporal variations in the deformational processes of this area and better locate the intersec¬tion of the DSF and CGFS. To accomplish these tasks, we performed an integrated geological and geophysical study of faulting in this region.

The study area extends from the Faria Fault in the south, through the Faria Anticline, and to the southern part of the down-faulted basin of the Beit She'an Valley in the north (Fig. 1C). This area developed as an interaction zone between the NW-trending CGFS and the N-trending tectonic system of the DSF. The axis of the Faria Anticline is in the SSW—NNE direction. The rock formations exposed in this region belong to several stratigraphic groups, ranging from the Jurassic Arad Group to the Quaternary Dead Sea Group. While old rock formations, i.e., Jurassic and Lower Cretaceous rock formations, are exposed at the core of the Faria Anticline, younger rocks of the Upper Cretaceous to Paleocene Mount Scopus and the Eocene Avedat Groups are exposed along the eastern flank of the anticline, forming the western boundary of the Jordan Valley (Mimran, 1984; Shaliv et al., 1991). The Faria Anticline is traversed by several NW-trending faults that demonstrate vertical displacements ranging from 500 m to 800 m. The major faults are the Faria Fault in the south and the southern boundary of the Beit She'an Valley in the north. Between the Faria Fault and Beit She'an Valley, less prominent faults also oriented NW, such as the Buqea and Tayasir Faults (Fig. 1), are observed (Shaliv et al., 1991; Mimran et al., 2016).

The Faria Anticline developed during the late Turonian—early Eocene as part of the Syrian arc structures (Krenkel, 1924) and was reactivated during the Oligocene—early Miocene as part of a tectonic phase that followed plate separation along the Red Sea (Mimran, 1984; Shaliv et al., 1991). Later, during the Miocene, this region was faulted along the north-south-trending DSF and along the NW-trending CGFS to form a major intersection area (Freund et al., 1970; Garfunkel, 1981; Mimran, 1984; Shaliv et al., 1991; Rozenbaum et al., 2016). The simultaneous activity of faulting of the two systems is still ongoing and divides the Sinai Plate into two tectonic domains south and north of the CGFS (Ben-Avraham and Ginzburg, 1990; Hofstetter et al., 1996; Sadeh et al., 2012; Dembo et al., 2015; Gomez et al., 2020; Hamiel and Piatibra-tova, 2021). Early Miocene faulting was accompanied by extensive volcanic activity, which is termed the Lower Basalt (e.g., Schulman, 1962). Near the Gilboa Fault (Fig. 1B), this volcanic phase started at ca. 17.5 Ma (Shaliv et al., 1991). Normal faulting accompanying this phase probably occurred under the same tensile regime that triggered volcanic eruptions during the Lower Basalt period (e.g., Hatzor and Reches, 1990; Shaliv et al., 1991; Dembo et al., 2015). The deposition of the Hordos conglomerate during the early—middle Miocene on the eastern flank of the Faria Anticline marks the development of a deep tectonic basin along the Jordan Valley at that time. Most of the clastic materials deposited within this basin were derived from the eastern flank of the Faria Anticline and expose Late Cretaceous to Eocene rocks. Similar conglomerates were derived from the high, faulted flanks of the Faria, Buqea, and Tayasir Grabens, which indicates the formation of these grabens that dissected the axis of the Faria Anticline in the early Miocene (Bentor, 1961; Schulman and Rosenthal, 1968; Shaliv et al., 1991).

During the late Miocene, a new phase of normal faulting occurred. These faults, striking to the NW-SE direction, ruptured only the northern part of the Faria Anticline and enhanced the tectonic relief along pre-existing faults that originally developed in the early Miocene (Shaliv et al., 1991). One of the larger tectonic basins that developed at this stage is the Beit She'an basin, which accumulated a thick sequence of the late Miocene Bira Formation (ca. 10-7 Ma; Rozenbaum et al., 2019) and was followed by the deposition of the late Miocene to early Pliocene Gesher Formation (ca. 7-5 Ma; Rozenbaum et al., 2019).

