Spectral decomposition of the Engineering Strong Motion (ESM) flat file: regional attenuation, source scaling and Arias stress drop

Abstract

We perform a spectral decomposition of the Fourier amplitude spectra disseminated along with the Engineering Strong Motion (ESM) flat file for Europe and Middle East. We apply a non-parametric inversion schema to isolate source, propagation and site effects, introducing a regionalization for the attenuation model into three domains. The obtained propagation and source components of the model are parametrized in terms of geometrical spreading, quality factor, seismic moment, and corner frequency assuming a ω 2 source model. The non-parametric spectral attenuation values show a faster decay for earthquakes in Italy than in the other regions. Once described in terms of geometrical spreading and frequencydependent quality factor, slopes and breakpoint locations of the piece-wise linear model for the geometrical spreading show regional variations, confirming that the non-parametric models capture the effects of crustal heterogeneities and differences in the anelastic attenuation. Since they are derived in the framework of a single inversion, the source spectra of the largest events which have occurred in Europe in the last decades can be directly compared and the scaling of the extracted source parameters evaluated. The Brune stress drop varies over about 2 orders of magnitude (the 5th, 50th and 95th percentiles of the ∆σ distribution are 0.76, 2.94, and 13.07 MPa, respectively), with large events having larger stress drops. In particular, the 5th, 50th and 95th percentiles for M > 5.5 are 2.87, 6.02, and 23.5 MPa, respectively whereas, for M < 5.5, the same percentiles are 0.73, 2.84, and 12.43 MPa. If compared to the residual distributions associated to a ground motion prediction equation previously derived using the same Fourier amplitude spectra, the source parameter and the empirical site amplification effects correlate well with the inter-event and inter-station residuals, respectively. Finally, we calibrated both non-parametric and parametric attenuation models for estimating the stress drop from the ratio between Arias intensity and significant duration. The results confirm that computing the Arias stress drop is a suitable approach for complementing the seismic moment with information controlling the source radiation at high frequencies for rapid response applications.

A regionally-adaptable ground-motion model for shallow crustal earthquakes in Europe

“…At a first glance, it may appear as if the spatial patterns are due to the predominant focal mechanisms in the regions, but the diversity of focal mechanisms within each region (especially among smaller events) dissuades this hypothesis. Spectral decomposition of the ESM dataset by Bindi and Kotha (2020) revealed a much larger stress-drop of the M6.45 Friuli earthquake compared to M6.5 Norcia earthquake-the δL2L l of regions containing these earthquakes also contrast similarly (Bindi et al 2019). Once again, correlating the random-effects obtained in response spectra domain (as here) to the earthquake parameters (such as stress-drop) Fig.…”

Empirical nonergodic shaking scenarios based on spatial correlation models: An application to central Italy

“…The Reasenberg's algorithm 35 is used to group the events: the earthquakes are counted within a time window of 10-days bins and considering a Poisson distribution. We obtain 20 clusters, which are further grouped based on the following criteria: (1) belonging to the same specific seismogenic setting (according to the source zones identified by DISS 3.2.1 catalogue, 2018), and (2) similarity in the stress drop 36,37 associated with the events of each cluster. In the end, six source areas are found ( Table 2 and Figure 2A To account for the source-to-site paths 2 , we divide the whole region into squared cells (0.2 • -spaced), following the approach by Dawood and Rodriguez-Marek.…”

Simulation of non-stationary stochastic ground motions based on recent Italian earthquakes

“…R is the distance (R JB ) and Q is the quality factor, proportional to the frequency as: Q = Q 0 f n , where Q 0 = 250 and n = 0.29. The values of the parameters adopted for the Brune model have been taken from Bindi et al (2018) and Bindi and Kotha (2020).…”

“…where f c is the corner frequency, Δσ is the stress drop in bars, M is the moment magnitude, and β is the shear-wave velocity in kilometers per second. The adopted values are Δσ = 50 bars and β = 3.5 km/s (Bindi and Kotha 2020).…”

“…A 0 is the source term representing the effect of both P and S waves; α is a parameter equal to 2 for body waves; Q c is the coda Q-value, assumed equal to Q S under the hypothesis of a common wave attenuation mechanism for the S and the coda wave (Roecker et al 1982). We adopted Q s = 250•f 0.29 (Bindi and Kotha 2020). Figure 8 shows the graphical definition of the envelope function Pa(t) proposed in this study in comparison with that adopted in SP96.…”