Search

Current “visual” integration approaches to map geology from geophysics are challenging, biased toward user assumptions, and time consuming. Applying a supervised machine learning Random Forest method to airborne geophysical data sets is an alternative solution to quickly generate interpretative maps, which can later be validated through targeted and therefore more efficient field campaigns. The approach is tested in the Malartic–Cadillac area of the Pontiac subprovince, Canada. Available airborne geophysical data sets include magnetic, frequency-domain electromagnetic, and radiometric data. Unlike the western part of the area, many studies have been done in the eastern part where the world-class Canadian Malartic gold deposit is located. The eastern part has abundant field observations and well-documented geology that are used as training data for Random Forest learning. A predicted map is built after preprocessing the data, gridding the area into a mesh of nodes and training using known geological knowledge. The results provide new information about the spatial distribution of lithological units, including diabase dikes, felsic-intermediate igneous rocks, mafic–ultramafic igneous rocks, and sedimentary rocks. Comparison between the predicted map and previous geological maps could suggest locations where future field mapping efforts should focus in the Malartic–Cadillac area.

The Canadian Malartic gold deposit was produced by a hydrothermal system that altered the host metasedimentary rocks of the Pontiac Group, metavolcanic rocks of the Piche Group, and associated porphyritic intrusions and mafic dykes. The isotopic footprint in Pontiac metasedimentary rocks is marked by an increase in δ2H values from − 91‰ near mineralization to background values of ca. − 59‰, whereas δ18O values remain relatively constant near 10‰. The isotopic footprint in mafic dykes is marked by an increase in δ2H values from − 84‰ near mineralization to background values of ca. − 73‰, and a decrease in δ18O values from 9.8‰ near mineralization to background values of ca. 8.3‰. Kriging of oxygen and hydrogen isotopic compositions around the Canadian Malartic deposit delineates the isotopic footprint of the deposit. In greywackes, the − 59‰ δ2H isopleth encircles the mineralized domain and is elongated towards the northwest and southeast. In mafic dykes, the − 73‰ δ2H isopleth encircles the mineralized domain as well as a large area to the south and southeast, whereas the 8.3‰ δ18O isopleth encircles the center of the mineralized system and extends over an area towards the southeast. Host rock oxygen and hydrogen isotope compositions therefore define a cryptic alteration footprint up to 2.5 km outside of the Canadian Malartic deposit. The isotopic composition of the metamorphic fluid in equilibrium with the average least altered greywacke at upper greenschist to amphibolite conditions of ca. 550 °C is approximately δ18O = 8.6‰ and δ2H = − 19‰. This is close to the composition of the mineralizing fluid at the Canadian Malartic deposit reported by previous studies. Fluid/rock oxygen and hydrogen isotope exchange modeling indicate that the isotopic footprint formed at relatively low fluid/rock ratios up to 0.5 at temperatures near 350–400 °C.

Magnetic and gravity inversions are used to create 2D or 3D models of the magnetic susceptibility and density, respectively, using potential field data. Unconstrained inversions generate an output based on mathematical constraints imposed by the inversion algorithm. Constrained inversions integrate lithological, structural, and petrophysical information in the inversion process to produce more geologically meaningful results. This study analyses the validity of this assertion in the context of the NSERC-CMIC Mineral Exploration Footprints project. Unconstrained and constrained geophysical inversions were computed for three mining sites: a gold site (Canadian Malartic, Québec), a copper site (Highland Valley, British Columbia), and a uranium site (Millennium – McArthur River, Saskatchewan). After initially computing unconstrained inversions, constrained inversions were developed using physical property measurements, which directly link geophysics to geology, and lithological boundaries extracted from an interpreted geological model. While each derived geological model is consistent with the geophysical data, each site exhibited some magnetic complexity that confounded the inversion. The gold site includes regions with a strong magnetic signature that masks the more weakly magnetic zone, thereby hiding the magnetic signature associated with the ore body. Initial unconstrained inversions for the copper site yielded solutions with invalid depth extent. A consistency between the constrained model and the geological model is reached with iterative changes to the depth extent of the model. At the uranium site, the observed magnetic signal is weak, but the inversion provided some insights that could be interpreted in terms of an already known complexly folded geological model.

Structurally complex zones within orogenic terranes typically correspond to areas where there is interference between multiple fold generations and are known to be favorable pathways for fluid flow because of their higher permeability. In the Canadian Malartic district, gold anomalies have been linked with zones of structural complexity that have been quantified by outcrop bedding orientation measurements and calculation of bedding variance maps. In this work, historical apparent resistivity and induced polarization data in the Canadian Malartic district were reprocessed and combined with new surveys to create a compilation of inverted chargeability and resistivity, which were then interpreted together with airborne electromagnetics and outcrop structural data. The results indicate chargeability anomalies, up to five times the background value, associated with the sulfide mineral content in monzodioritic dikes that are thickened in folds and hydrothermally altered. Although the airborne apparent half-space resistivity is mostly sensitive to conductive surficial cover, the inverted ground resistivity method is sensitive to deeper structure and likely represents bedrock signal at depths greater than 25 m. Inverted ground resistivity exhibits strong anisotropy in areas of subvertical bedding, where measured resistivities can vary by up to a factor of two, over the same location, depending on whether the survey lines are perpendicular or parallel to the strike of bedding. This result is observed at scales of 50 cm up to 100 m. Analysis of inverted ground resistivity together with bedding variance indicates a strong correlation between structurally complex zones with high bedding variance and a decrease in resistivity at depths greater than 25 m. This suggests that in places where the presence of disseminated gold cannot be directly detected, or where the outcrop exposure is limited due to overburden cover, geophysical data may still succeed in identifying structural complexity zones that could potentially host mineralization.

