Volcanic Metallogeny
How volcanism builds the world's ore deposits. Most of the metal we mine was concentrated by magmatic and hydrothermal processes tied to volcanism — and the tectonic setting of a volcano predicts which kind of deposit can form near it. This page explains those volcano→metal links in detail, and maps every Holocene volcano on Earth tagged by the deposit style its setting favours. It is geological context for exploration — not a hazard or eruption-risk product.
🌋 The volcano → metal connection
Each deposit family forms in a specific magmatic plumbing system. The metals, the depth, the host and the tell-tale alteration are all set by how and where the magma rises.
Porphyry Cu-Mo-Au
The planet's main copper source. Large, hydrous arc magmas stall as a pluton; as they cool, magmatic-hydrothermal fluids exsolve and deposit metals in dense stockwork veinlets over the cupola. Huge tonnage, low grade, concentric alteration (potassic → phyllic → argillic → propylitic).
Continental & oceanic subduction arcs.
Epithermal Au-Ag (HS / LS)
Shallow (<1.5 km) hot-spring systems above arc volcanism. High-sulfidation = acidic magmatic fluids → vuggy silica + alunite. Low-sulfidation = neutral, meteoric-dominated → adularia-sericite + banded quartz veins. Bonanza gold grades possible.
Subaerial felsic–intermediate arc volcanoes & calderas.
VMS — volcanogenic massive sulfide
Sulfides precipitated on or just below the seafloor at submarine volcanic centres — fossil "black smokers". Kuroko-type form in felsic island-arc/back-arc settings; Cyprus-type in ophiolitic (oceanic-crust) settings.
Submarine arcs, back-arcs & spreading ridges.
Skarn
Where an arc intrusion bakes and replaces carbonate wall-rock, building calc-silicate skarns charged with metal. Often flanks porphyry systems.
Arc intrusions cutting limestone/dolomite.
Alkaline complex / carbonatite
Deep, CO₂-rich alkaline magmas rising along continental rifts crystallise carbonatites and alkaline complexes — the world's premier source of rare-earth elements and niobium.
Continental rift / intraplate magmatism.
Greenstone-hosted & komatiite (ancient)
Archean–Proterozoic greenstone belts — ancient submarine volcanic piles — host the great orogenic gold camps (gold remobilised into shear-hosted quartz veins) and komatiite-hosted nickel. Not Holocene, so not on the live map — but it is the setting of much of GeaSpirit's second-chance gold (e.g. Marvel Loch), and where buried impact structures like Ora Banda can further focus metal.
Archean greenstone belts (Yilgarn, Abitibi, Barberton).
🗺️ World volcano map — tagged by deposit style
Every Holocene volcano (Smithsonian GVP), coloured by the ore-deposit style its tectonic setting favours. Click a volcano for its setting, rock type and metallogenic potential. Click a legend chip to filter.
Volcano data: Smithsonian Institution, Global Volcanism Program (GVP) — Holocene Volcano List (open data). Deposit-style tags are GeaSpirit's interpretation of tectonic setting and are indicative geological context, not a resource or discovery claim. Metallogenic models summarised from standard economic-geology literature.
📖 How to read this geological information
Every volcano, deposit and occurrence record shows the original scientific values and a plain-language interpretation. Scientific values are never changed; where a field is missing it reads “Not reported”.
Worked example — Cabezo María lamproite volcano▾
Scientific view
- Name / type
- Cabezo María — lamproite volcano, Almería, SE Spain
- Coordinates
- 37.2119, −1.9325
- Epoch
- Miocene
- Age
- ~8.6 ± 0.3 to ~7.9 Ma
- Stage (source label)
- Tortonian–Messinian
Plain-language view
- Name / type
- An ancient lamproite volcano (a rare, potassium-rich volcanic rock).
- Coordinates
- Latitude 37.2119°N, longitude 1.9325°W (negative longitude = west of Greenwich).
- Age
- Roughly 8.6 to 7.9 million years old. The “±0.3” is scientific uncertainty: 8.6 ± 0.3 Ma means most likely between about 8.3 and 8.9 million years.
- Stage
- Within the Miocene Epoch. The numeric dates shown (8.6–7.9 Ma) fall mainly in the Tortonian Stage under the current international timescale; the source's “Tortonian–Messinian” label describes a broader regional interval.
Field-by-field explainer▾
Why geological age matters▾
Age helps relate volcanic, tectonic, hydrothermal and mineral-forming events to one another — whether an eruption, an intrusion and an ore-forming fluid pulse could plausibly belong to the same system.
Age alone does not prove an economic deposit. Whether useful mineralisation exists depends on the wider context: tectonic setting, host rock, structure, hydrothermal alteration, the commodity in question, grade, tonnage, depth, and modern verification.
Understanding uncertainty (±)▾
Geological ages are estimates, not exact dates. Uncertainty comes from analytical precision, sample quality, alteration, the dating method, calibration and interpretation.
8.6 ± 0.3 Ma means a best estimate of 8.6 Ma with a margin of about ±0.3 — most likely between roughly 8.3 and 8.9 Ma. It is one age with a margin of error, not two separate ages.
Glossary — geological & metallogenic terms▾
🌍 Plate tectonics — the deep engine of metallogeny
Volcanism is the surface symptom; plate tectonics is the cause. The type of plate boundary sets the heat, fluids and structures available — and therefore which ore deposits can form. Below: how each boundary type builds (or fails to build) mineral systems, with real examples and citable sources. The 241 present-day boundaries are a live overlay on the main asset map — toggle 🌐 Tectonic Plates.
Convergent — subduction zones
An oceanic plate dives beneath another plate. Water driven off the slab fluxes the mantle wedge, generating hydrous, oxidised arc magmas — the single most productive metallogenic setting on Earth. These magmas feed porphyry Cu-Au-Mo systems, shallow epithermal Au-Ag, skarns and arc VMS.
