Earth's Layers Explained: Crust, Mantle, Outer Core and Inner Core

Earth is built in four main layers: a thin, brittle rocky crust (about 5 km thick under the oceans and about 30 km on average under continents), a solid but slowly flowing rock mantle about 2,900 km thick, a liquid iron and nickel outer core about 2,200 km thick, and a solid iron inner core with a radius of about 1,221 km. Pressure climbs from one atmosphere at the surface to roughly 364 GPa at the centre, and the inner core boundary sits near 6,000 K.

Nobody has ever been there. The deepest hole ever drilled, the Kola superdeep borehole, reached 12,262 m, which is about 0.2 percent of the way to the centre. Everything below that comes from earthquake waves, the planet's mass and spin, meteorites, rare deep rocks and laboratory experiments that squeeze iron and rock between diamonds. This post walks through each layer with numbers from primary sources, shows how each number was obtained, flags where the sources disagree, and finishes with a short walkthrough you can run to see the core's seismic shadow for yourself.

Key takeaways

• The crust is about 1 percent of Earth's volume, the mantle about 84 percent and the core about 15 percent, according to the US Geological Survey (USGS).
• Oceanic crust is thin (about 5 km) and basaltic. Continental crust averages about 30 km and can reach 100 km under large mountain ranges.
• The Moho, discovered by Andrija Mohorovičić in 1909, is the boundary where earthquake waves speed up sharply as they leave crust and enter mantle.
• The lithosphere (crust plus the cold, rigid top of the mantle) is broken into about a dozen plates that move a few centimetres a year over the softer asthenosphere.
• The mantle is solid rock that flows over geological time. It is not a sea of magma.
• R. D. Oldham identified the core in 1906. Inge Lehmann found the solid inner core in 1936 from P waves arriving where the liquid outer core should have left a shadow.
• Laboratory work puts the melting point of iron at the inner core boundary at 6,230 ± 500 K, at about 330 GPa.
• Recent peer reviewed studies disagree on how the inner core rotates relative to the mantle and how old it is. Treat headlines on both with care.

Earth's layers at a glance

Numbers below come from the USGS, NASA's Earth Fact Sheet and peer reviewed abstracts listed in Sources. Where a value is our own back of envelope estimate rather than a published one, the table says so. Depths are measured from the surface; NASA gives Earth's mean radius as 6,371 km and its core radius as 3,485 km, which puts the core mantle boundary near 2,886 km.

Two quick checks make the table less abstract. First, Earth's mean density is 5,513 kg per cubic metre according to NASA, which is roughly double the density of surface rocks. Isaac Newton had already reasoned, according to the USGS, that the interior must hold much denser material. Second, NASA lists Earth's moment of inertia factor as 0.3308. A uniform ball would score 0.4. The lower number means the mass is concentrated toward the centre, which is exactly what a dense iron core predicts.

How the pressure estimates in the table work

The "our estimate" pressures use the simplest rule in geophysics: pressure at depth equals density times gravity times depth. For the continental crust, take a density of about 2,800 kg per cubic metre, gravity of 9.8 metres per second squared and a depth of 35 km:

P = rho * g * h
  = 2800 kg/m^3 * 9.8 m/s^2 * 35000 m
  = 9.6e8 Pa
  = about 0.96 GPa (roughly 9,500 times sea level air pressure)

The same rule gives about 3 GPa at 100 km and roughly 14 to 23 GPa across the transition zone. It breaks down deeper, because density rises with depth and gravity changes inside the planet, so for the deep mantle and core the table uses published experimental and modelling values instead.

What is the crust, and how do continental and oceanic crust differ?

The crust is the only layer anyone has sampled directly, and it is astonishingly thin. The USGS compares the layered Earth to a boiled egg, then points out that the crust is thinner than an eggshell in proportion. It accounts for about 1 percent of Earth's volume.

There are two very different kinds.

Oceanic crust sits under the deep sea floor. The USGS says it "varies little in thickness," generally about 5 km. Off the island of Hawaii it is about 5 km thick. It is made almost entirely of basalt, a dark volcanic rock, and it is constantly being made at mid ocean ridges and destroyed where it sinks back into the mantle at subduction zones. That recycling keeps oceanic crust young compared with the continents.

