Earth’s mantle is divided into three layers- an upper region that extends to about 410 km below Earth’s surface, a mantle transition zone (MTZ) from about 410 km to 660 km depth, and a lower region that extends to the core–mantle boundary. The top and bottom boundaries of the MTZ are coincident with mineralogical phase transitions caused by the high temperature and pressure of Earth’s interior. Seismic S and P waves bounce off those discontinuities, as illustrated in the figure, and researchers use detections of those waves to glean information about Earth’s interior properties. But turning seismic observations into temperature and composition estimates is difficult; models also incorporate mineralogical predictions, and seismological data sets have gaps and uncertainties.

Underside reflections of S and P waves from the 410 km and 660 km discontinuities provide near-global coverage of the MTZ boundaries. The arrival of reflected waves from both discontinuities precedes the arrival of primary peaks in S-wave signals (the top waveform in the figure). But the 660 km boundary rarely appears in P-wave signals (the bottom waveform). The missing reflection has led to debate about the lower discontinuity’s thermal and compositional nature and how it affects mantle convection and dynamics.

By combining new data-mining techniques, mineral-physics modeling, and a database of more than 100 000 seismic observations, Lauren Waszek at James Cook University in Douglas, Australia, and her collaborators have obtained a thermal model for the MTZ. Their results reveal that just four areas—three in the Pacific Ocean and one at the North Pole, which together constitute 0.6% of the globe—produce visible P-wave reflections. Those areas are more likely to correspond to unusually high MTZ-boundary temperatures in excess of 2100 K.

Such high temperatures stabilize garnet-rich mineral assemblages that are otherwise absent in the MTZ. Using synthetic S and P waveforms predicted for a range of compositions and temperatures, the researchers found that those materials determine whether P-wave reflections are visible. A statistical analysis showed that an unequilibrated mixture of two chemical components, basalt and harzburgite, that were introduced to the mantle through subduction provided a more consistent description of the bulk mantle than a well-mixed chemically homogeneous composition.

The stability of garnet at very high temperatures results in efficient material transport up through the 660 km boundary, so Waszek and collaborators conclude that such enhanced flux occurs in only the few small regions that produce detectable P-wave reflections. In cooler areas that make up most of the MTZ, another mineralogical reaction—the breakdown of ringwoodite into bridgmanite—dominates. But that lower-temperature process impedes upward flow and causes hot upwelling material to pool near the 660 km boundary. Such plume stagnation would lead to heterogeneous semilayered convection and poor mixing in the MTZ. (L. Waszek et al., Nat. Geosci., 2021, doi-10.1038/s41561-021-00850-w.)