We combine available GPS data in and around the Aegean region to model plate boundary deformation and earthquake cycle effects in the observed velocity field. Typically, GPS data in the region have been used as evidence that a southern Aegean block behaves coherently and rigidly (e.g. Nyst and Thatcher, 2004; Reilinger et al., 2006). These first-order models match the observed data quite well, suggesting little intra-plate or plate-boundary deformation. There has been little suggestion that the GPS data record anything other than rigid plate motion. However, the M8.4 (Shaw et al., 2008) Crete earthquake of AD 365 (Figure 2.4) and some other historic earthquakes likely occurred along the Hellenic subduction zone (Burton et al., 2004), which implies that a substantial portion of the subduction thrust may be locked. Depending on the locking required to generate earthquakes of this magnitude, a measureable elastic strain signal reaching far into the overriding plate should be evident in the surface velocity field. Alternatively, slow accumulation of elastic strain and dominantly aseismic creep on the subduction thrust may generate a signal different from the pattern commonly observed along subduction zones. It is also possible that the surface velocities generated by convergence at a locked subduction zone are masked by simultaneous back-arc extension, creating the illusion of rigid block motion. We consider multiple possibilities in an attempt to interpret the current geodetic signal in the region and its implications for earthquake hazard assessment. We use a block modeling approach that considers both rigid rotations of plates and elastic strain fields along plate boundary faults to examine the possible trade-off between these components. As many of the observations are located away from the plate boundaries in question, it is difficult to constrain boundary parameters, such as locking depth and dip, using only the GPS data. We generate multiple models to explore the solution space of all reasonable parameters. Our modeling suggests that it is possible for coeval extension and convergence to occur at opposite ends of a plate masquerading as rigid block motion. Eventually, precisely determined vertical motions of GPS stations above the Hellenic subduction zone are needed to resolve this important question.
Our primary data in this study is a solution of 166 stations, concentrated primarily in Western Europe, processed by Enrico Serpelloni. Processing details can be found in (Serpelloni et al., 2007). Selected, globally distributed IGS sites were used to define an ITRF00 reference frame.
In addition to our own analysis, we integrate GPS-station velocities from published work in Africa, Central Greece, and Turkey. We integrate these published solutions (Stamps et al., 2008; Reilinger et al., 2006; Hollenstein et al., 2008; Clarke et al., 1998) with our own solution. These velocities were combined with our solutions by rotating them into a common reference frame. We combine velocities published in an ITFR00 reference frame into our own solution by minimizing the misfit at co-located stations. After the combination, we compare our combined solution the published values to estimate misfit. For most sites the RMS is 1-2 mm/yr, well within the 95% confidence intervals for these sites.
Plate boundary locations are critical for characterizing GPS velocities and the plate kinematics of a particular region. While some plate boundaries in the Aegean region are well defined by active fault traces, youthful geomorphology, and abundant local seismicity, others appear more diffuse and ambiguous. We draw on the distribution and kinematics of 20th century seismicity, local geology, and mapped faults, and the GPS velocity itself to define our block model boundaries. Within this paper the term plate (and microplate) refers to the rigid, coherent, lithospheric entity defined by faults, seismicity etc. The term block is the specific implementation of these data into a parameterized set of variables within our block model (e.g. Apel et al., 2006).
We define our blocks as rigid entities on a spherical earth bounded by dislocations in an elastic halfspace and invert for poles and rates of rotation that minimize the misfit to the GPS velocities using the block modeling code by Meade and Hager (2005). Because our inversion combines rigid block rotation with elastic strain accumulation effects, the parameterization of the block boundary geometry is critical. Geometry of the block boundaries is based heavily on seismicity and adopted from prior analyses (Nyst and Thatcher 2004; Reilinger et al., 2006) or adjusted as indicated by the geodetic data.
Some block motions are well defined and vary little within our model. The Eurasian block and Nubian block rotation parameters are defined primarily by the sites that lie within the stable interior and are affected very little by plate boundary strain. The inferred motion of smaller blocks (Aegean and Anatolian) can change based the parameterization of the boundaries of these blocks. The stability of the major blocks provides robust constraints on far-field motions and allows us to test variable segment geometries along those boundaries used develop our preferred model.
Figure 2.5 shows observed and predicted velocities through a model of simultaneous extension and contraction within the Aegean block. We model the subduction zone with variable locking dips which generate a series of non-unique fits to our data. It remains unclear whether or not the rigid block model with no elastic boundaries actually fits the horizontal GPS data better than models that include elastic block boundaries. At present, our models suggest that it possible to accumulate some amount of elastic strain along the subduction zone. Our continuing research includes incorporating vertical uplift rates in models to fully capture and constrain any amount of elastic strain that is accumulating along the Hellenic subduction zone.
Seismicity of the Aegean region for depths less than 100km. Arrows show relative block motions in a fixed Nubian reference frame.
Velocity profile of measured (circles) and predicted velocities (dashed lines) through the Aegean region. The inset figure shows the block configuration used in the inversion.
Apel, E. V., R. Bürgmann, G. Steblov, N. Vasilenko, R. King, and A. Prytkov, Independent active microplate tectonics of northeast Asia from GPS velocities and block modeling, Geophysical Research Letters, 33, L11303, doi:10.1029/2006GL026077, 2006.
Burton, P., X. Yebang, C. Qin, G. Tselentis, and E. Sokos, A catalogue of seismicity in Greece and the adjacent areas for the twentieth century, Tectonophysics, Vol 390, pg. 117–127, 2004.
Clarke, P. et al., Crustal strain in central Greece from repeated GPS measurements in the interval 1989-1997, Geophysical Journal International, Vol 135, pg. 195-214, 1998.
Hollenstein, Ch., M. Müller, A. Geiger, and H.-G. Kahle, Crustal motion and deformation in Greece from a decade of GPS measurements, 1993–2003, Tectonophysics, Vol 449, pg. 17–40, 2008.
Meade, B. J., and B. H. Hager, Block models of crustal motion in southern California constrained by GPS measurements, Journal of Geophysical Research, 110, B03403, 10.1029/2004JB003209, 2005.
Nyst, M., and W. Thatcher, New constraints on the active tectonic deformation of the Aegean, Journal of Geophysical Research, 109, B11406, doi:10.1029/2003JB002830, 2004.
Reilinger R., et al., GPS constraints on continental deformation in the Africa-Arabia-Eurasia continental collision zone and implications for the dynamics of plate interactions, Journal of Geophysical Research, 111, B05411, doi:10.1029/2005JB004051, 2006.
Shaw B., et al., Eastern Mediterranean tectonics and tsunami hazard inferred from the AD 365 earthquake, Nature - Geoscience, Vol 1, doi:10.1038/ngeo151, 2008.
Stamps, S., E. Calais, E. Saria, C. Hartnady, J.-M. Nocquet, C. Ebinger, and R. Fernandes, A kinematic model for the East African Rift, Geophysical Research Letters, 35, L05304, doi:10.1029/2007GL032781, 2008.