Continental crust is thicker, lighter and older. The USGS gives an average of about 30 km, NASA about 31 km. Under large mountain ranges such as the Alps or the Sierra Nevada the base of the crust can sit as deep as 100 km. Under eastern California alone the thickness ranges from about 25 km under the Great Valley to 60 km under the Sierra Nevada, according to the USGS. USGS Circular 532 (1966) describes continental crust as largely granite like rock in its upper part with denser basaltic rock below.

Because continental crust is less dense than the mantle, thick crust floats higher, which is part of why mountain ranges have deep roots. Oceanic crust is denser, which is why it can sink back into the mantle while continents mostly cannot.

What is the Moho?

The Mohorovičić discontinuity, usually shortened to the Moho, is the boundary between crust and mantle. USGS Circular 532 records that the Croatian seismologist Andrija Mohorovičić found it in 1909 by noticing a sharp increase in the speed of earthquake waves as they passed from crust into mantle. No one has seen the Moho. It is detected only by that change in wave speed, which the USGS explains as a change in rock type.

The Moho is shallow under oceans (a few kilometres below the sea floor) and deep under continents. In the 1960s scientists proposed drilling through it at sea, where the crust is thinnest; the 1966 USGS circular mentions the planned "Mohole" site with about 4 miles of crust under about 2.5 miles of water. The Moho has still not been reached by drilling.

What are the lithosphere and the asthenosphere?

Crust and mantle are labels based on composition. Lithosphere and asthenosphere are labels based on how the rock behaves mechanically. The two sets of names overlap, and confusing them is one of the most common mistakes in school geology.

The lithosphere is the crust plus the cooler, rigid top of the mantle. The USGS describes it as averaging at least 80 km thick over much of the Earth, thinnest under the oceans and in volcanically active continental regions such as the western United States. It behaves like one stiff shell, and that shell is broken into plates.

The asthenosphere lies beneath it. The USGS describes a zone between about 100 and 200 km down where the rock is near its melting point, earthquake waves slow slightly, and hot rock can soften and flow over geological time. Seismologists call this the low velocity zone. Molten rock erupted by some volcanoes originates in this region, according to the USGS. The asthenosphere is not liquid. It is solid rock that is weak enough to deform slowly, and the plates ride on it.

What is inside the mantle?

The mantle is the bulk of the planet: about 84 percent of its volume and about 2,900 km thick. The USGS calls it a dense, hot layer of semi solid rock containing more iron, magnesium and calcium than the crust. NASA's description of it as "a hot, viscous mix of molten rock" is a simplification; the seismic evidence is that it carries shear waves almost everywhere, and liquids cannot do that. Think of it as solid rock that creeps, like a glacier, over millions of years.

The mantle has internal structure that comes from pressure changing the crystal structure of the same chemical ingredients.

Upper mantle

The upper mantle runs from the Moho to about 410 km. Its dominant mineral is olivine, a magnesium and iron silicate, inside a rock called peridotite. The asthenosphere sits in the upper part of this zone.

The mantle transition zone (410 to 660 km)

Between about 410 and 660 km, olivine is squeezed into denser crystal structures called wadsleyite and then ringwoodite. These phase changes cause two jumps in seismic wave speed, the "two discontinuities" the USGS describes in the transition zone.

The transition zone matters for water. Theory and experiments show that wadsleyite and ringwoodite can hold water up to around 2.5 percent of their weight. In 2014, Pearson and colleagues reported in Nature the first natural sample of ringwoodite from Earth, a tiny inclusion inside a diamond from Juína, Brazil. Its infrared signature showed about 1 weight percent water, which the authors called direct evidence that, at least locally, the transition zone is hydrous. They were careful about scope: earlier geophysical probes had given conflicting results, and one inclusion shows local conditions, not a global average.

Lower mantle

Below 660 km the minerals change again into still denser iron and magnesium silicates and oxides. The USGS describes these as relatively simple minerals that "change gradually with depth to very dense forms." This is the largest single region of the planet by volume. It convects extremely slowly, carrying heat out of the core and returning sinking slabs of old ocean floor toward the bottom.

The D double prime layer and the core mantle boundary

The lowest 200 to 300 km of the mantle is called D double prime (often written D''). Seismologists noticed a jump in wave speed a few hundred kilometres above the core and strong directional differences in wave speed there. In 2004, Murakami and colleagues reported in Science that MgSiO3 perovskite turns into a new "post perovskite" structure above 125 GPa and 2,500 K, conditions they equate with about 2,700 km depth. A companion first principles study by Iitaka and colleagues in Nature the same year found the post perovskite phase becomes stable above 98 GPa at zero temperature and could explain the D double prime discontinuity and its anisotropy. Both papers say "may": it is a strong candidate explanation, not a closed case.

