Most of the subsurface is covered by soil, water, vegetation or younger rock. Drilling everywhere would cost a fortune. Geophysics gets round this by measuring physical fields at or above the surface. A body that is denser, more magnetic or more conductive than its surroundings, or that carries sound at a different speed, disturbs those fields slightly, and a sensitive instrument can pick up the disturbance.

What matters is the contrast between a body and its host rock, not the absolute value. A dense ore body inside equally dense basalt is invisible to gravity. The same body inside light sediments stands out clearly.

Each method senses one physical property, so the first question is always: what makes my target different? The second is where the signal comes from. Passive methods measure fields that nature already provides, like taking a photo in daylight. Active methods supply their own signal, like using a torch: more control, more cost.

Five methods do most of the work. Each card jumps to its section; hover over one to see it in action. The table after them compares these five with the more specialised tools.

Gravity is the simplest of them. Everything pulls on everything else, and denser rock pulls slightly harder. A gravimeter is essentially an extraordinarily sensitive spring balance: carry it over a buried dense body and the spring stretches a tiny bit more, by about one part in a million over a large ore body. Gravity is also the clearest example of how readings become a picture of the subsurface, so it gets a full worked example, with its own simulator, in Inversion & uncertainty.

The same instrument can be flown, carried or lowered down a hole, and each choice trades coverage against detail.

A few minerals, above all magnetite (an iron oxide), are strongly magnetic. Earth's magnetic field magnetises them, so a magnetite-rich rock behaves like a weak bar magnet buried in the ground, adding its own small field to Earth's. A magnetometer, usually towed by a plane, helicopter or drone, measures the total field to a small fraction of a nanotesla. The property that matters is magnetic susceptibility: how easily a rock becomes magnetised.

Basalt usually contains plenty of magnetite. Sand, limestone and most sediments contain very little. That makes magnetics the cheapest, fastest way to map volcanic rocks, dikes, faults and iron ores hidden under soil, and the reason most exploration programmes start with an airborne magnetic survey.

Magnetics has two quirks that gravity doesn't. First, Earth's field is tilted. It is vertical at the magnetic poles (pointing down in the north, up in the south), lies flat at the magnetic equator and dips at an angle in between, and the buried magnet is tilted the same way. So the anomaly is usually a lopsided high-and-low pair, not a single bump over the body. Second, some rocks carry remanent magnetism, a "fossil" magnetisation frozen in when they cooled, which can point in a quite different direction. It is the same effect that records magnetic reversals in the seafloor (see The deep Earth).

The third big property is how well rock conducts electricity. Most rock-forming minerals are insulators, so current flows mainly through salty water in pores and cracks, through clay, and through the few minerals that conduct well, such as sulphides and graphite. Geophysicists usually quote resistivity, in ohm-metres (Ωm), the inverse of conductivity: low resistivity means a good conductor. Dry granite can be tens of thousands of Ωm; massive sulphide can be well below 1 Ωm. That enormous range is what makes electrical methods so useful.

There are three ways to get current into the ground: let nature induce it (MT), induce it yourself with a changing magnetic field (controlled-source EM), or push it in directly through electrodes (resistivity and IP).

Earth is constantly bathed in natural electromagnetic signals. Lightning, striking dozens of times every second somewhere on the planet, sends out high-frequency signals that travel around the globe. Gusts of the solar wind, the stream of charged particles from the Sun, shake Earth's magnetic field and make slow, low-frequency ripples. Both induce weak electric currents in the ground, called telluric currents. A magnetotelluric (MT) station sits in one spot for hours to days, recording the electric field (with electrodes buried in the soil) and the magnetic field (with sensitive coils). The ratio between the two tells you how well the ground conducts.

The key idea is the skin depth. Signals fade as they soak into conductive ground, and how fast depends on their frequency. It's like music through a wall: the bass gets through and the treble doesn't. Low frequencies reach deep, high frequencies die out near the surface, and conductive rock absorbs signal faster than resistive rock.

Recording a spread of frequencies is like taking readings at a spread of depths, without drilling a hole. The result is plotted as an apparent resistivity curve: the resistivity the ground seems to have, averaged down to the depth each frequency reaches. MT reaches deeper than any other electrical method, from about a hundred metres to hundreds of kilometres. It is the tool for fluid-rich zones in the crust, geothermal reservoirs and the deep plumbing of large mineral systems. Its weaknesses are blurry resolution, which gets worse with depth, and electrical noise from power lines and railways.

