Probabilistic Tsunami Risk Assessment from Incomplete and Uncertain Historical Impact Records: Mediterranean and Connected Seas.

Tsunami risk is considered as the probability of a particular coastline being struck by a tsunami that may cause a certain level of impact (destructiveness). The impact metric of a tsunami is expressed in terms of tsunami intensity values, K, assigned on a 12-degree scale. To calculate tsunami risk we are based on the tsunami history of the region codified in tsunami catalogues. The probabilistic model adopted was used successfully for hazard assessment of earthquakes (Kijko et al. in Bull Seismol Soc Am 79:645–654, 2016) and of tsunamis (Smit et al. in Environmetrics 30:e2566, 2019) by considering seismic magnitude and tsunami height as metrics of the respective hazards. In this model, instead of hazard metrics we inserted risk metric, i.e. wave impact in terms of intensity values. The procedure allows utilization of the entire data set consisting not only from the complete (recent) part of tsunami catalogue but also from the highly incomplete and uncertain historical part of the catalogue. Risk is assessed in terms of probabilities of exceedance and return periods of certain intensity values in specific time frames. We applied the model using catalogues for the Mediterranean and connected seas. Sensitivity analysis showed that using complete data sets generally provided more realistic results than using entire data sets. Results indicated that the risk level depends on the seismicity level and not on the size of individual ocean basin. The highest tsunami risk level was found in the eastern Mediterranean (EM), with a significantly lower risk in the western Mediterranean (WM). In the Marmara Sea (MS), the tsunami risk was low, and the lowest was in the Black Sea (BS). The risk in the small Corinth Gulf (CG, Central Greece) was comparable to that of WM. The return period of damaging tsunamis (i.e. K ≥ 7) was 22 years in the entire Mediterranean basin and 31, 118, 135, 424, and 1660 years in the EM, WM, CG, MS, and BS basins, respectively. [ABSTRACT FROM AUTHOR]