The bottom of the mantle is also lumpy on a continental scale. Seismic images show two continent sized regions of slow wave speed there, the large low velocity provinces. In 2023, Yuan and colleagues argued in Nature that these could be buried remains of the mantle of Theia, the body whose collision with the young Earth formed the Moon. Their simulations show it is possible; the abstract uses "may" throughout, and this is one hypothesis among several.

Smaller patches right at the core mantle boundary, ultralow velocity zones, slow seismic waves even more. In 2010, Fiquet and colleagues melted natural peridotite at up to 140 GPa in a laser heated diamond anvil cell and found it melts at 4,180 ± 150 K at core mantle boundary pressures, a value that matches estimated temperatures there. Their conclusion: molten regions may exist at the base of today's mantle.

Going from mantle to core is the biggest change inside the planet. The USGS describes a drop of about 30 percent in earthquake wave speed and an increase of about 30 percent in density across this boundary.

What is the outer core made of, and why is it liquid?

The outer core is a shell of liquid metal about 2,200 to 2,250 km thick. The USGS describes the whole core as nearly twice as dense as the mantle because it is metallic rather than stony, mostly iron with about 10 percent of an alloying element such as oxygen, sulfur or nickel, or a mix. USGS Circular 532 gives a core density of about 11 grams per cubic centimetre and lists the reasons iron is the best candidate: it has about the right density and P wave speed, it is abundant in the Sun and stars, it is common in meteorites, and it conducts electricity well enough to make a magnetic field.

Pure iron is too dense to match the seismic data, which is how we know light elements are in there. A paper published in PNAS in April 2026 by Gan and colleagues measured shock temperatures of molten iron up to about 364 GPa. Comparing pure molten iron with the observed outer core, they found the outer core has a density deficit that grows with depth to about 9 percent at the inner core boundary. They also reported that the lowest roughly 280 km of the outer core behaves differently, consistent with a partially crystallised "F layer" topped by a gradient in light element content.

We know the outer core is liquid because it stops shear waves. The USGS puts it plainly: the outer core is presumed liquid because it does not transmit S waves and because P waves slow sharply when they enter it. The next section shows how that creates a shadow on the far side of the planet.

What is the geodynamo?

The liquid outer core is where Earth's magnetic field is made. NASA describes the iron and nickel moving in whirlpools driven by Earth's rotation, generating electrical currents hundreds of miles wide, and calls this the geodynamo. Convection is driven partly by heat leaving the core and partly by the inner core freezing: Pham and Tkalčić note in Nature Communications that as the inner core grows, the latent heat and light elements released by solidification drive outer core convection, which keeps the dynamo going.

At the surface the field is roughly a dipole with a strength between about 0.24 and 0.66 gauss, according to NASA's Earth Fact Sheet. It is not fixed. ESA's three satellite Swarm constellation, launched in November 2013, confirmed in its first results in June 2014 a general weakening of the field, strongest over the Western Hemisphere, with strengthening in other areas such as the southern Indian Ocean, and continued movement of magnetic north toward Siberia. NASA also describes the South Atlantic Anomaly, a weak spot where charged particles come closer to the surface and can damage spacecraft electronics, while not affecting life on the ground.

What is the inner core, and how can it be solid when it is the hottest part?

The inner core is a solid ball of iron alloy with a radius of about 1,221 km. NASA gives 1,221 km; the USGS This Dynamic Earth page gives 1,250 km. It holds less than 1 percent of Earth's volume, according to Pham and Tkalčić.

It is solid because pressure wins over temperature. Iron's melting point rises steeply with pressure. At the inner core boundary, pressure is about 329 to 330 GPa. Anzellini and colleagues used laser heated diamond anvil cells and fast X ray diffraction in experiments up to 200 GPa, then extrapolated to conclude in Science (2013) that iron melts at 6,230 ± 500 K at the inner core boundary. Below that boundary the pressure is even higher, so iron stays solid despite being extremely hot. NASA describes the inner core as about as hot as the Sun's surface, which is a fair comparison for a number near 6,000 K.

The crystal form of that iron is also studied in the lab. In 2010, Tateno and colleagues reported in Science that hexagonal close packed (hcp) iron is stable up to 377 GPa and 5,700 K, matching inner core conditions, and that its directional stiffness could explain why seismic waves travel faster along some directions through the inner core than others.

