SEISMIC ALBEDO FROM THE TOTAL SEISMIC ENERGY-DEPENDENCE ON HYPOCENTRAL DISTANCE IN SOUTHERN CALIFORNIA

Scattering and absorption of seismic waves have been studied in southern California to determine their respective contributions to the measured apparent attenuation. We measured the total seismic energy from the onset of the S waves for local earthquakes recorded at stations on granitic rock-reduced to a common source by eliminating the source and site effect by the coda method, then applied Wu's method based on the radiative transfer theory, which has the facility to separate the intrinsic absorption from the scattering effect. The observed total energy as a function of hypocentral distance exhibits a striking behavior, increasing to a maximum at about 50 km, then decreasing exponentially as a function of hypocentral distance. Decay of the total energy as a function of distance is remarkably similar for all frequency bands, centered at 1.5, 3.0, 6.0 and 12.0 Hz, in good agreement with earlier results on Q of coda and S waves in southern California. Comparison with Wu's theoretical plots, which shows curves that differ in shape because of differences in seismic albedo, B0 = eta-s/(eta-s + eta-a), where eta-s and eta-a are the scattering and absorption coefficients, respectively, leads to the conclusion that the S-wave energy is primarily redistributed in space by scattering rather than intrinsically absorbed by the Earth. Our results from southern California agree with those of Toksoz et al., who found seismic albedo, B0, to be 0.8-0.9 in eastern North America. Coda Q in southern California, however, closely agrees with scattering Q using Wu's model, in disagreement with the conclusions of Toksoz et al. for eastern North America, where coda Q is close to the absorptive Q obtained from the total energy study. The agreement between Wu's theoretical curves and our observations is not perfect. Departure of the observed curves from the theory may be due in part to inhomogeneous distribution of scatterers and absorbers as well as non-isotropic scattering. We need more realistic models which can also account for inhomogeneous bounded media.