11:00 AM - 11:15 AM
[SSS07-06] Amplitude dependence of energy dissipation in the crust and upper mantle
Keywords:quality factor, intrinsic attenuation
While the Earth (rocky planet) can be approximated as an elastic body, it also possesses anelasticity, Q-1, which causes energy dissipation (intrinsic attenuation) during deformation. Rock deformation experiments have demonstrated that Q-1 shows a marked amplitude dependency for strain of > ~10-6, but seismic-waves-induced strain is generally much smaller than the threshold value. Therefore, all seismological analyses have assumed amplitude-independent attenuation. However, theoretical models of dislocation-induced attenuation have predicted amplitude-dependent attenuation under high temperatures and pressures equivalent to the crust and the uppermost mantle conditions. This study investigates whether attenuation is actually independent of seismic-wave amplitudes by systematic analysis of a large number of spectral amplitudes of co-located earthquake pairs with different observed amplitudes.
Here we analyzed M3–5 intraslab (depth: 60–80 km), plate-boundary (10–50 km), and crustal (<~10 km) earthquakes that occurred beneath Tohoku, Japan. First, we calculated the spectral amplitudes using the vertical component of P waves and took the spectral ratio of any pairwise aftershocks in inter-event distances of ≦5 km at stations located within 200 km. Then, we divided the spectral ratio by the travel times to remove the effect of geometrical attenuation. Finally, we calculated the observed spectral ratios by stacking all pairwise earthquakes whose observed amplitude ratios fell within a certain range (every 0.2 in the ordinary logarithm of the amplitude ratio). The stacked spectral ratios were then fitted by the theoretical model to determine the Q-1 difference (ΔQ-1) with respect to different observed amplitude ratios.
The obtained results suggest that ΔQ-1 is zero for earthquake pairs that have identical observed amplitude, whereas ΔQ-1 monotonically increases as the amplitude ratio increases. This positive correlation between the observed amplitude ratio and ΔQ-1 suggests that seismic waves are amplitude dependent for intrinsic attenuation. Furthermore, we found that the intraslab earthquakes show the strongest positive correlation between the amplitude and Q-1. Our model calculation suggests that a proportionality between Q-1 and the seismic amplitude, A, can be represented as Q-1∝An. Here, n is estimated to range from 0.05 to 0.2, increasing with depth. We therefore infer that the physical mechanism causing amplitude-dependent attenuation works more effectively at higher temperature and pressure conditions.
Here we analyzed M3–5 intraslab (depth: 60–80 km), plate-boundary (10–50 km), and crustal (<~10 km) earthquakes that occurred beneath Tohoku, Japan. First, we calculated the spectral amplitudes using the vertical component of P waves and took the spectral ratio of any pairwise aftershocks in inter-event distances of ≦5 km at stations located within 200 km. Then, we divided the spectral ratio by the travel times to remove the effect of geometrical attenuation. Finally, we calculated the observed spectral ratios by stacking all pairwise earthquakes whose observed amplitude ratios fell within a certain range (every 0.2 in the ordinary logarithm of the amplitude ratio). The stacked spectral ratios were then fitted by the theoretical model to determine the Q-1 difference (ΔQ-1) with respect to different observed amplitude ratios.
The obtained results suggest that ΔQ-1 is zero for earthquake pairs that have identical observed amplitude, whereas ΔQ-1 monotonically increases as the amplitude ratio increases. This positive correlation between the observed amplitude ratio and ΔQ-1 suggests that seismic waves are amplitude dependent for intrinsic attenuation. Furthermore, we found that the intraslab earthquakes show the strongest positive correlation between the amplitude and Q-1. Our model calculation suggests that a proportionality between Q-1 and the seismic amplitude, A, can be represented as Q-1∝An. Here, n is estimated to range from 0.05 to 0.2, increasing with depth. We therefore infer that the physical mechanism causing amplitude-dependent attenuation works more effectively at higher temperature and pressure conditions.