Rigano, R.

Mt. Etna is characterized by flank instability of the eastern to south-western portions of the volcanic edifice, producing down-slope movements with rates up to several decimeters in a month during eruptive events of the 2002-2003 activity. The unstable sector is bounded to the North by a E-W transtensive fault (the Pernicana fault system), extending from the NE Rift to the coastline of the Ionian Sea for a length of >18 km. The western portion of the Pernicana fault system (close to Piano Pernicana area) is characterized by the most intense deformation. In this area we have performed volcanic tremor measurements on a dense grid along and across the fault zone. Ambient vibration measurements are also performed along a second fault (Tremestieri fault) which confines the slip of the eastern flank to the south-east. The analysis using both microtremors and local earthquakes recorded in these faults shows persistent polarization of ground motion. Horizontal-to-vertical spectral ratios (HVSR) show large directional resonances of horizontal components within the damaged fault zones. The resonance occurs around 1Hz at Piano Pernicana, and around 4 Hz in the Tremestieri fault zone. The resonance amplitude in the HVSRs seems to be fairly well correlated to soil gas anomalous concentrations (in particular, radon and CO2 both considered tracer gases of major crustal discontinuity) in the two fault zones, suggesting that both the effects are linked to local fracturing conditions. According to previous results on velocity anisotropy in the shallow crust, we believe that a role on polarization could be played by stress-induced anisotropy and micro-fracture orientation in the near-surface lavas. The occurrence of directional resonances, if confirmed in other faults, can be a powerful tool to map buried damaged fault zones on the Mt. Etna volcano.

A study of some earthquakes (M > 5.3) affecting Southeastern Sicily was performed to define their seismic sources and to estimate seismic hazard in the region. An analysis of historical reports allowed us to reassess intensities of the 1542, 1693, 1818, 1848 and 1990 earthquakes by using the new European Macroseismic Scale ’98. The new intensity data were used to define parameters and the orientation of seismic sources. The sources obtained were compared with the ones computed using the MCS intensities retrieved from the Catalogue of Strong Italian Earthquakes. The adopted procedure gives results that are statistically significant, but both the epicentre location and source azimuth, in some cases, are strongly affected by the azimuthal gap in the intensity distribution. This is evident mainly for the 1693 January earthquakes. For these earthquakes the macroseismic data uncertainty gives significantly different solutions, and does not allow the events to be associated with known active faults. By handling the new estimated intensity data and using the site seismic histories, the seismic hazard for some localities was calculated. The highest probability of occurrence, for destructive events (I = 10), was obtained in the area between Catania, Lentini and Augusta, suggesting that the seismogenic sources are located near the Ionian coast.

In this paper we investigate ground motion properties in the western part of the Pernicana fault. This is the major fault of Mount Etna and drives the dynamic evolution of the area. In a previous work, Rigano et al. (2008) showed that a significant horizontal polarization characterizes ground motion in fault zones of Mount Etna, both during earthquakes and ambient vibrations. We have performed denser microtremor measurements in the NE rift segment and in intensely deformed zones of the Pernicana fault at Piano Pernicana. This study includes mapping of azimuth-dependent horizontal-to-vertical spectral ratios along and across the fault, frequency–wave number techniques applied to array data to investigate the nature of ambient vibrations, and polarization analysis through the conventional covariance matrix method. Our results indicate that microtremors are likely composed of volcanic tremor. Spectral ratios show strong directional resonances of horizontal components around 1 Hz when measurements enter the most damaged part of the fault zone. Their polarization directions show an abrupt change, by 20° to 40°, at close measurements between the northern and southern part of the fault zone. Recordings of local earthquakes at one site in the fault zone confirm the occurrence of polarization with the same angle found using volcanic tremor. We have also found that the directional effect is not time-dependent, at least at a seasonal scale. This observation and the similar behavior of volcanic tremors and earthquake-induced ground motions suggest that horizontal polarization is the effect of local fault properties. However, the 1-Hz resonant frequency cannot be reproduced using the 1-D vertically varying model inferred from the array data analysis, suggesting a role of lateral variations of the fault zone. Although the actual cause of polarization is unknown, a role of stress-induced anisotropy and microfracture orientation in the near-surface lavas of the Pernicana fault can be hypothesized consistently with the sharp rotation of the polarization angle within the damaged fault zone.

During local and regional earthquakes, an evident amplification of horizontal ground motion is observed at two seismological stations near the Tremestieri fault, on the southeastern flank of Mt. Etna volcano. Rotated-component spectral ratios show a narrow spectral peak around 4-Hz along a N40°E direction. A conventional polarization analysis using the eigenvectors of the covariance matrix confirms the very stable directional effect enhancing the approximately NE-SW elongation of the horizontal ground motion in the fault zone. The effect is evident during the entire seismogram and independent of source backazimuth as well as distance and depth of earthquakes. The same polarization is observed in ambient noise as well. This consistency allowed us to use microtremors for checking ground motion polarization along and across the Tremestieri fault zone with a high spatial resolution. The result is a stable polarization of horizontal motion in the entire area, interesting a broad frequency band. To check whether this ground motion property is recurrent and understand a possible relationship with fault strike, faulting style, or orientation of fractures, ambient noise was recorded on other mapped faults of the Mt. Etna area, the Moscarello, Acicatena and Pernicana faults. The latter, in particular, is characterized by different strike and faulting style. A systematic tendency of ambient noise to be polarized is found in all of the faults. A picture emerges where normal faults of the eastern flank show a E-W to NE-SW polarization that changes on the Pernicana fault, which develops approximately E-W and is characterized by a prevailing NW-SE to NS polarization. Directions of polarization were never parallel to the fault strike. Moreover, polarization persists too far away from the fault trace, excluding an effect limited to a narrow low velocity zone hosted between harder wall rocks. Both these observations rule out an interpretation in terms of fault-trapped waves. The cause of observed polarizations will be the subject of future studies. However, the consistency with recent results of velocity anisotropy in a part of the investigated area suggests a possible role of attenuation anisotropy on horizontal amplitude variations versus azimuth.

Experimental data and numerical modelling were used to study the effect of local geology on the seismic response of the Catania area. The town extends on a marly clays bedrock and terraced deposits made up by coastal sands and alluvial conglomerates. This sedimentary substratum is deeply entrenched by paleo-valleys filled by lava flows and pyroclastics. Available borehole data and elastic parameters were used to reconstruct a geotechnical model in order to perfome 1D numerical modeling. Seismic urban scenarios were simulated considering destructive (M w = 7.0), strong (M w = 6.2) and moderate (M w = 5.7) earthquakes to assess the shaking level of the different outcropping formations. For each scenario seven real accelerograms were selected from the European Strong Motion Database to assess the expected seismic input at the bedrock. PGA and spectral acceleration at different periods were obtained in the urban area through the equivalent linear numerical code EERA, and contour maps of different levels of shaking were drawn. Standard and horizontal-to-vertical spectral ratios were achieved making use of a dataset of 172 seismic events recorded at ten sites located on the main outcropping lithotypes. Spectral ratios inferred from earthquake data were compared with theoretical transfer functions. Both experimental and numerical results confirm the role of the geological and morphologic setting of Catania. Amplification of seismic motion mainly occurs in three different stratigraphic conditions: (a) sedimentary deposits mainly diffused in the south of the study area; (b) spots of soft sediments surrounded by lava flows; (c) intensely fractured and scoriaceous basaltic lavas.