Scaling laws and spectral parameters of the ground motion in active extensional areas in Italy

Abstract

Abstract The horizontal components of 75 strong-motion accelerograms recorded in Italy in the Central and Central-Southern Apennines on the occasion of moderate to strong earthquakes (4 ≲ M L ≲ 7) have been analyzed. This data set has been selected as being representative of a seismotectonic environment generating earthquakes with prevalent normal faulting, at depths varying roughly from 5 to 15 km. In order to obtain the parameterization of the spectrum of motion in the band of engineering interest (0.1 ≲ f ≲ 20 Hz) for seismic risk purposes, physical quantities describing source, propagation, and site effects have been investigated. Spectral analysis reveals that local site response characterizes the frequency behavior of the ground motion at the available stations. In spite of slight variations in the spectrum at the same site for different magnitude earthquakes, significant differences appear at different stations for the same earthquake. In this context, the use of a simple source model has been preferred to more complex ones. The theoretical omega-square spectrum S ( f ; M 0 , f c ) has been used for describing source radiation as a function of earthquake size. Apart from specific site frequency response, on average, the optimal expression describing the acceleration spectrum of shear waves recorded at hypocentral distances, R , up to 100 km from the focus is A ( f ) = S ( f ; M 0 , f c ) R e − π κ 0 f e − ( π R / v Q 0 ) . κ 0 and Q 0 are two attenuation parameters which have been estimated to be, on the average, 0.07 and 100 sec, respectively; v is shear-wave velocity in the lithosphere. The corner frequency, f c , depends on seismic moment M 0 following the relation f c = 4.9 x 10 6 v ( 〈 Δ σ 〉 M 0 ) 1 / 3 where 〈Δσ〉 is the mean value of the Brune stress drop, that has been estimated to be roughly 200 bars for the earthquakes in analysis . The repetitive character of the frequency response shown by stations triggered by different earthquakes independently of azimuth and magnitude seems to suggest the use of small-size earthquakes ( M L ∼ 4) for successful predictions of the strong ground motions at specific sites (at least for magnitudes M L ∼ 7). Moreover, the proposed spectral model allows numerical simulations of the peak values of motion versus earthquake size, in good agreement with the average trend found by other authors for the Italian earthquakes in the magnitude range of engineering interest .