Simulations of 100 million years of mantle flow best match the New England chain’s shape and seamount ages when one plume, now under the Azores, drifts across three plates, along a path the team chose.

Peer reviewed · Modelling study · No new rock samples or measurements

A single mantle plume, the hot upwelling now beneath the Azores, may have built the New England Seamounts, a chain of extinct undersea volcanoes off the US east coast. So proposes a study published in Science Advances on 13 November 2026 by Mengxue Liu of the China Earthquake Administration and colleagues in Beijing and Zurich. Their 3D computer models suggest the same plume also made the Corner and Great Meteor seamounts. The catch: the plume had to wander along a path the researchers chose to fit the data, and rock chemistry does not cleanly link the chains.

What you need to know

  • Finding: The model that best matches the chain’s shape and dated volcano ages has the Azores plume moving northeast, then turning northwest after about 70 million years ago, feeding volcanoes on the North American, African and Eurasian plates in turn.
  • Method: 3D thermomechanical simulations of the whole mantle in the CitcomS code, driven by reconstructed plate motions at 1-million-year steps, with simplified melting to predict where lava would erupt.
  • Source: Liu et al., “The Azores plume origin of the New England Seamounts chain”, Science Advances, 13 November 2026 (open access).
  • Why it matters: The source of this 124-million-year-old volcanic track has been argued over for decades, and the model points to mantle flow driven by a sinking slab under western North America as the main force bending the plume.
  • Limitation: No new rocks were dated or analysed; fixed-plume models fit poorly, and the winning plume path is a tuned input, so a good fit supports the idea without proving it.

Evidence at a glance

Evidence summary for the Azores plume and New England Seamounts modelling study
ItemAssessment
Evidence statusPeer reviewed. Modelling study: numerical simulations compared with published seamount ages, seafloor shape and seismic images. No new field data.
Primary sourceMengxue Liu, Shijie Xie, Dinghui Yang, Taras Gerya, Lijun Liu and four co-authors, “The Azores plume origin of the New England Seamounts chain”, Science Advances 12(46), eaee1780. Received 25 November 2025, accepted 9 October 2026, published 13 November 2026. CC BY 4.0.
Evidence strengthModerate for the claim that an Azores plume source is physically plausible, because the model reproduces the chain’s shape and ages better than fixed-plume alternatives. Limited for the claim that it actually happened, because the plume’s path was chosen among several tested options and the chemistry of the lavas is not a clean match.
Principal limitationThe plume’s movement at 600 km depth is imposed, not calculated. The authors tried several paths and kept the one that fit best.
Coverage at a glance1 distinct credible publisher found as of 18 November 2026, 09:25 UTC (Phys.org), from SciQuest web searches for the paper title, DOI and topic. The AAAS press release on EurekAlert (from the journal’s publisher) was read but not counted; aggregators and AI-written encyclopedia pages were excluded. Coverage is not evidence that a finding is correct.

Sources: the Science Advances paper (full text via PMC13763836), the AAAS/EurekAlert release and SciQuest’s coverage search at the time stated.

Is this a big deal?

It’s a solid modelling step in a long argument, not a settled answer. The Azores plume idea was already on the table, raised in 2003 from rock chemistry and seafloor shape. What this paper adds is a physical test of whether one plume could have drawn the whole track, given what plates and sinking slabs were doing. In the model it can, but only if it drifted in a particular way. That makes the idea more credible and gives other teams a specific path to test.

What scientists found

The New England Seamounts run southeast from the edge of the North American continent into the deep Atlantic. NOAA describes the chain as more than 30 volcanic peaks stretching nearly 1,200 km, the longest in the North Atlantic. Dated rocks along the wider track get steadily younger to the east, from about 124 million years at the Monteregian Hills in Quebec to 82.7 million years at Nashville Seamount and 78 million years at the Corner Seamounts. That steady ageing is the classic sign of a hotspot track: the plate slides over a fixed or slow-moving source of hot rock, and a line of volcanoes forms.