During the late Miocene and the early Pliocene, regional extensional deformation occurred in northern Israel. This is evidenced by the extensive volcanic activity that formed the Cover Basalt Formation in northern Israel (Shaliv et al., 1991; Heimann et al., 1996). In the northeastern sector of our study area, near Manna Feiyad (Figs. 2A and 2C), several volcanic bodies of this phase are exposed and dated to 5.65-5.90 Ma (Shaliv et al., 1991; Dembo et al., 2015). A tectonic phase of faulting in the northern Faria Anticline occurred after the deposition of the Cover Basalt. This tectonic phase triggered the incision of the drain-age system on which the early Quaternary Wadi Malih Formation was deposited, especially at the outlet of the Wadi Malih drainage system to the southeastern part of Beit She'an Valley (Fig. 2), forming an extensive fan (Mimran, 1984; Shaliv et al., 1991; Mimran et al., 2016). Farther north, a similar fan was deposited at the outlet of Nahal Bezeq draining the southern flank of the Gilboa Mountains into the Beit She' an Valley (Figs. 1-2; Hatzor, 2000). The late Quaternary is characterized by tectonic faulting along the western margin of the Jordan Valley, followed by deep erosion and later deposition of the late Pleistocene Lisan Formation (e.g., Begin et al., 1974). Recent tectonics are demonstrated by the deformation of Lisan and younger sediments.

Based on the studies mentioned above, Mimran et al. (2016) published an updated version of the geological map of the northern part of the region. However, the southern part of our study area is only described by a regional scale (1:200,000) geological map (Sneh et al., 1998), which in some places does not correlate well with the map of Mimran et al. (2016). Furthermore, previous works (e.g., Schulman and Rosenthal, 1968; Mimran, 1984; Shaliv et al., 1991) and geological maps (Sneh et al., 1998; Mimran et al., 2016) correlate the Neogene sequences exposed in the study area to the Neogene sequence exposed ~30-50 km to the north, near the Sea of Galilee, which was described by Picard (1943) and Schulman (1962). This long-distance correlation ignores the large variety of local facies changes observed within the stratigraphic sequence studied and therefore may cause misleading correlations.

In the present work, we demonstrate that careful examination and revision of this correlation leads to a better understanding of the geological sedimentary sequence and sheds new light on the Neogene-Quaternary tectonic evolution of our study area and in particular on the DSF-CGFS intersection.

Our observations indicate dramatic changes in the tectonic setting of the intersection between the DSF and the CGFS since the late Mio cene. This change is manifested in both the stratigraphy and the faulting architecture of this intersection area. As described above and summarized here, there are several major differences between our new map and previous geological maps of this area: (1) based on lithological characteristics, dating, and field relations, large outcrops along the northern part of the Faria Anticline, which border the Beit She'an Valley and were previously mapped as undivided outcrops of the Hordos and Umm Sabune Formation of early-late Miocene Age (Shaliv et al., 1991; Mimran et al., 2016), were found to belong to younger Neogene 2016), were found to belong to younger Neogene and late Miocene–early Pliocene Gesher Formations. (2) Some of these previously mapped outcrops (Shaliv et  al., 1991; Mimran et  al., 2016) were found to be pseudo-conglomerates of pedogenic origin, locally known as “Nari” of Pliocene to early Quaternary age, which formed due to weathering and pedogenesis on the exposed outcrops. (3) New outcrops of the Quaternary Wadi Malih Formation were found and mapped. (4) Based on lithological characteristics and field relations, lage outcrops along and near the Faria Graben, which were previously mapped as unclassified Neogene–Quaternar units (Sneh et al., 1998), were found to belong to the early–middle Miocene Hordos Formation. (5) New outcrops of the Hordos Formation com-posed of polymictic conglomerates cemented by carbonates were found and mapped in the southern part of the Faria Anticline, along and between the Faria and Buqea Grabens. (6) New faults that rupture Neogene and Quaternary units were found and mapped, especially in the northern sector of the Faria Anticline. Most of them are normal faults, and two are oblique faults, which demonstrate a combination of sinistral and normal faulting. (7) Faults that rupture Pliocene and Quaternary units were only found along the northern part of the Faria Anticline and the south-ern part of the Beit She’an Valley. We show that since the late Miocene, the fragmentation of the Sinai Plate along the CGFS has been accompanied by northward migration and localization of deformaWe show that since the late Miocene, the fragmentation of the Sinai Plate along the CGFS has been accompanied by northward migration and localization of deformation to the southern boundary of the Beit She'an Valley. Figure 13 summarizes the spatial distribution of the main phases of tectonic evolution. During the early—middle Miocene, the CGFS was composed of an —35-km-wide deformation belt that stretched from the Faria to Beit She'an Valleys. This belt was dominated by normal faulting and created NW-trending grabens, such as Faria, Buqea, and Tayasir Grabens (Fig. 13). Then, during the late Miocene and the Pliocene, tectonic activ¬ity migrated northward to a deformation belt of —6 km in the northern Faria Anticline (Fig. 13). Since the late Pleistocene, the deformation has been localized to a zone of —1-2 km along the southern boundary of the Beit She'an Valley (Fig. 13). At this stage, both normal and sinistral faulting can be clearly identified (Figs. 6, 10, and 12). A major fault in this localized zone is the Wadi Malih Fault, which ruptured the late Pleistocene Lisan Formation (Figs. 3, 6, 8, and 13). It marks a clear lineament that crosses the Jordan Valley as it branches from the DSF (at the eastern side of the Jordan Valley), where a right-lateral step and a change in the strike of the DSF are observed near the uplifted (push-up) area of Tel Al-Qarn (e.g., Ferry et al., 2007; Figs. 2A and S1). This fault defines the current boundary between the Faria Anticline and Beit She' an Valtion to the southern boundary of the Beit She’an Valley. Figure 13 summarizes the spatial distribution of the main phases of tectonic evolution. During the early–middle Miocene, the CGFS was composed of an ~35-km-wide deformation belt that stretched from the Faria to Beit She’an Valleys.

