The upper mantle stretches from the base of Earth’s crust down to about 660 kilometers depth, giving it a total thickness of roughly 630 to 650 kilometers depending on where you measure. That range exists because the crust itself varies in thickness: it can be as thin as about 10 kilometers under the oceans and as thick as 50 kilometers beneath mountain-bearing continents. Since the upper mantle begins wherever the crust ends, its thickness shifts accordingly. But the bottom boundary is far more consistent, defined by a sharp change in mineral structure that seismic waves reliably detect at 660 kilometers worldwide.
Where the Upper Mantle Begins
The upper boundary of the upper mantle is a feature called the Mohorovičić discontinuity, usually shortened to Moho. It marks the place where the relatively light, rigid rocks of the crust give way to the denser, hotter rock below. Under continents, the Moho sits about 30 to 50 kilometers down, while beneath ocean basins it lies roughly 10 kilometers below sea level.1Nature. The Mohorovičić Discontinuity That difference is why the upper mantle is thicker beneath oceans (about 650 km) than beneath continental mountains (closer to 610 km).
The Moho sounds like a neat line on a diagram, but in reality it is a more complicated boundary than textbooks suggest. Detailed studies in Canada have shown that it has different signatures at different scales and in different tectonic settings. In some regions the Moho appears to be ancient, possibly dating from the time the crust originally formed, while in other areas it seems to be a younger feature stamped onto older rock by later tectonic activity.2Canadian Journal of Earth Sciences. How the crust meets the mantle: Lithoprobe perspectives on the Mohorovičić discontinuity and crust–mantle transition So while the depth of the Moho can be measured precisely at any given spot, thinking of it as a single uniform surface everywhere on Earth oversimplifies things.
Layers Within the Upper Mantle
The upper mantle is not one homogeneous slab. It contains several distinct zones stacked inside it, each with different physical behavior, and understanding these layers matters for everything from plate tectonics to volcanic activity.
The topmost part is the lithospheric mantle. This is the portion of the mantle that is cool and rigid enough to behave as part of Earth’s tectonic plates. Together with the crust above it, it forms the lithosphere, which is the layer that cracks, slides, and collides. Under old, stable continental interiors called cratons, the lithospheric mantle can extend down to around 175 kilometers or more. Kimberlite pipes, the volcanic conduits that carry diamonds to the surface, bring up chunks of this deep rock. Studies of these rock fragments from pipes in Lesotho have documented sampling depths from about 100 to 175 kilometers,3International Journal of Earth Sciences. Multistep evolution of harzburgitic mantle underneath pipe 200 kimberlite, northern Lesotho and pipes in Yakutia, Russia, reveal a layered lithospheric mantle with chemically distinct zones at different pressures.4Geosystems and Geoenvironment. Mantle xenoliths from Komsomolskaya kimberlite pipe, Yakutia: Multistage metasomatism
Below the lithosphere sits the asthenosphere, a zone where the rock is hot enough to flow slowly over geological time. This is the lubricating layer that allows tectonic plates to move. Within the asthenosphere, at depths of roughly 150 to 200 kilometers, there is a feature called the low-velocity zone, where seismic waves slow down. Research published in 2024 linked this zone to melting of recycled oceanic crust that has been carried deep into the mantle and transformed into a rock type called eclogite. The resulting melts are dense and sluggish, so they accumulate and sit at those depths rather than rising to the surface, creating the velocity dip that seismologists observe.5Geology. Melting of eclogitic oceanic crust for the low-velocity zone within Earth’s upper asthenosphere
Another subtle boundary, the Lehmann discontinuity, appears at roughly 220 kilometers depth beneath stable continents. Analysis of long-period seismic waves passing through Australia and the western Pacific supports the idea that this discontinuity marks the transition from anisotropic rock in the lithosphere (where crystals are aligned in a preferred direction, a relic of past deformation) to more randomly oriented material below.6PubMed. Lehmann discontinuity as the base of an anisotropic layer beneath continents The Lehmann discontinuity is not seen everywhere and is most prominent under old continental shields, which makes it a regional feature rather than a global one.
The 410-Kilometer Discontinuity
At about 410 kilometers depth, the upper mantle hits a major transition. The mineral olivine, which dominates the shallower mantle, is squeezed under such enormous pressure that its crystal structure rearranges into a denser form called wadsleyite. This is a phase change, essentially the same atoms repacking themselves more tightly, and it produces a sharp jump in seismic wave speed that instruments can detect worldwide.
Laboratory experiments reproducing mantle pressures and temperatures have pinned down the conditions of this transformation precisely. By matching the pressure of the olivine-to-wadsleyite transition with the observed depth of the 410-kilometer discontinuity, researchers have estimated the temperature at that depth to be about 1760 K (roughly 1490 °C), give or take about 45 degrees.7Journal of Geophysical Research: Solid Earth. Olivine‐wadsleyite transition in the system (Mg,Fe)2SiO4 The 410 marks the top of what geophysicists call the mantle transition zone, a region of rapid change that extends down to 660 km.
