Structural expression of a fading rift front, a case study from the Oligo-Miocene Irbid rift of northwest Arabia

. 10 Not all continental rifts mature to form a young ocean. The mechanism and duration of their cessation depend on the crustal structure, modifications in plate kinematics, lithospheric thermal response, or intensity of sub-crustal flow (e.g., plume activity). The cessation is recorded in the structure and stratigraphy of the basins that develop during the rifting process. This architecture is lost due to younger tectonic inversion, severe erosion or even burial into greater depths that forces their detection by low-resolution geophysical imaging. The current study focuses on a uniquely preserved Oligo-Miocene rift that was 15 subsequently taken over by a crossing transform fault system and mostly due to that died out. We integrate all geological, geophysical and results from previous studies from across the Southern Galilee to unravel the structural development of the Irbid failing rift, of Northwest Arabia. Despite tectonic, magmatic and geomorphologic activity postdating the rifting, its subsurface structure northwest of the Dead Sea Fault is preserved at depths of up to 1 km. Our results show that a series of basins subsided at the rift front, i.e. rift termination, across the southern Galilee. We constrain the timing and extent of their 20 subsidence into two main stages, based on facies analysis and chronology of magmatism. Between 20-9 Ma grabens and half-grabens subsided within a larger releasing jog, following an NW direction of a deeper presumed Principal Displacement Zone. The basins continued to subside until a transition from the transtenssional Red Sea to the transpressional Dead Sea stress regime occurred. With the transition, the basins ceased to subside as a rift, while the Dead Sea Fault split the jog structure. Between 9-5 Ma basin subsidence accentuated and an uplift of their margins accompanied their overall elongation to the NNE.

Rifting cessation may result from modifications in plate kinematics, or in lithospheric thermal re-equilibration (e.g., along the Ordovician-Silurian Transbrasiliano lineament; Oliveira and Mohriak, 2003). It could also reflect a decay in plume intensity (e.g, Delhi basin; Sharma, 2009) or variations in rheological properties (Lyakhovsky et al., 2012). In this case, the extensional 5 strain is accommodated by localized deformations over a wider region than the original rift axis (Van Wijk andBlackman, 2005, Segev et al., 2014).
The mechanism and duration of the cessation vary from one case to another. A rapid stop reacts to extensional stress decay, acquaintance with a more rigid crust, or a newly established stress regime, different enough to mute the rifting process. Fading away of the dominant rifting stress leads to attenuation and eventually rift failure. In the Potiguar rift (Brazil) case, Precambrian 10 basement faulting patterns dictated the Neocomian-Barremian syn-rift grabenization style. Magnetic, gravity and resistivity data track transform boundaries inside an intraplate setting, generating fault-controlled depressions. Both the NE-trending (parallel to rift axis) oblique-slip faults and the NS-trending en-echelon normal faults die out in the post-rift sedimentary units (de Castro and Bezerra, 2015). In southeastern Australia, propagation of the transform fracture zone cuts across preexisting basement structures. Folds and foliations of previous structural stages present unfavourable orientations for reactivation under 15 the present stress field (Lesti et al., 2008).
Preservation of failed rift structures in the geological record depends on the intensity and efficiency of later tectonic and erosion processes. In cases, the internal architecture of the basins comprising a failed rift may be lost due to tectonic inversion, severe erosion or even burial into greater depths (Beauchamp et al., 1996;Guiraud and Bosworth, 1997;Beauchamp et al., 1999;Dézes et al., 2004). The reconstruction of the architecture depends on the geophysical imaging resolution (d' Acremont et al., 20 2005;Enachescu, 2006;de Vicente and Muñoz-Martin, 2013;Melo et al., 2016). The current study focuses on the structural development of a rift front during its failure and later preservation. We concentrate on the Irbid Rift that developed across the Arabian plate and into the Sinai sub-plate during the Oligocene-Miocene (Schattner et al., 2006a;Segev et al., 2014;Fig. 1).
Despite tectonic, magmatic and geomorphologic activity post-dating the rifting, the original subsurface structure of the failed rift is preserved at depths of up to 1 km (Fig. 1). 25
Compressional stresses kept the margin at shallow depths, while the syn-tectonic chalks of the Santonian-Paleocene Mount Scopus Gr. covered the late Cretaceous relief. During the Paleogene-Eocene tectonic and thermal quiescence led to vertical 20 subsidence of NW Arabia. The resulting transgression submerged the entire Galilee under more than 1000 m of ocean water.

Morpho-tectonics of the southern Galilee basins
The southern Galilee Neogene basins extend across ~50 km, between the DSF and the Levant continental margin (Fig. 3). The Carmel-Gilboa and Zurim fault systems in the south and north (respectively) bound the Southern Galilee basins (Schattner et al., 2006a, b). Their N-S extent narrows westwards from ~35 to ~10 km in a low relief that exhibits sporadic highs dividing 15 local valleys. The surface of westernmost, Yizre'el (B2) lays at 30-70 m above sea level. To the east Kesulot (B3) and Taanach (B5) valleys are at 60-100 m, Harod valley (B6) is between and Bet Shean (B7) is at -250 m. The low relief of the southern Galilee basins divides between two segments of the Mesozoic Syrian arc fold belt (Krenkel, 1924) that raised by ~500 m since the Pliocene. Lower Cretaceous (Kurnub Group) and Jurassic (Arad Group) exposures appear in limited areas.
The current study integrates all the previous results with unpublished data to address fundamental questions regarding the origin and development of the lower Galilee: is it a single continental basin that accumulated sediments from its surrounding rims (Picard, 1943;Schulman, 1962;Shaliv, 1991)? Alternatively, maybe a full graben bounded by longitudinal faults, Zurim and Carmel-Gilboa from the north and south respectively (as suggested by Kafri and Ecker, 1964;Mero, 1983) or possibly a 5 couple of half grabens bounded by these faults (as proposed by May 1987; Matmon et al., 2003)? What is the structural and tectonic association between the southern Galilee basins development and the nearby DSF and Levant continental margin? In what manner does the structural development of the southern Galilee basins relate to the regional volcanic events?

