This electronic supplement includes additional information about the dataset used and the ground-motion prediction equations (GMPEs) that were preselected and implemented. Figures showing regression analysis for peak ground acceleration (PGA) values, location map, earthquake catalog, and summary table of parameters corresponding to each GMPE implemented.
The Akkar and Bommer (2010) ground-motion prediction equation (GMPE) for horizontal (geometric mean) peak ground acceleration (PGA), hereafter referred to as AB-2010, represents an update that corrects the shortcomings identified in Akkar and Bommer (2007a,b) and that then was applied to the strong-motion database of seismically active areas of Europe, the Mediterranean region, and the Middle East. The parameterization of AB-2010 contains magnitudes, distances, site types, and styles of faulting. Moment magnitudes between 5 and 7.6 and the Joyner–Boore (RJB) distances (Joyner and Boore, 1981) to 100 km are used. The site classes for the Eurocode 8 (2004; EC8) classification were A, B, C, and D. Normal, reverse, and strike-slip focal mechanisms are considered. In our implementation, the epicentral distances have been applied in place of RJB distances because of some incompleteness of the Italian database used (see Data and Resources). Because the calibration of AB-2010 was based on a dataset largely different from the one that we used in the main article, we set dk = 0 (see equation 7 of the main article), and σ = 0.279.
The Bindi et al. (2011) GMPE (B-2011) provides the geometrical mean of horizontal components for PGA. B-2011 was calibrated using the first version of the ITalian ACcelerometric Archive (ITACA) database (see Data and Resources), the strong-motion database for Italy. The regressions are performed for moment magnitudes between 4.1 and 6.9, for distances (RJB or epicentral for Mw < 5.5) up to 200 km, and depth shallower than 35 km. The focal mechanisms that are taken into account are normal, reverse, and strike slip. In this study, the recording sites are classified into the five EC8 classes: A, B, C, D, and E. In our implementation, only the epicentral distances have been considered. Because the calibration of B-2011 extensively used the data contained in ITACA-1.1, we set dk = 14 (see equation 7 of the main article), which corresponds to the number of parameters extracted to construct the attenuation law by the same data used in this work. The assumption of dk = 14 represents a penalizing factor for the performance evaluations. The value of σ is 0.337.
The Cauzzi and Faccioli (2008) GMPE (CF-2008) has been calibrated mostly (82.5% of the data) using digital records of the Japanese K-NET strong-motion network database (see Data and Resources) because of the high quality of the available accelerograms and the detailed information provided for each recording site. The remaining part of the data is from California, Europe, Iran, and Turkey. The horizontal PGA, defined as the geometric mean of the maximum values of the two horizontal components, is analyzed as a function of the magnitude, distance, site geology, and focal mechanisms. The moment magnitude range is 5.0≤ Mw ≤7.2, in which the lower bound is defined by the probabilistic deaggregation analyses of seismic hazard at many sites in Italy and the upper bound is representative of the largest estimated magnitudes in the historical earthquakes catalog in Italy. CF-2008 requires the hypocentral (focal) or fault distances. In our application, we use the hypocentral distance for focal depth less than 22 km. Because the calibration of CF-2008 was based on a dataset largely different from the one that we used in the main article, we set dk = 0 (see equation 7 of the main article), and σ = 0.344.
The Malagnini et al. (2008) GMPE (M-2008) is a more physically based model that takes regionalization into account. The computation of theoretical PGA considers hypocentral distances and moment magnitudes. The remaining part related to the attenuation effects depends on the specific crustal regional model attenuation in which source and receivers are embedded. The data used for calibration are peak ground velocities from narrowband-pass-filtered digital seismograms and Fourier amplitudes of the recorded spectra (see Malagnini et al., 2007, for details). The empirical results are modeled for the 1D regional propagation using the frequency-dependent quality factor Q(f) = 100f0.40 (Malagnini et al., 2008). The theoretical PGA values are derived from these attenuation relations for each moment magnitude and hypocentral distance, associating a theoretical PGA value to each 1 km distance (between 10 and 200 km). There are no imposed limits for magnitude and distance ranges, and we choose to use values consistent with the other values implemented (Mw ≥4.1; hypocentral distances between 10 and 200 km). We associated the corresponding PGA theoretical value to each kilometer. The focal depth is 22 km. To respect the regional attenuation relations required, we restricted our implementation to the region in central Italy where sources and sites are located (Fig. S2), and the theoretical PGA values applied are related to this attenuation model. Because the calibration of AB-2010 was based on a dataset largely different from the one we used in the main article, we set dk = 0 (see equation 7 of the main article); the associated value of σ is 0.23.