We show that since the late Miocene, the fragmentation of the Sinai Plate along the CGFS has been accompanied by northward migration and localization of deformation to the southern boundary of the Beit She'an Valley. Figure 13 summarizes the spatial distribution of the main phases of tectonic evolution. During the early—middle Miocene, the CGFS was composed of an ~35-km-wide deformation belt that stretched from the Faria to Beit She'an Valleys. This belt was dominated by normal faulting and created NW-trending grabens, such as Faria, Buqea, and Tayasir Grabens (Fig. 13). Then, during the late Miocene and the Pliocene, tectonic activity migrated northward to a deformation belt of ~6 km in the northern Faria Anticline (Fig. 13). Since the late Pleistocene, the deformation has been localized to a zone of ~1-2 km along the southern boundary of the Beit She'an Valley (Fig. 13). At this stage, both normal and sinistral faulting can be clearly identified (Figs. 6, 10, and 12). A major fault in this localized zone is the Wadi Malih Fault, which ruptured the late Pleistocene Lisan Formation (Figs. 3, 6, 8, and 13). It marks a clear lineament that crosses the Jordan Valley as it branches from the DSF (at the eastern side of the Jordan Valley), where a right-lateral step and a change in the strike of the DSF are observed near the uplifted (push-up) area of Tel Al-Qarn (e.g., Ferry et al., 2007; Figs. 2A and S1). This fault defines the current boundary between the Faria Anticline and Beit She'an Valley and shows clear evidence of normal faulting in the subsurface data (Fig. 12B). Farther NW of our study area, the observed post-late Miocene faults are connected to the Gilboa Fault (Fig. 1; Shaliv et al., 1991; Sneh et al., 1998; Dembo et al., 2015).

