Quaternary evolution of the Corinth Rift and its implications for the Late Cenozoic evolution of the Aegean

SUMMARY We present geological and morphological observations at different scales to constrain rates of faulting and the distribution of deformation in the seismically active Aegean region. We focus first on the 130 km long Corinth Rift, an asymmetric graben where a flight of terraces of marine origin are uplifted. We show that the edges of the terraces lie in the footwall of the normal fault bounding the Corinth Rift and correspond to sea-level highstands of laic Pleistocene age. Using a detailed analysis of aerial and SPOT imagery supported by field observations, we have mapped 10 terrace platforms and strandlines ranging in elevation from 10 to 400 m over distances of 2 to 20 km from the fault. The elevation of the terraces' inner edges was estimated at 172 sites with an error of ±5m. This data set contains a precise description of the uplift and flexure of 10 different palaeohorizontal lines with respect to the present sea level. To date the deformation, we correlate the Corinth terraces with late Pleistocene oxygen-isotope stages of high sea-level stands and with global sea-level fluctuations. Using a thick elastic plate model consistent with our current understanding of the earthquake cycle and a boundary-element technique we reproduce the geometry of the shorelines to constrain both mechanical parameters and the slip on the fault. We show that the seismogenic layer behaves over the long term as if its elastic modulus were reduced by a factor of about 1000. All the terraces are fitted for fault slip increasing in proportion to terrace age, and the component of regional uplift is found to be less than 0.3 mm yr−1. The best fits give a slip rate of 11±3 mm yr−1 on the main rift-bounding fault over the last 350 kyr. Other geological and morphologic information allows us to estimate the total age of the main fault (∼1 Ma) and to examine the mechanical evolution of the Corinth Rift. The minimum observed sediment thickness in the Gulf places an extreme check on the results of the modelling and a lower bound on slip rate of 6–7 mm yr−1 (40 per cent less than estimated with modelling). Even this slip rate is nearly 10 times higher than for comparable features in most of the Aegean and elsewhere in the world. At a larger scale, the spacing and asymmetry of the rift systems in the Aegean suggest strain localization in the upper mantle, with slow extension starting 15 Myr ago or earlier. The more recent (1 Myr), rapid phase of rifting in Corinth partly reactivated this earlier phase of extension. The younger faulting in Corinth appears to result from its present location in the inhomogeneous stress field (process zone) of the south-westward propagating tip of the southern branch of the North Anatolian Fault. We extend these relations to propose a mechanical model for the Late Cenozoic evolution of the Aegean. As the Arabia/Europe collision progressed in eastern Turkey it caused Anatolia to move to the west and the North Anatolian Fault to propagate into the Aegean, where the early slow extension started to be modified about 5 Ma ago. The process of propagation dramatically increased the activity of some but not all of the earlier rifts. The model we present is compatible with tectonic observations, as well as with the seismicity, the palaeomagnetic rotations and the displacement field now observed with GPS and SLR.