Geomorphic evolution of the Malta Escarpment and implications for the Messinian evaporative drawdown in the eastern Mediterranean Sea

ACCEPTED MANUSCRIPT 1 Manuscript submitted to Geomorphology: Geomorphic evolution of the Malta Escarpment and implications for the Messinian evaporative drawdown in the eastern Mediterranean Sea. Aaron Micallefa*, Angelo Camerlenghib, Aggeliki Georgiopoulouc,d, Daniel GarciaCastellanose, Marc-André Gutscherf, Claudio Lo Iaconog, Veerle A.I. Huvenneg, Joshu J. Mountjoyh, Charles K. Paulli, Timothy Le Basg, Daniele Spatolaa, Lorenzo Facchinb, Daniela Accettellab a Marine Geology and Seafloor Surveying, Department of Geosciences, University of Malta, Msida, MSD 2080, Malta. b Istituto Nazionale di Oceanografia e di Geofisica Sperimentale (OGS), Trieste, Italy. c UCD School of Earth Sciences, University College Dublin, Dublin, Ireland. d UCD Earth Institute, University College Dublin, Dublin, Ireland. e Instituto de Ciencias de la Tierra Jaume Almera, CSIC, Barcelona, Spain. f Laboratoire Géosciences Océan, University of Brest/CNRS, IUEM, Pl. N. Copernic, Plouzané, 29280, France. g Marine Geoscience, National Oceanography Centre, University of Southampton Waterfront Campus, European Way, Southampton, UK. h National Institute of Water and Atmospheric Research, Wellington, New Zealand. i Monterey Bay Aquarium Research Institute, Moss Landing, CA, USA. *Corresponding author: E-mail: aaron.mica[email protected]; Telephone: +356 23403662. ACCEPTED MANUSCRIPT Aaron Micallef, Angelo Camerlenghi, Aggeliki Georgiopoulou, Daniel Garcia-Castellanos, Marc-André Gutscher, Claudio Lo Iacono, Veerle A.I. Huvenne, Joshu J. Mountjoy, Charles K. Paull, Timothy Le Bas, Daniele Spatola, Lorenzo Facchin, Daniela Accettella , Geomorphic evolution of the Malta Escarpment and implications for the Messinian evaporative drawdown in the eastern Mediterranean Sea. Geomor (2018), https://doi.org/10.1016/j.geomorph.2018.11.012 Geomorphology 327: 264-283 (2019) ACCEPTED MANUSCRIPT 2 ABSTRACT Carbonate escarpments are submarine limestone and dolomite cliffs that have been documented in numerous sites around the world. Their geomorphic evolution is poorly understood due to difficulties in assessing escarpment outcrops and the limited resolution achieved by geophysical techniques across their steep topographies. The geomorphic evolution of carbonate escarpments in the Mediterranean Sea has been influenced by the Messinian salinity crisis (MSC). During the MSC (5.97-5.33 Ma), the Mediterranean Sea became a saline basin due to a temporary restriction of the Atlantic-Mediterranean seaway, resulting in the deposition of more than one million cubic kilometres of salt. The extent and relative chronology of the evaporative drawdown phases associated to the MSC remain poorly constrained. In this paper we combine geophysical and sedimentological data from the central Mediterranean Sea to reconstruct the geomorphic evolution of the Malta Escarpment and infer the extent and timing of evaporative drawdown in the eastern Mediterranean Sea during the MSC. We propose that, during a MSC base-level fall, fluvial erosion formed a dense network of canyons across the Malta Escarpment whilst coastal erosion developed extensive palaeoshorelines and shore platforms. The drivers of geomorphic evolution of the Malta Escarpment after the MSC include: (i) canyon erosion by submarine gravity flows, with the most recent activity taking place <2600 cal. years BP; (ii) deposition by bottom currents across the entire depth range of the Malta Escarpment; (iii) tectonic deformation in the southern Malta Escarpment in association with a wrench zone; (iv) widespread, small-scale sedimentary slope failures preconditioned by oversteepening and loss of support due to canyon ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 3 erosion, and triggered by earthquakes. We carry out an isostatic restoration of the palaeoshorelines and shore platforms on the northern Malta Escarpment to infer an evaporative drawdown of 1800 – 2000 m in the eastern Mediterranean Sea during the MSC. We interpret the occurrence of pre-evaporite sedimentary lobes in the western Ionian Basin as suggesting that either evaporative drawdown and canyon formation predominantly occurred before salt deposition, or that only the latest salt deposition at the basin margin occurred after the formation of the sedimentary lobes. Keywords: Malta Escarpment; geomorphic evolution; submarine canyon; palaeoshoreline; sea level drawdown; Messinian salinity crisis Declarations of interest: none. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 4 1. INTRODUCTION 1.1 The Malta Escarpment Carbonate escarpments are submarine limestone and dolomite cliffs with relief in excess of 1 km; they can be up to 650 km long and have average slope gradients exceeding 40° (Schlager and Camber, 1986; Twichell et al., 1990). Their morphologies are complex and variable, comprising scarps, overhangs, vertical walls, terraces and wide canyons (e.g. Paull et al., 1987; Twichell et al., 1991; 1996; Eberli et al., 2005). Carbonate escarpments have been documented in numerous sites around the world, e.g. offshore Florida (Paull and Dillon, 1980; Paull et al., 1990a), Bahamas (Freeman-Lynde et al., 1981), Yucatan Peninsula (Paull et al., 2014), north-west Australia (Veevers, 1974), Malta (Scandone et al., 1981), and the Balearic Islands (Acosta et al., 2001). Carbonate escarpments were the focus of several studies in the 1970s and 1980s. Understanding of their formation has been limited by difficulties in assessing their outcrops and the limited resolution achieved by 2D seismic reflection and single beam echo sounding techniques across steep topographies. An origin based on carbonate platform growth and erosion of the escarpments’ faces has been put forward to explain the topography of the world’s largest carbonate escarpments (Dillon et al., 1987; Paull et al., 1990a; 1991; 2014). This contrasts with the predominantly structural origin attributed to carbonate escarpments in the Mediterranean Sea (e.g. Emile Baudot Escarpment ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 5 (Mauffret et al., 2004), Malta Escarpment (Cita et al., 1980; Scandone et al., 1981; Casero et al., 1984; Jongsma et al., 1985; Reuther et al., 1993)). The Malta Escarpment is a steep, 290 km long submarine cliff with a relief in excess of 3 km. It extends from the eastern margin of Sicily southwards to the Medina Seamounts (Fig. 1). The most detailed surveys of the Malta Escarpment were primarily aimed at constraining the age, stratigraphy and nature of this feature using single-beam echosounders, seismic reflection profilers, gravity coring, dredge sampling and submersible dive observations (Cita et al., 1980; Scandone et al., 1981; Biju-Duval et al., 1983; Casero et al., 1984). The Malta Escarpment is arguably the best sampled of all carbonate escarpments worldwide. Nevertheless, its fine-scale geomorphology is still poorly constrained, and gaps remain in our understanding of the geomorphic evolution of the Malta Escarpment and its key driving forces. In-depth understanding of these aspects is necessary to better assess the hazard that the Malta Escarpment poses to seafloor infrastructure and densely populated coastal settlements in its vicinity, and to refine the tectonic and oceanographic history of the region. FIGURE 1 1.2 Messinian salinity crisis The Messinian salinity crisis (MSC) is an extraordinary event in the geological history of the Mediterranean region that had local to global consequences (Ryan, 2008a; Roveri et ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 6 al., 2014a; Flecker et al., 2015). Between 5.97 to 5.33 Ma, the Mediterranean Sea became an enormous saline basin due to a temporary restriction of the Atlantic-Mediterranean seaway, resulting in the deposition of more than one million cubic kilometres of salt that can locally exceed 3 km in thickness (Ryan, 2008b). Among the many open questions related to the MSC, the extent and relative chronology of the drawdown phases, following the deposition of the primary gypsum and prior to any significant sea-level change (i.e., the second step sensu Clauzon et al. (1996)), is poorly understood and three different scenarios have been proposed: (i) Only the shallow Messinian continental shelves were exposed as a consequence of a moderate drawdown (Roveri et al., 2001; 2008; 2014a). Deep basin unconformities in seismic reflection profiles from Mediterranean margins are interpreted as originating from different processes (e.g. cascading brines (Roveri et al., 2014b) or deep water evaporite dissolution (Gvirtzman et al., 2017)). The halite body and the overlying upper evaporites are therefore interpreted as the result of deposition in a deep, saturated, brine-filled basin. (ii) A kilometre-scale drawdown occurred following a period of deep-water brine concentration (Ryan, 2008b; Lofi et al., 2011b). According to this scenario, the erosional surfaces are the product of subaerial processes. In the shallow margins, these surfaces pre-date the salt deposition, whereas in the deep margins and basin floors, the erosional surfaces are coeval to the salt deposition during the “evaporative shrinking” of the basin. Evaporite emplacement started in deep water environments and ended in shallow water environments. