At the Gibraltar porphyry Cu-Mo deposit in south-central British Columbia, Canada, geochemical and mineralogical anomalies in till around a cluster of mineral occurrences form amoeboid-shaped dispersal patterns controlled by three phases of ice movement. The glacial dispersal patterns defined by elevated concentrations of ore (Cu, Mo) and pathfinder (Ag, Zn) elements and alteration oxides (Al2O3, K2O) in the till matrix (clay and silt plus clay fractions) extend over areas of 6–36 km2 and are centered on the main economic mineralization. Mineralogical dispersal trains were identified by the high abundances of chalcopyrite, pyrite, jarosite, titanite, and epidote, which were derived from the mineralization and its associated oxidation and alteration zones. These mineralogical anomalies in till extend approximately 30 km2 for chalcopyrite, 62 km2 for jarosite, and up to 136 km2 for epidote. Epidote grains in till with a trace element composition of >10 ppm As, >4 ppm Sb, <200 ppm total rare earth elements (ΣREEs), <45 ppm Y, <100 ppm Sc + Cr + Y, and/or <2 ppm Th + Hf are interpreted as being either principally derived from the intrusive rocks or from hydrothermal alteration of the porphyry mineralization. Rare epidote grains with >30 ppm Cu, interpreted to be related to the porphyry Cu mineralization, were detected in till up to 7 km down-ice from the economic mineralization. This study demonstrates that till geochemistry and mineralogy, combined with geochemical analyses of specific minerals such as epidote, are efficient methods for identifying porphyry Cu mineralization in terrains covered by glacial sediments.

The Canadian Cordillera is well known for its potential for porphyry Cu mineralization predominantly associated with felsic to intermediate intrusions of the Quesnel and Stikine arc terranes (Fig. 1). There, porphyry Cu mineralization occurs in intrusions related to two specific time windows: (1) a preaccretionary period extending from the Late Triassic to Early Jurassic and (2) a postaccretionary Late Cretaceous to Eocene episode (McMillan et al., 1995; Logan, 2013; Logan and Mihalynuk, 2014). In their assessment of porphyry Cu mineralization in the Canadian Cordillera, Mihalasky et al. (2011) estimate (based on 2010 data) that in addition to the 66.8 Mt of Cu in known deposits, approximately 49 Mt of Cu remain undiscovered. The prospective geologic environment in the Canadian Cordillera is part of a glaciated landscape with poor bedrock exposure and a nearly continuous cover of glacial sediments—hence, the need to develop exploration methods for deposits under cover.

To overcome the challenge of seeing through the glacial sediments, we present the results of a combined till geochemistry and mineralogy regional survey conducted around the Gibraltar Cu-Mo porphyry deposit in south-central British Columbia, Canada. We demonstrate that within the glaciated landscape of the Interior Plateau, characterizing the composition of glacial sediments directly deposited by glaciers (till) and the trace element content of epidote and zircon recovered from till can identify the presence of porphyry Cu mineralization defined by anomalies that extend over 100 km2.

Located 50 km north of Williams Lake, British Columbia, the Gibraltar mine is operated by Taseko Mines Ltd. and is currently the second largest open-pit mine in Canada (www.tasekomines.com). It is a calc-alkaline porphyry Cu-Mo deposit hosted by the Late Triassic Granite Mountain batholith (Fig. 1; van Straaten et al., 2013, 2020) with combined past production and current resources of 1.44 billion tonnes (Gt) with a Cu grade of 0.28% and an estimated Mo grade of 0.009% (van Straaten et al., 2020). The Granite Mountain batholith is compositionally zoned with a predominance of tonalite and lesser quartz diorite (Fig. 2; Drummond et al., 1973, 1976; Bysouth et al., 1995; Ash et al., 1999a; Ash and Riveros, 2001; Schiarizza, 2014, 2015). The batholith is divided into four phases based on rock types, texture, and alteration, which are, from southwest to northeast, the border phase, mine phase, Granite Mountain phase, and Burgess Creek stock (Fig. 2). For this paper, and to facilitate correlation with previous geologic maps, this fourfold classification is maintained. In the absence of distinct contacts between the four intrusive phases and their similar mineral modalities and textures, Kobylinski et al. (2020) suggested a simpler twofold classification scheme for the Granite Mountain batholith: Cu-bearing tonalite and barren tonalite. Zircon geochronology presented by Kobylinski et al. (2020) for the Granite Mountain batholith defines multiple intrusive phases starting at ca. 229 Ma and extending over 25 m.y. How the Granite Mountain batholith is subdivided does not affect our interpretation of till composition.

