Since late 2023, the Reykjanes Peninsula in southwest Iceland has erupted repeatedly, forcing evacuations from the fishing town of Grindavik and sending rivers of lava across roads and geothermal infrastructure. The eruptions are dramatic on the surface, but the force behind them originates far deeper than most people realize. A growing body of research, capped by a 2026 geochemical study published in Nature Communications, has now traced the heat source all the way down to the boundary between Earth’s mantle and its outer core, roughly 1,800 miles beneath the island. It is the most completely mapped mantle plume on the planet, a column of abnormally hot rock that rises from the deepest accessible layer of Earth’s interior to feed the volcanoes above.
A slow-motion geyser rooted at Earth’s core
The story begins at a depth of about 2,900 kilometers, where the rocky mantle meets the liquid iron of the outer core. Seismic waves generated by distant earthquakes slow sharply when they pass through a patch of material directly beneath Iceland, a feature geophysicists call an ultra-low-velocity zone, or ULVZ. That slowdown indicates rock that is either extremely hot, partially molten, or chemically different from the surrounding mantle. The linked 1997 study by Williams and Garnero was among the foundational papers establishing ULVZs as a recognized feature of the core-mantle boundary; subsequent decades of research have built on that framework. In Iceland’s case, the ULVZ sits directly below the surface hotspot, forming the base of a vertical conduit that channels heat upward through nearly the entire thickness of the mantle.
Whole-mantle seismic tomography, a technique that uses earthquake waves to build three-dimensional images of Earth’s interior, has confirmed that broad, plume-like structures extend through the lower mantle beneath several major hotspots. A 2015 study in Nature by seismologists Barbara Romanowicz and Scott French showed that these conduits are not narrow jets but wide columns hundreds of kilometers across, and that Iceland’s is among the most clearly resolved. The combination of a ULVZ at the base and a continuous low-velocity conduit reaching the surface makes Iceland’s plume the most thoroughly imaged from bottom to top.
Drilling into the plume’s history
Seismic images show the plume’s current structure, but they cannot reveal how it has behaved over millions of years. That is where the 2026 study comes in. Researchers analyzed basalt samples drilled from the seafloor south of Iceland during International Ocean Discovery Program Expedition 395. The cores captured a volcanic record stretching back millions of years, and the chemistry of those basalts told a story of a plume that pulses: periods when the flow of hot material surged, producing thicker crust and more voluminous eruptions, alternating with quieter intervals when the melt supply dropped.
“The geochemical record from these cores gives us a direct timeline of how the plume’s output has waxed and waned,” said one of the study’s lead researchers in a summary accompanying the Nature Communications paper. That timeline is built from isotopic ratios of strontium, neodymium, lead, hafnium, tungsten, and noble gases in the drilled basalts, signatures that point to a source region isolated deep in the mantle for billions of years. A 2023 review in Nature Reviews Earth & Environment detailed how these isotopic fingerprints are used to trace mantle source regions, and Iceland’s signatures are consistent with material that has sat near the core-mantle boundary since early in Earth’s history before being swept into the rising plume.
When two fundamentally different approaches, seismic imaging of structure and geochemical analysis of composition, independently point to the same deep origin, the case becomes substantially stronger than either line of evidence alone.
What the plume model still cannot explain
Not all scientists are convinced that a single deep plume accounts for everything happening beneath Iceland. A perspective catalogued by the U.S. Geological Survey laid out several persistent critiques. The temperature anomaly measured at Iceland’s surface does not always match what deep-plume models predict. The geographic pattern of volcanism over time does not trace a clean hotspot track the way Hawaii’s chain does. And seismic anomalies in the upper mantle beneath Iceland are broader and more diffuse than classic plume theory would suggest.
Some researchers argue that the spreading of the Mid-Atlantic Ridge, which runs directly through Iceland, combined with small-scale convection in the upper mantle, can explain much of the island’s volcanic output without requiring a structure rooted at the core-mantle boundary. Iceland sits on a tectonic plate boundary that is offset and segmented, with variable crustal thickness shaped by past rifting episodes. Numerical models can reproduce certain aspects of Icelandic volcanism with a deep plume, others with plate-driven flow alone, and still others with a hybrid of both. No single model yet captures all the observations, including eruption rates, crustal structure, seismic anomalies, and geochemical diversity, without leaving significant gaps.
There are also limits to what current instruments can measure at such extreme depths. No study has directly recorded the temperature or melt fraction inside the deep conduit. Seismic tomography infers those properties from how much waves slow down, but converting wave speed to temperature depends on assumptions about rock composition and mineral behavior under pressures found 1,800 miles below the surface. The ULVZ beneath Iceland could represent partial melt, chemically dense material swept to the mantle’s base, or some combination. Distinguishing between those possibilities will require advances in imaging resolution and laboratory experiments that simulate deep-mantle conditions.
Why it matters beyond geology
Understanding the plume beneath Iceland is not purely academic. The Reykjanes eruptions that began in 2021 and intensified through 2024 and into 2025 have displaced residents, threatened critical infrastructure including the Svartsengi geothermal power plant, and disrupted regional planning. Knowing whether the deep heat source is surging or waning, as the IODP drilling record suggests it does over longer timescales, could eventually help scientists assess whether the current eruptive cycle is a brief flare or the opening phase of a more sustained period of activity on the peninsula.
The Iceland plume also serves as a benchmark for studying deep-mantle processes elsewhere. Hawaii, Yellowstone, and the Galapagos are all thought to be fed by mantle plumes, but none has been mapped as completely from core-mantle boundary to surface. Techniques refined on Iceland’s well-imaged conduit, including joint seismic-geochemical analysis and ocean-floor drilling, are being applied to those other hotspots to test whether they share the same deep architecture.
Three open questions driving the next round of research
First, what triggers the plume’s pulses? The 2026 drilling study documents cycles of collapse and resurgence but does not resolve whether instabilities at the core-mantle boundary set the tempo or whether interactions with the Mid-Atlantic Ridge modulate the flow from above. Second, how does the plume couple to the complex tectonic setting at the surface? Iceland straddles a spreading ridge, and separating the plume’s contribution from the ridge’s own magma production remains a challenge. Third, can higher-resolution seismic networks and additional drilling narrow the uncertainties about temperature, composition, and melt fraction inside the conduit itself?
Each new eruption on the Reykjanes Peninsula, each seismic survey, and each core pulled from the North Atlantic seafloor adds detail to a picture that now stretches from the edge of Earth’s molten iron core to the glaciers and fresh lava fields above. The plume beneath Iceland is the clearest window scientists have into how the planet’s deepest interior shapes the ground we stand on.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