Slab dehydration → flux melting → hydrous arc magmatism → magmatic-hydrothermal fluid release.
Convergent — continental collision (orogens)
When two continents collide the crust thickens, heats and melts. Metamorphic dewatering drives orogenic gold into shear-hosted quartz veins; crustal anatectic granites concentrate tin and tungsten; pegmatites carry Li-Ta.
Crustal thickening → regional metamorphism → metamorphic/anatectic fluids → vein & granite-hosted deposits.
Divergent — mid-ocean ridges & back-arcs
Plates pull apart; seawater convects through hot young crust and precipitates seafloor massive sulphides / VMS at black smokers. Where that oceanic crust is later obducted onto land (ophiolites), it also hosts podiform chromite.
Decompression melting → seawater hydrothermal convection → sulphide precipitation on/under the seafloor.
Divergent — continental rifts
A continent begins to split. Deep, CO₂-rich alkaline magmas rise along the rift and crystallise carbonatites and alkaline complexes — the world's premier source of rare earths and niobium. Rift basins and brines also concentrate lithium and sediment-hosted copper.
Lithospheric extension → low-degree alkaline/carbonatite melting → REE-Nb enrichment; evaporitic/brine Li.
Transform boundaries
Plates slide laterally past one another with little magmatism, so transforms rarely create deposits directly. Their value is structural: deep transform and strike-slip faults act as long-lived plumbing that can focus fluids from adjacent systems and localise pull-apart basins.
Lateral shear → crustal-scale fault conduits → fluid focusing & pull-apart basins (rather than melt generation).
Intraplate & ancient cratons
Away from boundaries, two engines dominate. Mantle plumes / large igneous provinces deliver magmatic Ni-Cu-PGE sulphides; and thick, cold, ancient cratonic keels are where diamond-bearing kimberlites survive their fast ascent from the deep mantle.
Plume magmatism (LIP Ni-Cu-PGE) · deep kimberlite ascent through stable cratonic lithosphere (diamond).
Boundary type → deposit type, at a glance
| Boundary / setting | Signature deposits | Key metals |
|---|---|---|
| Subduction arc | Porphyry · epithermal · skarn · arc VMS | Cu Au Mo Ag |
| Continental collision | Orogenic gold · Sn-W granites · Li pegmatites | Au Sn W Li |
| Mid-ocean ridge / back-arc | VMS / seafloor massive sulphide · ophiolite chromite | Cu Zn Au Cr |
| Continental rift | Carbonatite REE-Nb · Li brines · sed-hosted Cu | REE Nb Li Cu |
| Transform | No primary class — structural fluid pathways | — |
| Intraplate / craton | LIP Ni-Cu-PGE · kimberlite diamond · IOCG | Ni Cu PGE Diamond |
♻️ The Wilson cycle — settings are recycled
Plate settings are not permanent — they cycle. An ocean basin opens at a rift, widens at a ridge, then closes by subduction until two continents collide; later the suture can rift apart again. This Wilson cycle means a single piece of crust can pass through several ore-forming settings over hundreds of millions of years. It is also why ophiolites — slices of old ocean floor, carrying their VMS and chromite — end up stranded high in mountain belts.
📌 Why deposits cluster — metallogenic belts
Deposits are not scattered at random: they cluster in metallogenic belts and provinces that trace a former boundary. The Andean porphyry belt follows an active subduction margin; the Iberian Pyrite Belt traces a Variscan collision; the Central African Copperbelt follows a Proterozoic rift basin. Find the right boundary at the right age and you have narrowed the search from a continent to a belt.
🔋 Critical metals — where each one forms
Lithium — continental rifts, salar brines and collision-zone pegmatites. Rare earths & niobium — carbonatites along old rifts. Copper — subduction arcs. Nickel-cobalt-PGE — mantle-plume large igneous provinces. Tin-tungsten — collisional granite belts. Every critical-metal search starts by asking which boundary, at which age, built it.
🗺️ Interactive plate-boundary map
The 241 PB2002 boundaries, coloured by type. Click any segment for the deposit families that type controls. Click a legend chip to show/hide a type.
☄️ Impacts — craters & meteorites
Confirmed terrestrial impact structures (Wikidata / Earth Impact Database) and recorded meteorite landings (NASA / Meteoritical Society, 51,776 catalogued, 38,246 with a published location). Impacts drive shock metamorphism (coesite, stishovite, shocked quartz, suevite) and locally impact diamonds (Popigai). Click a feature for its fiche; click a chip to toggle a layer. Click a mineral to open it in the Gem & Mineral database.
🌎 Antarctic polar view — the meteorites Mercator cannot show
This view uses Antarctic Polar Stereographic (EPSG:3031), in which the South Pole is an ordinary point on the map. The world map above cannot show these records honestly: Web Mercator stops at 85.0511°, so everything further south is clamped onto that line rather than placed. You can switch between NASA Blue Marble polar imagery, served natively in this projection by NASA EOSDIS GIBS, and a dark analytical basemap that draws a simplified cartographic outline of Antarctica for contrast — the meteorite positions are identical in both. Zoom in and the markers grow with the view. Where thousands of specimens share one published coordinate the marker stays one marker, because that is one collection area, and clicking it opens the records that belong to it rather than scattering them across positions nobody recorded.
❄️ Why are so many meteorites found in Antarctica?
Meteorites do not fall more frequently in Antarctica than elsewhere on Earth. Antarctica is exceptional because its cold, dry environment preserves them, while the slow movement of the ice can transport and concentrate meteorites in blue-ice areas known as meteorite stranding zones. Wind ablation and local ice-flow conditions expose old ice at the surface, where dark meteorites are comparatively easy to spot against the ice. As a result, Antarctica contains an unusually large share of the meteorites recovered on Earth.
- Preservation — cold, dry conditions can preserve meteorites for very long periods.
