This research is aimed at the ongoing debate over whether or not slabs at subduction zones penetrate into the lower mantle below about 670 km depth. The technique will be to measure travel times from WWSSN and Canadian network seismograms and model them using a fully 3-D finite difference algorithm. The area to be studied will be the Kurile and Japan subduction zones. Use of only relative times of arrival at the same stations will reduce the influence of choice of a reference earth model, station delays, and long-wavelength lower mantle anomalies. The question of slab penetration below 670 km is key to formulating convection models of the earth's mantle.