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 7 (iii)A kilometre-scale drawdown occurred before salt deposition, with the latter occurring in a shallow water environment (Bache et al., 2009; 2015). The main evidence for this scenario is the interpretation of the erosional surface extending below the Messinian evaporites in the Gulf of Lion. For scenarios (ii) and (iii), attempts at quantifying the maximum drawdown from seismic data around the rim of the Mediterranean that contained the evaporite pinch-out generated are as follows (Fig. 1a):  Ebro Margin: ~1300 m (Urgeles et al., 2011); ~2000 m (Stampfli and Höker, 1989);  Valencia Basin: ~1500 m (Maillard and Mauffret, 1993);  Gulf of Lion: ~2000 m (Ryan, 1976; Lofi, 2002; Steckler et al., 2003);  Levant Margin: 1000 m (Druckman et al., 1995); ~1500 m (Ben-Gai et al., 2005; Tibor and Ben-Avraham, 2005);  Nile Delta: 1500-2250 m (Gargani and Rigollet, 2007); 3000-4000 m (Barber, 1981). Numerical modelling by Blanc (2000) produced drawdown estimates of 1140 m and 1000 m in the western and eastern Mediterranean basins, respectively, while those of Ryan (2008b) suggest a maximum drawdown of ~1800 m and ~2400 m in the western and eastern Mediterranean basins, respectively. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 8 What is conspicuously absent in the above studies are observations from the central Mediterranean Sea. Reconstructions of past sea-level changes (Imbrie et al., 1989) and stratigraphic analyses (Max et al., 1993; Osler and Algan, 1999) suggest that, after the Mediterranean Sea was refilled at the end of the MSC, the central Mediterranean Sea has experienced low sedimentation rates (~6 cm ka-1) because it has largely remained isolated from inputs of fluvial and littoral sediments. The Malta Escarpment and western Ionian Basin should thus host topographic and sedimentary signatures of the key stages of the MSC in the eastern Mediterranean Basin. 1.3 Objectives In this study we analyse geophysical and sedimentological data from the Malta Escarpment and adjacent western Ionian Basin to: (i) reconstruct the geomorphic evolution of the Malta Escarpment since the MSC, and (ii) infer the extent and timing of evaporative drawdown during the MSC in the eastern Mediterranean Sea. 2. REGIONAL SETTING The Malta Escarpment is the north-eastern boundary of the Pelagian Platform, which consists of continental crust forming part of the African continental plate (Finetti, 1982). The north-eastern Pelagian Platform comprises the 100-150 m deep, gently sloping Malta Plateau (Micallef et al., 2011; 2016) (Fig. 1b). Here, parallel-bedded units of PlioPleistocene terrestrial, pelagic and hemipelagic sediments, which are up to 300 m thick, ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 9 overlie >4500 m thick older sedimentary sequences (Jongsma et al., 1985; Max et al., 1993; Torelli et al., 1995; Osler and Algan, 1999). These older sequences outcrop across the Malta Escarpment and were dredged during a number of expeditions (Cita et al., 1980; Scandone et al., 1981; Casero et al., 1984) (Fig. 2). They comprise Triassic to Cretaceous shallow platform carbonates and Cretaceous to Miocene shelf edge carbonate build ups (Scandone et al., 1981; Pedley et al., 1993). The upper Malta Escarpment is covered by Tortonian to Recent pelagic deposits (Biju-Duval et al., 1982). The Mesozoic succession of the Malta Escarpment is punctuated by extensive depositional hiatuses, as well as tuffs and pillow lavas deposited during a number of volcanic episodes (Scandone et al., 1981; Pedley et al., 1993; Bosellini, 2002). FIGURE 2 The Malta Escarpment was thought to have originated by rifting in the Upper PermianTriassic, followed by spreading from the Jurassic until the Upper Cretaceous-early Tertiary (Ben-Avraham and Grasso, 1991; Grasso, 1993; Catalano et al., 2000a). More recently, Catalano et al. (2000b) suggested that continental rifting took place from the pre-Triassic till the Early Cretaceous, and was followed by spreading in the Early Cretaceous. Adam et al. (2000) propose that, since the onset of plate convergence between Africa and Europe during the Late Cretaceous, the Malta Escarpment was transformed from a passive margin into a mega-hinge fault system with an additional sinistral strike-slip component. For this reason, the Malta Escarpment exhibits both Mesozoic normal block faulting and more recent sinistral strike-slip movement ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 16 FIGURE 3 FIGURE 4 4.1 Northern Malta Escarpment Our dataset covers a 120 km-long section of the northern Malta Escarpment (Fig. 3a). Here, the escarpment is oriented north-northwest to south-southeast and faces eastnortheast. The total relief on the escarpment face increases to the south, reaching a maximum of 3200 m. The depth of the top and base of the escarpment increases southwards from 100 m to 600 m, and from 2600 m to 3700 m, respectively. The maximum overall slope gradient of the escarpment is 11°, although the highest local mean slope gradient is 74°. The steepest slopes are characteristically found along the lower escarpment. The seismic signature of the escarpment is predominantly comprised of acoustically transparent seismic facies topped by a high amplitude and irregular reflector (facies C) (Figs. 5, 6). On the steeper sections of the escarpment (>10°), facies C is overlain by a thin (<0.05 s Two Way Travel Time (TWTT)) package of high amplitude, sub-parallel and continuous reflectors (facies A) (Fig. 6a). Gentler slope gradients are draped by sequences of sub-parallel, convex upward, continuous reflectors (facies B) (Figs. 5, 6b). The thickness of this drape is up to 0.3 s TWTT. The backscatter response is low where the escarpment is covered by facies A and B (Fig. 3b). ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 17 FIGURE 5 FIGURE 6 4.1.1 Submarine canyons (a) Fine-scale morphology We define submarine canyons as steep-sided valleys on the continental shelf and slope. Submarine canyons are the most pervasive landforms across the northern Malta Escarpment (Fig. 4). The escarpment hosts more than two hundred canyons (60 canyons and 140 tributaries). The canyons have V-shaped cross-sections, sharp interfluves and pointed heads (Fig. 7a); their morphometric attributes are listed in Table 2. The heads of the canyons connect, or are located close, to the top of the northern Malta Escarpment (Fig. 7a). The canyon thalwegs are characterised by high backscatter, whereas their walls and interfluves have a low backscatter response (Fig. 3b). Some canyons host individual or series of crescent-shaped steps along their thalwegs (Fig. 7a). In general, the shapes of the canyon thalweg profiles are concave (Fig. 7b). This is in contrast with the convex longitudinal profile of the escarpment face that is not associated with canyons (Fig. 7b). FIGURE 7 TABLE 2 ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 18 The northern Malta Escarpment includes two larger canyons – Noto Canyon in the north, which is 27 km long, 15 km wide, and 1.5 km deep (Fig. 7c), and Cumecs Canyon in the south, which is 37 km long, 39 km wide, and 3 km deep (Fig. 7e). The morphology of Noto Canyon is unusual for the Malta Escarpment as it has a U-shaped cross-section and a theatre-shaped head that is up to 700 m high and 70° in overall slope gradient (Fig. 7d). Upslope of Noto Canyon, the seafloor has a mean slope gradient of 7° and hosts a dendritic network of tributary canyons. The alignment of the main tributary of Noto Canyon changes by 90° at a distance of 8 km upslope from its mouth. The northernmost tributaries of Cumecs Canyon are oriented northwest to southeast (Fig. 7e), which contrasts with the southwest to northeast orientation of the other tributaries along the same canyon and the majority of the canyons nearby. The seafloor on the bed or downslope from the mouths of these large canyons exhibits series of ellipsoidal depressions up to 1.5 km in width (Figs. 7c, e). A distinctive feature on the northern Malta Escarpment is a promontory that