The localization of deformation near the Beit She'an Valley and the Gilboa Fault agrees with current GPS observations (Hamiel and Piatibra-tova, 2021). The results of this study highlight the significant deformation near the intersection of the DSF and the CGFS and suggest that most of the deformation in the eastern section of the CGFS occurs along the southern boundary of the Beit She'an Valley (i.e., near the Wadi Malih Fault and the Mehola Fault), and the current contribution to deformation of the Faria Fault is negligibly small. The localization of deformation near the Beit She'an Valley and the Gilboa Fault is also in agreement with paleomagnetic observations and mechanical models (Dembo et al., 2015). Dembo et al. (2015) showed localization of deformation near the Cannel and Gilboa Faults at sites younger than ca. 8 Ma. Previous studies suggest that dramatic changes in plate kinematics occurred in the Levant during the late Miocene—early Pliocene (e.g., Garfunkel, 1981; Joffe and Garfunkel, 1987; Marco, 2007). Such studies divided the deformation along the DSF and surrounding areas into two main phases: before and after —5 m.y. ago. At this transition stage, a shift in the location of the Sinai-Arabia Euler pole took place, leading to changes in the style and rate of deformation as well as the internal structure and localization of the DSF system (e.g., Garfunkel, 1981; Joffe and Garfunkel, 1987; Marco, 2007). Farther north of our study area, along the DSF and within the Sea of Galilee and Hula depressions, studies have shown that changes and reorganization of the main and marginal faults occurred at ca. 4-5 Ma (e.g., Heimann and Ron, 1993; Hurwitz et al., 2002; Schattner and Weinberger, 2008; Heimann et al., 2009; Matmon and Zilberman, 2017). Other studies suggest that major changes in the regional plate tectonics and the structure of the DSF occurred at ca. 10 Ma. Around this time, the collision of Arabian and Eurasian plates along the Bitlis suture initialized in SE Anatolia (e.g., McQuarrie and van Hinsbergen, 2013), the DSF was fully developed as a plate boundary (e.g., Gomez et al., 2020), and the tectonic activity along the intraplate Sinai—Negev Shear Zone terminated (Weinberger et al., 2020). Similar to these studies, in our study area, a major transition in tectonic deformation was found to have started at ca. 10 Ma, when deposition of the Bira Formation first began. Spatial analysis of the new geological map (Figs. 3, 8, and S2) shows the migration of post-late Miocene faulting from SW to NE, which deepened the Beit She'an Valley toward the NE. The volcanic activity in the SE sector of our study area (Fig. 3), which is dated to 5.65-5.90 Ma (Shaliv et al., 1991; Dembo et al., 2015), is probably related to the same tectonic phase that started during the late Miocene and continued until the Pliocene. The strike of most post-late Miocene faults is in the NW direction, some are in the NNW direction, and a small minority of faults are in the NE direction. Many of the NW faults are probably rejuvenated faults that initiated during the early—middle Miocene tectonic phase or even before, as indicated by the early stages of faulting along the CGFS (e.g., Shaliv et al., 1991; Segev et al., 2014). Finally, we propose that the northward migration and localization of the intersection between the CGFS and the DSF that has occurred since the late Miocene is probably related to the above-mentioned regional tectonic and stress field changes, the transition in the Sinai-Arabia Euler pole location, and the localization of deformation along the DSF. We also propose that our observations, along with current GPS measurements (e.g., Gomez et al., 2020; Hamiel and Piatibratova, 2021), suggest the fragmentation of the Sinai plate since the late Miocene.

To address one of the central questions of plate tectonics — How do large transform systems work and what are their typical features? — seismic investigations across the Dead Sea Transform (DST), the boundary between the African and Arabian plates in the Middle East, were conducted for the first time. A major component of these investigations was a combined reflection/refraction survey across the territories of Palestine, Israel and Jordan. The main results of this study are:

1. The seismic basement is offset by 3–5 km under the DST
2. The DST cuts through the entire crust, broadening in the lower crust
3. Strong lower crustal reflectors are imaged only on one side of the DST
4. The seismic velocity sections show a steady increase in the depth of the crust-mantle transition (Moho) from ~26 km at the Mediterranean to ~39 km under the Jordan highlands, with only a small but visible, asymmetric topography of the Moho under the DST.

The DST is a system of left-lateral strike-slip faults that accommodate the relative motion between the African and Arabian plates. Except for a mild compressional deformation starting about 80 Ma, the larger Dead Sea region has remained a stable platform since the early Mesozoic. Approximately 17 Ma, this tectonic stability was interrupted by the formation of a transform, the DST, with a total left-lateral displacement of 105 km until today (Quennell 1958; Freund et al. 1970; Garfunkel 1981).