For many years, the Sabetta and Pugliese (1996) GMPE (SP-1996) was the most commonly used in seismic-hazard studies in Italy. SP-1996 was set using the strong-motion database containing 17 Italian earthquakes, and a regression analysis was used to examine the attenuation response of the larger of the two horizontal PGA components as a function of magnitude, distance, and site geology. The magnitude is surface-wave magnitude (Ms) when both local magnitude (ML) and Ms are greater than or equal to 5.5; otherwise, it is ML. SP-1996 cannot be used for magnitudes less than 4.6 or greater than 6.8 and at distances greater than 100 km; SP-1996 uses both epicentral and RJB distances. The recording sites are classified into three categories based on site geology: stiff, shallow alluvium, and deep alluvium. The response spectra of the horizontal components compared with the spectral shapes suggested by EC8 for soil classes A and B correspond roughly to stiff and deep alluvium. The ITACA-1.1 database (see Data and Resources) used for our implementation is provided by moment and local magnitudes only. To respect the type and magnitude ranges required by the authors, we convert Mw to Ms using the relation applied for the last Italian hazard map (Mappa di Pericolosità Sismica [MPS] Working Group, 2004, appendix 1). Finally, we set dk = 9; because SP-1996 was calibrated only with Italian data that were available at that time, we assumed that the nine parameters used to construct this attenuation law are derived from data very similar to ours. The assumption of dk = 9 represents a penalizing factor for the performance evaluations, and σ = 0.190.
Figure S1. Regression analysis for peak ground acceleration (PGA) values as a function of distances for the attenuation laws implemented in this study. Each plot is related to a GMPE applied while strictly respecting the parameter ranges required. Real (black dots) and predicted (colored dots) PGAs and residuals (black crosses) are displayed: (a) AB-2010, (b) B-2011, (c) CF-2008, (d) M-2008, and (e) SP-1996.
Figure S2. The region considered for the implementation of M-2008. The shaded area indicates the attenuation region; red and yellow dots represent sites and hypocenters, respectively, extracted from the ITalian ACcelerometric Archive (ITACA)-1.1 database (see Data and Resources).
Table S1. Complete list of 137 earthquakes used in the reference database ITACA-1.1 from 1972 to 2009.
Table S2. Schematic summary of the range of parameters corresponding to each GMPE implemented in this study.
Site classes, focal mechanisms, location of events, and peak ground acceleration (PGA) data were provided by the Italian Strong Motion Database (ITalian ACcelerometric Archive [ITACA]-1.1), available online at http://itaca.mi.ingv.it/ItacaNet/ (last accessed June 2013). The Japanese K-NET strong-motion network database is available online at http://www.kyoshin.bosai.go.jp/ (last accessed February 2015). Many of the figures were prepared using the Generic Mapping Tools package (http://gmt.soest.hawaii.edu/, last accessed June 2013) and MATLAB software (http://www.mathworks.it/products/matlab/, last accessed December 2014).
Akkar, S., and J. Bommer (2007a). Prediction of elastic displacement response spectra in Europe and the Middle East, Earthq. Eng. Struct. Dynam. 36, no. 10, 1275–1301.
Akkar, S., and J. Bommer (2007b). Empirical prediction equations for peak ground velocity derived from strong-motion records from Europe and the Middle East, Bull. Seismol. Soc. Am. 97, no. 2, 511–530.
Akkar, S., and J. Bommer (2010). Empirical equations for the prediction of PGA, PGV, and spectral accelerations in Europe, the Mediterranean region, and the Middle East, Seismol. Res. Lett. 81, 195–206.
Bindi, D., F. Pacor, L. Luzi, R. Puglia, M. Massa, G. Ameri, and R. Paolucci (2011). Ground motion prediction equations derived from Italian strong motion data-base, Bull. Earthq. Eng. 9, 1899–1920.
Cauzzi, C., and E. Faccioli (2008). Broadband (0.05 to 20 s) prediction of displacement response spectra based on worldwide digital records, J. Seismol. 12, 453–475.
Eurocode 8 (2004). Eurocode 8, design of structures for earthquakes resistance, part 1: General rules, seismic actions and rules for buildings, EN 1998-1, European Committee for Standardization (CEN), Brussels, Belgium, http://www.cen.eu/cenorm/homepage.htm (last accessed February 2016).
Joyner, W. B., and D. M. Boore (1981). Peak horizontal acceleration and velocity from strong-motion records including records from the 1979 Imperial Valley, California, earthquake, Bull. Seismol. Soc. Am. 71, no. 6, 2011–2038.
Malagnini, L., K. Mayeda, R. Uhrhammer, A. Akinci, and R. B. Herrmann (2007). A regional ground motion excitation/attenuation model for the San Francisco region, Bull. Seismol. Soc. Am. 97, 843–862.
Malagnini, L., L. Scognamiglio, A. Mercuri, A. Akinci, and K. Mayeda (2008). Strong evidence for non-similar earthquake source scaling in central Italy, Geophys. Res. Lett. 35, no. 17, doi: 10.1029/2008GL034310.
Mappa di Pericolosità Sismica (MPS) Working Group (2004). Redazione della mappa di pericolosità sismica prevista dall’Ordinanza PM 3274 del 20 marzo 2003, Rapporto conclusivo per il Dipartimento della Protezione Civile, INGV, Milano-Roma, April 2004, 65 pp. + 5 appendice.
Sabetta, F., and A. Pugliese (1996). Estimation of response spectra and simulation of nonstationary earthquake ground motions, Bull. Seismol. Soc. Am. 86, no. 2, 337–352.
[ Back ]