Liu’s team tested where that source sat. With the plume source fixed beneath the Great Meteor Seamounts, the simulated track missed the observed one, and the Corner and Great Meteor seamounts fell outside the zone where lava erupted. A plume fixed beneath the Azores fitted even worse. A slow northward drift of 1 cm a year, based on earlier estimates for the Azores plume, also failed.

The best fit, called model 3B, has the plume source move northeast first, then turn northwest after about 70 million years ago. In that run the plume sits under the North American Plate and builds the New England and Corner seamounts, crosses beneath the Mid-Atlantic Ridge to feed the Great Meteor Seamounts on the African Plate, and around 10 million years ago passes under the Eurasian Plate to reach the Azores. The authors report that its predicted track matches the chain’s shape, its ages match the dated rocks, and its plume conduit tilts toward the Mid-Atlantic Ridge, as seismic images of the Azores plume show.

How they found it

A mantle plume is a column of unusually hot rock rising slowly through the mantle, the solid but slowly flowing layer between Earth’s crust and core. Modelling one over 100 million years means modelling everything around it too. The team built a regional model covering 120° of latitude and 210° of longitude, reaching from the surface to the core, with a grid spacing of about 26 km in depth. Its surface was moved according to published plate reconstructions, updated every million years, so that the Farallon Plate sank under western North America as the reconstructions say it did.

The plume was placed at 600 km depth, either fixed or moving at 0.1° per million years, which is about 11 km per million years, or roughly 1.1 cm a year. Its excess temperature was set at 250 to 350 °C, values the authors take from observations of upper-mantle plumes. Where rock got hot enough to melt past a 4% threshold, a share of the melt “erupted” at the surface and was carried along with the plate as a model volcano. That let the team compare predicted volcano positions and ages with real ones, rather than just projecting the deep source straight upward.

Show the maths: how far did the plume source wander?

In the moving-plume runs, the source at 600 km depth moves 0.1° of arc per million years. One degree of arc at Earth’s surface is about 111.2 km, so that is 0.1 × 111.2 ≈ 11.1 km per million years.

Converting: 11.1 km = 1,110,000 cm, spread over 1,000,000 years, gives about 1.1 cm a year.

Kept up for the roughly 100 million years the model covers, that adds up to about 10° of arc, or roughly 1,100 km travelled along its path. Because model 3B’s path bends, the straight-line distance between start and end is shorter.

The speed is close to the 1 to 2 cm a year that Arnould and colleagues inferred in 2019 for a northward-drifting Azores plume. The new result differs in direction, not speed: a steady 1 cm-a-year drift due north did not reproduce the chain, while the northeast-then-northwest path did. SciQuest calculated these figures from the rate stated in the paper’s methods.

One plume, three plates: the proposed Azores plume track Dated volcanoes along the New England chain get younger to the east; the model links them to the plume now under the Azores Land: Quebec and New England 70°W 60°W 50°W 40°W 30°W 30°N 35°N 40°N 45°N Mid-Atlantic Ridge NORTH AMERICAN PLATE AFRICAN PLATE EURASIAN PLATE Gap: no clear seamount trail (model: dispersed melting, ~80–30 Ma) Monteregian Hills 124 Ma White Mountains 122.5 Ma Bear Seamount 103.4 Ma Nashville Seamount 82.7 Ma Corner Seamounts 78 Ma 5 more dated seamounts, 96–83.5 Ma Great Meteor Seamounts 10–30 Ma Azores: plume today Dated volcano (age in millions of years) Hotspot location shown in the paper Plate boundary Schematic: positions approximate What the best-fitting model (3B) proposes, in millions of years ago (Ma) 124: oldest dated rocks ~100: model run begins ~80: plume meets the ridge ~70: plume source turns NW 80–30: dispersed melting ~10: enters Eurasian Plate 0: under the Azores Blue bar: span simulated (~100 Ma). Green: stretch with no clear seamounts. Sources: Liu et al., Science Advances (2026), doi:10.1126/sciadv.aee1780 (Fig. 1 ages and positions; model 3B results). Model results are simulations, not observations. SciQuest original graphic.
Dated volcanoes along the New England chain (yellow) get younger from Quebec toward the Corner Seamounts. The study’s best-fitting model links them, across a gap with no clear seamount trail, to the Great Meteor Seamounts and the Azores (red). The timeline shows what that model proposes; it is a simulation, not an observation. Ages and approximate positions from Liu et al., Science Advances (2026), Fig. 1. SciQuest original graphic.

