Abstract
For the same volume of rock, I compare the attenuation obtained by seismic pulse transmission over the frequency range 1-150 kHz with that obtained by vibrational resonance techniques over the frequency range 1-50 kHz. The initial studies were performed on a rectangular block of medium-grained granite which was large enough to permit the installation of a seismic pulse transmission array over a 1.8 m path length, yet small enough to permit whole-body resonance.A Q of 82, for the P wave, was derived from the vibrational resonance results, whereas a Q of 15 was derived from the pulse transmission results. In light of models proposed for the viscoelastic, geometric, and elastic scattering attenuation mechanisms, the experimental results suggest that this large discrepancy in Q values is due to elastic scattering by grain clusters (rather than individual grains) within the granite. Scattering is significant in the high-frequency pulse transmission tests, but is considered insignificant in the lower frequency resonance tests. Furthermore, this scattering is represented approximately by a constant-Q loss mechanism, which makes it difficult to separate unambiguously elastic scattering and viscoelastic losses.Subsequent studies performed on a large block of fine-grained norite yield a resonance Q of 89 and a pulse Q of approximately 102 and suggest a negligible scattering loss for this material. The experimental results for the norite imply that the constant-Q theory of seismic pulse attenuation provides a reasonable description of wave attenuation in a dry, fine-grained crystalline rock over the frequency range 1-150 kHz.