Applications of the spectral induced polarization (SIP) method to mineral exploration are limited by our knowledge of the relationships among rock texture, mineral composition, and electrical properties. Laboratory SIP responses were measured on rock samples from the Canadian Malartic gold deposit. Field SIP responses were also measured at the outcrop scale, along a profile that intersects a well-studied mineralized zone. The mineralogy and the texture of sedimentary rocks from this deposit were quantitatively determined with mineral liberation analysis. A systematic decrease (Pearson [Formula: see text]) in total chargeability with increasing fraction of the sulfide mineral interfaces associated with feldspar minerals (namely, K-feldspar and albite) was observed. On the other hand, total chargeability increased with the fraction of sulfide mineral interfaces associated with carbonates and micas (Pearson [Formula: see text]). At Canadian Malartic, proximal alteration in the mineralized zones is marked by rocks that lack a foliation plane and that were subjected to pervasive K-feldspar, albite, and pyrite alteration. In contrast, distal alteration in sedimentary rocks is marked by biotite, albite, carbonate, and pyrite that are oriented along the regional [Formula: see text] foliation. In the least-altered (LA) sedimentary rocks, quartz and biotite are associated with pyrrhotite and ilmenite as the main sulfide and oxide mineral phases, respectively. SIP measurements conducted at district and outcrop scales and along a drill core indicated that proximally altered sedimentary rocks were characterized by low total chargeability values ([Formula: see text] to [Formula: see text] in the laboratory and [Formula: see text] in the field). In contrast, the LA sedimentary rocks were characterized by total chargeability values up to [Formula: see text] in the laboratory and [Formula: see text] in the field. We conclude that mineralized zones associated with this type of ore deposit are characterized by low chargeability anomalies.

The Pontiac Subprovince is located in the Superior Province, south of the Abitibi Subprovince. The metasedimentary rocks of the Pontiac Group are characterized by a Barrovian metamorphic gradient increasing from north to south from biotite- through garnet- to staurolite-zone conditions. The Pontiac Subprovince has been interpreted as an accretionary wedge or alternatively as an exotic terrane that was tectonically docked to the Abitibi Subprovince during Archean subduction. However, few studies have attempted to characterize its pressure-temperature-time path to test this hypothesis and little is known about the timing relationship between regional metamorphism and the gold mineralization. An E-W lepidoblastic ductile micaceous foliation defines the regional fabric related to the second episode of deformation. Garnet and staurolite porphyroblasts are interpreted to be late- to post-kinematic to the regional fabric. Locally, cordierite or andalusite porphyroblasts fully pseudomorphed by a polymineralic assemblage of muscovite, feldspar and quartz are wrapped by the main foliation indicating that a potential low pressure - high temperature metamorphism (M1) preceded the main episode of deformation and Barrovian metamorphism (M2). Pressure and temperature (P-T) forward thermodynamic modelling on a water-saturated pseudosection yielded peak conditions of 550–600 °C and 5–6 kbar during a prograde, clockwise P-T path for M2. Lu-Hf dating of garnet from three different locations within the staurolite zone yielded a weighted average age of 2657.5 ± 4.4 Ma (95% confidence level) that is inferred to be representative of these conditions. The age of garnet growth calls into question the previous tectonic interpretation of the Abitibi and Pontiac Subprovinces and is coherent with the new rafted ribbon continent model of Bedard (2018) . Because the Canadian Malartic gold mineralization is older or synchronous with the garnet growth episode in the staurolite zone south of the ore body, this garnet age also provides an estimate of the minimum age of the mineralization event, previously dated at 2664 ± 11 Ma with Re-Os on molybdenite.

International audience; The world-class, oxidized intrusion-related Canadian Malartic gold deposit, with reserves estimated at 5.56 Moz Au grading 1.10 g/t Au, and a total geological endowment of 16.3 Moz Au, is one of the largest gold deposits in the Archean Superior Province of Canada. The gold mineralization is hosted predominantly by Pontiac Group metasedimentary rocks, Piché Group metavolcanic rocks, and quartz monzodiorite to granodiorite porphyritic intrusions. The ore takes the form of a low-grade envelope of disseminated pyrite (0.35 to 1 g/t Au) grading inwards into higher grade (>1 g/t Au) stockwork and breccia zones. Hydrothermal alteration in the metase-dimentary rocks is zonally distributed around the fluid pathways. Proximal alteration is characterized by a microcline±albite-quartz replacement-type assemblage, with lesser phlogopite, calcite±Fe-dolomite, pyrite and rutile. The distal alteration assemblage comprises biotite, microcline±albite, phengite, quartz, calcite, pyrite and rutile. In this study, we assess the magnitude and distribution of fluid-rock interaction in the metasedimentary rocks of the Malartic district. The metaturbidites are separated into four lithotypes based on grain-size to reduce the effects of primary depositional processes on mass change calculations. Despite the variability in protolith composition, the metasedimentary rocks define a geochemically consistent, cogenetic sequence. The results of the mass transfer calculations indicate progressive gains in CO 2 , S, K 2 O and LOI, as well as Au, Te, W, Ag, As, Be, Sb, Bi, Mo and Pb, from background, to distal and proximal alteration zones (relative to the least-altered samples). Molar element ratio analysis (alkali/aluminum) indicates an increase in alkali metasomatism (K and Na) adjacent to the main hydrothermal fluid pathways, which is manifested by the progressive stabilization of microcline and albite at the expense of oligoclase, biotite and white mica. Ore-associated pathfinder elements delineate broad enrichment patterns around the deposit, and are used to understand hydrothermal fluid circulation in the Malartic district. A statistical approach based on a comparison of the mass change results with the background composition provides robust constraints on the magnitude and extent of the lithogeochemical haloes. Generally, the alteration forms envelopes that extend along the S 2 fabric, with the largest lithogeochemical anomalies (e.g., Au, W, Te and Ag) reaching up to 10 km in length, and 2 km in width. The results of this study demonstrate that whole-rock lithogeochemistry can provide a valuable tool with which to define vectors toward gold mineralization in a regional exploration context.