Dataset and Methodology
Geological reconstruction of the structure and development of the southern Galilee basins relies on an integrated interpretation 10 of all available geophysical and geological datasets from the study area. The new database was constructed on the Kingdom Suite (IHS) platform. It includes 85 multi-channel seismic reflection profiles, 506 boreholes, outcrop data, and previous interpretations. The seismic reflection data were acquired between the 1970's through 2000's. The profiles cover a total length of 800 km. The average penetration depth is 500-1000m below the seismic datum (sea level). The boreholes depth ranges between 35-2390 m below surface. Seismic resolution enables the interpretation of geological units starting from the upper 15 Cretaceous (Fig. 2).
Previous geological mapping was used to extend the structural model from sea level datum (elevation of 0m) up to the presentday topography (30-550m asl). Using a 2000 m/sec velocity for the shallow, near-surface beds (weathered beds), enabled correlation between depth and time domains. Completion of the structural model relied upon digitization of truncation surfaces 25 from previous studies in ArcMap (ESRI) (Weiler, 1968;Dicker, 1969;Dekel, 1988;Shaliv, 1991;Shaliv, 2003;Sneh, 2008).
Outcropping truncated surfaces are considered as layers within a specific unit rather than its top (due to erosion). Control points were added from boreholes. Integration of all datasets yielded a coherent database and a three-dimensional geological grid model of the Galilee subsurface, extending from a depth of 2500m to the present-day surface topography.

Results
The results section describes the sedimentary fill of the basins in chronological order. It is followed by a description of the structural elements. Local names of the basins and the structural highs are abbreviated to simplify the description (Fig. 3). The geographical location of all sites mentioned in text appear in the Google Earth TM supplementary material, herein referred to as GE. The sedimentary fill is bounded between two temporal and structural markers. The basin floor is marked by the Oligo-5 Analysis of the entire database indicates that the type section is located along the axis of the Southern Galilee basins (B2, B4, 10 B6, and B7). Further details from B3, B5, B8-B9 basins complete the section. Additional information from B1, B10, and B11 is provided in the discussion.

Basin fill
The oldest formations deposited above the RTS are the contemporaneous Lower Basalt and Hordos Fms. (Fig. 2). Today, these formations appear in the subsurface, and also outcrop across marginal areas and local highs (Fig. 6). The Hordos Fm. predates the Lower Basalt Fm., yet their seismic appearance is similar. They resemble in reflection frequency, amplitude, and continuity. 5 Some differences between these formations appear in parts of B7. Seismic and borehole data ( Numerous seismic, borehole and outcrop datasets indicate that the Lower Basalt Fm. generally thickens towards the center of each of the basins (Fig. S6). The thickening is also indicated by the arrangement of main faults, dikes, and volcanic feeders 15 (Figs. 5,10,11,. In B3-7 and H2 the thickness exceeds 100 m. In B2, the Lower Basalt Fm. fills a Cretaceous syncline while onlapping its flanks. It thickens from a few meters over H1 to a constant ~125 m at the center of B2. The thickness of the Lower Basalt Fm. reaches 400-600 m adjacent to H2 ( Fig. 10). At the western part of B4, a borehole crossed 630 m of the Lower Basalt Fm. (Table 1). However, this is a minimal value since the base of the formation has not been reached. B3 is divided into two sub-basins by H2. The eastern part of B3 accumulated 50-100 m of Lowe Basalt Fm., while 20 the western part accumulated at least 350 m (base of the formation was not reached). In the eastern border of B4 and B5, the  8,11). Bira Fm. covers this unconformity over the H2, H3 structural highs and across B6 (Fig. 9). In the eastern Galilee (B8-9) and B7, the top Lower Basalt unconformity is either directly overlain by the Cover Basalt Fm. at 30 elevated terrains (e.g. Yisachar-Gazit and Hashita-Geva blocks of B8 (location: Fig. 10a)  The clastic formations of the Dead Sea Gr. overlie the truncated top of Lower Basalt Fm. (Figs. 8,9,11,Figs. S1-6, S8, S10, S14). Data indicate that the group accumulated during the upper Miocene-Pliocene in a lacustrine/fluvial environment. 5 Appearances of lumachelle ostracods at the Bira Fm. indicate an episodic connection to the marine environment. Interchanging paleosol horizons and volcanic remains crossed in boreholes point to exposed continental environments. Um Sabune Conglomerate Fm. overlies Lower Basalt Fm. at H1, the margin of B2 (Kishon 1 borehole, Fig. 8, GE), and in the eastern Galilee. The conglomerates appear near the margins of the basins and volcanic centers. They are bounded by the intersection between Gevat and Nazareth faults (Fig. S7). Um Sabune Conglomerate Fm. contains basaltic pebbles derived from the Lower 10 Basalt Fm., as well as alluvial carbonate and basaltic pebbles that experienced extensive mechanical reworking.
The Clay Series Fm. is contemporaneous to Um Sabune Conglomerate Fm. (Figs. 2, Fig. S7). The grain size of both formations decreases upwards as well as towards the depocenters of each basin. The geographic coverage of these formations defines the present spatial extent of basins B2-6 (Fig. 3). The Clay Series Fm. appears at the center of B2-B7. In places, it directly overlays the Lower Basalt Fm. (e.g., Taanach 4 borehole, Fig. S3, GE). Its thickness is relatively constant along the axis of the central 15 basins B2 (400 m), B4 (200 m) and it reduces towards B6. In more peripheral areas it ranges around tens of meters (Figs. 8,9, 11, S1-6, Table 1). The thickness differences may point to differential subsidence while deposition.   Lower Basalt Fm. is covered by three younger formations: Bira Fm., Gesher Fm. and locally by the Cover Basalt Fm. (Fig. 2).
Seismic resolution does not allow to differ between the Bira Fm. and the Gesher Fm. so these two units are generally termed Bira Fm. in seismic profiles shown here. The Bira Fm. consists mostly of marls, but also of marine and lacustrine limestones, 15 gypsum and salt. Its thickness ranges between 0-200 m (Fig. 9, Fig. S5). Bira Fm. also overlies Um Sabune Conglomerate and Clay Series Fms. in places (Figs. 2,5,(8)(9)11,Figs. S1,. In seismic data Bira Fm. appears as a continuous set of reflectors, detectable across the basins (Figs. 5, Fig. S4) even in folded and faulted regions (Figs. 8,11). Reflectors at the base of the formation onlap an unconformity. The top of Bira Fm. is an unconformity surface (Fig. S4). In places, it is overlaid with paraconformity by the Cover Basalt Fm. (Fig. 9). Bira Fm. is missing over topographic and structural highs (Figs. 5,9,Fig. 5 S5).

