How our planet works — from core to crust
We've never directly sampled anything below about 12 km (the Kola Superdeep Borehole). Earth's radius is 6,371 km. So how do we know what's inside? The answer is one of the great detective stories in science — using earthquake waves as probes.
Earthquakes generate two main types of body waves: P-waves (primary, compressional — like pushing a slinky) travel through solids and liquids. S-waves (secondary, shear — like shaking a rope) travel only through solids. Both waves change speed and direction when they cross boundaries between materials of different densities.
In 1906, Richard Oldham noticed that S-waves completely disappear in a "shadow zone" on the far side of the Earth from an earthquake. Since S-waves can't travel through liquid, there must be a liquid layer inside Earth — the outer core. In 1936, Inge Lehmann noticed faint P-wave arrivals inside the S-wave shadow zone that could only be explained by a solid inner core reflecting them. The entire interior structure was mapped this way — from patterns of wave arrivals at seismograph stations around the world.
Try this analogy: imagine you can't see inside a box, but you can tap on one side and listen on the other. Hard material transmits the tap quickly and clearly. Soft material absorbs it. Liquid changes the sound completely. By tapping in different places and listening carefully, you could build a picture of what's inside — without ever opening the box. That's seismology.
Two types: oceanic (5–10 km thick, basalt, density ~3.0 g/cm³) and continental (30–70 km thick, granite, density ~2.7 g/cm³). Continental crust is lighter and more buoyant, which is why it floats higher — forming continents above sea level. The boundary between crust and mantle is the Mohorovičić discontinuity (Moho), identified by a sharp increase in P-wave velocity from ~6.5 to ~8 km/s.
The mantle is ~84% of Earth's volume. It's solid rock (mostly peridotite) but flows over geological time at rates of 1–10 cm/year. The asthenosphere (100–250 km depth) is partially molten and relatively weak — this is the "lubricating layer" on which plates slide. Below it, the lower mantle extends to 2,900 km, getting progressively denser under pressure. At the core-mantle boundary, temperature reaches ~4,000°C. The mantle convects — hot material rises, cools, and sinks — driving plate tectonics.
Liquid iron-nickel alloy at 4,500–5,500°C. Density: ~10–12 g/cm³. We know it's liquid because S-waves can't pass through it. This conducting, convecting fluid generates Earth's magnetic field via the geodynamo. Flow rates are estimated at 10–30 km/year — detectable through changes in the magnetic field over decades. The outer core is cooling at about 100°C per billion years.
Solid iron-nickel, radius ~1,220 km. Temperature: ~5,400°C (comparable to the Sun's surface). Pressure: ~360 GPa. Despite the extreme heat, pressure forces iron into a solid hexagonal close-packed crystal structure. The inner core is growing as the outer core slowly cools and solidifies onto it — at about 1 mm/year. This solidification releases latent heat that helps drive outer core convection.
Earth's interior isn't just layers — it's a heat engine. Radioactive decay and leftover primordial heat drive mantle convection, which drives plate tectonics. The inner core's growth releases heat that drives outer core convection, which generates the magnetic field. The magnetic field protects the atmosphere, which enables liquid water and life. Remove any one link and Earth as we know it changes fundamentally.
Plate tectonics is to geology what evolution is to biology — the framework that makes everything else make sense. It wasn't accepted until the late 1960s, despite Alfred Wegener proposing continental drift in 1912. Wegener had the right idea (continents move) but the wrong mechanism (he thought they plowed through oceanic crust). The discovery of seafloor spreading and mantle convection provided the mechanism.
In the 1960s, surveys of the ocean floor found alternating stripes of normal and reversed magnetic polarity, symmetric about mid-ocean ridges. This proved that new crust forms at ridges and moves outward — seafloor spreading.
Identical fossils of Mesosaurus (a freshwater reptile) appear in both South America and Africa — continents now separated by 5,000 km of ocean. It couldn't have swum across. The continents must have once been joined.
When earthquake locations are plotted globally, they form narrow lines that outline the plate boundaries. This was only visible once a global seismograph network existed in the 1960s.
Since the 1990s, GPS has directly measured plate movements to millimetre precision. The Atlantic is widening at ~2.5 cm/year. India pushes into Asia at ~5 cm/year. These measurements match predictions from plate theory.
Water is critically important for plate tectonics — and this isn't widely appreciated. Water weakens rock, lowering the viscosity of the mantle and allowing plates to move. It lubricates subduction zones, enabling one plate to slide under another. It lowers the melting point of rock, creating the magma that forms volcanic arcs. Earth's plate tectonics may depend on its oceans. Venus, which lost its water, appears to lack plate tectonics entirely — this may not be a coincidence.
If your child asks "will the continents keep moving?" — yes! In about 250 million years, they're predicted to reassemble into a new supercontinent (sometimes called "Pangaea Ultima"). The Atlantic will close, and the Americas will collide with Africa and Eurasia. Show them an animation of continental drift online — seeing 4.5 billion years of drift in 30 seconds makes it visceral.
