Development of an earthquake loss model for Turkish catastrophe insurance

Following the devastating Kocaeli and Düzce earthquakes of August andNovember 1999, the Turkish Government was faced with an enormousfinancial burden as a result of its statutory obligation to cover the full costsof rebuilding. In order to offset this liability in the future – which has hadan adverse effect on the Government's economic programme – acompulsory earthquake insurance scheme has been introduced for allhouseholders in Turkey. A key element for successful implementation ofthis novel and ambitious programme is the transfer of the earthquake riskabsorbed by the Turkish Catastrophe Insurance Pool (TCIP) to theinternational reinsurance market. An earthquake loss model, described inthis paper, has been developed for the TCIP to serve as a basis for thedecision-making process with respect to the pricing of its insurance policy,risk control, the purchase of reinsurance, and the transfer of seismic risk.Sample results of the loss calculations are presented.     

3D Hybrid Ray-FD and DWN-FD Seismic Modeling for Simple Models Containing Complex Local Structures

Hybrid approaches find broad applications wherever all-in-one modelling of source, path, and site effects is too expensive. Our new 3D hybrid approach allows to compute the seismic wavefield in elastic isotropic models containing a complex local structure embedded in a large, but considerably simpler, regional structure. The hybrid modelling is realized in two successive steps.In the 1^st step, the ray or discrete wave number (DWN) method is used to compute the seismic wavefield due to the source and simple regional structure. The complex local structure is not present. Thus, the excitation contains the source and regional path effects. The time history of this wavefield (excitation), recorded at the points of so called excitation box, is stored on a disk. The excitation box envelopes a small portion of a computational domain.The 2^nd step of the hybrid method, now containing the complex local structure, is computed by finite differences (FD) inside the excitation box and its close vicinity. The excitation from the 1^st step is now used to inject the 1^st step wavefield into the 2^nd step computation. After that, the hybrid combination of the 1^st and 2^nd steps contains the source, regional path, and local structure effects at reasonably lower computational costs than in case of all-in-one modelling.The 3D ray-FD method is tested on models in which the locally complex structure is the well-known Volvi lake basin, embedded in various 1D structures. The wavefield is excited by the point source situated outside the basin. Although the structure outside the excitation box may be less dimensional (2D, 1D, homogeneous), the whole problem is actually 3D due to the 3D features of the structure inside the excitation box, 3D shape of the excitation box, and arbitrary source — excitation-box configuration. Simple (1D) structures outside the excitation box allow for comparison with the alternative hybrid DWN-FD results. However the ray method is suitable for computation of 3D regional structures outside the excitation box. The results from both approaches show a very good agreement for realistic crustal and local structural models.     

Middle Pleistocene to Holocene activity of the Gondola Fault Zone (Southern Adriatic Foreland): Deformation of a regional shear zone and seismotectonic implications

Recent seismicity in and around the Gargano Promontory, an uplifted portion of the Southern Adriatic Foreland domain, indicates active E–W strike-slip faulting in a region that has also been struck by large historical earthquakes, particularly along the Mattinata Fault. Seismic profiles published in the past two decades show that the pattern of tectonic deformation along the E–W-trending segment of the Gondola Fault Zone, the offshore counterpart of the Mattinata Fault, is strikingly similar to that observed onshore during the Eocene–Pliocene interval. Based on the lack of instrumental seismicity in the south Adriatic offshore, however, and on standard seismic reflection data showing an undisturbed Quaternary succession above the Gondola Fault Zone, this fault zone has been interpreted as essentially inactive since the Pliocene. Nevertheless, many investigators emphasised the genetic relationships and physical continuity between the Mattinata Fault, a positively active tectonic feature, and the Gondola Fault Zone. The seismotectonic potential of the system formed by these two faults has never been investigated in detail. Recent investigations of Quaternary sedimentary successions on the Adriatic shelf, by means of very high-resolution seismic–stratigraphic data, have led to the identification of fold growth and fault propagation in Middle–Upper Pleistocene and Holocene units. The inferred pattern of gentle folding and shallow faulting indicates that sediments deposited during the past ca. 450 ka were recurrently deformed along the E–W branch of the Gondola Fault Zone.We performed a detailed reconstruction and kinematic interpretation of the most recent deformation observed along the Gondola Fault Zone and interpret it in the broader context of the seismotectonic setting of the Southern Apennines-foreland region. We hypothesise that the entire 180 km-long Molise–Gondola Shear Zone is presently active and speculate that also its offshore portion, the Gondola Fault Zone, has a seismogenic behaviour.     