The Transition Zone and Its Internal Structure
Between 410 and 660 kilometers, the mantle undergoes a cascade of mineral transformations. Wadsleyite changes to yet another, even denser crystal structure called ringwoodite at around 520 kilometers depth. This 520-kilometer discontinuity is weaker and harder to detect than the ones at 410 and 660 km, but recent work using ambient seismic noise beneath the contiguous United States has mapped it in new detail, finding a strong reflection signal under the central U.S. and a weaker one under the eastern U.S.8Geophysical Research Letters. Internal Structure of the Mantle Transition Zone Beneath the Contiguous U.S.: Insights From the 520‐km Discontinuity Revealed by Ambient Noise Correlations
The 520-kilometer feature turns out to be more than a simple mineral change. Experiments have shown that at similar depths, calcium-rich garnet also transforms into a new phase, and these two transformations can produce what looks like a “split” discontinuity when viewed with seismic waves. The degree of splitting varies from region to region and appears to reflect differences in calcium concentration in the mantle rather than differences in temperature.9PubMed. Splitting of the 520-kilometer seismic discontinuity and chemical heterogeneity in the mantle That makes the 520 a useful probe of large-scale chemical variety in the deep mantle, essentially revealing how “fertile” the mantle is in different places and how much recycled oceanic crust has been mixed in.
The 660-Kilometer Floor
The bottom of the upper mantle is the sharpest and most globally consistent of all its internal boundaries. At around 660 kilometers, ringwoodite breaks down into two different minerals: bridgmanite and ferropericlase. This reaction is known as the post-spinel transition, and new high-resolution experiments using synchrotron X-ray diffraction under controlled pressures have confirmed that the transition pressure lines up almost perfectly with the observed 660-kilometer depth.10Scientific Reports. Complete agreement of the post-spinel transition with the 660-km seismic discontinuity This result affirms the standard model of the upper mantle’s base.
The 660 does not sit at exactly the same depth everywhere, however. In regions where cold tectonic slabs are sinking into the mantle, the discontinuity can be pushed deeper. Mapping of the transition zone beneath the Ibero-Maghrebian region found that the cold Gibraltar-Alboran and Alpine-Tethys slabs thickened the transition zone by as much as 35 kilometers along the Mediterranean coast.11Earth and Planetary Science Letters. High-resolution mapping of the mantle transition zone and its interaction with subducted slabs in the Ibero-Maghrebian region In those areas, the 410 is pushed slightly shallower (because the cold slab favors the phase change at lower pressure) while the 660 is pushed deeper (because its mineral reaction has the opposite temperature dependence). The net effect is a fatter transition zone wherever cold material is descending.
Whether those sinking slabs punch through the 660 or pile up on top of it has been one of the big questions in geophysics. Computer models of mantle convection, when compared against seismic images of real slab geometries, suggest that Earth’s convection is more like a whole-mantle circulation, meaning material does eventually pass through the 660, rather than two completely separate convection loops stacked on top of each other.12Journal of Geophysical Research: Solid Earth. Three‐dimensional spherical models of layered and whole mantle convection But “eventually” is doing heavy lifting in that sentence: some slabs stall at the 660 for tens of millions of years before sinking further, creating temporary barriers that complicate the picture.
A Hidden Water Reservoir
One of the most surprising discoveries about the upper mantle in recent decades is that the transition zone can hold enormous quantities of water, not as liquid water or ice, but as hydrogen atoms locked inside the crystal structures of its minerals. Wadsleyite and ringwoodite, the dominant minerals between 410 and 660 km, can incorporate up to about 3 weight percent water in their structures.13PubMed Central. A nearly water-saturated mantle transition zone inferred from mineral viscosity More refined experiments at realistic mantle temperatures found that ringwoodite can store about 0.8 to 1.2 weight percent of water under actual transition-zone conditions.14Earth and Planetary Science Letters. High water solubility of ringwoodite at mantle transition zone temperature
Even at the lower end of those estimates, the total volume of the transition zone is so vast that it could hold a quantity of water comparable to the world’s surface oceans. Whether it actually does contain that much is still debated, but viscosity measurements of transition-zone minerals suggest the zone could be close to water-saturated in some regions. This water influences how easily rock flows, how readily it melts, and how efficiently tectonic plates are recycled. It may also help explain why the transition zone acts as a partial barrier to convection: water-rich minerals above and below the 410 and 660 boundaries have very different capacities for storing hydrogen, so water released during phase changes could trigger localized melting at those boundaries.
How Scientists See Through 660 Kilometers of Rock
Nobody has drilled anywhere close to the upper mantle. The deepest borehole ever completed barely scratched 12 kilometers. So everything known about the upper mantle’s thickness and internal layering comes from indirect methods, and it is worth understanding how they work because their limitations shape what we can and cannot say with confidence.
