THE 10 CORE DISCIPLINES OF MINING.

Mineral deposits are rare. Ordinary crustal rock contains gold at roughly four parts per billion; an economic gold deposit concentrates that metal by a factor of hundreds or thousands. That concentration is the product of specific geological processes — magmatic fluids rising through fractures, hydrothermal circulation along shear zones, weathering that strips away waste and leaves resistant metals behind, or sedimentary systems that sort heavy minerals into layers.

In Tanzania, the Archaean greenstone belts of the Lake Victoria goldfields are the classic example. Ancient volcanic and sedimentary sequences were deformed, intruded and then plumbed by gold-bearing fluids that deposited metal in quartz veins and altered wall rock. Elsewhere in the country, the Karagwe-Ankole belt hosts pegmatite-related tin, tantalum and lithium, while the Mozambique belt supplies gemstones and graphite.

The practical consequence is that geology dictates everything downstream. It controls where you drill, how deep the ore continues, how hard the rock is to break, whether the metal will respond to gravity, flotation or leaching, and how much waste you must move for every tonne of ore. A mine plan written without a defensible geological model is a schedule of expensive guesses.

Good geological practice is therefore cumulative rather than heroic. Mapping, sampling, geophysics and drilling each reduce uncertainty in a different way, and the model is revised every time new evidence arrives. The best exploration teams are not the ones with the strongest opinions — they are the ones who change their minds fastest when the rock tells them to.

Exploration begins at regional scale. Satellite imagery, airborne magnetics and radiometrics, historic records and artisanal workings are used to identify belts and structures with the right ingredients. From tens of thousands of square kilometres, a team narrows to a few hundred through desk study alone.

Field programmes then take over: geological mapping, soil and stream sediment geochemistry, rock chip sampling, trenching and ground geophysics such as induced polarisation or electromagnetics. Each survey answers a narrow question. Geochemistry asks whether metal is present near surface. Geophysics asks whether the structures or sulphide bodies that host it continue at depth.

Only then does drilling begin, because drilling is the most expensive way to learn anything. Reverse-circulation drilling gives cheap, fast coverage; diamond core gives structural and metallurgical detail. Assay results are logged, quality-controlled with standards and blanks, and fed into a three-dimensional model that is tested against the next hole.

Modern programmes increasingly run drones ahead of the geologists. High-resolution orthomosaics and photogrammetric terrain models give centimetre-scale base maps for planning access, locating collars, and documenting outcrop before it is disturbed. Machine learning applied to combined geophysical and geochemical layers ranks targets by similarity to known deposits, which changes the order in which holes are drilled and shortens the path to discovery.

The discipline of exploration is knowing when to stop. Most projects should be abandoned, and abandoning them quickly is what funds the one that works.

A mineral resource is an estimate of metal in the ground with reasonable prospects for eventual economic extraction. It is reported as Measured, Indicated or Inferred depending on the density and quality of the data behind it. An ore reserve is the portion of a Measured or Indicated resource that a study has shown can be mined and processed profitably after applying mining dilution, recovery, costs, prices and permitting reality.

Estimation itself is a statistical exercise. Assay intervals are composited to consistent lengths, extreme values are treated so a single spectacular sample cannot inflate a block, and the spatial continuity of grade is measured with a variogram. Grade is then interpolated into a block model using kriging or inverse-distance methods, constrained by geological wireframes so that grade never migrates across a fault or lithological boundary where it does not belong.

Classification follows confidence, not optimism. Reconciliation against production is the honest test: if the mill consistently receives less grade than the model promised, the model, the dilution assumptions or the grade control practice is wrong.

Reporting must follow a recognised code — JORC, NI 43-101, SAMREC or PERC — signed by a competent person. These codes exist because investors were once routinely misled. They impose transparency, materiality and competence on every published number, and any project seeking finance should be built to satisfy them from the first drill hole.

The first question is open pit or underground. That is answered by depth, geometry, grade distribution and rock strength, expressed through the stripping ratio — the tonnes of waste that must be moved for each tonne of ore. Shallow, broad deposits favour open pits. Narrow, deep, high-grade bodies favour underground methods such as sublevel stoping, cut-and-fill or, at scale, block caving.

