Constraining sediment transport to deep marine basins through submarine channels: The Levant margin in the Late Cenozoic

Marine Geology 347 (2014) 12–26

Contents lists available at ScienceDirect

Marine Geology journal homepage: www.elsevier.com/locate/margeo

Constraining sediment transport to deep marine basins through submarine channels: The Levant margin in the Late Cenozoic Z. Gvirtzman a,⁎, I. Csato b,1, D. Granjeon c a b c

Geological Survey, Israel Collin College, USA IFP Energies nouvelles, France

a r t i c l e

i n f o

Article history: Received 26 September 2012 Received in revised form 21 October 2013 Accepted 27 October 2013 Available online 2 November 2013 Communicated by D.J.W. Piper Keywords: Levant continental margin Levant Basin submarine channels submarine canyons stratigraphy sediment supply

a b s t r a c t The recent world-class gas discoveries in Early Miocene sand units offshore Israel raises the question of their origin. Apparently, the simplest explanation is to relate them to a fluvial system that arrived from Arabia at that time. This system predated the modern (Pliocene) Nile River supply and existed until captured by the Dead Sea valley. Interestingly, however, very little sedimentation occurred along the Levant continental margin before the Pliocene in spite of its stepped structure that provided much space for accommodation. The only way that sediments could have bypassed the continental margin and arrive at the deep basin without being trapped in the middle is through submarine channels that crossed the continental margin. Here we explore this possibility using 3-D stratigraphic modeling techniques that quantify the sediment load and the water discharge required to fill the basin by pushing enough sediment through submarine channels. We show that such a scenario requires a fluvial system in the order of the largest rivers that exist today on earth in terms of drainage area and water discharge. Alternatively, it requires extreme hydraulic conditions in terms of diffusion coefficients and an elevated drainage basin that could not have existed in the study area. We therefore challenge the traditional view of Arabia as the main source for Oligo-Miocene deposits in the Levant Basin and suggest that the basin was mainly fed by a proto-Nile system that transported clastic material to the North African margin and then farther east by ocean currents. In a wider view we demonstrate how numerical modeling can constrain sediment transport through submarine channels as a function of basin geometry and hydraulic conditions, and how paleogeographic knowledge can be combined with current data on world rivers to evaluate if modeling results are plausible. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Drilling in the deep Levant Basin offshore Israel finally settled the debate regarding the age of the exceptionally thick sedimentary sequence below the Messinian salt layer. Three wells were drilled since 2009 in the deep Levant Basin: Tamar, Dalit, and Leviathan, located 90, 40, and 135 km offshore Israel, respectively (Fig. 1). All three wells are consistent with the seismic interpretation of an exceptionally thick Late Cenozoic section (Gardosh et al., 2008; Gvirtzman et al., 2008; Gardosh et al., 2010; Steinberg et al., 2011) and challenge other interpretations claiming thick Cretaceous (Peck, 2008) or Paleocene–Eocene (Gardosh and Druckman, 2006; Roberts and Peace, 2007) sequences. Consequently, the Late Cenozoic deep-water sediments of the Levant Basin became a great interest to the industry as well as to the scientific community. To understand the depositional history of these sediments and how sands were transported to the deep basin hundreds of kilometers away from the ancient coastline (the reconstructed Oligocene coastline is shown in Fig. 1), the paleogeography of the circum eastern Mediterranean area at that time must be understood. ⁎ Corresponding author. 1 Present address: MOL Oil and Gas, Plc, Hungary. 0025-3227/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.margeo.2013.10.010

Although drilling proved the existence of siliciclastic deposits down to the Late Oligocene (public releases, Noble Energy Inc.), interpretation of seismic data indicates that the significant change in the nature of deposition occurred in the Late Eocene when the sedimentation rate in the deep Levant Basin accelerated by nearly 20 times (Steinberg et al., 2011, Fig. 2). This fundamental observation raises fundamental questions. Why did the sedimentation rate increases? Where did the sediments come from? Where are the ancient sedimentary pathways into and in the deep sea? What was the mode of sedimentary transport and dispersal? A priori, there are at least two feasible source-to-sink scenarios marked by different arrows in Fig. 1. (1) The large amounts of terrigenous material that began entering the Levant Basin in the Late Eocene originated in Africa and was transported via the Egyptian continental margin that ~25 Ma later (Pliocene) evolved into the Nile River cone; or (2) originated in Arabia (plus the Sinai Peninsula?) and was transported via the Levant continental margin. The existence of a pre-Pliocene east-to-west transport system, which reached Israel from Arabia across the area that eventually developed into the Dead Sea rift valley and continued farther west to the Levant Basin, has been well established. The earliest indication for such transport is turbidite deposits found within Oligocene outcrops

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

13

Fig. 1. Location map with present topography and main Middle East rivers. Reconstructed Oligocene shoreline shown by a thick broken line indicates that the extensive deposition in the Levant Basin occurred while the north Arabian Platform was still under water, excluding the possibility of sediment supply from the northeast. Black arrows schematically show expected drainage directions. Estimated area that had drained to the Levant Basin is marked by light gray. Recent gas wells in Oligo-Miocene sand units are marked by red dots. Red rectangle marks the location of the model of Fig. 5. Red line outlines the location of the geological section of Fig. 2. Paleogeography after Steinberg et al. (2011).

of the Lower Saqiye Group in the Judea foothills (Buchbinder et al., 2005) and within Late Eocene deposits of the lowermost part of the Saqiye Group in the subsurface of the coastal plain (Buchbinder et al., 2005). Noteworthy, these turbidites were transported farther to the Levant Basin through deep submarine channels (Druckman et al., 1995; Gardosh and Druckman, 2006; Gardosh et al., 2008; Bar, 2009) that were incised in the Israeli continental margin in the Late Eocene

(El-Arish, Afiq, Ashdod, Hanna, and another unnamed canyons marked in Fig. 3). Ten to fifteen million years later, the nature of the transport from east changed. In the Early Miocene, large amounts of coarse siliciclastic sediments (Hazeva Formation) transported from distances of hundreds of kilometers were trapped in several inland basins (Garfunkel and Horowitz, 1966; Zilberman, 1991; Calvo and Bartov, 2001). The finer

Fig. 2. Geological cross section from the inland Levant region to the Eratosthenes Seamount. Note that the Late Cenozoic section (red colors) representing 35 Ma of deposition is thicker than the deeper part (gray) which was deposited during 250 Ma (Triassic). Also note the Late Eocene–Miocene section (lowest red unit) was deposited mostly in the deep Levant Basin block and very little over the intermediate margin block that was buried only by the topmost unit (Pliocene to recent) after the Messinian Salinity Crisis (spatial details in Figs. 3–4). From Steinberg et al. (2011); location in Figs. 1 and 3.

14

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

Fig. 3. Map of the study area that combines the present topography in the east with a subsurface structural map of the base Saqiye Group in the west. This map approximately describes the Early Oligocene relief, highlighting morpho-structural steps and incised canyons (discussed in text). Broken rectangular line indicates location of the model shown in Fig. 5. Note that the inland side of the rectangular approximately follows the Oligocene shoreline today located at the Judea western mountain front (Gvirtzman et al., 2011). The structural map is taken from Bar et al. (2013) based on Steinberg et al. (2011) with minor modifications in its eastern part from Fleischer and Gafsou (2003). Outlines of the El-Arish Canyon are modified after Gvirtzman et al. (2008). Black lines are faults of the Continental Margin Fault Zone, from Gvirtzman et al. (2008).

terrigenous materials were carried farther to the sea, from where the fine clasts were further transported to the deep basin through the same submarine channels that were partly filled in the Oligocene and re-incised in the Early Miocene (Druckman et al., 1995). Interestingly, however, the amount of sediments trapped along the Levant continental margin during the 30 Ma period from the Late Eocene to the end of the Miocene is only a few hundred meters, whereas in the deep basin a nearly 4 km-thick section accumulated at the same time (excluding the 1.5 km thick Messinian salt layer; Fig. 2 after Steinberg et al., 2011). This fundamental observation emphasizes the difficulty of an easterly supply scenario that requires bypassing of the continental margin and jumping over two morphological steps at the sea floor without being trapped in between (see steps in Fig. 3 and more details below). The only possible transport mechanism for such a scenario is through the submarine channels that crossed the two steps and connected the ancient continental shelf (today's coastal plain and foothills, Fig. 3) with the deep basin.