- Ice transport — meteorites buried in the ice are carried along as the ice sheet flows.
- Stranding zones — where ice flow slows or is obstructed and surface ice is removed, meteorites can become concentrated and exposed in blue-ice areas.
- Visibility — dark rocks stand out strongly against blue and white ice, and there is little terrestrial rock or vegetation to hide them.
These concentrations therefore reflect preservation, ice transport and search conditions — not a higher extraterrestrial impact rate.
Sources: NASA Earth Observatory; ANSMET (Antarctic Search for Meteorites), Case Western Reserve University.
🛢️ Petroleum provinces — subsurface context
The world's petroleum provinces, as defined and assessed by the USGS. They are here for one reason: a province where petroleum has been assessed is a province whose deep architecture has been studied far more than almost anywhere else on Earth — decades of seismic, wells, formation tops, temperature, pressure and basin modelling. That record describes sedimentary basins, stratigraphy, faults, traps, depth, thermal history and fluid migration, and all of it is useful when reading a completely different mineral system. Click a province for what the assessment recorded.
🛢️ Oil & gas fields — subsurface context
Individual documented fields, from national regulators — not provinces. A petroleum province is a basin-scale assessment domain; an oil or gas field is one accumulation, with a name, an operator and a licence. The province above says where the deep subsurface has been investigated at all; a field says where that investigation is at its most intense, because a producing field is drilled, logged, pressure-tested and modelled for decades. Click a field for what its regulator publishes.
🎯 Mineral prospectivity — where the controls converge
A Weights-of-Evidence model learns, from the current eligible metallic-mineral corpus — 193,674 documented occurrences, deposits and mines — how strongly each geological control co-occurs with a known deposit — then produces a relative prospectivity surface across the modelled domain. Below: the learned strengths, the held-out result, and the most under-explored convergence zones. It is a relative screening signal, not a probability: it ranks ground against other ground, never states a chance that a deposit exists, and only drilling confirms.
⏳ Paleo-tectonics — deposits live on ancient boundaries
Most ore was made at a boundary that no longer exists where the deposit now sits. Iberia's tin-tungsten belt formed in a Variscan collision ~300 million years ago; Bayan Obo's rare earths formed in a Proterozoic rift. To explore by setting you have to reconstruct the plates back through time. Plate-reconstruction models (GPlates / PALEOMAP) let you rewind continents from today to ~540 Ma and ask: what boundary was here when the metal formed? A paleo-reconstruction time-slider is available on the asset map, under Paleo-tectonics: it draws reconstructed coastlines for selected ages. It is interpretive context at a coarse scale, not a reconstruction of where any particular deposit sat when it formed.
🌋 Volcanic provinces — Spain as a worked example
Volcanism matters to metallogeny not because lava is metal, but because a cooling volcanic pile drives the hydrothermal circulation that concentrates it. Spain carries six volcanic provinces spanning almost the whole range — one active, two with Holocene volcanism, and three that stopped long enough ago to be eroded down to the level where ore actually sits. Read them as stages of one process, not as a list of scenic cones.
“Dormant” is a word to use carefully. Popular summaries put all six Spanish provinces under it, and that flattens about fifteen million years. La Palma erupted in 2021; Campo de Calatrava last erupted in the mid-Holocene, under 5,500 years ago; Cabo de Gata has been extinct since the Miocene. Three different geological situations, and only the first carries present volcanic hazard. Each province below states its evidenced status instead of the convenient single label.
Canary Islands — ACTIVE. Spain's only province with present volcanic hazard; intraplate, on oceanic crust. La Palma erupted in 2021, El Hierro submarine in 2012, Lanzarote in 1824 — its 1730–36 fissure eruption was the archipelago's largest historical event — and Tenerife in 1909, where Teide sits inside the 10 × 17 km Las Cañadas caldera. Metallogenic relevance: alkaline and phonolitic magmatism with active hydrothermal systems; the economic story here is geothermal energy and hazard management, not mining.
La Garrotxa / Olot, Girona — HOLOCENE, no historical eruption. About 500 km² of Strombolian scoria cones and alkali-basalt flows, part of the NE Volcanic Province continuing into the Empordà and La Selva. Croscat is among the youngest cones of mainland Iberia. Best preserved, not largest — Campo de Calatrava covers roughly ten times the area. Metallogenic relevance: low. Young, uneroded, and with no exposed hydrothermal root.
Campo de Calatrava, Ciudad Real — HOLOCENE. The largest field in mainland Spain: on the order of 350 volcanoes and 250 maars across about 5,000 km², basaltic to foiditic, in a continental-rift setting. Its last eruption was mid-Holocene at Columba volcano, under 5,500 years ago, and CO₂-rich springs (hervideros) still degas today. Metallogenic relevance: the best mainland example of a fluid system caught while still running — phreatomagmatic craters plus magmatic CO₂ are the plumbing that mineralises, seen before erosion removes it.
Cabo de Gata-Níjar, Almería — EXTINCT, Miocene (~15–7.5 Ma). Calc-alkaline calderas, domes and pyroclastic sequences, deeply eroded. This is the province that matters most here, and erosion is the reason: the edifice has been stripped back to the level where its hydrothermal system sat. Rodalquilar is a high-sulphidation epithermal gold system in a ~11 Ma caldera — on the order of a tonne of gold through the Denver cyanidation plant (1956–66), closed on grade, not on geology. Widespread alunite–kaolinite alteration maps the old fluid pathways.
Columbretes Islands, Castellón — QUATERNARY, inactive. A small alkaline archipelago in the Gulf of Valencia, largely submarine in origin, emergent as Illa Grossa, Illa Ferrera and Illa Foradada, today a marine reserve. Metallogenic relevance: none demonstrated. It is listed for completeness of the provincial picture, not because evidence links it to mineralisation.