The Malartic gold camp is located in the southern part of the Archean Superior Province and straddles the Larder Lake-Cadillac fault zone that is between the Abitibi and Pontiac subprovinces. It comprises the world-class Canadian Malartic deposit (25.91 Moz, including past production, reserves, and resources), and smaller gold deposits located along faults and shear zones in volcanic and metasedimentary rocks of the Abitibi subprovince. North of the Larder Lake-Cadillac fault zone, the Malartic camp includes 2714 to 2697 Ma volcanic rocks and ≤2687 Ma turbiditic sedimentary rocks overlain by ≤2679 to 2669 Ma polymictic conglomerate and sandstone of the Timiskaming Group. South of the fault, the Pontiac subprovince comprises ≤2685 Ma turbiditic graywacke and mudstone, and minor ultramafic to mafic volcanic rocks and iron formations of the Pontiac Group. These supracrustal rocks were metamorphosed at peak greenschist to lower amphibolite facies conditions at ~2660 to 2658 Ma, during D2 compressive deformation, and are cut by a variety of postvolcanic intrusions ranging from ~2695 to 2640 Ma. The Canadian Malartic deposit encompasses several past underground operations and is currently mined as a low-grade, open-pit operation that accounts for about 80% of the past production and reserves in the camp. It dominantly consists of disseminated-stockwork replacement-style mineralization in greenschist facies sedimentary rocks of the Pontiac Group. The mineralized zones are spatially associated with the Sladen fault and ~2678 Ma subalkaline to alkaline porphyritic quartz monzodiorite and granodiorite. Field relationships and isotopic age data for ore-related vein minerals indicate that gold mineralization in the Canadian Malartic deposit occurred at ~2665 to 2660 Ma and was contemporaneous with syn- to late-D2 peak metamorphism. The smaller deposits in the camp include auriferous disseminated-stockwork zones of the Camflo deposit (1.9 Moz) and quartz ± carbonate-pyrite veins and breccias (0.6 Moz) along faults in chemically and mechanically favorable rocks. The age of these deposits is poorly constrained, but ~2692 Ma postmineral dikes, and ~2625 Ma hydrothermal titanite and rutile from the Camflo deposit highlight a long and complex hydrothermal history. Crosscutting relationships and regional geochronological constraints suggest that an early episode of pre-Timiskaming mineralization occurred at >2692 Ma, shortly after the end of volcanism in the Malartic camp, and postmetamorphic fluid circulation may have contributed to concentration or remobilization of gold until ~2625 Ma. However, the bulk of the gold was concentrated in the Canadian Malartic deposit during the main phase of compressive deformation and peak regional metamorphism.

Most currently constructed 3D geological models are to a first order the result of transformations of data with different dimensionality into 3D: • 0D (e.g. outcrop data, at the regional scale), • 1D (e.g. drill hole data, at the mine scale), • 2D data (e.g. satellite data, at the regional scale) or • 3D data (e.g. seismic data, when high resolution 3D geophysical data are available, such as in basins), • 4D models (3D evolutions with time). The datasets used to project between dimensions vary according to the scenario, however they generally consist of a mixture of 0D observations and local temporal or spatial relationships (their topology). Modern software systems are able to use a sub-set of these relationships (fault-stratigraphy relationships for example) to build 3D geological models, however the relationships are not typically used as an independent constraint on how much of the 3D model is constrained by observations, and how much is generated by the end user (or the algorithms they use). This study explores the relationships between topological observations in 1, 2 and 3D in order to better understand how these may be used in the future as inputs to a revised 3D modelling workflow. We have investigated both synthetic cases, where we have full control, and natural examples, which permit alternate hypotheses. This approach has potential relevance to mine-scale and regional 3D models where the 3D topologies are poorly defined by the existing data, but 1D and 2D constraints are available.

Ore-forming hydrothermal fluids react differently with different country rocks, but few studies have applied this knowledge with the specific goal of expanding the size of hydrothermal footprints in mineral exploration. To develop this concept, 122 metamorphosed mafic dykes from the world-class Canadian Malartic gold district (18.6 Moz Au) were sampled and analyzed for mineralogy, physical properties, and lithogeochemistry (partial and total digestion). The mafic dykes intrude mainly metasedimentary rocks, post-date D1 deformation, and cross-cut early-D2 quartz monzodiorite intrusions, but they were deformed and altered during D2 deformation, mineralization, and metamorphism. They can be subdivided into three groups: Group 1 dykes are least-altered, characterized by amphibole-rich regional metamorphic assemblages and distributed throughout the Pontiac Subprovince. Group 2 and 3 dykes underwent ore-related hydrothermal biotite–calcite–pyrite alteration and are associated with density- and volume-adjusted concentration gains of over 100% in Au–W–Te–C–S–Ag–Cs–Mo–Cu–K–Rb–Se–U–Pb–Ba–F–Bi–Sn. They define a metasomatic footprint up to 6 km away from the deposit toward the SE and up to 2 km away from the deposit toward the SW. Fifty-eight variables that define halos around the Canadian Malartic deposit have been identified and integrated using principal component analysis. PC1 explains 30% of the variance, separates least-mobile elements from ore-related elements, and is interpreted to reflect the ore-forming alteration process. PC2 and PC3 represent igneous processes. PC4 highlights the calcite–pyrite and biotite alteration. Spatial variations of the modal abundances of amphibole, biotite, calcite, and pyrite are the simplest expression of the metasomatic footprint, and they can easily be documented during exploration at the camp scale.

The Canadian Malartic deposit is a world class intrusion-related Archean gold deposit hosted in the Pontiac Subprovince, Superior Province, in Quebec, Canada. Laboratory petrophysical properties measurements were performed on 824 rock samples collected from the various rock types observed within the ore body and peripheral host rocks. The various rock types present in the Malartic District, mainly meta-sedimentary rocks, felsic-intermediate intrusive rocks and mafic dykes have contrasting grain densities and magnetic susceptibilities. Using support vector machines, it is shown that these two physical properties can be used to predict the rock type of a sample with an average precision and recall rate of 89%. Within the meta-sedimentary rocks class, variations in magnetic susceptibility are due to the changes in mineralogy associated with hydrothermal alteration. These are caused by the destruction of iron-bearing silicate minerals and magnetite in unaltered rocks (10−4 to 10−3 SI) to form pyrite, carbonates, K-feldspar and Fe-depleted hydrothermal biotite in altered rocks (10−5 to 10−4 SI). Within the felsic-intermediate intrusive rocks, grain densities below 2.7 g/cm3 and magnetic susceptibilities in the 10−6 to 10−4 SI range yield the highest probabilities that a rock has been submitted to carbonate and pyrite alteration. However, magnetic susceptibility and grain density of these intrusive rocks are also dependent on their Fe2O3/Al2O3 and TiO2/Al2O3 ratios, which are not related to the hydrothermal alteration footprint, but are rather due to distinct protolith compositions. Within the mafic dykes, grain density is the best indicator of hydrothermal alteration. Unaltered mafic dykes (2.95 to 3.10 g/cm3) are mostly composed of amphibole, whereas altered mafic dykes (2.70 to 2.95 g/cm3) have reduced amphibole contents and higher abundances of carbonates, pyrite, quartz, and biotite alteration. The support vector machine classifier is extended to predict if meta-sedimentary rocks, felsic-intermediate intrusive rocks and mafic dykes have undergone hydrothermal alteration with average F1 scores of 73%, 69% and 93%, respectively. In altered meta-sedimentary rocks, the integration of grain density and magnetic susceptibility allows the identification of altered but unmineralized samples. The classifier is further extended to predict if the gold content of meta-sedimentary rocks is above or below threshold values of 0.01, 0.1 and 1 ppm with average F1 scores of 83%, 80%, and 76%, respectively. Using conceptual models of the rock physical properties at the Malartic District scale, it is shown that ground magnetic surveys are the most promising geophysical tool for early-stage greenfield exploration of this type of deposit. However, depending on the scale at which the surveys are conducted, magnetic susceptibility contrasts between the various investigated rock types can overshadow the specific signatures of hydrothermally altered rocks. This in part explains why past airborne geophysical exploration campaigns for this type of deposit in the Malartic District were inconclusive. Finally, the machine learning process used in this case study can be applied in advanced exploration stages, during which drilling and either subsequent laboratory petrophysical analyses of core samples or downhole geophysical surveys produce large amounts of data that can be used to train prediction models.