The purpose of this study is to examine this possibility. Utilizing a 3-D stratigraphic model, we constrain the sediment load and the water discharge required to transport enough sediment into the deep basin through the submarine channels. We show that the water discharge required for such a scenario is unreasonably high, thus indicating the need to consider a major source from the southwest (i.e., from Africa). While the practical implications of our modeling are relevant to revealing sediment pathways and tracing potential reservoirs, its wider implications relate to paleo-climate and paleo-geography during an important period in which the ancient Tethys Ocean was shrinking and the marine gateways connecting the Persian Gulf and the Mediterranean Sea closed. Moreover, the question of how much sediment can be transported through submarine channels, how far it reaches, and how widely it is distributed is a fundamental problem in understanding lowstand system tracts and their dependency on tectonics and climate. If the shelf-to-basin relief is relatively low as in many mature passive

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

continental margins, progradational development oceanwards is likely. But when the shelf-to-basin relief is steep and particularly when it consists of high morphological steps, as in the reactivated Levant margin (Gvirtzman et al., 2008), progradation of the shelf is difficult and more aggradational and buildup patterns are likely (Hadler-Jacobsen et al., 2007). Another important factor is the width of the continental shelf; wide shelves capture considerable amounts of coarse grain sediments while narrow shelves are more easily bypassed. Therefore, the distal, deep marine deposits along narrow shelves will typically contain coarse-grained fans while deep marine deposits along wide shelf will contain mostly fine-grained deposits (Bouma, 2000). In this context the pre-Messinian Levant continental margin is challenging. It consisted of a relatively wide block that had separated the narrow ancient continental shelf (today's foothill area) from the deep Levant Basin (Fig. 3). This intermediate block could have captured course grained sediments, but two steep morphological steps that bounded it from both sides (details below) may have prevented lowstand and highstand progradational patterns. Here we explore whether the sand-prone, gas-rich, deep marine fans were fed by fluvial material that had bypassed this challenging region or, alternatively, were fed from elsewhere. In addition to sea-level stand and margin physiography, we note that the Oligo-Miocene sequence we study here was deposited during and mostly after the Eocene–Oligocene transition that changed the earth's climate from a greenhouse regime to an icehouse regime (Zachos et al., 2001; Miller et al., 2005; Fielding et al., 2008). Accordingly, the volume of sediment supply into the basin, the timing and the size of slope failure events that produce debris flows and turbidity currents (Morehead et al., 2001) may have all been influenced by climate. To constrain the amount of sediments that crossed the Levant continental margin at that time, we utilize numerical modeling tools that can predict stacking patterns as a function of basin geometry and hydraulic conditions. The results are then compared to present day fluvial systems in various climate conditions and relief. 2. Geological background 2.1. Regional setting The closure of the Neo-Tethys since the Late Eocene was accompanied by a significant change in the paleogeography of the Middle East region. Along with the collision in the Bitlis-Zagros thrust zone, an extensive area previously submerged under water for tens of millions of years, rose above sea level (Adams et al., 1983; Buchbinder, 1996; Ziegler et al., 2001). The exposure of this land mass, hundreds of kilometers to the south of the collisional plate boundary, created the continental region today occupied by Iraq, Syria, Jordan, Israel, and Lebanon and disconnected the Mediterranean basin from the Mesopotamian basin. As a result, the contours of Arabia changed and shorelines that had previously extended from Egypt eastwards towards the Persian Gulf changed their course northwards towards Turkey along the present day Mediterranean coasts (Gvirtzman et al., 2011). These processes were accompanied by the formation of an Oligocene erosional surface whose remnants have been described in Sinai, southern Israel, and Jordan (Picard, 1943; Garfunkel and Horowitz, 1966; Zilberman, 1991; Bar, 2009; Avni et al., 2012) and by the formation of a drainage system towards the eastern Mediterranean. Approximately at the same time, farther to the south, East Africa was also uplifting along with the vast volcanism that accompanied the Red Sea opening (Bosworth et al., 2005, and references therein). This inland uplift provided another major source for sediments to the eastern Mediterranean. However, differing from the Israeli margin, the much wider fluvial system in northern Egypt prograded the coastline of North Africa hundreds of kilometers northwards (Fig. 1, Salem, 1976; Said, 1981; Burke, 1996), indicating that the African source of sediments was much larger than the Arabian source.

15

The depositional expression of the vast amounts of fine clasts that began entering the basin in the Late Eocene and Early Oligocene was a shift from pelagic chalks (Avedat Group) that had characterized the ~50 Ma of the Santonian–Mid Eocene, to hemipelagic marls of the Saqiye Group (Gvirtzman and Buchbinder, 1978), which have prevailed since. In addition, sedimentation rates in the deep basin increased from ~5 m/Ma in the Paleocene to ~100 m/Ma in the Late Eocene–Oligocene (Steinberg et al, 2011). 2.2. Basin structure and step formation The Neo-Tethys closure and Red Sea opening were accompanied by reactivation of the Levant continental margin (Gvirtzman et al., 2008; Gvirtzman and Steinberg, 2012) after being passive for more than 100 Ma. Three morpho-structural steps, formed by transpression, reshaped the moderate, gradually westward descending seafloor. These steps are evident in the map of Fig. 3, which combines the present topography in the east with a subsurface structural map in the west. We used the nearly base Oligocene structural map as a proxy for the Early Oligocene relief, which was controlled by the newly formed steps. The eastern step, which accentuated an old Syrian Arc structure (Bar, 2009) forms the present-day Judea Western Mountain Front Step (WMFS). The central step, now buried under Neogene sediments below the present day coastal plain of Israel (CPS, Coastal Plain Step), accentuated the Cretaceous continental slope (Bar, 2009). The western step, bounding the Levant Basin on the east approximately below the present day continental slope, vertically accentuated a Santonian– Eocene, 10 km-wide fault zone. This step was termed by Gvirtzman et al. (2008) as the Levant Continental Margin Fault Zone (CMFZ) and is interpreted as the incipient northwest plate boundary of Arabia that later jumped inland to the Dead Sea Transform (Gvirtzman and Steinberg, 2012). The formation of steep morphological steps at the bottom of the sea rapidly caused a prominent incision by submarine canyons. The Ashdod, Afiq, El-Arish and Hanna canyons (Fig. 3), which are each 1200–2000 m deep (Druckman et al., 1995; Buchbinder et al., 2005; Bar, 2009), approximately indicate the heights of the morpho-structural steps that interrupted the previously gradually deepening sea floor. In the following Oligocene–Miocene times, massive sedimentation in the deeper Levant Basin diminished the morphological expression of the CMFZ (Steinberg et al., 2011), while the CPS continued to dominate sea-floor morphology for much longer. During the Messinian Salinity Crisis (e.g., Hsu et al., 1973) the level of the Mediterranean Sea drastically dropped and deep canyons were incised again across the continental margin. However, the Messinian incision was not as deep as the former Late Eocene–Early Oligocene incision and did not reach the canyon bottoms, as evident by the thick Oligo-Miocene sections preserved in the canyons (e.g., Druckman et al., 1995). Finally, during the Pliocene Nile-derived sediments buried the coastal plain step. These sediments extended the continental shelf westwards and built the present-day continental slope. 2.3. Late Tertiary sedimentation and depositional patterns The increased sedimentation rate since the Late Eocene formed several stratigraphic sequences as described by Steinberg et al. (2011). The first sedimentary package that filled the basin during the ~13 Ma of the Late Eocene–Oligocene period accumulated almost solely in the deep basin with an apparent depocenter located in the southwestern part of the study area (Fig. 4a). Differing from the gradual northern and western thinning, this unit abruptly thins to the east from a 1500m-thick section at the base of the CMFZ step to a thin or absent section on the intermediate block. Sediments of this age also filled a canyon that incised the western step (Hanna-1 well; Gardosh et al., 2008; Bar, 2009), indicating canyon incision that either predates or is coeval with this package.

