Crampin emphasizes two important points: first, surface waves in an anisotropic media, will, in general, possess three dimensional particle motion; and second, a rigorous means of inverting data concerning anisotropy is needed. However, surface waves which have travelled oceanic paths tend to have predominantly Love-type or Rayleigh-type particle motion and clear cases of three-dimensional motion are difficult to find. This may be because the observing stations are located on continents rather than within the anisotropic region, or because transverse motion associated with a wave which is predominantly Rayleigh-type may be misinterpreted as a component of radial motion in a direction deviating slightly from the great circle path to the epicentre. For a few paths in the Pacific, I have observed vertical particle motion coupled to short-period (6-8 s) waves of predominantly Love type, similar to the observation in Eurasia reported by Crampin (1967), but the primary evidence for anisotropy comes from the azimuthal variation of velocity and the inconsistency of Love and Rayleigh velocities. Only if no isotropic structure can be found which fits both Love and Rayleigh wave dispersion simultaneously should an anisotropic structure be considered. My inversion of the data in terms of SV and SH velocities was intended primarily to indicate the degree to which the observations are inconsistent with isotropic structures and was obviously not intended to be a complete description of the 21 independent elastic constants. Before rigorously tackling the inverse problem, a solution is needed to the forward problem of surface wave propagation in a spherical, gravitating earth with a liquid layer. To the best of my knowledge, the general problem has been solved only for a solid, flat-layered earth (Crampin 1970). Perhaps the most fruitful approach to inversion will be some combination of Smith & Dahlen’s (1973) perturbation technique with the forward approacb used by Crampin.
Read full abstract