Nat Hazards (2010) 54:583–603 DOI 10.1007/s11069-009-9469-x REVIEW ARTICLE

Geohazards rooted from the northern margin of the Sea of Marmara since the late Pleistocene: a review of recent results Naci Go¨ru¨r • M. Namık C ¸ ag˘atay

Received: 13 March 2009 / Accepted: 6 October 2009 / Published online: 27 October 2009 Ó Springer Science+Business Media B.V. 2009

Abstract Owing to its special geodynamic setting on the western extension of the North Anatolian Fault (NAF) and oceanographic setting between Mediterranean and the Black Seas, the Sea of Marmara (SoM) is prone to various geohazards, such as earthquakes, submarine landslides, tsunamis and hypoxia. The NAF is a major transform plate boundary that has produced devastating historical earthquakes. The most active northern strand of the _ NAF cuts across sea floor, delimiting the northern margin of the SoM. With the 1999 Izmit (M 7.4) and 1912 S¸arko¨y (M 7.4) earthquakes having occurred on its eastern and western ends, a large part of the SoM basin presently constitutes a seismic gap. The northern margin is composed of the Palaeozoic shales, weakly cemented Tertiary sedimentary rocks, and overlaying wedge of Quaternary sediments on the shelf edge. The mechanically weak lithology of the margin, coupled with steep slopes (20°–26°) and the activity of the NAF, makes the SoM susceptible to submarine landslides. The mass-wasting was more common during times of sea level lowstands when brackish-water lacustrine conditions prevailed than under highstand marine periods. It is most likely that gas escape and gas hydrate dissociation also contributed to the mass-wasting in the SoM during the sea level lowstands. Historical records reveal that more than 30 tsunami events occurred in the past two millennia. It is more likely that most tsunamis in the SoM have been associated with submarine landslides triggered by large earthquakes. However, the normal faulting on the southern margin of the C¸ınarcık Basin might have also caused tsunamis. Successive development of hypoxic conditions extending to the shelf areas occurred every time the SoM was flooded by marine waters during the interglacial periods. The SoM has been affected by the geohazards in the past and will likely face them again in the future. Long-term multi-disciplinary seafloor observatories are needed to monitor the geohazards in real time in the SoM. Keywords Geohazard
Sea of Marmara
Earthquake
Submarine landslide
Tsunami
Hypoxia N. Go¨ru¨r
M. N. C¸ag˘atay (&) General Geology Section, Department of Geological Engineering, Eastern Mediterranean Centre of Oceanography and Limnology (EMCOL), Faculty of Mines, Istanbul Technical University, Maslak, 34469 Istanbul, Turkey e-mail: [email protected]

123

584

Nat Hazards (2010) 54:583–603

1 Introduction The Sea of Marmara (SoM) is located on the western extension of the North Anatolian Fault (NAF) that is the major continental transform boundary between the Eurasian and _ Anatolian plates (Fig. 1). After the 1999 Izmit (M 7.4) and Du¨zce (M 7.2) earthquakes in the eastern Marmara region, alarm of a future devastating earthquake has been given _ within the SoM, close to Istanbul. This warning has been based mainly on the fact that the NAF usually creates earthquakes with a westward progression, in which one event promoting the next (Tokso¨z et al. 1979; Stein et al. 1997; Armijo et al. 1999; Hubert-Ferrari et al. 2000; Parsons et al. 2000; Parsons 2004; Meade et al. 2002). The large earthquakes on the northern branch of the NAF in the SoM show ca. 250 years recurring period and 243 years have already passed since the last earthquake occurred on this branch in 1766 (S¸ engo¨r et al. 2004). Later marine geological and geophysical studies confirmed the appropriateness of the warning and showed that the fault system in the SoM is capable of _ producing large earthquakes (Le Pichon et al. 2001, 2003; Imren et al. 2001; Armijo et al. 2002, 2005). Coulomb stress transfer, dynamic finite element modelling and historical earthquake studies indicate that occurrence of such an earthquake (Marmara earthquake) may take place any time within the next 30 years or so from the 1999 onwards (Parsons et al. 2000; Parsons 2004; Oglesby et al. 2008). Although the expected Marmara earthquake represents itself a threat, it may also cause other geohazards, such as submarine mass movements (landslides and turbidity currents), volcanic eruptions, fluid venting and gas hydrate dissociation and tsunamis (e.g., Cochonat et al. 2007). These geohazards may pose a great threat to costal settlements and installations and offshore infra-structures, including platforms, tunnels, pipelines, communication and power cables and sub-sea installations. Therefore, the study of such phenomena is important in risk assessment and disaster management and mitigation. Particularly, the

Fig. 1 Active tectonics of the eastern Mediterranean showing the Sea of Marmara located on the North Anatolian Transform Fault zone

123

Nat Hazards (2010) 54:583–603

585

knowledge of their types, magnitudes, locations and possible impacts before the disaster are critical issues for better communication between authorities, industry and researchers, leading to effective disaster preparedness. Evaluation of the future geohazard risks in the SoM requires knowledge of the geological and morphological evolutions of its margins that are characterized by sea level changes, active faulting and associated sediment accumulation. It also requires identification of the records of past events and mapping of disaster-prone areas. In order to accumulate such knowledge, we studied the geological evolution of its northern margin with an emphasis on the geological processes that have created hazardous conditions throughout its evolution. The study was made by means of the stratigraphic and structural interpretations of closely spaced seismic reflection lines, chirp profiles and core data. This work revealed that owing to its recent origin and the geodynamic setting, the northern margin has been prone in the course of its evolution to various geohazards beside the earthquakes. The aim of this article is to characterize them, and thereby to contribute an effective assessment of the future geohazards in this basin.

2 Oceanography, bathymetry and morphotectonic features The SoM is located between the Aegean and Black seas to which it is connected with _ C¸anakkale (Dardanelles) and Istanbul (Bosporus) straits having sill depths of -65 and -35 m, respectively. It has a two-way flow system with upper waters of the Black Sea (salinity: 18%) and lower waters of the Mediterranean (salinity: 37%) origins with the pycnocline is at -20 m (Bes¸ iktepe et al. 1994). The northern margin of the SoM consists of an east–west trending narrow shelf (\20 km) and a steep slope (Fig. 2). The shelf extends southward to the shelf break at water depths of about -90 m. It is divided geomorphologically by the submarine extension _ of the Istanbul Strait (Paleo Bosporus) into two parts. The western part extends between the

Fig. 2 Morphotectonic map the Sea of Marmara showing the active faults, submarine landslides and the locations of chirp sub-bottom profiles (bathmetry after Le Pichon et al. 2001; faults modified after Armijo et al. 2002)

123

586

Nat Hazards (2010) 54:583–603

_ Istanbul Strait and Gaziko¨y (Tekirdag˘) near the confluence of the C¸anakkale Strait and is characterized mainly by two embayments, namely Silivri and Tekirdag˘ Bays. The maximum width of the western shelf is about 15 km over these embayments. The eastern shelf _ includes the Prince Islands and covers the area between the Istanbul Strait and the entrance _ of the Gulf of Izmit (Fig. 2). It is wider near the Bosporus (ca. 15 km) and becomes narrow towards Tuzla (ca. 7 km). The overall shelf dips gently towards the shelf break at about -90 m (C¸ag˘atay et al. 2009). Analysis of Chirp sub-bottom and multibeam bathymetric data shows that the shelf displays irregular bottom topography with some erosional, tectonic and constructional features (Eris¸ et al. 2007; C¸ag˘atay et al. 2009). The erosional features contain wave-cut notches, channels, amphitheatre-like landslide scarps forming submarine canyon heads, terraces, platforms and escarpments (Figs. 2, 3). Some of these features follow certain isobaths and represent paleoshorelines. Two paleoshorelines are found on the western shelf, off C¸ekmece. One is located at ca. -85 m and the other one is seen at -93 m (Figs. 4, 5; C¸ag˘atay et al. 2009). The -85 m paleoshoreline can also be traced on the eastern part of the northern shelf, between the Bu¨yu¨kada and Tuzla (Fig. 6). Here, it is observed as a 10-m high south-facing escarpment with a linear southeast trend (C¸ag˘atay et al. 2009). Tectonic features are found south of C¸ekmece, west of the Istanbul Strait and the south of Prince Islands shelf. They are represented by WNW–ESE trending scarps and steps on the seafloor and have fault origin (Fig. 2). On the seismic profiles, most of these faults are observed as strike-slip faults in part with dip-slip components (Fig. 7). In the south of Bu¨yu¨kada, a 6.1-km long and 12-m high fault scarp represents such a fault along which the Palaeozoic basement is uplifted (Figs. 6, 7; C¸ag˘atay et al. 2009). The ESE trend of these faults suggests that they are reactivated Hercinian structures (A.M.C. S¸engo¨r, pers. comm). Constructional features occur as thick sediment wedges on the shelf edge (Figs. 3, 5, 10, 12; see also Sect. 3).

