1971 年 24 巻 3 号 p. 228-239
A deep earthquake occurred in the SW part of Fukui Prefecture on 14 August 1967. This shock produced a clear longitudinal phase about 8 sec before the normal ScS phase at stations URA and KMU, Hokkaido. On careful examination of seismograms of other nearby deep earthquakes at KMU, it was found that similar forerunners were recorded on several seismograms. These forerunners may not be interpreted as PKiKP phases because of their time differences. A possible interpretation is the conversion from ScS to ScSp at some plane of discontinuity in the upper mantle. Assuming the vertical incidence of ScS waves and a velocity contrast at the discontinuity (the upper layer has a velocity 10% smaller than that of the lower layer), a region in which the conversion plane may be located is determined from the ScSp incident angle, the time difference, and the amplitude ratio. The conversion plane has a dip of about 35°-39° and is nearly parallel to the deep seismic zone. The conversion plane is located near the upper boundary of the deep seismic zone and may coincide with the boundary between the dipping high-Q, high-V zone and the landside low-Q, low-V zone. It is interesting that the seismic activity is low under this conversion plane, while the surface volcanism is present above this plane.