One way to picture it: slide a sheet of paper slowly over a candle and you get a line of scorch marks, oldest at one end. That is the textbook hotspot. This study says the candle was also being carried sideways, and a draught was bending its flame. The comparison breaks down quickly, since a plume is solid rock creeping at centimetres a year, not a flame, but it captures the three things the model juggles: the moving plate, the moving source and the sideways “wind” in the mantle.

Dated volcanoes along the proposed Azores plume track, oldest first
Volcano or groupPlateAge (million years)
Monteregian Hills, QuebecNorth American124
White Mountains, New EnglandNorth American122.5
Bear SeamountNorth American103.4
Atlantis II SeamountNorth American96
Vogel SeamountNorth American91.2
Gosnold SeamountNorth American90.8
Allegheny SeamountNorth American83.8
Michael SeamountNorth American83.5
Nashville SeamountNorth American82.7
Corner SeamountsNorth American78
Great Meteor SeamountsAfrican10 to 30
AzoresEurasianPlume beneath today

Measured ages from earlier dating studies, as compiled in Liu et al. (2026), Fig. 1. The model runs cover only about the last 100 million years, so the two land sites were not simulated.

Why it matters to you

This won’t change anything in your daily life. It is a question about how the ground under the North Atlantic was built. It does have a familiar edge for people in the northeastern US and Quebec: NOAA counts the Monteregian Hills near Montreal and the magma intrusions of New Hampshire’s White Mountains as the oldest, onshore end of this same volcanic track. If the model is right, rocks some hikers walk on and the Great Meteor Seamounts, roughly 4,000 km away in the eastern Atlantic, share a source that now sits under Portugal’s Azores islands.

What the evidence does not show

The model does not show that the plume really moved this way. The researchers imposed each plume path and kept the one that fit, after a fixed plume and a simple northward drift failed. A different combination of inputs, such as another plate reconstruction or viscosity structure, might fit as well. Model 7B, a different two-stage path, gave very similar results, differing mainly in seamount ages older than about 90 million years.

The chemistry is also not a neat fit. Lavas from the New England Seamounts carry a different lead-isotope signature from Azores lavas, and in 2003 Pascal Gente and colleagues used differences in lava chemistry to argue that two separate hotspots were involved. Liu’s team notes that long-lived plumes can change chemically over time, as at Hawaii and Tristan-Gough, but they write that they do not use the geochemical data “as definitive proof of a common plume source.”

Two more limits matter. Between the Corner Seamounts and the Great Meteor Seamounts there is no clear trail of volcanoes, and the model explains that gap after the fact. And the melting scheme is simplified: it records where and when melt would erupt, not whether there was enough of it to build a seamount.

How this fits what we already knew

For decades the leading view, as the paper summarises it, was that the New England chain and the Great Meteor Seamounts were two halves of one track made by a plume beneath Great Meteor, on the African Plate. That view drew on modelling, plate reconstructions, chemistry and rock ages. Other proposals put an extinct plume under North America, or, following Gente’s 2003 work, a source beneath the Azores. Seismic imaging has found no upper-mantle plume under either the New England or the Great Meteor seamounts, while a clear hot anomaly sits beneath the Azores that may mark a plume reaching at least 150 to 300 km down.

