Compressional velocity in source regions of deep earthquakes: An application of the master earthquake technique

The technique of earthquake location relative to a master event is used to estimate near-source velocity and take-off angles for rays travelling to selected stations. Computations of a reconnaissance nature were carried out with arrival times of P and pP from deep earthquakes beneath the northwest corner of the Fiji plateau, the Peru-Brazil border region and the basin separating Fiji from the Tonga arc. These data yield estimates of compressional velocity of 11.2 ± 0.4, 11.4 ± 0.7 and 10.7 ± 0.3 km/sec respectively. Each of these velocities and the other parameters of each model space are essentially independent of their starting values. The corresponding depth ranges are 600–660, 580–650 and 540–600 km. These in-situ velocities are 5–10% higher than those of the Helmberger and Wiggins model. To account for such high velocities by a thermal effect alone would require an improbably high thermal contrast of 1000°C between “normal” mantle and the cooler earthquake zones. Spinels of proposed mantle composition would have compressional velocities of about 10.4 km/sec at temperatures that are taken as normal for these depths. If the high values of near-source velocity are explained by the addition of a post-spinel assemblage, then by implication this transformation occurs at shallower depths in those seismic zones than in the “normal” mantle.