Preliminary evidence for low S-velocity in the D" under the south Pacific superswell

S-wave heterogeneity in the D" region under the South Pacific Superswell is examined using SKKS-SKS differential travel times observed by a temporary broad-band seismometer networks of the SPASE (Wiens, 1995) and a new permanent network of the SPANET (Ishida et al., 2000). Seismograms of deep earthquakes in the south America were collected. SKKS-SKS differential travel times were measured by the cross-correlation between band-passed (10 s to 100 s) waveforms of SKS and SKKS on the radial component after correction of phase shift and polarity reverse. The residuals of the differential travel times were calculated with respect to PREM. The SKKS rays pass across the core-mantle boundary (CMB) under the receiver side distribute evenly around the central point of (30\deg S, 150\deg W) within 500 km radius, whereas those under the source side concentrate around the two points of (20\deg S, 100\deg W) and (35\deg S, 90\deg W). The SKKS-SKS residuals vary from 1 s to 6 s. This variation is reflected mainly the lateral variation of S-velocity under the receiver-side because that ray paths of SKS and SKKS under the source side are very close each other with respect to the wavelength. The SKKS-SKS residuals, of which SKKS rays pass across the CMB in west of 150\deg W longitude lines, are around 2s consisted with prediction from 3D S-wave heterogeneity models of SKS12WM13 by Dziewonski et al. [1997] and S16U6L8 by Liu and Dziewonski [1998]. However, the SKKS-SKS residuals of which SKKS rays pass the CMB in east of 150\deg W longitude lines are approximately 4 s, and some ones reach to 6 s. They are much larger than those predicted by the above models. When the S16U6L8 are considered, the residuals of 3 to 5 s are remained. These SKKS passing points with the large positive residuals are distributed beneath the South Pacific superswell including the hot spots of Tahiti and Macdonald. S-wave heterogeneity models by SKS12WM13 and S16U6L8, SB4L18[Masters et al., 2000] and SAW24B16[Megnin and Romanowicz, 2000] do not show an extreme of a low-velocity anomaly in the D" region under this region, this is probably due to the lack of sampling. A new data set suggests strongly the existence of a very slow anomaly in D" region under the South Pacific superswell.