Faults
Three types of faults appear in the database: (1) major marginal faults that bound the southern Galilee basins from north and south; (2) faults dividing between basins, sub-vertical to the basin axis. Their orientation ranges between NE to NNE; and (3) Through-going faults that cross the basins. The current study focuses on the first two types, while the third is at the center of Wald et al. (under review). 5

Major faults
Three major marginal faults define the southern rim of the Southern Galilee Basins. In the NW, the Carmel fault down-throws B1 by ~1500 m. Further ESE, a series of normal faults, includes the Yoqneam fault, whose downthrown side is B2. The throw decreases southeastwards from ~200 m to ~50 m (Figs. 7, 10C, GE). The trace of Yoqneam fault diminishes to the SE until it 10 intersects with Gideon and Hayogev faults in Megiddo region (western margin of B4-5; Fig. 8). The Umm El-Fahm fold plunges NE towards B5, where it appears at a depth of 150-200 m below surface (Fig. S9). Given the poor seismic imaging, a southern bounding fault is marked as suspected (Figs. 10b-c, Figs. S10, S11). However, this discontinuity of reflectors may be ascribed to an apparent structural throw, termed Dotan flexure herein (Figs. 10c, 13, Fig. S10) between Umm El-Fahm anticline (Fig. 3, Figs. S10, S11) and Shekhem syncline (Fig. 3, Fig. S9), of the upper Cretaceous Syrian Arc fold belt (Fig. 3). 15 The amount of displacement increases again along the Gilboa fault in the southeast. The Gilboa fault extends from the middle of H2 southeastwards (Fig. 10, Figs. S9, S11). In the NW the Gilboa fault appears in the subsurface of northern B5, where the entire package of reflectors of the basin fill is dipping northwards, towards B4. It downthrows B4 by 400 m relative to B5. The fault downthrows B6 about the Gilboa block footwall (Figs. 7, Fig. S11, GE). The fault is detectable across the shallow subsurface, up to the seismic datum (mean sea level), and exposed in places. This suggests it was active at least through the 20 Plio-Pleistocene. In the east, the Gilboa fault also appears in the subsurface of B7, where it forms a flower structure, attesting to a lateral component of displacement. Vertical displacement along the fault is in the range of 100 m (Fig. 10, Fig. S11). In the southeast, Tayassir, Bardala, and Bezeq faults bound B7 from the south (Figs. 10, Figs. S9, S11, S14). These faults divide between the basin and the NNE trending Faria anticline that plunges from the south. At the eastern boundary of the study area, DSF truncates the eastern part of B7 (Fig. S14). 25 The northern border of the Galilee basins is the E-W trending Bet Hakerem fault system (including Zurim escarpment) and Ahihud fault (e.g., Matmon et al., 2003;Schattner et al., 2006b;Figs. 3, 10c). The Neogene basins mapped here pinch out northwards and do not reach these faults. Therefore, the E-W trending Tur'an, Bet Netofa and Bet Hakerem valleys are excluded from the current analysis (locations: Fig. 3). A series of NW to W trending faults divides between the latter E-W valleys and the Neogene basins. The western segment of Bet Qeshet fault borders H1 from the north. Further east, three step 30 faults downthrow B2 (Zarzir, Timrat, Nahalal faults; Figs. 10c, 11, Fig. S2). The NE-trending Nazareth fault downthrows B3 southwards, while B3 fill is dipping to the north (Fig. S4). East of B3, the Tavor horst (T in Fig. 6) is uplifted along the eastern segment of Bet Qeshet fault (Figs. 3, 10c, 12, GE). The fault divides the horst from the Sirin-Qama block (B9- Fig. 3, location of fault: Fig. 10c, GE). Neogene exposures extend up to the northeastern corner of the southern Galilee basins (Fig. 6, 10).
However, in this area, the delimitation of southern Galilee basins is less clear, due to later displacements.

