Mapping source of volcanic activity deep within the Earth

When geophysicist Xiyuan Bao first moved to Cambridge, he often relied on navigation software to find his way through the unfamiliar network of roads. One day as he sat behind the wheel, he realized the same technology could help him find something else: a solution to a mystery deep within the Earth.

For more than half a century, geophysicists have struggled to understand enigmatic structures called mantle plumes — upwellings of hot rock that produce volcanic “hotspots” such as the Hawaiian Islands or Iceland. Located far from plate boundaries, these hotspots are believed to originate thousands of kilometers beneath the surface. Although these subterranean structures have been mapped in detail, there has been no reliable way to trace their paths to the surface — until now.

In a recent paper in Science, Bao and his colleagues show how these structures can be mapped by graph theory — the same algorithms that power modern navigation systems.

“This is the first joint geophysical-geochemical zoning map of the lowermost mantle,” said Bao, Reginald A. Daly Postdoctoral Research Fellow in the Department of Earth and Planetary Sciences. “In effect, the navigation map identifies broad regions deep inside Earth with both an address and a chemical identity.”

Mystery runs deep

Most volcanoes erupt near the boundaries of tectonic plates. Sometimes, however, they emerge from the middle of the plates — a phenomenon attributed to pillars of hot rock rising from deep within the Earth to punch through the overlying material like red-hot probes.

The theory of mantle plumes was first proposed in the early 1960s to explain the emergence of the Hawaiian Islands. In 1971, geophysicist W. Jason Morgan of Princeton — a pioneer of plate tectonics — proposed that hotspots resulted from plumes of hot mantle material originating just above the Earth’s core. (After retiring from Princeton, Morgan became a visiting scholar at Harvard until his death in 2023.)

Now scientists have developed more detailed maps of these structures up to 3,000 kilometers (1,864 miles) below the surface — nearly halfway to the center of the Earth — just above the boundary between the predominately liquid iron outer core and solid rock mantle.

For years, researchers assumed that plumes traveled straight upward. More recently, seismic tomography — which uses the vibrations of earthquakes to image the interior of the Earth, like CT scans in medicine — has revealed that these structures follow irregular paths with odd turns, subterranean ponds, anomalous shapes, and links to other plumes in interconnected networks.

Convection occurs like it does in a pot of boiling water: hot material rises and cold material falls. Because the mantle is solid rock, this process happens very slowly over millions of years, and the material only becomes molten as it nears the surface.

Until now, there has been no reliable way to connect these points of origin with their ultimate destinations — in other words, to trace the plausible routes those materials took before they erupted onto the surface.

Enter Bao.

A driven scientist

Bao has conducted research on mantle plumes since he was a Ph.D. student at the University of California, Los Angeles. In 2022, he authored another Science paper on the relative temperatures of these hotspots. Trained as a geophysicist, he has enjoyed experimenting with techniques borrowed from other disciplines.

“I just get obsessed with so many methods,” he said with a smile. “I’m not limited to any traditional or existing method. I just invent methods on the fly or adopt methods from other fields.”

When he moved to Harvard two years ago for a postdoctoral fellowship, Bao found himself stressed by the Boston area’s unfamiliar network of roads and tunnels — not to mention the notoriously impatient local drivers. In a single day, he was honked at by six separate motorists. “In Southern California, I got the same amount of horn in six years,” he recalled with a chuckle.

To stay out of trouble, he relied on his navigation system. “One day I clicked the Google Maps, and I realized this is just a different type of plume — like a two-dimensional version of it,” he said.

He adapted the algorithms for the three-dimensional problem of mantle plumes and applied them to a region with good coverage by seismic tomography — East Africa and the Indian Ocean. He focused on eight hotspots, two on land and six in the sea.

In a navigation algorithm, a map is turned into a series of nodes connected by roads. Each segment is assigned a weight such as the distance divided by the speed limit. The fastest route is the one with the lowest sum of weights. Similarly, mantle plumes can be tracked from their origins to their destinations, but the weighting involves distance and seismic velocity.

In Bao’s method, a series of path options are ranked and grouped with a technique called cluster analysis. Bao tested these groupings by comparing them with a geochemical analysis of rocks produced by the plumes. He enlisted his colleague, Andreas Stracke from Germany, to verify the geochemical analysis — and sure enough, the results matched his own.

They found rocks from the eight hotspots fell into three groups, likely an indication of shared origins: one linked East Africa and the Comoros islands, a second for Réunion island, and a third for several hotspots in the Indian Ocean. The study produced the first-ever map that combines locations with chemical signatures of mantle plumes. The results suggest that these structures form what Bao described as a “mosaic of at least three compositionally distinct source regions.”

Rock of ages

Bao’s work reveals the complexity of mantle plume networks. A single hotspot may trace its origins to multiple sources. A mantle plume may produce multiple hotspots.

It also suggests the unusual dynamics of mantle plumes are not driven solely by hot temperatures. Indeed, these structures have sharp boundaries, suggesting they are composed of different, denser materials than the surrounding lower mantle.

Bao said these results may help illuminate the deep history of the Earth.

“Rocks that reside in the mantle have circulated for billions of years,” he said. “If large chemical differences exist at depth, it suggests that these regions have never been completely mixed. The new map gives scientists a framework for asking whether these hidden regions preserve material from Earth’s early separation into layers, later recycling of crust, or some combination of the two.”

(Varied hypotheses have been suggested to explain the origins of these structures: the collision of young Earth with another planet, the accumulation of dense material over time, or the separation of material soon after formation of the Earth.)

The study also offers intriguing clues about previously unknown geography deep within the Earth. Many mantle plumes bend or merge about 1,000 kilometers beneath the surface, and these features may reveal the locations of hidden geography and obstacles.

“Xiyuan’s paper is transformative in the study of mantle dynamics and long-term Earth system evolution,” said Jerry Mitrovica, Frank B. Baird, Jr. Professor of Science. “It’s a beautiful paper, and it will change the way we think of the long-term evolution of the mantle and plate tectonics — all within a completely innovative, modern spin of graph theory.”

Bao’s paper demonstrates the technique in only one region, but he says the method could be applied to the entire globe.

“This provides a new tool to obtain a full map of the mantle,” he said. “You can extend it to whatever depth you want.”