In recent years, a number of simultaneous advances in different areas of geophysics has led to a huge resurgence of interest in the topic. The most important of these were experimental results from mineral physics [ Ito and Takahashi, 1989; Ito et al., 1990; Akaogi and Ito, 1993] establishing the 660 km discontinuity as primarily a phase boundary with a strongly negative Clapeyron slope, and seismological observations of slabs [ Zhou and Anderson, 1989; Zhou and Clayton, 1990; van der Hilst et al., 1991; Fukao et al., 1992; Lay, 1994] showing that in many areas, e.g., the Japan, Southern Kurile, and Izu-Bonin arcs, subducted slabs do not immediately penetrate into the lower mantle, but are deflected at 660 km depth, whereas in other areas, e.g., the northern Kurile, Marianas, Tonga, and Java arcs [ Fischer et al., 1991], slabs appear to penetrate immediately.
Numerical modeling of phase change modulated mantle convection was revived in two influential papers [ Machetel and Weber, 1991; Peltier and Solheim, 1992], later expanded upon [ Solheim and Peltier, 1994], both of which modeled compressible mantle convection in axisymmetric spherical geometry, at high Rayleigh numbers and with realistic Clapeyron slopes. These studies found a pattern of intermittent layered convection, characterized by accumulation of cold downwelling material above the 660 km phase transition followed by catastrophic avalanches of the pooled cold material into the lower mantle, compatible with the intermittent layering found by Christensen and Yuen [1985]. Thus, it became clear that with realistic parameters, the transition zone is likely to have a major effect on the pattern of mantle flow.