The new paper also argues about what bent the track. It finds the main control was not the drag of the moving plate but a broad westward flow in the upper mantle, set up as the old Farallon Plate sank steeply beneath western North America and drew mantle toward the trench. A bend in the predicted track at about 75 to 85 million years ago looks much the same whichever path the plume takes, which the authors tie to a counterclockwise turn of the North American Plate. That contrasts with Hawaii’s famous bend, where plume motion is thought to account for up to about 20°, roughly a third of the turn. Their earlier model of Hawaii, published in 2024, used the same approach.

For the gap, the authors propose that from about 80 to 30 million years ago the plume met the slow-spreading Mid-Atlantic Ridge, and its hot material spread sideways and along the ridge instead of building tall volcanoes. They point to similar gaps in the Kerguelen and Réunion hotspot tracks, which also crossed ridges between plates.

Independent perspective

SciQuest found no independent comment on this paper at the time of writing and did not obtain one. Co-author Taras Gerya of ETH Zurich is a geodynamicist, but as an author he is not an independent voice. The study’s own framing is careful: “Our models test whether a dynamically plausible connection could exist among the New England–Corner seamounts, the Great Meteor Seamounts, and the Azores seamounts.”

How other headlines framed it

Only one independent outlet had covered the paper when we checked. Phys.org framed it as a possible explanation, saying one wandering plume “may have” made several Atlantic chains, and noted near the end that the simulations do not prove a common origin. The AAAS press release from the journal’s publisher went further, saying the study “pinpoints” the source. The paper itself uses “propose” and “suggest” for its main claim, and SciQuest has followed that.

What happens next

The clearest tests are on the seafloor. The authors call for combined isotope and trace-element work to settle whether the lavas share a source, and more dated samples from the gap between the Corner and Great Meteor seamounts would show whether the predicted ages hold. Sharper seismic images of the Azores plume could check the predicted tilt of its conduit. The paper does not name a specific expedition or date. Model inputs and results are archived on Zenodo, so other groups can rerun them.

In one sentence

Computer models suggest the hot plume now under the Azores could have built the New England Seamounts and the Great Meteor Seamounts by drifting across three plates, but only along a path the team chose to fit, so the idea is plausible, not proven.

Quick check: did it click?
  1. Did the team date any new rocks? No. They ran computer simulations and compared them with ages and seismic images from earlier studies.
  2. Which model fitted best: a fixed plume or a moving one? A moving one, travelling northeast and then northwest; fixed plumes under Great Meteor or the Azores fitted poorly.
  3. What does the model say mainly bent the plume? Upper-mantle flow toward the trench where the Farallon Plate sank under western North America, more than the drag of the moving plate.

Read next

Sources and disclosure

  1. Liu, M., Xie, S., Yang, D., Gerya, T., Liu, L., Qi, R., Shan, X., Dong, X. and Xu, X. The Azores plume origin of the New England Seamounts chain. Science Advances 12, eaee1780 (2026). Full text: PMC13763836. Data: Zenodo.
  2. AAAS, “One mantle plume to rule them all: Study pinpoints the source of a submarine volcano track”, EurekAlert, 13 November 2026.
  3. NOAA Ocean Exploration, “Geological History of the New England and Corner Rise Seamounts”, 23 June 2021.
  4. Gente, P., Dyment, J., Maia, M. and Goslin, J. Interaction between the Mid-Atlantic Ridge and the Azores hot spot during the last 85 Myr. Geochemistry, Geophysics, Geosystems 4, 8514 (2003).
  5. Arnould, M., Ganne, J., Coltice, N. and Feng, X. Northward drift of the Azores plume in the Earth’s mantle. Nature Communications 10, 3235 (2019).

Disclosure: This article was drafted with AI assistance and checked against the sources above. The 1.1 cm-a-year and 1,100 km figures (from the paper’s 0.1° per million years) and the 4,000 km distance (from approximate map positions) are SciQuest calculations; the graphic and table are SciQuest originals based on the paper’s Fig. 1 and results. The study was funded by China’s Deep Earth Probe and Mineral Resources Exploration National Science and Technology Major Project, and the authors declare no competing interests. Spotted an error? See our corrections policy.


Discover more from SciQuest

Subscribe to get the latest posts sent to your email.