Pit design starts with an economic shell: the surface that maximises undiscounted value at a given price. That shell is then engineered into something buildable — benches sized for the loading fleet, ramps at gradients the trucks can climb, wall angles supported by geotechnical analysis, and haul routes that minimise cycle time.

Scheduling turns the design into years. Cut-off grade strategy, ore blending, waste dump and stockpile placement, fleet size and pre-strip requirements are optimised together, because the sequence that produces the most metal is rarely the sequence that produces the most value. Bringing high-grade tonnes forward improves cash flow but can shorten mine life and complicate blending.

Plans are living documents. Survey pickups, blast performance, grade control and equipment availability feed weekly and monthly reconciliations. Drone survey has become the standard method for this: routine flights measure stockpile volumes, dump capacity, bench advance and highwall condition in hours rather than days, which means the plan is corrected against reality while the correction still matters.

A blast is designed as a pattern: hole diameter, burden, spacing, sub-drill, stemming, explosive type and initiation timing. Powder factor — kilograms of explosive per tonne of rock — is the headline number, but timing is what controls the outcome. Millisecond delays between holes and rows determine how rock moves, how well it fragments, how much vibration reaches surrounding structures and how stable the resulting wall is.

Fragmentation matters economically because crushers and mills consume energy roughly in proportion to the size reduction they must perform. Coarse, poorly fragmented muck slows loading, damages tyres and buckets, and lowers mill throughput. Over-blasting, on the other hand, wastes explosive, throws material where it is not wanted, generates fines that hurt recovery in some circuits, and damages the pit wall.

Control comes from measurement. Hole deviation surveys, high-speed video, vibration monitoring and photographic fragmentation analysis reveal whether the design was executed as drawn. Drone photogrammetry before and after firing quantifies the blasted volume and the actual face geometry, which closes the loop between design intent and delivered rock.

Safety and community obligations are inseparable from blast design. Exclusion zones, magazine control, licensed shot-firers, flyrock prediction and vibration limits at the nearest receptor are engineering constraints, not paperwork.

The load-and-haul system is a queueing problem. Excavators or loaders fill trucks, trucks travel a cycle to a tip, and every mismatch between loader capacity and truck capacity, or between truck fleet size and route length, shows up as idle equipment. Matching passes per load, sizing the fleet for the longest planned haul, and maintaining roads so trucks can run at design speed are the fundamentals.

Cost is dominated by fuel, tyres, labour and maintenance. Haul road rolling resistance is the lever operators most often ignore: a poorly maintained surface can add double-digit percentages to fuel burn and cycle time across an entire fleet. Ramp gradient, corner radius, drainage and dust suppression are therefore production issues.

Dispatch and telemetry systems assign trucks dynamically, track payload, and record delays by reason code. When that data is combined with regular drone survey of the pit and dumps, an operator can reconcile what was planned, what was moved, and what remains — by bench, by material type, by week.

Ore loss and dilution happen here. Mining outside the designed boundary sends waste to the mill; mining inside it leaves paid-for metal in the ground. Good grade control marking, disciplined operators and accurate as-mined survey are what keep the mill feed close to the model.

Comminution comes first. Crushing and grinding reduce rock until the valuable mineral grains are physically liberated from the gangue. This is energy-intensive — often the largest power consumer on site — so grind size is always a trade-off between recovery and cost.

Separation methods follow the mineralogy. Gravity concentration exploits density and works well for coarse free gold and heavy mineral sands. Flotation exploits surface chemistry, using collectors and frothers to attach sulphide particles to air bubbles; it is the workhorse for copper, lead, zinc and refractory gold concentrates. Magnetic and electrostatic separation handle iron and some industrial minerals.

Hydrometallurgy dissolves the metal. Cyanide leaching with carbon-in-pulp or carbon-in-leach recovery remains standard for gold, while heap leaching handles low-grade oxide ores at large scale and copper is often recovered by solvent extraction and electrowinning. Refractory ores — where gold is locked inside sulphides — need oxidative pre-treatment such as roasting, pressure oxidation or bio-oxidation before leaching will work.

