Thirty-four teleseismic events, recorded on the deep horizontal sediments of central Alberta, using one fixed and one movable station, have been analyzed as a test of the Haskell matrix formulation applied to short period P waves. Only limited agreement is obtained between averaged experimental vertical-horizontal spectral ratio curves and those calculated theoretically using known layer thicknesses and velocities.
Scattering in the crust and upper mantle is indicated by large transverse amplitudes including distict phases and by lower coherency for smaller epicentral distances where the travel path is confined to the crust and upper mantle. Anomalous SV/P ratios are believed to contribute to the difficulties. A study of 20 events in the azimuth range 285° to 310° indicates an apparent azimuth approximately 18° more northerly than the true azimuth. Localized dips of approximately 15° on the crustal boundaries are required to explain this deviation.
It is concluded that this region for which the sediments are horizontally layered does not fulfill the requirements of the Haskell matrix theory due to scattering and anomalous PS conversions in the crust and upper mantle.