Fig. 3 Chirp sub-bottom profile TOCIN-1 cutting across the palaeo-Bosporus and showing the Holocene mud drape, erosional marine terraces, channel fills, -85 m palaeoshoreline, basement structures and secondary extensional faults. Location of the profile is given in the Fig. 2

123

Nat Hazards (2010) 54:583–603

587

Fig. 4 Chirp sub-bottom profile Marm-10 showing the Holocene mud drape (Unit S1) and 1–2 m-high mound-like bioherms atop Unit S3 that is deposited above erosional unconformity (b1). For location of the profile see Fig. 2

Fig. 5 Sub-bottom Chirp profile MARM-18 showing the thickening wedge of late Quaternary sediments deposited near the shelf edge on the northern shelf off C ¸ ekmece (modified from C¸ag˘atay et al. 2009). These units extend back to at least Marine Isotope Stage 7 (MIS-7) (ca. 160 ka BP). For location of the profile see Fig. 2

The northern shelf passes into the basin slope at depths of -95 m (Fig. 2). The con_ tinental slope extends in an east–west direction between Gaziko¨y and the Istanbul Strait _ with a characteristic multi-cuspate shape. At the Istanbul Strait, it makes a sharp turn to the _ southeast to connect with the Gulf of Izmit. It ranges in width from 4 to 9 km and is delimited at the base by the deep northern trough of the SoM. This trough marks the seismically most active northern branch of the NAF zone and consists of three deep depressions (from east to west: C¸ınarcık, Central and Tekirdag˘) and two intervening highs (central and western highs) (Fig. 2; Wong et al. 1995; Rangin et al. 2001). Where the slope directly leads to the depressions, it is steep and has a slope angle more than 18°. The basin slope displays a large number of submarine canyons and landslides (Fig. 2). Although they are observed almost everywhere on the northern slope, the canyons seem to be more common on the slopes leading to the C¸ınarcık and the Central basins. Their length is mostly confined with the width of the basin slope and therefore varies from 4 to 9 km on

123

588

Nat Hazards (2010) 54:583–603

Fig. 6 Palaeo-bathymetric map of the shelf off the Prince Island, showing the -85 m palaeoshoreline, basement structures and secondary extensional faults below the Holocene mud drape (modified from C ¸ ag˘atay et al. 2009)

Fig. 7 Sub-bottom Chirp profile Pi-59 on the Prince Islands shelf, showing the various morphological features and structures, such as the -85 m paleoshoreline, basement uplift and secondary extensional faults. For location of the profile see Fig. 2

the basin margins and shorter on the intervening Central and Western highs. They are straight in most of their course but show branching towards the shelf edge. Landslides are common on the northern slope of the C¸ınarcık Depression, where slope angles vary between 20° and 29°. The largest one is located south of Tuzla, covering a total area of ¨ zeren et al. submitted). 32 km2 (Fig. 8; O

3 Stratigraphy 3.1 Basement The basement sequence of the SoM is defined as the rocks predating the formation of the SoM basin, which started around the late Middle Miocene (Serravallian) (S¸ engo¨r et al. 1985; Go¨ru¨r et al. 1997a). East of the old city walls of Istanbul, this sequence consists of

123

Nat Hazards (2010) 54:583–603

589

Fig. 8 Bathymetric map showing the submarine landlides on opposite slopes of the eastern part of the C ¸ ınarcık Basin (a). The one located in the north, south of Tuzla Peninsula, is the largest of two and is dated at ca. 17 ka BP using the chronostratigaphy of Core TSU-04 (b)

Ordovician to Carboniferous clastic and carbonate rocks unconformably overlain by Triassic redbeds and limestones, which in turn are succeeded unconformably by Upper Cretaceous to Lower Tertiary carbonates (Go¨ru¨r et al. 1997b). The Palaeozoic part of this succession represents a north-south striking, west-vergent segment of the Hercynian Orogen. The slope south of the Prince Islands consists mainly of Devonian shales, as observed during the Nautile dives of the Marnaut cruise (Marnaut Cruise Report 2007). West of the old city walls around C¸ekmece, the basement sequence consists as far as C¸atalca of Eocene reefal carbonates, Oligocene mudstone, sandstone, limestone, deltaic sandstones and conglomerates (Figs. 2, 9; Tu¨rkecan and Yurtsever 2002). Across the Bu¨yu¨k C¸ekmece Lagoon, the greenschists of the Mesozoic C¸atalca Massif forms the basement. West of C¸atalca, a large accretionary complex of the Upper CretaceousPaleocene age, together with the overlying Eocene-Oligocene siliciclastic sediments of the Thrace Basin, constitutes the basement (Go¨ru¨r and Okay 1996; Go¨ru¨r et al. 1997a, b; Turgut and Eseller 2000). 3.2 Margin sediments The margin sediments started depositing with the beginning of the formation of the SoM basin during the late Middle Miocene (Serravalian) (Go¨ru¨r et al. 1997a; C¸ag˘atay et al. 2006). On land, i.e. around C ¸ ekmece, the Middle Miocene rocks unconformably overlie the basement rocks and consist mainly of fluvial to lacustrine (Paratethyan) sandstone, mudstone, marl and Mactra-bearing limestones (Fig. 9; Sakınc¸ et al. 1999; Turgut and Eseller 2000). Further west, in the S¸ arko¨y area, they begin at the base with variegated coarse basal conglomerates and sandstones, passing upward into fine-grained and coal-bearing clastics and carbonates. Here, these rocks are succeeded by Upper Miocene sediments through a

123

590

Nat Hazards (2010) 54:583–603

Fig. 9 Generalized stratigraphic section between the C ¸ ekmece and the Tekirdag˘ coastal areas, west of Istanbul (compiled from Sakınc¸ et al. 1999 and Turgut and Eseller 2000)

thin sedimentary unit of Upper Serravallian age, comprising calcareous sandstones and laminated shales with various fossils of Mediterranean origin (Sakınc¸ et al. 1999). The Upper Miocene is characterized by oolitic and bioclastic limestones, rich in Mactra, gastropods and ostracods. The presence of abundant Mactra suggests a Paratethyan (The Black Sea) origin for these carbonates (Go¨ru¨r et al. 1997a). They are unconformably overlain by the Upper Pliocene conglomerates and pebbly sandstones in part with freshwater bivalves. The early Pliocene appears to be an erosional period during which the Miocene sediments were eroded in large part on the northern margin. The Upper Pliocene lithologies are succeeded by Pleistocene sediments across an angular unconformity (Go¨ru¨r et al. 1997a; Sakınc¸ et al. 1999). On the northern shelf, within the depth penetrated by the Chirp sub-bottom profiles and the cores, the Pleistocene to Holocene sediments are divided into seven units (Unit S1 to Unit S7 in descending order) that are bounded by unconformities (Fig. 10). They thicken