Seismic waves are the primary tool. Earthquakes and controlled explosions send waves through the planet, and every time those waves cross a boundary where rock properties change, part of the energy bounces back or converts from one wave type to another. By recording these arrivals at surface stations, geophysicists build images of the subsurface. A technique called receiver function analysis tracks the conversions between compressional and shear waves at each discontinuity, and it has been the workhorse for mapping the 410 and 660 boundaries around the world.15Earth and Planetary Science Letters. Receiver function imaging of upper mantle complexity beneath the Pacific Northwest, United States Newer methods using sparse signal-processing techniques are improving the resolution of these images, making it possible to detect subtler features like the Lehmann discontinuity and the 520 with greater confidence.16Geophysical Journal International. On the detection of upper mantle discontinuities with radon-transformed receiver functions (CRISP-RF)
Seismic methods are complemented by electromagnetic techniques. Magnetotelluric surveys measure natural variations in Earth’s electric and magnetic fields to map the electrical resistivity of the subsurface. Since hot, partially molten, or water-rich rock conducts electricity differently than cold, dry rock, these surveys can illuminate features that seismic methods might miss. A magnetotelluric transect across Colorado, for instance, imaged the crust and upper mantle to depths exceeding 150 kilometers, revealing variations in hydration and melt content beneath the Rocky Mountain Front.17Journal of Geophysical Research: Solid Earth. Magnetotelluric Imaging of Lower Crustal Melt and Lithospheric Hydration in the Rocky Mountain Front Transition Zone, Colorado, USA
Then there are the physical samples. Volcanic eruptions, especially kimberlite eruptions, rip pieces of mantle rock from depths of 100 to over 200 kilometers and carry them to the surface as xenoliths. These chunks are the only direct samples of the upper mantle anyone has ever held, and their mineral chemistry provides ground truth for the models built from remote sensing. The upper mantle is the source of nearly all magmas that reach Earth’s surface, but most mantle samples come from the relatively shallow lithospheric portion; the deeper asthenosphere and transition zone partially melt and chemically change before any material can reach us, making direct sampling of those regions extremely difficult.18GeoScienceWorld (Elements). The Upper Mantle and Transition Zone
The Upper Mantle Is Not Permanent
It is easy to picture the upper mantle as a fixed feature of the planet, but its composition and even its effective thickness have changed over geological time. In the Archean eon, when Earth’s interior was significantly hotter than today, converging oceanic plates produced far greater volumes of heavily depleted rock (mantle rock from which melt had already been extracted). Geodynamic modeling shows that at mantle temperatures 200 to 250 °C hotter than the present, depleted peridotites could have made up more than half of the upper mantle in as little as 100 to 200 million years. At modern temperatures, that buildup of depleted rock does not happen to nearly the same degree.19PubMed Central. Depletion of the upper mantle by convergent tectonics in the Early Earth The Archean upper mantle, in other words, was chemically quite different from the one we have now.
Even in more recent geological time, the upper mantle has been actively reshaped. Analysis of kimberlite distributions across Africa, combined with seismic velocity anomalies at 110 to 150 kilometers depth, indicates that large portions of African cratonic lithosphere have been destroyed or substantially thinned during the last 200 million years, likely by interaction with mantle plumes.20PubMed Central. African cratonic lithosphere carved by mantle plumes The thick, cold roots that once extended deep beneath cratons like the southern Kaapvaal and western Man-Léo have been eroded from below, shrinking the lithospheric mantle in those regions. The upper mantle’s layered structure is not a static feature etched permanently into the planet; it is a snapshot of ongoing processes that continuously build, thin, and chemically alter it.
What About Other Planets
Earth is the only rocky planet where the upper mantle’s structure has been mapped in any real detail. Mars has constraints on its overall interior from geodetic data, but the precision is vastly lower: the radius of Mars’s core, for example, is uncertain by about 150 kilometers, roughly ten times worse than the precision achieved for Earth’s core over a century ago.21Elsevier (Planetary and Space Science). The deep interior of Venus, Mars, and the Earth: A brief review and the need for planetary surface-based measurements Venus is in even worse shape: the most conservative model of its interior is essentially a scaled copy of Earth, adjusted for Venus’s slightly different radius and mass. Without seismometer networks on other planets’ surfaces, the kind of discontinuity-by-discontinuity mapping that defines Earth’s upper mantle simply cannot be done elsewhere. NASA’s InSight lander placed a single seismometer on Mars and detected marsquakes, providing the first direct glimpse of another planet’s interior layering, but a single station gives far less detail than the thousands of stations that cover Earth.
The question of whether other rocky planets even have an upper mantle in the Earth-like sense, with olivine-dominated mineralogy, a transition zone driven by the same phase changes, and a 660-style floor, depends on their bulk composition, interior temperature, and pressure profiles. A planet with a very different iron-to-magnesium ratio or a substantially cooler interior might not undergo the same mineral transformations at the same depths, or at all. Earth’s upper mantle, with its particular thickness and internal architecture, is as much a product of this planet’s specific chemistry and thermal history as it is a universal feature of rocky worlds.