16

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

Fig. 4. Isopach maps from Steinberg et al. (2011) based on seismic interpretation and well data. Pliocene sediments offshore Egypt from Segev et al. (2006). Note that Late Eocene to Miocene units of maps a and b were mostly deposited in the deep Levant Basin and very little along the continental margin.

The second sedimentary package is composed of Miocene sediments (Fig. 4b) below the Messinian evaporites, and represents a period of ~16 Ma and thins from ~2000 m in the deep basin to a few hundred meters along the continental margin. This thinning pattern indicates that the CMFZ step still existed in the Miocene, though gradually disappeared. The gradual thinning of the Mid-Late Miocene unit in Fig. 3 (not shown separately in Fig. 4) indicates that at the end of the Miocene the CMFZ step was completely buried and inactive; instead, short wavelength thickness variations documents Miocene folding. The third sedimentary section includes the rapid deposition of a thick salt layer during the Messinian Salinity Crisis (Fig. 4c), when the Atlantic– Mediterranean gateway was restricted (reviewed by Ryan, 2009). This layer gradually thins landward from ~2 km in the Levant Basin until almost completely pinching out at the proximal–distal boundary (Bertoni and Cartwright, 2005). East of this point, the Messinian evaporites form a thin layer that is poorly imaged on seismic data. The fourth and uppermost unit is composed of Pliocene–recent sediments (Fig. 4d). Unlike older units, which are thicker in the Levant Basin, this unit is substantially thicker along the continental margin. Offshore Egypt, this unit builds the Nile Delta, reaching a thickness of ~4 km (Said, 1981; Segev et al., 2006). Offshore Israel, this unit

completely fills the Jaffa Basin (Gvirtzman et al., 2008) and obscures the CPS under the present-day continental shelf. In this study, we focus on the first two packages deposited between 37 Ma (beginning of Late Eocene) and 7 Ma (just before the Messinian Salinity Crisis). At that time sediments were transported to the Levant Basin both from Africa and from Arabia. Theoretically, one would want to be able to distinguish between the relative contributions of the two sources. However, given that there are no geological constraints for realistic modeling, we compromise with being able to eliminate the possibility that the eastern source was the major source. As explained above, the only way to explain basin filling by an easterly system that bypassed the continental margin is by massive transport limited to the submarine canyons. According to this scenario (Gardosh et al., 2008), sediments were transported through the canyons without being dispersed before reaching their final destination. The question examined here is whether or not such canyons are capable of transporting the required amount of sediments. In what follows we begin with showing that “common” or “normal” conditions are certainly not enough to transport the required amounts of material and that canyons overflow with much sedimentation along the continental margin, in contrast to observations. We proceed with

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

showing that increasing the water discharge through the canyons to extreme values or alternatively using extremely high diffusion coefficients could solve the problem. Though such conditions cause very rapid flow of sediments without their being caught in the middle, they are unrealistic. 3. Method: principles of forward modeling by Dionisos Dionisos is a 3D stratigraphic forward model aiming to simulate geometry and facies of sedimentary units on a regional spatial scale from several tens to hundreds of kilometers and of geological time scales from tens of thousands to hundreds of millions of years (Granjeon, 1996; Granjeon and Joseph, 1999, Granjeon and Wolf, 2007; Granjeon, 2009). This numerical model accounts for accommodation, sediment supply and sediment transport. Since the early work of Culling (1960), and Carson and Kirkby (1972), diffusion laws have been used in geomorphology and geology to represent large-scale, spatially averaged transport of sediment by creep, overland flow and channel flow processes. The concept of diffusive sediment transport states that the transport capacity of a water flow is proportional to the local basin slope and water discharge. The diffusion equation was derived from empirical and conceptual hydraulic equations by Begin et al. (1981) and Paola et al. (1992). It has been widely used in various forms to model sediment transport along alluvial fans, rivers and floodplains (Begin et al., 1981; Murray and Paola, 1994; Parker et al., 1998; Coulthard, 1999), mountains and foreland basins (Jordan and Flemings, 1991; Tucker and Slingerland, 1994). Since the Late 1990s and early 2000s when high resolution bathymetric studies began showing that seabed morphology is curiously similar to lowland landscapes (Adams et al, 1998; Galloway, 1998; Schlager and Adams, 2001; Mitchell and Huthnance, 2008), laws of landscape evolution have been extended to the seascape to simulate progradation of deltas (Kenyon and Turcotte, 1985) and continental shelves (Jordan and Flemings, 1991; Kaufman et al, 1991; Granjeon, 1996; Rivenaes, 1997; Mitchell and Huthnance, 2008). A particularly challenging task is to model turbidity currents by numerically solving the Navier–Stockes equations for conservation of momentum, water, and sediment (e.g. Parker et al., 1986; Bradford, 1996; Skene et al., 1997; Syvitski and Hutton, 2001; Kassem and Imran,

17

2004; Cantero, 2007). But these high-resolution and physically based models are difficult to implement for large areas over geological time scales. Therefore, for long-term modeling of a series of turbidity flows, the diffusive sediment transport approach was used in many studies (Kaufman et al., 1991; Steckler et al., 1999; Schlager and Adams, 2001; van Heijst et al., 2001; Lai and Capart, 2007; Mitchell and Huthnance, 2008; Gerber et al., 2009; Spinewine et al., 2011). Following these studies, Dionisos extends the diffusion equation used in fluvial environments to marine environments. In other words, our landscape model is used as a seascape model to simulate large-scale and long-term evolution of turbiditic complexes. We note, however, that this approach cannot be applied for modeling an individual turbidity current flow, where the assumption of equilibrium between buoyancy driving forces and friction forces is not valid. Another issue that should be treated carefully is the connection between the fluvial and submarine systems. As rivers reach the sea their ability to transport sediments suddenly decreases and coarse grains immediately settle down to form a delta. In some cases the remaining water flow is still denser than the seawater and a hyperpycnal flow continues downstream plunging down the seabed and giving birth to density currents (Bates, 1953; Mulder and Alexander, 2001). However, this process is usually not continuous. Only in rare cases, where steeply sloping canyons enter the river mouth or estuary, riverfed hyperpycnal flows traverse the shelves without interruption (Wright and Friedrichs, 2006). In most cases, sediments temporarily settle down on the delta front or inner shelf and only later they are transported downslope by gravity-driven processes. In other words, the fluvial discharge either continues seaward immediately as hyperpycnal flow, or deposits first at the mouth of rivers as a buoyant sediment plug. In the latter case, the progressive increase of sediment load eventually causes slope failure and turbidity currents. As the detailed modeling of the transformation of fluvial discharge into either dense bottom currents or hyperpycnal flows is far beyond the scope of this paper, we assumed that all fluvial discharge continues seaward as gravitational flows, either fed directly by rivers as hyperpycnal flow, or by slope failures. Following the classical approach used in landscape evolution model (Willgoose et al., 1991; Tucker and Slingerland, 1994), Dionisos combines two large-scale processes: a hill-slope creeping and a faster

Fig. 5. Model setup including three structural steps that represent the Early Oligocene steps shown in Fig. 3. Two submarine canyons, conduits of sediments transported from the east, were assumed to exist at 37 Ma. Abbreviations: WMF—the Judea western mountain front; CPS—the coastal plain step; CMFZ—the continental margin fault zone step.

