Is the Earth's Interior Really Made of Runny Magma?
― The mantle is solid rock. But over hundreds of years, it slowly flows
Plenty of people believe there is a sea of red, molten magma under the ground. But the "mantle", which fills more than 80% of the Earth's volume, is solid rock that has not melted. Even so, it is also said to "flow". Solid, yet flowing. The key to this seeming contradiction is how long a time you look over.
You have probably seen a picture of the Earth cut in half in a picture book. The inside is coloured orange and red, and looks hot enough to be molten.
On the news, red lava pours out of a crater and flows downhill. It is natural to think the underground is packed with that stuff.
But if there were a sea of liquid just below the ground, the continents would have to be thin rafts floating on it. Is that really so?
There are two reasons we can say it is "not molten"
Some kinds of shaking cannot travel through liquid. Those waves come through the mantle just fine, so the mantle must be solid.
To a shake lasting a few minutes, rock bounces back like a spring. To a force lasting more than a few hundred years, it flows like syrup. The misunderstanding comes from mixing up these two faces.
So "the mantle is runny" is only half right. It is not melted. It stays solid and changes shape very, very slowly.
Earthquake waves cannot travel through liquid
When an earthquake strikes, its shaking travels through the Earth and reaches distant observatories. There are two kinds of shaking: one that squeezes and stretches, and one that shears sideways.
The sideways kind travels because the material pushes back to restore its shape. A liquid has no such push. If you shear water sideways, it simply flows. That is why the shearing waves cannot travel through liquid.
When we observe them, these shearing waves pass through the mantle without trouble, down to a depth of about 2900 km. They stop only when they reach the "outer core" below. Look at Figure 1. The only large liquid part of the Earth is the molten iron of the outer core.
The mantle's temperature is said to be around 1000℃ near the top and over 3000℃ at the bottom. It still does not melt, because the weight of the rock above presses on it with enormous pressure. The higher the pressure, the harder it is for rock to melt.
Magma forms only where this balance breaks down. One place is where rock rises slowly and the pressure drops. Another is where water seeps out of a sinking ocean plate and makes the rock easier to melt. Even there, only part of the rock melts, and the liquid that seeps out gathers together. It is not a spreading sea of magma.
Solid yet flowing ― the key is "how long you look"
It sounds odd to say that hard rock flows. But there is a familiar example. A glacier on a snowy mountain stays solid ice, yet it creeps down its valley by tens of centimetres to several metres a day.
What matters is that everything has a "switchover time". Under a force shorter than that, it bounces like a spring. If you keep pushing for longer, it flows like syrup. For mantle rock, this time is estimated at a few hundred years.
Try the slider in Figure 2 to change the length of time you look over. To an earthquake that lasts only tens of seconds, the mantle is just hard rock. But over tens of thousands or tens of millions of years, a block of rock pushed on its top face tilts further and further, and never returns.
There is real evidence that the mantle flows. Scandinavia in northern Europe is still rising by up to about 1 cm a year. About 20,000 years ago, a sheet of ice several kilometres thick lay on it. The ground that its weight pressed down is now slowly rebounding, as the mantle below flows in underneath.
From this rebound rate, the mantle's "stickiness" was first estimated, in 1935. The value is said to be a trillion times a trillion times that of water. Plates also move only a few centimetres a year, about as fast as your fingernails grow.
At the University of Queensland in Australia, a lump of "pitch", a relative of tar, has been sitting in a funnel since 1927. It is hard enough to shatter when hit with a hammer, yet drops still fall from the tip of the funnel. So far only nine have fallen. Pitch is said to be more than 100 billion times stickier than water, and the mantle is stickier still by about another trillion times.
Summary
Under our feet is not a sea of molten magma. The mantle beneath the crust is solid rock, and the proof is that shearing earthquake waves pass through it. The only large liquid part is the iron of the outer core, below a depth of about 2900 km. Magma forms only in limited places near the surface, where the pressure drops or water is added. Still, the solid mantle does flow when you look over more than a few hundred years. It is this "slow flow" that moves the plates and lifts the land once the glaciers are gone.
