P wave attenuation structure in the fault area of the 1995 Kobe earthquake

Abstract

The three‐dimensional (3‐D) P wave attenuation structure in the Kobe epicentral area is determined using aftershock records from the 1995 Kobe (Hyogo‐ken Nanbu) earthquake. Waveform data from both permanent seismic networks and portable stations, set up following the Kobe mainshock, were used. The observed P wave amplitude spectra were inverted for the spectral plateau value, the source corner frequency, and a frequency‐independent t* operator assuming an ω2‐type source model. The calculated t* operators were subsequently used for the computation of the 3‐D absorption structure using a 3‐D velocity model. Only high‐quality amplitude spectra were used for the estimation of reliable attenuation parameters, resulting in a final data set of ∼4100 t* operators. Two regions of high absorption were found in the upper part of the crust. One is below the northern part of Awaji Island and reaches down to a depth of 5 km. The other region, northeast of Kobe city, is along the Arima‐Takatsuki tectonic line and extends along the Rokko, Kashiodani, and Yamada Faults. It extends northward to the Hokusetsu‐Sanchi area. In both regions a high density of fault lineations is mapped, and high Poisson ratios are known. This suggests that partially saturated cracks cause the observed high attenuation.