In the northeast, the batholith intrudes Nicola Group sedimentary and volcanic rocks (Quesnel terrane). It is in fault contact with basalt, chert, and limestone of the Carboniferous to Lower Jurassic Cache Creek Complex in the southeast and with the Early Cretaceous Sheridan Creek stock to the south (Schiarizza, 2014, 2015). The western sector of the intrusion is overlain by an overlap assemblage of Eocene to Neogene basaltic volcanic and sedimentary rocks. Postmineralization regional greenschist facies metamorphism and ductile deformation affected the Granite Mountain batholith (Bysouth et al., 1995; Ash et al., 1999b; van Straaten et al., 2013, 2020).

The Gibraltar deposit includes five major mineralized zones: Gibraltar East and West, Pollyanna, Granite, and Connector (Fig. 2) (van Straaten et al., 2020). Bysouth et al. (1995) defined sulfide zonation at Gibraltar as a Cu-Mo core (chalcopyrite, molybdenite, ± bornite) grading outward to a Cu zone (chalcopyrite ± molybdenite), a Cu-Zn-Ag zone (chalcopyrite, sphalerite), and a Zn zone (sphalerite ± chalcopyrite) extending at least 6 km to the west of the Pollyanna and Granite Lake mineralized zones. In addition to the main deposits, 19 porphyry mineral showings and prospects are reported within the Gibraltar mine region (British Columbia Geological Survey, 2020) (Fig. 2).

As commonly observed in porphyry Cu deposits (e.g., Lowell and Guilbert, 1970; Sinclair, 2007; Sillitoe, 2010; Cooke et al., 2014b, 2020a), propylitic alteration, in which epidote and chlorite formed as part of the hydrothermal alteration process, is present at the Gibraltar deposit and forms a >100 km2 halo centered on mineralization (Drummond et al., 1976; Bysouth et al., 1995; van Straaten et al., 2013; Kobylinski et al., 2020). Kobylinski et al. (2020) note an increasing abundance of plagioclase replacement by epidote, from <5 to ~30%, within 3 km of the main ore zones.

The Gibraltar deposit is located in the Interior Plateau physiographic region of British Columbia (Holland, 1976). The region has a subdued topography with isolated mountain ranges separated from the contrasting topographic barriers of the Coast Mountains to the west and the Cariboo Mountains to the east (Fig. 1). The deposit is located on the western flank of Granite Mountain, which has a rounded summit at 1,390 m above sea level (a.s.l.)—somewhat higher than the surrounding region, which is at an average elevation of 850 to 1,150 m a.s.l. Glacial sediment cover is nearly continuous with a predominance of till that can reach a thickness of 10–15 m as exposed in the periphery of the mine pits (Fig. 3). Lesser amounts of glaciofluvial sand and gravel deposits are confined to valleys and lower areas (Plouffe and Ferbey, 2015). Bedrock exposures are limited. Before mining started, only 5% of the mineralization from the southern part of Gibraltar West deposit and the limonitic leach caps of the Pollyanna and Gibraltar East deposits were exposed at surface, but the northern deposits Gibraltar West and Granite Lake were completely covered by glacial sediments (Rotherdam et al., 1972; Bysouth et al., 1995; van Straaten et al., 2013). In other words, most of the mineralization was under a glacial sediment cover.

