Seismological constraints on the gabbro-eclogite transition in subducted oceanic crust

Abstract

Dispersion of seismic waves travelling along the strike of the subducted Pacific plate beneath Tonga-Kermadec shows evidence for a high-velocity layer of crustal thickness extending to a depth of at least 450 km. The opposite dispersion, with low frequencies arriving first, is observed in Japan and has been taken, together with inferences from converted phases, as evidence of a thin, low-velocity layer. Both layers have, in previous studies, been interpreted as subducted oceanic crust which has been transformed to eclogite, but in Tonga-Kermadec the eclogite must have high seismic velocity and beneath Japan it must have a low seismic velocity: an apparent contradiction therefore exists. Here we find theoretical estimates of the seismic velocities of eclogite and normal mantle using the data and equations of Duffy and Anderson and a variety of mineral assemblages and bulk compositions. We find a range of both low and high seismic velocities; the high end of the range depending mainly on the contents of jadeite, pyrope and grossular. It is therefore possible, in principle, to explain both the high-velocity layer beneath Tonga-Kermadec and the low-velocity layer beneath Japan as being consistent with an eclogitic crust, provided they have significantly different compositions. However, we can find no clear reason for such a difference between that part of the Pacific Plate currently subducting beneath Japan and that beneath Tonga-Kermadec.