The 1500 km-long North Anatolian Fault runs across northern Turkey and represents a major transform plate boundary. Near Istanbul (population: 12 millions), it accommodates 23 mm/yr (7.5 feet per century) of right-lateral motion between the Eurasian and Anatolian plates. While the North Anatolian fault is narrowly defined for most of its length, it is splitting into several branches in northwest Turkey (Figures 1 and 2). Cumulative slip on two of these fault branches is responsible for the formation of the Marmara Sea. The northern branch is the most active and bisects the entire Marmara Sea from east to west, passing only 20 km south of Istanbul. Bends in that northern branch results in the formation of a series of ~1 km-deep basins that are both rapidly deepening and accumulating sediments. In contrast, the Central Branch consists of many strands distributed across the shallower southern Marmara shelf. Most of these strands slip at slower rates than the Northern Branch, and are associated with basins that are subsiding more slowly, and that are therefore not as deep. This project aimed to understand the past history of motion on these multiple fault strands, as well as the relations of their motion to basin formation, sedimentation, and earthquake hazards. Indeed, similarly distributed transform fault systems are present in other densely populated parts of the world, including the San Andreas Fault system in southern California, the El Pilar fault system in northern Venezuela, and the fault system that crosses Haiti and the Dominican Republic. Understanding the spatial and temporal behavior of multi-stranded transform fault systems is thus highly relevant to seismic hazard assessment.

As a joint project with scientists at Dokuz Eylül University in Izmir (Turkey), some of the US investigators for this project participated in a research cruise to the southern Marmara Sea aboard the Turkish research vessel K. PIRI REIS. This field expedition was funded by the Scientific and Technological Research Council of Turkey, known as TÜBTITAK. Despite equipment problems and bad weather (including a gale), thanks to the professional expertise of the Turkish team, geophysical data of excellent quality were acquired along more than 1,600 km of shp track (see Figure 2). New, high-resolution bathymetric data reveal subtle fault scarps at the seafloor, in areas where others have questioned whether any faults were seismically active. The bathymetric data also reveal pockmarks and mounds as large as a few 100 m across, features typically associated with the venting of fluids and gas from the sediments. In particular, one pockmark field extends for at least 10 km along a fault scarp. A sub-bottom profiler (chirp) sonar imaged geological structures as deep as 30 m below the seafloor, with a sub-meter resolution. These chirp data highlight the vertical offsets of sediment layers across faults, informing about their geological history. They also image an erosion surface that formed around the Marmara some 14,000 years ago when its level dropped to 87 m below present sea level, as well as some "gas flares" produced by the venting of bubbles from pockmarks and mounds (see Figure 3). Lastly, multichannel seismic (MCS) reflection data have imaged geological structures as deep as 2 km below the seafloor. The acquisition and processing of MCS data are particularly effort-intensive, but allow for reconstructing the evolution of a fault system back into deep geological time than chirp sonar data, as far back as 1 million years in this case.

Thanks to these new data, a series of geological strata ("seismic horizons") identified from MCS data acquired previously north of the study area can be confidently correlated to seismic horizons in the southern Marmara Sea (see Figure 4). This provides a solid time-frame for interpreting the history of...

Please report errors in award information by writing to: awardsearch@nsf.gov.