extends 8 km northeast from the general trend of the escarpment’s base (Fig. 3a). The promontory is the only section of the escarpment not cut by canyons. (b) Seismic facies The canyon thalwegs are characterised by acoustically chaotic to transparent seismic facies (facies D) that are up to 0.2 s TWTT thick and that overlie strata that are acoustically transparent and topped by a high amplitude and irregular reflector (facies C) ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 19 (Fig. 7d). The mouth of Cumecs Canyon is buried beneath 0.4 s TWTT of high amplitude, sub-parallel and continuous reflectors (facies A) (Fig. 7f). Along the northern flank of Cumecs Canyon, a package of sub-parallel, convex upward continuous reflectors (facies B) is slightly offset at three sites (Fig. 6b). (c) Sediment cores One 0.4 m-long core was retrieved from Cumecs Canyon (CU14_04; Fig. 8a). It predominantly consists of mud, whereas the core catcher contained gravel, pebbles and a ~10 cm-long limestone cobble (Fig. 8b). FIGURE 8 Five cores were retrieved from Noto Canyon – two from upslope of the tributary canyons (CU14_06 and _07), one from below the head (CU14_08), and two from the mouth (CU14_01 and _02). The two cores from upslope of the tributary canyons (CU14_06 and _07), taken from water depths of 133 m and 230 m, contain a distinct coarse layer composed of an aggregate of shells, sand and mud, which is overlain by a fining-upwards sequence (Fig. 8a). The radiocarbon date for the shell in this layer is 19,811 cal. years BP. Since that the shell is intact, it must have either been deposited in situ or it has not been transported far from its living position. We therefore consider the age of the coarse layer as <19,811 cal. years BP. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 20 The other cores display intervals of the canyon’s most recent activity. CU14_01 features a homogeneous sandy layer within the top 0.5 m of the core. This layer, deposited after ~2,600 cal. years BP (age derived from dating of foraminiferal assemblages of mixed planktonic species), is characterised by a sharp erosional base with parallel lamination. These features are evident in the X-ray and coincide with an increase in P-wave velocity, gamma ray density and magnetic susceptibility. This layer is not present in CU14_02. However, the physical property logs of CU14_02 show a relative increase of magnetic susceptibility at nearly the same depth below the seafloor (marked by a dashed line in fig. 8a). Likewise, we see a P-wave velocity peak in cores CU14_06 and _07 at a similar depth below the seafloor. About one metre below this event, core CU14_01 contains a fining-upward sandy layer that contains some shell fragments and that caps a layer of mixed mud clasts. The latter is separated by a sharp, straight boundary from an interval of thinly laminated mud and silt below, which is occasionally punctuated by thin, coarser, sandy layers (laminated facies). The sharp boundary at the top of the laminated facies is also very clearly identified in core CU14_02, but it lacks the clast-rich interval and fining up sand layer above it. The laminated facies is fully penetrated in core CU14_02, where it is 2 m thick and shows thicker and coarser sandy intervals punctuating it downcore. CU14_08, which was retrieved from the base of the Noto Canyon head, contains an interval that is nearly 0.5 m thick, comprising contorted and sheared mud layers (Fig. 8a). The core does not penetrate through the base of the layer, so this can only be inferred to be the minimum thickness. It is difficult to determine whether this layer correlates with the sandy layer and clast-rich layer of core CU14_01. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 21 We estimate an indicative sedimentation rate of 4.85 cm ka-1 based on the radiocarbon ages obtained from core CU14_06. Based on this sedimentation rate, the contorted interval in CU14_08 is 7,800 yr old. The homogeneous sandy layer and the fining upward sandy layer in CU14_01 are 2,600 yr and 20,000 yr old, respectively. Attempts at coring the smaller canyon thalwegs proved unsuccessful, and the corer returned clean and undamaged (Table 1). 4.1.2 Re-entrants A number of re-entrants cluster along the base of the southern half of the northern Malta Escarpment (Fig. 4). Re-entrants have U-shaped cross-sections and theatre-shaped heads; their morphometric attributes are listed in Table 2 (Fig. 7g). The base of the re-entrants is characterised by low backscatter, whereas their walls have a high backscatter response (Fig. 3b). 4.1.3 Scars More than two hundred scars have been identified across the northern Malta Escarpment (Fig. 4). The headwalls of the scars are generally linear to arcuate, steep (10-25°), and characterised by high backscatter. At their distal limit, most scars connect to either another scar or a canyon. The seabed within these scars is typically relatively smooth, ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 22 planar and characterised by low backscatter. The smaller scars are located along the heads and walls of submarine canyons (Fig. 9f). They have headwalls <50 m high and an average width of 1.3 km. The larger scars are located along the top of the escarpment (Fig. 7e), which are up to 23 km wide, with headwalls up to 200 m in height. FIGURE 9 4.1.4 Breaks of slope and terraces Thirteen breaks of slope were identified across the northern Malta Escarpment (Fig. 4). We define a break of slope as a change in profile curvature of > 0.4° m-1 (concave) or <- 0.4° m-1 (convex), occurring along a length of >2 km and not associated with a scar. The longest of these breaks of slope are concave in profile and include a 70 km terrace at a mean depth of 2478 m ± 119 m (1 S.D.) extending between Noto Canyon and Cumecs Canyon, and a 25 km-long terrace occurring within Cumecs Canyon at a mean depth of 2545 m ± 70 m (1 S.D.) (Fig. 9d). The surfaces of these terraces are smooth, up to 1 km in width, and slope at <5° (Fig. 9a). In seismic profiles, the terraces consist of a high amplitude and irregular reflector marking the top of an acoustically transparent seismic facies (facies C) (Fig. 9c). The terraces are generally draped by a mounded facies with sub-parallel, convex upward, continuous reflectors (facies B). A similar facies with comparable thickness occurs along the top of the escarpment. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 23 The remaining eleven breaks of slope are convex in profile, <7 km long, are not associated with a terrace, and occur in 350 – 2000 m water depths (Figs. 4, 9b, 9d). 4.2 Central Malta Escarpment The central Malta Escarpment is 90 km long, up to 16 km wide, faces northeast and is arcuate-shaped in plan view (Fig. 3a). The top and base of the escarpment occur at an almost invariable mean depth of 800 m and 3400 m, respectively. The maximum relief of the central Malta Escarpment is 2600 m and the maximum overall slope gradient is 8°, with local mean slope gradients of 50°. The fine-scale morphology is difficult to characterise in view of the lower resolution of the available bathymetric grid covering this section. However, the central Malta Escarpment hosts at least 35 canyons that share similar characteristics to those of the northern Malta Escarpment (Fig. 7h). The morphometric attributes of these canyons are listed in Table 2. 4.3 Southern Malta Escarpment 4.3.1 Fine-scale morphology The southern Malta Escarpment is dominated by Heron Canyon, which is the largest submarine canyon along the entire Malta Escarpment (Fig. 10a). Heron Canyon is 100 km long, up to 10 km wide and 1 km deep. Its depth ranges from 450 m at the head to 4000 m at the mouth, and its orientation changes from west-northwest to east-northeast, ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 24 to southwest to northeast, at 40 km from its mouth. The longitudinal profile of Heron Canyon is linear and characterised by three steps in its lower half; these are up to 30° in slope gradient and have heights of 130 m, 110 m and 110 m (Fig. 10b). Heron Canyon hosts 20 tributary canyons along its southern flank, which share similar morphologic characteristics to those in the northern and central Malta Escarpment; morphometrics are shown in Table 2. The upper half of Heron Canyon is characterised by three linear escarpments oriented northwest to southeast, which are up to 50 km long and 250 m high (Fig. 10a). FIGURE 10 Along most of its length, Heron Canyon has an asymmetric cross-section with a steeper northern flank. The downstream half of the northern flank is contiguous with a 20 km wide upland that has an average depth of 1500 m. Here, the flank is characterised by an escarpment that is up to 1200 m high and has an overall slope gradient of 25°. The eastern flank of the upland features a 13 km-long concave break of slope, which is associated with a 3 km wide terrace and which occurs at a mean depth of 2165 m ± 67 m (1 S.D.) More than 60 scars were mapped on Heron Canyon walls (Figs. 4, 10a). These scars have arcuate headwalls that are up to 100 m high, 25° steep, 7 km wide and characterised by high backscatter. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 25 4.3.2 Seismic facies We identify three facies in the multichannel seismic reflection profiles located upslope of Heron Canyon (Figs. 10c, d): (i) Facies with high amplitude, sub-parallel and continuous reflectors (facies A), which has a maximum thickness of 0.2 s TWTT. Reflectors are clearly offset vertically to sub-vertically by up to 0.1 s TWTT in places. These offset seismic reflectors are often associated to a step in the seafloor. (ii) Facies with low amplitude, sub-parallel and discontinuous reflectors (facies G); the latter are vertically to sub-vertically offset down to 2 s TWTT below facies A. (iii)Acoustically chaotic to transparent seismic facies, (facies D), which has a maximum thickness of 0.7 s TWTT and fills a buried depression. 4.4 Western Ionian Basin 4.4.1 Bedforms The surface of the Ionian Basin seafloor at the foot of the Malta Escarpment is generally smooth with a south-southeast slope of 0.6° (Fig. 3a). Twenty ellipsoidal depressions, up to 2 km wide and 125 m deep, occur along the base of the northern Malta Escarpment, mostly in proximity to the promontory and near the base of the Noto Canyon walls (Figs. 4, 9e). The depressions generally coincide with the mouths of submarine canyons. One ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 32 would have been sourced by slope failures on the outer Malta Plateau (Micallef et al., 2016) and the Malta Escarpment canyon walls (section 5.2.3). During sea-level lowstands, when the coastline would have been located close to the upper northern Malta Escarpment (Fig. 3a), subaerial fluvial discharge and possibly hyperpycnal flows, associated to wave or tidal suspension, could have generated gravity flows. We assign the high backscatter signature of the canyon thalwegs (Fig. 3b) to bedrock exposures and/or coarse sediment (Collier and Brown, 2005). These observations, combined with the thin Plio-Pleistocene sedimentary cover along canyon thalwegs (Fig. 6b), indicate that post-MSC pelagic and hemipelagic sedimentation within the canyons has been low, likely a result of transport and erosion by gravity flows. The above inferences do not apply to Noto Canyon, however. The unusual morphology of Noto Canyon is similar to that reported for bedrock canyons that were rapidly eroded by megafloods (Lamb et al., 2008; 2014). A buried 4 km wide and ~400 m deep erosional channel is located upslope of Noto Canyon, whilst the megaflood deposit (facies D) is located downslope of Noto Canyon (Micallef et al., 2018). For these reasons, Noto Canyon has been attributed to erosion of the Malta Escarpment during the passage of the Zanclean megaflood at the end of the MSC (Micallef et al., 2018). ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 33 5.2.2 Palaeoshoreline and shore platform development The concave breaks of slope and associated terraces, which correspond to the MSC erosional surface (Fig. 9c), are interpreted to have formed during the MSC and subsequently became draped by Plio-Pleistocene contouritic deposits. In view of their morphology and the low standard deviation associated with their depths, we propose that the concave breaks of slope and terraces are palaeoshorelines and shore platforms formed by wave erosion during a Messinian sea-level lowstand. The high salinity contrast between a brine-filled western Ionian Basin and freshwater discharge along the palaeocoastline could have enhanced dissolution processes (Plummer, 1975; Hanshaw and Back, 1980). The shoreline recession rate is estimated to be 0.003 m/a (based on a 3000 m recession in 100 ka (Gargani and Rigollet, 2007)), which is comparable to, or lower than, rates measured for coastal limestone cliff erosion (e.g. 0.0013 m/a in a coral atoll in the Indian Ocean (Trudgill, 1976); 0.018 m/a in Sweden (Rudberg, 1967); 0.07 m/a in Wales (Williams and Davies, 1987); 0.5 m/a in the Black Sea coast (Zenkovich et al., 1965)). An alternative explanation for the concave slope breaks and terraces may be the occurrence of lithologies with contrasting erosional resistances. However, neither the boreholes nor the dredge samples (Scandone et al., 1981; Casero et al., 1984) show significant changes in lithology at the depths at which the breaks occur. A lithological change is also unlikely to maintain a similar depth over a distance of almost 100 km, particularly if we consider that the eastern Pelagian Platform strata are not horizontal, as ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 34 indicated by the 500 m change in bathymetry along the top of the northern Malta Escarpment. Thus, a lithological contrast does not explain the development of concave slope breaks and terraces at a quasi-uniform depth along the escarpment. 5.2.3 Slope failures The numerous scars on the escarpment are evidence for slope instability. The style of deformation is either translational sliding, as indicated by the linear to arcuate steep headwalls and the smooth and planar scars (Locat and Lee, 2002), or debris flow, because most of the failed material has been evacuated from the scars. The slope failures occurred in stratified, fine-grained contouritic or hemipelagic/pelagic sediments deposited across the Malta Escarpment canyon walls and heads during the Plio-Pleistocene. Micallef et al. (2016) have shown that slope failures across the eastern Malta Plateau occur along specific stratigraphic horizons in the Plio-Pleistocene sediment package; this is likely to also be applicable to mass movements across the Malta Escarpment, because the sedimentological and stratigraphic characteristics of the material that has failed are similar. The most likely preconditioning factor for slope failure is over-steepening and loss of support as a result of canyon erosion. Seismicity, on the other hand, is likely to be the main trigger, because ground shaking associated with distal earthquakes of EMS-98 intensities >VIII has the potential of initiating slope failures in the eastern Malta Plateau (Micallef et al., 2016). Slope failures have played a key role in widening and extending canyons upslope and facilitating tributary development. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 35 5.2.4 Bottom current activity The contouritic drift deposits in the western Ionian Basin (facies B in fig. 11a) are indicative of the establishment of strong and long-term bottom currents, which flowed parallel to the Malta Escarpment, after the Mediterranean Sea was refilled at end of the MSC. The horizontal truncation of the underlying Zanclean megaflood deposit (facies D) to the WSW of the contouritic drift deposit (Fig. 11a) is indicative of erosion by the bottom currents. The ridges and troughs located at the seafloor along the base of the northern Malta Escarpment and the underlying undulating sub-parallel seismic reflectors, have been interpreted as sediment waves (Gutscher et al., 2016; Munari et al., 2016) (Figs. 3a, 11b). The occurrence of the above features, together with contouritic drift deposits on the escarpment (Figs. 4-6) and the outer Malta Plateau (Micallef et al., 2016) suggest that bottom currents flowing parallel to the escarpment have been active across the entire depth range of the Malta Escarpment. The unusually large depression located at the tip of the promontory (Fig. 9e) may be either a zone of non-deposition between the escarpment and the sediment wave field, or an erosional zone excavated by bottom current flow that was locally modified by the promontory. 