With the increasing size and complexity of geoscientific databases, statistical tools and machine learning algorithms are needed to highlight subtle patterns within dense data clouds and reduce human-related bias in geological mapping. This study presents a new tool that maps structural complexity using circular variance and spherical dispersion in two Archean greenstone belts near Dryden and Timmins in the Superior Province, Ontario. Structural complexity quantitatively highlights areas where structures are folded, juxtaposed, and/or parallel. Bedding measurements and autodetected aeromagnetic lineaments are used as inputs to test if structural complexity can refine or add confidence to existing geological interpretations. Linear, low variance trends frequently delineate the location and extent of deformation zones, which likely reflect parallel transposed fabrics in high strain corridors. Areas of high variance spatially correlate with fold traces. Sharp boundaries between high and low variance anomalies are interpreted as faults and/or shear zones that juxtapose structural blocks. The near continuous and representative nature of autodetected magnetic lineaments was effective in capturing structural complexity at a regional scale; however, spherical dispersion captured the highest resolution of structural complexity at local scales ( μ + 1σ). Additionally, the gold grade of large deposits (> 1 Moz) increases with proximity to high variance anomalies. When paired with expert knowledge, these techniques will increase repeatability in future exploration endeavors, making exploration a more rigorous process with increased confidence.

Understanding the architecture of the West African craton at depth is essential to be able to reconstruct its evolution. Our study focuses on the crustal imaging of structures and geometries characterizing the crust of the Leo-Man shield with broadband and long period magnetotelluric data collected in southern Burkina Faso and covering a 220 km long profile. The resulting 3D resistivity crustal model highlights the distribution of the granite-greenstone assemblages with depth showing excellent correlation with mapped surficial lithologies. The whole crust of southern Burkina Faso is resistive, with lateral as opposed to vertical major resistivity contrasts, reflecting the location of major-scale shear zones characterizing this part of the Baoule-Mossi domain. Ground gravity data acquired along the same line as the MT data were also modeled and show relatively good correspondence with the resistivity model. The new resistivity and gravity models compared with results from joint inversion of gravity and aeromagnetic data highlight significant changes between the greenstone belts and granitoid domains along the profile. The comparison of the geophysics with the geology enables us to define new depth constraints of the main tectonic features in the area. The observed large scale dipping shear zones favour the model of crustal building through major parallel thrust faults.

The structural setting of the Pontiac Subprovince in the vicinity of the world-class Canadian Malartic gold deposit has been revisited by combining and reinterpreting airborne geophysical surveys together with a century of structural observations. Felsic-intermediate intrusive bodies are a key component of this deposit. Defining the regional and local favorable structural setting for intrusive rock emplacement within the clastic meta-sedimentary rocks of the Pontiac Group, south of the Cadillac Larder Lake Deformation Zone, may further assist gold exploration in similar tectonic environment. Three structural domains are interpreted in the area based on the geometry of the bedding, folds and structural fabrics related to the three major phases of deformation. During these events four phases of magmatism and one metamorphic episode occurred. The North domain, which hosts the Canadian Malartic deposit, is characterized by highly variable bedding orientations produced by the interference patterns of isoclinal F1 folds overprinted by open to tight, steeply dipping, F2 folds. The bedding in this domain is cut by a penetrative S2 biotite foliation, which possibly built on rheological changes enhanced by metasomatism in the footprint of the Canadian Malartic deposit. By contrast, the Central and South domains display homogeneous bedding orientations cross-cut by a discrete S2 biotite foliation and syn- to late- D2 metamorphic porphyroblasts. In all domains the L2 stretching lineation consistently plunges at about 60 degrees toward the east. D3 is a minor deformation event in the Pontiac Subprovince, which possibly correlates with late dextral transcurrent movement along the Cadillac Larder Lake Deformation Zone. Intrusive bodies were emplaced throughout the first and second deformation events. Phase 1 monzonite, quartz-monzodiorite and granodiorite plutons (ca. 2683–2680 Ma) intruded into consolidated Pontiac sedimentary rocks during D1. Phase 2 quartz-monzodiorite bodies (ca. 2679–2676 Ma) predominantly formed in the North Domain along F1 fold axial surfaces prior to or at the onset of D2. Phase 3 basic dykes (ca. 2675–2673 Ma) cross-cut earlier felsic-intermediate intrusions across all domains and subsequently developed an S2 metamorphic hornblende foliation. Phase 4 magmatism (ca. 2672–2662 Ma) is related to the Decelles Batholith S-type granite and pegmatite, which is interpreted to be contemporaneous with the peak of regional metamorphism. The Decelles Batholith may project at depth underneath the Canadian Malartic deposit and could be associated with magmatic-hydrothermal mineralizing fluids in the Canadian Malartic deposit. Gold mineralization at Canadian Malartic is spatially located on the contact of Phase 2 quartz-monzodiorite bodies. The proximity to the Cadillac Larder Lake Deformation Zone combined with the rheological contrast between steeply dipping clastic meta-sedimentary rocks and quartz-monzodiorite intrusions favored the protracted failure of the contacts between these two rock masses, thus forming favorable conduits for hydrothermal fluids (e.g., the Sladen Fault Zone). Such specific rheological behavior is demonstrated by domains of structural complexity, emphasized by the variance of the bedding dip. These structurally complex zones systematically host gold mineralization in the Canadian Malartic district of the Pontiac Subprovince proximal to the Cadillac Larder Lake Deformation Zone.

