Abstract:
The assessment of the distribution of potential earthquake damage at the urban scale provides
a basis for emergency planning and seismic risk reduction policies. Skikda is located in a
region of moderate seismicity. However, the level of risk is high due to the concentration of
the population and the legacy of old buildings designed without a seismic code. This work is
aimed at assessing the seismic vulnerability of the city of Skikda, using two different
methods, the first is a classical European method of the vulnerability indice Risk-UE, based
on an in situ survey, and the second is a method based on the technique of data mining
(Association of Learning Rules (ARL)) is to minimize, to avoid the costly process in seismic
vulnerability analysis and to use basic building attributes (number of floors or period of
construction) easily collected from visual observation, in order to create links between
building attributes and the EMS-98 vulnerability classes. The learning (proxy) of vulnerability
of Skikda of the ARL method is validated by the comparison with the Risk-UE method.
Concerning the scenarios of seismic risk, we are based on two earthquake scenarios, a
deterministic scenario (historical event) and the other probabilistic scenario (i.e. characterized
by return periods), in accordance with the study of seismic micro zoning of the city. The
results of the seismic risk assessment in the city are represented by the damage to buildings.
Thus, physical damage was also used as a risk-related variable to predict economic and
human losses and other quantities, such as restoration costs and debris quantities. Despite the
moderate seismic risk in Skikda, the results show a high risk due to the high exposure and
vulnerability of the buildings, which generally consist of unreinforced masonry buildings.
