GNGTS 2017 SeSSione 2.2

397

sIte effeCts In the pollIno reGIon from speCtrAl And polArIzAtIon AnAlyses of seIsmIC noIse And eArthquAkesF. Napolitano1, A. Gervasi2,3, M. La Rocca2, I. Guerra2, R. Scarpa1

1 Università degli Studi di Salerno, Italy 2 Università della Calabria, Italy3 Istituto Nazionale di Geofisica e Vulcanologia, Italy

Introduction. Site effects occur at any places where the propagation of seismic waves is affected by the local geological structure. This may produce changes in amplitude, duration, waveform and polarization of ground motion, as observed for a large number of earthquakes during the last decades (Bonilla et al., 1997; Clemente-Chavez et al., 2014). Site effects can increase significantly the damage produced by earthquakes (Cantore et al., 2011; Cultrera et al., 2016). The H/V spectral ratio (HVSR, Nakamura, 1989) of background seismic noise is probably the most used and efficient technique to estimate the site response in case of simple layered structures, overcoming the use of a reference site. Site effects can also modify the polarization of particle motion due to the proximity of fault damage zones (Pischiutta et al., 2017; Panzera et al., 2017) and topographic irregularities (Rigano et al., 2008; Pischiutta et al., 2010; Formisano et al., 2012). Accross fault zones the horizontal polarization of seismic signals usually shows high angle respect to the fault strike direction. Instead, topographic effects usually manifest as a horizontal polarization nearly perpendicular to the ridge direction, or nearly parallel to the slope direction. In this paper we evaluated site effects in the Mt. Pollino area (Southern Italy) using HVSR method applied to both seismic noise and earthquakes. Furthermore we computed the polarization analysis of seismic noise to look for possible relationship with fault zones and topography. The Pollino area is located at the junction between the southern Apennines northeast-verging collision orogen and the Calabrian rollback subduction zone. Till a couple of decades ago it was generally considered as a gap in the crustal seismicity. More recent seismic catalogs and papers include maximum magnitudes in the 5.2 – 6.0 range (e.g., 1693 and 1708 earthquakes; Rovida et al., 2011; Tertulliani and Cucci, 2014). Moreover, paleoseismological investigations of the southern slope of the Pollino area have identified faults capable of earthquakes of magnitude 6.5 – 7 (Cinti et al., 2002). It seemed appropriate to perform a detailed study of site response to improve seismic hazard assessment.

Data and methods. We applied the HVSR method to at least 20 hours of seismic noise for each site (Fig. 1), chosen at different time during night and day, for different week days

398

GNGTS 2017 SeSSione 2.2

during several months, in order to make the analysis reliable. Estimated HVSR curves are reliable and representative of each studied site, as confirmed by the stability of results and by the small values of standard deviation. We confined HVSR analysis in the 0.5 – 20 Hz frequency range, enough to investigate the main contribution from the geological structures present in the area and to make the analysis useful for engineering purposes. Te second step was the HVSR analysis applied to 83 local and regional earthquakes recorded by the 15 seismic stations shown in Fig. 1, for a total amount of 276 HVSR curves. Regional earthquakes were selected by requiring the SNR greater than 3, while highest SNR local earthquakes in the 0.5 – 20 Hz frequency band were chosen among a large number of recorded events. Signal spectra were computed on 20 s windows containing the direct S wave and early coda for regional earthquakes, while 10 s window length was used for local events. Then polarization analysis was computed on a moving window on bandpass filtered signals in different frequency bands, in order to investigate the frequency dependence: 1 – 2 Hz and 2 - 4 Hz, 3 - 5 Hz and 4 - 8 Hz. The stacking of polarization azimuth is plotted in rose diagrams with intervals of ten degrees.

Results. The analysis of site effects, taken out by computing the HVSR curves on the background seismic noise and local and regional earthquakes at 15 sites, has been improved by computing the polarization analysis. Fig. 2 shows the average HVSR curves obtained by noise recordings (red lines), compared with that computed from earthquakes recordings (black lines). The general characteristics of the two types of diagrams at the some station appear very similar at the most studied sites, significant differences being evident only at MMN, MMN2 and MMN4.

We correlated polarization results with topographic irregularities and known faults (Guerra et al., 2015) in the area, in order to explain those differences in a wide range of frequencies (1 - 2 Hz, 2 - 4 Hz, 3 - 5 Hz and 4 - 8 Hz). In Fig. 3 we show the rose diagrams computed

Fig. 1 - Station location.

GNGTS 2017 SeSSione 2.2

399

in two of the above-mentioned frequency bands, 2 - 4 Hz and 3 -5 Hz. The results show that at nine sites (MMN, MMN2, MMN4, MMN9, MMNB, T0711, T0716, T0713, MMN7) the polarization direction is nearly transversal to the ridge direction, at four sites (MMN1, MMN3, MMN5, MMN6) it is roughly parallel to it, while at only two sites (MMNA, T0714) it does not show a well defined polarization direction. Hereafter, we took into account the known faults near the analyzed sites, as shown in Fig. 1, considering only those known faults that are closer than 250 m to the analyzed sites. Because of this restriction, only 5 sites are taken into account (MMN4, MMN9, MMNB, T0711, T0713). At three of these 5 sites (MMN4, MMN9, T0711) the polarization direction is roughly normal to the fault strike, as observed by many authors in other regions (Rigano et al., 2008; Pischiutta et al.., 2017), while the last two sites show polarization oblique to the fault strike. At the three sites where the HVSR from noise and earthquakes are significantly different (MMN, MMN2, MMN4) the earthquake HVSR curves do not show the peaks that characterize the noise HVSR. This feature could indicate that the peaks in the noise HVSR is an effect related mainly with the topographic irregularities, represented by mountainous reliefes. MMN and MMN2 deserve careful attention, due to their location inside the built-up area of Mormanno, therefore being of particular interest in the view of seismic