The Vazante deposit, which is the world's largest zinc silicate deposit, occurs in brecciated dolomite and comprises mainly willemite with various proportions of hematite, Fe-carbonate, minor franklinite and magnetite. Exploration for this type of deposit is more challenging than zinc sulfide deposits, as they do not exhibit similar geophysical anomalies. To improve the application of geophysical surveys to the exploration of hypogene silicate zinc deposits, data from 475 samples were investigated from drill holes representative of the various types of ore and host rocks as well as barren zones of known geophysical anomalies in the Vazante District. Lithogeochemical and mineralogical (optical, SEM and MLA) data were integrated with physical rock properties (density, magnetic susceptibility and K U Th gamma-ray spectrometry) to assist in exploring for this type of deposit. The most distinct physical property of the ore is density, compared with the host rocks due to high proportion of denser minerals (hematite and willemite). However, barren hematite breccias also have high densities. The zinc ore and hematite breccias yielded higher magnetic susceptibilities than the surrounding host rocks, with the highest values associated with greater proportions of franklinite and magnetite. The density and magnetic susceptibility contrasts are a result of hydrothermal fluids interacting with and altering the carbonate host rocks. Zinc ore also yielded elevated U concentrations relative to the various host rocks, yielding higher gamma-ray spectrometric values. The results of this investigation indicate that an integration of magnetic, gravimetric and radiometric surveys would be required to identify zinc silicate ore zones.

The West-African craton is defined by a combination of Archean and Palaeoproterozoic rocks that stabilised at ~2 Ga towards the end of the Paleoproterozoic Eburnean Orogeny, and therefore may reflect the transition from Archean to modern tectonic processes. Exploring its present lithospheric architecture aids further understanding of not only the craton’s stability through its history but also its formation. We investigate the lithospheric structure of the craton through analysing and modelling magnetotelluric (MT) data from a 500-km-long east-west profile in northern Ghana and southern Burkina Faso crossing part of the Baoule-Mossi Domain and reaching the Volta Basin in the south-eastern part of the craton. Although the MT stations are along a 2D profile, due to the complexity of the structures characterising the area, 3D resistivity modelling of the data is performed to obtain insights on the thermal signature and composition of the subcontinental lithosphere beneath the area. The thermal structure and water content estimates from different resistivity models highlight a strong dependence on the starting model in the 3D inversions, but still enable us to put constraints on the deep structure of the craton. The present‐day thermal lithosphere‐asthenosphere boundary (LAB) depth is estimated to be at least 250 km beneath the Baoule-Mossi domain. The area likely transitions from a cold and thick lithosphere with relatively low water content into thinner, more fertile lithosphere below the Volta Basin. Although the inferred amount of water could be explained by Paleoproterozoic subduction processes involved in the formation of the Baoule-Mossi domain, later enrichment of the lithosphere cannot be excluded

Canadian Malartic, with a total endowment of 16.3 Moz Au, is an important example of a large-tonnage, low-grade Archean gold deposit (current reserves of 204 Mt @ 1.08 g/t Au). It is located in the southern Superior Province in contact with, and immediately south of the east-west trending Cadillac-Larder Lake fault zone, which delineates the boundary between the Pontiac and Abitibi subprovinces. The deposit is hosted by Pontiac Group metaturbidites, Piche Group mafic-ultramafic metavolcanics and by porphyritic quartz monzodiorite to granodiorite that intruded these lithologies. The metamorphic grade increases southward from upper greenschist facies, which characterizes the immediate metamorphic environment of the Canadian Malartic deposit, to mid-amphibolite facies; the garnet isograd is crossed 1.5 km south of the deposit, and the staurolite isograd about one kilometer further south. Textural, structural and geochronological observations are consistent with a syn-kinematic (D2), early- to syn-peak metamorphic (M2) timing for the main ore-forming event. Mineralization in the metasedimentary rocks and porphyritic intrusions consists of disseminated native gold and minor gold-tellurides within stockworks of quartz-biotite-microcline-carbonate±pyrite veinlets (v2) and associated pervasive microcline-albite-biotite±(white mica)-carbonate-pyrite alteration. A zonal distribution of alteration is centered on structures that acted as preferential pathways for the hydrothermal fluids (e.g., Sladen Fault, NW-SE deformation zones). In the metasedimentary rocks, the proximal potassic alteration zone is dominated by microcline-albite±quartz with variable proportions of phlogopite (±white mica), carbonate minerals (calcite-ankerite±Fe-dolomite), quartz, pyrite and rutile. It grades outwards into a distal potassic-sericitic alteration zone characterized by relatively abundant Mg-rich biotite and phengitic white mica, as well as microcline, albite, quartz, calcite, pyrite and rutile. The zonal distribution of alteration features was due in part to a decrease in the total activity of sulfur species (∑aS) and oxygen fugacity (fO2) away from the main hydrothermal corridors, which is manifested by systematic changes in Fe-sulfide and (Fe-Ti)-oxide mineralogy. The effects of sulfidation (∑aS), oxidation (fO2) as well as K+, Fe2+ and H+ activity on biotite and white mica compositions were assessed from silicate-oxide-sulfide equilibria. Increasing ∑aS-fO2 conditions proximal to the hydrothermal fluid pathways caused pyrite to be stabilized over biotite, and iron sequestration in the sulfide phase in turn promoted the stability of magnesian biotite. Increasing sulfur metasomatism towards the hydrothermal centers was associated with systematic increases in biotite Mg# [Mg/(Fetotal+Mg)] and fluorine concentration, in agreement with the Fe-F avoidance principle. Alteration was also associated with a decrease in Al in biotite and white mica, coincident with increases in Si and Fe + Mg concentrations. These compositional trends were controlled by a Tschermak exchange reaction and are consistent with a gradual decrease in aK+ and/or pH upon progressive rock-buffering of a mildly alkaline, potassium-rich ore-forming fluid. Multicomponent phase equilibria (pseudosections), constrained for specified bulk rock compositions typical of non-mineralized Pontiac Group metasedimentary rocks, are used to model the expected composition of mica solid solutions under varying metamorphic P-T conditions. The thermodynamic modelling of biotite and white mica compositions suggests that the district-scale compositional trends documented for biotite and white mica along the metamorphic gradient, including the progression of the Tschermak substitution towards more aluminous compositions southwards, were likely the product of increasing metamorphic P-T conditions, rather than a distal effect of hydrothermal alteration. Mica mineral chemistry is clearly a sensitive indicator of hydrothermal and metamorphic processes at Canadian Malartic. Our results show that biotite and white mica compositions provide valuable tools that can be used to define hydrothermal fluid pathways in and around the deposit. These factors should be particularly useful for mapping the zonation of alteration features that characterize the footprint of major gold deposits in metamorphic terranes where micas typically display a spatial distribution ideal for defining mineral-chemical vectors to ore.