123

Nat Hazards (2010) 54:583–603

591

Fig. 10 Generalized seismic and core stratigraphy of the Upper Pleistocene to Holocene sediment units on the northern shelf of the Sea of Marmara

towards the shelf edge and wedge out to the north towards the shore (Figs. 3, 5, 7, 11). The Pleistocene units form transgressive and progradational packages with characteristic internal structures. Their sequence stratigraphic boundaries can be matched with global sealevel curve (Fig. 12; Shackleton and Opdyke 1989; C¸ag˘atay et al. 2009). Units S7 and S6 are encountered in the north-eastern shelf, south of the Prince Islands (Fig. 11). They are not seen in the western shelf probably due to the thick cover of the younger units and the limited penetration of the seismic lines. The Unit S7 could not be cored and therefore

123

592

Nat Hazards (2010) 54:583–603

Fig. 11 Sub-bottom Chirp profile TOCIN-3 on the Prince Island shelf, showing the various units deposited in the last ca. 150 ka. These units are separated by shelf-crossing unconformities, and together form a thickening wedge towards the shelf edge. Note the cores (MD-2745, PIC-40, TSU02-01) used for chronoand litho-stratigaphical correlations along the seismic section. For location of the profile see Fig. 2

Fig. 12 Timing of deposition of stratigraphic units on the northern shelf of the Sea of Marmara in relation to the global sealevel curve of Shackleton and Opdyke (1989). The seismic stratigraphic units (S1–S6) are delimited mainly by shelf crossing unconformities (b1–b4) and correlated with lithostratigraphic units (L1– L5). The sapropel layers are shown as MSAP2–MSAP4. The presence of the MSAP-5 is assumed on the basis of high sea level during marine isotope stage 5e

its lithology and age are not known. Chirp profiles indicate, however, that the Unit S7 is dipping below an irregular unconformity surface (Fig. 5). On the Prince Islands shelf, it rests on older sediment units above an angular unconformity (Fig. 11). The overlying Unit S6 forms a regressive package with clinoforms dipping southward, towards the shelf edge

123

Nat Hazards (2010) 54:583–603

593

(Figs. 11, 5). It accumulated probably during the Marine Isotope stage 6 (MIS-6) of the pleistocene as suggested its position below Unit S5 that was deposited during MIS-5 (Figs. 10, 12; C¸ag˘atay et al. 2009). In core C-17 on the C¸ekmece shelf, Unit S6 consists mainly of alternation of grey mud beds and brackish-water Dreissena banks with some other Neouxinian molluscs. The Unit S5 is encountered in the C¸ekmece and eastern shelf around the Prince Islands (Figs. 5, 11). It shows parallel bedding and fills depressions above a prominent unconformity surface (b4) on the C¸ekmece shelf (Fig. 5). Unit S5 is about 15 m thick in the outer shelf of Prince Islands and thins toward the inner shelf. In the Chirp sub-bottom profiles, it onlaps all the older sediments and displays a transparent internal configuration. Lithologically, it is represented by a green mud and laminated and cross-laminated silt in the lower part and two sapropel layers (Marmara Sapropels 4 and 3: MSAP-4 and MSAP-3) in the upper part. The two sapropels are separated by Dreissena-bearing silt and sand. The lower green mud is rich in fully marine foraminifera and rare euryhaline molluscs. On the basis of the seismic, core and sealevel data, the age of this unit is given as late Marine Isotope Stage 5 (MIS-5), spanning 78 and 129 ka (Figs. 10, 12; Eris¸ 2007; C¸ag˘atay et al. 2009). Units S4 and S3 are seen as prograding clinoforms on the shelf edge and parallel internal reflections on the inner shelf (Figs. 5, 11). The clinoforms are much thicker on the western shelf, off the Bu¨yu¨k and Ku¨c¸u¨k C¸ekmece lagoons (Fig. 5). Core data indicate that these units represent a regressive facies represented by a Dreissena-bearing sandy silt–mud of lacustrine origin (Fig. 10). These two units show variable thickness; their total thickness is more than 20 m on the C¸ekmece shelf and about 7 m on the Prince Islands shelf. The age of the Unit 4 is inferred to be MIS-4 (59–78 ka BP) (Figs. 10, 12; Eris¸ 2007; C¸ag˘atay et al. 2009). The age of the upper boundary of Unit 3 is well constrained by a radiocarbon date, whereas that of the lower boundary is based on the relationships between seismic stratigraphy, facies and global sealevel change (Fig. 12) (Eris¸ 2007; C¸ag˘atay et al. 2009). Unit S2 shows similarities in seismic stratigraphy to the underlying Units S3 and S4. It consists of two subunits (Fig. 10). It displays a progradational wedge with southward dipping clinoforms at the shelf edge (Fig. 5). It is about 4 m thick and shows strong internal reflections on the shelf (Figs. 5, 11). It is characterized mainly by Dreissenabearing grey mud, passing upward into well sorted and shelly fine to coarse-grained sands with a mixed euryhaline to brackish water bivalves. The age of the Unit 2 is 30–12 ka BP according to its stratigraphic position and the radiocarbon ages (Figs. 10, 12; Eris¸ 2007; C¸ag˘atay et al. 2009). Unit S1 is a ca. 2.5-m thick transgressive sequence, displaying transparent to parallel internal reflections in the Chirp profiles (Figs. 4, 5). The Unit 1 onlaps all the older units and the basement rocks with an unconformity. Radiocarbon dating assigns to this unit an age ranging from 12 ka BP to present (Figs. 10, 12; C¸ag˘atay et al. 2000). In the entrance of the Bosporus to the SoM, equivalent sediments of this unit appear to be a valley infill with channel, levee and associated delta deposits (Eris¸ et al. 2007). It comprises at the base dark grey to green pebbly sand, grading upward into green silty mud with abundant marine bivalves and foraminifera. These sediments are also associated with ca. 1.5 m thick bioherms and a sapropel layer (MSAP-2) on the C¸ekmece and the Prince Islands shelves, respectively (Figs. 4, 12; C¸ag˘atay et al. 2000, 2009). The MSAP-2 sapropel was deposited during 10.6–6.3 ka BP (C¸ag˘atay et al. 2000) and was characterized by a dark green and faintly laminated mud, containing up to 2.3 wt% Corg and a benthic foraminifer fauna of suboxic–dysoxic conditions. Another sapropelic sediment unit dated 4.7–3.2 ka BP is also observed over the shelf areas (MSAP-1) (C ¸ ag˘atay et al. 1999; Tolun et al. 2002).