hazard assessment. In summary, for 11 sites

out of the 15 studied the earthquake HVSR curves are very similar to the average noise HVSR. The most evident differences between noise and earthquakes are seen at MMN, MMN2, MMN4. The good agreement of noise and earthquake results for many sites is an important indication of the reliability of the HVSR method to estimate site effects through the analysis of seismic noise.

Fig. 2 - Average HVSR curves.

Fig. 3 - Rose diagrams.

400

GNGTS 2017 SeSSione 2.2

On the other hand, the three sites where we observe significant differences between noise and earthquake results remind us that much care must be taken when studying site effects. The results of this work indicate that in mountainous regions where the local geology is far from a horizontally layered structure, the H/V spectral ratio alone is not sufficient to characterize the site response. More exhaustive analysis need to be performed for a satisfying interpretation of the results obtained from seismic noise. It is also very important the fact that earthquake HVSR curves at sites with marked topographical effects do not show the same peaks, and often have lower amplitude. This result suggests overstimated site effects given by the noise HVSR. ReferencesBonilla, L. F., H. J. Steidl, G. T. Lindley, A. G. Tumarkin, and R. Archuleta; 1997: Site amplification in the San

Fernando Valley, California: Variability of the site-effect estimation using the S-wave, coda, and H/V methods, Bull. Seismol. Soc. Am. 87, no. 3, 710–730.

Cantore, L., A. Oth, S. Parolai, and D. Bindi; 2011: Attenuation, source parameters and site effects in the Irpinia–Basilicata region (southern Apennines, Italy). Journal of Seismology, 15(2), 375-389.

Cinti, F. R., M. Moro, D. Pantosti, L. Cucci, and G. D’Addezio; 2002: New constraints on the seismic history of the Castrovillari fault in the Pollino gap (Calabria, southern Italy), J. Seismol. 6, 199–217.

Clemente-Chavez, A., F. R. Zúñiga, J. Lermo, A. Figueroa-Soto, C. Valdés, M. Montiel, O. Chavez, and M. Arroyo; 2014: On the behavior of site effects in central Mexico (the Mexican volcanic belt. Natural Hazards and Earth System Sciences, 14(6), 1391.

Cultrera, G., E. D’Alema, S. Amoroso, B. Angioni, P. Bordoni, L. Cantore, ... and G. Di Giulio; 2016: Site effect studies following the 2016 Mw 6.0 Amatrice earthquake (Italy): the Emersito Task Force activities. Annals of Geophysics, 59.

Formisano, L.A., M. La Rocca, E. Del Pezzo, D. Galluzzo, C. Fischione, and R. Scarpa; 2012: Topography effects in the polarization of earthquake signals: a comparison between surface and deep recordings. Bollettino di Geofisica Teorica ed Applicata, 53(4), 471-484.

I. Guerra, A. Borghi, S. Del Mese, F. Filice, A. Gervasi, A. Mendicelli, L. Seeber, G. Stellato; 2015: Contribution to the Seismotectonic Studies in the Area at the Calabria-Lucania Border by Means of Geophysical Monitoring and Analyses, Atti 34° Conv. Ann. Gr. Naz. Geofis. Terra Solida, Trieste, 17-19 Novembre 2015.

Nakamura, Y.; 1989: A method for dynamic characteristics estimations of subsurface using microtremors on the ground surface, Q. Rep. Railway Tech. Res. Inst. Japan 30, 25–33.

Panzera, F., G. Lombardo, E. Longo, H. Langer, S. Branca, R. Azzaro, V. Cicala, and Trimarchi, F. (2017a). Exploratory seismic site response surveys in a complex geologic area: a case study from Mt. Etna volcano (southern Italy). Natural Hazards, 86(2), 385-399.

Pischiutta, M., G. Cultrera, A. Caserta, L. Luzi, and A. Rovelli; 2010: Topographic effects on the hill of Nocera Umbra, central Italy. Geophysical Journal International, 182(2), 977-987.

Pischiutta, M., M. Fondriest, M. Demurtas, F. Magnoni, G. Di Toro, A. Rovelli (2017). Structural control on the directional amplification of seismic noise (Campo Imperatore, central Italy). Earth and Planetary Science Letters 471, 10–18.

Rigano, R., F. Cara, G. Lombardo, and A. Rovelli; 2008: Evidence for ground motion polarization on fault zones of Mount Etna volcano. Journal of Geophysical Research: Solid Earth, 113(B10).

Rovida, A., R. Camassi, P. Gasperini, and M. Stucchi; 2011: CPTI11, la versione 2011 del Catalogo Parametrico dei Terremoti Italiani, Milano, Bologna.

Tertulliani, A., and L. Cucci; 2014: New insights on the strongest historical earthquake in the Pollino region (southern Italy), Seismol. Res. Lett. 85, 743–751, doi: 10.1785/0220130217.