by N.Seeber and C.McHugh
Multibeam bathymetry, multichannel seismic profiling, high-resolution CHIRP, gravity coring, and remote operated vehicle (ROV) observations were collected by the R/V Urania in the Marmara Sea during May and June 2001 to address several current issues in tectonics and seismogenesis along the North Anatolian transform in northwestern Turkey. This fast-moving (22 mm/y) young continental plate boundary has produced many destructive earthquakes recorded in the long histoy of northwestern Turkey. The latest examples are the pair of M7+ earthquakes in 1999 that ruptured several segments of the plate boundary over a distance of 160 km. These ruptures are the last in a series of related earthquakes that have been progressing from east to west along the transform. The Marmara Sea segment is expected to rupture next in this progression. The likelyhood and nature of such an earthquake are critical for the future of the city of Istanbul on the northern shore of the Marmara Sea. Furthermore, this plate boundary offers unique opportunities for understanding processes that link past to future earthquakes.
The tectonic observations will be interpreted within the framework of the paleocanographic changes undergone by the Marmara Sea in response to glacioeustacy. These changes that have been recorded in its sediments and the position of the shorelines and delta complexes as they migrate across the margin include documenting when the Marmara Sea was in a lake stage without connections with the Mediterranean and Black Seas and in locating possible outlets of discharge between these systems. Other localized environmental conditions such sources of drainage, that can be quite different from one shoreline to another of the Marmara Sea, will also be evaulated.
The 2001 R/V Urania cruise will address these tectonic and paleoceanographic topics by focusing on young strata of Holocene to Pleistocene age that have recorded the most recent deformational and paleoclimatic changes. For this purpose fault structure and kinematics are imaged at unprecedented resolution and the strata will be dated to the greatest precision by radioisotopic methods for correlation of tectonic motions with paloecanographi factors. This will permit to establish a datum with which to address the stability of the tectonic regime; possible variations in tectonic conditions and Holocene fault slip rates. The drive for high resolution is pushed to new limits in the first attempt at resolving individual earthquakes in the geologic record by high density geophysical imaging, coring, and dating. Events adjacent to a fault scarp and across the basin will be identified from visual and x-ray observations in the core record, dated and correlated to known historical earthquakes. The ultimate goal is to resolve time, location, and amount of slip of single submarine fault ruptures. In the western mostly submerged portion of the North Anatolian plate boundary, new high-resolution submarine data are essential to reconstruct the evolution of the fault system through the late Quaternary and the seismogenic history through several earthquake cycles.
The first leg (May 28-June 7) of the R/V Urania 2001 cruise was primarily dedicated to acquiring geophysical data in five key areas in preparation for coring during the second leg scheduled from June 7 ?18. The results from each site are briefly discussed in the order they were visited by the Urania.
Ganos fault. The Ganos segment of the North Anatolian transform ruptured in 1912 between the Gulf of Saros and the western Marmara sea. The trace of the fault was imaged on the shelf of the Marmara just offshore the town of Gazikoy, with multibeam bathymetry and CHIRP surveys. A furrow and two small basins clearly delineate the fault on the shelf. These basins have been designated coring targets for paleoseismology in the second leg. This area coincides with a sharp (15°) bend in the strike of he North Anatolian transform. Secondary structures suggest a sharp transition from transpression to transtension from south to north across the fault in the area of the survey and, generally, from the ENE-striking Ganos segment to the EW-striking Marmara segment. This spatial differentiation in the tectonic regime may correspond with a transition in time for a block that translates around the bend. Accordingly, active normal faults and rapid subsidence characterize the block north of the fault where terraces, folds, and thrust faults attest to former uplift and shortening. The nature of the transition from transpression to transtension is relevant to both the formation of the basins in the Marmara as well as to the segmentation of the transform in discrete segments and earthquake ruptures.
Izmit Gulf. Most of the plate motion (20 mm/y) is taken up by the northern branch of the transform which is confined within the Izmit Gulf. The 1999 rupture entered the eastern end of the gulf and continued westward through all of it, according to SAR results. The R/V Urania's first leg mapped the central basin and completed bathymetric mapping and CHIRPprofiling of the western basin initiated by the 2000 R/V Odin Finder cruise. In addition, the R/V Urania covered both basins with a high-resolution CHIRP survey. The chirp imaged a progressively tilted sequence of turbidites in the central basin so clearly that individual layers could be correlated by their thickness between adjacent profiles. The latest turbidite is the only horizontal one; it is 0.5m thick and covers about 2 km2 in the deepest portion of the central basin. This layer may have been derived from the 1999 rupture. Such a thick deposit might be possible in a single event because the turbidite area is small compared to the submarine cachment area of the basin and because the rupture traverses the length of the basin. In any case, turbidite sequences in the central basin and in other small basins along the gulf east of the Hersek Promontory are very likely to be triggered by earthquakes and to contain a long paleoseismic record. A series of unconformities and paleoshorelines suggest that the paleolevel of the water in the basin has been repeatedly controlled by the sill north of Hersek Peninsula, delimiting the central basin to the west. Progressively older shorelines can be associated with a stack of three uncorformities. These horizontal paleoreference levels have implications for paleoceanography, paleoclimate, and as reference horizons of tectonic deformation. Sampling these surface is another high priority in the Urania 2001 second Marmara leg.