123

594

Nat Hazards (2010) 54:583–603

4 Geological evolution and geohazard development The northern margin started its development in the mid-Miocene, some 14 to 10 Ma ago, when the North Anatolian Shear Zone originated and a variety of structures formed making up the shear zone (S¸ engo¨r et al. 1985; Go¨ru¨r et al. 1997a). These structures were NW–SE striking normal faults, WNW-striking right-lateral Riedel shears and NNW-striking leftlateral anti-Riedel shears. They seem to have combined to form a roughly E–W structure with dominant NW-striking extensional segments. Because of that the northern shelf has a multi-cuspate southerly margin. Seismic reflection profiling shows the northern shelf as a whole tilted to the north. Near Bu¨yu¨k C¸ekmece, the shelf edge is offset right-laterally for some 25 km (S¸ engo¨r et al. 2004). The northern shelf emerged during the glacials and was submerged during interglacials that greatly influenced its stratigraphic development. During most of the MIS-2, MIS-3, MIS-4 and MIS-6, it was lacustrine (Paratethyan) environment with fresh to brackish water salinities, suggesting that the Mediterranean Sea level dropped below the sill depth of the C¸anakkale Strait. The global glacio-eustatic sealevel fluctuations coupled with the activity of the NAF Zone have been the major control on the geohazard development on the northern margin. Anoxia, earthquakes, slope failures and tsunamis took place on this margin evidently throughout its geological history. Occurrence of another geohazard, namely gas hydrate melting, was inferred from the recent finding of gas hydrate and unusual water and gas expulsion activities along the trace of the active faults on the seafloor (Marnaut Cruise Report 2007). All these geohazards are briefly described below along with the geological evolution of the northern shelf. 4.1 Anoxia ¨ nlu¨ata et al. 1990). Presently the deep SoM waters contain 1–3 mg/l of dissolved oxygen (U However, in the past, there was successive development of suboxic/anoxic conditions even on the northern shelf as shallow as -70 m. This is evidenced by the presence of the four sapropelic layers (MSAP-4 to MSAP-1) in Units S5 and S1 of the Upper Pleistocene to Holocene sequence (Figs. 10, 12; C¸ag˘atay et al. 1999, 2000, 2009). The fossil content, transgressive nature, seismically transparent and onlapping characteristics of these units indicate that all the sapropel/sapropelic sediment units formed during the initial stages of sealevel highstands after the connection of the SoM with the Mediterranean Sea (Fig. 12). Invasion of the brackish SoM by the Mediterranean Sea water caused density stratification and increased organic productivity. These resulted in bottom water oxygen depletion, which in turn allowed the preservation of organic matter in the sediments. The presence of some benthic foraminifera in the MSAP-4 and MASP-2 indicates that the bottom water conditions on the northern shelf were suboxic to dysoxic during their development at 105–100 ka BP (MIS-5a) and 10.6–6.3 ka BP (MIS-1), respectively, whereas their absence in the MSAP-3 suggests that the bottom waters were anoxic during the 86–83 ka BP (MIS-5c) (C ¸ ag˘atay et al. 2009). Although not yet recovered in cores in the SoM, the MIS-5e sediments also are expected to include a sapropel layer (Fig. 12). The latest low-oxygen bottom water conditions over the SoM shelves occurred during 4.7–3.2 ka BP (MSAP-1) (C ¸ ag˘atay et al. 1999). 4.2 Earthquakes Being located on the NAF zone, the SoM is a tectonically very active basin. The NAF is a major transform plate boundary that has produced devastating historical earthquakes along

123

Nat Hazards (2010) 54:583–603

595

its 1500 km length (Guidoboni et al. 1994; Ambraseys and Finkel 1995; Ambraseys 2002). The most active northern branch of the NAF delimits the northern margin of the SoM from the south, generally following the toe of the northern slope. The earthquake events along the course of the NAF appear to have a westward progression with 60-year sequence of rupturing toward Istanbul, in which one event promoting the next (Stein et al. 1997; _ Parsons et al. 2000; Parsons 2004). With the last 1999 Izmit (M 7,4) and 1912 S¸arko¨y (M 7,4) earthquakes having occurred on the western and eastern ends of the SoM, a large part of this basin constitutes a seismic gap awaiting to be filled by the next earthquake events. There are many primary and secondary features on the sea-floor and in the sediments indicating that the NAF has been producing earthquakes of various magnitudes since it was originated in the SoM basin since the Pliocene. The earthquake-related primary features are structural in nature and formed due to rupture of the sea-floor. They include fault scarps, tilting, and offset of some submarine features, such as strata, palaeoshoreline, canyons, etc. The secondary earthquake-related features originated as a result of shaking and are represented mainly by mass wasting and gravitational mass flow deposits accumulated in the deep depressions of the SoM (McHugh et al. 2006; Sarı and C¸ag˘atay 2006; Beck et al. 2007). The earthquake scarps are in very good state of preservation. They are observed in the deep depressions (Armijo et al. 2005). Between the Tekirdag˘ and Central basins, the ruptures and the associated scarps appear to be continuous and fresh over 60 km. They are attributed to the eastward extension of the 1912 S¸arko¨y earthquake (Ms 7.4; Armijo et al. 2005). Two other ruptures are found along the northern and the southern bounding faults of the C¸ınarcık depression. The rupture in the north is a 20–30-km long fresh scarp correlated tentatively with the 18 October 1963 (Ms 6.4) earthquake. The one in the south is 50 km long and probably formed during the 10 July 1894 earthquake (Ms 7.0; Armijo et al. 2005). Large sea-floor rupturing earthquakes seem to have played an important role also in the filling of the deep depressions of the SoM. These basins are filled predominantly with intercalated sequences of turbidite–homogenite units characterized by a basal sand layer and an overlying mud layer (Beck et al. 2007). The sand-rich layers commonly have an erosional base and occur as stacked, millimetre to centimetre thick bands or laminae that thin and fine upward (Sarı and C¸ag˘atay 2006; C¸ag˘atay et al. 2008). These sequences are analogues to ‘‘homogenites’’ of the Mediterranean and can be directly linked to earthquakes and tsunamis (Beck et al. 2007). Recently, by studying the turbidite–homogenite units in the basins of the SoM, evidence for a number of large historical earthquakes was documented (McHugh et al. 2006; Sarı and C¸ag˘atay 2006; C¸ag˘atay et al. 2008). According to these studies, large earthquakes occurred in the SoM in 181 AD, 740 AD, 1063 AD, 1343 AD, 1509 AD, 1766 AD, 1894 AD and 1912 AD. 4.3 Submarine landslides Assessment of high-resolution bathymetric map of the SoM shows several scars and deposits of paleo-landslides originated from the steep slopes of the deep basins (Fig. 2). As described in Sect. 3, the shelf break and slopes of the SoM are susceptible to mass-wasting processes. The shelf break and upper slope areas were characterized by relatively high rate of sediment deposition during the Quaternary sealevel lowstands. Moreover, the northern margins of the deep basins are composed of clayey Palaeozoic rocks in the east and weakly cemented Tertiary siliciclastic rocks in the west. The Quaternary stratigraphy of the northern shelf reveals a number of depositional sequences bounded above and below by unconformities (Fig. 10). These unconformities

123

596

Nat Hazards (2010) 54:583–603

are the result of sealevel fall that caused the subaerial erosion of the exposed sediments deposited during the earlier depositional cycle. Considering the age of the shelf sediments, the sealevel fluctuations took place during the glacial–interglacial cycles and were therefore mainly eustatic in nature (Aksu et al. 1999; C¸ag˘atay et al. 2003, 2009). During the time of relatively low sealevel, the SoM became a fresh–brackish water lake and its shelves were largely exposed and eroded. This resulted in a high sediment flux to the shelf break, slope and the deep basins. The presence of the thick channel fills in Unit S5 off the Bu¨yu¨k C¸ekmece and Ku¨c¸u¨k C¸ekmece lagoons and the palaeo valley that incised the shelf from the _ Istanbul Strait to the shelf break and upper slope (palaeo-Bosporus) may be cited as evidence for this interpretation (Figs. 3, 5). These sediments cannot be derived from the land in the north, because the drainage area of the northern shelf is very small and no important streams flow from here to the SoM. Although it has been proposed that the Maritsa River in the Balkans was once flowing into the western SoM and carrying sediments here through the Ergene River, it was diverted into the Aegean Sea before the early Quaternary (Okay and Okay 2002). A large number of submarine canyons on the basin slope indicate that most of the sediments were transported into the C¸ınarcık, Central and Tekirdag˘ basins through these canyons. The sediment produced during sealevel lowstands on the northern shelf were probably first deposited on the outer shelf, shelf edge and the basin slope, and then transferred into the deep depressions of the SOM by slope failure and other mass-wasting processes. This is suggested by the existence of a large number of submarine canyons on the northern slope. The slope failures may have been triggered by sediment loading, earthquakes and gas escape. The NAF has been active in the area since the Pliocene (Go¨ru¨r et al. 1997a; Yaltırak 2002; S¸ engo¨r et al. 2004) and most probably played an important role in the slope failures. The deposition in response to glacial–interglacial variability formed on the northern margin sedimentary successions susceptible to slope failures. As described above, the late Pleistocene successions of this margin consist mainly of muds alternating with silt and sand sediment units, which are rich in smectite. Furthermore, the basement rocks forming the slope south of the Prince Islands are mainly the Devonian shales rich in the swelling illite–smectite mixed layer clays. The clayey sediments and sedimentary rocks on the shelf margin and slopes of the SoM were more susceptible to slope failure during times of lowstands, when fresh to brackish water lacustrine conditions prevailed, than under highstand marine conditions. Such lowstand periods might have also easily generated unfavourable shear stress and fluid escapes from the sediments, thus causing slope failure (Bryn et al. 2002, 2003). Indeed, the deep basin sedimentary records of the SoM indicate that the slope failure processes in the form of submarine landslides, turbidites and homogenites were more common during the last lowstand lacustrine period prior to 12 ka BP than during the Holocene highstand marine period (Beck et al. 2007). For example, the ca. 15-m thick homogenite deposition in the Central Basin (Beck et al. 2007) and the large landslide on the NE slope of the C¸ınarcık Basin south of Tuzla (Fig. 2; ¨ zeren et al. submitted) occurred during late glacial period. O There are several paleo- and potential-landslide areas on the northern margin of the SOM, ranging in depth from -100 m to -1200 m (Fig. 2). Their sizes vary in area from ca. 10 to 32.5 km2 and in volume from ca. 0.4 to 1.6 km3. The large landslide south of Tuzla lies at depths between -250 and -1200 m and has a distinctive triangular shape with 7 km width and ca. 50 m thickness (Fig. 8a). It includes Palaeozoic basement shale of Devonian age, which is rich in the swelling illite–smectite mixed-layer clay. The landslide covers a total area of 32, 5 km2 and includes a slid mass of ca. 1.62 km3. The 14C analysis of a core sample and the well-dated MSAP-2 sapropel above the chaotically mixed

