Palaeoseismology on Jumping Faults: the Case of the Kamena Vourla Fault, Central Greece.

The basinward migration of normal faults is a widely documented phenomenon around the world, but just how and how fast this happens is still unclear. In Greece, active normal fault systems are often located along the sea and their most active segment lies underwater (eg 1981 Gulf of Corinth earthquakes). Palaeoseismologists are confronted with a difficult problem where the fault segment that they can trench on land is not necessarily the one that reflects the most recent fault activity. The Kamena Vourla fault offers the opportunity to study this phenomenon of `fault jumping' along the southern shore of the Gulf of Evvia in central Greece. Despite a dramatic tectonic expression on land, no historical or palaeoseismic earthquake can be positively ascribed to it. At its eastern tip, the Agios Konstantinos-Livanates segment exhibits a remarkably fresh assemblage of tectonic landforms. Uplifted sedimentary basins in its footwall, tilted wave-cut platforms, and dislocated stratigraphic marker horizons are among the clearest evidence of late Quaternary activity. Remarkable exposures of the main limestone fault plane also preserve signs of Holocene reactivation. However the occurrence of uplifted coastal sediments in its immediate hangingwall contests the idea that the onshore Agios Konstantinos-Livanates segment is the main active fault trace. Furthermore, identical Holocene marine fauna are presently exposed one meter above sea level identically on both sides of the fault. The tectonic geomorphology of the area strongly implies that an offshore segment, the Arkitsa segment, has gradually taken over some of its the activity during the late Pleistocene and the Holocene. To unravel the switching of activity between those two segments, we present a quantitative analysis combining a high-resolution photogrammetric Digital Elevation Model (DEM), field data with elastic half-space modeling. From the DEM, tilted former wave-cut platforms are automatically identified and their dips and azimuths measured. Comparing their azimuth trends with simple elastic half-space dislocations show that more than one fault is involved in the deformation. Although a prominent stratigraphic marker horizon, a fresh-water `beach-rock' containing Vivipara vivipara, is displaced by 30 m across the Agios Konstantinos-Livanates segment, this is insufficient to produce the tilt observed in lacustrine sediments in its hangingwall. Elastic modeling suggests that while the onshore Agios Konstantinos-Livanates fault probably had slip in excess of 200m since that lacustrine episode, the offshore Arkitsa fault segment needs an equivalent displacement to raise the hangingwall coastal terrace at Arkitsa to its current position. This indicates that both segments have coexisted for perhaps as long as one million years. The point of this poster is to illustrate that where fault activity switches back and forth between a dying onshore fault and a new submarine offshoot, new palaeoseismic tools are needed. Tectonic geomorphologic reconstruction that couples high-resolution topographic data with simple fault models offers one way of understanding the evolving fault behaviour.