Does the inner core have its own inner core?

Possibly. In February 2023, Pham and Tkalčić reported waves that bounced back and forth along Earth's diameter up to five times, picked out by stacking records from a growing global seismic network. From the travel times they inferred an innermost ball about 650 km in size with P waves about 4 percent slower at around 50 degrees from the rotation axis, wrapped in an outer shell with much weaker directional effects. They describe this as strengthening the evidence for a distinct innermost inner core that could be a fossil record of a major event in Earth's past.

Is the inner core spinning differently from the rest of the planet?

This is a live research question, and the peer reviewed papers do not all say the same thing.

• In January 2023, Yang and Song reported in Nature Geoscience that seismic paths which had shown changes in earlier decades showed little change over the past decade. They read this as the inner core's differential rotation recently pausing, as part of an oscillation of roughly seven decades with an earlier turning point in the early 1970s, which lines up with changes in the length of day and the magnetic field.
• In June 2024, Wang, Vidale and colleagues reported in Nature on 143 pairs of repeating earthquakes in the South Sandwich Islands between 1991 and 2023. Waveforms that changed and later reverted told them when the inner core returned to an earlier position. Their reading: the inner core super rotated (spun slightly faster than the mantle) from 2003 to 2008, then from 2008 to 2023 sub rotated two to three times more slowly back along the same path. They say new models will be needed to explain it.

Both studies describe very small relative motions over years and decades. Neither means the core has "stopped spinning". The whole Earth, inner core included, still rotates once a day; the debate is about a tiny difference in rate between the inner core and the mantle.

How old is the inner core?

Also disputed. Both studies below appeared in Nature on 2 June 2016 and point in different directions.

• Ohta and colleagues measured the electrical resistivity of iron at up to 4,500 K and megabar pressures, inferred a high thermal conductivity for the core, and concluded that the core cooled quickly and the inner core is young, less than 0.7 billion years old.
• Konôpková and colleagues measured the thermal conductivity of solid iron directly and placed it near the low end of earlier estimates, at 18 to 44 watts per metre per kelvin. That result, they say, agrees with evidence that the geodynamo has run since early in Earth's history and allows an inner core as old as the dynamo.

Until those measurements converge, any single number for the inner core's age should be read as model dependent.

How do scientists know what is inside the Earth?

The USGS lists the main lines of evidence: the paths and character of earthquake waves; laboratory experiments on rocks and minerals at high pressure and temperature; surface geology; Earth's motions in the Solar System, its gravity and magnetic fields; and the flow of heat from the interior. Here is how each one works.

Earthquake waves: P waves and S waves

An earthquake sends out two kinds of body waves. USGS Circular 532 describes them simply: compressional P waves, in which rock moves back and forth along the direction of travel, and shear S waves, in which rock moves across it. P waves go through solids and liquids. S waves need a material that resists shearing, so they cannot cross a liquid.

Wave speed depends on what the rock is made of and how compressed it is. Because speed changes with depth, rays bend as they travel, and some reflect off boundaries. Seismologists name the paths: PKP is a P wave that crosses the core, SKS is an S wave that converts to a P wave inside the liquid core and back to S on the way out, and PKIKP is a P wave that crosses the inner core. By timing thousands of these paths from many earthquakes at many stations, they solve for wave speed at every depth. A widely used result is the Preliminary Reference Earth Model (PREM) of Dziewonski and Anderson, published in 1981 and still bundled with seismology software today.

The shadow zones

The shadow zones are the cleanest single piece of evidence for a liquid outer core. According to IRIS (now part of EarthScope), S waves do not reappear beyond an angular distance of about 103 to 104 degrees from the earthquake, because they are stopped by the liquid outer core. P waves do not arrive directly between about 103 and 140 degrees, because they bend sharply where they enter the core. IRIS notes that this pattern led scientists in the early 1900s to deduce a liquid outer core.

Lehmann and the inner core (1936)

USGS Circular 532 records that R. D. Oldham discovered the core in 1906 from earthquake records, and that "Miss I. Lehmann" discovered the inner core in 1936 from its higher P wave speed. The Danish seismologist Inge Lehmann noticed faint P waves arriving inside the P wave shadow zone where a simple liquid core model said there should be none. She explained them as waves reflected and refracted by a smaller, faster inner body. That body is the solid inner core. The walkthrough later in this post reproduces the effect with a modern Earth model: PKIKP arrivals appear in the shadow zone at distances where the direct P wave is gone.