Secondary faults
A series of NNE to NE-trending normal faults divide between the basins and structural highs of the southern Galilee. The faults are nearly perpendicular to the axis of the basins complex. Seismic data show that displacements across these faults are 5 mainly vertical with a horizontal component. Regional numerical modelling of Lyakhovsky et al (2012) followed by a review of rift-transform interaction adjacent to continental margins , has predicted rift-perpendicular features.
Locally, these faults, structural highs, and basins between them are evident from the structural map of top Avedat Gr. that consist the floor of most of the basins (Fig. 12). The following paragraphs describe the division along the major axis, from NW to SE. 10 The structural and topographic transition between H1 and B2 occurs along a lineament associated with Sede Yaakov and Aloney Abba faults. These faults are derived from the geological map (Sneh et al., 1998;Segev et al., 2006)  down-throwing B4. Normal displacement along this fault is ~100 m in its northern and southern margins. It reaches ~500 m in the middle (main axis of the basins). Correlation between seismic data and Gideon 1, 2, and 5 wells (Figs. 5,9,GE,Fig. S5) show uneven thickness between the fault flanks, suggesting that it was active several times during the mid and late Miocene, at least until the end of deposition of Bira Fm. (Figs. 5, 7, 9, Fig. S5).
Three structural elements separate B3 from B4. Afula fault vertically throws Lower Basalt Fm. reflectors northward by app. 25 200 m (Figs. 7, 10). East of the fault the volcanic Givat Hamore and Ein Dor blocks separate B3 from B4. (Fig. 10, location: Gilboa fault defines the boundary between B4 and B5 to the south, off the axis of the southern Galilee basins. Data indicate 30 that the B5 fill thickens northwards towards Gilboa fault (Fig. 12, Fig. S16). B5 is bounded by H2 in the west and H3 in the east. Avital fault crosses the NW corner of B5 (Fig. 10c, Fig. S10). Displacements along this sub-vertical fault are mainly horizontal. They are associated with branching into secondary faults and local folding (Figs. S10, S16).  (Shaliv 2003(Shaliv , 2005. Seismic data show that the northern limit of B6 is downthrown along Hashita and En Harod faults relative to Hashita-Geva/Zevayim block (Figs. 7, 10, 12, 13). Sub-vertical normal faults downthrown B7 relative to the eastern flank of H4 and Hashita-Geva block (Figs. 9, 13; Figs. S11, S13). Bet Shean fault is the easternmost of this series. It downthrows 5 the Lower Basalt Fm. 200 m on its eastern side (Fig. S14). However, the basin fill thickens and tilts to the east, where its original structural boundary is unclear. Similarly, the structural transition from B7 northwards into B8 is vague.