Cinarcik Basin. The very steep northeast wall of the Cinarcik Basin terminates downward into a characteristic apron that rises from the basin with relatively gentle slope. Chirp profiles show this apron to consist of basin sediments tilted toward the basin. Panels of sediments progressively steeper away from the basin (toward NE) are separated by sharp kinks. Tilting begins below the basin floor and older and deeper turbidites are therefore progressively more tilted within a section - a striking example of growth kinking. When the panel is raised above the basin floor, sedimentation is only pelagic. Thus older turbidites are close to the surface on the apron and they can be traced deep in the section below the active basin. This section offers the opportunity of sampling a long continuous sequence of turbidites by a set of cores through overlapping progressively older sections from the basin to the apron. The turbidites are likely to have paleoseismic origin. Their timing, if it can be determined from the cores, can be compared with paleoseimic results from around the Marmara Sea and with historic data. The ages and geometry of the turbidites can be modeled for a (minimum) tectonic subsidence rate of the basin.
South Boundary Fault. The South Boundary Fault feeds into the main branch of the transform in Ismit Gulf and may be a main contributor to the plate motion. The Urania mapped a key 4x8 km area east of Imrali Island where the fault takes a broad bend from ENE to WNW, east to west. A complex system of down-to-the-north normal faults was clearly imaged by chirp profiles and multibeam bathymetry. Six fault scarps were sampled for paleoseismic purposes. The morphology of the fault system suggests normal faulting is dominant over the right-lateral component. It seems unlikely that this fault system takes on a major portion of the transform motion.
Central Marmara Fault. The Main Marmara Fault is marked by sea-floor scarps and seismicity across the western Marmara Sea (Imren et al., 2001). In the saddle between the central Marmara High and the northern shelf, the fault is clearly expressed by two parallel narrow furrows separated about 200 meters y a pressure ridge. This morphology suggests a compressional step over. Urania mapped this saddle with chirp profiles and multibeam bathymetry and than sampled the fault zone with two cores in each strand. A core was also retrieved from the top of the Central Marmara antiform.
Coring Operations
Initial groundtruthing of the geophysical data was obtained by gravity coring in the South Boundary and Central Marmara Faults in the Imrali and Büyükçekmece Areas, respectively. Coring in the South Boundary Fault took place June 5 2001 from 11:10 UTM to 17:18 and June 6 2001 from 05:08 UTM to 6:25:18 in water depths ranging from 41 to 298 meters (Table 1). Eleven gravity cores ranging in length from .20 to 4.11 m were recovered with a 1.2 Tons weight. Core recovery ranged from 46 to 80.34 The main objective of the paleocanographic coring in the Central Marmara Fault was to penetrate two highly reflective surfaces shown on the CHIRP records. These surfaces are presumed unconformities related to relative sea-level changes. They form scarps that are considered to be paleoshorelines. The deeper reflective surface is about four meters below the sea-bed and the second one is at the sea bed. The coring objective was successful in that we recovered dreissena shells of Neuxinian fauna affinity at the base of cores MARM 01-IM02; MARM 01-IM02 BIS and MARM01-IM05. These sediments are thought to represent the lake phase and last regressive paleoshoreline. These lacustrine sediments underlying the Holocene mud drape are older than 12, 000 years B.P. This datum was penetrated at 3.5 meters below the sea-bed. The sediment is composed of dark gray to black, organic-rich sandy silty muds.
There were two attempts to penetrate the deeper seismic reflection interpreted as older than the last glacial maxiumum (Oxygen Isotope Stage 2 (OIS)). The first was not successful and we recovered a 2-cm long fragment of carbonate cemented shell fragments of euryhaline origin. In the second attempt, using a trigger core weight we recovered 20 cm of coarse sand, carbonate cements and shells. The cements are coated with serpulid tube-worms. None of these two attempts penetrated the deeper target.
The Central Marmara Fault was sampled in the Büyükçekmece Cekmece Area on June 6 2001 from 11:50 UTM to 17:24 in water depths ranging from 328 to 396.7 m. Six, gravity cores ranging in length from 3.0 to 4.16 meters were recovered with a head weight of 1.2 Tons. Core recovery ranged from 47.6 to 67.1