18

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

water discharge driven transport (Willgoose et al., 1991; Tucker and Slingerland, 1994). The former is simulated by a linear slope-driven diffusion equation where the transport is proportional with the gradient of the slope. The later is simulated by a non-linear water- and slope-driven diffusion equation. The combination of these two transports laws leads to the following sediment transport equation:   n m−1 ! ∇h Q s ¼ − Ks þ K w Q w S where Qs is the sediment flux [km2/year]; h [m] is the topographic elevation; Ks and Kw are the diffusion coefficient respectively for the slow creeping transport and the faster water-driven process [km2/year]; Qw is dimensionless local water discharge at the cell (that is the local water discharge normalized by 100 m3/(s.km)) [–]; S is the local gradient of the basin slope [–]; n and m are constant, usually between 1 and 2. Sediment is assumed to be composed of a finite number of grain-size fractions. Transport equation and mass conservation are applied to each grain-size fraction, leading to the definition of the local sedimentation or erosion rate. Sedimentation occurs at a point in the basin if the transport capacity decreases (either because the slope decreases or the water flow spreads). Contrarily, erosion occurs if the transport capacity increases. The stratigraphic architecture of a sedimentary basin, such as the Levant Basin, is controlled by the transport of sediment, but also by the accommodation space and the supply of sediment. The accommodation is determined by basement motion (subsidence/uplift) and sea-level variation together. The same stratigraphy may be produced in more than one way, for example, the same accommodation may be created by sea level as well as regional subsidence, and the resulting stratigraphy may look the same, although the two histories and causal processes are profoundly different. Forward modeling provides the possibility of

running a series of scenarios of possible basin histories, providing the user with a number of outputs to analyze and test against geological constraints. Following this procedure, likely and unlikely sets of parameters, i.e., scenarios of geological history may be determined. 4. Model setup In order to focus on the capability of submarine canyons to transport enough sediments to fill the Levant Basin as observed, a simplified rectangular model, 300 km × 265 km, was built in a way that one of its sides schematically represents the Levant continental margin (model shown in Fig. 5, location marked in Figs. 1 & 3). At that side of the model (approximately east), three morphological steps were introduced in a way that their sizes resemble the real paleo-bathymetric steps formed in the Late Eocene–Early Oligocene. For simplicity, the steps in the model are parallel with no along-strike variations. As described above, the morphological steps played a crucial role in controlling sediment transport paths and largely contributed to the thickness variations across the Levant (Bar et al., 2013). At 37 Ma the Judea Western Mountain Front step was 500 m high, the coastal plain step was 1000 m, and the Continental Margin Fault Zone was a 1500 m feature. At that time, submarine canyons already crossed the steps and transported sediments into the deep basin (Fig. 3). Thus, we introduced to the initial simplified rectangular model two canyons with dimensions similar to the real canyons (Fig. 5). From the initial conditions shown by Fig. 5 onwards, the Judean upland uplifted while the Levant Basin subsided. These vertical motions were introduced into the model by three simplified elevation maps of the 37 Ma sea floor (base Saqiye Group), which is considered as the model's basement. The three maps are represented by the three cross sections of Fig. 6a (based on Bar et al., 2013) with no along-strike variations: (a) the 37 Ma surface before its burial (sea floor); (b) the

Fig. 6. (a) Schematic cross sections representing the vertical motions introduced into the model by three elevation maps of the model's basement (that is, the base Saqiye Group unconformity surface). Blue—37 Ma; Yellow—16 Ma; Black—present. (b) Reconstructed cross section for 7 Ma showing what simulation results should look like, that is approximately 4 km thick section and 1900 m of water column at the western side of the basin and very little deposition nearshore.

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

37 Ma surface at its buried position in 16 Ma; and (c) the 37 Ma surface at its a present day position. From these maps, we defined the subsidence history of the “basement” by linear interpolation of the motion. Compaction was simulated using burial depth–porosity curves Φ = Φo e−z/D, where Φ is the porosity of the sediment, z the burial depth, Φo is the initial porosity (at the time of the deposition), and D a constant characterizing the decay of the porosity. Φo and D are equal to 20% and 2000 m, and 80% and 500 m, for sand and shale respectively. An Airy-type isostasy was assumed with no flexural forces included. For sea-level changes, the Haq et al. (1987) eustatic curve was used.

19

Our modeling strategy is to infer the sediment influx from the volume of the sedimentary section and the duration of its deposition. Then, we deduce from modeling the water discharge needed to produce the pattern of deposition. The challenge in our case is to get a parameter set (sediment and water supplies) that (1) transports sediments through the submarine channels into the distal basin, (2) with only minor deposition along the proximal continental margin (Fig. 6b). As described below such a pattern of transport and deposition requires huge values of water discharge, the feasibility of which is examined in light of data about drainage area and relief.

Fig. 7. Examples of unsuccessful (right-hand side) and successful (left-hand side) models using a common values of diffusion coefficient K(sand) = 0.2 km2/ka. Each model represents four snapshots for 36 Ma, 23 Ma, 16 Ma, and 7 Ma. The bottom panels at 7 Ma are compared to geological constraints as shown in Fig. 6b. d is considered unsuccessful because most of the sediments were caught nearshore in contrast with the geological data (Fig. 6b). h is considered successful because most of the sediments were transported and deposited in the deep basin as required by the geological constraints.

20

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

Our modeled cube contains 4 km thick section in the deep part of the basin, which is 265 km long (N–S) and 175 km wide (E–W) plus a 0.3 km thick section over the intermediate block, which is 265 km long and 75 km wide. This totals (185,500 + 5962 km3) 191,462 km3 deposited during 30 Ma, that is Qs = 6383 km3/Ma. For simplicity we use a round number of 6500 km3/Ma. Note that at present the modeled section (base Late Eocene to base Messinian) is only ~3.6 km thick and the value of 4 km approximates the decompacted thickness of that section prior to its burial by a ~2 km thick Messinian–present section.

In addition, minor amounts of carbonates were also added. We assume that while siliciclastic supply from the emerging continents increased dramatically at 37 Ma, carbonate production remained in its previous rate. Based on the thickness and duration of the Santonian to Middle Eocene section (Steinberg et al., 2011) we used a production rate of 10 m/Ma for pelagic carbonates (100–4000 m depth range) and 50 m/Ma for platform carbonates (0–90 m depth range). With these rates during the 30 Ma period modeled, pelagic carbonates contribute 300 m to the 4300 m thick section modeled; and platformal

Fig. 8. Examples of unsuccessful (right-hand side) and successful (left-hand side) models using extremely high diffusion coefficient K(sand) = 20 km2/ka. Each model represents four snapshots for 36 Ma, 23 Ma, 16 Ma, and 7 Ma. The bottom panels at 7 Ma are compared to geological constraints as shown in Fig. 6b. d and h are unsuccessful and successful results as explained in the captions of Fig. 7.

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

carbonates, which are limited by the available accommodation space, are even less important. However, for completeness purposes, we included these components in the model. The diffusion approach for fluvial (Paola et al., 1992) and marine (Kaufman et al., 1991; Steckler et al., 1999; van Heijst et al., 2001; Lai and Capart, 2007; Gerber et al., 2009; Spinewine et al., 2011) transport has been validated both theoretically and empirically, yet, finding numerical values for parameters in real cases is still a challenge. The diffusion coefficient depends on a number of factors, such as grain size, lithology, roundness, water depth, water discharge, vegetation, river type, etc., that are difficult to quantify. Not surprisingly, the values of the coefficient vary greatly by case study. To list a few: 1.0–6.9 km2/ka (Paola et al., 1992); 0.05–1.10 km2/ka (Rivenaes, 1992); 0.2–0.6 km2/ka (den Bezemer et al., 2000); 0.01 km2/ka (Gawthorpe et al., 2003); 3.9 × 106–1.6 × 107 km2/ka (Burgess et al., 2006). In those cases the authors used a linear diffusion model, however, if a non-linear model (the sediment flux is proportional to a power of the water discharge) is applied, the coefficient may become much lower than in the linear case. Since transport distance highly depends on the diffusion coefficients and since we are interested in marine transport, we ran simulations with Kw values extending over three orders of magnitude. For each Kw value, we gradually increased the water discharge, from 100 m3/s to 100,000 m3/s, until reaching a stratigraphic model that fit the geological constraints. Furthermore, we assumed that clastic sediments are composed of sand and shale particles. In all cases Kw,shale was set as 5 times Kw,sand. 5. Simulation results