Meteorites

USGS Circular 532 cites the large proportions of iron in meteorites as one reason iron is the leading candidate for the core. Meteorites give a sample of the kind of material planets formed from, and iron meteorites show what metal separated from rock in early planetary bodies looks like.

Deep samples from below the crust

Volcanoes and kimberlite pipes bring up fragments of the mantle. The most dramatic example is the ringwoodite in a Brazilian diamond reported by Pearson and colleagues in 2014. Its preservation, the authors note, also tells us some kimberlites must have their sources in the transition zone, hundreds of kilometres down.

Laboratory experiments

The diamond anvil cell squeezes a tiny sample between two diamond tips and heats it with lasers, reaching pressures of hundreds of gigapascals. Every deep number in this post's table that is not seismic comes from experiments like these: post perovskite (Murakami 2004), peridotite melting at the core mantle boundary (Fiquet 2010), hcp iron at inner core conditions (Tateno 2010), the melting point of iron at the inner core boundary (Anzellini 2013) and the conductivity of core iron (Ohta 2016, Konôpková 2016). Shock compression experiments, such as the 2026 PNAS work of Gan and colleagues, reach similar pressures for a split second.

Boreholes: how deep have we actually drilled?

The Kola superdeep borehole SG-3 in northwest Russia reached 12,262 m, according to a 2020 Scientific Reports paper by Prokofiev and colleagues that studied gold bearing rocks of Archaean age found in its core between 9,500 and 11,000 m. The USGS describes it as a 12 km Soviet hole and notes that drilling cost rises enormously with depth. Twelve kilometres is well short of the roughly 30 km average thickness of continental crust and about 0.2 percent of the distance to Earth's centre. Kola did not reach the Moho, and no hole has.

A dated timeline of how we learned about Earth's interior

The 1966 numbers are worth a second look. Three million atmospheres is about 304 GPa, and 8,000 °F is about 4,700 K. Today's laboratory values for the inner core boundary are about 330 GPa and about 6,230 K. The pressure estimate has held up well. The temperature estimate rose by about 1,500 K as experiments improved. That is normal science: the method stays the same and the error bars shrink.

How does plate tectonics connect to the layers?

Plate tectonics is what the lithosphere does on top of a convecting mantle. The USGS defines a plate as a large, rigid slab of solid rock and describes the outermost layer as fragmented into "a dozen or more" plates moving relative to one another as they ride on hotter, more mobile material. The plates include the crust and part of the upper mantle.

The idea had a long road. The USGS traces continental drift to Alfred Wegener, whose theory was introduced in 1912. According to the USGS, it was poorly received because Wegener could not explain what force could move such large masses of solid rock; the geophysicist Harold Jeffreys argued that continents could not plow through the ocean floor without breaking up. Wegener died in 1930 on the Greenland ice cap. Later evidence from ocean floor exploration revived the idea and led to plate tectonics.

How fast do plates move? The USGS gives some benchmark figures:

The USGS also notes that plates move slower than fingernails grow, and that their edges are marked by concentrations of earthquakes and volcanoes. Collisions build mountains like the Himalayas. Sinking slabs at subduction zones carry old ocean floor, and water, back into the mantle, which is one reason the transition zone may hold water.

Step by step: see the core's shadow and Lehmann's discovery yourself

You do not need a lab to reproduce the reasoning that found the core. The open source ObsPy library ships TauP, a travel time calculator, with several published Earth models including PREM. The steps below are safe, read only and run in a throwaway virtual environment.

1. Create an isolated Python environment

python3 -m venv taupenv
./taupenv/bin/pip install obspy
./taupenv/bin/pip show obspy | grep -E '^(Version|Requires)'

When we ran this on 23 October 2026, pip installed ObsPy 1.5.1, which pulled in numpy, scipy, matplotlib, lxml, sqlalchemy, requests, decorator and setuptools. That list matters later.