The crystalline basement of the area represents the NW part of the Arabo-Nubian Shield (ANS) and consists of mainly juvenile Late Proterozoic rocks (Stoeser & Camp 1985; Stern 1994). Regarding both isotopic and chemical data of xenoliths, the involvement of older crustal material in the lower crust of the ANS seems improbable (Henjes-Kunst et al. 1990; Stern 1994; Ibrahim & McCourt 1995). There is a gradual transition from the continental crust of the ANS with thicknesses of 35–40 km (El-Isa et al. 1987; Makris et al. 1983; Al-Zoubi & Ben-Avraham 2002) to the crust of the eastern Mediterranean, that is assumed to be partly underlain by typical oceanic crust with thicknesses smaller than 10 km (Ginzburg et al. 1979a; Makris et al. 1983; Ben-Avraham et al. 2002), see also Fig. 1(b).

The Precambrian basement is usually overlain by an Infracambrian to Early Cambrian volcano-sedimentary succession of variable thickness. Whereas coarse-grained clastics (Saramuj and Elat conglomerates) are restricted to fault-bounded basins, fine-grained clastics, mostly consisting of arkosic sandstones and associated volcanic rocks (Zenifim Formation, Haiyala Volcaniclastic Unit and equivalent rock units) have been observed in large parts of the Israel and Jordan subsurface (Weissbrod & Sneh 2002). In boreholes close to the WRR and NVR profiles (Fig. 1b) the Zenifim Formation was determined to be several hundreds of metres thick (> 500 m in Maktesh-Qatan, MQ, and >2000 m in Ramon-1, R1), although its base has not been encountered in any of the Israeli boreholes.

The position of the study region at the NW edge of the ANS, i.e. at a passive continental margin since early Paleozoic times, is reflected in facies changes and varying sedimentary thicknesses along the seismic profiles (Fig. 2). The Phanerozoic along the northwestern part of the profile is dominated by Cretaceous and Tertiary rocks underlain by Jurassic, Triassic and Permian sequences that thin out towards the east. East of the DST, however, Permian to Triassic strata are missing and Lower Cretaceous rocks unconformably overlie Ordovician and Cambrian sandstones. The crystalline basement rocks (calc-alkaline granitoids and rhyolites) cropping out in the Jebel Humrat Fiddan area east of the DST are thought to be equivalent to the basement rocks of the Timna region in Israel, near Elat.

Previous crustal-scale wide-angle reflection/refraction experiments in the study area (Fig. 1b) include that in Israel in 1977 (Ginzburg et al. 1979a), the onshore-offshore experiment between the northwestern end of the WRR profile and Cyprus in 1978 (Makris et al. 1983; Ben-Avraham et al. 2002) and that in Jordan in 1984 (El-Isa et al. 1987). However, there was no profile, which crossed the DST to provide a complete image across this structure. Moreover, deep seismic reflection data only exist from the area between the Dead Sea and the Mediterranean (Yuval & Rotstein 1987; Rotstein et al. 1987), and high-resolution seismics (Frieslander 2000) are mostly limited to the western part of the Arava Valley. Further south within the Afro-Arabian rift system seismic profiles, which cross the rift structures, have proved to be very useful to understand the crustal structure as, for example, the E-W profiles crossing the Kenya rift (Maguire et al. 1994; Braile et al. 1994; Birt et al. 1997). Seismic crustal investigations of the Afro-Arabian Rift system are summarized by Mechie & Prodehl (1988) and Prodehl et al. (1997).

Our study provides the first whole-crustal image across the Dead Sea Transform (DST), one of the Earth’s major transform faults. Under the Arava Fault (Fig. 11), the main fault of the southern DST system, the seismic basement is offset by several kilometres, but the Moho depth increases steadily from ~26 km at the Mediterranean to ~39 km under the Jordan highlands, except for a small but visible, asymmetric topography under the Arava Valley. The general trend of continuous Moho-depth increase is confirmed by the interpretation of potential field data (Al-Zoubi & Ben-Avraham 2002) and the results of a receiver functions study (A. Mohsen, personal communication, 2003). Based on the interpretation of the NVR data, we infer that the AF cuts through the crust, becoming a broad zone in the lower crust, and reaches down to the mantle. This agrees with the results of the thermo-mechanical modelling of Sobolev et al. (2003) and the SKS-splitting observations by Rumpker et al. (2003), which suggest that the DST cuts through the whole lithosphere, thus accommodating the motion between the African and the Arabian plates (Fig. 1). The lack of significant uplift of the Moho under the Arava Valley speaks strongly against a potential rift structure in this area. We therefore conclude that rift-type deformation (fault perpendicular extension) did not play a dominant role in the dynamics of the DST, a fact corroborated again by the results of Sobolev (unpublished data). Although the shallow structure at the DST differs significantly from the structure at the San Andreas Fault, the deep reaching deformation zones accompanied by a strong asymmetry in subhorizontal lower crustal reflectors appear to be similar for both fault zone systems. We therefore suggest that these deep features are common for large continental transform plate boundaries