First stage (20-9 Ma)
The first stage of subsidence initiated during the early-mid Miocene. The subsidence occurred mainly near the eastern part of the southern Galilee basins, across B6-11 (Fig. 3). Subsidence and faulting developed while the conglomerate member of 5 Hordos Fm. accumulated in topographic lows (Schulman and Rosenthal, 1968;Garfunkel, 1989). A composite section crossing the basins along a WNW trajectory shows that Hordos Fm. accumulation in B6-8 was accompanied with normal faulting and folding (Fig. 9). However, remains of Hordos Fm. are not restricted the subsiding basins. They appear in sporadic outcrops, such as Marma Fayad and Ein Gev (thickness exceeds 200 m; location: Fig. 6, GE); in various elevations on the northern flank of Faria anticline; across the tilted blocks of the eastern Galilee; and across southern B9. Above mentioned evidence suggest 10 that the current shape of the southern Galilee basins was formed by younger deformations, while preceding Miocene basins extended further south of their present-day structure. It also indicates that these remains were displaced by younger faults (Figs. 10, 12;Shaliv et al., 1991) that were active during the initiation of motion along the DSF (Freund, 1978;Garfunkel, 1981;Garfunkel, 1989).
The Spatial and temporal provenance of the lower to mid-Miocene conglomerate of Hordos Fm. are still debated. Conglomerate 15 accumulation of the Hordos Fm. suggests that basin subsidence predates the Lower Basalt Fm., although in several localities it inter-fingers with it ( Fig. 9, Figs. S8, S14). Temporal emplacement therefore is tricky. Outcrop and seismic data from B2 show that normal faults displace a conglomerate unit, before the Lower Basalt Fm. accumulated (Fig. 11). Sandler et al. (2004) associates the conglomerate unit to Bet Nir Fm., suggesting it is concurrent with the Lower Basalt Fm. (17-9 Ma). Our integrative morpho-structural analysis bridges over the spatial gap between the isolated patches of the conglomerates (e.g. 20 Kafri (2002) provenance study), suggesting that Bet Nir and Hordos Fms. accumulated at the same time frame. Together they are products of the same paleo-drainage system that extended from the east to the west across the low relief of the Galilee, immediately before the subsidence of the basins.
The southern Galilee basins accumulated an up to 650 m thick section of volcano-clasts and flows of Lower Basalt Fm. during their subsidence (Fig. 10). In general, the thickness of a basaltic unit is expected to increase close to its source. This assumption 25 guided the identification of volcanic sources across the study area. The seismic and borehole database provided evidence for thickness variations and information about the lithology. Previous studies provided basalt dating from outcrops and wells, along with mapping of tilted blocks and faults (Fig. 10;GE;Segev et al., 2006;Dicker, 1964;Schulman, 1962;Shaliv, 1991).
During the mid-Miocene, normal displacements along faults facilitated deepening of the basins (Figs. 7, 10c, 12). Structural signature of the left-lateral displacement along the DSF enhanced between 12-14 Ma. Bosworth et al. (2005) suggest that the movement started at ~14 Ma in association with the transition of Red Sea opening. In response, the slip along DSF shifted 5 from a N60°E opening motion, perpendicular to the Red Sea axis, to a N15°E motion, diagonal to that axis but parallel to the axis of the DSF. Others estimate the initiation of DSF displacement in the study area to 13 Ma (Shaliv, 1991). Northward channelling of the Afar plume (Ritesma et al., 1999;Chang et al., 2011;Hansen and Nyblade, 2013) along with geodetic and structural research (Bellahsen et al., 2003;Bosworth et al., 2005;ArRajehi et al., 2010) suggest a transition in stress regime.
Three-dimensional analogue models of the Red Sea-Gulf of Aden rift system point at an increase of 70% in the rotational 10 relative motion between Africa and Arabia since 13 Ma (Molnar et al., 2017). This pronounced shift at 13 Ma has left footprints in the Galilee branch.
The association between volcanism and tectonics specifically around 13 Ma appears in several studies across the Arabian plate (e.g., Bayer et al., 1989;Camp and Roobol, 1992;Ebinger and Casey, 2001). Until 13 Ma volcanic activity closely follows the faulting event. A marked shift in volcanism is noted at ~13 Ma. In the western Arabian plate, volcanic fields renewed their 15 activity after a cessation of 9 Ma (Bohannon et al. 1989;Camp and Robool 1992;Ilani et al. 2001;Krienitz et al. 2009). In contrast, magmatic activity in the Galilee was relatively continuous. K-Ar dating bound the volcanic activity across B2 between 16-9 Ma (i.e., the Lower Basalt Fm.; Shaliv, 1991). Further to the east across B6-B11, H3 and Mt. Gilboa, older K-Ar ages of 17-15 Ma were retrieved (Shaliv, 1991;3,5,14, 19 in Fig. 10a). Updated 40Ar/39Ar dates yield a lower limit of 17 Ma for the Lower Basalt Fm. (Rozenbaum et al., 2016;Sandler et al., 2015). Since 13 Ma, volcanism was active across Harrat-A-Sham-20 western Arabia and the Galilee. It was active during the subsidence of the southern Galilee basins and accumulation of conglomerates.
Integration of all above evidence indicates that during the first stage an E-W trending paleo-drainage system developed across the southern Galilee, accumulating conglomerates. Shortly after, this drainage pattern ceased during the relief accentuation due to subsidence of a series of <10 km wide grabens and half-grabens. The basins collected conglomerates, separately, along 25 with the Lower Basalt Fm.. The basins subsided along an NW-trending axis (Fig. 12). Within this general trend, some individual basins trend to the WNW and W. These basins continued to sink, extend and even merge during the transition to the second stage of subsidence.

