Mantle layering from ScS reverberations: 2. The transition zone


  • Justin Revenaugh,

  • Thomas H. Jordan


This is the second paper in a four-part sequence that examines the nature of mantle layering using the ScSn phases and internal reflections observed within the reverberative interval of SH-polarized seismograms. Mantle reflectivity profiles for 18 seismic corridors sampling portions of Indonesia, Australia, and the western and central Pacific are constructed from low-frequency (25-mHz) ScS reverberations; i.e, ScSn-type phases reflected once from internal mantle discontinuities. Modeling the reflectivity profiles using synthetic seismograms yields precise, path-specific estimates of the travel times to the 410-km and 660-km discontinuities and their reflection coefficients. The data are consistent with an olivine-dominated model of the transition zone, marked by two major phase transitions having average apparent depths of 414±2 km and 660±2 km and impedance contrasts of 0.046±0.010 and 0.072±0.010, respectively. The interpath variations observed in the reverberation travel times imply that the topography on the two discontinuities is of the order of 20 km (peak to peak) and is negatively correlated, consistent with an endothermic transition at 660 km whose Clapeyron slope magnitude is comparable to that of the 410-km (exothermic) transition. In the case of the 660-km discontinuity, we have been able to supplement these long-wavelength (1500–5000 km) observations with an estimate of intrapath topography, inferred from the frequency dependence of the 660-km reflection coefficient, which gives peak-to-peak variations less than 40 km on scale lengths of 500–1500 km. For the 660-km discontinuity, the observed topography is significantly less than the dynamic topography expected for a simple compositional interface in regions of subduction. However, we do observed an intriguing negative correlation between the apparent depth and the reflection coefficient of the 660-km discontinuity which may involve small compositional variations. Plausible explanations for this correlation include heating of an initially cool chemical boundary layer gravitationally trapped above an endothermic phase transition, loss of reflected energy due to local curvature (or roughness) of the reflector, and/or extreme topography occurring on a small compositional component to the discontinuity (≤40% of the total impedance contrast). By stripping the 410-km and 660-km peaks from the composite path reflectivity profile, we identify three minor reflectors at mean depths of 520, 710, and 900 km. The shallowest may mark the β-phase → γ-spinel phase transition. The 710-km discontinuity can be attributed to the ilmenite (gamet) → perovskite phase transition, although changes in perovskite symmetry can potentially provide explanations for both the 710-km and 900-km features.