Earth’s mantle spans a temperature range from roughly 1,000°C just beneath the crust to perhaps 3,700°C or more at its base, where it meets the outer core. That range is wider than many people expect, and the numbers shift depending on depth, location, and the type of rock involved. Two main fuel sources keep the mantle hot: leftover heat from the planet’s violent formation about 4.5 billion years ago, and the steady decay of radioactive elements scattered throughout the rock. Neither source alone explains the full thermal budget, and pinning down precise temperatures hundreds of kilometers below the surface remains one of the harder problems in Earth science.
The Temperature Profile From Crust to Core
The shallowest mantle, right below the crust, sits at temperatures in the neighborhood of 1,000°C under the oceans and somewhat less under old, thick continental roots. Temperature climbs with depth, but not at a constant rate. At around 380 km depth, estimates based on seismic data and high-pressure mineral experiments converge on roughly 1,400°C.1Journal of Geophysical Research: Solid Earth. The temperature profile of the upper mantle From there, through the transition zone between about 350 and 655 km, the temperature rises to approximately 1,600°C. That increase is not smooth. Minerals in the mantle undergo structural rearrangements at specific pressures, and these phase changes produce steep jumps in the thermal gradient at depths around 390–415 km and again near 500–530 km.2Geophysical Research Letters. A temperature profile of the mantle transition zone
Below the transition zone, in the lower mantle, temperatures continue to rise more gently along what geophysicists call an adiabatic gradient, meaning the increase is mainly due to compression rather than heat flowing in from outside. By the time you reach the core-mantle boundary at about 2,900 km depth, temperatures are estimated at somewhere between 3,500°C and 4,000°C, though the exact figure depends on which mineral-physics dataset and which geodynamic model you trust. The temperature drop across the thin boundary layer right at the base of the mantle is dramatic, on the order of 1,000 K, because the liquid iron outer core sitting below is significantly hotter still.3Journal of Geophysical Research: Solid Earth. Why Are Plume Excess Temperatures Much Less Than the Temperature Drop Across the Lowermost‐Mantle Thermal Boundary Layer?
Radioactive Decay and Primordial Heat
The mantle stays hot because of two overlapping heat supplies. The first is radioactive decay. Uranium, thorium, and potassium are sprinkled throughout the silicate rock of the mantle and crust, and as their nuclei break apart they release energy. Combined analyses of neutrino detectors, which can actually catch the anti-neutrinos produced by radioactive decay deep inside the planet, favor a present-day total radiogenic power output of about 20 terawatts.4Earth and Planetary Science Letters. Quantifying Earth’s radiogenic heat budget That is a substantial fraction of the total heat the planet loses at its surface, which sits around 46 terawatts.
The second source is primordial heat: energy stored from Earth’s formation, when countless planetesimals slammed together and their kinetic energy converted to thermal energy. The separation of iron from silicate to form the core released even more gravitational energy. Modeling based on the fact that Earth’s core remains largely molten today suggests at least 7% of the accretional energy was retained as primordial heat inside the planet.5Journal of Geophysical Research: Planets. Initial Thermal States of Super‐Earth Exoplanets and Implications for Early Dynamos Over billions of years, that stored energy has been slowly leaking outward through the mantle. The heat flowing across the core-mantle boundary itself, fed by the cooling core, is estimated at roughly 5 to 17 terawatts, with recent geodynamic models placing it closer to 13–15 terawatts.6Geophysical Journal International. Changes in core–mantle boundary heat flux patterns throughout the supercontinent cycle
In short, Earth is losing heat faster than radioactive decay alone can supply. The gap is made up by the planet drawing down its primordial thermal reserves. Present-day heat loss is roughly twice present-day heat production, and the uncertainty on the resulting cooling rate is around 15%.7Earth and Planetary Science Letters. Thermal evolution of the Earth: Secular changes and fluctuations of plate characteristics
How Anyone Can Know What It Is Like Down There
No one has drilled deeper than about 12 km into the Earth, which barely scratches the crust. So how do scientists arrive at temperature estimates for rock hundreds or thousands of kilometers below? Several independent methods exist, and the fact that they broadly agree is what gives researchers confidence.