The Ashanti greenstone belt in southwest Ghana hosts many gold deposits distinguished by different timing and structural contexts. This study investigates the evolution of the Wassa system by integrating field and geophysical observations. This 4 million ounces (past production and current resources) gold deposit is interpreted to represent the oldest gold mineralization event in West Africa with gold-bearing pyrites aligned and stretched within the S1 ductile fabric. Mineralized quartz-carbonate veins were strongly deformed during the D1 deformation event. Three additional folding events are characterized by hectometer-scale tight to isoclinal folds, by a kilometer-scale synform fold centered on the mine and by a late recumbent metric-scale folds. Because of its early timing, the Wassa system represents a new poly-deformed deposit type in West Africa and highlights a potential for new discoveries in the underexplored meta-volcanic and meta-sedimentary Sefwi Group. Timing of the gold mineralization at the Wassa mine makes this deposit type a possible candidate for the source of the gold contained in the Tarkwa paleoplacer.

The Ashanti Belt in Ghana hosts numerous multi-million ounce gold deposits and is one of the most richly gold endowed Paleoproterozoic belts of the West African Craton. This work shows that the Wassa mineralized intrusion is part of the Sefwi Group. This unit at Wassa is strongly magnetic and show a distinctly high response in regional magnetic data sets compared to other units of equivalent age within the belt. The unit is inferred to be a lateral extension of an exposed fragment of what defines the substrate to the Tarkwa Basin sediments. The Wassa deposit, located in the eastern limb of the belt, is hosted within mafic to intermediate volcanic flows that are interbedded with minor horizons of volcaniclastics, clastic sediments. The clastic sediments include wackes and magnetite rich sedimentary layers, presumably derived from banded iron formations. The previously described sequence is intruded by syn-volcanic mafic intrusives and felsic porphyries rocks that are all part of the Birimian stratigraphy. Two new key SHRIMP II U–Pb ages were determined as part of this study: a new age of 2191 ± 6 Ma was determined on magmatic zircon grains of the Wassa porphyry host rock, which now represents the oldest known felsic intrusion hosting gold mineralization in the Ashanti Belt region. The Benso gold deposit system, which is located in the eastern limb of the Ashanti Belt approximately 38 km southwest of Wassa is hosted within a series of volcanic units intruded by mafic to intermediate units. A SHRIMP II U–Pb age of 2157 ± 5 Ma was determined from magmatic zircons obtained from a granodiorite of the G-Zone of the Benso deposit. This granodiorite is the main host rock for gold mineralization and thus the age provides an upper constraint for mineral emplacement. The newly determined ages provide an upper constraint for the gold mineralization within this region of the Ashanti Belt. They also support recent structural studies that have interpreted that the Wassa mineralization is related to the Eoeburnean period, defined to have occurred between 2191 and 2158 Ma (D1 event) prior to the main phase of gold mineralization. This new U–Pb age on the host rock provides an upper age limit to the Wassa mineralization of ca. 2191 Ma and is compatible with the proposed Eoeburnean early mineralization model for the Wassa deposit. The age of the intrusion along with the location of the deposit in the Sefwi Group give a unique characteristic that distinguishes Wassa from the other gold deposits along the Ashanti belt which are believed to have been the result of a mineralizing event between ca. 2090 and 2070 Ma.

Studies of mafic dyke swarms may simultaneously provide information on the mechanical, geochemical, geochronological and magnetic environments at the time of their formation. The mafic intrusive history of different cratons can also be potentially used to unravel their assembly into their current configuration. The identification and classification of dykes is a first step to all these studies. Fortunately, even in regions with poor outcrop, we can use the strong magnetic response of mafic dykes to identify and map their extent. In West Africa the first maps of mafic dyke distribution were made over 40 years ago, but there are still large areas where there are almost no published data. In this paper we present a significantly updated map of mafic dykes for the West Africa Craton based in large part on new interpretations of the regional airborne magnetic database. This map includes the locations of over three thousand dykes across the craton, which locally shows several orientation clusters that provide a minimum estimate for the total number of dyke swarms in this region. Whilst we will have to wait until systematic dating of the different swarms is completed, we can demonstrate that there is a long and complex history of mafic magmatism across the craton, with up to 26 distinct dyke swarms mapped based according to their orientation. The mapping and dating of these swarms will provide key constraints on the assembly of the fragments that make up the modern continents.

International audience; We estimate using gravity data the thickness of post-glacial unconsolidated sediment filling two major glacial valleys in northern Pyrenees: the Gave de Pau valley between Pierrefitte-Nestalas and Lourdes, and the Garonne valley between Saint-Béat and Barbazan. One hundred and eighty-four new gravity data complete 74 measurements obtained from the International Gravimetric Bureau database. Negative residual anomalies resulting from the presence of small-density unconsolidated sediment approach 4 mgal in both the Gave de Pau and the Garonne valleys. Estimating the sediment thickness requires knowing the density contrast between Quaternary sediments and the underlying bedrock. Supposing this density contrast is 600 kg/m3, the maximum estimated thickness of post-glacial sediment is ∼ 230 and 300 meters, and the volume of sediment is 2.1 and 3.2 km3 in the Gave de Pau and Garonne valleys, respectively. In both valleys, the depth of Quaternary sediment suddenly increases at the confluence between two major glacial valleys (Gave de Pau – Gave de Cauterets, and Garonne – Pique confluences). Overdeepened basins are less deep downstream when approaching terminal moraines (Lourdes and Barbazan area), illustrating that the efficiency of glacial erosion depends on the ice flux flowing through valleys.