Kw,

Fig. 7 presents a “non-successful” and “successful” simulations using 2 sand = 0.2 km /ka, which is a reasonable value adopted from the

21

literature (Rabineau et al., 2005; Burgess et al., 2006; Alzaga-Ruiz et al., 2009; Somme et al., 2009 and Csato et al., 2012). The simulation that fails to produce a model that fits the geological constraints (lefthand side of Fig. 7) is the one that was run with Qw = 2830 m3/s (the water discharge of the present-day Nile River). The first snapshot (Fig. 7a) taken only 1 Ma after startup already shows that with these parameters sediments are not transported to the distal basin, but rather, get stuck close to the shoreline, where they build a progradational delta. Accordingly, the final result at 7 Ma (Fig. 7d) is very different from the observed stratigraphy. The simulation that does produce a successful model is the one that was run with Qw = 52,000 m3/s. Only with this extremely high water discharge value (see Discussion section below) are sediments transported all the way through the submarine channels to the deep basin. The final result of this model at 7 Ma (Fig. 7h) fits the geological data, that is, a ~4 km thick section in the deep basin with a water column of 1.9 km and only a few hundred meters of section over the intermediate block. Examples for extremely large Kw, sand = 20 km2/ka (two orders of magnitude larger than the previous example) are shown in Fig. 8, which also presents non-successful and successful models. For Qw = 100 m3/s the model fails to meet the observations (the left-hand side of Fig. 8), whereas for Qw = 2830 m3/s (present-day Nile River) a good match with the geological constraints is achieved. The results of the models of Figs. 7–8 plus the results of many other simulations are summarized in Figs. 9–10. Fig. 9 demonstrates the stratigraphy of the final model at 7 Ma as a function of K and Qw. Note that for Kw, sand = 0.2 km2/ka most of the sediments accumulate at the nearshore zone even for a large Qw of 10,000 m3/s; similarly for Qw = 100 m3/s, sediments are stuck nearshore even for unreasonably high Kw, sand of 20 km2/ka. Detailed analysis of model sensitivity to varying K's and Qw's is further shown in Fig. 10, which plots the water

Fig. 9. Summary of modeling results at 7 Ma for increasing K (leftwards) and Qw (upwards) values. Note that for Kw, sand = 0.2 km2/ka most of the sediments accumulate at the nearshore zone even for very high Qw of 10,000 m3/s; similarly for Qw = 100 m3/s sediment are caught nearshore even if Kw, sand is as extremely high as 20 km2/ka. Models that fit the geological observations are only at the right upward corner.

22

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

depth and sediment thickness obtained for the deepest (westernmost) part of the basin as a function of Qw and K. The geological data (Fig. 6b) require that these parameters approach ~1900 m and ~4000 m, respectively, that is, successful models are those that fall within the area marked by gray bands. The plots of Fig. 10a,b highlight the range of Qw that strongly affects the results (the steep part of the curves) as opposed to the range of variations that do not affect the results (flat parts of the curves). For a constant Kw, sand of 0.2 km2/ka, successful models are obtained for Qw N 30,000 m3/s; for constant Kw, sand of 2 km2/ka, successful models are obtained for Qw of 4000–10,000 m3/s; and for constant Kw, sand of 20 km2/ka, successful models are obtained for Qw of 1000–3000 m3/s. The cases presented in Figs. 7 and 8 are represented by colored circles. In a similar way Fig. 10c,d indicates that for a constant Qw of 2830 m3/s successful models are obtained only for Kw, sand between 4 and 7 km2/ka and that for larger K values the curves flatten. 6. Discussion To evaluate the feasibility of the eastern supply model for the Levant Basin we use global relationships between sediment loads, water discharge, and drainage area (Milliman and Syvitski, 1992; Dai and Trenberth, 2002; Syvitski and Milliman, 2007). Fig. 11a shows that sediment discharge of 6500 km3/Ma (deduced from the thickness and duration of the sequence studied) is presently measured in rivers with a drainage area varying over three orders of magnitude from 1000 km2 to 5,000,000 km2; thus, this factor alone is not very useful for testing model results. However, the high water discharge values (2830– 52,000 m3/s) required to satisfy the observed depositional pattern

narrows the range of possible drainage areas to 30,000–5,000,000 km2 (Fig. 11b). Considering world data about the elevation of drainage basins, Fig. 11a further indicates that for a given sedimentary load, as drainage area decreases elevation increases. This means that in order to explain the given sedimentary load with a drainage area in the order of 30,000 km2 (point C in Fig. 11), not only do we need to have exceptionally high diffusion coefficient (Kw,sand = 20 km2/ka), but we also need to have an elevated mountain range of at least 2500 m above sea level that will supply the required amount of sediments from a relatively small area. In other words, a model with common diffusion coefficients requires a fluvial system in the order of the largest rivers that exist today on earth in terms of drainage area and water discharge (point A in Fig. 11). Alternatively, models with high or very high diffusion coefficients (points B and C, respectively) can explain the geological observations using rivers that are an order of magnitude smaller in terms of water discharge and drainage area, but require elevated drainage basins. Are these two alternatives possible for the studied case? A priori, the first alternative (point A) sounds impossible. A drainage area of 2,500,000 km2 is almost as large as that of the Nile basin (the longest river on earth with a drainage area of 3,254,555 km2) and ~3 times larger than that of the Tigris (375,000 km2) and Euphrates (500,000 km2) basins together. This means that almost the entire Arabian Peninsula of that time (see reconstructed shoreline in Fig. 1) and certainly an area larger than the Kingdom of Saudi Arabia (2,000,000 km2) drained northwestwards to the Levant Basin. In terms of water discharge (52,000 m3/s) it requires an enormous amount of water more than ten

Fig. 10. Sensitivity analysis of modeling to varying Qw (a, b) and K (c, d). Model results are presented by plotting the sediment thickness (a, c) and water depth (b, d) at the deepest (westernmost) part of the basin. Gray zone marks the range of sediment thickness (~1900 m) and water depth (~4000 m) required by the geological constraints (Fig. 6b). The two “successful” and two “unsuccessful” cases of Figs. 7 and 8 are shown by colored circles.

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

times the sum of the Tigris, the Euphrates (1750 m3/s when they converge at the Shatt al Arab), and the Nile (2830 m3/s) together; that implies, a prevailing climate that produces water ten times more than the precipitation over the entire Middle East region today. The question remaining is whether the alternatives with the high (Kw, sand = 2 km2/ka) and very high (Kw, sand = 20 km2/ka) diffusion coefficients are possible. Considering only the size of the drainage area, the possibility that 250,000 km2 (point B) to 30,000 km2 (point C) of the northern part of the Arabian Peninsula had drained to the Mediterranean at that time cannot be excluded. However, water discharge of 10,000 or 2800 m3/s, respectively, requires that precipitation over these areas alone had produced an amount of water that presently feeds the Nile. In addition, it requires an elevated drainage area that is 1200–2500 m above sea level for point B (A ~ 250,000 km2) and

23

N2500 m above sea level for point C (A ~ 30,000 km2). In reality, this region has been elevating since the Late Eocene (Bar, 2009; Avni et al., 2012, reference) and still has not reached these elevations except for exceptional mountain peaks (Fig. 1). Thus, this alternative is unlikely for the study area. To summarize, filling of the Levant Basin from eastern sources alone does not seem plausible. Even a Nile-size river discharge is not enough to produce the observed stratigraphy unless extreme hydraulic conditions are assumed. Alternatively, for normal range hydraulics, extremely large water discharges and extremely large drainage areas are necessary. This conclusion is consistent with the southward thickening of Late Eocene to Early Miocene rocks within the deep Levant Basin (Macgregor, 2011; Steinberg et al., 2011) that hint at a major