123

Nat Hazards (2010) 54:583–603

597

landslide surface suggest that it occurred ca. 17 14C ka BP when the SOM was lacustrine ¨ zeren et al. submitted). with the water level between -85 and -95 m (Fig. 8b; O 4.4 Tsunamis Tsunamis usually occur in connection with large earthquakes that are associated with faulting with a vertical displacement of the seafloor. They also result from submarine landslides. Considering the fact that the predominant faulting style along the NAF is strikeslip, it is more likely that most tsunamis in the SoM have been associated with landslides on the steep slopes as result of strong earthquake shaking. However, the normal faulting on the southern margin of the C¸ınarcık Basin might have also caused tsunamis. Numerous paleo-landslides on the steep slopes of the SoM suggest that tsunamis associated to these submarine landslides also took place. The shelf break and upper slope areas have been the depositional sites for unstable sediments that could have repeatedly slumped down the slopes during large earthquakes, creating damaging tsunamis. For example, modelling studies indicate that the slid mass south of the Tuzla peninsula ¨ zeren et al. (Figs. 2, 8) could have had created a tsunami with a wave height of ca. 10 m (O submitted). The most recent tsunami event was associated with the Go¨lcu¨k earthquake of August 1999, when a wave with up to 2.5 m run-up inundated coastal areas in the central and _ eastern parts of the Gulf of Izmit (Altınok et al. 2001; Alpar 1999, Yalc¸ıner et al. 1999, ¨ ztu¨rk et al. 2000; Tappin 2000; Tappin et al. 2002). The source of the tsunami was 2002; O on the south side of the Gulf between Go¨lcu¨k and Karamu¨rsel. Although the main trace of the NAF runs through this area, the presence of several significant seabed slumps here suggests that the tsunami was not solely due to seabed faulting, but sediment slumping was also involved (Yalc¸ıner et al. 1999, 2002). This interpretation is also supported by the widespread coastal subsidence along the southern shore of the Gulf during the earthquake. The 1999 Go¨lcu¨k tsunami is not exceptional for this region. Historical records reveal that more than 30 tsunami events occurred in the past two millennia, with heights up to about 6 m in the coastal areas (Papazachos et al. 1986; Ambraseys and Finkel 1990, 1991, 1995; Guidoboni et al. 1994; Altınok et al. 2003; Demirkent 2001; Ozansoy 2001; Yalc¸ıner et al. 2002). In the Byzantine period, the following tsunami events are mentioned in connection with major earthquakes: 358 Izmit, 447 Istanbul, 478 Istanbul and Izmit, 554 Istanbul and Izmit, 989 Izmit, 1332 Istanbul, 1343 Istanbul, 1344 Western Marmara and 1419 Izmit. The Ottoman record includes 1509 Istanbul, 1754 Gulf of Izmit, 1766 Istanbul, 1878 Izmit and 1894 Istanbul earthquakes and associated tsunamis after the conquest of Istanbul. 4.5 Fluid venting and gas hydrate dissociation Detailed marine geological and geophysical surveys carried out using the ROV Victor 6000 (Armijo et al. 2005; Zitter et al. 2008), manned submersible Nautile (Marnaut Cruise Report 2007; Ge´li et al. 2008) and RV L’Atalante revealed the widespread occurrence of fluid venting along the submerged segments of the NAF in the SoM (Fig. 13). During the Marnaut cruise gas hydrate was also discovered close to the seafloor on the Western High at depth of -660 m. The occurrence of this substance at such a shallow depth can be explained by its unusual composition. It consists of 82% methane and 18% ethane, propane and butane. Its hydrocarbon and and C-isotope compositions indicate that the gases are of thermogenic origin and similar to the natural gas from the Tertiary Thrace Basin that partly extend in to the SoM.

123

598

Nat Hazards (2010) 54:583–603

Fig. 13 Map showing the locations of the Nautile submersible dives where cold seeps are observed (red circles) and where no such seeps are present (white circles) along the active fault trace (Ge´li et al. 2008). Note the location of the first gas hydrate discovery at 660 m water depth on the Western High and the mantle helium on the west of Tekirdag˘ Basin. The red circle south of C¸ekmece, where thermal methane emission is observed, is not on the fault, but on an anticline a few km south of the fault. The fault segment south of Silivri-C ¸ ekmece has not ruptured since 1766 AD, and is characterized by a very low microseismicity and fluid activity

The presence of gas hydrates, together with sea level fluctuations, suggests that gas hydrate dissociation and related mass-flows (e.g. Sultan et al. 2004; Nixon and Grozic 2006) might have occurred especially during the low water level (i.e. lacustrine) periods of the SoM. Changes in the water levels of the SoM during the glacial–interglacial period might have caused changes in the sea-bottom pressures and temperatures, which in turn might have affected gas hydrate stability. Gas venting and gas hydrate dissociation from fault plain at the base of slope may also occur during an earthquake cycle and destabilize the slope sediments. Unravelling the triggering mechanism (earthquake, gas hydrate dissociation or fluid venting) of individual landslides requires detailed geological and geophysical studies, and long term monitoring and observations, involving in situ measurements of pore pressure, strain and tilt.