Cofrentes / Cerro de Agrás, Valencia — PLEISTOCENE, extinct. A basaltic cone and small field, mapped as part of the Pliocene–Pleistocene Gulf of Valencia province. No present hazard. Metallogenic relevance: none demonstrated at deposit scale; its value is as a control point on the same intraplate province as Columbretes.
The pattern across the six is the thing to carry away. Young provinces show you the machine still running and tell you little about ore; old, eroded provinces show you what the machine left behind. Cabo de Gata is mineralised and La Garrotxa is not — and the difference is not how much magma each erupted. It is fifteen million years of erosion cutting one of them down to the level where the gold was deposited.
Sources: Smithsonian Global Volcanism Program (province summaries, eruption dates, rock types, settings); Ancochea & Huertas 2021, Journal of Iberian Geology 47:209 (Campo de Calatrava); López-Ruiz et al. 2002 (Gulf of Valencia province); Fernández-Soler volcanic stratigraphy, and Braga & Martín (Cabo de Gata); GeaSpirit's own Rodalquilar sample report. One widely quoted figure was deliberately not used: that Spain contains about 200 volcanic structures. Campo de Calatrava alone accounts for roughly 350 volcanoes and 250 maars, so that number cannot be a count of structures, and its basis could not be established from an authoritative source.
🔩 Structural controls — the plumbing of mineral systems
The section above asked four questions; this one answers the second. Structure makes no metal. What it decides is where fluid could go, where permeability opened, where pressure fell, and therefore where the metal that was already moving had a reason to stop. That makes structure the dominant control on pathway and a frequent control on trap — and no control at all on source. It is also the part of geology most often misread from a map, because a line on a map is easy to see and the four things that make it matter are not.
Faults. A fault is not one thing through time. The same surface can be a conduit while it is moving, a seal once its gouge compacts, and a conduit again when it reactivates a hundred million years later. Normal faults open under extension and are the classic feeders in rift and basin systems; thrusts stack and can both carry and cut off a system; strike-slip faults create dilation at bends and jogs. What decides whether a fault mattered: did it exist at the time, was it permeable during mineralisation, did it connect a source to a trap, did the right fluid move through it, and did later deformation leave the result intact. What it does not prove: most large faults are barren. Proximity to one is not evidence — the four conditions are.
Shear zones. Where rock deforms by flowing rather than breaking, strain localises into ductile shear zones; nearer the surface the same corridor behaves brittle-ductile and fractures as well as flows. These are the great fluid highways of the crust: they are long, they are deep, they stay active across whole orogenies, and they can pump enormous volumes. Orogenic gold is the system built on them — regional metamorphism dehydrates the pile, fluid rises along a crustal-scale shear, and gold drops where a second-order structure opens a dilational site and the chemistry changes. What it does not prove: shear zone is not gold. The world has vastly more shear corridors than gold deposits, and the difference is timing, chemistry and a trap.
Structural intersections. The most consistently useful structural target is not a big line but a crossing: fault against fault, fault against a fold hinge, fault against a lithological contact, fault against an intrusion margin, fault against an unconformity, fault against a permeable unit. Intersections concentrate dilation, drop pressure locally, fracture rock more intensely than either structure alone, and bring two different fluids or two different chemistries into contact. That is four trapping mechanisms available in one place, which is why intersections are habitually ranked high in targeting. What it does not prove: geometric complexity is not mineralisation. Crossings are common; deposits are not.
Breccias. Rock broken into angular clasts, and the reason it matters is not the drama but the arithmetic: fragmentation multiplies permeability and reactive surface area at the same time. Tectonic breccia is ground up by faulting; hydrothermal breccia is shattered by fluid overpressure; magmatic-hydrothermal breccia is driven by an exsolving intrusion; collapse breccia forms where support is dissolved away. Porphyry, epithermal, IOCG, skarn and carbonate-replacement systems all commonly sit in or beside breccia bodies. What it does not prove: not every breccia is hydrothermal, and not every hydrothermal breccia is mineralised. A fault can grind rock for a hundred kilometres without a gram of metal passing through it.
Fluid corridors. A deposit usually occupies a very small part of something much larger — a fault system, a shear belt, a fracture network, a permeable stratigraphic unit, a basin-margin structure, an intrusion's fracture halo. Exploration works down that hierarchy: regional fluid corridor → favourable structure → local trap → ore shoot. Each step narrows by orders of magnitude, and each step is a different question. Confusing them is the commonest scale error in targeting: a corridor is a reason to look at a region, never a reason to drill a point.
Folds. Folding localises strain, and strain makes fracture permeability. Hinges are where extension concentrates on the outer arc, where competent beds crack while incompetent ones flow, and where an already-favourable bed can be brought repeatedly against a structure. Fold-hinge sites matter in orogenic gold, in some sediment-hosted systems and in carbonate-replacement systems. What it does not prove: folds are everywhere in deformed terrain. A hinge is only interesting when a mineral system was passing through it at the time.
Unconformities. A surface where the record stops and resumes — and in mineral-system terms, a permeability contrast, a chemical contrast and a reactivation surface stacked on one plane. Basin fluids move along them, redox can change sharply across them, and they are preferentially reactivated by later faulting. Unconformity-related uranium is the system named after this control: oxidised basin brine moves along the unconformity, meets reduced basement or graphitic structure, and uranium precipitates at the contact. What it does not prove: unconformities are among the most common surfaces in geology.
Timing — the dimension a map cannot show
A structurally perfect fault is worthless if it formed after the metal moved. This is the most common way structural targeting goes wrong, and a map will never tell you, because a map shows geometry and not chronology. The questions are: when did the structure form, when was it active, when did the mineralising fluid move, and when did the alteration form? The evidence is relative before it is absolute — cross-cutting relationships, vein generations, alteration overprinting one another, offsets that displace the ore or fail to. Radiometric dating settles it where it exists. This is where structure rejoins the previous section: pathway is not a place, it is a place at a time.
