Multicomponent processing of channel waves for anisotropy

Combining information derived from mode polarization and dispersion may help to limit inherent non-uniquenesses when interpreting channel wave data in anisotropic waveguides. However, estimates of these attributes require careful processing and cannot be fully understood by direct inspection of the time-series. A multicomponent analysis method is developed to separate these frequency-dependent attributes when data containing no more than a few overlapping modes is excited by several source directions. The method, a complex similarity transformation, may be viewed as the channel wave equivalent of real rotation algorithms currently employed for use with body waves. The analysis is illustrated by application to field data from an in-seam seismic survey in the U . K . displaying I dominant second generalized mode (Love-type motion). A favourable comparison is made between the observed and synthetic eigenmode dispersion, and the polarization ellipticity, sense of rotation and orientation.