GNGTS 2013 - Atti del 32° Convegno Nazionale

The results of HVNSR analysis are shown in Fig. 2 for all the deployed stations. First we examinated HVNSR for station AR1, in order to verify the possibility of using this bedrock site as a reference one. This control is necessary since the choice of a reference site is an important step in the analysis of seismic event with standard spectral techniques (Steidl et al. , 1996; Bordoni et al. , 2010). As clearly shown by figure 2 AR1 station presents H/V values with amplitude ranging between one and two in the entire analyzed frequency band. This observation confirms the hypothesis that this station can be considered as a reference site suggesting the idea that the limestone outcropping in the historical centre of Arpino can be really considered as seismic bedrock. The AR2H/V results indicate the presence of a clear peak centred at about 3.3Hz with average amplitude of about 4.5 – 5.0. This peak suggests the presence of some strong amplification effect at this site that is located only few tenths of meters away from the AR1 reference site. At AR3 H/V ratio presents a value close to the unity at low frequencies and reaches an average value between 2.0 and 2.5 in the frequency range between 1 and 10 Hz. This result suggest the presence of some slight amplification effect in the area and put into evidence a sensible difference with respect to AR2 site even tough the two stations are located on the same geological unit. The presence of low amplitude broadband curves instead of sharp peaks in the H/V curve can be related to the absence of strong impedance contrasts in the surface layers. For AR4 station H/V ratio shows a clear peak at a frequency of about 1.2 Hz with average amplitude of about 7. In this case some strong amplification effect can be inferred due to a probable impedance contrast between the soft alluvial sediments related to the river valley and the bedrock possibly due to the Santopadre formation. The second step of the seismic data analysis consisted in the application of standard spectral techniques (standard spectral ratio, SSR) (Bocherdt, 1970) to the aftershocks recorded by the four installed seismic stations. The seismic sequence of started on February 2013 was recorded by the Italian seismic network (RSNC) managed by INGV. In order to improve the quality of hypocentral localizations INGV installed some extra instrument in the area just after the mainshock of February 16. This allowed to both improve the quality of the event location and to lower the threshold of detectable magnitude. As a consequence of what above-mentioned INGV people was able to select a list of more than one hundred well located aftershocks with maximum local magnitude of 4.8. This information was used to extract from the continuous recordings at the AR stations the aftershocks time windows. Due to their purposes, the AR stations were installed in sites where the signal to noise ratio was not always so favourable, and this strongly influenced the quality of the collected aftershocks at least for low magnitude events. Due to these considerations the selection of the events for evaluating SSR spectral ratio involves a careful signal to noise analysis and up to now it is still going on. In order to derive at least some preliminary information on the transfer function of AR stations, using SSR approach and AR1 as reference station, we present the results for the magnitude Ml=3.1 aftershock of February 23 (17:05 UTC). The event was cut using a 30 seconds time window starting one second before the arrival of S waves picked on the horizontal components of the velocity time histories. A Fourier transform of the selected windows was evaluated and a smoothing algorithm was applied to reduce the oscillation on the spectral ratio. The spectral ratio was then evaluated on the smoothed spectra for all AR2, AR3 and AR4 stations, using AR1 as a reference site. The obtained results are shown in Fig. 3. For AR2 station SSR results show some clear amplification with two clear peaks. The first is centred between 3.0 and 3.5 Hz, the second, with bigger amplitude, between 4.0 and 4.5 Hz. The first peak corresponds, in frequency, with the H/V peak, while the second is only slightly visible in H/V data. As often found in spectral analysis, SSR results show amplitude much higher than H/V data. In our analysis this difference can also be enlarged by the lack of any statistical analysis on SSR data. 393 GNGTS 2013 S essione 2.3