Partial Melt Found Beneath Tectonic Plates, But It May Not Make Them Move

A team of seismologists has gone through 184,340 measurements taken at 716 permanent stations around the world, looking for a boundary roughly 150 kilometers down, about 93 miles. They found it beneath many of those stations. The signal marks a jump in seismic velocity with depth, the sort of change expected when waves leave a slower layer above. The researchers read that layer as rock holding a small amount of melt.

The work was published in Nature Geoscience by scientists from Brown University, the University of Texas at Austin and Cornell University. Junlin Hua began it as a Brown doctoral student and led it as a postdoctoral researcher at UT Austin. Karen Fischer, a Brown geophysicist and Hua's doctoral adviser, is a co-author, as is Thorsten Becker of UT Austin.

Some background frames the question. Tectonic plates are the crust and the rigid uppermost mantle, together called the lithosphere. Beneath them lies the asthenosphere, a weaker mantle region that lets the plates move relative to deeper rock. It is mostly solid, yet over geological time it can flow. Why it is so weak has been disputed. Partial melting looked like a plausible answer, since liquid between mineral grains can change how rock behaves. Earlier studies supported the presence of melt, but how much there is and how much it matters mechanically stayed contested.

The team went after the lower edge of the slow seismic region, which is harder to resolve than the upper edge. They used receiver functions, which isolate signals made when seismic waves meet changes inside Earth. Specifically, they tracked S waves converting into P waves at underground boundaries, and the timing and strength of those conversions gave boundary depth and the change in wave speed. Where both edges were detected, the slow layer had a median thickness of about 65 kilometers. Evidence for the boundary was strong at 160 stations and slight at another 240, with the rest less certain. That patchiness supports widespread melting rather than a continuous liquid shell.

The boundary appeared most often where other seismic models showed relatively hot upper mantle. It turned up beneath oceans and continents and well beyond volcanic hotspots. The team calculated how waves should behave under different temperatures, compositions and water contents, and models with partial melt reproduced the key features, with melt fractions around 1% or less depending on the model. Dropping stations near descending slabs left the main relationship largely intact, and oceanic and continental stations showed broadly similar patterns. Chemistry of basalts erupted worldwide also indicated melting beginning near the depth of the boundary. Extrapolating to areas with comparable mantle velocities, the researchers estimated the boundary could lie under about 44% of Earth's surface. That is a model-based figure, and since most stations sit on continents, limited ocean coverage could affect it.

The central test was whether the melt weakens the mantle. The team compared regions with and without the boundary using seismic anisotropy, the variation of wave speed with direction, which can reflect crystal alignment as rock deforms. Mantle-flow models in one and three dimensions predicted that a strong drop in viscosity would leave a distinct anisotropy signature. It did not appear. Regions with the boundary had lower velocities, but their anisotropy profiles were not significantly different from those without it. That argues against the melt sharply reducing long-term viscosity.

Timescale helps explain this. Earthquake waves probe how rock responds over seconds, while plate motion reflects deformation over vastly longer periods. Melt can slow a wave without being the main control on slow flow. The authors point instead to gradual variations in temperature and pressure as primary controls on the asthenosphere's weakness.

The study stops short of calling melt irrelevant everywhere; Becker noted local effects could still matter. It also does not say when plate tectonics began, since present-day mantle properties cannot settle that. The open task is to explain the asthenosphere's weakness while reconciling what earthquake waves record with what long-term deformation shows.