Jl. Thirot et al., UPPER-MANTLE SEISMIC DISCONTINUITIES IN A SUBDUCTION ZONE (JAPAN) INVESTIGATED FROM P TO S CONVERTED WAVES, Physics of the earth and planetary interiors, 108(1), 1998, pp. 61-80
The lateral variations of the upper mantle discontinuities in a subduc
tion context are investigated by the use of broadband wave data provid
ed by the GEOSCOPE network at the Inuyama station (Japan). Differentia
l arrival times of P and P-to-S conversions on the discontinuities are
used. These converted phases, though of weak amplitude, can be observ
ed by stacking many seismograms. A stack of 50 data with high signal-t
o-noise ratio is performed by different techniques. The dominant perio
d is around 7-10 s in this study, and the lateral resolution is estima
ted to be of the order of 300 km. This is an intermediate scale betwee
n short wavelength studies, which can resolve very small details, and
long wavelength studies, for which the Fresnel zone is wider than abou
t 1000 km. The first stacking technique allows to retrieve the mean de
pth of the discontinuities below the station. The main result is that
the well-known discontinuity close to 660 km is depressed down to abou
t 700 km, while the 410-km discontinuity is observed close to its usua
l depth, and that a discontinuity close to 520 km is visible. However,
in such a heterogeneous region, the different traces are not coherent
, and it is useful to account for the variability of arrival times of
converted phases. Using alternative stacking techniques, we show that
the 410-km discontinuity is elevated in the slab area of about 60 km,
and the 660-km discontinuity is depressed by more than 50 km. This dep
ression of more than 50 km found for the 660-km discontinuity yield to
a value of -3 GPa K-1 of the Clapeyron slope for the phases change Ol
ivine-gamma-Magnesiowustite + Perovskite. Although the 520-km disconti
nuity clearly shows up with all the techniques used, it is difficult t
o reliably access its depth variation in the slab area. The evidence f
or a 520-km discontinuity in this period range, while it is not observ
ed at shorter periods, suggest that this discontinuity has a thickness
between 15 and 25 km. The large elevation of the 410-km discontinuity
suggests that the thermally controlled depth of the phase change may
affect the dynamic behaviour of the slab. (C) 1998 Elsevier Science B.
V. All rights reserved.