Normal Size

Description	Image	Source
Fig. 1a -	Fig. 1a Seismic experiments in the Middle East. The 260 km long wide-angle reflection/refraction profile (WRR, dark blue dots) crosses Palestine, Israel and Jordan. The near-vertical seismic reflection profile (NVR, cyan) coincides with the inner 100 km of the WRR. A red line and two red arrows indicate the Dead Sea Transform (DST) between the Dead Sea and the Red Sea. The white arrows indicate the left-lateral motion of 105 km between the African and Arabian plates. Red stars mark large earthquakes. (Inset) Tectonic setting of the DST. Weber et al. (2004)	Weber et al. (2004)
Fig. 1b -	Fig. 1b Previous wide-angle reflection / refraction experiments in the study area (dashed black; Ginzburg et al. 1979a; Ginzburg et al. 1979b; Makris et al. 1983; El-Isa et al. 1987) together with the WRR profile of DESERT (blue) and the DST (red). The black circles are boreholes used in the interpretation HD = Helez Deep-1A MQ = Maktesh-Qatan R1 = Ramon-1 EJ = El-Jafr-1 Weber et al. (2004)	Weber et al. (2004)
Fig. 2 -	Fig. 2 Location map for the WRR experiment. During the experiment 13 shots (red stars) were executed and recorded by 99 three-component instruments (blue dots) spaced 1–4.5 km apart along the whole length of the profile and 125 vertical component geophone groups (black dots between shots 5 and 51) with 100 m spacing along a 12.5 km long section of the profile in the Arava Valley. Simplified geological map of the NVR profile area (modified after Sneh et al. 1998; Bartov et al. 1998). The red line is the smoothed reflection line of the NVR experiment. Six dominant faults (Saad Nafha, Ramon, Baraq, Zofar, Arava, Al Quwayra) are indicated in black. Geological cross-section along the NVR profile. The arrow indicates the Arava Fault (AF), the dominant fault of the DST. Weber et al. (2004)	Weber et al. (2004)
Fig. 11 -	Fig. 11 2-D P-velocity model (velocities in km s−1) for the WRR experiment. The shots (triangles at top) were recorded by 99 three-component instruments along the whole profile and 125 vertical component geophone groups in the Arava Valley (see also Fig. 2a). For accuracy estimates of velocities and interface depths see text. Only the central area inside the diagonal lines is resolved in this study. Outside this area the model is based on previous work (Fig. 1b; Ginzburg et al. 1979a; Ginzburg et al. 1979b; Makris et al. 1983; El-Isa et al. 1987). The hatched symbols near the Moho, the border between yellow and red colours, indicate the location of bands of strong reflections in the NVR migrated section (b). Note vertical exaggeration of 2:1. Automatic line drawing of the depth-migrated seismic CDP section of the NVR experiment (Fig. A5). The red band indicates the location of the Moho (crust/mantle boundary) as derived from the WRR experiment in (a). The black arrows mark the break-off of reflectivity, generally interpreted as the Moho in seismic reflection data. Weber et al. (2004)	Weber et al. (2004)
Fig. 13 -	Fig. 13 Sketch of the DST dynamics along the DESERT profile, based on results shown in Fig. 11 and the results of Sobolev (unpublished data). Crustal structure before the beginning of transform motion, i.e. 17 Ma ago. The Phanerozoic sedimentary cover is shaded, dark shading indicates sections which today are facing each other and were crossed by the DESERT profile. Transform motion of ~105 km results in a crustal cross-section with a significantly different structure east and west of the Arava Fault. The shaded domain shows the region of ductile shear deformation (Sobolev, unpublished data). Wavelike symbols indicate lower crustal regions of high reflectivity (Fig. 11) possibly related to lower crustal flow. Associated minor tension (~4 km) produces subsidence and flexure in the western block accommodated by uplift of the eastern block, with a similar small amplitude flexure at the Moho (Fig. 11). Erosion and sedimentation produces the present day structure. Weber et al. (2004)	Weber et al. (2004)