Glacially streamlined landforms (flutings, crag and tails, and drumlins) and glacial striations, measured on outcrops in the Gibraltar mine region, are key to the interpretation of the ice-flow history and glacial transport directions of the last glaciation (Plouffe and Ferbey, 2015) (Fig. 4). Fifteen kilometers north of the Gibraltar mine, an unnamed mountain ridge, which is marked with fresh cirques and arêtes at 1,240–1,455 m a.s.l., likely acted as an ice source region at the onset of the last glaciation (Fig. 4A). Striations and rattails, measured on a single outcrop 3.5 km south of the Gibraltar mine, indicate an ice movement to the southeast from this unnamed ridge. The summit of Granite Mountain was probably unglaciated during the early phase of glaciation, piercing through the ice surface as a nunatak (Fig. 4A). Our glacial reconstruction shows that only the peripheral subeconomic mineral occurrences were covered by ice at that time, not the main deposits. Evidence for this first phase of ice flow is scant, having been either eroded by subsequent ice movements or covered by glacial sediments of the subsequent ice-flow phases.

As glaciation intensified, ice from the Cariboo Mountains advanced southwesterly over the Gibraltar region (Fig. 4B). The westward extent of ice during this second phase of ice flow is uncertain. At glacial maximum, ice from the Coast and Cariboo mountains coalesced over the Interior Plateau, resulting in the formation of an ice divide at about 52° N latitude, south of the Gibraltar mine. Ice from this ice divide flowed to the north to northwest in the Gibraltar region (Fig. 4C). This third and dominant phase of ice flow is well represented by the streamlined glacial landforms, as depicted in the surficial geology map of the Gibraltar region (Plouffe and Ferbey, 2015) (Fig. 3B). The last two phases of ice flow have been recognized elsewhere over the Interior Plateau of British Columbia, west of the Cariboo Mountains (Plouffe et al., 2011, 2016).

In summary, the Gibraltar region was subjected to three ice-flow movements during the last glaciation: (1) to the southeast, (2) to the southwest, and (3) to the north to northwest. Metal-rich debris eroded by ice from a cluster of mineral occurrences was likely entrained by any combination of these three ice movements.

Till samples for geochemical and mineralogical analyses were taken following field protocols described by Plouffe et al. (2013) and McClenaghan et al. (2020). Surface glacial sediment samples (n = 97) were collected from a well-compacted, sandy silt to silty sand diamicton that contains abundant striated and faceted clasts, including distal rock types. The diamicton shows poorly to well-developed fissililty and is interpreted to be a subglacial till. Till samples were collected in roadside exposures along forestry roads at an average depth of 1 m from the natural land surface, well below the oxidized soil horizon (B horizon). Sample spacing along roads was approximately 1 km, taking advantage of existing till exposures. At four locations, where more than 3 m of till was exposed, at least two samples were collected at different depths to assess the variability of till composition with depth. Two samples were collected at every sample site: a 9- to 15-kg sample for heavy mineral processing, and a 1- to 2-kg sample for matrix geochemical analyses. At each till sampling site, clasts from the dug pit and the till exposure were carefully examined to detect the presence of mineralization (e.g., chalcopyrite). Very rare clasts with evidence of mineralization were observed, but their low abundance was inconclusive as an indication of glacial dispersal from mineralization.

The clay (<0.002 mm) and silt plus clay (<0.063 mm) fractions were prepared from the 1- to 2-kg samples in the Sedimentology Laboratory of the Geological Survey of Canada (GSC, Ottawa, Ontario) following procedures outlined in McClenaghan et al. (2020). The clay fraction was separated by decantation and centrifugation, and the silt plus clay fraction by dry sieving. For geochemical analyses, silica blanks, duplicates, and primary standards were inserted in between routine samples (i.e., regional till samples) to establish the precision and accuracy of the data. Analyses were completed at Bureau Veritas Laboratory (Vancouver, British Columbia). Trace element contents (Ag, Cu, Mo, and Zn) were determined after an HCl/HNO3 leach (1:1) followed by inductively coupled plasma-mass spectrometry (ICP-MS). Major oxides (Al2O3, K2O) were measured after an LiBO2/Li2B4O7 leach, followed by fusion at 980°C, dissolution in 5% HNO3, and ICP-emission spectrometry (ICP-ES) analyses. All analytical results and quality assurance and quality control measures are reported in Plouffe and Ferbey (2016). The analytical precision for the six elements/oxides listed above is <±10%.

