The modules on ore deposits, magmatic, hydrothermal and sedimentary deposits explained how nature concentrates metals. Exploration is the practical business of finding those concentrations. It runs as a sequence of steps. Each one costs more per square kilometre than the last, so each step's real job is to decide where the next, more expensive step is worth spending money.

The odds are brutal because a mine needs many things to be true at once: enough metal, at a high enough grade (concentration), shallow enough, in a shape that can be mined, in rock the metal can be extracted from, near roads and power, with permits, finance and community consent. Exploration is the process of finding out which of those fail, as cheaply and as early as possible.

The cheapest data is data someone has already collected. A desktop study gathers published geological maps, old reports, academic papers and records of past mining before anyone goes into the field.

In India the starting point is the Geological Survey of India (GSI). Its online portal, Bhukosh, serves geological maps alongside data from two nationwide programmes: the National Geochemical Mapping programme (systematic stream-sediment and soil sampling) and the National Geophysical Mapping programme (gravity and magnetic surveys), plus newer national airborne surveys. The National Mineral Exploration Policy (2016) aimed to make this baseline data public and to draw private explorers in. In 2023 the government named 30 critical minerals, from lithium to rare earths, and began auctioning blocks for them.

Old workings are another classic clue. Miners in India dug for gold, copper, lead and zinc centuries ago. Modern operations such as the Hutti gold field in Karnataka and the Zawar zinc–lead mines in Rajasthan sit on ground worked in ancient times. Old pits, slag heaps and village names that refer to metals all point to mineralised ground.

Then you go and look. Field mapping means walking traverses (planned lines across the country) and recording every outcrop, where bare rock pokes through soil. At each one the geologist notes the lithology, measures the orientation of layers, faults and veins (the structures from Folds, faults & deformation), and looks for alteration: rock that hot fluids have changed chemically, often bleached, clay-rich or stained.

One of the most valuable things to spot is a gossan: a rusty, spongy cap of iron oxides left where sulphide minerals have weathered at the surface. The sulphur and many metals wash away, but the iron stays behind as red-brown limonite and goethite. A gossan is the weathered top of a sulphide body, and it can sit directly above ore.

Weathering slowly releases metals from a deposit into the soil and streams around it, creating a dispersion halo far larger than the deposit itself. Geochemistry measures this halo. A deposit a few hundred metres across can leave a detectable chemical signal kilometres downstream.

Stream-sediment surveys are the classic reconnaissance tool. A sample of fine sediment from a stream is a natural average of everything upstream of it: its whole catchment, the area that drains to that point. One sample can screen tens of square kilometres. When a sample comes back high, you sample further up each tributary until the signal disappears. The source lies between the last high sample and the first low one.

Once a catchment lights up, the work gets more detailed. Soil grids sample the ground on a regular pattern, say every 50 m along lines 200 m apart, to map the halo directly above the source. Rock-chip samples test individual outcrops and veins.

Explorers rarely measure only the metal they want. Gold occurs at parts per billion and is patchy, so they also measure pathfinder elements: elements that travel with the ore but are more abundant or more evenly spread, so their halo is larger and easier to detect. Arsenic and antimony are classic pathfinders for gold. Molybdenum points to porphyry copper systems (see Hydrothermal deposits).

Every mineral reflects sunlight in its own pattern across the visible and infrared, like a barcode. That pattern is its spectral signature. Clays and micas formed by alteration absorb strongly near 2.2 micrometres because of the bonds in their hydroxyl (OH) groups. Iron oxides such as hematite absorb blue light and have a broad dip near 0.9 µm, which is why they look red. A sensor that measures reflected light in the right wavelength bands can map these minerals from orbit.

Remote sensing only sees the top few micrometres of the surface. Soil, vegetation and monsoon cloud all get in the way, which is a real limitation in much of India. To see deeper you need geophysics, covered in Geophysical methods.

Every surface clue is indirect. Only a drill hole brings back rock from depth, so drilling is where most targets die. It is also the largest cost in exploration. There are two main methods.

Core is laid out in trays and logged: a geologist records lithology, alteration, veins, minerals and structures metre by metre, and photographs every tray. The core is then sawn in half. One half is sent for assay and the other is kept as a permanent record.

Assay numbers are only as good as the process behind them, so every batch carries hidden checks called QA/QC (quality assurance and quality control). These are commonly inserted at around one in every 20 samples:

With enough holes, geologists build a 3D model of the ore body and estimate its tonnes and grade between the holes. How much that estimate can be trusted depends on how closely spaced the holes are and how continuous the geology is. Reporting codes sort estimates by confidence. Australia's JORC Code, Canada's CIM Definition Standards (required under its NI 43-101 rules) and others are aligned with a common template set by an international committee, CRIRSCO. (India's own system is based on the UN Framework Classification.)

Why so much formality? Because exploration companies raise money from investors who cannot check the rocks themselves. The cautionary tale is Bre-X. In the mid-1990s this small Canadian company claimed an enormous gold discovery at Busang in Indonesia, and its share value rose to several billion dollars. In 1997 independent check drilling found almost no gold. The crushed core samples had been "salted" with gold from elsewhere. The fraud destroyed billions in savings, and it helped lead to Canada's NI 43-101 rules in 2001 and tighter codes worldwide.

Today a public report of resources or reserves must be signed off by a Competent Person (a "Qualified Person" in Canada). This is a member of a recognised professional body with relevant experience in that type of deposit, who is personally accountable for the numbers. Chain of custody, QA/QC and transparent methods are what make those numbers defensible.