Second stage (9-5 Ma)
Tectonic displacements that acted during the first stage of subsidence continued during the second, along with erosion. A series of blocks and depressions depicted from the structural map of the Lower Basalt Fm. points at the continuance of vertical motions. Basins continued to subside, forming local topographic lows that accumulated the erosion products. Conglomerates 15 of the Um Sabune Fm. settled close to the edges of the basins (Fig. 8, Figs. S2, S7). Their composition includes pebbles of Lower Basalt Fm. as well as older carbonates (Sandler et al., 2004). Grain size of the conglomerates decreases upwards (Schulman, 1962), indicating a moderation of tectonic activity along the rims of the basins with time. Um Sabune Fm. outcrops tilt southwards along the northern rim of B2 (Kafri, 2002) border ; compose the upper part of B8-9 inter fingering with the Bira Fm. (see below). The Um Sabune Fm. appears to thicken within the incised channels that drain B6. The thickening could result from two factors. Syn-tectonic magmatism allowed the Lower Basalt Fm. to accumulate within subsiding basins on the one hand whereas other parts of the formation were uplifted across their rims. The basins deepened while their margins were gradually elevated (Dicker, 1964). Therefore, elevated terranes and basinal margins were the provenance of the Um Sabune Fm.. Ongoing subsidence of B7-8 5 during the mid-late Miocene facilitated the accumulation of thick section of Um Sabune Fm. near the margins of the basins (e.g., Bet-Yosef, Neve Ur and Zemach wells, Fig. 8; Fig. S7; locations: GE), the Clay Series Fm. was deposited within their depocenters Figs. 8,9,11;Figs. S1, S2, S4-6). The Clay Series Fm. has been preserved since most of the tectonic activity focused on the edges of the basins. According to tens of water wells this formation is verified as a local aquiclude (Wishkin, 1973). 10 Deposition of the Bira Fm. occurred during the volcanism that produced the Intermediate Basalt Fm.. This volcanic formation mainly follows faults (Shaliv, 1991) and due to its minor occurrences (thin sections of few to tens of meters) seismic resolution does not permit its interpretation. It occurs cross H3 , along Rewaya and Gefet faults (Fig. 9, Figs. S5, S8, S14) in B7-10 and the central Jordan Valley (Schulman, 1962;Rozenbaum et al., 2016). With time, accumulation of Bira Fm. moderated the rugged relief of the Galilee until it became almost flat at the end of the Miocene (Fig. S5). The outcrops of Bira 15 Fm. appear today close to faults that were active during the second stage of subsidence, and in places cover these faults. This evidence suggests that Bira Fm. recorded the cessation of subsidence of the southern Galilee basins. The cessation might be associated with a short-term tectonic quiescence across Sinai plate and its nearby Levant margin, allowing marine transgressions to cover the low relief of the southern Galilee.
Our data show that chalks and limestones were deposited in shallow basins at the Yisachar and Poriyya area, while conglomerates accumulated along the rims of the basins (Fig. S15). This flooding is contemporaneous with the onlap of the 5 Pattish Fm. reefal limestones along the Israeli coastal plain.
At the end of the second stage, shallow brackish water lakes occupied the topographic lows above the basins. Limestones and chalks of the Gesher Fm. accumulated in lakes Rozenbaum et al., 2016). The thickness of Gesher Fm., merely reach tens of meters, slightly above the seismic resolution limit. Bira and Gesher Fms. sealed the southern Galilee basins and formed a relatively flat relief. Similar to the RTS at the base of the basins, the relatively flat top of the Bira and 10 Gesher Fm. serve as a marker for tectonic activity that deformed the study area during the Plio-Pleistocene.
Data presented in this study suggest that the uplift of Mount Carmel, Tivon and Shefar'am occurred close to the end of the second stage, between 5-6 Ma (Figs. 3, 12). The uplifts placed topographic barriers between the Mediterranean Sea and the inland lakes, diverting possible marine transgressions to regions south of the Galilee. These observations stand in line with Shaliv (1991). Gvirtzman et al. (2011) suggest that the Carmel area was submerged under marine conditions before the upper 15 Miocene. They base this deduction on a single outcrop located in Bet-Rosh that contains a continuous marine succession from the Eocene to mid-Miocene. These authors accept the possibility that the Galilee was exposed and claim that it resembled the Carmel in the timing of initiation of vertical displacements during the upper Miocene. The integrative geological-geophysical data presented here show differently. Our results attest to hundreds of meters thick lacustrine-fluvial infill that accumulated during the early and mid-Miocene displacements, while tectonics were active . 20 In summary, the pattern of subsidence of separated and localized basins continues from stage one to stage two. However, during the second stage, the basins also elongated along an NNE trend, while keeping the elevated structural highs in between ( Fig. 12). Numerical modelling of deeper sections of the lithosphere predicted such relief pattern, of rift axis perpendicular faulting (Lyahovsky et al., 2012;Segev et al., 2014). The subsidence extended beyond the area studied here into the regions that were uplifted and eroded during the Plio-Pleistocene, for example, over H3 and the tilted blocks of the eastern Galilee 25 (Figs. 3, 6, 7).

Tectonic classification of the basins during the two stages
The Galilee basins developed during the Neogene through two major structural processes. Extensional regime during the first stage formed the Galilee basins. The thinning of the Lower Basalt Fm. to the northwest (Figs. 6,10) points to shallowing of the basin floor in that direction, and hence to a reduction in regional extension towards the continental margin in the west. At 30 this stage, the structure of the basins and their dimensions are equivalent to the definition of intraplate basins that form during rifting (Evison, 1959;Bosworth, 1994;Busby and Ingersoll, 1995;Allen and Allen, 2005;Morley et al., 2004) as well as to the grabens and half-grabens of intra-continental rifts (Bosworth, 1994). Previous studies showed the development of the Qishon-Sirhan rift during the Oligocene-Miocene in a northwesterly direction (Shaliv, 1991;Schattner et al., 2006a;. Results of the present study claim this rift comprises the southern Galilee Basins during their first stage of development. The second stage of subsidence (9-5 Ma) marks a transition of the extensional stress regime into transtension along a primary NNE direction and a secondary WNW direction. Basins subside vertically and extend perpendicularly to the principal axis of 5 the first stage basins, while structural highs separate between them (Fig. 12). The highs are accommodation zones, structurally equivalent to the separators between basins along the East African Rift (Bosworth, 1985;Bosworth et al., 1986;Rosendahl, 1987;Ebinger et al., 1987;Burgess et al., 1988;Ebinger et al., 1989;Morley et al., 1990). These studies also show that basins along a forming rift accumulate sediments while tectonic subsidence is in action. As a system, some of these rifts may succeed and continue to drift, while others fail. The two stages recorded here occurred alongside the initiation of motion along the 10 nearby DSF plate boundary. While motion took place along the entire boundary between the Arabian and Sinai plates, intracontinental basins subsided only across the southern Galilee.

