Changes in the crust and upper mantle near the Japan‐Bonin Trench

Abstract

Depths and reflection times to mantle have been computed in the west Pacific from 60 sonobuoy refraction solutions, many of which could be compared with observed mantle reflection depths from multi‐channel data obtained at the same time. After repicking some of these sonobuoy records, all were eventually adjusted to agree within 0.05 s with the observed mantle reflection times. This added constraint produces solutions that are clearly more reliable. Crustal velocities (exclusive of water and sediment) from the study area are rather tightly distributed about a mean value of 6.53 km/s with a standard deviation of only 0.23 km/s (n = 47). Results show that the crust thickens in a westerly direction from the west Pacific basin, where mantle depths are 11–11.5 km to a belt 200 km east of the Japan trench, coinciding with the outer gravity high, where mantle is at an average depth of 14 km. Several sonobuoys in the zone of maximum crustal thickness just east of the outer slope of the Japan trench record two deep reflectors about 0.6 s apart in the vicinity of the upper mantle. Two values of interval velocity obtained from a reduced T2/X2 analysis of the layer bounded by these reflectors are 7.5 and 7.2 km/s. These sonobuoys and a few others with weaker double reflections are all located within the outer gravity high. To the south a well‐observed mantle reflection and its strong 8.2‐km/s refraction disappear from our records just as the crust begins its descent into the Bonin trench. Within the outer trench slope a 7.3‐km/s refractor, which is a weak arrival elsewhere, becomes the dominant refractor. The peculiar double reflector near mantle and the marked change in velocity structure and upper mantle reflectivity at the edge of the outer slopes seem to confirm the changes in upper mantle refraction velocity reported by Talwani et al. (1977) in the Curacao trench, Caribbean Sea, but our evidence is not conclusive. In spite of the ambiguity it is clear that velocity structures near active trenches cannot be confidently defined by correlating sections with similar velocities.