2003 Seattle Annual Meeting (November 2–5, 2003)

Paper No. 2
Presentation Time: 8:15 AM


MULLER, Jordan R. and AYDIN, Atilla, Geological and Environmental Sciences, Stanford Univ, Stanford, CA 94305, mullerj@pangea.stanford.edu

Since 1939 the North Anatolian fault (NAF) in Turkey has produced twelve M>6.7 earthquakes with epicenter locations that have progressed predominantly westward. These events have combined to rupture over 1100 km of the ~1500 km long NAF. The most hazardous seismic gap along the NAF is now recognized as being within the Marmara Sea with Istanbul, a city of over 12 million inhabitants, lying on its northern shore. In order to determine the fault structure within the Marmara Sea, geophysical surveys over the past six years have provided high-resolution multi-beam bathymetry and a variety of seismic reflection data sets. These data reveal a complex pull-apart basin fault geometry composed of multiple orientations of strike-slip and oblique-slip faults. Despite the increased quantity of data, the definitive 3-D fault geometry within the Marmara Sea remains elusive. As does the issue of which faults within that geometry may be considered susceptible to a future earthquake.

This paper presents a method for characterization of 3-D fault geometry within the Marmara Sea from observed displacements on the seafloor or stratigraphic horizons. Our method, based on linear elastic fracture mechanics, utilizes three-dimensional boundary element modeling software (Poly3D). In our model, faults within a specified 3-D geometry are allowed to slip in response to regional stress loading conditions for the Marmara Sea region. Then, fault slip distributions and subsequent displacements on grid surfaces are computed. These grid surface displacements are compared to observed Marmara Sea bathymetry as well as the deformed pre-transform/syn-transform stratigraphic horizons mapped from 2-D seismic reflection profiles. Results to date for the western Marmara Sea suggest that the fault geometry in the Tekirdag basin region requires lateral bounding faults as well as the large east-west trending faults to the north and south to produce the basin geometry observed in the seismic and bathymetry data. This predicted geometry is in contrast to the simpler single fault hypothesis put forth recently by several authors for the western Marmara Sea.