Earth’s mantle may be hotter than thought

Temperatures across Earth’s mantle are about 60 degrees Celsius higher than previously thought, a new experiment suggests. Such toasty temperatures would make the mantle runnier than earlier research suggested, a development that could help explain the details of how tectonic plates glide on top of the mantle, geophysicists report in the March 3 Science.

“Scientists have been arguing over the mantle temperature for decades,” says study coauthor Emily Sarafian, a geophysicist at the Woods Hole Oceanographic Institution in Massachusetts and at MIT. “Scientists will argue over 10 degree changes, so changing it by 60 degrees is quite a large jump.”
The mostly solid mantle sits between Earth’s crust and core and makes up around 84 percent of Earth’s volume. Heat from the mantle fuels volcanic eruptions and drives plate tectonics, but taking the mantle’s temperature is trickier than dropping a thermometer down a hole.

Scientists know from the paths of earthquake waves and from measures of how electrical charge moves through Earth that a boundary in the mantle exists a few dozen kilometers below Earth’s surface. Above that boundary, mantle rock can begin melting on its way up to the surface. By mimicking the extreme conditions in the deep Earth — squeezing and heating bits of mantle that erupt from undersea volcanoes or similar rocks synthesized in the lab — scientist can also determine the melting temperature of mantle rock. Using these two facts, scientists have estimated that temperatures at the boundary depth below Earth’s oceans are around 1314° C to 1464° C when adjusted to surface pressure.

But the presence of water in the collected mantle bits, primarily peridotite rock, which makes up much of the upper mantle, has caused problems for researchers’ calculations. Water can drastically lower the melting point of peridotite, but researchers can’t prevent the water content from changing over time. In previous experiments, scientists tried to completely dry peridotite samples and then manually correct for measured mantle water levels in their calculations. The scientists, however, couldn’t tell for sure if the samples were water-free.

The measurement difficulties stem from the fact that peridotite is a mix of the minerals olivine and pyroxene, and the mineral grains are too small to experiment with individually. Sarafian and colleagues overcame this challenge by inserting spheres of pure olivine large enough to study into synthetic peridotite samples. These spheres exchanged water with the surrounding peridotite until they had the same dampness, and so could be used for water content measurements.

Using this technique, the researchers found that the “dry” peridotite used in previous experiments wasn’t dry at all. In fact, the water content was spot on for the actual wetness of the mantle. “By assuming the samples are dry, then correcting for mantle water content, you’re actually overcorrecting,” Sarafian says.
The new experiment suggests that, if adjusted to surface pressure, the mantle under the eastern Pacific Ocean where two tectonic plates diverge, for example, would be around 1410°, up from 1350°. A hotter mantle is less viscous and more malleable, Sarafian says. Scientists have long been puzzled about some of the specifics of plate tectonics, such as to what extent the mantle resists the movement of the overlying plate. That resistance depends in part on the mix of rock, temperature and how melted the rock is at the boundary between the two layers (SN: 3/7/15, p. 6). This new knowledge could give researchers more accurate information on those details.

The revised temperature is only for the melting boundary in the mantle, so “it’s not the full story,” notes Caltech geologist Paul Asimow, who wrote a perspective on the research in the same issue of Science. He agrees that the team’s work provides a higher and more accurate estimate of that adjusted temperature, but he doesn’t think the researchers should assume temperatures elsewhere in the mantle would be boosted by a similar amount. “I’m not so sure about that,” he says. “We need further testing of mantle temperatures.”