What can actually be observed
Field and map data — mapped faults, contacts, fold axes, veins, breccias, alteration. Remote sensing — lineaments, alteration mineral patterns, geomorphic expression. Elevation models — scarps, linear valleys, offset drainage, structural ridges. Geophysics — magnetic gradients, gravity boundaries, conductivity contrasts, seismic structure where it exists. Historical mining — the alignment of old workings, vein orientations and ore-shoot trends, which is often the sharpest structural data available for a district because it was mapped by people following the metal. And the chain of warnings that goes with all of it: a lineament is not a fault, a fault is not a fluid pathway, and a fluid pathway is not ore.
How this works inside GeaSpirit. The temptation is a fault score: measure the distance to the nearest mapped structure and award points as it shrinks. That would be worse than useless, because mapped-fault density tracks survey coverage — well-mapped countries would score high for having been looked at. Structural evidence is used contextually instead, as an argument about pathway and trap that has to survive four further questions: is there a source it could have connected to, is there a trap at the other end, does the timing work, and what argues against it. Mapped structure, tectonic boundaries, elevation-derived lineaments, geological contacts, the alignment of historical workings, alteration and geophysics all feed that argument. None of them is a point. The Evidence Compass is built to weigh them, and to keep the counter-evidence visible while it does.
Five systems, seen as plumbing
Orogenic Au · regional shear corridor → second-order fault splay → dilational jog → quartz-carbonate vein → gold at the pressure drop
Epithermal Au-Ag · volcanic-hydrothermal system → fault and fracture network → boiling at a pressure drop → banded vein mineralisation
MVT Pb-Zn · basin brine → basin fault into permeable carbonate → chemical trap where sulphur is available → sulphide replacement of the host
Porphyry Cu · intrusion → stockwork fracture network generated by the intrusion itself → hydrothermal fluid through its own plumbing → alteration zoning and sulphide deposition
Sediment-hosted Cu · basin-scale oxidised fluid → permeable unit or fault → reduced stratigraphic trap → copper precipitated at the redox contact
What structural evidence does not prove. A large fault is not a deposit. A fault intersection is not a deposit. A lineament is not a fault — it is a line someone drew on an image, and it may be a road, a dyke, a shadow or a fence. A breccia is not necessarily hydrothermal. A shear zone is not gold. A vein is not economic ore. Dense fracturing is not mineralisation. In many systems the right structural setting is necessary; it is almost never sufficient, and treating a necessary condition as evidence of the result is the single most productive source of wasted exploration money in the industry.
Sources: Sibson on fault-valve behaviour and fluid redistribution in fault zones; Cox on structural controls and permeability in hydrothermal systems; Groves et al. and Goldfarb et al. on orogenic gold and crustal-scale shear systems; Sillitoe on porphyry copper systems and their fracture networks; Simmons, White & John on epithermal vein systems; Leach et al. on structural controls in sediment-hosted lead-zinc; Jefferson et al. on unconformity-related uranium; Geoscience Australia's mineral-systems framework and the USGS mineral deposit models series for the deposit-family context. Full entries are listed in Sources & further reading.
🧭 The mineral-system view — source → pathway → trap → preservation
Every section above this one names a kind of deposit: porphyry, MVT, SEDEX, orogenic gold, Ni laterite, lithium brine. Those labels are useful, and they are not the question exploration actually turns on. The question is whether a complete system existed at that place: metal had to come from somewhere, travel somehow, stop for a reason, and survive to the present. Break any one of the four and there is no deposit — no matter how favourable the other three look. This is the frame the rest of Metallogeny hangs from.
SOURCE — where did the metal come from? Nothing concentrates a metal that was never there. A source can be the mantle, an arc magma, the crust, metamorphic dehydration, a sedimentary basin's brines, seawater, groundwater, or the weathering of a fertile parent rock. Porphyry copper draws on arc magma; orogenic gold on metamorphic fluids; nickel laterite on the ultramafic rock directly beneath it; lithium brine on weathered volcanic and crustal input carried in by water. Evidence GeaSpirit can read: lithology, geochemistry, intrusive and ultramafic rocks, basin architecture, the metallogenic province, what the district historically produced. What it does not prove: a fertile source is a necessary condition and nothing more. Most fertile rock hosts no deposit at all.
PATHWAY — how did it move? Metal sitting in its source is inert. Something has to carry it, and that requires connected permeability at the moment it mattered: faults, fractures, shear zones, porous sandstone, carbonate aquifers, magma conduits, a weathering profile, or a basin's circulating brine. Evidence GeaSpirit can read: mapped faults and shear corridors, lithological contacts, permeable stratigraphy, structural intersections, alteration halos that trace where fluid actually went. What it does not prove — and this is the most misread line on the page: a large fault near a mine is not evidence of a deposit. A fault only counts if it was open during mineralisation, connected the source to the trap, and carried the right fluid. Timing and connection decide it, not proximity.
TRAP — why did it stop? This is the step that turns a metal-bearing fluid into an ore body, and it is usually a change the fluid could not survive: cooling, boiling, a pressure drop, mixing with another fluid, a shift in pH, a redox boundary, sulphur becoming available, a reactive carbonate, adsorption onto clay, evaporation, or simply a permeability contrast that forces the fluid to slow. Roll-front uranium precipitates at an oxidation-reduction front in sandstone. MVT lead-zinc drops out where brine meets carbonate and sulphur. Epithermal gold precipitates where the fluid boils. What it does not prove: a trapping mechanism explains why metal accumulated — it says nothing about how much, or at what grade.