The workhorse approach uses seismic waves. Earthquakes send vibrations through the planet, and the speed of those waves depends on the density, stiffness, and temperature of the rock they pass through. Hotter rock slows waves down; cooler rock speeds them up. By measuring millions of travel times, scientists build three-dimensional images of seismic velocity throughout the mantle. Converting those velocities to temperatures requires knowing how each candidate mineral responds to heat and pressure, and that conversion is the tricky part. Different assumptions about composition, the presence of partial melt, and how rock deforms over long timescales can shift the inferred temperature by a couple of hundred degrees.8Physics of the Earth and Planetary Interiors. Mantle temperature and density anomalies: The influence of thermodynamic formulation, melt, and anelasticity
A second method involves magnetotellurics, which measures natural electromagnetic signals that penetrate deep into the Earth. Because the electrical conductivity of mantle minerals is temperature-dependent, mapping conductivity variations at depth can yield temperature estimates. This technique has been applied in places like Svalbard and the Canadian Cordillera to produce independent temperature constraints on the upper mantle.9Geochemistry, Geophysics, Geosystems. Magnetotelluric Constraints on the Temperature, Composition, Partial Melt Content, and Viscosity of the Upper Mantle Beneath Svalbard10Earth and Planetary Science Letters. Upper mantle temperature determined from combining mineral composition, electrical conductivity laboratory studies and magnetotelluric field observations: Application to the intermontane belt, Northern Canadian Cordillera
A third approach takes advantage of xenoliths, fragments of mantle rock carried to the surface by volcanic eruptions. Because these chunks last equilibrated at mantle pressures and temperatures before being launched upward, their mineral chemistry records the conditions at depth. A global compilation of xenolith data yields a best-fit mantle potential temperature of about 1,315°C.11Geophysical Research Letters. Xenolith Constraints on the Mantle Potential Temperature and Thickness of Cratonic Roots Through Time That number is a useful benchmark because it represents an average for stable continental interiors, smoothing over local anomalies.
Finally, laboratory experiments using laser-heated diamond anvil cells allow researchers to squeeze tiny mineral samples to pressures matching the deep mantle while simultaneously heating them with focused laser beams. These experiments pin down melting points and phase boundaries of mantle minerals at extreme conditions, which in turn calibrate the temperature scales used by all the other methods.12Journal of Geophysical Research: Solid Earth. Measurement of melting temperatures of some minerals under lower mantle pressures
Hot Spots and Cold Slabs
The temperatures described so far are averages, but the mantle is far from uniform. Some regions are hundreds of degrees hotter or cooler than their surroundings, and these thermal anomalies drive much of the geological activity visible at the surface.
The most conspicuous hot anomalies are mantle plumes, columns of unusually warm rock rising from deep in the mantle. Hawaii is the classic example. Estimates based on the chemistry of lavas erupted there point to a mantle potential temperature around 1,688°C, while Iceland comes in at about 1,637°C. Compare those numbers to the ambient mantle beneath typical mid-ocean ridges, where potential temperatures run closer to 1,453–1,475°C.13Geochemistry, Geophysics, Geosystems. Mantle potential temperatures at Hawaii, Iceland, and the mid‐ocean ridge system, as inferred from olivine phenocrysts: Evidence for thermally driven mantle plumes That means the rock feeding Hawaiian volcanoes is roughly 200–250°C hotter than the rock feeding a typical stretch of the Mid-Atlantic Ridge. Even within the mid-ocean ridge system, potential temperatures are not constant: recent global analyses of ridge basalt chemistry show they range from about 1,250°C to 1,520°C depending on location.14Journal of Petrology. Global Variation of Mid-Ocean Ridge Basalt Chemistry: Conditions and Composition of the Source Mantle and Physical Properties of the Ridges
On the cold side, subducting slabs of oceanic lithosphere plunge back into the mantle at convergent plate boundaries and carry relatively cool rock deep into the interior. The thermal evolution of these slabs is complex and depends on how long subduction has been going on. Early in a subduction zone’s life, the slab cools rapidly at a given depth; over time the cooling rate slows as the system matures.15Geochemistry, Geophysics, Geosystems. Slab Temperature Evolution Over the Lifetime of a Subduction Zone These cold slabs can remain seismically visible deep into the lower mantle, hundreds of millions of years after they began sinking. The temperature contrast between a subducted slab and the surrounding mantle at the same depth can be several hundred degrees.
These hot and cold anomalies are not just curiosities. They are the engines and brakes of mantle convection. Plumes carry heat upward from the core-mantle boundary; slabs carry cold material downward from the surface. Together they keep the mantle churning in a pattern that, at the surface, we experience as plate tectonics.
A Slowly Cooling Planet
Earth’s mantle was substantially hotter in the distant past. The clearest evidence comes from komatiites, ultra-hot lavas found almost exclusively in rocks older than about 2.5 billion years. These magmas formed by extremely high degrees of partial melting, on the order of 30–50%, which requires mantle temperatures well above anything seen today.16PubMed Central. Archean komatiite volcanism controlled by the evolution of early continents Their near-disappearance from the geological record after about 2 billion years ago is widely interpreted as a sign that the mantle cooled below the threshold needed to produce them.
More precise thermal-history modeling shows a steady decline in average mantle temperature over roughly the last 2.5 billion years.17Geology. Correlating mantle cooling with tectonic transitions on early Earth The rate of cooling is modest by everyday standards. Present-day estimates range from about 7 to 210 K per billion years depending on the model and assumptions, with the wide spread reflecting honest uncertainty in the mantle’s bulk composition, its viscosity, and how efficiently plate tectonics transports heat.6Geophysical Journal International. Changes in core–mantle boundary heat flux patterns throughout the supercontinent cycle Even at the high end, a drop of a couple of hundred degrees per billion years is imperceptibly slow from a human perspective, but over geologic time it reshapes how the planet works.