2. Compute which seismic phases arrive at each distance

Save this as shadow.py in a separate folder from the environment:

from obspy.taup import TauPyModel

model = TauPyModel(model="prem")
phases = ["P", "S", "Pdiff", "PKP", "PKIKP", "SKS"]

for distance in (30, 60, 90, 100, 110, 120, 130, 140, 150, 170):
    arrivals = model.get_travel_times(
        source_depth_in_km=10,
        distance_in_degree=distance,
        phase_list=phases,
    )
    names = sorted({a.name for a in arrivals})
    first = min(a.time for a in arrivals) / 60 if arrivals else None
    print(distance, names, round(first, 1) if first else None)

Run it:

./taupenv/bin/python -I shadow.py

3. Read the result

Our run with PREM and a 10 km deep source printed:

30 ['P', 'S'] 6.1
60 ['P', 'S'] 10.1
90 ['P', 'S', 'SKS'] 13.0
100 ['Pdiff', 'S', 'SKS'] 13.7
110 ['Pdiff', 'SKS'] 14.5
120 ['PKIKP', 'Pdiff', 'SKS'] 15.2
130 ['PKIKP', 'Pdiff', 'SKS'] 15.9
140 ['PKIKP', 'Pdiff', 'SKS'] 16.7
150 ['PKIKP', 'PKP', 'Pdiff'] 17.4
170 ['PKIKP', 'PKP'] 20.1

Three things to notice:

• The direct P wave is gone by 100 degrees, replaced by Pdiff, a weak wave diffracted around the core.
• The direct S wave is gone beyond 100 degrees. Only SKS, which crosses the core as a P wave, gets through. That is the S wave shadow, and it is the evidence for a liquid outer core.
• PKIKP, the P wave that crosses the inner core, shows up at 120 to 140 degrees, inside the P shadow zone. That is the kind of arrival Lehmann explained in 1936.

The last number on each line is the first arrival time in minutes. A wave takes about 20 minutes to cross to the far side of the planet.

4. Pull real earthquakes to test against

The USGS publishes a free event API. This read only query lists the three largest earthquakes of magnitude 6.5 or more in September 2026:

curl -s "https://earthquake.usgs.gov/fdsnws/event/1/query?format=geojson&starttime=2026-09-01&endtime=2026-10-01&minmagnitude=6.5&orderby=magnitude&limit=3"

Each result includes latitude, longitude and depth. Feed the depth into source_depth_in_km and the distance to a seismic station into distance_in_degree, then compare predicted times with records from the open seismic data services run by EarthScope.

5. Pin what you installed

Research code is software, and it depends on other software. Record exact versions so a result can be reproduced and checked later:

./taupenv/bin/pip freeze > requirements.txt

If your team uses Safeguard, safeguard scan --dir . in the project folder builds an inventory of those dependencies and matches them against vulnerability databases, with results written as JSON under ./output. That checks the code around the science; it says nothing about whether your Earth model is right.

Where Safeguard fits, honestly

Safeguard does not do geoscience. It does not process seismograms, model the mantle or check whether a travel time is correct. The overlap is narrow and real: the instruments that measure the Earth run software. Seismic station firmware, data processing pipelines, magnetometer ground segments and research notebooks are built from open source components, as the eight dependencies pulled in by one ObsPy install show. Safeguard scans repositories, container images and SBOM files, matches components against vulnerability databases, ranks findings with reachability analysis, and shows which projects carry a given component and version. That is useful when a numerical or network library gets an advisory and an observatory needs to know which processing hosts and containers run it. The planet's layered structure is also a fair picture of defence in depth: no single shell carries the load, and you learn what is inside by watching how signals pass through each boundary.

Common mistakes

• "The mantle is a sea of magma." The mantle is solid rock that transmits S waves. It flows only over geological time. Magma forms in limited places, such as near the 100 to 200 km zone the USGS describes, under ridges and above subduction zones.
• "Crust and lithosphere are the same thing." Crust is defined by composition; lithosphere by rigidity. The lithosphere includes the top of the mantle and averages at least 80 km thick, far thicker than oceanic crust.
• "The asthenosphere is liquid." It is weak, hot, solid rock near its melting point, possibly with small amounts of melt.
• "The inner core is solid because it is cold." It is about as hot as the Sun's surface. It is solid because pressure raises iron's melting point faster than temperature rises with depth.
• "Lava comes from the core." Volcanic magma comes from the upper mantle and crust. The core is metal; lava is molten silicate rock.
• "The inner core stopped spinning." The whole Earth still rotates. Studies disagree about tiny changes in the inner core's rotation relative to the mantle over years and decades.
• "Kola reached the mantle." At 12,262 m it did not reach the Moho, and no borehole has.
• "Everyone agrees on the thickness numbers." Published values differ by tens of kilometres. The USGS gives 2,200 km for the outer core and 1,250 km for the inner core; NASA gives 2,250 km and 1,221 km. Quote the source with the number.
• "The table's temperatures are measured." Deep temperatures are inferred from experiments and models, with error bars of hundreds of kelvin.