Magnified

Description	Image	Source
Fig. 1a -	Fig. 1a Seismic experiments in the Middle East. The 260 km long wide-angle reflection/refraction profile (WRR, dark blue dots) crosses Palestine, Israel and Jordan. The near-vertical seismic reflection profile (NVR, cyan) coincides with the inner 100 km of the WRR. A red line and two red arrows indicate the Dead Sea Transform (DST) between the Dead Sea and the Red Sea. The white arrows indicate the left-lateral motion of 105 km between the African and Arabian plates. Red stars mark large earthquakes. (Inset) Tectonic setting of the DST. Weber et al. (2004)	Weber et al. (2004)
Fig. 1b -	Fig. 1b Previous wide-angle reflection / refraction experiments in the study area (dashed black; Ginzburg et al. 1979a; Ginzburg et al. 1979b; Makris et al. 1983; El-Isa et al. 1987) together with the WRR profile of DESERT (blue) and the DST (red). The black circles are boreholes used in the interpretation HD = Helez Deep-1A MQ = Maktesh-Qatan R1 = Ramon-1 EJ = El-Jafr-1 Weber et al. (2004)	Weber et al. (2004)
Fig. 2 -	Fig. 2 Location map for the WRR experiment. During the experiment 13 shots (red stars) were executed and recorded by 99 three-component instruments (blue dots) spaced 1–4.5 km apart along the whole length of the profile and 125 vertical component geophone groups (black dots between shots 5 and 51) with 100 m spacing along a 12.5 km long section of the profile in the Arava Valley. Simplified geological map of the NVR profile area (modified after Sneh et al. 1998; Bartov et al. 1998). The red line is the smoothed reflection line of the NVR experiment. Six dominant faults (Saad Nafha, Ramon, Baraq, Zofar, Arava, Al Quwayra) are indicated in black. Geological cross-section along the NVR profile. The arrow indicates the Arava Fault (AF), the dominant fault of the DST. Weber et al. (2004)	Weber et al. (2004)
Fig. 11 -	Fig. 11 2-D P-velocity model (velocities in km s−1) for the WRR experiment. The shots (triangles at top) were recorded by 99 three-component instruments along the whole profile and 125 vertical component geophone groups in the Arava Valley (see also Fig. 2a). For accuracy estimates of velocities and interface depths see text. Only the central area inside the diagonal lines is resolved in this study. Outside this area the model is based on previous work (Fig. 1b; Ginzburg et al. 1979a; Ginzburg et al. 1979b; Makris et al. 1983; El-Isa et al. 1987). The hatched symbols near the Moho, the border between yellow and red colours, indicate the location of bands of strong reflections in the NVR migrated section (b). Note vertical exaggeration of 2:1. Automatic line drawing of the depth-migrated seismic CDP section of the NVR experiment (Fig. A5). The red band indicates the location of the Moho (crust/mantle boundary) as derived from the WRR experiment in (a). The black arrows mark the break-off of reflectivity, generally interpreted as the Moho in seismic reflection data. Weber et al. (2004)	Weber et al. (2004)
Fig. 13 -	Fig. 13 Sketch of the DST dynamics along the DESERT profile, based on results shown in Fig. 11 and the results of Sobolev (unpublished data). Crustal structure before the beginning of transform motion, i.e. 17 Ma ago. The Phanerozoic sedimentary cover is shaded, dark shading indicates sections which today are facing each other and were crossed by the DESERT profile. Transform motion of ~105 km results in a crustal cross-section with a significantly different structure east and west of the Arava Fault. The shaded domain shows the region of ductile shear deformation (Sobolev, unpublished data). Wavelike symbols indicate lower crustal regions of high reflectivity (Fig. 11) possibly related to lower crustal flow. Associated minor tension (~4 km) produces subsidence and flexure in the western block accommodated by uplift of the eastern block, with a similar small amplitude flexure at the Moho (Fig. 11). Erosion and sedimentation produces the present day structure. Weber et al. (2004)	Weber et al. (2004)