Abstract Palermo is a populous city situated on the northern coast of Sicily, bordered by the Tyrrhenian Sea. This central part of the Mediterranean Sea features dramatic bathymetry, numerous subaqueous landslides and is also the epicentre to many subaqueous earthquakes. As such, the region is an area prone to tsunamis. This investigation uses the Cornell Multi-Grid Coupled Tsunami (COMCOT) tsunami modelling package to simulate five near-field landslides, and five near-field earthquakes regarded as worst-case credible scenarios for Palermo. The seismic simulations produced waves on a very small scale, the largest being c. 5 cm at its maximum height, and none of the earthquake-generated tsunami waves produced any measurable inundation. The landslide simulations produced larger waves ranging from 1.9 to 12 m in maximum height, two of which resulted in inundation in areas surrounding the Port of Palermo. Sensitivity analysis identified that fault width and dislocation as well as landslide-specific gravity did have significant influence over maximum wave height, inundation and maximum run-up wave height. There are methodological issues limiting the extent to which this study forms a comprehensive tsunami hazard assessment of Palermo, such as gaps in bathymetric data, computational restrictions and lack of a probabilistic element. These issues are counteracted by the fact that this study does serve as a robust first step in identifying that landslides in the region may produce larger tsunami waves than earthquakes, and that the directionality of mass movement is critical in landslide-driven tsunami propagation in the southern Tyrrhenian region.

ABSTRACT Large-scale correlations and sequence stratigraphic analyses have been carried out in the central Mediterranean region, a tectonically active area crossing the extensional margin of the southern Tyrrhenian, the compressional front of the Siculo-Maghrebian Tertiary chain and the North African foreland. The Plio-Pleistocene marine record has been subdivided in sequences and systems tracts on the basis of both original data and correlations. We provide seismic, well-log and outcrop data supporting the occurrence of regional unconformities of constant age, related to glacio-eustatic oscillations. Evidence of transgressive-regressive facies cycles having different orders of duration, major erosional truncations and basin starvation events contributed to the construction of a new sea-level cycle chart based on the available Mediterranean high-resolution biochronology and magnetostratigraphy. We largely used the deep-sea correlative conformities of sequence boundaries in order to improve the age calibration of the cycle chart. The chart, based on a new Plio-Pleistocene time scale, can resolve boundary ages up to 5th-order paracycles based on correlations to the high-frequency oscillations of the deep-sea record. Outcrop evidence of correlations between individual parasequences and 41 ky astronomical and climatic oscillations of the deep-sea record is supported by high-resolution biochronology. A comparison with the Mediterranean Plio-Pleistocene sequence chart confirms the general validity of the Gulf of Mexico cycle chart of Wornardt and Vail (1991), except for minor differences in age and number of 4th-order sequences. The sequence stratigraphic subdivisions are recognizable even in active sectors where stratal analysis separates the eustatic from the tectonic component; from this perspective, our experience support the regional synchroneity of sequences and systems tracts occurring in the studied interval independently of local tectonic factors.