PRESERVATION — did it survive? A system can form perfectly and leave nothing you can find. Erosion can strip it away, burial can put it out of reach, metamorphism can redistribute it, oxidation can disperse it, faulting can cut it in half, and younger sediment or lava can hide it completely. Erosion is double-edged: too little and the ore is still buried; too much and it is gone. Cabo de Gata is the useful comparison in this page — fifteen million years of erosion cut it down to the level where its hydrothermal system sat, which is why Rodalquilar's gold is exposed, while La Garrotxa is young and uneroded and shows nothing at all. Two errors to avoid, in both directions: no known deposit does not prove the system never formed, and favourable preserved geology does not prove an economic one exists.
Why this frame governs how GeaSpirit reasons. The weak version of this platform would ask four shallow questions — is there a volcano nearby, a fault nearby, the right rock, a mine nearby — and add up the yeses. That produces a number that rises with data density rather than with geology. The mineral-system view asks instead: source? pathway? trap? preservation? and requires independent evidence for each. Those four are reasoning dimensions, not four scores to sum. A target with three strong dimensions and one missing is not 75% of a deposit; it is a system with a hole in it, and the hole is the finding. Counter-evidence and missing evidence stay in the record precisely because this frame makes their absence conspicuous. Everything the platform already carries feeds one of the four — geology and historical production into source, structure and alteration into pathway, lithological and chemical contrasts into trap, topography and erosional level into preservation — and the Evidence Compass exists to combine them into one reasoned geological conclusion, not four independent points.
Five systems, read through the same frame
Porphyry Cu · source fertile arc magma · pathway the intrusion itself plus its fracture network · trap cooling and pressure drop driving hydrothermal reaction · preservation the erosional level decides whether the system is exposed, buried or already gone
Orogenic Au · source metamorphic dehydration fluids · pathway a crustal-scale shear corridor · trap pressure drop and dilation at a structural jog, plus chemical reaction with the host · preservation later uplift and erosion bring it into reach
MVT Pb-Zn · source metal-bearing basin brine · pathway permeable formations and basin faults · trap reaction with carbonate where sulphur is available · preservation basin evolution and erosional level
Ni laterite · source the ultramafic parent rock directly below · pathway the weathering profile and downward-percolating water · trap residual and secondary concentration as magnesium and silica leave · preservation a stable tropical regolith that was neither stripped nor buried
Lithium brine · source Li-bearing volcanic and crustal rock · pathway groundwater draining into a closed basin · trap evaporation under hydrological closure · preservation a long-lived arid basin that never flushed
What none of this proves. Source is not a deposit. A fault is not a pathway unless its timing and permeability fit. A pathway is not mineralisation. Alteration is not ore. A trapping mechanism is not grade. Preserved favourable geology is not a resource. And several favourable indicators together are still not a discovery — stacking weak evidence does not make it strong, it only makes it longer. Only sampling, drilling, mineralogy, assay and engineering can establish that an economic occurrence exists. GeaSpirit ranks ground against other ground; it never states that metal is there.
Sources: the mineral-system approach as formalised by Wyborn, Heinrich & Jaques (1994) and developed through Geoscience Australia's mineral-systems framework; McCuaig & Hronsky on mineral-system exploration targeting; Hronsky & Groves on targeting under uncertainty; the USGS mineral deposit models series; Groves et al. on orogenic gold systems; Sillitoe on porphyry copper systems; Leach et al. on sediment-hosted lead-zinc; Ore Geology Reviews, Economic Geology and Mineralium Deposita for the deposit-family literature cited in the sections above. Full entries are listed in Sources & further reading.
🛢⛏ Basin Mineral Systems — where petroleum and mineral systems intersect
An oil field is not a sign of metal. That is the first thing to say and the last thing to forget: the great majority of ore deposits form by processes a petroleum system has nothing to do with, and a well on a map tells you about hydrocarbons, not mineralisation. But a sedimentary basin is a single machine. It buries rock, heats it, squeezes water out of it, and drives that water along faults and permeable beds for tens of millions of years. That machine can make a petroleum system and, in the right chemistry, certain mineral systems as well — and in a few of them the hydrocarbons are not bystanders but participants. This section sets out where the relationship is real, where it is only shared address, and where it should not be expected at all.
Three different relationships, routinely confused. Shared context — a petroleum system and a mineral system occupy the same basin, the same faults and the same aquifers because the same burial history produced both. Neither caused the other; they have a common cause. Direct interaction — hydrocarbons or organic matter take part in the chemistry that precipitates the ore, most often by supplying the reduced sulphur that metal-bearing brine cannot carry with it. The fluid is the resource — the water produced alongside oil and gas carries dissolved elements in recoverable concentration, and no ore body is involved at all. Calling all three “a link between oil and minerals” is how a real mechanism gets buried under a coincidence.
Basinal brines — the transport system both systems use
Bury a sedimentary pile and water is expelled from it. Compaction drives it upward and sideways; heat drives reactions that change what it can dissolve; evaporite beds make it hypersaline, and a hot chloride brine is an excellent solvent for base metals. These brines reach 90 to 200 °C in ordinary basins — hot enough to leach lead and zinc from basement and from the detrital sediment shed off it, cool enough that nothing magmatic is required. The same expulsion and the same faults move hydrocarbons. When the literature says an ore-forming brine is “similar to the formation waters of large sedimentary basins, differing mainly in carrying significant zinc and lead”, it is describing the fluid a petroleum geologist meets every day.
Burial and expulsion. Sediment compacts, pore water is driven out, and the basin begins to circulate. The volume matters: an ore body is a small amount of metal collected from an enormous amount of rock.
Salinity and heat. Evaporites and burial produce a hot, chloride-rich brine. Metals travel as chloride complexes, which is why salinity, not temperature alone, sets what a basin fluid can carry.
Pathway. Faults and permeable formations focus the flow. The same architecture that traps hydrocarbons focuses metal-bearing brine, because it is the same architecture.
Trap. Metal in solution stays in solution until the chemistry changes. In basin-hosted ore systems the change is usually reduction — and that is where organic matter and hydrocarbons enter the story.