The clay fraction (<0.002 mm) from a limited number of samples was analyzed at the Mineralogy Laboratory of the GSC (Ottawa, Ontario) by X-ray powder diffraction analysis (XRD). Oriented smear mounts were prepared by pipetting 40 mg of material mixed in 1 mL of distilled water onto glass slides that were then air dried overnight. X-ray patterns were recorded using a Bruker D8 advance powder diffractometer equipped with a Lynx-Eye detector, Co radiation set at 35 kV and 40 mA. The samples were also X-rayed following saturation with ethylene glycol and heat treatment (550°C). Semiquantitative mineral analyses were completed based on the Reference Intensity Ratio (RIR) method (EVA software: Bruker.com).

The larger till samples (9–15 kg) were submitted for heavy mineral separation and identification to Overburden Drilling Management Ltd. (Ottawa, Ontario). In-house blank and spiked samples were processed and picked at the beginning and throughout the batch to monitor the precision and accuracy of the data (Plouffe et al., 2013). Duplicate samples of till were collected in the field to monitor the sediment heterogeneity. Each bulk till sample was wet sieved to <2 mm and processed on a shaking table to generate a heavy mineral preconcentrate (Fig. 5A). The preconcentrate was further processed in a heavy liquid (methylene iodide) diluted with acetone to specific gravities (SG) of 2.8 and 3.2 to generate mid- (2.8–3.2 SG) and high-density (>3.2 SG) fractions of the 0.25- to 2.00-mm material. Ferromagnetic minerals were removed using a hand magnet. The mid- and high-density fractions were further separated into three size fractions: 0.25–0.50, 0.5–1.0, and 1–2 mm. The high-density (>3.2 SG) 0.25- to 0.50-mm fraction was further separated with an electromagnet set at different amperages (< 0.6, 0.6–0.8, 0.8–1.0, and >1.0 amp) to facilitate heavy mineral identification based on the magnetic character of mineral grains. Mineralogists used a binocular microscope to visually identify heavy mineral species based on physical attributes such as color, luster, cleavage, and crystal habit. The identification of some mineral grains was verified with a scanning electron microscope (SEM). Mineral abundances are reported as the number of grains per 10 kg of <2 mm material or as percentage values. Epidote abundance in each sample, herein reported, was determined on the high-density (>3.2 SG) 0.25- to 0.50-mm fraction (Epidote1 in Fig. 5A). Plouffe et al. (2016) reported the abundance of epidote in the >3.2 SG, 0.25- to 0.50-mm, 0.8- to 1.0-A fraction (Epidote2 in Fig. 5A). All mineralogical results, including results of the quality assurance and control measures, and field duplicates are provided in Plouffe and Ferbey (2016), who show that the data is fit for geologic interpretation. The total abundance of epidote in till is included in the Appendix data file. For this paper, this mineral identification procedure is termed the optical method.