5.2.5 Structural and tectonic processes The southern Malta Escarpment features long escarpments and offset seismic reflectors that provide clear evidence of recent fault activity. We interpret the several vertical and sub-vertical offsets of reflectors upslope of Heron Canyon (Fig. 10c-d) as normal faults ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 36 that displace sequences of Plio-Pleistocene pelagic/hemipelagic material and carbonate bedrock, and occasionally bound half-graben structures. Evidence of strike-slip deformation, on the other hand, is provided by twenty six earthquakes, which were recorded in the region since 1985 and which have a strike-slip focal mechanism (Figs. 2; 10a). We therefore infer that the interpreted faults are associated to both extensional and strike-slip kinematics. Faulting appears to have had a direct control on the morphology of Heron Canyon, in particular its dimensions, orientation, cross-sectional asymmetry, and linear longitudinal profile (as in Covault et al., 2011; Micallef et al., 2014a). We therefore infer that Heron Canyon constitutes the eastern extremity of the Sicily Channel Rift Zone, and that the latter likely corresponds to a wrench zone. The palaeoshoreline and shore platform identified in the southern Malta Escarpment are 335 m shallower than those recorded in the northern Malta Escarpment. This difference is similar to the total height of the steps measured along the bed of Heron Canyon (350 m), which we interpret as knickpoints formed due to changes in base level (e.g. Mitchell, 2006; Micallef et al., 2014a). We propose that the difference in depth is a response to tectonic deformation. Namely, regional base level was lowered due to uplift of the southern Malta Escarpment in association to wrench faulting (Jongsma et al., 1987). The vertical displacement was at least 335 m higher than that for the northern and central Malta Escarpment during the last 5.3 Ma. Other explanations may include differential subsidence due to lower sedimentation rates in the southern Malta Escarpment, or base level fluctuations in the eastern Mediterranean during the MSC and the formation of palaeoshorelines and shore platforms at multiple depths. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 37 The section of the Malta Escarpment between Siracusa and the southern Malta Escarpment lacks obvious topographic and seismic evidence of widespread, recent fault activity, confirming inferences made previously that this is concentrated north of Siracusa (Argnani and Bonazzi, 2005). Evidence of recent fault activity in our seismic reflection data is restricted to offsets in the Plio-Pleistocene sedimentary cover in the lower northern Malta Escarpment and north of Cumecs Canyon (Fig. 6b), which we interpret as normal faults. Our data do provide evidence of the geomorphic response to the occurrence of fault structures, however. The perpendicular change in orientation of Noto Canyon (Fig. 7c), as well as the linear, NW-SE oriented tributary of Cumecs Canyon (Fig. 7e), coincide with faults inferred from seismic data by Casero et al. (1984) (Fig. 2). We therefore propose that such abrupt changes in canyon orientation are fault-controlled (e.g. Noda et al., 2008; Micallef et al., 2014a). The shallow, convex breaks of slope identified across the northern Malta Escarpment may also be seafloor expressions of faults. 5.2.6 Hydrogeological processes The morphologic and backscatter characteristics of the re-entrants are not consistent with the action of either sediment gravity flows or subaerial fluvial incision. Rather, they share their characteristics with canyons reported on the Florida Escarpment and New Jersey margin. Because of their similarity to subaerial box canyons, such canyons are thought to have been formed by groundwater seepage (Paull and Neumann, 1987; Dunne, 1990; Paull et al., 1990b; Robb, 1990; Twichell et al., 1990; Nagihara, 1996). According to this ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 38 model, canyons are retrogressively eroded, both mechanically and chemically, into a seepage face. Corrosion, which is enhanced by mixing of solutions of different salinities (Plummer, 1975; Hanshaw and Back, 1980), is thought to be the most important erosional process associated with groundwater seepage in bedrock (Twichell et al., 1990). Groundwater seepage at the base of carbonate escarpments is mainly a result of fluid circulation driven by lateral thermal/salinity differences (e.g. Kohout, 1965; Paull and Neumann, 1987; Chanton et al., 1991; Hughes et al., 2007). Groundwater seepage may be an important agent forming re-entrants across the northern Malta Escarpment. Erosion by groundwater corrosion would explain the general absence of debris aprons at the mouths of the re-entrants. Fluid circulation across the escarpment is plausible, either due to a combination of excess pore pressure (due to overburden) and faulting, or geothermal convection. There are a number of uncertainties associated with the role of groundwater seepage in eroding re-entrants, however. First, we do not have direct or indirect evidence of fluid seepage at the base of the Malta Escarpment. Second, and perhaps more important, is the fact that the efficacy of fluid seepage as an erosive agent, both in subaerial and submarine contexts, still needs to be validated (Irwin et al., 2006; Lamb et al., 2006). 5.3 Implications for the MSC in the eastern Mediterranean Sea 5.3.1 Extent of evaporative drawdown in the eastern Mediterranean Basin ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 39 The submarine canyons, palaeoshorelines and shore platforms across the Malta Escarpment are all indicative of base-level fall in the eastern Mediterranean Basin during the MSC. By taking into consideration the results of the isostatic restoration, and assuming that the eastern Pelagian Platform has not been tectonically displaced since the Late Miocene (Argnani and Bonazzi, 2005), the palaeoshorelines point to a drawdown of 1800 – 2000 m in the eastern Mediterranean. This value is in line with the observations by Gargani and Rigollet (2007) from the Nile Delta, and by Ryan (1976), Stampfli and Hoker (1989), Lofi (2002), and Steckler et al. (2003) for the western Mediterranean Basin. 5.3.2 Timing of evaporative drawdown in the eastern Mediterranean Basin The thick lobes in the western Ionian Basin only occur in facies F, pre-date the evaporite sequence and spatially correlate with canyon mouths. We interpret these lobes as sediment deposits sourced by canyon erosion of the Malta Escarpment during the second step of the MSC, thus forming what has been called a Complex Unit (Clauzon et al., 1996; Lofi et al., 2011a). The fact that the complex palaeo-seafloor topography associated with these lobes is infilled and smoothened by subsequent deposition of the salt layer (facies E2; Fig. 11c) may suggest that the large-scale drawdown occurred before the salt deposition, in agreement with scenario (iii) described in section 1.2 and proposed by Bache et al. (2009; 2015). However, the salt layer at this location is thin, possibly deposited in a marginal setting, not necessarily in lateral continuity with the thicker and deeper basin salt body buried below the Ionian Basin. A similar situation is present on the lower slope of the steep Emile Baudot Escarpment on the southern margin of the Balearic ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 40 Promontory in the western Mediterranean Sea (Dal Cin et al., 2016; their fig. 10). In this case, the deposition of the marginal, shallower salt deposition occurred above the clastic lobes deposited simultaneously with the evaporative shrinking in the deeper basin, in accordance with the models by Ryan (2008b) and Lofi et al. (2011b). 5. CONCLUSIONS We used high resolution geophysical and sedimentological data from the central Mediterranean Sea to reconstruct the geomorphic evolution of the Malta Escarpment and infer the extent and timing of Messinian evaporative drawdown in the eastern Mediterranean Basin. Our main conclusions are the following: (i) Submarine canyons are the most common landform across the Malta Escarpment and were predominantly driven by subaerial fluvial erosion during the MSC. After refilling of the Mediterranean Basin at the end of the MSC, submarine canyons continued to be eroded by submarine gravity flows. The latter were sourced by slope instability in the outer Malta Plateau and canyon walls and heads on the Malta Escarpment. During sea-level lowstands, subaerial fluvial discharge and possibly hyperpycnal flows could have also generated gravity flows. Episodic gravity flow activity has taken place as recently as <2600 cal. years BP. (ii) The submarine canyons, palaeoshorelines and shore platforms are indicative of base-level fall during the MSC. Isostatic restoration of the palaeoshorelines points to a drawdown of 1800 – 2000 m in the eastern Mediterranean Sea. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 41 (iii)The thick pre-evaporite sedimentary lobes in the western Ionian Basin margin are interpreted as part of a Complex Unit. The occurrence of this unit would suggest either that evaporative drawdown and canyon formation occurred predominantly before salt deposition (in agreement with scenario (iii) proposed by Bache et al., 2009; 2015), or that only the latest salt deposition at the basin margin, near the lateral pinch-out of the evaporites, occurred after the formation of the sedimentary lobes. (iv) The post-MSC sedimentary environment in the Malta Escarpment and western Ionian Basin is dominated by hemipelagic sedimentation and deposits associated with bottom current activity, such as sediment waves and contouritic drift deposits. The latter involved strong and long-term bottom currents that flowed along the entire depth range of the Malta Escarpment after the end of the MSC. (v) Slope failures – predominantly translational slides and debris flows - are widespread and affect Plio-Pleistocene sediments. They are preconditioned by over-steepening and loss of support due to canyon erosion, and triggered by earthquakes. However, they are small in scale and not a significant escarpment formation process. (vi) Recent tectonic activity is restricted to the southern Malta Escarpment and was associated with extensional and strike-slip kinematics. Heron Canyon likely comprises the eastern extremity of a wrench zone. We estimate that, during the last 5.3 Ma, uplift in the southern Malta Escarpment could have been at least 335 m higher than the northern and central Malta Escarpment. This would explain the shallower depth of palaeoshorelines in the southern Malta Escarpment, although ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 48 Clauzon, G., Suc, J.P., Gautier, F., Berger, A., Loutre, M.F., 1996. Alternative interpretation of the Messinian salinity crisis, controversy resolved? Geology, 24, 363-366. Collier, J.S., Brown, C.J., 2005. Correlation of sidescan backscatter with grain size distribution of surficial seabed sediments. Marine Geology, 214, 431-449. Covault, J.A., Fildani, A., Romans, B.W., McHargue, T., 2011. The natural range of submarine canyon-and-channel longitudinal profiles. Geosphere, 7(2), 313332. Dal Cin, M., Del Ben, A., Mocnik, A., Accaino, F., Geletti, R., Wardell, N., Zgur, F., Camerlenghi, A., 2016. Seismic imaging of Late Miocene (Messinian) evaporites from Western Mediterranean back-arc basins. Petroleum Geoscience, 22(4), 297-308. Dart, C.J., Bosence, D.W.J., McClay, K.R., 1993. Stratigraphy and structure of the Maltese graben system. Journal of the Geological Society, 150, 1153-1166. de Voogd, B., Truffert, C., Chamot-Rooke, N., Huchon, P., Lallemant, S., Le Pichon, X., 1992. Two-ship deep seismic soundings in the basins of the eastern Mediterranean Sea (Pasiphae cruise). Geophysical Journal International, 109, 536-552. Dillon, W.P., Valentine, P.C., Paull, C.K., 1987. The Blake Escarpment - A product of erosional processes in the deep ocean. In: R. Cooper (Ed.), Symposium Series ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 49 for Undersea Research. National Oceanic and Atmospheric Administration, Washington, D.C., pp. 177-190. Druckman, Y., Buchbinder, B., Martinotti, G.M., 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. Dunne, T., 1990. Hydrology, mechanics, and geomorphic implications of erosion by subsurface flow. In: C.G. Higgins, D.R. Coates (Eds.), Groundwater Geomorphology: The Role of Subsurface Water in Earth-Surface Processes and Landforms. Special Paper Geological Society of America, pp. 1-25. Eberli, G.P., Anselmetti, F.S., Betzler, C., Van Konijnenburg, J.H., Bernoulli, D., 2005. Carbonate platform to basin transitions on seismic data and in outcrops: Great Bahama Bank and the Maiella Platform margin, Italy. AAPG Memoir(81), 207-250. Finetti, I., 1982. Structure, stratigraphy and evolution of the central Mediterranean. Bollettino di Geofisica Teorica e Applicata, 24, 274-315. Finetti, I., 1984. Geophysical study of the Sicily Channel Zone. Bollettino di Geofisica Teorica ed Applicata 26(101-102), 3-28. Finetti, I., Del Ben, A., 2005. Crustal tectono-stratigraphic setting of the Pelagian Foreland from new CROP Seismic data. In: I. Finetti (Ed.), CROP Project: ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 50 Deep Seismic Exploration of the Central Mediterranean and Italy. Elsevier, pp. 581-595. Finetti, I., Morelli, C., 1973. Geophysical exploration of the Mediterranean Sea. Bollettino di Geofisica Teorica e Applicata, 15, 263-341. Flecker, R., Krijgsman, W., Capella, W., de Castro Martíns, C., Dmitrieva, E., Mayser, J.P., Marzocchi, A., Modestou, S., Ochoa, D., Simon, D., Tulbure, M., van den Berg, B., van der Schee, M., de Lange, G., Ellam, R., Govers, R., Gutjahr, M., Hilgen, F., Kouwenhoven, T., Lofi, J., Meijer, P., Sierro, F.J., Bachiri, N., Barhoun, N., Alami, A.C., Chacon, B., Flores, J.A., Gregory, J., Howard, J., Lunt, D., Ochoa, M., Pancost, R., Vincent, S., Yousfi, M.Z., 2015. Evolution of the Late Miocene Mediterranean–Atlantic gateways and their impact on regional and global environmental change. Earth-Science Reviews, 150, 365392. Freeman-Lynde, R.P., Cita, M.B., Jadoul, F., Miller, E.L., Ryan, W.B.F., 1981. Marine geology of the Bahama Escarpment. Marine Geology, 44(1-2), 119-156. Gallais, F., Graindorge, D., Gutscher, M.-A., Klaeschen, D., 2013. Propagation of a lithospheric tear fault (STEP) through the western boundary of the Calabrian accretionary wedge offshore eastern Sicily (Southern Italy). Tectonophysics, 602, 141-152. Gallais, F., Gutscher, M.A., Graindorge, D., Chamot-Rooke, N., Klaeschen, D., 2011. A Miocene tectonic inversion in the Ionian Sea (central Mediterranean): ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 51 Evidence from multichannel seismic data. Journal of Geophysical Research, 115, B12108. Garcia-Castellanos, D., 2002. Interplay between lithospheric flexure and river transport in foreland basins. Basin Research, 14(2), 89-104. Garcia, M., Maillard, A., Aslanian, D., Rabineau, M., Alonso, B., Gorini, C., Estrada, F., 2011. The Catalan margin during the Messinian Salinity Crisis: Physiography, morphology and sedimentary record. Marine Geology, 284, 158-174. Gardiner, W., Grasso, M., Sedgeley, D., 1995. Plio-Pleistocene fault movement as evidence for mega-block kinematics within the Hyblean-Malta Plateau, Central Mediterranean. Journal of Geodynamics, 19(1), 35-51. Gargani, J., Rigollet, C., 2007. Mediterranean Sea level variations during the Messinian salinity crisis. Geophysical Research Letters, 34, L10405. 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, 114, F03002. Govers, R., Wortel, M.J.R., 2005. Lithosphere tearing at STEP faults: Response to edges of subudction zones. Earth and Planetary Science Letters, 236, 505-523. Granot, R., 2016. Palaeozoic oceanic crust preserved beneath the eastern Mediterranean. Nature Geoscience, 9, 701. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 52 Grasso, M., 1993. Pleistocene structures along the Ionian side of the Hyblean Plateau (SE Sicily): Implications for the tectonic evolution of the Malta Escarpment. In: M.D. Max, P. Colantoni (Eds.), Geological Development of the SicilianTunisian Platform. UNESCO Technical Reports in Marine Sciences. UNESCO, Urbino, pp. 49-54. Groupe Escarmed, 1982. Donnges nouvelles sur les marges du bassin Ionien profond. Revue de l'Institut Francais du Petrole, 37, 713-731. Gutscher, M.-A., Dominguez, S., de Lepinay, B.M., Pinheiro, L., Gallais, F., Babonneau, N., Cattaneo, A., Le Faou, Y., Barreca, G., Micallef, A., Rovere, M., 2016. Tectonic expression of an active slab tear from high-resolution seismic and bathymetric data offshore Sicily (Ionian Sea). Tectonics, 35(1), 39-54. Gutscher, M.A., Kopp, H., Krastel, S., Bohrmann, G., Garlan, T., Zaragosi, S., Klaucke, I., Wintersteller, P., Loubrieu, B., Le Faou, Y., San Pedro, L., Dominguez, S., Rovere, M., Mercier de Lepinay, B., Ranero, C.R., Sallares, V., 2017. Active tectonics of the Calabrian subduction revealed by new multi-beam bathymetric data and high-resolution seismic profiles in the Ionian Sea (Central Mediterranean). Earth and Planetary Science Letters, 461, 61-72. Gvirtzman, Z., Manzi, V., Calvo, R., Gavrierli, I., Gennari, R., Lugli, S., Reghizzi, M., Roveri, M., 2017. Intra-Messinian truncation surface in the Levant Basin explained by subaqueous dissolution. Geology, 45(10), 915-918. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 53 Hanshaw, B.B., Back, W., 1980. Chemical mass wasting of the northern Yucatan Penninsula by ground-water dissolution. Geology, 8, 222-224. Hsü, K.J., Montadert, L., Bernoulli, D., Bizon, G., Cita, M.B., Erickson, A.J., Fabricius, F.H., Garrison, R.E., Kidd, R.B., Melieres, F., Mueller, C., Wright, R.C., 1978. Site 374 - Messina abyssal plain, DSDP Deep Sea Drilling Project. Hughes, J.D., Vacher, H.L., Sanford, W.E., 2007. Three-dimensional flow in the Florida platform: Theoretical analysis of Kohout convection at its type locality. Geology, 35(7), 663-666. Illies, J.H., 1981. Graben formation - The Maltese Islands, a case study. Tectonophysics, 73, 151-168. Imbrie, J., McIntyre, A., Mix, A.C., 1989. Oceanic response to orbital forcing in the Late Quaternary: Observational and experimental strategies. In: A. Berger, S.H. Schneider, J.C. Duplessy (Eds.), Climate and Geosciences: A Challenge for Science and Society in the 21st Century. Reidel Publishing Company. Irwin, R.P., Howard, A.D., Craddock, R.C., 2006. Theater-headed valleys: The roles of overland flow and groundwater sapping. Lunar and Planetary Science, 1912. Jongsma, D., Van Hinte, J.E., Woodside, J.M., 1985. Geologic structure and neotectonics of the north African continental margin south of Sicily. Marine and Petroleum Geology, 2, 156-177. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 54 Jongsma, D., Woodside, J.M., King, G.C.P., Van Hinte, J.E., 1987. The Medina Wrench: A key to the kinematics of the central and eastern Mediterranean over the past 5 Ma. Earth and Planetary Science Letters, 82, 87-106. Kohout, F.A., 1965. A hypothesis concerning cyclic flow of salt water related to geothermal heating in the Floridan aquifer. New York Academy of Sciences Transactions, 28, 249-271. Lachenbruch, A., Morgan, P., 1990. Continental extension, magmatism and elevation; formal relations and rules of thumb. Tectonophysics, 174, 39-62. Lamb, M.P., Dietrich, W.E., Aciego, S.M., DePaolo, D.J., Manga, M., 2008. Formation of Box Canyon, Idaho, by megaflood: Implications for seepage erosion on Earth and Mars. Science, 320, 1067-1070. Lamb, M.P., Howard, A.D., Johnson, J., Whipple, K.X., Dietrich, W.E., Perron, J.T., 2006. Can springs cut canyons into rock? Journal of Geophysical Research, 111, E07002. Lamb, M.P., Mackey, B.H., Farley, K.A., 2014. Amphitheater-headed canyons formed by megaflooding at Malad Gorge, Idaho. Proceedings of the National Academy of Sciences of the United States of America, 111(1), 57-62. Le Bas, T.P., Hühnerbach, V., 1998. PRISM Processing of remotely-sensed imagery for seafloor mapping handbook, Southampton Oceanography Centre. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 55 Lermusiaux, P.F.J., Robinson, A.R., 2001. Features of dominant mesoscale variability, circulation patterns and dynamics in the Strait of Sicily. Deep-Sea Research I Oceanographic Research Paper, 48(9), 1953-1997. Lipparini, L., Scrocca, D., Marisili, P., Morandi, S., 2009. Offshore Malta licence in the Central Mediterranean Sea offers hope of hydrocarbon potential. First Break, 27, 105-116. Locat, J., Lee, H.J., 2002. Submarine landslides: advances and challenges. Canadian Geotechnical Journal, 39, 191-212. Lofi, J., 2002. La crise de salinité messinienne: Incidences directes et différées sur l´évolution sédimentaire de la marge du golfe du Lion. PhD, Lille, 277 pp. Lofi, J., Deverchere, J., Gaullier, V., Gillet, H., Gorini, C., Guennoc, P., Loncke, L., Maillard, A., Sage, F., Thinon, I., 2011a. Seismic Atlas of The "Messinian Salinity Crisis" markers in the Mediterranean and Black Seas. Lofi, J., Gorini, C., Berné, S., Clauzon, G., Tadeu Dos Reis, A., Ryan, W.B.F., Steckler, M., 2005. Erosional processes and paleo-environmental changes in the Western Gulf of Lions SW France) during the Messinian Salinity Crisis. Marine Geology, 217, 1-30. Lofi, J., Sage, F., Déverchère, J., Loncke, L., Maillard, A., Gaullier, V., Thinon, I., Gillet, H., Guennoc, P., Gorini, C., 2011b. Refining our knowledge of the Messinian salinity crisis records in the offshore domain through multi-site seismic analysis. Bulletin Societé Géolgique de France, 182, 163-180. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 56 Maesano, F.E., Tiberti, M.M., Basili, R., 2017. The Calabrian Arc: three-dimensional modelling of the subduction interface. Scientific Reports, 7(1), 8887. Maillard, A., Mauffret, A., 1993. Structure et volcanisme dans la fosse de Valence (Méditerranée Nord-occidentale). Bulletin de la Société Géologique de France, 164, 365-383. Mauffret, A., Frizon de Lamotte, D., Lallemant, S., Gorini, C., Maillard, A., 2004. E-W opening of the Algerian Basin (Western Mediterranean). Terra Nova, 16(5), 257-264. Max, M.D., Kristensen, A., Michelozzi, E., 1993. Small scale Plio-Quaternary sequence stratigraphy and shallow geology of the west-central Malta Plateau. In: M.D. Max, P. Colantoni (Eds.), Geological Development of the Sicilian-Tunisian Platform. UNESCO, Urbino, pp. 117-122. Micallef, A., Berndt, C., Debono, G., 2011. Fluid flow systems of the Malta Plateau, Central Mediterranean Sea. Marine Geology, 284, 74-85. Micallef, A., Berndt, C., Masson, D.G., Stow, D.A.V., 2007. A technique for the morphological characterization of submarine landscapes as exemplified by debris flows of the Storegga Slide. Journal of Geophysical Research, 112, F02001. Micallef, A., Camerlenghi, A., Garcia-Castellanos, D., Cunarro Otero, D., Gutscher, M.- A., Barreca, G., Spatola, D., Facchin, L., Geletti, R., Krastel, S., Gross, F., ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 57 Urlaub, M., 2018. Evidence of the Zanclean megaflood in the eastern Mediterranean Basin. Scientific Reports, 8(1), 1078. Micallef, A., Georgiopoulou, A., Mountjoy, J., Huvenne, V., Lo Iacono, C., Le Bas, T., Del Carlo, P., Cunarro Otero, D., 2016. Outer shelf seafloor geomorphology along a carbonate escarpment: The eastern Malta Plateau, Mediterranean Sea. Continental Shelf Research, 131, 12-27. Micallef, A., Mountjoy, J., Barnes, P.M., Canals, M., Lastras, G., 2014a. Geomorphic response of submarine canyons to tectonic activity: Insights from the Cook Strait canyon system, New Zealand. Geosphere, 10(5), 905-929. Micallef, A., Paull, C.K., in prep. Mesozoic carbonate escarpments: State-of-knowledge and future directions. Earth-Science Reviews. Micallef, A., Ribó, M., Canals, M., Puig, P., Lastras, G., Tubau, X., 2014b. Space-fortime substitution and the evolution of a submarine canyon–channel system in a passive progradational margin. Geomorphology, 221, 34-50. Miller, K.G., Mountain, G.S., Wright, J.D., Browning, J.V., 2011. A 180-million-year record of sea level and ice volume variations from continental margin and deep-sea isotopic records. Oceanography, 24(2), 40-53. Millot, C., Taupier-Letage, I., 2005. Circulation in the Mediterranean Sea, The Handbook of Environmental Chemistry. Springer, Berlin/Heidelberg, pp. 29-66. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 64 Torelli, L., Grasso, M., Mazzoldi, G., Peis, D., Gori, D., 1995. Cretaceous to Neogene structural evolution of the Lampedusa shelf (Pelagian Sea, Central Mediterranean). Terra Nova, 7, 200-212. Trudgill, S.T., 1976. The marine erosion of limestones on Aldabra Atoll, Indian Ocean. Zeitschrift für Geomorphologie, 26, 164-200. Twichell, D.C., Dillon, W.P., Paull, C.K., Kenyon, N.H., 1996. Morphology of carbonate escarpments as an indicator of erosional processes. In: J.V. Gardner, M.E. Field, D.C. Twichell (Eds.), Geology of the United States Seafloor: The View from GLORIA. Cambridge University Press, Cambridge, pp. 97-108. Twichell, D.C., Parson, L.M., Paull, C.K., 1990. Variations in the styles of erosion along the Florida Escarpment, eastern Gulf of Mexico. Marine and Petroleum Geology, 7(3), 253-266. Twichell, D.C., Paull, C.K., Parson, L.M., 1991. Terraces on the Florida escarpment: Implications for erosional processes. Geology, 19(9), 897-900. Urgeles, R., Camerlenghi, A., Garcia-Castellanos, D., De Mol, B., Garces, M., Verges, J., Haslam, I., Hardman, M., 2011. New contraints on the Messinian sealevel drawdown from 3D seismic data of the Ebro Margin, western Mediterranean. Basin Research, 23(2), 123-145. Veevers, J.J., 1974. Western continental margins of Australia. In: C.A. Burke, C.L. Drake (Eds.), The Geology of Continental Margins. Springer, New York, pp. 605616. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 65 Watts, A.B., 2001. Isostasy and Flexure of the Lithosphere. Cambridge University Press, Cambridge. Williams, A.T., Davies, P., 1987. Rates and mechanics of coastal cliff erosion in Lower Lias rocks, Coastal Sediments '87. American Society of Civil Engineers, pp. 1855-1870. Winnock, E., 1981. Structure du Bloc Pelagien. In: Wezel, I.F.C. (Ed.), . In: I.F.C. Wezel (Ed.), Sedimentary Basins of Mediterranean Margins. Tecnoprint, Bologna, pp. 445-464. Zenkovich, V.P., Ionin, A.S., P.A., K., 1965. Abrasion as the resources of debris in the sea shore zone. Institute of Oceanography, Academy of Sciences of the USSR, 76, 103-125. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 66 9. FIGURE LEGENDS Figure 1: (a) Location map of the central Mediterranean Sea. 1: Valencia Basin, 2: Ebro Margin, 3: Gulf of Lions, 4: Nile Delta, 5: Levant Margin. (b) Bathymetric map of the eastern margin of the Pelagian platform and western Ionian Basin. The map displays the principal morphological features, accretionary wedge thrust structures (solid white lines, with white teeth at deformation front), structures proposed as surface expressions of STEP fault and Alfeo-Etna fault (yellow line), and the Ionian fault (orange line) (Gutscher et al., 2016; Polonia et al., 2016). The main pathways of Modified Atlantic Water (MAW) and Levantine Intermediate Water (LIW) in eastern Sicily Channel are denoted by purple lines (Béranger et al., 2004; Ciappa, 2009). Background bathymetry is from Gutscher et al. (2017) and EMODnet bathymetry (http://www.emodnetbathymetry.eu). Figure 2: Map of the Malta Escarpment showing the spatial coverage of new and published multibeam echosounder, multi-channel reflection seismics, gravity cores and dredged samples used in this study. Location of figures 6, 11 and Oreste_001 well is indicated. Map includes location of earthquakes with a strike-slip focal mechanism recorded since 1985 (http://info.terremoti.ingv.it/). The background bathymetry is from Gutscher et al. (2017) and EMODnet bathymetry (http://www.emodnet-bathymetry.eu). The mapped faults (black lines) are from Lipparini et al. (2009), Gardiner et al. (1995) and Casero et al. (1984). ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 67 Figure 3: (a) Compilation of multibeam bathymetry data from the study areas. The pink line indicates the coastline during the Last Glacial Maximum (22 ka BP). The background bathymetry is from Gutscher et al. (2017) and EMODnet bathymetry (http://www.emodnet-bathymetry.eu). (b) Multibeam backscatter data from the northern Malta Escarpment. It includes enlarged sections of backscatter maps of examples of canyons and re-entrants. The isobaths, denoted in red, are at 1000 m intervals. The location of figures 7, 9 and 10 is indicated. Figure 4: Map of principal morphologic elements and seismic facies and features identified across the study area, draped on a shaded relief map of the seafloor. Figure 4a displays the north-western part of the study area; its location roughly coincides with that of figure 3b. Figure 4b shows the south-eastern continuation. Figure 5: Multichannel seismic reflection profile MEM07-104 correlated with Oreste 001 well (189 m water depth) and DSDP 374 (4088 m water depth; 100 km to the south-east of profile MEM07-104). (Facies A = high amplitude, sub-parallel, continuous reflectors; facies B = sub-parallel, convex upward reflectors; facies C = transparent facies; facies D = chaotic to transparent facies; facies E1 = highly reflective, discontinuous facies; facies E2 = reflector-less facies with basin-fill geometry; facies F = lenses of intermediate amplitude and sub-parallel reflectors). Location of profile MEM07-104 is in the middle of the northern Malta Escarpment. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 68 Figure 6: (a) Multichannel seismic reflection profile CIR-03 intersecting the northern Cumecs Canyon, showing thin parallel seismic reflector packages (facies A), or absence thereof, on the steeper parts of the northern Malta Escarpment. On gentler gradients, these packages have a mounded morphology (facies B) and reach a thickness of 0.3 s TWTT. Location of profile CIR-03 is shown in figure 2. Figure 7: (a) Multibeam bathymetric map of canyons. Inset shows enlarged section of crescent-shaped scars along the bed of a canyon. (b) Profiles of canyon thalwegs (location in figures a and g). (c) Multibeam bathymetric map of Noto Canyon. (d) Profile CIR-07, which provides a cross-section of Noto Canyon. Location of profile in figure 7c. (e) Multibeam bathymetric map of Cumecs Canyon. (f) CROP profile M-23A, which provides a cross-section of the mouth of Cumecs Canyon. Location of profile in figure e. (g) Multibeam bathymetric map of re-entrants at the base of the northern Malta Escarpment. (h) Multibeam bathymetric map of the central Malta Escarpment. Locations of figures a, c, e, g and h are shown in figure 3. (Facies A = high amplitude, sub-parallel, continuous reflectors; facies C = transparent facies; facies D = chaotic to transparent facies). Figure 8: (a) CUMECS-2 gravity core correlation panels, including core photographs, xradiographs, description, and P-wave velocity, gamma ray density and magnetic susceptibility plots. (b) Gravel, pebbles and a cobble retrieved from the core catcher at CU14_04. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 69 Figure 9: (a) Slope gradient map of the promontory in the northern Malta Escarpment showing a concave break of slope (denoted by black arrows) and associated terrace below. Isobaths at 1 km interval shown as black lines. (b) Slope gradient map of a convex break of slope (denoted by white arrows) in the northern Malta Escarpment. Isobaths at 1 km interval shown as white lines. (c) Multichannel seismic reflection profile CIR-04 intersecting the break of slope (denoted by dotted black line) and associated terrace. (Facies B = sub-parallel, convex upward reflectors; facies C = transparent facies). Location in (a). (d) Histogram of the depth of breaks of slope in the northern Malta Escarpment. (e) Depressions (denoted by black arrows) located at the base of the central part of the promontory. (f) Shallow arcuate scars (denoted by black arrows) on the northern interfluve of Cumecs Canyon. Location of figures is shown in figure 3. Figure 10: (a) Multibeam bathymetry map of the southern Malta Escarpment. The strongest earthquake recorded on the northern flank of Heron Canyon (M4.4; 30-102016) (http://info.terremoti.ingv.it/) is shown. (b) Longitudinal profile along the lower half of Heron Canyon, showing the three morphological steps along the canyon axis. Location in (a). (c) Interpreted multichannel seismic reflection profile MS-20 located upslope of Heron Canyon. (d) Interpreted multichannel seismic reflection profile MS-14 located upslope of Heron Canyon. Location of MS profiles in figure 2. (Facies A = high amplitude, sub-parallel, continuous reflectors; facies D = chaotic to transparent facies; facies G = low amplitude, sub-parallel and discontinuous reflectors). ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 70 Figure 11: (a) Interpreted multichannel seismic reflection profile CIR-04, intersecting the base of the central Malta Escarpment and the western Ionian Basin. (b, c) Interpreted multichannel seismic reflection profile CUMECS-3, located in the western Ionian Basin and parallel to Malta Escarpment. Location of profiles in figures 2 and 9e. (Facies A = high amplitude, sub-parallel, continuous reflectors; facies B = sub-parallel, convex upward reflectors; facies C = transparent facies; facies D = chaotic to transparent facies; facies E1 = highly reflective, discontinuous facies; facies E2 = reflector-less facies with basin-fill geometry; facies F = lenses of intermediate amplitude and sub-parallel reflectors). Figure c is redrawn from Micallef et al. (2018). Figure 12: Map showing isolines of post-MSC subsidence (in metres) of the seafloor across the study area, based on a flexural isostatic compensation of the weight of postMSC sediment and the water column, and using an equivalent elastic thickness (EET) of 50 km and 40 km. The tilting towards the east reflects the deeper water column loaded in that region relative to the Pelagian Platform in the west. Isobaths at 1 km interval shown as black lines. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 71 10. TABLES Table 1: Station Latitude (°N) Longitude (°E) Water depth (m) Recovery (m) CU-01 36.823872 15.370824 2307 2.25 CU-02 36.814108 15.373017 2281 2.85 CU-03 36.716713 15.473141 2835 0 CU-04 36.043433 15.556455 2578 0.42 CU-05 36.150213 15.588011 2915 0 CU-06 36.841167 15.210168 133 1.38 CU-07 36.84200 15.232400 230 2.16 CU-08 36.83330 15.31000 1830 1.11 ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 72 Table 2: Mean length (km) Mean length:width ratio Mean spacing (km) Mean bifurcation angle (°) Mean wall gradient (°) Northern Malta Escarpment - canyons 7.9 2.38 0.7 20 22 Northern Malta Escarpment – re-entrants 1.5 0.75 5.0 45 52 Central Malta Escarpment - canyons 8.7 3.21 2.2 18 18 Southern Malta Escarpment - canyons 9.1 3.31 2.18 21 18 11. TABLE LEGENDS Table 1: Location, depth and recovery of the gravity cores collected during the CUMECS-2 oceanographic survey. Table 2: Morphometric parameters of submarine canyons and re-entrants. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT 73 Highlights  We reconstruct the geomorphic evolution of the Malta Escarpment  During MSC, subaerial and coastal erosion formed canyons and palaeoshorelines  Gravity flows, bottom currents and tectonic deformation shaped escarpment after MSC  We infer a drawdown of 1800-2000 m in the eastern Mediterranean during MSC ACCEPTED MANUSCRIPT Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12