Mississippi Valley-type Pb-Zn — the clearest case
MVT deposits sit in carbonate platforms at basin margins and forelands, with no igneous body anywhere near them. Oxidised basinal brine carries lead and zinc as chloride complexes; the metals precipitate as galena and sphalerite the moment they meet reduced sulphur. The brine cannot bring that sulphur with it — a fluid carrying both metal and sulphide would have dropped its load long before arriving — so the sulphur has to be waiting, and the usual way to make it is thermochemical sulfate reduction: sulphate in the brine reduced to H₂S by hydrocarbons, by heating kerogen in the host carbonate to give methane and hydrogen, or by the solid bitumen that is found altered and intergrown with the ores at many deposits. This is not a coincidence of location. It is a step in the reaction.
SOURCE · lead and zinc leached from basement and its erosional detritus across a large volume of basin
PATHWAY · hot saline basinal brine focused along basin faults and permeable carbonate
TRAP · reduced sulphur, commonly generated by thermochemical sulfate reduction mediated by hydrocarbons or kerogen
PRESERVATION · shallow, low-temperature, and rarely overprinted — MVT ores survive well once formed
And in at least one deposit, the petroleum did more than reduce. The standard role for hydrocarbons in this family is chemical: supply the reductant, make the sulphide, let the brine drop its metal. At Laisvall in Sweden, lead isotopes tell a stronger story. Comparing the isotopic composition of the ore sulphides with bitumen extracted from the Alum Shale source rock and with the underlying basement granite, the authors resolved two distinct lead sources and attributed 40 to 60 per cent of the deposit’s lead budget to petroleum-associated fluids rather than to basement leaching. That is the petroleum carrying the metal, not merely precipitating it. It is also, for now, one deposit: the authors present it as opening a way to re-evaluate the role of hydrocarbons in genetic models, not as an established property of sediment-hosted ores. Treat it as a demonstrated possibility with a name and a number attached, and not as a general basin process.
What this does not say. It does not say hydrocarbons are required for MVT ore. Reduced sulphur can also come from bacterial sulfate reduction or from pre-existing sulphide, and sulphur sources vary from district to district. It says that where a basin generated hydrocarbons and also moved metalliferous brine through carbonate, one of the main obstacles to forming an ore body — getting reduced sulphur to the right place — may already have been solved.
Sediment-hosted copper — a redox boundary, drawn by organic matter
Sediment-hosted copper works on the same logic with the redox sense reversed. Copper travels in an oxidised brine, typically one that has passed through red beds, and precipitates where that brine meets reduced ground: organic-rich shale, pyrite, or hydrocarbons occupying a palaeo-aquifer. The ore is a stain on a chemical boundary, and its geometry is the geometry of that boundary rather than of any intrusion. The European Kupferschiefer is the textbook case — a thin organic-rich marine shale above red beds, inside a basin that is also a petroleum province. The controls the deposit models list are basin controls: oxidised source beds, a permeable aquifer, a reductant, and a structure to focus the flow.
Sandstone-hosted uranium — the reduction front made visible
Uranium is soluble when oxidised and almost insoluble when reduced, which makes it the most legible member of this family. Oxygenated groundwater carries uranium through a permeable sandstone until it meets something that will reduce it — carbonaceous plant debris, pyrite, or reducing gas and hydrocarbons migrating up from below — and the uranium drops out along a curved front that then creeps slowly downflow. The ore is not where the uranium came from and not where it is going; it is the surface where the chemistry changed. Where the reductant is petroleum-derived, the petroleum system and the uranium system are physically coupled.
Organic-rich black shales — same environment, two products
Some black shales are enriched in cobalt, chromium, molybdenum, nickel, platinum-group elements, rhenium and vanadium, and the same units are studied as petroleum source rocks and drilled as tight-oil plays. It is tempting to read that as a link between oil and metal. It is not. An anoxic, organic-rich seafloor preserves organic carbon and scavenges certain metals from seawater, because sulphidic bottom water does both at once. The petroleum and the metals are siblings, not parent and child. The practical consequence for exploration is real all the same: the mapping, geochemistry and stratigraphy generated by decades of shale-gas work describe, in detail, rocks that are also the host of a metal-enrichment problem.
Produced waters — when the petroleum system is the mineral resource
Every barrel of oil comes with water, and in some basins that water is a concentrated solution of elements that are mined elsewhere from rock. Lithium is the clearest example: brines in the Smackover Formation of southern Arkansas have been measured above 400 milligrams per litre, and the USGS calculates that between 5.1 and 19.0 million tonnes of lithium are present in them. Present is not recoverable, and the range itself — a factor of nearly four — is the honest measure of how much is still unknown. Bromine, iodine, boron, strontium, rubidium and caesium appear in produced waters elsewhere. This is a different kind of target: there is no ore body, no drilling for the deposit, and the material is already being lifted to surface as a waste stream. It is also a different kind of proof. A concentration is not a resource. Recoverability needs the concentration and the produced volume, the flow rate, the full water chemistry, a working extraction route, a cost that survives the co-product economics, and the rights and environmental position to do anything with it. GeaSpirit treats produced-water chemistry as evidence about a fluid, never as a reserve.
Where no relationship should be expected
A theory that only predicts where things are related is not a theory. The same reasoning that connects basin brines to MVT lead-zinc predicts, just as firmly, that most of the world's metal has nothing to do with petroleum. And if a map ever seemed to show otherwise, the first suspect would not be geology. A petroleum basin has been drilled more, logged more, sampled more and mapped more than almost any other ground on Earth: more wells, more subsurface geochemistry, more seismic, more stratigraphy. Anywhere people have looked harder, they have found more of everything — including mineral occurrences. An apparent overlap between oil fields and mineral records can be a record of exploration effort rather than of geology, and that possibility has to be excluded before it is interpreted.
