GNGTS 2022 - Atti del 40° Convegno Nazionale

GNGTS 2022 Sessione 1.1 21 NEAR-SOURCE ATTENUATION AND SPATIAL VARIABILITY OF THE SPECTRAL DECAY PARAMETER KAPPA IN CENTRAL ITALY R.R. Castro 1 , L. Colavitti 2 , C.A. Vidales-Basurto 3 , F. Pacor 2 , S. Sgobba 2 , G. Lanzano 2 1 Centro de Investigación Cientifica y de Educación Superior de Ensenada (CICESE), División Ciencias de la Tierra, Departamento de Sismología, Ensenada, Baja California, Mexico 2 National Institute of Geophysics and Volcanology (INGV), Section of Seismology applied to Engineering, Milano, Italy 3 Universidad Autónoma de Zacatecas, Unidad Académica de Ciencia y Tecnología de la Luz y la Materia (LUMAT), Zacatecas, Mexico Introduction. The attenuation decay parameter (κ) is broadly used in many earthquake engineering applications, as for removing the path and site attenuation effects of ground motion prediction equations (Ktenidou et al. , 2014; Ktenidou et al. , 2015). The observed values of κ are usually the results of path effects, which are distance dependent, and attenuation near the source and the recording site. As written in Anderson (1991) and Ktenidou et al. (2014), the distance dependence of κ is: κ ( r ) = κ 0 + κ s + κ avg ( r ), in which κ 0 is the near-site attenuation parameter, κ s is the near-source attenuation parameter, and κ avg ( r ) is the average attenuation along the S-wave path. In this work (more details are in Castro et al. , 2022a), we compare estimates of κ s determined from two different techniques, and we also study the spatial variability of κ avg ( r ) using data from the dense array of Central Italy, which is characterized for being a seismically active region and tends to generate long sequence, as the case of Amatrice-Visso-Norcia sequence (Chiarabba et al. , 2018). Data. Over the last 25 years, several earthquakes had occurred in this region, from the 1997- 1998 Umbria-Marche sequence, to the 6 April 2009 M w 6.1 L’Aquila mainshock, and the 2016- 2017 M w 6.5 Amatrice-Visso-Norcia sequence. For the scopes of this research, we analyzed both accelerograms and velocity records since 2008, which are used for phase picking and earthquake locations (Pacor et al. , 2016; Castro et al. , 2021). To analyze the spatial variability of κ , we used a dataset in which we considered events coming from different source areas based on the spatial clustering performed by Sgobba et al. (2021) for the Central Apennines region. We selected events detected by at least three stations, and we set a threshold in which the ratio between the standard deviation of κ and the mean of κ was less than 50%. In this dataset, the selected events have magnitudes fromM 3.2 to 6.3, focal depths between 6.1 and 17.1 km, and the hypocenter distances ranging between 7.1 and 168.8 km, for a total of 266 earthquakes, 353 stations, and 13,952 observations of κ (Fig. 1a). We formed four groups of stations distributed by different quadrants, taking as axis of symmetry the average strike of the Apennine faults, which is approximately 150° along the NW-SE direction (Improta et al. , 2019; Vignaroli et al. , 2020) and follows the epicentral distribution of most of the selected earthquakes (Fig. 1b). The reason for this division of the κ avg ( r ) estimation in each quadrant is to analyze where there is any spatial variability of the spectral decay parameter. METHOD. Estimates of kappa. We compute κ following the method proposed by Anderson and Hough (1984), in which the logarithm of the high-frequency S-wave spectral amplitude acceleration is least squared fitted and κ is computed from the slope of the linear fit. We used a semiautomatic technique to estimate κ from each record, which consists in precompute the slopes over 11 frequency bands, with variable length in the range between 8 and 38 Hz (Lanzano et al. , 2022), and then averaging these values to obtain the final estimate. The median value is