5 Future threats From the preceding discussions, it is clear that the SoM was the source of some severe geological hazards, including anoxia, earthquakes, mass flows, tsunamis and possibly sudden fluid venting and gas hydrate dissociation. The areas that have been affected by these hazards in the past are likely to face them again in the future. Therefore, the communities in the Marmara region are exposed to risk from these natural hazards. As a matter of fact soon after the 1999 earthquakes, an earthquake threat to Istanbul has been claimed by the earth scientists in Turkey. This warning has been given on the basis of the stress transfer mechanism, because during the twentieth century the NAF was ruptured by a series of earthquakes from east to west as a result of this mechanism (Stein et al 1997). It is

123

Nat Hazards (2010) 54:583–603

599

now believed that the sequence will continue with a future earthquake located in the SOM. Historical earthquake data suggest that this marine realm is the only place along the NAF system that did not experience a large earthquake during the twentieth century. It thus presents an exceptionally high earthquake risk (Ambraseys and Jackson 2000; HubertFerrari et al. 2000; Parsons et al. 2000). Multichannel seismic and bathymetric studies reveal that an earthquake with a magnitude of more than 7 may strike the region (Le Pichon et al. 2001; 2003; Rangin et al. 2001, 2004; Imren et al. 2001; Armijo et al. 2002; 2005; Demirbag˘ et al. 2003; Carton et al. 2007). Stress transfer analysis indicates that it will take place within 30 years with a probability of 62% (Parsons et al. 2000). Our recent marine and submarine studies indicate that the steep submarine scarp of the northern margin of the SoM has a potential to create a large-scale slope failure. There are various processes in the SoM to trigger the submarine slope to fail suddenly. Earthquakes are likely triggers as are the gas hydrate dissociation and sudden release of gases. Methane outflow in the water column associated with slope instabilities has been observed in the Gulf of Izmit after the 1999 earthquake (Alpar 1999; Kus¸ cu et al. 2002, 2005). This has in turn caused anoxic bottom water conditions (Balkıs 2003), probably caused by anaerobic methane oxidation. A huge slope failure could generate a tsunami that could propagate in any direction depending on the location of the source, path of propagation and near shore morphology, thus forming a risk to the entire coastal area. Geohazards are of significant interest from many aspects. They represent major threats for offshore infrastructures (platforms, pipelines, cables and sub-sea installations) and onshore facilities (e.g. Tappin et al. 2001; Longva et al. 2003; Sultan et al. 2004). A critically important perspective on these risks is forecasting the likelihood of future events and their threat to populated areas. In order to achieve this task a comprehensive and pragmatic approach is needed. Although a substantial amount of geological and geophysical information on the fault system of the SoM has been gathered since the 1999 earthquakes, further site studies involving high resolution bathymetric mapping, seismic reflection and coring are needed for assessing the future geohazards in this basin. Long-term, real-time seafloor observations combining seismology with strain, pore pressure, and fluid flux and composition measurements are necessary for monitoring purposes. Such future experimental studies are planned under the ESONET NoE (European Seafloor Observatory Network of Excellence) Marmara-Demonstration Mission project.

6 Conclusions Being located on a transform fault boundary between the Eurasian and Anatolian plates, the SoM is tectonically very active with fast deformation rates (25 mm/a horizontal and 5–6 mm/a vertical), measurable on annual time scales. Morphotectonics of its northern margin with steep slopes (18°–29°) are controlled mainly by reactivated Hercynian structures. It is therefore prone to earthquakes, submarine lanslides and associated tsunamis. The next major (M [ 7) earthquake(s) is expected within the SoM. The submarine mapping and coring studies indicate that submarine mass movements were especially common during low sealevel periods, which were also the lacustrine periods of the SoM. Anoxic–suboxic events occurred during marine transgressions, following marine connection with the Mediterranean Sea. These events affected the shelf areas to depths of -70 m. _ The semi-enclosed narrow gulfs and bays of the SoM, such as the Izmit and Gemlik gulfs, are prone to bottom-water anoxia during the earthquakes because of the anaerobic

123

600

Nat Hazards (2010) 54:583–603

oxidation of methane emitted as a result of crustal deformation. Permanent, real-time seafloor observations combining seismology with strain, pore pressure, and fluid flux and composition measurements are necessary for monitoring purposes. Acknowledgements The marine geological and geophysical data discussed in this study were obtained during the various cruises of R/Vs Le Suroit, L’Atalante, Marion Dufresne, Urania, Sismik-1 and Odin Finder. We would like to thank the captain and crew of these RVs. We also thank scientists Xavier Le Pichon, Celal S¸ engo¨r, Rolando Armijo, Luca Gasperini, Alina Polonia, Pierre Henry, Louis Geli, Leonardo Seeber, and Milene Cormier, Cecilia McHugh who took part in these cruises and contributed immensely to our present knowledge about the Sea of Marmara. Thanks are also extended to the Turkish Academy of Sciences (TUBA) for supporting the studies of N. Go¨ru¨r.

References Aksu AE, Hiscott RN, Yasar D (1999) Oscillating Quaternary water levels of the Marmara Sea and vigorous outflow into the Aegean Sea from the Marmara-Black Sea drainage corridor. Mar Geol 153:275–302 Alpar B (1999) Underwater signatures of the Kocaeli Earthquake (August 17th 1999). Turkish J Mar Sci 5:111–129 Altınok Y, Tinti S, Alpar B, Yalc¸ıner AC, Ersoy S, Bortolucci E, Armigliato A (2001) The tsunami of August 17, 1999 in Izmit Bay, Turkey. Nat Hazards 24:133–146 Altınok Y, Alpar B, Yaltırak C (2003) Sarko¨y-Murefte 1912 earthquake’s tsunami: extension of the associated faulting in the Marmara Sea, Turkey. J Seismol 7:329–346 Ambraseys NN (2002) The seismic activity of the Marmara Sea Region over the last 2000 years. Bull Seismol Soc Am 92:1–18 Ambraseys NN, Finkel CF (1990) The Marmara Sea earthquake of 1509. Terra Nova 2:167–174 Ambraseys NN, Finkel CF (1991) Long term seismicity of Istanbul and the Marmara Sea region. Terra Nova 3:527–529 Ambraseys NN, Finkel CF (1995) The seismicity of Turkey and adjacent areas: a historical review. Eren, Istanbul, pp 1500–1800 Ambraseys NN, Jackson JA (2000) Seismicity of the Sea of Marmara (Turkey) since 1500. Geophys J Int 141:F1–F6 Armijo R, Meyer B, Hubert A, Barka A (1999) Eastward propagation of the North Anatolian Fault into the northern Aegean: timing and kinematics. Geology 27:267–270 Armijo R, Meyer B, Navarro S, King G, Barka A (2002) Asymmetric slip partitioning in the Sea of Marmara pull-apart: a clue to propagation processes of the North Anatolian Fault? Terra Nova 14:80–86 Armijo R, Pondard N, Meyer B, Mercier de Lapinay B, Ucarkus G, MARMARASCARPS Cruise Party (2005) Submarine fault scarps in the Sea of Marmara pull-apart (North Anatolian Fault): implications for seismic hazard in Istanbul. Geochem Geophys Geosyst 6:1–29 _ _ Balkıs N (2003) The effect of Marmara (Izmit) Earthquake on the chemical ceanography of Izmit Bay, Turkey. Mar Pollut Bull 46:865–878 Beck C, Mercier de Lapinay B, Schneider JL, Cremer M, Cagatay N, Wendenbaum E, Boutareaoud S, Menot G, Schmidt S, Webe O, Eris K, Armijo R, Meyer B, Pondard N, Gutcher MA, Turon JL, Labeyrie L, Cortijo E, Gallet Y, Bouquerel H, Gorur N, Geravis A, Castera MH, Londeix L, de Resseguier A, Jaouen A (2007) Late Quaternary co-seismic sedimentation in the Sea of Marmara’s deep basins. Sediment Geol 199:65–89 ¨ zsoy E, Abdul Latif M, Og˘uz T, U _ O ¨ nlu¨ata U ¨ (1994) The circulation and hydrography Bes¸ iktepe S¸ T, Sur IH, of the Marmara Sea. Prog Oceanogr 34:285–334 Bryn P, Berg K, Lien R, Solheim A, Ottensen D, Rise L (2002) The Storegga Geomodel and its use in slide risk evaluation: geological and geotechnical site investigations in the Storegga Slide Area, in Offshore Site 151. Investigation and Geotechnics: Diversity and Sustainability, London, UK Bryn P, Solheim A, Berg K, Lien R, Forsberg CF, Haflidason H, Ottesen D, Rise L (2003) The storegga slide complex; repeated large scale sliding in response to climatic cyclicity. In: Locat J, Mienert J (eds) Submarine mass movements and their consequences. Kluwer, Dordrecht, pp 215–222 C ¸ ag˘atay MN, Algan O, Sakınc¸ M, Eastoe CJ, Egesel L, Balkis N, Ongan D, Caner H (1999) A mid-late Holocene sapropelic sediment unit from the southern Marmara shelf and its palaeoceanographic significance. Quat Sci Rev 18:531–540