Porphyry Cu-Au-Mo · arc magma, magmatic-hydrothermal fluid, and a fracture network above a cooling intrusion. No generic petroleum-system relationship is expected: nothing in the model requires basin fluid. A porphyry can sit inside a petroleum basin without the two having anything to do with each other.
Epithermal Au-Ag · shallow magmatic-hydrothermal systems, boiling and mixing. No generic relationship with hydrocarbon generation is expected, and spatial coexistence would carry no genetic weight.
Orogenic Au · metamorphic dehydration fluids in crustal-scale shear zones, forming far hotter and deeper than any petroleum system operates. No generic relationship is expected.
Carbonatite REE-Nb · alkaline and carbonatitic magmatism in rifted lithosphere — mantle-derived, with no sedimentary step. No generic relationship is expected, even where a rift hosts both.
Ni laterite · tropical weathering of ultramafic rock — a surface process, with no basin fluid involved. No generic relationship is expected.
The GeaSpirit rule, stated plainly. Proximity to an oil or gas field carries zero automatic mineral-prospectivity weight, and no score anywhere on this platform is raised because a well is nearby. Petroleum information becomes relevant only when a mineral-system hypothesis that already stands on its own evidence requires something a petroleum dataset can speak to — basin architecture, basinal brines, a fluid pathway, a reductant, or produced-water chemistry. A well near a porphyry target means nothing. A carbonate platform at a basin margin, with evaporites, a fault set, documented brine migration, organic matter and lead-zinc showings, is a different sentence entirely — and the petroleum data is one of several things being read, not the reason for the conclusion.
What this looks like as evidence, not as a score
Supporting · carbonate host at a basin margin; evaporites in the section; a fault set linking basin to platform; documented formation-water salinity; organic matter or bitumen in the host; known Pb-Zn showings.
Counter-evidence · no carbonate host; no plausible migration path; peak burial temperature too low for thermochemical sulfate reduction; no sulphate available; mineralisation dated to a time when the basin was not circulating.
Missing evidence · brine chemistry; fluid-inclusion temperatures and salinities; sulphur isotopes to identify the sulphur source; an age for the mineralisation; burial history.
Next best evidence · sulphur isotopes, because they separate bacterial from thermochemical sulphate reduction — and that single measurement decides whether the hydrocarbon story applies at all.
Timing decides it. Two systems sharing a map do not share a history. Mineralisation, brine migration and hydrocarbon generation have to have overlapped in time before any of this is more than geography, and a deposit that formed two hundred million years before the basin generated oil is simply an older feature under a younger one. Spatial overlap without temporal compatibility is not genetic evidence, and it is the easiest error to make from a map.
Sources: Leach et al. and the USGS deposit model for Mississippi Valley-type lead-zinc ores (USGS SIR 2010-5070-A) for basinal-brine transport, carbonate hosts and sulphur sources; experimental and field work on thermochemical sulfate reduction in MVT systems (Thom & Anderson, Geofluids) for hydrocarbon-mediated generation of reduced sulphur; Migdisov et al., “Petroleum as source and carrier of metals in epigenetic sediment-hosted mineralization” (Scientific Reports, 2019), a lead-isotope study of the Laisvall deposit in Sweden, for the case in which petroleum-associated fluids carried 40–60 per cent of the lead — a result the authors present for that deposit and as an opening for re-evaluating other systems, not as an established general mechanism; the USGS descriptive and assessment models for sediment-hosted copper, including the Kupferschiefer of the Southern Permian Basin, for oxidised brines meeting reduced sediment; USGS work on sandstone-hosted uranium of the Colorado Plateau for redox fronts and reductant sources; the USGS Geology, Geophysics and Geochemistry Science Center programme on critical metals in black shales for Co, Cr, Mo, Ni, PGE, Re and V in organic-rich shales that are also petroleum source rocks; and the USGS Oil and Gas Waters Project and National Produced Waters Geochemical Database for lithium and other commodities in produced water, including the Smackover Formation brines of southern Arkansas. Full entries are listed in Sources & further reading.
❄️ The Arctic — mining, mine waste and oil & gas, region by region
Eleven Arctic and sub-Arctic regions and what GeaSpirit holds in each. Mining assets open in the Asset Map, mine waste in Analyze, and oil and gas fields in their own layer below — never as mineral evidence. Svalbard is listed under Norway, its sovereign under the Svalbard Treaty.
Low data coverage is not low resource potential. A region with few records here may simply be under-documented in the open sources GeaSpirit can lawfully use: the counts describe this platform's data, not the geology. GeaSpirit has ingested the identifiable records from the verified open sources covered by this release; this is not a complete Arctic database.
Counts are computed from the stores for each region's box and countries. A named asset links to its own record; an asset whose position no source publishes is not drawn.
Sources & further reading
- Bird, P. (2003). An updated digital model of plate boundaries. Geochemistry, Geophysics, Geosystems 4(3), 1027. — the PB2002 boundary model used in the map overlay (public domain).
- Sillitoe, R.H. (2010). Porphyry copper systems. Economic Geology 105(1), 3–41.
- Groves, D.I. et al. (1998). Orogenic gold deposits. Ore Geology Reviews 13, 7–27.
- Hronsky, J.M.A. et al. (2012). A unified model for gold mineralisation in accretionary orogens. Mineralium Deposita 47.
- Müller, R.D. et al. (2018). GPlates: building a virtual Earth through deep time. G-Cubed 19. — open plate-reconstruction software/service.
- Scotese, C.R. PALEOMAP Project — paleogeographic reconstructions (educational use).
- Macrostrat · USGS plate-tectonics datasets · NASA SEDAC plate boundaries (open data).
Plate-boundary geometry: PB2002 (Bird 2003) via the public-domain tectonicplates dataset. Boundary→deposit links summarise standard economic-geology literature and are indicative geological context, not a resource or discovery claim.