The Earth's interior is not melted.
It simply flows at a speed no human lifetime can see.
You can read how shaking told us about the Earth's inside in Nobody can travel to the Earth's centre, so how do we know what is inside?, and about the heat source that keeps warming the mantle in Why hasn't the Earth's interior cooled down after 4.6 billion years?. For how magma erupts, head to Why do volcanoes erupt?.
- Roll a store-bought slime (or slime made with laundry starch) into a ball and drop it gently on a desk. For an instant you can see it bounce a little, like a ball.
- Leave the same ball on the desk for about 10 minutes. The ball that bounced spreads out flat.
- Finally, pull it with both hands in one go. Pulled slowly, it stretches. Pulled quickly, it snaps apart.
Solid under a short force, liquid under a long one. Slime switches over in a few seconds; for the mantle it is a few hundred years. Wash your hands after playing with slime.
For those who want to know more ― terms, equations and links to textbooksEach part is labelled with its level, from middle-school science to university courses
- Middle schoolCovered in middle-school science
- High schoolCovered in high-school earth science or physics
- High school+Advanced high-school material, or a textbook sidebar
- UniversityUniversity-level specialist courses (geophysics, rheology) not taught in high school
- ResearchTopics researchers are still investigating, not yet taught as settled even at university
Middle schoolTerms: this phenomenon has names
- Mantle: the rocky layer from beneath the crust down to a depth of about 2900 km. It makes up more than 80% of the Earth's volume.
- S waves (shear waves): the part of earthquake shaking that shears sideways. They do not travel through liquid. The squeezing-and-stretching kind are called P waves (compressional waves).
- Viscoelasticity: the property of acting like a spring under a short force and like a liquid under a long one. Slime, glacier ice and mantle rock all have it.
- Post-glacial rebound: the slow rising of land that was pressed down, after the thick ice of an ice age has gone.
Middle schoolHigh schoolChecking with an equation: the time when the mantle switches from "solid" to "liquid"
Divide the stickiness (viscosity) by the stiffness (shear modulus), and you get the time at which the spring face and the liquid face swap. This time is called the Maxwell time. The symbols mean the following. The units are seconds, pascals and pascal-seconds.
| τ | Switchover time (Maxwell time, in seconds) |
| η | How sticky it is (viscosity, in pascal-seconds) |
| G | Stiffness against shearing (shear modulus, in pascals) |
| In symbols | τ ＝ η ÷ G |
| In words | Switchover time ＝ stickiness ÷ stiffness |
| Where it comes from | It comes from a model (the Maxwell model) with a spring (stiffness G) and a liquid damper (stickiness η) connected in series: the rate at which the deformation stored in the spring drains away through the liquid's flow. |
| Mantle viscosity η (estimated from post-glacial rebound) | about 1×10^21 pascal-seconds |
| Shear modulus G of the upper mantle | about 7×10^10 pascals |
| Length of one year | about 31.54 million seconds |
| Speed of mantle convection (about the same as plate speed) | about 5 cm per year |
| The number part of τ (1 ÷ 7) | 1 ÷ 7 ≒ 0.143 |
| The "power of 10" part of τ (21 − 10) | 21 − 10 ＝ 11 |
| Together: 0.143×10^11 s ＝ about 14.3 billion seconds. Convert to years | 14300000000 ÷ 31540000 ≒ 453 |
| What fraction of τ is an earthquake's shaking (about 1 minute ＝ 60 s)? | 14300000000 ÷ 60 ≒ 238000000 |
| Convert the mantle's thickness, 2900 km, to cm | 2900 × 100000 ＝ 290000000 |
| Years to flow across that thickness at 5 cm per year | 290000000 ÷ 5 ＝ 58000000 |
| How many times τ (453 years) is that? | 58000000 ÷ 453 ≒ 128000 |
The switchover time is about 450 years. Earthquake shaking lasts only about one 200-millionth of that, so the mantle carries the waves as a spring. On the other hand, it takes about 58 million years for rock to flow across its own thickness, which is over 100,000 times the switchover time. Over that span, the mantle shows its liquid face completely. The numbers in Figure 2 also use this τ ＝ about 14.3 billion seconds.