Fig. 11. (a) Sediment load (Qs) vs. area of drainage basin (A) in present world rivers based on “Database of World Rivers and their Sediment Yields”, AquaStat, Food and Agriculture organization of the United States (www.fao.org/nr/water/aquastat/sediment/). Gray horizontal band represents the range of possible drainage area that corresponds with the approximate value of sediment load (6500 km3/Ma) used for modeling the Levant Basin. Color code represents basin relief in meters (R). Note that a given value of sediment load corresponds to a large range of drainage areas, depending on their relief. (b) Water discharge (Qw) vs. area of drainage basin in present world rivers based on Dai and Trenberth (2002) downloaded from www. cgd.ucar.edu/cas/catalog/surface/dai-runoff/. Gray horizontal band represents the range of possible drainage areas that correspond with three Qw values (2830, 10,000, and 52,000 m3/s) that yield models that fit the geological data (Figs. 7–9). These Qw values correspond to Kw, sand = 20, 2, and 0.2 km2/ka, respectively. Blue circles correspond to possibilities discussed in text. Point A represents a combination which requires a river in the order of the largest rivers on earth. Points B and C represent conditions which fit the geological data with less extreme Qw and A values, but instead, require very high K and R values (see text). Point R represents a more realistic combination in which only 1/6 of sediment supply to the Levant Basin arrived from the east.

24

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

sedimentary source from Africa. Nonetheless, some supply from Arabia definitely arrived to the basin through submarine canyons, as indicated by seismic studies showing northwest trending drainage patterns (Gardosh et al., 2008). However, the exact amount of sediment that arrived from the east is hard to evaluate. Based on shoreline reconstruction in Israel (Gvirtzman et al., 2011), Egypt (Salem, 1976) and Saudi Arabia (Ziegler et al., 2001), the map of Fig. 1 schematically illustrates the main drainage directions (thick black arrows) from Arabia at that time and the area that may have drained into the Mediterranean (gray polygon), which is in the order of 100,000–200,000 km2 (approximately 400 km × 400 km). A regional truncation surface recently described by Bar (2009) and Avni et al. (2012) indicates that the Oligocene topography in this area was relatively flat. According to the Qs–A relationships of Fig. 11a for areas with a relief of 200–600 m or 600–1200 m, Qs is expected to be 400–1100 km3/Ma (point R in Fig. 11a), that is, less than 1/6 of the sediment that entered the modeled cube (which does not include the entire basin). Assuming effective rainfall of 200 mm per year, the water discharge from an area of 100,000– 200,000 km2 should have been ~1000 m3/s (point R in Fig. 11b). To illustrate the influence of such a supply from the east on the deep basin stratigraphy and on the stacking patterns along the Levant margins, we reran our model with eastern sources of Qs = 1000 km3/Ma, Qw = 1000 m3/s and Kw,sand = 2 km2/ka (point R in Fig. 11) and south and west sources of Qs = 5600 km3/Ma and Qw = 5600 m3/s. Note however that the south and west boundaries of the modeled cube do not reach land area and the parameters used are not realistic. They were used to artificially replace ocean currents that presumably transported sediments that arrived to the eastern Mediterranean from the coasts of North Africa. The aim of this model is to illustrate the relatively minor role of the Levant sources, which approximately amounts to 1/6 of the total sediment supply. For better modeling one needs to include the entire basin and use geological constraints from wells of the Nile cone and the deep Levant Basin. At this stage, however, while

these data are confidential, our model (Fig. 12) schematically illustrates that sand units are spread in the entire basin with increasing amounts towards the south (under the Pliocene Nile cone, which is not modeled here) and also towards the mouth of the Levant margin canyons. 7. Conclusions In the Late Cenozoic large amounts of terrigenous material entered the Levant Basin from two main sources: (a) a proto-Nile fluvial system that transported sediments from Africa through the Egyptian continental margin; and (b) an easterly fluvial system that transported sediments from Arabia through the Israeli continental margin. Geological evidence indicates that prior to the blocking of the easterly transport system by the Dead Sea rift valley and before the establishment of the modern Nile Cone (that is, before the Pliocene) deposition in the Levant Basin was mostly concentrated in its deeper part and very little along its Levant margin. This means that most of the sediments that reached the Levant from Arabia were either trapped inland and did not reach the continental margin or that they bypassed the continental margin through several known submarine canyons and settled down in the deep part of the basin. 3D stratigraphic forward model carried out in this study indicates that filling of the Levant Basin only by sediments that were transported through the submarine canyons is possible only by assuming an unreasonably huge fluvial system in the order of the largest rivers that exist today on earth in terms of drainage area and water discharge. Such rivers obviously did not exist in the Levant area at that time. Alternatively, models with extreme hydraulic conditions in terms of diffusion coefficients can explain the observed stratigraphy with rivers that are an order of magnitude smaller, but this requires an elevated drainage basin that couldn't have existed in the study area. We conclude that the Levant Basin was fed mainly from Africa as also indicated by thickening of units southwards.

Fig. 12. Simulation results for eastern sources amounting to Qs = 1000 km3/Ma, Qw = 1000 m3/s, and Kw,sand = 2 km2/ka and south and west sources amounting to Qs = 5600 km3/Ma, Qw = 5600 m3/s. Note however that the south and west boundaries of the modeled cube do not reach land area and the parameters used are not realistic. They were used to artificially replace ocean currents that presumably transported sediments that reached the eastern Mediterranean from the coasts of North Africa. The aim of this model is to illustrate the relatively minor role of the Levant sources, which approximately amounts to 1/6 of the total sediment supply. Model results at 7 Ma are color coded by sand ratio (a) and facies (b). Cross sections along the mouth of the Levant margin submarine canyons (c) and across the southern canyon (d) are shown to illustrate how supply from the east influenced stacking patterns and lithology along the Levant slope.

