Multidisciplinary investigations using historical data, specific experimental surveys, numerical simulations and earthquake data to assess seismic hazard in a densely urbanized city: the study case of Palermo

Abstract

The city of Palermo (southern Italy) was severely damaged in the past by moderate-magnitude earthquakes located tens of kilometres offshore. The historical monumental heritage and the high density of population motivated large efforts for the seismic risk assessment. We present the geological and seismological studies performed in downtown Palermo as a study case to show how the complexity of an urban environment can be approached with multidisciplinary investigations. Downtown Palermo is characterized by sea deposits in the coastal zone and the alluvial deposits of two rivers (Papireto and Kemonia) of about 150 m width, which were buried and filled during the XVII century. The difficulty of surface geological surveys was compensated through an analysis of aerial photos and more than 2000 borehole data organized in the City-GIS of the Department of Geology and Geodesy of the University of Palermo. A previous study on the well-documented historical damage indicated the major role played by the two river valleys and the sea deposits in controlling the damage distribution, above the assumption of a fairly homogeneous vulnerability of the existing buildings in downtown. To test the feasibility of using ambient noise for recognizing the presence of alluvial deposits in a densely urbanized environment, a large microtremor measurement campaign was performed in Palermo across several profiles. The frequency peaks inferred from the horizontal-to-vertical spectral ratio were compared with numerical simulations to assess the seismic velocity profile and the soil stratigraphy. Moreover, noise data were analyzed through a statistical approach to establish a possible correlation between damage, resonance frequency and amplitude, and geology. After the moderate earthquake of September 6, 2002 (Mw=5.9, 50 km far away), the analysis of the aftershock sequence provided a well documented estimate of the variation of ground motion within the city in the case of linear soil response. Using these aftershocks we computed also synthetic accelerograms of the main shock through Empirical Green’s Functions that provided ground accelerations as large as 50 gals, consistently with the documented EMS-98 intensity. Synthetic accelerograms showed a large variability of horizontal ground motion within the city (a factor of 3 – 4) that confirms the role of local geology in causing an increase of the seismic hazard on sea and alluvial deposits. Finally, we discuss the comparison between the acceleration response spectra calculated for different soil categories and the design elastic spectra provided by EC8.