FAQ

What are the four layers of the Earth?

The crust, the mantle, the outer core and the inner core. The crust is thin rock, the mantle is about 2,900 km of solid but slowly flowing rock, the outer core is liquid iron and nickel about 2,200 km thick, and the inner core is a solid iron ball with a radius of about 1,221 km.

How thick is the Earth's crust?

About 5 km under the oceans and about 30 km on average under continents, according to the USGS. Under large mountain ranges such as the Alps, the base of the crust can be as deep as 100 km.

How hot is the Earth's core?

Laboratory experiments put the melting point of iron at the inner core boundary at 6,230 ± 500 K, about 5,960 °C, at a pressure of about 330 GPa (Anzellini and colleagues, Science, 2013). The top of the outer core is cooler, around 4,000 K. NASA compares the inner core to the surface of the Sun.

Why is the inner core solid if it is so hot?

Pressure. The inner core is under 329 to 364 GPa, more than three million times sea level air pressure. At that pressure iron's melting point is higher than the actual temperature, so it freezes solid even near 6,000 K.

How do we know what is inside the Earth if no one has been there?

Mostly from earthquake waves, whose speeds, paths and shadow zones reveal layers and their state. Earth's mass, spin and magnetic field, meteorites, deep rock and diamond samples, and high pressure laboratory experiments fill in composition and temperature.

What is the deepest hole ever drilled?

The Kola superdeep borehole SG-3 in northwest Russia, at 12,262 m. That is about 0.2 percent of the distance to the centre of the Earth.

Who discovered the Earth's inner core?

Inge Lehmann, a Danish seismologist, in 1936. She explained faint P waves arriving inside the shadow zone of the core as waves bent by a solid inner core.

Is the Earth's magnetic field weakening?

ESA's Swarm satellites confirmed in 2014 a general weakening trend, strongest over the Western Hemisphere, while the field strengthened in some other areas and magnetic north kept moving toward Siberia. The field is generated by the geodynamo in the liquid outer core, and NASA notes that the magnetic poles are continually moving.

Fix it without waiting for upstream

The Earth's layers have no dependencies, but the software that watches them does: seismic processing libraries, station firmware build chains and the containers that run observatory pipelines all ship open source components. The fit here is indirect, so read what follows as the same patch discipline applied to scientific and sensor software, not to geoscience.

Whatever the vulnerability, most of the risk sits in two gaps. The first is the time between a fix being published and that fix actually running in your production builds. The second, and worse, is the time before any fix exists at all, while a maintainer works out a patch and a release schedule. Waiting through both gaps is a habit, not a requirement. Safeguard gives teams three ways to close them.

1. Autonomous remediation: pull requests, not tickets

When a fixed version exists, Safeguard's Griffin models open a tested, policy gated pull request that moves the affected dependency to the safe version. You can run it on one repository or in bulk across every project that carries the vulnerable version. Remediation covers eight categories, including dependencies, licenses, security posture and findings. Reachability analysis ranks the vulnerable code paths your application actually calls first, so the first pull requests you merge are the ones that matter most. Autonomous remediation ships switched off, with every category set to Manual, so your team decides how much runs on its own and what still needs a human approval.

2. Gold Open Source: the safe version in seconds, free

gold.safeguard.sh is free and needs no login. Search any CVE or package to see affected and fixed versions, CVSS, EPSS, CISA KEV status and a recommended safe version, or use the "Am I affected?" checker. The same data is available through a free read API, README status badges and RSS threat feeds. The safeguard-sh/gold-scan-action GitHub Action can fail a build on a known exploited vulnerability, on any known vulnerability, or never, so a version you have just removed cannot quietly come back in next week's merge.

3. Gold Registry: start from components that are already fixed

registry.safeguard.sh provides hardened, drop in replacements for upstream packages and container images. Every Gold artifact is held continuously to zero exploitable critical or high CVEs, zero malware and a completed license check, with SLSA L3 or higher provenance, a Sigstore signature and a public evidence page you can hand to an auditor. An artifact loses its Gold badge if it regresses. Teams can point their own package directory at the registry so hardened equivalents resolve without changing how developers install anything.

When upstream has no fix yet, or will not ship one, Safeguard first works to get the fix accepted upstream. If the issue is severe or upstream will not move, Safeguard builds a fork, fixes the vulnerability there and publishes a hardened Gold build, so you are not stuck on a maintainer's release schedule. You can also request a hardened, compatibility tested build of a specific package. This fork and patch flow is in early access today, so ask which packages it covers before you plan around it.

