The deepest hole humans have ever dug, the Kola Superdeep Borehole on the Kola Peninsula in Russia, reached approximately 12 kilometres before the heat, pressure, and engineering challenges of drilling overwhelmed the equipment. Earth’s mantle begins at approximately 35 kilometres depth and continues to approximately 2,890 kilometres — 1,800 miles — where it meets the boundary of the molten outer core. Every layer beneath the upper crust of Earth is, in practical terms, inaccessible to direct observation. Everything humans know about the interior of the planet has been inferred from seismic waves passing through it: their travel times, their reflections, their refractions, the patterns of speed and direction by which they emerge at the surface after earthquakes generate them on the other side of the globe. The most striking thing humans have learned about the planet’s interior from these inferences, over the past several decades, is that the lowest layer of the mantle is not uniform. It contains two enormous, continent-spanning, distinct structures.
According to a comprehensive reference summary of LLSVPs in the geophysical literature, the two structures are formally known as Large Low-Shear-Velocity Provinces (LLSVPs), a name derived from the technical observation that distinguishes them — they slow down seismic shear waves by approximately 1 to 3 percent relative to the surrounding mantle. They were discovered in the late 20th century as global seismic tomography improved enough to resolve them. The African LLSVP, sitting beneath Africa and the eastern Atlantic, was eventually named Tuzo in honour of the Canadian geologist John Tuzo Wilson, one of the founding figures of plate tectonics. The Pacific LLSVP, sitting beneath the central and western Pacific Ocean, was named Jason in honour of the American geophysicist W. Jason Morgan, who developed the modern theory of mantle plumes. The names were proposed in 2011 by the geologist Kevin Burke. The two structures together occupy approximately 8 percent of the volume of Earth’s mantle, or roughly 6 percent of the total volume of the planet.
How big they actually are
The Pacific LLSVP, Jason, extends approximately 3,000 kilometres laterally — nearly 1,900 miles across, comparable to the width of the contiguous United States. The African LLSVP, Tuzo, is similarly continental in scale, extending across much of the African plate and into the southern Atlantic. Both structures rise vertically from the core-mantle boundary by up to approximately 1,000 kilometres — taller than the height of the Earth’s atmosphere by a factor of ten. Each is, in absolute terms, substantially larger than any feature visible on the planet’s surface. The Himalayas, if dropped into Tuzo or Jason, would be one of the smaller features within the structure. The entire continent of Australia, if shrunk down to fit, would occupy a modest fraction of either LLSVP. The structures are not visible, not directly accessible, not detectable except via seismic inference, and not part of the standard educational treatment of Earth’s interior — but they are, in raw scale, among the largest individual structures the species has ever identified inside any planetary body.
As reported by a 2018 paper in Geoscience Frontiers by the geochemist Yaoling Niu on the origin of LLSVPs, the two structures are not symmetrically distributed around the planet. They are antipodal — sitting on roughly opposite sides of Earth from each other — and equatorial, with their combined centre of mass aligned with the planet’s rotation axis. This geometric arrangement is one of the proposed reasons for their long-term stability. A pair of large mass anomalies positioned this way is, dynamically, in a configuration that resists lateral migration during the planet’s rotation. Whatever Tuzo and Jason are, they have been sitting essentially where they currently sit for an extraordinarily long time.
How long they have been there
The temporal stability of the LLSVPs is, by paleogeological standards, remarkable. The continents on Earth’s surface have moved substantially during the past several hundred million years — Pangaea broke apart approximately 200 million years ago, the Atlantic Ocean has been opening since, India collided with Asia approximately 50 million years ago, and the basic configuration of the surface looked entirely different in geological deep time. The LLSVPs, in contrast, appear to have stayed put. The evidence comes from a clever geological argument. Hotspots — the volcanic features at the surface where mantle plumes break through, such as Hawaii, Iceland, the Canary Islands, and Réunion — appear to originate predominantly at the edges of the LLSVPs, in regions geologists now call Plume Generation Zones. When ancient hotspot tracks are reconstructed and mapped back through 300 million years of plate motion, more than 90 percent of them line up with the current locations of the LLSVP margins. Some reconstructions extend the pattern back to 540 million years ago, into the early Cambrian. This suggests Tuzo and Jason have been holding their shape, and their positions, since before the appearance of complex life on land.
What they actually are
What Tuzo and Jason are made of, and how they originally formed, is one of the major unresolved questions in deep-Earth geophysics. Per a 2016 review in Nature Geoscience by Garnero, McNamara, and Shim on continent-sized anomalous zones in the deep mantle, several distinct hypotheses are currently being investigated. The thermal interpretation holds that LLSVPs are simply zones of unusually hot mantle, where the slower shear wave velocities reflect higher temperature. The chemical interpretation, supported by the difficulty of explaining the LLSVPs’ apparent density excess with temperature alone, holds that they are compositionally distinct — possibly enriched in iron, possibly containing material from subducted oceanic crust accumulated over hundreds of millions of years, possibly remnants of a primordial layer that survived from early Earth differentiation.
As covered in a Daily Galaxy summary of a recent hypothesis on LLSVP origin from early-Earth core-mantle material exchange, one increasingly explored hypothesis is that LLSVPs may date back to the very early period of Earth’s formation, approximately 4 billion years ago, when the entire planet was a global magma ocean and material was exchanging freely between what would eventually become the core and what would eventually become the mantle. The 2025-2026 simulations supporting this interpretation suggest that material contaminated by core composition could have accumulated in two stable, antipodal locations during the cooling of the early magma ocean — eventually crystallising into the dense, hot, chemically distinct structures that are still sitting in those locations 4 billion years later. Whether this interpretation, the subducted-crust interpretation, the primordial-layer interpretation, or some combination turns out to be correct will require more seismic data, more mineral physics experiments at the relevant pressures and temperatures, and probably the kind of multi-decade research programme that deep-Earth geophysics has historically required.
What is no longer in doubt is that the structures exist, that they are extraordinarily large, that they have been stable for hundreds of millions of years, and that they appear to control where most of the planet’s hotspot volcanism happens at the surface. The Hawaiian Islands are erupting where they are because Jason’s edge is underneath. The Canary Islands erupted in 2021 because Tuzo’s edge is underneath. The largest geological structures in human experience are, in a literal sense, downstream effects of much larger structures sitting at the base of the mantle, which no human will ever see and which the species nonetheless now knows are there.
