During the months of August 1984 and October 1985, experiments were undertaken in the Charlevoix Seismic Zone to record high-frequency body waves from earthquakes. The seismograms were recorded by digital three-component seismographs. The experiments were conducted on the north shore of the St. Lawrence River only and covered nine sites. All but the two most northeasterly recording sites exhibited P-wave arrivals that arrived from azimuths that differed from the theoretical azimuth based on the computed hypocenter-station relationship. These differences amounted to several tens of degrees. Neither the errors in hypocenter location nor in the orientation of the horizontal seismometers are believed to be the cause of this difference.
Theoretical computations based on a two-dimensional model with strong lateral velocity gradients or a discontinuity in velocity produce deviations of the direction of arrival of the P wave in a horizontal plane normal to the anomaly that can amount to several tens of degrees. Both the lateral gradient and discontinuity models satisfy the observed deviation of azimuth versus azimuth deflection. However, only the discontinuity satisfies the deflection versus hypocentral distance distribution, at least for distances to 20 km.
Comparing the observed data with the theoretical values suggests that the anomaly in velocity occurs in a near-vertical plane striking parallel to the St. Lawrence River. This is also the direction of ancient extension faults in the region. Other structures were introduced 350 m.y. ago when the Charlevoix impact structure was formed. A number of anorthosite bodies with very high seismic velocities also outcrop in the region. Because of the high frequencies involved, the scale of the anomaly may well be only a few hundred meters, and the anomaly may most likely be located very close to, even directly beneath, the station.