Page 4 - Iannucci_alii_2015

GNGTS 2015 SESSIONE 3. 2

According to Vames (1978) and Hutchinson (1988), the resulting landslide process should be
defined as a complex-type one since the failure mechanism is a com bination oflateral spreading,
rock fall and rock topples. Moreover, the two main isolated blocks show a counter-slope top,
indicating that a roto-translational component is also present in the spreading process.
Data acquisition and processing. In May of 2015, a field-campaign was carried out in the
western part of Cala Rossa Bay t o record seismic am bient noise. Over an area of approximately
2
0.01 km , 25 single-station measurements were deployed to cover the aforementioned landslide
zones (Fig. l b). More in particular, 5 stations were placed in each of three unstable zones while
10 stations were deployed on the stable plateau area, at various distances from the f3 ground
crack.
Each station was equipped with a 3-component seismometer: 20 measurements were carried
out using a LE-3D/Ss seismometer by Lennartz Electronic GmbH coupled with a REFTEK
130-01 datalogger, setto a 250Hz sampling frequency; the other 5 measurements were carried
out using a 1.5 Hz SL06 acquisition unit by SARA Electronic Instmments, set to a 200 Hz
sampling frequency. Ambient noise records of 1-hour duration were acquired in each station
in two days characterized by different weather conditions: on May 27 with a strong wind (on
average 20 knots according to the Trapani forecast station) and sea waves di.rectly pounding
against the cliff and on May 30 with a weaker wind (on average 15 knots) blowing on the
opposite direction and so avoiding the generation of sea waves against the cliff.
The seismic noise records were processed by the use of Geopsy software (www.geopsy.
org). The 1-hour time histories were divided into non-overlapping windows of 40 sand the Fast
Fourier Transform (FFT) in the frequency range between 0.4 and 60.0 Hz was computed for
each component. By averaging over the windows, the amplitude spectra and the HIV spectral
ratio, as well as the distribution of their values in horizontal p lane (i. e. spectrum rotate and HN
rotate), were finally achieved for each single record.
The HVSR (Nakamura, 1989) analysis is worldwide used to predict the resonance frequency
of a site, particularly when layers having low shear-wave velocity cause a sharp impedance
contrast with the bedrock. The presence of a resonance peak in the HVSR curve has been
interpreted both in tenns of SH-wave resonance in soft surface layers, or in tenns of the
elli.pticity of particle motion when the ambient noise wave train is made up predominantly
of surface waves (Bonnefoy-Claudet et al., 2006). In practice, the wavefield is expected to be
a combination of both types and the HVSR curve contains infonnation about the shear wave
velocity profile in shallow sediments (Galea et al., 2014).
Results. By analyzing the FFT of the Cala Rossa Bay records (Fig. 2), a significant energy
contribution clea:rly appears a t frequencies lower than l Hz in May 27, when a strong wind
and sea waves acted against the cliff; on contrary, in May 30, with more favorable weather
conditions, such a spectral contribution disappears.
The HVSR curves show a peculiar shape both in the unstable and in the plateau zones, a
cave can be observed in the frequency range between l and 2 Hz, corresponding to a marked
amplification of the vertical ground motion component. The HVSRs also show an asymmetrical
peak between 5 and 6Hz and several peaks at higher frequencies (10-60Hz). Even though it is
not always significant according to SESAME guidelines (Bard, 2005), the HVSR peak at 5-6
Hz frequency is present in the measurements carri ed out within and in proximity of the unstable
zones, while it is not visible in the plateau area.
Discussion. Tlze Anchor Ba)l test sile. A comparison between the results obtained at Cala
Rossa Bay in Favignana Island and the ones of Anchor Bay in Malta was possible taking into
account the data already published by Galea et al. (2014).
The Anchor Bay test si te was chosen for such a comparison because of the lithotechnical
setting very similar to the one of Cala Rossa Bay. According to Pedley et al. (2002), in Anchor
Bay stiff limestones of Late Miocene age (Upper Coralline Limestone, UCL), with a thickness
of about 20 m, overlie blue clays of the Middle Miocene (Blue Clay fonnation, BC) about 30

59