January 6, 1997

Roots of "hot spots" may extend to earth's core-mantle boundary

"Hot spots," the isolated patches of volcanism unrelated to plate tectonics, may spring from surprisingly deep within the planet: the turbulent boundary between earth's mantle and its core.

That conclusion, sure to put scientists on the spot at the AGU meeting, has arisen from intense study of a layer at the base of the mantle that apparently contains partially molten rock. Researchers at UCSC and elsewhere analyzed a peculiar type of seismic wave that skims along the mantle side of the sharp core-mantle boundary. In several regions, something bogs down the speed of the waves by about 10 percent--a huge amount by geophysical standards. The most likely cause, the researchers maintain, is that a small fraction of melted material bathes the mantle rock and transforms it into a thick mush.

Seismologist Edward Garnero of UCSC has worked with Donald Helmberger of Caltech and UCSC's Justin Revenaugh to characterize the layer. So far, the seismologists have probed about 45 percent of the core-mantle boundary in search of the partial melt. Temperature constraints at the boundary would require that if the layer exists anywhere, it would encircle the globe. However, the layer appears substantially thicker in some regions--most notably under the south-central Pacific, Iceland, East Africa, and the Azores. The seismic evidence thus far points to a layer of partial melt no thicker than about 20 kilometers in those zones. Elsewhere, it would be much thinner--5 kilometers or less--and would evade seismic detection entirely.

These first few swaths of partial melt, notes UCSC mineral physicist Quentin Williams, lie 2,900 kilometers beneath some of the most well-known hot spots on the planet. That correlation is not yet perfect, but it is no coincidence, he believes. "This may be the smoking gun that hot spots originate from the core-mantle boundary," Williams says. Although others have proposed such a deep genesis for hot spots, no one has yet unveiled solid evidence to support that theory.

Williams envisions a partially molten layer that becomes unstable at irregular intervals above the core-mantle boundary, welling up into thicker, stubby plumes. These zones would help heat flow with great dispatch out of the liquid outer core and into the lower mantle. "This may be the first image we have of a process that spans the mantle," Williams says. "Feeder zones of plumes that start at the base of the mantle would rise all the way to the surface and create hot spots."

This line of reasoning has several fascinating implications. For instance, if there is indeed a partially molten layer at the base of the mantle, it would be about a trillion times less viscous than the solid mantle immediately above. "That's geodynamically very interesting," Williams observes. Further, the scenario raises a "chicken and egg" conundrum: Would upwellings in the partially molten layer trigger and sustain the hot spots, or would hot spots draw material up through the mantle and cause the upwellings? Stay tuned as the debate heats up.

--Robert Irion

Return to the Currents home page.

Go to UCSC's home page.