Structural transitions along the plate boundary
The on-going Afro-Arabian and Eurasian convergence (Letouzey, J. and Tremolieres, 1980) induced three major stress regimes across the Galilee. (1) The Syrian Arc compressional stress regime (Krenkel, 1924) produced a WNW shortening during the 15 Turonian (Eyal, 1996;Eyal et al., 2001). Compression-related folds plunge north towards the Carmel-Gilboa trajectory, are buried at the subsurface of the southern Galilee basins, and are exposed again across the northern Galilee (Fig. 3).
(2) The Red Sea extensional regime prevailed during the Oligocene to the early Miocene (Steckler and tenBrinck, 1986;Khalil and McClay, 2002;Younes and McClay, 2002;Bosworth et al., 2005;Khalil and McClay, 2016). It resulted in the coeval opening of the parallel Red Sea and Sirhan rifts Schattner et al., 2006a). The ENE axis extension (McClay and Khalil, 20 1998;Younes and McClay, 2002;Bosworth et al., 2005) later shifted during the Neogene (Garfunkel and Bartov, 1977) to the NNE (N15°E, Bosworth et al., 2005). The NW trending faults developed across the study area are part of a larger set of the western Arabian plate. Fault systems of the Suez-Red Sea (Steckler and ten Brink, 1986), Sirhan (Schattner et al., 2006a) and Karak (Bender, 1974) reactivated traces of the Precambrian Najd fault system (Stern, 1985;Agar, 1987;Stern, 1994). Our data show that the Red Sea regime provided sufficient conditions for the first stage of subsidence of the southern Galilee basins, at 25 the northwestern tip of the Sirhan rift. The failure of this rift during the early-to-mid Miocene is closely associated with the emergence of the third, Dead Sea, stress regime (Schattner et al., 2006b;Segev et al., 2014).
Convergence between the Arabian and Eurasian plates transformed into collision and slowed down during the mid-Miocene (14-12 Ma). This short recess resulted in tectonic quiescence in Suez (Bayer et al., 1989), the southern equivalent of the Galilee basins. In between the two rift systems, the Negev ceased to subside (Zilberman and Calvo, 2013;location-Fig. 1); while the 30 Judea region was elevated by 400 m above the Miocene coastline (Sneh and Buchbinder, 1984;Bar, 2013;location-Fig. 1).
During the same time window, a numerical simulation shows a depression that subsided along the Sirhan trajectory, still not entirely affected by the displacement along the intersecting DSF. This depression extended from Irbid structural low in the east (NW Jordan) to Beteha-Sea of Galilee-Kinnarot basins in the west (Location- Fig. 1; Fig. 10 in Segev et al., 2014). During that time, volcanic activity stopped in Syria (Mouty et al., 1992).
The tectonic transition between the Red Sea and Dead Sea stress regimes was accompanied by up to 50% decrease in the relative velocity of the African plate around 11 Ma (Reilinger and McClusky, 2011), and a geometric rearrangement of the plates around 9 Ma (McQuarrie et al., 2003;Faccenna et al., 2013). This transition corresponds with the first to second 5 subsidence stage shift of the southern Galilee basins (Fig. 12). The DSF cuts through all previous structures along its ~1000 km trajectory. These include the Sirhan rift. As a result, the southern Galilee basins, isolated from their original system, continued to extend along an orientation tangential to the new stresses. This extension appears as the second stage of subsidence of the southern Galilee basins (Fig. 12).
Previous studies widely agree on an N-S extension of the Galilee during the upper Miocene. Schulman (1962) and Horowitz 10 (1979) suggest that the Galilee basins continued to extend during the upper Miocene. Freund (1970) calculates the finite N-S extension based on exposed faults in the Galilee. His results indicate an increase from 0% along the Mediterranean coast, through 5% across the central Galilee, and 7% in B7 near the DSF. This distribution pattern of displacement also corresponds to the exposure of Lower Basalt Fm. that decreases westwards. Freund (1970) related the differential N-S extension to the displacement along the nearby DSF. Ron and Eyal (1985) suggest that during the Miocene to early Pliocene an N-S extension 15 with E-W compression prevailed across the Galilee. These stresses resulted in lateral shear along conjugate faults, accompanied by block rotation. The NNE trending extensional basins defined in our results are in line with these deductions. The separation between first (17-9 Ma) and second (9-5 Ma) stages suggested here for the first time explains the structural relations between the declining Irbid rift and the emergence of DSF dominance. The NE extension of the Galilee during the declining rifting decreases in the second stage and shifts to NNE. However, NNE extension, including an E-W compression component, prevails 20 into the Pliocene (Figs. 12, 13).

