We have developped an a priori model of the upper mantle (Nataf H-C. and Y. Ricard, 3SMAC : an a priori tomographic model of the upper mantle based on geophysical modeling, Phys. Earth Planet. Inter., 95, 101-122, 1996).
Several well-known tomographic studies published in the 1990's (van der Hilst et al., 1991; Fukao et al., 1992) captured a large-scale subhorizontal high velocity anomaly at the bottom of the upper mantle in the northwestern Pacific. These models suggest that a large volume of subducted slab is stagnant in the transition zone.
The development and interpretation of seismic mantle tomographic results has usually proceeded under the assumption that fast and slow velocity anomalies reflect a spatially heterogeneous temperature field controlling, or being controlled by, mantle convection. Implicit in this approach is an assumption that the effect of anisotropy on seismic velocities is small in comparison with isotropic thermal or compositional effects, or that the tomographic results represent the average isotropic heterogeneity, even if individual seismic observations are affected by anisotropic structure. Velocity anomalies in the oceanic mantle are commonly interpreted as reflections of the progressive cooling (and localized re-heating) of a mechanical and thermal boundary layer consisting of the rigid oceanic lithosphere and the underlying less viscous asthenosphere. We show here that the interpretation of seismic velocity anomalies is considerably more complicated for the mantle beneath the central Pacific: in a broad area, with its center near Hawaii, seismic data reveal a regional anomaly in elastic anisotropy which produces variations of seismic velocities that are at least as large as those due to thermal effects. Seismic anisotropy is an independent indicator of strain in Earth materials, and our new tomographic results therefore provide constraints on both the buoyancy forces (thermal effects) and flow patterns in the mantle.