LAUSR

Abstract The extensive loss of human lives and building damage sustained during past tsunamis drove development of the first tsunami risk assessment in Israel. Although past tsunamis have occurred along the eastern Mediterranean coasts, no data is available for empirical hazard and risk analysis. Thus, our assessment focused on a simulated tsunami generated by an Mw 8.2 earthquake along the eastern Cypriot Arc (about 200 km NW of Haifa) in Bat Galim, the most vulnerable neighborhood in the city of Haifa, situated along the Mediterranean coast in northern Israel. This tsunami event represents the worst-case scenario threatening Bat Galim. Exposure data were taken from the National Archives of the Survey of Israel and was verified with a field survey. Two approaches were taken to derive the most appropriate vulnerability models for the study: first, already published Damage Criteria (DC) were developed based on information from past tsunami events around the world; second, damage functions were applied using the Hazus Tsunami Model (HTM), a multi-parameter loss estimation tool built using detailed engineering knowledge. Risk was considered the product of hazard, exposure, and vulnerability. Following several parametric studies and sensitivity tests, we produced loss assessment results from the two most realistic model outcomes and arrived at conclusions meaningful for tsunami preparedness in northern Israel. Most importantly, we note that the loss of life far surpasses building damage. Increasing community preparedness is the simplest and most cost-effective way to significantly reduce the risk of casualties during a tsunami. We also note that upgrading the seismic design of existing structures according to required standards in Israel will decrease vulnerability to tsunamis and expand possibilities for vertical evacuation. We discuss the accuracy and reliability of our analysis and demonstrate that a simple DC approach is useful for preliminary investigation in cases of limited data. The HTM however, requires detailed preparation and data inputs but produces more realistic loss estimates. These insights may aid others during future implementation of tsunami loss modeling in threatened coastal communities.