Geophysical inversion employs various methods to minimize the misfit between geophysical datasets and three-dimensional petrophysical distributions. Inversion techniques rely on many subjective inputs to provide a solution to a non-unique underdetermined problem, including the use of a priori model elements (i.e. a contiguous volume of the same litho-stratigraphic package), the a priori input model itself or inversion constraints. In some cases, inversion may produce a result that perfectly matches the observed geophysical data, but can still misrepresent the geological system. A workflow is presented here that offers objective methods to provide inputs to inversion: (1) simulations are performed to create a model suite that contains a range of geologically possible models; (2) stratigraphic variability is determined via uncertainty analysis to identify low certainty model regions and elements; (3) geodiversity analysis is then conducted to determine geometrical and geophysical extremes and commonalities within the model space; (4) geodiversity metrics are simultaneously analysed using principal component analysis to identify the contribution of different model elements toward overall model suite uncertainty; (5) principal component analysis also determines which models exhibit diverse or common geological and geophysical characteristics which (6) facilitate the selection of models as inputs to geophysical inversion. This workflow is applied to a three-dimensional model of the Ashanti Greenstone Belt, southwestern Ghana in West Africa in order to reduce the subjectivity incurred during decision making, explore the range of geologically possible models and provide geological constraints to the inversion process to produce geologically and geophysically robust suites of models. Results further suggest that three-dimensional uncertainty grids can optimize inversion processes and assist in finding geologically reasonable solutions.

Canadian Malartic is a large-tonnage, low-grade Archean gold deposit (16.3 Moz, 1.08 g/t Au) located in the Abitibi region of Quebec, Canada. A large part of the mineralization is hosted in the Pontiac Group metasedimentary rocks, which consist of mudstones to greywackes at upper greenschist to amphibolite facies. In exploration and production environments, these lithologies are challenging to characterize by conventional core logging methods, while in a research setting thin section-sized samples (2 cm × 4 cm) do not capture the full extent of mineralogical variability, which can extend from centimeters to meters away from mineralized zones. Here, high-resolution hyperspectral imagery (0.2–1.0 mm/pixel) in both shortwave infrared (SWIR, 1000–2500 nm) and longwave infrared (LWIR, 8000–12000 nm) is acquired for over two thousand meters of drill core, and is used to visualize changes in mineralogy and mineral chemistry related to metamorphism and hydrothermal alteration. Unaltered metasedimentary rocks contain metamorphic white mica with AlVI contents varying between 1.90 and 1.75 apfu (2195 to 2203 nm), depending on metamorphic grade. Hydrothermal alteration is characterized by white mica which becomes progressively more phengitic with increasing alteration intensity, with AlVI contents ranging from 1.70 to 1.50 apfu (2204 to 2212 nm). Phengitic white mica extends from meters to tens of meters away from major mineralized zones, and can be used as a vector towards mineralization in an exploration setting. White mica composition is correlated to Au content, and can be used to discriminate between unmineralized ( 0.3 g/t Au, 2205–2208 nm), and highly mineralized (>1.0 g/t Au, >2208 nm) samples, which is a simple metric that can directly be applied to sort ore in a production environment. The Mg# (molar Mg/[Mg + Fe]) of biotite, on the other hand, is unaffected by metamorphic grade, and consistently is Mg# 55–60 (2251–2250 nm) in unaltered samples. In mineralized samples, biotite is Mg-rich (Mg# >65

Geophysical inversion employs numerical methods to minimise the misfit between three-dimensional petrophysical distributions and geophysical datasets. Inversion techniques rely on many subjective inputs to provide a solution to a non-unique problem, including use of an a priori input model or model elements (a contiguous volume of the same litho-stratigraphic package) and inversion constraints. Inversions may produce a result that perfectly matches the observed geophysical data but still misrepresents the geological system. A workflow is presented that offers objective methods to provide inputs to inversion. First, simulations are performed to create a model suite that contains a range of geologically possible models. Next, uncertainty analysis is performed using stratigraphic variability to identify low certainty model regions and elements. Geodiversity analysis is then conducted to determine the geometrical and geophysical extremes within the model space. Next, geodiversity metrics are then simultaneously analysed using principal component analysis to identify the geometrical and geophysical aspects that contribute most toward model suite variability. Principal component analysis determines which models exhibit common or diverse geological and geophysical characteristics, facilitating selection of models subjected to geophysical inversion. We apply this workflow to the Ashanti Greenstone Belt, southwestern Ghana in west Africa. The workflow described in this manuscript reduces the subjectivity during decision making, explores the range of geologically possible models and provides geological constraints to the inversion process with the aim of producing geologically and geophysically robust suites of models associated with an uncertainty grid.

International audience; The process of 3-D modelling forces the operator to consider data collection and processing error, while simultaneously making assumptions about geology during interpretation, to arrive at the most likely or logical geological scenario. These kinds of ambiguities lead to situations where multiple model realizations can be produced from a single input data set. Decisions are typically made during the modelling process with the aim of reducing the number of possible models, preferably to produce a single geological realization. These types of decisions involve how input data are processed and what data are included, and are always made without complete knowledge of the system under study. This regularly, if not always, results in natural geometries being misrepresented by the model, which can be attributed to uncertainty inherent in the modelling process. Uncertainty is unavoidable in geological modelling as complete knowledge of the natural system is impossible, though we use many techniques to reduce the amount introduced during the modelling process. A common technique used to reduce uncertainty is geophysical forward modelling, and the misfit between the calculated and observed response provides a means to gauge whether changes in model architecture improve or degrade the quality of the model. Unfortunately, geophysical data are ambiguous and provide a non-unique solution, with different model geometries able to produce the same geophysical response. We propose a process whereby multiple models, collectively known as the `model suite', are produced from a single data set that allows an exploration of geological model space. Various `geodiversity' metrics have been developed to characterize geometrical and geophysical aspects of each model. Geodiversity measurements are combined into multivariate analysis to reveal relationships between metrics and define the boundaries of the geological possibility. A previous study using geodiversity metrics on the Gippsland Basin is extended here by including geophysical metrics. We use the Ashanti Greenstone Belt, southwestern Ghana in West Africa, as a case study to assess the usefulness of the technique. A critical assessment of the 3-D model is performed and aspects of the model space are identified that could be of interest to gold explorers.