Failed rifts and magmatism
The low extension rate (<7%) in the Galilee corresponds to similar values in other failed rifts, such as Lake Tanganyika (Morley et al., 1990;Rosendahl, 1987). The extension is also associated with dike emplacement. Dikes may focus the strain to detachment faults (Rosenbaum et al., 2008). In Afar and Ethiopia (eastern Africa) normal faults developed during the initial stages of rifting and were abandoned 10 Ma later. Extensional stresses there have focused on a narrow region that contains 15 faults and magmatic intrusions (Ebinger and Casey, 2001). In the Gulf of Aden, the magmatic activity was smaller. d'Acremont  (2005) show an abandonment of older detachment faults within the rift environment replaced by the formation of a newer, shorter segmentation along the central axis of the rift. Rift associated magmatism therefore commences in regions distant from the rift axis, and is dependent on fault distribution. In systems where extension is localized to narrow zones, dikes may follow extension lineaments. Examples of such cases are the Gulf of California (Lizarralde et al., 2007) and along the magmatic boundary of the north Atlantic (White et al., 2008). In both areas, the basaltic intrusions appear within the narrow 50-100 km 5 outline of the rift. Hence evidence for magmatic intrusions and their spatial arrangement may hint at rifting orientation and associated extensional stresses.
During the first stage of subsidence in the Galilee volcanism arrived mainly through extensional lineaments associated with normal faulting, along with the subsidence of the basins (Fig. 12). Syn-tectonic volcanism supplied the thick sections of Lower Basalt and Hordos Fm. in B4,B6,and B7 (Figs. 6,9,10b). A volcano in the southern margin of B6 and possible sources along 10 H2 supplied additional volcanics that accumulated in B2, B4, and B5. The magmatic intrusions in H3 (Givat Hamore: location: Figs. 3, 10, 7) were dated to 15 Ma and associated with an NW to WNW faulting system (Fig. 7;Dicker, 1964;Shaliv, 1991).
Volcanism continued during the second stage of subsidence, along with the vertical and horizontal displacement of the study area. The Intermediate Basalt Fm. dated to ~6 Ma arrives through normal faults bounding H3 from the NE, and perhaps through a volcano located in the Rewaya block (Shaliv, 1991;Fig. S8). The directional correlation between faulting and volcanic 15 centers and lineaments (Figs. 7,10,12) obeys to a similar regional tendency. Equivalent correlation appears in Karak graben (Bender, 1974), Miocene dikes across Sinai (Bartov et al., 1980;Baldridge et al., 1991), and across Harrat-A-Shaam volcanic field (Feraud et al., 1985;Mor, 1986;Giannérini et al., 1988;Brew et al., 2001;Al Kwatli et al., 2012). The strips of alkaline volcanism across the Arabian plate represent the beginning of Miocene volcanism (Camp and Roobol, 1992;Weinstein, 2000;Ilani et al., 2001). We, therefore, suggest that the faulting and volcanism of the southern Galilee also follow weak lineaments 20 in the lithosphere.
The timing of regional volcanism is noteworthy. Between 18-12 Ma volcanic activity ceased across the Arabian plate and was dominant across the southern Galilee basins (Lower Basalt Fm.). This shift may represent an NW propagation of extension and volcanism across the Arabian plate (Weinstein, 2000). The northwestern Arabia volcanism was renewed at 14-12 Ma (Bohannon et al., 1989;Camp and Robool 1992;Ilani et al., 2001;Krienitz et al., 2009). Several studies link the renewal and 25 activity with structural aspects (Bayer et al., 1988;Camp and Roobol, 1992;Ebinger and Casey, 2001). However, other studies suggest that the lateral slip along the DSF decreased during the upper Miocene (Hempton, 1987;Bayer et al., 1989;Reilinger and McClusky, 2011;Faccenna et al., 2013), while drift across the NW trending Irbid rift was active Segev et al., 2017). Our results suggest that this decrease also enabled the subsidence of the southern Galilee basins during the second stage, as part of the hybrid Red Sea -Dead Sea stress regime. With enhancement of motion along the DSF during the lower 30 Pliocene around 5 Ma, the Dead Sea stress regime became dominant, laterally shifting the southern Galilee basins, and structurally isolating them from their first association to the Irbid rift.

Conclusions
The Galilee basins subsided along the northwestern front of the Sirhan rift. Integration of geological and geophysical data bounds the subsidence of the basins between two major surfaces: the Oligocene Regional Truncation Surface (RTS) and the top of Bira Fm. unconformity. The subsidence is divided into two stages. 5 During the first stage  the Galilee basins subside along the main trend of the Oligo-Miocene Sirhan rift system. They subsided as grabens and half-grabens, bounded by normal faults and structural saddles. Larger subsidence was recorded along the main NW trending rift axis. Smaller basins subsided off the main axis. The subsidence occurred along with extensive volcanism that arrived through fault planes that bound the basins. The spatial arrangement of the rift basins suggests that they follow a larger Principal Displacement Zone (PDZ). The major boundary faults mapped here are the surface expression of the 10 PDZ strands that bound the basin complex of the rift from north and south. The complex originally formed as a releasing jog along a rift system. The structural change around 9 Ma is associated here with the gradual transition between the Red Sea and the Dead Sea stress regimes. With the initiation of shearing along the DSF, the jog and its basins were truncated. The transition elongated the basins, accentuated their subsidence, and uplifted their surrounding margins.
During the second stage (9-5 Ma) left lateral shearing of the entire study area results in subsidence of a series of NNE trending 15 basins, perpendicular to the major axis of basins from the first stage. Structural highs that divide between the first-stage basins remain high during the second stage. However, during the second stage, their bounding normal faults also exhibit a lateral component. The general shear distorts the original structure of the first stage basins north and south of the major NW-trending axis. The length of the basins decreases from ~60 km in the east to ~15 km in the west of the study area. The volcanism of the second stage arrives from weak zones and focusses on structural boundaries between the basins, and volcanic activity along 20 their margins.
Structural architecture of the southern Galilee indicates that the rift basins continued to subside while the Irbid rift was active.
Their shape and arrangement were constrained by two main rheological features -the bounds of a releasing jog along the PDZ and the acquaintance with a more cohesive crust at the peripheral area, perhaps a "locked zone" (see Lyakhovsky et al., 2012;Segev et al., 2014). However, neither of these seems to have caused the cessation of rifting. In fact, the basins at the rift tip 25 subsided until the jog was decapitated by the motion along the DSF. The main cause of the structural transition (and preservation) of the southern Galilee basins was the transition from one dominant stress regime to another. Our study provides a unique and detailed architecture of a rift basin complex. Based on this case study we suggest that the rift did not fail but rather faded and was taken over by a more dominant stress regime. Otherwise, basins of this failing rift could have simply died out.

Author contribution
This work is based on a profound chapter from Reli Wald's PhD thesis. Reli Wald has processed and analysed the datasets, including seismic interpretation and development of a 3D geological model. Amit Segev, Zvi Ben-Avraham and Uri Schattner have critically read and reviewed all the data following their participance in work as thesis advisors. Uri Schattner has contributed in writing and in figure graphics. Reli Wald has prepared the manuscript with major contribution from Uri 5 Schattner and with review of co-authors.