Metallurgical test work is what makes these choices defensible. Comminution indices, mineralogical study, bottle-roll and column leach tests, flotation locked-cycle tests and settling behaviour of tailings should be established on representative samples long before a flowsheet is committed to steel.

Every processing plant also has a tailings problem, and that problem is best solved in design rather than retrofitted after the fact.

The principal hazards are consistent across the industry: ground failure and rockfall, mobile equipment interaction, working at height, confined spaces, uncontrolled energy, dust, noise, vibration, heat, and exposure to chemical reagents such as cyanide, acids and explosives.

Control follows a hierarchy. Eliminate the hazard where possible, then substitute, then engineer barriers, then apply administrative controls, and only then rely on personal protective equipment. An operation that leans heavily on PPE and behavioural campaigns has usually skipped the engineering steps.

Systems make control repeatable: risk assessments, safe work procedures, permit-to-work, isolation and lock-out, pre-start inspections, competency-based training, emergency response and mine rescue capability. Leading indicators — hazard reports, near-miss investigations, critical control verifications — tell you more about future performance than lagging injury rates do.

Remote sensing has become a genuine safety tool rather than a novelty. Drone inspection of highwalls, tailings embankments, headframes, conveyors and plant structures removes people from unstable ground and from working at height, while producing a dated visual record that supports engineering decisions.

Occupational health deserves the same rigour as safety. Silica dust, diesel particulate, mercury exposure in artisanal gold processing, noise-induced hearing loss and heat stress cause harm slowly, which is exactly why they are neglected.

Environmental management starts with baseline data. Water quality and flow, groundwater levels, air quality, noise, soils, flora, fauna and existing land use must be characterised before disturbance, because without a baseline no later claim about impact can be substantiated.

The most serious technical risks are tailings storage and water. Tailings facilities must be engineered for the design storm and the design earthquake, with controlled water balance, instrumented embankments, independent review and a documented emergency plan. Recent global failures were failures of governance and monitoring as much as of geotechnics. Acid rock drainage is the second long-lived risk: sulphide waste exposed to air and water can generate acidity and mobilise metals for decades, so waste characterisation and encapsulation belong in the mine plan, not the closure plan.

Social performance is equally structural. Free, prior and informed consultation, transparent grievance mechanisms, fair resettlement and compensation, local employment and procurement, and honest reporting of payments to government determine whether a project is tolerated or supported. In Tanzania this is reinforced by local content and corporate social responsibility obligations under the Mining Act and its regulations.

Closure is designed at the start. Progressive rehabilitation, landform design that is stable without maintenance, revegetation with appropriate species, and a funded closure provision convert a mine from a liability into a handed-back landscape. Drone-based monitoring makes rehabilitation measurable — vegetation cover, erosion, subsidence and water bodies can be tracked year on year over the whole footprint.

A mining project is evaluated as a long-dated cash flow. Capital expenditure is spent up front, revenue arrives over years or decades, and the discount rate applied to that gap dominates the result. Net present value, internal rate of return, payback period and the all-in sustaining cost per unit of metal are the standard measures, and they are only as good as the assumptions behind them.

Costs split into operating and capital. Operating cost per tonne mined and per tonne milled, driven by fuel, power, reagents, labour, consumables and maintenance, sets the cut-off grade. Sustaining capital — fleet replacement, tailings lifts, pre-strip, deferred development — is real cost that early studies routinely understate.

Studies escalate deliberately: scoping (order of magnitude), pre-feasibility, then feasibility with accuracy tight enough to support project finance. Skipping stages to save time is the most reliable way to build the wrong mine. Sensitivity and scenario analysis on grade, recovery, price and capital cost show which variables actually control the outcome, and those are the variables worth spending money to define.

Fiscal and governance terms matter as much as geology. Royalties, corporate tax, clearing fees, withholding taxes, local content requirements, state participation and stability of tenure all sit between the metal and the shareholder. Commodity price cycles then test every assumption: operations that survive downturns are the ones with low unit costs, flexible plans, modest debt and clean licences.

The pattern behind all of it is information. Projects fail when decisions are made on hope rather than measured evidence, and they succeed when each stage of work buys enough certainty to justify the next.