123

Nat Hazards (2010) 54:583–603

601

C ¸ ag˘atay MN, Go¨ru¨r N, Algan A, Eastoe CJ, Tchapalyga A, Ongan D, Kuhn T, Kus¸ cu I_ (2000) Late GlacialHolocene palaeoceanography of the Sea of Marmara: timing of connections with the Mediterranean and the Black Sea. Mar Geol 167:191–206 ¨, C ¸ ag˘atay MN, Go¨ru¨r N, Polonia A, Demirbag˘ E, Sakınc¸ M, Cormier MH, Capotondi L, McHugh C, Emre O Eris¸ K (2003) Sealevel changes and depositional environments in the Izmit Gulf, eastern Marmara Sea, during the late glacial-Holocene period. Mar Geol 202:159–173 C ¸ ag˘atay MN, Go¨ru¨r N, Flecker R, Sakınc¸ M, Tu¨nog˘lu T, Ellam R, Krijgsman W, Vincent S, Dikbas¸ A (2006) Paratethyan–Mediterraenan connectivity in the Sea of Marmara region (NW Turkey) during the Messinian. Sediment Geol 188–189:171–188 C ¸ ag˘atay MN, Belucci L, Polonia L, Sancar U, Dikc¸e D, Eris¸ K, Damci E, Gasperini L, Go¨ru¨r N, Henry P, Zitter TAC, Geli L, Tryon M (2008) Sedimentary earthquake records in the Sea of Marmara. EGU General Assembly 2008, Vienna, EGU2008-A-05412 C ¸ ag˘atay MN, Eris¸ K, Sancar U, Polonia A, Akc¸er S, Biltekin D, Gasperini L, Ryan WBF, Gorur N, Bard E, Lericolais G (2009) Late pleistocene-holocene evolution of the Northern shelf of the Marmara Sea. Mar Geol 265:87–100 Carton H, Singh SC, Hirn A, Bazin S, de Voogd B, Vigner A, Ricolleau A, Cetin S, Ocakoglu N, Karakoc F, Sevilgen V (2007) Seismic imaging of the three-dimensional architecture of the Cinarcık Basin along the North Anatolian Fault. J Geophys Res. doi:10.1029/2006JB004548 Cochonat P, Masson D, Armigliato A, Bornhold B, Camerlenghi A, C ¸ ag˘atay N, Favali P, Kvalstad T, Kopf A, Lykousis V, Miranda JM, Person R, Esclasans RU (2007) History, monitoring and prediction of geohazards. In: Cochonat P et al (eds) Deep Sea Frontier. Science challenges for a sustainable future. Directorate general for research Environment/Management Natural Resources publication EUR22812EN, Luxembourg, pp 9–15 Demirbag˘ E, Rangin C, Le Pichon X, S¸engo¨r AMC (2003) Investigation of the tectonics of the Main Marmara Fault by means of deep towed seismic data. Tectonophysics 361:1–19 _ Demirkent I (2001) Bizans kaynaklarına go¨re IV-XI yu¨zyıllarda Istanbul ve c¸evresinde depremler. In: S¸ ahin I (ed) Tarih boyunca Anadoluda dog˘al afetler ve deprem semineri 22–23 Mayıs 2000. Globus Dunya _ Basım Evi, Istanbul, pp 51–65 Eris¸ KK (2007) Marmara Denizi’nde Orta Pleyistosen’den gu¨nu¨mu¨ze deniz seviyesi deg˘is¸ imleri, Doktora _ _ ¨ niversitesi, Avrasya Yerbilimleri Enstitu¨su¨, Iklim Tezi, Istanbul Teknik U ve Deniz Bilimleri Anabilim Dalı, Yer Sistemi Bilimi Programı, 208 s Eris¸ KK, Ryan WBF, C ¸ ag˘atay MN, Sancar U, Lericolais G, Me´not G, Bard E (2007) The timing and _ evolution of the post-glacial transgression across the Sea of Marmara shelf south of Istanbul. Mar Geol 243:57–76 Ge´li L, Henry P, Zitter T, Dupre´ S, Tryon M, C ¸ agatay MN, Mercier de Le´pinay B, Le Pichon X, Sengo¨r ¨ zeren S, Volker D, Gasperini L, Burnard P, Bourlange S, AMC, Go¨ru¨r N, Natalin B, Uc¸arkus G, O Marnaut Scientific Party (2008) Gas emissions and active tectonics within the submerged section of the North Anatolian Fault zone in the Sea of Marmara. Earth Planet Sci Lett 274(1–2):34–39 Go¨ru¨r N, Okay AI (1996) Origin of the Thrace Basin, NW Turkey. Geol Rundschau 85:662–668 Go¨ru¨r N, C ¸ ag˘atay MN, Sakınc¸ M, Su¨mengen M, S¸entu¨rk K, Yaltırak C, Tchapalyga A (1997a) Origin of the Sea of Marmara as deduced from Neogene to Quaternary palaeogeographic evolution of its frame. Int Geol Rev 39:342–352 Go¨ru¨r N, Monod O, Okay AI, S¸engo¨r AMC, Tu¨ysu¨z O, Yig˘itbas¸ E, Sakınc¸ M, Akko¨k R (1997b) Palaeogeographic and tectonic position of the carboniferous rocks of the Western Pontides (Turkey) in the frame of the variscan belt. Bull Soc Geol France 168:197–205 Guidoboni E, Comastri A, Trania G (1994) Catalogue of ancient earthquakes in the Mediterranean area up to 10th century. Instituto Nazionale di Geofisica, Rome Hubert-Ferrari A, Barka A, Jacques E, Nalbant S, Meyer B, Armijo R, Tapponier P, King GCP (2000) Seismic hazard in the Marmara Sea following the 17 August 1999 Izmit earthquake. Nature 404:269–272 _ Imren C, Le Pichon X, Rangin C, Demirbag˘ E, Ecevitoglu B, Gorur N (2001) The North Anatolian Fault within the Sea of marmara: a new interpretation based on multichannel seismic and multibeam bathymetry data. Earth Planet Sci Lett 186:143–158 _ Okamura M, Matsuoka H, Awata Y (2002) Active faults in the Gulf of Izmit on the North Kus¸ cu I, Anatolian Fault, NW Turkey: a high-resolution shallow, seismic study. Mar Geol 190:421–443 _ Okamura M, Matsuoka H, Go¨kas¸ an E, Awata Y, Tur H, S¸ ims¸ ek M, Kec¸er M (2005) Seafloor gas Kus¸ cu I, seeps and sediment failures triggered by the August 17, 1999 earthquake in the eastern part of the Gulf of Izmit, Sea of Marmara, NW Turkey. Mar Geol 215:193–214 _ Le Pichon X, S¸ engo¨r AMC, Demirbag˘ E, Rangin C, Imren C, Armijo R, Go¨ru¨r N, C ¸ ag˘atay N, Mercier de Lepinay B, Meyer B, Saatc¸ılar R, Tok B (2001) The active main Marmara fault. Earth Planet Sci Lett 192:595–616