In addition to the optical mineral identification method described above, mineral species in the nonmagnetic >3.2 SG, 0.125- to 0.180-mm fractions were identified using an automated technique that combines an SEM and a mineral liberation analysis (MLA) database. In this paper, it is referred to as the MLA-SEM method. These analyses were completed at the Micro Analysis Facility (Memorial University, St. John’s, Newfoundland). The separation procedure for the >3.2 SG, 0.125- to 0.180-mm fraction is shown in Figure 5B. Only a selected number of samples (n = 20) were submitted for MLA-SEM based on their location up-ice and down-ice from mineralization. Aliquots weighing 0.3 g were mounted as a monolayer on 30-mm-diameter epoxy pucks as described by Wilton and Winter (2012). Each mount, one mount per sample, contains an estimated 20,000 mineral particles (Wilton et al., 2017). Mineral mounts were analyzed with an FEI Quanta 400 environmental SEM equipped with a dual Bruker energy dispersive spectrometer (EDS) detector utilizing the MLA software developed at the University of Queensland Julius Kruttschnitt Mineral Research Centre (JKTech) in Australia. The SEM was operated at a voltage of 25 kV and a beam current of 13 nA. The working distance between the sample and detector was 12 mm and the spot size was approximately 5 μm, representing the minimum detectable particle diameter. The imaging dwell time (i.e., time spent acquiring data) was 16 μs, with frame resolution of 800 pixels, and X-ray collection at 12 ms. Minerals identified by MLA-SEM are reported as an area percent of the total area of all minerals mapped on a given mount. With a SEM-EDS spot size of 5 μm, corresponding to a circular particle area of 20 μm2 or an average of 5 × 10–6 % of the total area of all minerals on a mount, we conservatively report mineral abundance to three decimals. Several factors can influence the precision of automated quantitative mineralogical determination from sample heterogeneity, sample preparation (sieving and splitting), quality and type of grain mounts, instrumentation, and software. Some of these factors are discussed in Lastra and Paktunc (2016). Problematic EDS analyses at the boundary of mineral grains with similar backscattered electron brightness, typical in the analysis of petrographic thin sections (Sylvester, 2012), are partly attenuated in grain mounts because mineral grains are dominantly separated from each other. Lougheed et al. (2020) calculated a precision of <±1% on MLA-SEM analyses based on replicate analytical runs of the same sample using proper calibration measures between both runs. MLA-SEM analytical results are provided in Plouffe et al. (2021b).

Polished thin sections of bedrock samples containing epidote and polished mounts of epidote grains picked from till samples were analyzed by laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) at the Geological Survey of Canada (Ottawa, Ontario). Major and trace element measurements were acquired using a Photon Machines Analyte G2 (193-nm wavelength laser, equipped with a dualvolume cell) connected to an Agilent 7700x ICP-MS. Analyses were conducted using a laser fluence of 5 J/cm2, spot sizes of 30–65 μm (dependent on the size of the epidote), and a repetition rate of 10 Hz. Helium gas, at a flow rate of ~1 L/min, carried the ablated aerosol to the ICP-MS. Signals were acquired on the ICP-MS using a total mass cycle time of 456 ms and dwell times of 2–4 ms for major elements, 4–8 ms for most trace elements, and 12–16 ms for specific trace elements of interest (e.g., As, Sb). Total analytical time was 100 s comprising 40 s of background measurement (laser off) and 60 s of ablation (laser on). The U.S. Geological Survey GSE-1G standard was used for primary calibration and was analyzed twice every hour to correct for instrument drift. The secondary standards NIST-612, GSD-1G, BCR-2G, and Po689 were analyzed once per hour and routinely returned concentrations within 5–10% of their accepted values. Data calibration was undertaken using Glitter software (Griffin et al., 2008), during which analytical spectra were visually inspected and obvious mineral inclusions were avoided during signal selection. Elemental concentrations were calibrated using Ca as the internal standard, which was predetermined at the Department of Earth Sciences, University of Ottawa, on a JEOL8230 electron microprobe, equipped with five wavelength spectrometers. Additional details about the analytical procedures and all analytical results, including mean detection limits, are presented in Plouffe et al. (2021a).

