Kinematics of the Marmara Region: a fusion of continuum and block models

Özbey V., Özeren M. S. , Henry P., Klein E., Galgana G., Karabulut H., ...More

Mediterranean Geoscience Reviews, vol.3, pp.57-78, 2021 (Peer-Reviewed Journal)


The Sea of Marmara has been the focus of numerous marine geophysical studies during the last couple of decades. Determination of the interseismic slip distribution along the Main Marmara Fault (MMF) has observational difficulties due to the fact that the fault is under the sea. In addition to the onshore geodetic studies, two underwater geodetic studies were carried out recently and brought new clues about the locked/creeping status of two different fault segments. In this study, we investigate the interseismic deformation of the Marmara region from a kinematic perspective using both block and continuum kinematic models. Block model inversion yields relative motion along the block boundaries as well as the distribution of slip deficit on the faults bounding the blocks. Calculated long-term slip rates are 19–20 mm/yr for the MMF and 3–4 mm/yr for the Southern branch of the North Anatolian Fault (SNAF). In order to determine the sensitivity of the slip deficit
solution to arbitrary spatial discretizations, we perform a series of checkerboard tests and optimize node distribution on the MMF accordingly. We also explore the sensitivity of inversion results on the MMF to assumptions regarding SNAF long-term slip and interseismic locking. This study shows that the underwater geodetic data can be combined with the GPS data in a joint inversion, bringing a better-constrained slip deficit distribution on the MMF. MMF has aseismic creep in the west of the Marmara Sea, with creep rates approaching long-term slip rates in the Central Basin. On the other hand, the fault appears locked from ∼ 28.2°E eastward. In the best fitting model, the fault segment across Silivri is locked down to ∼ 10 km. The existence of a hazardous seismic gap in the Sea of Marmara, potentially yielding a magnitude 7.5 earthquake, is thus confirmed. The continuum approach is used to calculate strain rate fields from GPS data alone, from the output (slip rates and block average strain rates) of the best-fitting block model, and from the GPS residuals of this model. Our continuum models show that most of the accumulated strain in the Marmara region is indeed focused around the MMF, but also that strain distributed within the blocks cannot be neglected. On the other hand, the SNAF system differs from the MMF in that it appears as a 50 km wide zone of deformation with a patchy distribution of strain reflecting the activity of individual faults.