MT waits for nature, and nature's signal can be weak or drowned out by noise. Controlled-source electromagnetics brings its own transmitter: stronger, steadier and under your control, though it can't reach as deep. It comes in many forms. One of the most common, on the ground or slung under a helicopter, uses a big loop of wire. Drive a strong current round it, then switch it off abruptly. Because it works in the time after switch-off, practitioners call this time-domain EM (TEM or TDEM); the term CSEM is mostly kept for the marine and grounded-wire versions described below.

Switching off makes the loop's magnetic field collapse, and a collapsing field induces eddy currents in the ground. In ordinary rock these spread downwards and outwards and fade within a few thousandths of a second, like a smoke ring drifting away. A good conductor, such as massive sulphide ore, graphite or very salty water, holds on to its own eddy currents much longer. A receiver coil listens to the dying signal, and a slow decay is the giveaway. Many buried nickel, copper and zinc deposits have been found by listening for exactly that.

At sea, controlled-source EM, here properly called marine CSEM, is used the other way round, to find resistive layers. A ship tows a transmitter just above the seabed, and receivers sit on the seafloor at a range of distances. Sediment soaked in salty water is a good conductor and swallows the signal quickly. A layer whose pores hold oil, gas or injected CO₂ instead of salty water is resistive, so the signal travels along it with less loss and leaks back up to distant receivers. That extra signal at the far receivers is the clue.

The most direct electrical method skips induction altogether. Two metal stakes, the current electrodes, push a current through the ground, and two more, the potential electrodes, measure the voltage it produces. Ohm's law turns current and voltage into an apparent resistivity. Spreading the electrodes further apart sends the current deeper, so a line of readings at several spacings builds up a cross-section. It is a workhorse for groundwater, clay and the thickness of weathered rock.

Induced polarisation (IP) adds one more measurement. When the current is switched off, the voltage in most rock collapses at once. Where tiny grains of sulphide are scattered through the rock, each grain has stored a little charge at its surface, like a miniature battery, and the voltage dies away slowly instead. That property is the rock's chargeability. It matters because scattered grains barely change the rock's overall conductivity, so EM can miss them, while IP sees them clearly. That makes IP the classic tool for porphyry copper deposits, where sulphide is disseminated through a huge volume of rock.

Shout at a cliff and the echo's delay tells you how far away it is. Seismic surveys do the same with sound waves in rock. A source sends a pulse into the ground: a sledgehammer on a steel plate, a truck that vibrates the ground (a vibroseis), a small explosive charge, or an airgun at sea. Lines of geophones, small sensors that feel the ground move, record what comes back.

An echo happens wherever acoustic impedance changes. That is sound speed multiplied by density. The bigger the change, the stronger the echo, and if impedance drops instead of rising, the echo comes back upside down:

Seismic measures time, not depth. Each echo's two-way time (down and back up) only becomes a depth once you know how fast sound travels: depth = speed × time ÷ 2. Get the speed wrong and every layer shifts up or down. Try it below with the "speed you assume" slider. That trade-off between speed and depth is why seismic is usually tied to drill holes where the depths are known.

Seismic gives by far the sharpest pictures of layered rock, resolving layers metres to tens of metres thick, and it is the backbone of oil, gas and CO₂-storage work. It is also the most expensive. Earthquakes are nature's seismic source, and they are how we know the structure of the whole planet (see The deep Earth).

A cheaper variant, seismic refraction, ignores the echoes and times the first sound to arrive at each geophone. Near the source that is the direct wave, travelling through the slow top layer. But sound that dips down to a fast layer, such as fresh basalt under soil and weathered rock, can race along the top of it as a head wave and leak back up. Beyond a certain distance, the crossover distance, the detour pays off and the head wave arrives first. Plotting first-arrival time against distance gives two straight lines whose slopes give the two speeds (each slope is 1 ÷ speed), and where they meet tells you how deep the fast layer is. It is a standard way to measure depth to bedrock or the thickness of regolith, the blanket of soil and weathered rock above fresh bedrock.

The method follows the target. Ask what makes it physically different from its surroundings, how deep it is, how big, and how much ground you need to cover. Real programmes layer the methods: cheap airborne magnetics and EM over a whole district, then ground gravity, IP, MT or seismic over the best targets, then drilling to test the interpretation.