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

A preliminary estimation of sediment supply from the east, which is based on paleogeographic reconstructions, indicates that Arabian sources probably contributed less than 1/6 of the sediment supply to the Levant Basin and probably influenced only its eastern side close to the canyons mouths. The question of how much sediment can be transported through submarine channels, how far it reaches, and how widely it is dispersed is a fundamental problem in understanding lowstand system tracts and their dependency on tectonics and climate. Here we demonstrate how numerical modeling can quantify such problems as a function of basin geometry and hydraulic conditions; and how modeling results can be evaluated in light of global datasets of present-day rivers along with paleogeographic knowledge. References Adams, C.G., Gentry, A.W., Whybrow, P.J., 1983. Dating the terminal Tethyan event. In: Meulenkamp, J.E. (Ed.), Reconstruction of Marine Paleoenvironments. Utrecht Micropaleontological Bulletins, Utrecht, pp. 273–298. Adams, E.W., Schlager, W., Wattel, E., 1998. Submarine slopes with an exponential curvature. Sedimentary Geology 117, 135–141. Alzaga-Ruiz, H., Granjeon, D., Séranne, M., Lopez, M., Roure, F., 2009. Gravitational collapse and Neogene sediment transfer across the western margin of the Gulf of Mexico: insights from numerical models. Tectonophysics 470, 21–41. Avni, Y., Segev, A., Ginat, H., 2012. Oligocene regional denudation of the northern Afar dome: pre and syn-breakup stages of the Afro-Arabian plate. Geological Society of America Bulletin 124, 1871–1897. Bar, O., 2009. The shaping of the continental margin of Central Israel since the Late Eocene-tectonics. Morphology and Stratigraphy.Ben Gurion University, Be'er Sheva 207. Bar, O., Gvirtzman, Z., Feinstein, S., Zilberman, E., 2013. Accelerated subsidence and sedimentation in the Levant Basin during the Late Tertiary and concurrent uplift of the Arabian platform: tectonic versus counteracting sedimentary loading effects. Tectonics 32, 334–350. Bates, C.C., 1953. Rational theory of delta formation. AAPG Bulletin 37, 2119–2162. Begin, Z.B., Meyer, D.F., Schumm, S.A., 1981. Development of longitudinal profiles of alluvial channels in response to base-level lowering. Earth Surface Processes and Landforms 6, 49–68. Bertoni, C., Cartwright, J.A., 2005. 3D seismic analysis of circular evaporite dissolution structures. Eastern Mediterranean Journal of Geological Society 162, 909–926. Bosworth, W., Huchon, P., McClay, K., 2005. The Red Sea and Gulf of Aden basins. Journal of African Sciences 43, 334–378. Bouma, A.H., 2000. Coarse-grained and fine-grained turbidite systems as end member models: applicability and dangers. Marine and Petroleum Geology 17, 137–143. Bradford, S.F., 1996. Numerical Modeling of Turbidity Current Hydrodynamics and Sedimentation. University of Michigan 220. Buchbinder, B., 1996. Miocene carbonates of the eastern Mediterranean, the Red Sea and the Mesopotamian basin: geodynamic and eustatic controls. Buchbinder, B., Calvo, R., Siman Tov, R., 2005. The Oligocene in Israel: a marine realm with intermittent denudation accompanied by mass-flow deposition. Israel Journal of Earth Sciences 54, 63–85. Burgess, P., Lammers, H., van Oosterhout, C., Granjeon, D., 2006. Multivariate sequence stratigraphy: tackling complexity and uncertainty with stratigraphic forward modeling, multiple scenarios, and conditional frequency maps. American Association of Petroleum Geologists Bulletin 90, 1883–1901. Burke, K., 1996. The African plate. South African Journal of Geology 99, 341–409. Calvo, R., Bartov, Y., 2001. Hazeva Group, southern Israel: new observations, and their implications for its stratigraphy, paleogeography, and tectono-sedimentary regime. Israel Journal of Earth Sciences 50, 71–99. Cantero, M.I., 2007. Direct Numerical Simulation of Thermohaline and Particulate Density Currents. University of Illinois. Carson, M.A., Kirkby, M.J., 1972. Hillslope Form and Process. Cambridge University Press, Cambridge. Coulthard, T.J., 1999. Modelling Upland Catchment Response to Holocene Environmental Change. School of Geography, University of Leeds, U.K. Csato, I., Granjeon, D., Catuneanu, O., Baum, G.R., 2012. A three-dimensional stratigraphic model for the Messinian crisis in the Pannonian Basin, eastern Hungary. Basin Research 25, 121–148. Culling, W., 1960. Analytical theory of erosion. Journal of Geology 68, 336–344. Dai, A., Trenberth, K.E., 2002. Estimates of freshwater discharge from continents: latitudinal and seasonal variations. Journal of Hydrometeorology 3, 660–687. den Bezemer, T., Kooi, H., Kranenborg, J., Cloetingh, S., 2000. Modelling grain-size distributions. A comparison of two models and their numerical solution. Tectonophysics 320, 347–373. Druckman, Y., Buchbinder, B., Martinotti, G., Siman Tov, R., Aharon, P., 1995. The buried Afiq Canyon (eastern Mediterranean. Israel): a case study of a Tertiary submarine canyon exposed in Late Messinian times. Marine Geology 123, 167–185. Fielding, C.R., Frank, T.D., Birgenheier, L.P., Rygel, M.C., Jones, A.T., Roberts, J., 2008. Stratigraphic imprint of the Late Palaeozoic Ice Age in eastern Australia: a record of alternating glacial and nonglacial climate regime. Journal of the Geological Society 165, 129–140.

25

Fleischer, L., Gafsou, R., 2003. Top Judea Group digital structural map of Israel, Scale 1:200,000 (2 sheets). The Geophysical Institute of Israel. Galloway, W.E., 1998. Siliciclastic slope and base-of-slope depositional systems: component facies, stratigraphic architecture, and classification. AAPG Bulletin 82, 569–595. Gardosh, M., Druckman, Y., 2006. Stratigraphy and tectonic evolution of the Levantine Basin, offshore Israel. In: Robertson, A.H.F., Mountrakis, D. (Eds.), Tectonic Development of the Eastern Mediterranean. The Geological Society of London, London, pp. 201–227. Gardosh, M., Buchbinder, B., Druckman, Y., Calvo, R., 2008. The Oligo-Miocene deepwater system of the Levant Basin. Geological Survey of Israel. 73. Gardosh, M.A., Garfunkel, Z., Druckman, Y., Buchbinder, B., 2010. Tethyan rifting in the Levant Region and its role in Early Mesozoic crustal evolution. In: Homberg, C., Bachmann, M., Homberg, C., Bachmann, M. (Eds.), The Geological Society of London, pp. 9–36. Garfunkel, Z., Horowitz, A., 1966. Upper Tertiary and Quaternary morphology of the Negev, Israel. Israel Journal of Earth Sciences 15, 101–117. Gawthorpe, R.L., Hardy, S., Ritchie, B., 2003. Numerical modelling of depositional sequences in half-graben rift basins. Sedimentology 50, 169–185. Gerber, T.P., Amblas, D., Wolinsky, M.A., Pratson, L.F., Canals, M., 2009. A model for the long-profile shape of submarine canyons. Journal of Geophysical Research—Earth Surface 114. Granjeon, D., 1996. Modélisation stratigraphique déterministe—conception et applications d'un modèle diffusif 3D multilithologique, Géosciences Rennes. Université de Rennes 1, Rennes. Granjeon, D., 2009. 3D stratigraphic modeling of sedimentary basins. AAPG Annual Convention and Exhibition. AAPG Search and Discovery, Denver, Colorado, USA. Granjeon, D., Joseph, P., 1999. Concepts and applications of a 3-D multiple lithology, diffusive model in stratigraphic modeling. In: Harbaugh, J.W., Watney, W.L., Rankey, E.C., Slingerland, R., Goldstein, R.H. (Eds.), Numerical Experiments in Stratigraphy: Recent Advances in Stratigraphic and Sedimentologic Computer Simulations, pp. 197–210. SEPM Special Publications. Granjeon, D., Wolf, S., 2007. 3D stratigraphic modeling in complex tectonics area. AAPG Annual Convention and Exhibition. AAPG Search and Discovery Long Beach, California, USA. Gvirtzman, G., Buchbinder, B., 1978. The Late Tertiary of the Coastal Plain and continental shelf of Israel and its bearing of the history of the Eastern Mediterranean. In: Ross, D.H., Neprochnov, Y.P. (Eds.), Initial Reports of the Deep Sea Drilling Project. US Government Printing Office, Washington, pp. 1195–1222. Gvirtzman, Z., Steinberg, J., 2012. Inland jump of the Arabian northwest plate boundary from the Levant continental margin to the Dead Sea Transform. Tectonics 31, TC4003. Gvirtzman, Z., Zilberman, E., Folkman, Y., 2008. the Jaffa Basin offshore central Israel Reactivation of the Levant passive margin during the late Tertiary and formation of. Journal of the Geological Society 165, 563–578. Gvirtzman, Z., Steinberg, J., Bar, O., Buchbinder, B., Zilberman, E., Siman-Tov, R., Calvo, R., Grossowicz, L., Almogi-Labin, A., Rosensaft, M., 2011. Retreating Late Tertiary shorelines in Israel: implications for the exposure of north Arabia and Levant during Neotethys closure. Lithosphere 3, 95–109. Hadler-Jacobsen, F., Gardner, M.H., Borer, J.M., 2007. Seismic stratigraphic and geomorphic analysis of deep-marine deposition along the West African continental margin. In: Davies, R.J., Posamentier, H.W., Wood, L.J., Cartwright, J.A. (Eds.), Seismic Geomorphology: Applications to Hydrocarbon Exploration and Production. Geological Society London, pp. 47–84. Haq, B.U., Hardenbol, J., Vail, P.R., 1987. Chronology of fluctuating sea levels since the Triassic. Science 235, 1156–1167. Hsu, K.J., Cita, M.B., Ryan, W.B.F., 1973. The origin of the Mediterranean evaporites. In: Ryan, W.B.F., Hsu, K.J. (Eds.), DSDP Initial Report US Gov. Printing office, Washington DC, pp. 1203–1232. Jordan, T.E., Flemings, P.B., 1991. Large-scale architecture, eustatic variation and unsteady tectonism: a theoretical evaluation. Journal of Geophysical Research 96, 6681–6699. Kassem, A., Imran, J., 2004. Three-dimensional modeling of density current. II. Flow in sinuous confined and unconfined channels. Journal of Hydraulic Research 42, 591–602. Kaufman, P., Grotzinger, J.P., McCormik, D.S., 1991. Depth-dependent diffusion algorithm for simulation of sedimentation in shallow marine depositional systems. Bulletin of the Kansas Geological Survey 489–508. Kenyon, P.M., Turcotte, D.L., 1985. Morphology of a delta prograding by bulk sediment transport. Geological Society of America Bulletin 96, 1457–1465. Lai, S.Y.J., Capart, H., 2007. Two-diffusion description of hyperpycnal deltas. Journal of Geophysical Research—Earth Surface 112. Macgregor, D., 2011. Rift shoulder source to pro-delta sink: The Cenozoic development of the Nile drainage system. AAPG International Conference and Exhibition, Milano, Italy. Miller, K.G., Kominz, M.A., Browning, J.V., Wright, J.D., Mountain, G.S., Katz, M.E., Sugarman, P.J., Cramer, B.S., Christie-Blick, N., Pekar, S.F., 2005. The phanerozoic record of global sea-level change. Science 310, 1293–1298. Milliman, J.D., Syvitski, J.P.M., 1992. Geomorphic/tectonic control of sediment discharge to the ocean: the importance of small mountainous rivers. Journal of Geology 100, 525–544. Mitchell, N.C., Huthnance, J.M., 2008. Oceanographic currents and the convexity of the uppermost continental slope. Journal of Sedimentary Research 78, 29–44. Morehead, M.D., Syvitski, J.P., Hutton, W.H., 2001. The link between abrupt climate change and basin stratigraphy: a numerical approach. Global and Planetary Change 28, 107–127. Mulder, T., Alexander, J., 2001. The physical character of subaqueous sedimentary density flows and their deposits. Sedimentology 48, 269–299. Murray, A.B., Paola, C., 1994. A cellular model of braided rivers. Nature 371, 54–57. Paola, C., Heller, P.L., Angevine, C.L., 1992. The large-scale dynamics of grain-size variation in alluvial basins, 1: theory. Basin Research 4, 73–90.

