Map showing the location of the deep-mantle structures -- ultra-low velocity zones’ or ULVZs. Credit: James Atkins.
Jules Verne’s Journey to the Centre of the Earth remains the stuff of science fiction. Scientists will probably never be able to directly sample deep mantle rocks, but they are finding increasingly sophisticated ways to examine Earth’s interior.
A new map created by Cambridge earth scientists reveals a hidden landscape of broad, flattened mountains rising from the core-mantle boundary, each measuring up to hundreds of kilometres across and tens of kilometres high.
It’s the first map to capture the global distribution of these ‘ultra-low velocity zones’ or ULVZs, so-called because seismic waves released by earthquakes slow down dramatically when they encounter the dense, iron-rich zones. The findings show a link between the location of ULVZ and volcanic hotspots at the surface.
“The features are an enigma,” said James Atkins, lead author of the study and PhD student at Cambridge’s Department of Earth Sciences. “Scientists have lots of questions about ULVZs, their origin, composition and function. With our global map, we are a step closer to understanding them.”
Revealing a hidden landscape
The new map reveals Earth’s interior in a similar way to a medical CT scan. But rather than using X-rays to see hidden features, the technique relies on seismic waves. When an earthquake happens, seismic waves spread outward, taking a curving path through Earth’s interior before emerging at distant seismic stations on the other side of the planet.
“When seismic waves strike a patch of denser material, such as a ULVZ, they scatter – just like light passing through a lens,” explained Sanne Cottaar, senior author of the study, who leads the Deep Earth Explorers Group in Cambridge.
Larger, deep-focus earthquakes are better for spotting ULVZs, Cottaar explained, because their seismic waves penetrate deeper into the mantle. “The waves skim across the core-mantle boundary, giving us a good coverage.”
In total, Atkins examined 8,000 earthquakes going back as far as 1995, spending months visually unpicking millions of signals captured by seismic stations. For each earthquake, he measured the interval between the arrival of the first pulse of seismic waves (those caused by the main earthquake event), and the delayed secondary waves that have been slowed and scattered. “This delay time tells us the dimensions of the ULVZ,” said Atkins. “The larger the feature the longer the delay.”
Seismic waves released by an example Indonesian earthquake reach seismic stations in North America in two pulses (orange = first arrival, red = delayed waves).
Each of the millions of traces that Atkins studied traced a path through Earth’s interior, giving global coverage. Where these transects intersect, Atkins was able to pinpoint the location of each ULVZ. In total, he identified six main ULVZs around the globe, lying on the core-mantle boundary in vicinity of Hawaii, Iceland and Galapagos, for instance.
When Atkins compared the map of ULVZs with features at Earth’s surface, he found that they sat beneath unusual volcanic hotspots found in the middle of tectonic plates and far from the plate boundaries where volcanism typically occurs. Volcanic hotspots, which include Hawaii, are fed by an upwelling of magma rising from deep within Earth's mantle.
One long-standing theory is that ULVZs might be the deep-seated roots of the upwellings causing volcanic hotspots. “Each ULVZ we mapped has a volcanic hotspot lying above it,” said Atkins. “This affirms the theory that they must be some kind of feeder for surface volcanism.”
Illustration of Earth's interior, showing how ULVZ may feed hotspot volcanoes. Credit: Jenny Jenkins / Amalgam for Sedgwick Museum exhibit.
But the debate is still far from settled. One problem is that the number of hotspots exceeds the number of ULVZs (there may be as many as 50 hotspots globally). One explanation for this could be that some ULVZs may anchor an upwelling plume with a branching structure capable of feeding multiple hotspots, Atkins explained.
It’s also possible that the map might be missing smaller ULVZs that pepper the core-mantle boundary, said Cottaar. This is because their analysis used a specific seismic wave type and wavelength that may be less responsive to smaller features.
“But, at least for the bigger ULVZs, the link to hotspots is becoming clearer,” said Cottaar, who admits their work poses many unknowns. “Our findings really do beg the question – are all these zones the same thing?”
“If we tried to map Earth’s surface from thousands of kilometres away, then a sedimentary basin might look the same as a volcanic island. In the same way, our mapping likely misses some of the diversity in the deep mantle.”
Cottaar said that she and the Deep Explorers Team are now planning to build upon their map by using fast-moving seismic P-waves capable of travelling through liquid as well as solid. P-waves might identify different types of ULVZ and be able to establish if they contain molten material – helping further understand their composition and function.
“Ultimately, the only way we will ever know is by combining lines of evidence – by looking for clues not just in the seismic data but the geochemistry of hotspot rocks as well as Earth’s geomagnetic field.”
Reference: Atkins, J. R., Martin, C., & Cottaar, S. (2026). Global presence and absence of ultra‐low velocity zones as seen by Sdiff postcursors. Journal of Geophysical Research: Solid Earth, 131(4), e2025JB033171.