Statistical estimates of earthquake magnitudes are unreliable when based on very few historical data. Additional sources of information, such as geological data, are then necessary to update estimates of seismicity parameters. The Bayesian probability theory is a tool to combine prior information of seismicity obtained from geological data with historical observations. This theory is tested in the case of the Inner Messiniakos fault zone, southern Greece, for the estimation of the probability of occurrence of strong earthquakes. Prior estimates of seismicity are developed from slip rate measurements, obtained from offsets of geological formations, on the basis of both onshore and offshore neotectonic data. The analysis emphasizes the importance of the input seismicity parameters, particularly the significance of the upper bound magnitude in the estimation of the seismic potential of active faults

We present the results of a wide-aperture seismic onshore–offshore study (Crete Seismic Experiment) in the Cretan region as part of the Hellenic arc compressional system. Three seismic lines were carried out on and around the island of Crete in order to investigate the crustal structure of the region. Up to 119 three-component recording stations were deployed on each profile that observed seismic energy generated by a 48-l airgun array and eight 20-kg landshots. A total of 6208 shots were fired. Upon completing the fieldwork, the vertical components of all stations were evaluated; 300 Common-Receiver-Gather (CRG) sections of the ocean bottom seismographs (OBS) and land stations as well as 100 Common-Source-Gather (CSG) sections of the land shots and selected airgun shots were compiled and modeled in order to generate a 2D P-wave velocity–depth model for each profile. The accuracy of the model depends on the depth and position along the profiles and does not exceed 5% for both depth and P-wave velocity. We identified strong lateral variations in crustal and sedimentary thickness mainly in a north–south direction but also along strike (east–west). The crust is continental and has a maximum thickness of 32.5 km below northern central Crete. Its subdivision in an upper (vp = 5.8–6.3 km/s, locally up to 6.5 km/s) and a lower (vp = 6.4–6.9 km/s) part is justified by a first-order discontinuity with vp-velocity a contrast of up to 0.6 km/s. The eastern part of Crete shows a significantly thinner crust of 24 to 26 km. To the North, the crustal thickness decreases to 15 km below the central Cretan Sea. The prominent decrease of the Moho depth north of central Crete is interpreted to represent the northern end of a microcontinent that was subducted in Oligocene times and later surfaced by ‘buoyant escape’ (Sto¨ckhert et al., 1999; Thompson et al., 1999). The P–T–t–D history of the high-pressure rocks of Crete, Greece: denudation by buoyant escape. In: Exhumation Processes: Normal Faulting, Ductile Flow and Erosion. Ring, U., Lister, G.,Willet, S., Brandon, M. (Eds.), Spec. Publ. of the Geol. Soc. of London,p. 154]. To the south and southwest of the island, the continental crust gradually thins to a minimum of 17 km and atapproximately 100 km off the southern coast of Crete, it is in contact with oceanic crust below the Mediterranean Ridge. Upper mantle velocities were determined to be 7.7 km/s below the Cretan Sea and 8.0 km/s south of Crete. Below the continental Cretan crust, a 6- to 7-km-thick layer with vp-velocities between 6.6 and 7.1 km/s was identified on each line and could be followed by reflections to a depth of 42 km. It is decoupled from the overlying continental crust at central Crete and is interpreted as oceanic crust presently under subduction towards the NNE below the Aegean Sea.