The Late Pleistocene clastic deposits in the Romito Cave,southern Italy: a proxy record of environmental changes and human presence

Clastic sediments deposited in caves and rock shelters bear peculiar sedimentological characteristics and have seldom been considered as a high‐resolution proxy record of climatic or environmental changes. The Romito Cave has its entrance at 275 m above sea level, about 25 km from the Tyrrhenian coast of Calabria, southern Italy. New archaeological excavation performed since 2000 has revealed a sedimentary succession spanning the record of Gravettian to Late Epigravettian cultures (Late Pleistocene). The present study focuses on the lower part (2.5 m thick) of the succession, where three main unconformity‐bounded stratigraphic units have been recognised (labelled RM1–3). Each unit consists of water‐lain deposits indicating high‐ to low‐competence flow, capped with anthropogenic deposits. The gradual deactivation and reactivation of the water drainage between 23 475 ± 190 and 16 250 ± 500 cal. a BP is correlated with regional precipitation changes due to the onset of dry climatic conditions of the Last Glacial Maximum. However, the deactivation of cave drainage after the deposition of unit RM3, around 15 400 ± 500 cal. a BP, deviates from the regional hydrological trend of progressively increasing water discharges and is attributed to the drainage cut‐off by probable cave wall collapses. Copyright © 2008 John Wiley & Sons, Ltd.     

The rate of sedimentation from turbulent suspension: An experimental model with application to pyroclastic density currents and discussion on the grain‐size dependence of flow runout

Large‐scale experiments generating ground‐hugging multiphase flows were carried out with the aim of modelling the rate of sedimentation, of pyroclastic density currents. The current was initiated by the impact on the ground of a dense gas‐particle fountain issuing from a vertical conduit. On impact, a thick massive deposit was formed. The grain size of the massive deposit was almost identical to that of the mixture feeding the fountain, suggesting that similar layers formed at the impact of a natural volcanic fountain should be representative of the parent grain‐size distribution of the eruption. The flow evolved laterally into a turbulent suspension current that sedimented a thin, tractive layer. A good correlation was found between the ratio of transported/sedimented load and the normalized Rouse number of the turbulent current. A model of the sedimentation rate was developed, which shows a relationship between grain size and flow runout. A current fed with coarser particles has a higher sedimentation rate, a larger grain‐size selectivity and runs shorter than a current fed with finer particles. Application of the model to pyroclastic deposits of Vesuvius and Campi Flegrei of Southern Italy resulted in sedimentation rates falling inside the range of experiments and allowed definition of the duration of pyroclastic density currents which add important information on the hazard of such dangerous flows. The model could possibly be extended, in the future, to other geological density currents as, for example, turbidity currents.     

The European Upper Mantle as Seen by Surface Waves

We derive a global, three-dimensional tomographic model of horizontally and vertically polarized shear velocities in the upper mantle. The model is based on a recently updated global database of Love- and Rayleigh-wave fundamental-mode phase-anomaly observations, with a good global coverage and a particularly dense coverage over Europe and the Mediterranean basin (broadband stations from the Swiss and German seismic networks). The model parameterization is accordingly finer within this region than over the rest of the globe. The large-scale, global structure of our model is very well correlated with that of earlier shear-velocity tomography models, based both on body- and surface-wave observations. At the regional scale, within the region of interest, correlation is complicated by the different resolution limits associated to different databases (surface waves, compressional waves, shear waves), and, accordingly, to different models; while a certain agreement appears to exist for what concerns the grand tectonic features in the area, heterogeneities of smaller scale are less robustly determined. Our new model is only one step towards the identification of a consensus model of European/Mediterranean upper-mantle structure: on the basis of the findings discussed here, we expect that important improvements will soon result from the combination, in new tomographic inversions, of fundamental-mode phase-anomaly data like ours with observations of surface-wave overtones, of body-wave travel times, of ambient “noise”, and by accounting for an a-priori model of crustal structure more highly resolved than the one employed here.