The process of building a 3D model necessitates the reconciliation of field observations, geophysical interpretation, geological data uncertainty and the prevailing tectonic evolution hypotheses and interpretations. Uncertainty is compounded when clustered data points collected at local scales are statistically upscaled to one or two points for use in regional models. Interpretation is required to interpolate between sparse field data points using ambiguous geophysical data in covered terranes. It becomes clear that multiple interpretations are possible during model construction. The various interpretations are considered as potential natural representatives, but pragmatism typically dictates that just a single interpretation is offered by the modelling process. Uncertainties are introduced into the 3D model during construction from a variety of sources and through data set optimisation that produces a single model. Practices such as these are likely to result in a model that does not adequately represent the target geology. A set of geometrical ‘geodiversity’ metrics are used to analyse a 3D model of the Gippsland Basin, southeastern Australia after perturbing geological input data via uncertainty simulation. The resulting sets of perturbed geological observations are used to calculate a suite of geological 3D models that display a range of geological architectures. The concept of biodiversity has been adapted for the geosciences to quantify geometric variability, or geodiversity, between models in order to understand the effect uncertainty has models geometry. Various geometrical relationships (depth, volume, contact surface area, curvature and geological complexity) are used to describe the range of possibilities exhibited throughout the model suite. End-member models geodiversity metrics are classified in a similar manner to taxonomic descriptions. Further analysis of the model suite is performed using principal component analysis (PCA) to determine important geometrical characteristics. The configuration of the model space is determined through identifying ‘outlier’ model examples, which potentially represent undiscovered model ‘species’.

International audience; Including past production, current indicated and inferred resources, Wassa is a 5 Moz poly-deformed early-orogenic gold deposit located on the eastern flank of the Ashanti Belt, in southwest Ghana. It is hosted by metamorphosed volcanic, intrusive and sedimentary rocks of the Sefwi Group (ca. 2260–2160 Ma). Early mineralization has an Eoeburnean age (2164 ± 22 Ma, Re–Os on pyrite) and is characterized by quartz veins, by a carbonate alteration of the host rocks, and by deformed gold-bearing pyrite. Remobilization of this gold occurred during the late stages of the Eburnean Orogeny (~ 2.1 Ga) and is associated with quartz-carbonate veins with visible gold and euhedral pyrites.

Integration of regional geophysical datasets and detailed field observations provide new insights into the paleoproterozoic structural evolution of southwestern Ghana. The study area is dominated by three metavolcanic and metasedimentary packages known as the Sefwi Group, the Kumasi Group (Birimian) and the Tarkwa Group (Tarkwaian) that were intruded by abundant TTG granitoids during the Eoeburnean and Eburnean phases of an event termed the “Eburnean Orogeny”. This study identifies an Eoeburnean (pre-Tarkwaian) deformation event (D1) that produced significant deformation in the Sefwi Group metavolcanics. D1 is associated with N-S shortening manifested as regional scale folding in the southern Ashanti Belt. D1 synorogenic granitoids were intruded between 2187 Ma and 2158 Ma under greenschist metamorphic condition. Syn-D1 gold mineralisation associated with quartz veining could be the original source of Tarkwaian paleo-placers and/or remobilised gold concentrations along major shear zones. D2 represents an extensional phase associated with the Kumasi Group sedimentation (2154–2125 Ma) which could be related to activation of major structures such as the Ashanti Fault as low angle detachments that controlled the deposition of the Kumasi Group and the opening of the Kumasi and Akyem Basin. The Tarkwa Group (2107–2097 Ma) unconformably overlies the Birimian Supergroups and was deposited in response to D3 shortening. D3 resulted in the inversion of syn-D2 detachments faults within the Ashanti Belt. NW-SE D3 shortening produced regional scale folding within the Birimian and the Tarkwaian metasediments. D4 deformation corresponds with sinistral reactivation of D3 thrust faults, and is locally associated with macro-scale folding at Obuasi and Wassa gold mines. By the end of D4, the regional scale architecture was built and was only slightly modified by the two last events. D5 postdates the Eburnean metamorphic peak and corresponds to open recumbent folds associated with a subhorizontal crenulation cleavage. D6 is present as a subvertical crenulation cleavage and reverse faults associated with NE-SW shortening.

International audience; We solve the 3-D gravity inverse problem using a massively parallel voxel (or finite element) implementation on a hybrid multi-CPU/multi-GPU (graphics processing units/GPUs) cluster. This allows us to obtain information on density distributions in heterogeneous media with an efficient computational time. In a new software package called TOMOFAST3D, the inversion is solved with an iterative least-square or a gradient technique, which minimizes a hybrid L1-/L2-norm-based misfit function. It is drastically accelerated using either Haar or fourth-order Daubechies wavelet compression operators, which are applied to the sensitivity matrix kernels involved in the misfit minimization. The compression process behaves like a pre-conditioning of the huge linear system to be solved and a reduction of two or three orders of magnitude of the computational time can be obtained for a given number of CPU processor cores. The memory storage required is also significantly reduced by a similar factor. Finally, we show how this CPU parallel inversion code can be accelerated further by a factor between 3.5 and 10 using GPU computing. Performance levels are given for an application to Ghana, and physical information obtained after 3-D inversion using a sensitivity matrix with around 5.37 trillion elements is discussed. Using compression the whole inversion process can last from a few minutes to less than an hour for a given number of processor cores instead of tens of hours for a similar number of processor cores when compression is not used.

A three-dimensional model of the gold-rich Paleoproterozoic Ashanti Greenstone Belt has been built with a “hybrid” geological modeling method using Geomodeller and Gocad, then inverted using VPmg software. The model helps to identify the main regional-scale controls leading to abundant gold mineralization in the southeast of the belt by highlighting the lithological distribution and geometry of geological structures. The three-dimensional geological model shows that 85% of gold occurrences in the Ashanti Greenstone Belt are proximal to a specific tholeiitic metabasaltic horizon “BV1”. This result is noteworthy given that the relatively thin, 300 m thick, BV1 layer represents only 10%, by volume, of the modeled Sefwi Group stratigraphy. A map of distance between the topography and the BV1 layer is produced and identifies areas showing a high probability to host gold deposits in the southeast of the Ashanti Belt. Possible structural controls performed by the BV1 tholeiitic metabasaltic rocks on the location of gold occurrences are discussed, and we hypothesize that the BV1 layer acted as a mineralization trap or source of gold. Both hypotheses agree with examples from the literature and explain the observed distribution of the mineralized sites.