High schoolHigh school+The S-wave shadow, and the melting temperature that rises with pressure
High schoolAt observation points more than about 103 degrees away from the epicentre (measured as an angle), S waves do not arrive. This is called the "S-wave shadow", and it is taken as proof that the outer core is liquid. Inside the mantle, by contrast, S waves pass normally. This is covered in "Earth's internal structure" in Fundamental Earth Science.
High school+The temperature at which rock begins to melt (the solidus) rises as pressure rises. The mantle's actual temperature is below this solidus at every depth, so it stays solid. At mid-ocean ridges and hot spots, rock rises, the pressure falls, the rock drops below the solidus and partly melts (decompression melting). At subduction zones, added water lowers the solidus itself.
UniversityThe Maxwell model, mantle convection and isostasy
The property of bouncing under short forces and flowing under long ones is treated as viscoelasticity in rheology. The simplest model is the Maxwell model, a spring and a dashpot connected in series, and its time constant is the Maxwell time above. Real mantle rock flows by "solid-state creep", slow deformation through dislocations in crystals and the diffusion of atoms. When this flow is driven by the buoyancy of rock warmed by heat, it is mantle convection; the Rayleigh number is extremely large, so convection occurs even in a solid. The viscosity value was first estimated by Haskell in 1935, by analysing the rebound of the crust after the ice sheets vanished as a relaxation toward isostasy (the state in which the crust is balanced by buoyancy).
📖 Derivations and further reading: Viscoelasticity (Wikipedia, in Japanese) ／ Mantle convection (Wikipedia, in Japanese)
ResearchWhat is still not clear
- How much does stickiness differ with depth? Many estimates put the lower mantle tens of times stickier than the upper mantle, but the values vary widely and are thought to differ from place to place.
- Is the mantle one single convection system? Debate continues over whether the upper and lower parts are separated at a boundary about 660 km deep, or whether sunken plates fall all the way to the bottom and stir the whole mantle.
- What are the "giant blobs" at the base of the mantle? Large regions where shaking travels slowly have been found under Africa and the Pacific. Whether they are rock of a different composition or partly molten is still unsettled.
So this article, too, is "an explanation of what we know so far". Nobody has seen the Earth's inside directly. It is pieced together little by little from shaking, gravity and rock in laboratories.
Links to textbooks (by level)
| Level | Subject / unit | Where in this article |
|---|---|---|
| Middle school | Science Field 2, "Changes of the Earth's crust" (earthquake shaking, volcanoes) | The two kinds of shaking, and how magma forms |
| High school | Fundamental Earth Science, "Earth's internal structure" | The S-wave shadow and the outer core, the cross-section in Figure 1 |
| High school+ | Earth Science, "Plate motion and magma generation" | The solidus and decompression melting |
| University | Geophysics, rheology | Maxwell time, mantle convection, post-glacial rebound |
| Research | Dynamics of the Earth's interior | Viscosity by depth, the shape of convection, the giant blobs at the base |
| ― | Link to daily life | Slime, glaciers, the rising land of Scandinavia |
- Haskell, N. A. (1935) The Motion of a Viscous Fluid Under a Surface Load. Physics 6, 265.
- University of Queensland, School of Mathematics and Physics, "Pitch Drop experiment"
- Wikipedia (Japanese), "マントル対流" (Mantle convection)
- Wikipedia (Japanese), "粘弾性" (Viscoelasticity)
- National Astronomical Observatory of Japan (国立天文台), ed., 『理科年表』 (Chronological Scientific Tables), Maruzen Publishing (丸善出版) (figures for the Earth's internal structure)
- Turcotte, D. L. & Schubert, G. Geodynamics, Cambridge University Press (Maxwell time and mantle viscosity)
※This article is a popular-science explanation for general readers. The figures given are rough guides to help you understand how things work. The mantle's viscosity and temperature vary with depth, location and study.