One honest limit: when the vulnerable code lives in a vendor appliance, a hosted service or an operating system runtime you do not build yourself, the fix still has to come from that vendor or distribution. The steps above apply to the open source dependencies and container images you build and ship, which is where most teams carry the largest and least tracked share of their exposure.

See how Safeguard would help

The first demo sets up Safeguard's local runner on your own machine, approves it with a device code, and runs a scan from the terminal. On a workstation that processes seismic or magnetometer data, the same flow lists the packages a tool like ObsPy brought in. It covers the software, not the science.

Click through it at your own pace, or open it full screen.

The second demo is a real Safeguard CLI session: asked to fix a lodash CVE, it reads the finding and advisory, asks before each command, edits package.json to lodash 4.17.21, runs npm install and npm test with approval, prints a summary and pushes nothing. The same approve each step flow applies when a dependency in a research data pipeline or dashboard needs an upgrade.

Click through it at your own pace, or open it full screen.

The third demo connects Claude to the Safeguard MCP server with OAuth and lets you choose which tools Claude can use. That lets a team ask which of its projects carry a given package version while keeping control over what the assistant can touch.

Click through it at your own pace, or open it full screen.

To try it on your own projects, download Safeguard at safeguard.sh/download and read the CLI documentation.

Quick check

Sources

• USGS: This Dynamic Earth, Inside the Earth
• USGS: The Interior of the Earth (General Interest Publication)
• USGS Circular 532: The Interior of the Earth, an Elementary Description (Robertson, 1966, PDF)
• USGS: This Dynamic Earth, Understanding plate motions
• USGS: This Dynamic Earth, Historical perspective
• USGS: Earthquake Catalog event web service (FDSN)
• NASA NSSDCA: Earth Fact Sheet
• NASA Science: Earth's Magnetosphere, Protecting Our Planet from Harmful Space Energy
• ESA: Swarm reveals Earth's changing magnetism (19 June 2014)
• IRIS (EarthScope): Seismic Shadow Zone, Basic Introduction
• ObsPy documentation: obspy.taup travel time models
• Wang, Vidale et al., Inner core backtracking by seismic waveform change reversals, Nature 631, 340 (2024)
• Yang and Song, Multidecadal variation of the Earth's inner-core rotation, Nature Geoscience 16, 182 (2023)
• Pham and Tkalčić, Up-to-fivefold reverberating waves through the Earth's center and distinctly anisotropic innermost inner core, Nature Communications 14, 754 (2023)
• Pearson et al., Hydrous mantle transition zone indicated by ringwoodite included within diamond, Nature 507, 221 (2014)
• Anzellini et al., Melting of iron at Earth's inner core boundary based on fast X-ray diffraction, Science 340, 464 (2013), PubMed abstract
• Yuan et al., Moon-forming impactor as a source of Earth's basal mantle anomalies, Nature 623, 95 (2023), PubMed abstract
• Murakami et al., Post-perovskite phase transition in MgSiO3, Science 304, 855 (2004), PubMed abstract
• Iitaka et al., The elasticity of the MgSiO3 post-perovskite phase in the Earth's lowermost mantle, Nature 430, 442 (2004), PubMed abstract
• Fiquet et al., Melting of peridotite to 140 gigapascals, Science 329, 1516 (2010), PubMed abstract
• Tateno et al., The structure of iron in Earth's inner core, Science 330, 359 (2010), PubMed abstract
• Ohta et al., Experimental determination of the electrical resistivity of iron at Earth's core conditions, Nature 534, 95 (2016), PubMed abstract
• Konôpková et al., Direct measurement of thermal conductivity in solid iron at planetary core conditions, Nature 534, 99 (2016), PubMed abstract
• Prokofiev et al., Exceptional concentrations of gold nanoparticles in 1.7 Ga fluid inclusions from the Kola Superdeep Borehole, Scientific Reports 10, 1108 (2020), PubMed abstract
• Ritterbex and Tsuchiya, Viscosity of hcp iron at Earth's inner core conditions from density functional theory, Scientific Reports 10, 6311 (2020), PubMed abstract
• Gan et al., Molten iron at extreme conditions reveals compositional inhomogeneity in Earth's lower outer core, PNAS 123, e2533391123 (2026), PubMed abstract

Last reviewed: 23 October 2026.