One subtle effect of secular cooling is on ocean-floor depth. Even temperature variations as small as about 20 K, accumulated over the age of old oceanic crust, can measurably alter ocean bathymetry because deep topography is sensitive to the thermal state of the underlying mantle.18Journal of Geophysical Research: Solid Earth. A New Reference Model for the Evolution of Oceanic Lithosphere in a Cooling Earth
Water’s Outsized Effect on Melting
When people talk about the mantle being “hot enough to melt,” an important qualifier often gets lost. Whether rock melts depends not just on temperature but on pressure and volatile content, particularly water. The upper mantle contains small but consequential amounts of dissolved water in its minerals, typically in the range of 50 to 200 micrograms per gram. That sounds trivial, but relative to its low concentration, water has a disproportionate effect on the temperature at which mantle rock begins to melt.19PubMed. Experimental constraints on the damp peridotite solidus and oceanic mantle potential temperature
This matters because most of the mantle is not actually molten. It is solid rock that flows very slowly over millions of years. Partial melting, where a small percentage of the rock liquefies, occurs only in specific settings: beneath mid-ocean ridges where pressure drops as the mantle rises, above subduction zones where water released from the descending slab lowers the melting point of the overlying mantle wedge, and beneath hot spots where anomalously warm plumes push temperatures above the local solidus. The rest of the mantle stays solid despite being extraordinarily hot, because the immense pressure at depth raises the melting point faster than the temperature rises.
Mantle Heat and Earth’s Magnetic Field
The mantle’s temperature has consequences that extend well beyond volcanoes and plate motions. Earth’s magnetic field, generated by convection in the liquid iron outer core, depends critically on how much heat flows from the core into the base of the mantle. If that heat flow drops too low, thermal convection in the core weakens, and the rate at which the solid inner core grows, releasing the compositional buoyancy that also drives convection, slows as well. Below a certain threshold, the dynamo that sustains the magnetic field could stall.20Geophysical Research Letters. Estimates of heat flow in the deep mantle based on the power requirements for the geodynamo
The power needed to run the geodynamo thus places a floor on the heat passing across the core-mantle boundary, which in turn constrains models of Earth’s thermal history.21Geophysical Journal International. Can the Earth’s dynamo run on heat alone? In other words, the fact that we have a magnetic field today tells us something about how hot the deep mantle is and how efficiently it conducts heat away from the core. The magnetic field is not a passive byproduct of a hot interior; it is an active constraint on what the mantle’s thermal state can and cannot be.
How Venus Offers a Contrast
A useful way to appreciate Earth’s mantle temperatures is to compare them with those of a neighboring planet. Venus is nearly the same size and bulk composition as Earth, but it lacks plate tectonics. Its lithosphere appears to be a single rigid shell, with no subduction recycling cool surface rock back into the interior. Numerical models of mantle convection show that this difference alone would make Venus’s mantle hotter than Earth’s, because a rigid lid insulates the interior more effectively than a system of mobile plates that actively transport heat.22Journal of Geophysical Research: Solid Earth. Mantle dynamics in Mars and Venus: Influence of an immobile lithosphere on three‐dimensional mantle convection Venus also starts with a higher surface temperature, around 460°C thanks to its runaway greenhouse atmosphere, which further reduces the temperature gradient driving heat loss.
The implication is that plate tectonics is not just a consequence of a warm mantle but also a thermostat. By cycling cool material into the deep interior and bringing hot material to the surface where it can radiate heat to space, plate motion regulates how quickly a rocky planet cools. Without it, a planet of Earth’s size could retain significantly more internal heat over the same span of time, potentially altering whether and how long a magnetic dynamo can operate, and by extension, whether the planet remains habitable.
Mantle Viscosity and the Speed of Convection
Temperature alone does not tell the full story of how the mantle behaves. Viscosity, the resistance of rock to flow, drops steeply as temperature rises. A region of the mantle that is a few hundred degrees warmer than its surroundings can be orders of magnitude less viscous, which changes how fast convection currents move and how much stress they transmit to tectonic plates. Constraining the relationship between temperature and viscosity at mantle conditions remains one of the harder problems in geophysics, partly because viscosity depends not just on temperature but on pressure, grain size, water content, and rock composition.23PubMed Central. Constraining Earth’s nonlinear mantle viscosity using plate-boundary resolving global inversions
The practical result is that mantle convection is not a simple uniform circulation. It is a sluggish, uneven flow where warmer, less viscous regions move faster and cooler, stiffer regions resist deformation. The asthenosphere, the relatively weak layer in the upper mantle that lets tectonic plates slide around, exists partly because temperatures there are close enough to the solidus that the rock softens dramatically. Whether that softening is purely a temperature effect or also involves small amounts of partial melt or dissolved volatiles is still debated.24Geophysical Journal International. Thermal nature and resolution of the lithosphere–asthenosphere boundary under the Pacific from surface waves The answer has implications for how quickly plates can accelerate or slow down in response to changes in mantle flow, and for how lateral temperature variations translate into the forces that ultimately drive earthquakes and build mountains.