123

602

Nat Hazards (2010) 54:583–603

Le Pichon X, Chamot-Rooke N, Rangin C, S¸ engo¨r AMC (2003) The North Anatolian Fault in the Sea of Marmara. J Geophys Res. doi:10.1029/2002JB001862 Longva O, Janbu N, Blikra LH, Bøe R (2003) The 1996 Finneidfjord slide: seafloor failure and slide dynamics. In: Locat J, Mienert J (eds) Submarine mass movements and their consequences. Kluwer, Dordrecht, pp 531–538 Marnaut Cruise Report (2007) 13/05/2007-11/06/2007 Istanbul (Turkey), 70 pp McHugh CMG, Leonardo S, Cormier MH, Dutton J, Cagatay MN, Polonia A, Ryan WBF, Gorur N (2006) Submarine earthquake geology along the North Anatolian Fault in the Marmara Sea, Turkey: a model for transform basin sedimentation. Earth Planet Sci Lett 248:661–684 Meade BJ, Hager BH, McClusky S, Reilinger S, Ergintav S, Lenk O, Barka A, Ozener H (2002) Estimates of seismic potential in the Marmara Sea region from block models of secular deformation constrained by Global Positioning System measurements. Bull Seismol Soc Am 92:208–215 Nixon MF, Grozic JLH (2006) A simple model for submarine slope stability analysis with gas hydrates. Nor J Geol 86:309–316 Oglesby DD, Mai PM, Atakan K, Pucci S (2008) Dynamic models of earthquakes on the North Anatolian fault zone under the Sea of Marmara: effect of hypocenter location. Geophys Res Lett. doi:10.1029/ 2008GL035037 Okay N, Okay AI (2002) Tectonically induced Quaternary drainage diversion in the northeastern Aegean. J Geol Soc Lond 159:393–399 Ozansoy E (2001) Bizans kayıtlarına gore 1200–1453 Istanbul depremleri. In: S¸ ahin I (ed) Tarih boyunca _ Anadoluda dog˘al afetler ve deprem semineri, 22–23 Mayıs 2000. Globus Dunya Basım Evi, Istanbul, pp 1–29 ¨ zeren MS, C¸ag˘atay MN, Postacioglu N, S¸engo¨r AMC, Go¨ru¨r N, Eris¸ K (submitted) A late glacial subO marine landslide in the Sea of Marmara and the mathematical modeling of the associated tsunami. Geo Mar Lett ¨ ztu¨rk H, Koral H, Geist EL (2000) Intra-basinal water movements induced by faulting: the August 17, O 1999 Go¨lcu¨k (Image Izmit Bay) earthquake (Mw = 7.4). Mar Geol 170:263–270 Papazachos CB, Koutitas CH, Hatzidimitriou MP, Karacostas GB, Papaioannou ACH (1986) Tsunami hazard in Greece and the surrounding area. Ann Geophys 4:79–90 Parsons T (2004) Recalculated probability of M [ 7 earthquakes beneath the Sea of Marmara. J Geophys Res. doi:10.1029/2003JB002667 Parsons T, Toda S, Stein RS, Barka A, Dieterich JH (2000) Heightened odds of large earthquakes near Istanbul: an interaction based probability calculation. Science 288:661–665 Rangin C, Demirbag E, Imren C, Crusson A, Normand A, Le Drezen E, Le Bot A (2001) Marine Atlas of the Sea of Marmara (Turkey). Ifremer, Brest Rangin C, Le Pichon X, Demirbag E, Imren C (2004) Strain localization in the Sea of Marmara: propagation of the North Anatolian Fault in a now inactive pull-apart. Tectonics. doi:10.1029/2002TC001437 Sakınc¸ M, Yaltırak C, Oktay FY (1999) Palaeogeographical evolution of the Thrace Neogene basin and the Tethyan-Paratethyan relations at northwest Turkey (Thrace). Palaeogeogr Palaeoclimatol Palaeoecol 153:17–40 Sarı E, C ¸ ag˘atay MN (2006) Turbidites and their association with past earthquakes in the deep C ¸ ınarcık Basin of the Marmara Sea. Geo Mar Lett 26:69–76 S¸ engo¨r AMC, Go¨ru¨r N, S¸ arog˘lu F (1985) Strike-slip faulting and related basin formation in zones of tectonic escape; Turkey as a case study. Society of economic paleontologists and minerologists, Spec. Pub. 37. Research symposium, strike-slip deformation, basin formation and sedimentation, San Antonio, TX, pp 227–264 _ S¸ engo¨r AMC, Tu¨ysu¨z O, Imren C, Sakinc¸ M, Eyidog˘an H, Go¨ru¨r N, Le Pichon X, Rangin C (2004) The North Anatolian Fault: a new look. Annu Rev Earth Planet Sci 33:1–75 Shackleton NJ, Opdyke NG (1989) Oxygen isotope and palaeomagnetic stratigraphy of equatorial Pacific core V28–238: oxygen isotope temperatures and ice volumes on a 105 year and 106 year scale. Quat Res 3:39–55 Stein RS, Barka A, Dieterich JH (1997) Progressive failure on the North Anatolian fault since 1939 by earthquake stress triggering. Geophys J Int 128:594–604 Sultan N, Cochonat P, Canals M, Cattaneo A, Dennielou B, Haflidason H, Laberg JS, Long D, Mienert J, Trincardi F, Urgeles R, Vorren TO, Wilson C (2004) Triggering mechanisms of slope instability processes and sediment failures on continental margins: a geotechnical approach. Mar Geol 213: 291–321 Tappin DR (2000) Tsunamis in the Sea of Marmara (project proposal). British Geol Survey, 28 pp Tappin DR, Watts P, McMurty GM, Lafoy Y, Matsumoto T (2001) The Sissano, Papua New Guinea tsunami of July 1998: an offshore evidence of the source mechanism. Mar Geol 175:1–23

123

Nat Hazards (2010) 54:583–603

603

Tappin DR, Go¨ru¨r N, Watts P, Musson RMW, Evans R (2002) The tsunami threat in the Sea of Marmara, Turkey: a review. In: Go¨ru¨r N et al (eds) Integration of earth sciences research on the 1999 Turkish and Greek earthquakes. NATO Science Series. Kluwer, Dordrecht, pp 87–100 Tokso¨z MN, Shakal AF, Michael AJ (1979) Space–time migration of earthquakes along the North Anatolian Fault zone and seismic gaps. Pure Appl Geophys 117:1258–1270 Tolun L, C¸ag˘atay MN, Carrigan WJ (2002) Organic geochemistry and origin of late glacial-holocene sapropelic layers and associated sediments in Marmara Sea. Mar Geol 190:163–174 Turgut S, Eseller G (2000) Sequence stratigraphy and depositional history in eastern Thrace basin. Mar Petrol Geol 17:61–100 _ Tu¨rkecan A, Yurtsever A (2002) Geological map of Turkey, Istanbul Sheet, 1/500, 000 scale. MTA General Directorate, Ankara ¨ zsoy E (1990) On the physical oceanography of the Turkish Straits. In: Pratt ¨ nlu¨ata U ¨ , Og˘uz T, Latif MA, O U LJ (ed) The physical oceanography of sea straits. NATO/ASI Series. Kluwer, Dordrecht, pp 25–60 Wong HK, Lu¨dman T, Ulug A, Go¨ru¨r N (1995) The Sea of Marmara: a plate boundary sea in an escape tectonic regime. Tectonophysics 244:231–250 Yalc¸ıner AC, Synolakis CE, Borrero J, Altınok Y, Watts P, Imamura F, Kuran U, Ersoy S¸ , Kanog˘lu U, Tinti S (1999) Tsunami generation in Izmit Bay by the Izmit earthquake. Proceedings of ITU-IAHS Intern Conference on the Kocaeli Earthquake, 17 August 1999, pp 217–212 Yalc¸ıner AC, Alpar B, Altınok Y, Ozbay I, Imamura F (2002) Tsunamis in the Sea of Marmara: historical documents for the past, models for future. Mar Geol 190:445–463 Yaltırak C (2002) Tectonic evolution of the Marmara Sea and its surroundings. Mar Geol 190:493–529 Zitter TAC, Henry P, Aloisi G, G Delaygue, C ¸ ag˘atay MN, B Mercier de Lepinay, Al-Samir M, Fornacciari F, Tesmer M, Pekdeger A, Wallmann K, Lericolais G (2008) Cold seeps along the main Marmara fault in the Sea of Marmara (Turkey). Deep Sea Res I 55:552–570

123