NEW S-WAVE VELOCITY MODEL AND ANISOTROPY MEASUREMENTS FOR THE UPPER MANTLE BENEATH THE AEGEAN AND ANATOLIA: IMAGES OF A VERY COMPLEX SUBDUCTION SYSTEM

The Aegean-Anatolia region is a key area to investigate the relations between surface expressions of a long-lived and still active deformation zone and the dynamics of the underlying mantle. However, high-resolution data on the upper mantle structure remain scarce, in particular beneath Anatolia. To better constrain the geometry at depth of the Hellenic – Cyprus subduction and image the structure of the upper mantle beneath Western and Central Anatolia, we analyzed records of ~150 broadband permanent and temporary seismic stations compiled in the database of the SIMBAAD project. The result is an unprecedented 3-D S-wave velocity model of the crust and upper mantle to 300 km depth in the area [35-42°N; 20-39°E]. The Moho depth model obtained from ambient noise tomography is used to constrain the inversion of phase velocity measurements between 40-200s for upper mantle velocity. The upper mantle beneath Anatolia has low S wave velocities in the 50-200 km depth range, whereas most of the Aegean mantle has higher S wave velocities related to the Hellenic slab. Subduction-like high velocity anomalies underlay the low velocity anomaly of the Anatolian upper mantle. As they appear in the vicinity of the Cyprus trench in the 80-km shallowest depth slice of our mantle velocity model, we interpret these high velocity anomalies as the traces of the Cyprus subduction. It is characterized by weaker anomalies of more intricate shape than the Hellenic slab beneath the Aegean. The Hellenic-Cyprus slab is split into 2 parts by a ~200-250-km wide low-velocity anomaly located beneath southwestern Anatolia which we interpret as a vertical (along-dip) tear. We also measured seismic anisotropy from SKS splitting at the ~150 stations of our database. The location of the low velocity anomaly interpreted as a slab tear coincides with a sudden change in the orientation of the fast polarization direction of the split SKS phases. Fast split SKS waves have a fairly homogeneous N35 polarization in western and central Anatolia, except in the area of mantle low velocity anomaly in southwest Anatolia where the polarization turns to N145. We believe that this rotation documents a change in mantle flow associated with slab tear. Our new mantle model supplies a precise framework for future modeling of the contribution of mantle circulation to surface tectonics.