Abitibi Greenstone Belt
Mining Region Profile

The Kapuskasing Structural Zone at the belt’s western end exposes lower crust from 20–30 km depth, brought up by a Proterozoic thrust. Percival and West (1994) characterized the section as granulite-facies gneiss, anorthosite, and mafic-ultramafic cumulates. Fluid inclusion and trace element data from these rocks, measured by Laurentian and GSC researchers, show Au-S-CO₂ enrichment above levels documented in deep-crustal exposures beneath other Superior Province subprovinces. Per unit volume of rock undergoing metamorphic devolatilization, the Abitibi lower crust generated fluid with higher Au activity than equivalent source regions elsewhere in the craton. The Yilgarn’s Eastern Goldfields has similar structural complexity and metamorphic grade and has produced less gold per unit area. There is no deep-crustal exposure in the Yilgarn to compare source fertility directly. A mass-balance calculation comparing the volume of Kapuskasing source rock devolatilized (constrained by P-T paths in metamorphic assemblages) against total Au endowment of deposits fed from that source has not appeared in Econ Geol, Mineralium Deposita, or Ore Geology Reviews. The data for a first-order attempt exists in published form. Exploration implications follow from lateral variation in source fertility beneath the belt: MT-derived conductivity profiles could proxy for lower-crustal compositional variation along strike, and segments of the Cadillac-Larder Lake break overlying the most fertile source should carry the largest Au endowments. This has not been tested.

The camp spacing along the Cadillac-Larder Lake break is 25–35 km center-to-center between Kirkland Lake, Larder Lake, Malartic, and Val-d’Or, discussed at PDAC and GAC-MAC workshops without a formal statistically rigorous publication. Two post-2010 discoveries sit at predicted wavelength positions. Untested positions cluster near the Ontario-Québec border where incompatible provincial mapping created a data gap.

Deposits sit at dilational geometric sites along the fault systems. Sigma-Lamaque at Val-d’Or occupies a releasing bend. Kirkland Lake sits at a cross-fault intersection. Canadian Malartic is fracturing around a quartz monzonite buttressed against the break, ~1 g/t disseminated through large tonnage. Robert (1990) and Hodgson (1989) documented these controls. Porcupine assemblage turbidites around Timmins host or adjoin Hollinger-McIntyre, Dome, Borden Gold, with graphite in mudstones reducing Au(HS)₂⁻ and precipitating Au, and EM surveys detecting the graphitic horizons directly because graphite is conductive. Cadillac Group graphitic mudstones coincide with Au at Lapa, Goldex, LaRonde.

Now the syenite-gold association, which is where the belt’s geology gets specific enough to have direct consequences for where exploration money goes.

Canadian Malartic, Young-Davidson at Matachewan, Beattie, Holt-McDermott. All hosted by or adjacent to syenite-monzonite intrusions at dilational sites on the Cadillac-Larder Lake break. Robert (2001) mapped the spatial pattern. The question of whether the intrusions contributed Au from a mantle-metasomatized magmatic source or served as passive structural hosts filled by metamorphic fluids from elsewhere has been discussed in general terms for over twenty years. Helt et al. (2014) provided data that should have ended the discussion, or at least narrowed it sharply, and the fact that the discussion continues in the same general terms as before suggests that the data has not been fully absorbed by the exploration community.

The metamorphic volatile budget is H₂O-CO₂-H₂S, Au carried as Au(HS)₂⁻, and the Te content of mafic volcanic source rocks is too low by roughly an order of magnitude to generate the Te activity needed to stabilize calaverite at depositional P-T conditions. Alkaline magmas from metasomatized lithospheric mantle carry Te at the required concentrations because Te is chalcophile and concentrates in metasomatic sulfide phases in the mantle wedge.

The Te requires a magmatic source. The Au in these deposits could theoretically come entirely from metamorphic fluids while the Te comes from the magmatic system, producing a mixed deposit where the intrusion donates Te but not Au. That model requires two independent fluid sources converging at the same structural site and depositing Au-Te minerals together, which is a more complicated set of conditions than a single magmatic-hydrothermal fluid contributing both metals. The Cordilleran alkalic porphyry deposits at Mt. Milligan and Galore Creek have the same Te-V-Au fingerprint in alkaline hosts along arc-transverse structures, and their magmatic-hydrothermal origin is not debated.

