Weak-motion-based attenuation relationships for Israel

Abstract

We performed a regional study of earthquake ground motion scaling relations to provide the seismic hazard community of Israel with a new attenuation relationship that could be used for the prediction of earthquake-induced ground motion. Strong earthquakes are rather scarce in Israel and the existing collection of records from strong earthquakes is not sufficient to allowthe use of traditional regression methods to develop a regional attenuation relationship. We used velocity seismograms from the Israel Seismic Network to estimate the distance and frequency dependence of ground motion in the Israel region from frequent, smaller regional earthquakes. Our analyses included 4814 waveforms recorded by 30 stations of the Israel Seismic Network from 2000 to 2005. We restricted our analysis to 330 events recorded at five or more stations, with duration magnitudes ranging between 1.0 and 5.2. We derived empirical excitation, site and regional attenuation terms by regressing the peak amplitudes of narrowband-filtered seismograms around the shear wave arrivals and the rms Fourier spectral amplitudes taken around the specific sampling frequency. In order to optimize the attenuation parameters in our scaling model, we used a simple grid search. An optimal solution for minimal error between empirical and theoretical attenuation function was found for the quality parameter Q( f ) = 298 f^0.67 and the geometrical spreading g(r ) parametrized as a bilinear, piecewise function: r^−0.74 for r ≤ 60 km and r^−0.47 for r > 60 km. The spectral parameters κ of 0.015 s and stress drop increasing from 0.3 to 4 MPa were used to model the excitation spectra. A theoretical modelling effort based on Brune’s source spectrum and Random Vibration Theory (RVT) was performed on the attenuation and source parameters estimated in this study. Comparison of the attenuation relationship derived with locally measured ground motions shows excellent agreement with the data in the magnitude range forwhichwe have observations and seems to be adequate for predictions of earthquake ground motion for the Israel region. Comparison of Peak Ground Acceleration (PGA) predictions, based on our scaling relationship with those that have been recently used for seismic hazard analysis in Israel shows that our attenuation relationship predicts significantly lower ground motions than other relations.