Earth's magnetic field isn't produced by a permanent magnet — if it were, the Curie temperature (above which iron loses its magnetic properties, ~770°C) would demagnetise it at core temperatures. Instead, it's generated by the geodynamo: a self-sustaining process where convective motion of electrically conducting liquid iron in the outer core generates and maintains the magnetic field.
Three ingredients are needed: (1) a conducting fluid (liquid iron in the outer core), (2) a source of energy to drive convection (heat from the inner core's solidification and radioactive decay), and (3) rotation (the Coriolis effect from Earth's spin organises the flow into helical patterns). The flowing iron generates electric currents, which produce magnetic fields, which in turn influence the flow of iron — a self-reinforcing feedback loop. The process has run continuously for at least 3.5 billion years.
The magnetic field reverses (north and south swap) at irregular intervals — sometimes after 10,000 years, sometimes after tens of millions. The last reversal was ~780,000 years ago (the Brunhes-Matuyama reversal). During a reversal, the field weakens, becomes chaotic with multiple poles, then re-establishes in the opposite direction. The process takes 1,000–10,000 years. The paleomagnetic record (frozen in ocean floor rocks, lava flows, and sediments) provides a continuous history of reversals used for dating and correlation.
Mars had a global magnetic field for its first ~500 million years (we know from magnetised crustal rocks observed by orbiters). When its smaller core cooled and the dynamo stopped, the solar wind began stripping its atmosphere. Over the next few billion years, atmospheric pressure dropped from potentially Earth-like levels to just 0.6% of Earth's. Surface water was lost. The lesson: a planet's magnetic field isn't a nice-to-have; it's essential for long-term surface habitability. Earth's dynamo is powered by ongoing core cooling — and it will eventually stop too, but not for billions of years.
Knowing the ages of rocks is fundamental to geophysics. Without dating, we couldn't know when the continents split, when mass extinctions happened, or how old Earth itself is (4.54 billion years). The primary tool is radiometric dating — using the predictable decay of radioactive isotopes as natural clocks.
Certain atoms (radioactive isotopes) are unstable and transform into different atoms at a known, constant rate. The half-life is the time it takes for half of the atoms to decay. After one half-life, 50% remain. After two, 25%. After three, 12.5%. By measuring the ratio of parent to daughter isotopes in a rock, you can calculate when the rock formed.
Half-life: 5,730 years. Used for organic material up to ~50,000 years old. Formed in the atmosphere by cosmic rays, absorbed by living things, stops being replenished at death. Useless for geological timescales but invaluable for archaeology and recent geology.
Half-life: 1.25 billion years. Used for volcanic rocks. Argon is a gas that escapes when rock melts, so the "clock" resets with each eruption. Measuring trapped argon tells you when the lava last solidified. Works from 100,000 years to billions of years.
Half-life: 4.47 billion years. Used for the oldest rocks and minerals. Zircon crystals are especially useful — they incorporate uranium when they form but reject lead, so any lead present must be from decay. Earth's age (4.54 Ga) was determined using uranium-lead dating of meteorites.
Use M&Ms or coins. Start with 100. Each round, flip them all — remove any that land tails up (they've "decayed"). After round 1, about 50 remain. After round 2, about 25. After round 3, about 12. The number halves each round — that's a half-life. Now imagine each "round" takes a billion years, and instead of M&Ms you're counting atoms. That's radiometric dating.
Understanding Earth's geophysics becomes much deeper when you compare it to other rocky worlds. Each planet is like a natural experiment — same basic ingredients, different outcomes.
Nearly identical size and composition, but no plate tectonics, no magnetic field, and a runaway greenhouse atmosphere (465°C, 92 atm). Venus may periodically "resurface" itself through catastrophic global volcanism rather than steady-state tectonics. The lack of water appears critical — without it, the crust is too stiff to crack into plates. Venus demonstrates that size alone doesn't guarantee habitability.
Too small to retain its heat. Core cooled, dynamo stopped, magnetic field vanished, atmosphere was stripped by solar wind, water was lost. Mars had the right conditions early on (liquid water, thicker atmosphere) but couldn't sustain them. It's a poignant example of how a planet's geophysics determines its fate on billion-year timescales.
Too small for convection, plate tectonics, or a dynamo. Its interior cooled rapidly. No atmosphere, no magnetic field, no geological activity for ~3 billion years. The Moon's surface preserves ancient impact craters from the early Solar System — a record that Earth (with its active geology) has erased.
Europa, Enceladus, and Titan demonstrate that internal heat can come from tidal heating (gravitational flexing) rather than radioactive decay. This expands the habitable zone concept — liquid water oceans can exist far from any star, maintained by tidal forces. Geophysics isn't just about rocky planets.
Earth is habitable because of a specific combination of geophysical properties: (1) large enough to retain internal heat for billions of years, (2) has water to enable plate tectonics, (3) has a liquid outer core generating a protective magnetic field, (4) has plate tectonics acting as a carbon thermostat. Change any one of these and you get Venus, Mars, or the Moon. Understanding geophysics is understanding what makes a planet alive.