2318 Bulletin of the Seismological Society of America, Vol. 95, No. 6, pp. 2318–2327, December 2005, doi: 10.1785/0120040214 E Investigation of Surface Geology and Intensity Variability in the Palermo, Italy, Urban Area after the 6 September 2002 Earthquake by M. Stella Giammarinaro, Andrea Tertulliani, Gianfranco Galli, and Mario Leta Abstract For the first time, a high-density macroseismic survey has been carried out in the city of Palermo, Italy, after the 6 September 2002, M w 5.9 earthquake. The aim was to investigate the spatial relationships and correlations between intensity data and surface geology. A very dense database has been created to store a large amount of macroseismic, stratigraphic, and geotechnical information. A Geographic Information Systems (GIS) tool, called City-GIS, enables data processing by instru- ments and research keys devoted to seismic hazard. The results show the feasibility of such procedures, especially for large cities where the surface geology critically influences the ground motion. Our analysis, made on the basis of more than 2000 intensity observations scattered in the Palermo urban area, demonstrates the role of the variable geology in the variability of shaking. Significant variations of intensity are related to the geometry of the geologic formations present in the area. Intensities increase by about one intensity class inside those city sectors where the stratigraphic succession is characterized by a thin layer of calcarenites overlying very thick layers of silty clayey sands. Major amplification of up to two classes occurs in alluvial deposits in the southeastern part of Palermo, as expected. These results extend our knowledge of the seismic response in the outskirts of Palermo, which until now, was limited to the city center. Online material: Macroseismic survey form and data distribution. Introduction Figure 1. (a) Schematic geological map of the study area; the three rivers crossing the area are also indicated. The epicenter and focal mechanism of the 6 September 2002 earthquake and the high urbani- zation zone of Palermo (within the blue line) are in- dicated. (b) Total distribution of intensity classes from the macroseismic questionnaires. (c) Spatial dis- tribution of the intensity points in the city. Blue points indicate intensities lower than 5, and red points in- dicate intensities greater than or equal to 5. High variability of surface geology in urban areas can be responsible for a significant variation in ground motion. Studies on site effects in urban areas have increased in the recent past, especially after the occurrence of heavy destruc- tion in large cities caused by amplification phenomena: Mex- ico City in 1985 (Singh et al., 1988), Armenia in 1988 (Had- jian, 1993), and Izmit in 1999 (Bakir et al., 2002). The evaluation of site effects is commonly performed by com- paring spectral ratios from weak- or strong-motion record- ings with surface geology, at a fine scale when sufficient data are available. Unfortunately, strong-motion data are not always available in urban areas, especially at a fine scale, so the application of macroseismic analysis is a powerful tool for collecting a large amount of data by a fast and inexpen- sive technique (Toshinawa et al., 1997; Cifelli et al., 1999). In this work, we applied the macroseismic approach to the city of Palermo, which was hit by a seismic sequence that started on 6 September 2002, M w 5.9 (Fig. 1), with epi- center about 50 km offshore from the city. Palermo experi- enced Mercalli-Cancani-Sieberg (MCS) 6 shaking at maxi- mum intensity. In spite of its epicentral distance and moderate magnitude, the earthquake provoked much alarm and fear in Palermo. In some cases, light damage occurred, especially in historical buildings and churches in the center and in southeastern parts of the city (Azzaro et al., 2004). A recent study by Guidoboni et al. (2003) indicated that the city center of Palermo is prone to damaging site effects during earthquakes, such as occurred in 1726, 1823, and 1940. In fact Guidoboni et al. (2003) show a close correlation be- tween major damage and alluvial deposits, especially where the thickness of sediments exceeds 12 m. No information on the effects of past earthquakes on the north and south sectors is available, because they are recently urbanized areas.