Till geochemistry and clay mineralogy

Geochemical results for the two size fractions [clay (<0.002 mm) and silt plus clay (<0.063mm)] are presented for ore metals (Cu, Mo), pathfinder elements (Ag, Zn), and alteration elements expressed as their oxides (Al2O3, K2O). Results, presented on geochemical maps, are classified into the 50th, 70th, 90th, 95th, and 98th percentiles (Fig. 6). An empirical approach was used to define the threshold between background and anomalous concentrations in till by mapping concentrations and assessing their distributions based on bedrock geology, known mineralization, and reconstructed ice-flow directions. Histograms in the lower left corner of each map provide a graphic representation of the distribution of concentrations. Anomalous concentrations, i.e., concentrations indicative of mineralization or associated alteration, are defined as those only encountered near economic or subeconomic mineralization. In some cases, the anomalous concentrations appear on the histograms as the highest values of a bimodal population (e.g., Fig. 6E, H). Exceptions to this general rule are discussed below. The threshold is shown on the histogram of each map. For each element, the extent of glacial dispersal with anomalous concentrations is illustrated (dashed line in Fig. 6).

Concentrations of ore metals (Cu and Mo), pathfinder elements (Ag and Zn), and oxides (Al2O3 and K2O) above the 70th, 90th, or 95th percentiles, in both size fractions, are generally found near (≤6 km) economic and subeconomic mineralization (Fig. 6). The areal extent of the region with anomalous concentrations is largest for Cu and Zn (~36 km2) and smallest for Ag and Al2O3 (~6 km2). Although the spatial distributions of anomalous concentrations are similar for the same element, the concentrations are higher in the clay than in the silt plus clay fractions. Elevated concentrations of Ag, Mo, and K2O along a regional fault northeast of Gibraltar are not linked to known mineralization, and the sources remain undetermined.

The elevated concentrations of Al2O3 and K2O in both fractions of till samples collected near mineralization are interpreted to be related to minerals that formed in hydrothermal alteration zones associated with mineralization (Drummond et al., 1976; Bysouth et al., 1995; van Straaten et al., 2013). Sheet silicates produced during early potassic (biotite) or late-stage phyllic to argillic alteration (e.g., muscovite, chlorite, illite) (Cooke et al., 2014b) are either <0.063 mm in size in the source rocks or easily comminuted to clay-sized particles during glacial abrasion, leading to the elevated Al2O3 contents (>21.04 wt % in the clay and >15.96 wt % in the silt plus clay) in the till matrix. High K2O concentrations in the till matrix (>2.71 wt % in the clay and >1.66 wt % in the silt plus clay) are in samples collected near the economic mineralization and are likely derived from the potassic alteration zone, which has produced secondary biotite (Kobylinski et al., 2020).

To confirm these interpretations, the clay-sized fraction (<0.002 mm) of four till samples was analyzed by XRD: three samples contain ≥21.45 wt % Al2O3 and ≥3.18 wt % K2O, and one sample 15.13 wt % Al2O3 and 1.97 wt % K2O (sample locations are shown in Fig. 6I; semiquantitative mineralogy is given in Table 1). The four samples contain minor to abundant illite, chlorite, plagioclase feldspar, and quartz. Minor kaolinite and trace amphibole and dolomite occur in each sample; other mineral phases include K-feldspar, mixed-layer clay minerals (illite-smectite or chlorite-smectite), calcite, and jarosite, with the latter occurring in two samples in trace amounts. The greater abundance of illite and illite-smectite clay minerals in the three samples with high Al2O3 and K2O concentrations is interpreted to result from the glacial erosion of the alteration zones associated with the porphyry mineralization.

The upper and lower part of four till exposures show the general variability in the geochemical composition in the two size fractions with depth (Fig. 7). Near mineralization (<1 km, see profiles for sites 11PMA024 and 11PMA025 in Fig. 6; see Fig. 2 for profile site locations), concentrations of ore metals (Cu, Mo) and pathfinder elements (Ag, Zn) increase with depth. At approximately 4 km south of mineralization (Fig. 7, see profile 12PMA035), this relationship is reversed for Cu, Mo, and Ag in the clay fraction, with concentrations decreasing with increasing depth. At 6 km south of mineralization (Fig. 7, profile 12PMA042), ore metals and pathfinder element concentrations both increase and decrease with increasing depth. The concentrations of alteration elements (Al2O3 and K2O) do not vary substantially with depth.

