No abstract

Southern Italy is a key area for understanding the tectonic processes in the Africa‐Eurasia collision zone. We analyze new GPS measurements carried out between 1994 and 2001. The results are presented in terms of time series, trajectories and velocities of crustal motion, as well as a geodetic strain rate field. While central Italy, Corsica, Sardinia and the Tyrrhenian Sea move like the Eurasian plate, the overall motion of the Sicily Rift Zone region matches African plate motion. Unexpected are the north‐northwest directed motions of northeastern Sicily and the Eolian islands. Most striking on the deformation field are a north‐south oriented compression along the northern Sicilian coast, compression between Apulia and northwestern Greece as well as extension in the Sicily Rift Zone and the interior of Sicily.

No abstract

S U M M A R YAlthough the parameters used in seismic hazard analyses imply a long-term seismic strain rate, they are usually not checked against such alternative estimates. In this study, we determine hazard parameters for the eastern Mediterranean (a-value, b-value, m max and the corresponding long-term seismic moment rateṀ seis 0 ) consistent with seismicity data, tectonic information and geodetic strain rates. The dense data coverage in this region permits a detailed comparison of the horizontal seismic strain rate field,ε s , as recorded in the 500-yr long historical catalogue and the tectonic strain rate field,ε g , measured geodetically. We find thatε s is very similar in style over all magnitude ranges within each different tectonic regime in the study region. Furthermore,ε s is similar in style toε g . Except along the Hellenic Arc,ε s is consistent withε g in amplitude. We verify that for the high strain rates accommodated in the eastern Mediterranean and historical catalogues spanning at least 100-200 yr,ε s should reflect the long-term seismic strain release when averaged over each tectonic zone. To estimate such seismic strain reliably, accurate knowledge about the rates of recurrence of intermediate size events (M w = 4.5-6.5) is needed. For b ≥ 1, these events can accommodate up to 60 per cent of the strain. The combined analysis ofε g andε s provides an estimate of the seismic/total strain. The major strike-slip zones in the region, the Northern Anatolian Fault and the Kephalonian Fault, experience little to negligible aseismic deformation. In the remaining eastern Mediterranean up to 10-30 per cent of the total deformation is aseismic. The Hellenic Trench is largely uncoupled, with at least 50 per cent and up to 90 per cent of the compressive strain released aseismically. Only the extensional component of strain at the eastern end of this trench appears significantly seismically active.

For the insurance and reinsurance industries, earthquake loss estimation is crucial not only to adequately price its product but also to manage the accumulation risk in the face of the ever-increasing exposure in highly seismic regions. Changes in the built environment and a continuously evolving earthquake science make it a necessity for the industry to constantly refine earthquake loss estimation models. In particular, it has been recognized for a long time that the vulnerability of buildings to ground shaking is a key parameter in any earthquake risk model. Current methods tend either to rely on the limited historical damage and loss data or on the numerical simulation of the response of individual buildings to the ground-shaking produced by earthquakes. Although both methods have their advantages and pitfalls, we are proposing here a simple solution, using transparent input data, that can be realistically used for the needs of the insurance and reinsurance industry, whether detailed information about the insured structures is available or not. The resulting product is known as GEVES (Global Earthquake Vulnerability Estimation System). It is primarily intended for evaluating the mean damage ratio (MDR) suffered by a portfolio of buildings classified by use, under the action of a given earthquake scenario (i.e. an earthquake of given size at a given distance from the portfolio of buildings). A key assumption was that macroseismic intensity rather than spectral displacement would be the basis of loss estimation. The paper describes the model with emphasis on its structure and the justification for the assumptions made. In addition to a new set of earthquake vulnerability functions, the paper also provides recommendations on some aspects of the earthquake hazard, in particular about how to define macroseismic intensity at the site of interest, for a given earthquake scenario. This paper also discusses validation of the GEVES model against calculated vulnerability approaches, and the treatment of uncertainty within the model.