26

Z. Gvirtzman et al. / Marine Geology 347 (2014) 12–26

Parker, G., Fukushima, Y., Pantin, H.M., 1986. Self-accelerating turbidity currents. Journal of Fluid Mechanics 171, 145–181. Parker, G., Paola, C., Whipple, K.X., Mohrig, D., 1998. Alluvial fans formed by channelized fluvial and sheet flow: I: theory. Journal of Hydraulic Engineering 124, 985–995. Peck, J., 2008. Giant oil prospects lie in distal portion of offshore East Mediterranean basin. Oil and Gas Journal 106. Picard, l, 1943. Structure and Evolution of Palestine with Comparative Notes on Neighbouring Countries. Hebrew University of Jerusalem, Jerusalem. Rabineau, M., Berne, S., Aslanian, D., Olivet, J.L., Joseph, P., Guillocheau, F., Bourillet, J.F., Ledrezen, E., Granjeon, D., 2005. Sedimentary sequences in the Gulf of Lion: a record of 100,000 years climatic cycles. Marine and Petroleum Geology 22, 775–804. Rivenaes, J.C., 1992. Application of a dual-lithology, depth-dependent diffusion equation in stratigraphic simulation. Basin Research 4, 133–146. Rivenaes, J.C., 1997. Impact of sediment transport efficiency on large-scale sequence architecture: results from stratigraphic computer simulation. Basin Research 9, 91–105. Roberts, G., Peace, D., 2007. Hydrocarbon plays and prospectivity of the Levantine Basin, offshore Lebanon and Syria from modern seismic data. GeoArabia 12, 99–124. Ryan, W., 2009. Decoding the Mediterranean salinity crisis. Sedimentology 56, 95–136. Said, R., 1981. The Geological Evolution of the River Nile. Springer-Verlag, New York. Salem, R., 1976. Evolution of Eocene–Miocene sedimentation patterns in parts of Northern Egypt. American Association of Petroleum Geologists 60, 34–64. Schlager, W., Adams, E.W., 2001. Model for the sigmoidal curvature of submarine slopes. Geology 29, 883–886. Segev, A., Rybakov, M., Lyakhovsky, V., Hofstetter, A., Tibor, G., Goldshmidt, V., Ben Avraham, Z., 2006. The structure, isostasy and gravity of the Levant continental margin and the southeast Mediterranean. Tectonophysics 425, 137–157. Skene, K.I., Mulder, T., Syvitski, J.P.M., 1997. INFLO1: a model predicting the behaviour of turbidity currents generated at river mouths. Computers and Geosciences 23, 975–991. Somme, T.O., Helland-Hansen, W., Granjeon, D., 2009. Impact of eustatic amplitude variations on shelf morphology, sediment dispersal, and sequence stratigraphic interpretation: icehouse versus greenhouse systems. Geology 37, 587–590.

Spinewine, B., Sun, T., Babonneau, N., Parker, G., 2011. Self-similar long profiles of aggrading submarine leveed channels: analytical solution and its application to the Amazon channel. Journal of Geophysical Research—Earth Surface 116. Steckler, M.S., Mountain, G.S., Miller, K.G., Christie-Blick, N., 1999. Reconstruction of Tertiary progradation and clinoform development on the New Jersey passive margin by 2-D backstripping. Marine Geology 154, 399–420. Steinberg, J., Gvirtzman, Z., Folkman, Y., Garfunkel, Z., 2011. The origin and nature of the rapid late Tertiary filling of the Levant Basin. Geology 39, 355–358. Syvitski, J.P.M., Hutton, E.W.H., 2001. 2D SEDFLUX 1.0C: an advanced process-response numerical model for the fill of marine sedimentary basins. Computers and Geosciences 27, 731–753. Syvitski, J.P.M., Milliman, J.D., 2007. Geology, geography, and humans battle for dominance over the delivery of fluvial sediment to the coastal ocean. Journal of Geology 115, 1–19. Tucker, G.E., Slingerland, R., 1994. Erosional dynamics, flexural isostasy, and longlived escarpments: a numerical modeling study. Journal of Geophysical Research 10, 12229–12243. van Heijst, M., Postma, G., Meijer, X.D., Snow, J.N., Anderson, J.B., 2001. Quantitative analogue flume-model study of river-shelf systems: principles and verification exemplified by the Late Quaternary Colorado river-delta evolution. Basin Research 13, 243–268. Willgoose, G.R., Bras, R.L., Rodriguez-Iturbe, I., 1991. A physically based coupled network growth and hillslope evolution model, 1, theory. Water Resources Research 27, 1671–1684. Wright, L.D., Friedrichs, C.T., 2006. Gravity-driven sediment transport on continental shelves: a status report. Continental Shelf Research 26, 2092–2107. Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292, 686–693. Ziegler, P.A., Cloetingh, S., Guiraud, R., Stampfli, G.M., 2001. Peri-Tethyan platforms: constraints on dynamics of rifting and basin inversion. Memoires du Museum National d'Histoire Naturelle 186, 9–49. Zilberman, E., 1991. Landscape evolution in the central and northern Negev in the Neogene and the Quaternary. Geological Survey of Israel, Jerusalem.