Search

We investigate by means of a 3-D geomechanical model the relationship between structural elements and contemporary kinematics in the Marmara Sea region, northwest Turkey. The recently imaged fault system beneath the Marmara Sea is incorporated into the model as frictional surfaces with varying strike and dip. The Main Marmara Fault is implemented as through-going and is accompanied by mostly non-vertical second-order faults. Topography, basement-topography and the Moho become mechanically effective through changes in density and elastic parameters across these horizons. The model is subjected to gravity and kinematic boundary conditions. The ultimate goal of this study is to set up a 3-D model that is consistent with both, kinematic observations and stress data. The stress results are presented in a complementary paper. In this paper we present the modelled long-term 3-D kinematics in terms of fault slip rates, rotations, vertical motion and sense of fault slip. The model results agree with Global Positioning System velocities, geological fault slip rates, palaeomagnetic measurements and with the observed pattern of subsidence and uplift. Furthermore, our tectonically driven vertical velocities can be linked to landscape and basin evolution and to features of sedimentation. Our results indicate that the Main Marmara Fault can be interpreted as a through-going fault that slips almost purely in a strike-slip sense. Nevertheless, and not contradictory to the previous statement, there is significant dip-slip motion at some sections of the Main Marmara Fault. The agreement of the modelled 3-D kinematics with model-independent observations supports that the main structural details of the fault system are accounted for. Sensitivity analysis of model parameters reveals that changes in rock properties and the initial stress state have minor influence on the 3-D kinematics. We conclude that the 3-D structure of the fault system is the key control of the kinematics. The slip rate of the Main Marmara Fault from our model is lower than previous estimates and shows high variability along strike (12.8-17.8 mm a). The latter indicates that stress accumulation is non-uniform along strike. [ABSTRACT FROM AUTHOR]

We present the contemporary 3-D background stress field of the Marmara Sea region derived from a geomechanical model. The background stress field (i.e. the component of the absolute stress state that is time-independent over the seismic cycle) primarily depends on the distribution of density and elastic parameters, on the acting far-field stresses from plate boundary forces and on fault geometries. We take these into account to predict the 3-D background stress field including its spatial variations. Technically, our model concept involves the definition of an appropriate initial stress state accounting for the gravitational reference stress state of the crust, which is then changed by plate tectonics until the 3-D background stress field has evolved. The modelled stress field agrees well with observations from earthquake focal mechanism solutions and their formal stress inversion, with orientation of maximum horizontal stress and with the distribution of seismicity in the Marmara Sea. In particular in the vicinity of fault bends the stress field deviates considerably from the regional NW-SE oriented maximum horizontal stress and exhibits variability of the stress regime. Our model results are consistent not only with dynamic observations but also with kinematic ones. Various kinematic observations are understandable from the stress field. We show that a stress regime that indicates normal faulting in the basins is nonetheless reconcilable with almost pure strike-slip motion on the Main Marmara Fault. The distribution of seismicity in the Marmara Sea can be explained in first order by the distribution of critical differential stress, which is closely related to local fault geometries. We refer the wide absence of seismicity between the bend of the Main Marmara Fault near Istanbul and the Central basin to the relatively plane fault geometry of that segment. Normal stress on the Main Marmara Fault is highly variable along strike, which makes segment-wise rupture more likely than a rupture of the seismic gap at once. Comparably low normal stress on the Prince's Islands Segment indicates a shorter interevent time for this fault segment than for the central segment of the Main Marmara Fault. [ABSTRACT FROM AUTHOR]