For exploration targeting, this means syenites and monzonites along the Cadillac-Larder Lake break are compositionally important, not just mechanically important. An exploration program looking for the next syenite-gold deposit should be screening for alkaline intrusions with mantle-metasomatic trace element signatures (high Nb, high LREE, high Te in sulfide phases), not just looking for any competent body at a dilational site. A gabbro or a granodiorite at the same structural position is a weaker target because it lacks the magmatic metal contribution. This prediction could be tested by comparing Au-Te-V budgets in alkaline-hosted deposits versus non-alkaline-hosted deposits at equivalent structural positions along the same fault, using data already published across papers by Robert, Helt, Rowins, and others. The comparison has not been assembled as a single systematic study.

Kidd Creek. Over 150 Mt at ~6% Zn, ~1% Cu, five times the next largest Abitibi VMS deposit, operating since 1966, mining below 3,000 m. Hannington et al. (1999) argued for a mantle-recharged sub-volcanic heat source sustaining hydrothermal venting past 100,000 years based on isotopic data and the deposit’s internal Zn-Cu zonation recording multiple venting pulses. VMS deposit size, under this model, is primarily a function of how long the hydrothermal system stays active. A felsic magma chamber cooling as a closed system runs for tens of thousands of years. A system being recharged from the mantle runs for much longer and produces a proportionally larger sulfide mound. The exploration discriminant is stratigraphic: felsic volcanic centers recording repeated eruptive cycles with immobile-element evidence of magma replenishment, distinguishable from single-cycle centers using Zr-Y-Ti-Nb systematics through the volcanic section. Noranda and Matagami volcanic complexes have been characterized at this stratigraphic resolution. The rest of the belt’s felsic volcanic stratigraphy, hundreds of kilometers of it, has not.

The analytical issue attached to invisible gold has not received the attention it deserves from qualified persons signing NI 43-101 resource estimates. Fire assay with standard PbO flux underrecovers Au from arsenopyrite-rich samples when the fusion does not fully decompose the sulfide lattice. Undigested arsenopyrite fragments in the lead button retain their lattice-bound Au, which never reaches the cupel. This underrecovery applies to every arsenopyrite-bearing sample processed under those conditions. It shifts the entire assay population downward, which is a systematic bias, unlike the nugget effect from coarse visible Au, which is random scatter around a stable mean. A systematic negative bias of 10–20% on all assays within arsenopyrite-rich domains in a block model directly underestimates contained metal in those domains, and unlike random variance, it does not wash out with larger sample populations. How many resource estimates in the Abitibi Belt are built on fire assay data from arsenopyrite-rich rock without verification that the fusion conditions were adequate? Probably more than the QPs involved would be comfortable admitting, because verifying fusion adequacy for arsenopyrite decomposition requires running paired assays (standard fire assay vs. fire assay with pre-oxidation or vs. aqua regia digest with AA finish) on representative arsenopyrite-rich samples, and that verification step is not routinely done.

Identifying which other shelved prospects deserve re-assaying with refractory protocols requires cross-referencing historical drill logs against arsenopyrite-bearing lithologies along deformation zone corridors. That is a database exercise before it is a geological one, and government survey databases in Ontario and Québec now have the digital infrastructure to support it.

Pathfinder elements from routine 30–50 element ICP-MS: Sb halos extend 200–500 m beyond economic Au limits, more continuous than Au distribution, useful as directional vectors where coarse Au creates erratic grades. Sb gets dropped from analytical packages to save cost. Te at Kirkland Lake and Val-d’Or marks fluid pathways and now has co-product value for CdTe photovoltaic manufacturing. V in magnetite and biotite around alkaline intrusions is readable by portable XRF for cheap field mapping of potassic alteration footprints.

Heat flow 5–10 mW/m² above shield background. Kidd Creek rock temp 45°C past 3,000 m. LaRonde refrigeration costs heavy below 2,500 m. Battery-electric LHD and automated drills running at several mines. MT surveys show undrilled conductive anomalies at 3–5 km depth along the Cadillac-Larder Lake break. Historical exploration confined mostly to upper 500 m, fluid inclusion data from deep mines showing no weakening of alteration or Au tenor at current mining depths.