Koukouvelas, I.

We performed palaeoseismological investigations along the Aigion Fault, one of the main faults that bound the southern side of the Corinth Gulf. The mapped trace of the Aigion Fault onland is about 8 km long and may extend as much as 14 km if one includes its offshore trace. We made detailed studies at two sites adjacent to the Meganitas River. Although dating of faulted sediments was a bit problematic, we present a preliminary estimate of the faults short-term slip rate and recurrence interval. Slip rates range from 1.6 to 4.3 mmyr−1, with a maximum up to 6.3 mmyr−1. Three surface faulting events occurred in the seven centuries prior to 1888 AD, yielding an average (maximum) recurrence interval of 360 yr.

The positions and dimensions of fault segments within the Gulf of Corinth fault system have been identified by analysing spatial variation in fault displacements and fault kinematics. Growth of these fault segments is assessed by comparing their geometry and kinematics with the geometry and kinematics of the three sets of earthquake surface ruptures that are known to have affected the area in the last -200 years. Areas along the Gulf of Corinth fault system exhibiting low fault displacement (tens of metres) are identified as persistent segment boundaries which separate the fault system into a number of fault segments characterised by displacements which achieve maxima of -3 km. Fault-slip directions defined by lineations on fault planes vary systematicalIy with fault displacement, showing a converging pattem towards the hanging-walls of the fault segments: the fault-slip directions change by -90° across persistent segment boundaries. It is unclear where fault segments end and persistent segment boundaries begin, but if the persistent segment boundaries are considered to be -10-15 km across, the intervening fault segments achieve lengths of 30-35 km. In contrast known surface ruptures during the last -200 years, including those for the 1995 Egion earthquakes, have all been < 15 km in Length. these so-caIled earthquake segmens are, therefore, considerably shorter than the fault segmentss that hosted the earthquakes. Also, the positions of earthquake segments have varied relative to the positions of the fault segments during successive earthquakes. It appears, therefore, that a Modified Overlap Model is more appropriate than the Characteristic Earthquake Model to describe the seismological behaviour of fault segments around the Gulf of Corinth through a number of earthquake cycles. A pattern of coseismic slip vectors converging towards the hanging-wall has been measured for the surface ruptures to 1995 Egion earthquakes; a similar pattern was noted for the 1981 Alkyonides earthquake ruptures. Repetition of such ruptures in different positions along fault segments, in accordance with a Modified Overlap Model, will produce systematic variations in the scatter of fault-slip directions, with fault displacement. Thus, scatter in the orientations of lineations on fault planes may contain information concerning the lengths and positions of numerous pre-historic earthquake segments; information which may be used to constrain both the palaeoseismology and the future seismjcity in areas of active extension.