Separation of Intrinsic and Scattering Attenuation Using the TERRAscope Data in Southern California
Anshu Jin, Kevin Mayeda, David Adams, & Keiiti AkiPublished 1994, SCEC Contribution #100
A multiple lapse time window analysis was applied to three-component broadband seismograms recorded at five TERRAscope stations in southern California to separate scattering and intrinsic attenuation. Seismic energies were integrated over three consecutive lapse time intervals: 0–15, 15–30, and 30–45 s (measured from the S arrival) for approximately 30 earthquakes with hypocentral distances of less than 70 km from each station. Using the fundamental separability of source, site, and path effects for coda waves, the integrated energies for different magnitude earthquakes were normalized to a common source size at each station, and the effect of near-site amplification is removed. Subsequently, we constructed a group of geometric spreading-corrected normalized energy-distance curves for each station region over frequency bands 0.5–1, 1–2, 2–4, and 4–8 Hz for all five stations. Two more frequency bands, 8–16 and 16–32 Hz, were added at stations PAS and SVD, for which higher sample rate data were available. A theoretical model of body wave energy propagation in a randomly heterogeneous elastic medium was employed to interpret the observation. Two parameters describe the medium in this model. These are the scattering attenuation coefficient η s and the intrinsic attenuation coefficient η i . By assuming that scattering is isotropic and including all orders of multiple scattering, this model predicts the spatial and temporal distribution of seismic energy. A two-step least squares fitting procedure was used to find the best fitting model parameters. The result shows the following: (1) the seismic albedo, B 0 = η s /(η i + η s ), increases with decreasing frequency for all station regions. The difference in B 0 among stations is frequency dependent, and the difference disappears at 6.0 Hz. (2) Significant differences exist for the scattering attenuation coefficient η s and seismic albedo B 0 among stations at lower frequencies. Stations on or close to the fault zone(s), such as PAS, PFO, and SVD, show a stronger scattering at frequencies of 0.75 and 1.5 Hz, whereas station ISA, located far away from any major active faults, shows the lowest scattering. (3) The coda Q −1 determined from the decay rate of coda waves lies, in general, between the intrinsic Q −1 and total Q −1, sometimes closer to the former than the latter, and sometimes opposite, depending on the frequency and station.
Citation
Jin, A., Mayeda, K., Adams, D., & Aki, K. (1994). Separation of Intrinsic and Scattering Attenuation Using the TERRAscope Data in Southern California. Journal of Geophysical Research, 99(B9), 17385-17848.