Till mineralogy

The mid- (2.8–3.2 SG) and high-density (>3.2 SG) fine- to medium-sand fractions of till contain abundant porphyry Cu ore and alteration minerals formed during hydrothermal alteration or ore deposition processes. As with the geochemical results, mineral abundances in till indicative of mineralization in bedrock (anomalous abundances) are evaluated relative to bedrock geology, mineralization, and ice-flow directions (Fig. 8). Mineral abundance data are mapped using percentile categories and depicted in histograms in the bottom left corner of each map. The abundance of minerals identified by the optical method is compared to the abundance of minerals determined by MLA-SEM. The smaller number of samples (n = 20) analyzed by MLA-SEM limits the detailed interpretation of the glacial dispersal but does allow for the identification of specific minerals that have high abundances near mineralization. As for geochemical results, anomalous mineral abundances are defined here as those only encountered near economic or subeconomic mineralization and the related alteration zones.

Results of the optical and MLA-SEM mineral identifications show that chalcopyrite and pyrite, two sulfide minerals associated with mineralization, are more abundant in till south, west, and north of the main mineralized zones compared to the surrounding regions, which are devoid of mineralization (Fig. 8A-D). The total area with anomalous mineral abundances indicative of mineralization determined by the optical method is about 30 km2 for chalcopyrite and pyrite (Fig. 8A, C). The distribution of elevated chalcopyrite grain counts indicative of mineralization (>5 grains/10 kg) is similar to the distribution patterns of the high Cu concentrations (Figs. 6A, 8A). Molybdenite, the main sulfide ore mineral at Gibraltar, was not detected in till because of some combination of the following factors: (1) molybdenite was present in bedrock exposed to glacial erosion but was not sufficiently abundant to be detected in either the medium (0.25–0.50 mm) or the fine sand (0.125–0.180 mm) fractions; (2) molybdenite was glacially comminuted to particles <0.125 mm; or (3) molybdenite was present in the till but was destroyed during postglacial weathering (Plouffe and Ferbey, 2019).

Jarosite, a secondary sulfate mineral that forms from the oxidation of pyrite, has not been reported in the Gibraltar deposit, but it may be present in the oxidized zones in the upper 1–3 m of the deposit (Bysouth et al., 1995). Jarosite is present in trace amounts in the clay-sized fraction of till (Table 1). As observed by the optical method, it is also present in the heavy medium sand fraction of till within a 62-km2 area centered on mineralization but not in the surrounding regions (Fig. 8E). As determined by the MLA-SEM method, till samples with elevated jarosite (>0.002 area %) are located near the main ore zone but also at two sample sites approximately 11 km to the northwest, in the direction of the last phase of ice flow (Fig. 8F). The jarosite present in these two samples is likely derived from the main porphyry mineralization because no jarosite is found in the heavy medium sand fraction of the till (optical method) except near porphyry mineralization. In other words, the oxidized zone of the porphyry mineralization is likely the sole source of the jarosite in the till. Given that the abundance of jarosite was determined by the optical method in the 0.25- to 0.50-mm, 2.8- to 3.2-SG fraction and by MLA-SEM in the 0.125- to 0.180-mm, >3.2-SG fraction (Fig. 5), and that this mineral has a specific gravity that varies from 2.9 to 3.3 (http://webmineral.com/data/Jarosite.shtml; accessed January 2021), the results from both fractions represent partial abundances of what may be in the till.

Kobylinski et al. (2016) reported igneous and hydrothermal titanite in the Granite Mountain batholith based on its composition (<1.3 wt % Al2O3 in igneous and 2.0–6.0 wt % Al2O3 in hydrothermal titanite) and crystal habit (igneous titanite typically forms euhedral tabular crystals and hydrothermal titanite, needles <0.05 mm in length and <0.01 mm in width). In addition, Plouffe and Ferbey (2019) reported up to 80% titanite in the >3.2-SG, 0.25- to 0.5-mm fraction of disaggregated bedrock samples from the Gibraltar deposit. Till samples with >1% visually observed titanite only occur in a 25-km2 zone extending to the north and east of the economic mineralization (Fig. 8G). All of the other till samples, even those collected overlying other zones of the batholith, contain <1% titanite. In comparison, at two other porphyry Cu deposits in British Columbia, titanite is more abundant in till near mineralization (in trace amounts at the Mount Polley and 2–10% at the Highland Valley Copper deposits) compared to the surrounding region barren of mineralization (Plouffe and Ferbey, 2017). In contrast, the abundance of titanite in till determined by MLA-SEM is low proximal to Gibraltar deposits, and samples with high titanite area percent values are distal. Therefore, no titanite glacial dispersal train could be defined from the MLA-SEM method in the region of the Gibraltar deposit, most likely because titanite derived from the batholith and associated mineralization is in low abundance in the 0.125- to 0.180-mm fraction analyzed by MLA-SEM (Fig. 8H).

The presence of a large propylitic alteration halo around the Gibraltar deposit is reflected in the distribution of epidote in till. As determined by the optical method, samples with >45% epidote were collected above the Granite Mountain batholith and several kilometers to the north and south, extending over an area of at least 138 km2 (Fig. 8I; Plouffe et al., 2021b). These epidote abundances in the till (>3.2 SG, 0.25–0.50 mm; Epidote1 in Fig. 5) display distribution patterns very similar to those reported by Plouffe et al. (2016) for the >3.2-SG, 0.25- to 0.50-mm, 0.8- to 1.0-amp fraction (Epidote2 in Fig. 5). The presence of epidote in all till samples (an estimated amount of 0.1–45% in background areas; Fig. 8I) indicates that it is not only derived from the porphyry alteration zones but also from regional rock units, such as the Nicola Group volcanic and sedimentary rocks that are known to contain epidote formed during regional greenschist facies metamorphism (Greenwood et al., 1991; Panteleyev et al., 1996; Ash et al., 1999b).

Samples with >17 area % epidote, as determined by the MLA-SEM method, were collected over the Granite Mountain batholith as well as to the north and south, covering a minimum of 50 km2. This dispersal pattern is similar to that observed by the optical method, although it is less precisely defined because of the smaller number of samples examined (Fig. 8J).

A zircon dispersal train extends at least 5 km northwest (down-ice) from the northwestern edge of the Granite Mountain batholith (Fig. 8K). It is not by itself indicative of mineralization but, as discussed below, the trace element composition of zircon can provide an indication of intrusion fertility. Furthermore, the presence of large numbers of zircon grains in till in regions of porphyry deposits, reaching >100 grains/10 kg around the largest calc-alkaline intrusions, is in part due to the resistate nature of zircon (hardness = 7.5, no cleavage), which allows it to survive glacial erosion and transport and postglacial weathering (Plouffe and Ferbey, 2017).

The highest abundances of minerals associated with the porphyry mineralization (chalcopyrite, pyrite) and the oxidized zone (jarosite) as determined by the optical method occur in samples collected in the lower parts of the till profiles, near the mineralization (profiles 11PMA024 and 11PMA025, Figs. 2, 9), and in the upper part of the till at ~4 km from mineralization (profile 12PMA035, Fig. 9). Minerals partly derived from hydrothermal alteration (titanite and epidote) show either no trend or only subtle changes in abundances with depth. The interpretation of the mineralogical profiles is provided in the discussion.

Epidote chemistry

In addition to the regional till geochemistry and mineralogy, the trace element composition of epidote was assessed as a means of detecting buried hydrothermal alteration zones associated with porphyry Cu mineralization. As shown in Figure 8I and J, epidote being more abundant in till near porphyry mineralization can be an important indicator of hydrothermal alteration associated with this type of mineralization (see also Hashmi et al., 2015; Plouffe et al., 2016; Plouffe and Ferbey, 2017). The mixed provenance of epidote from porphyry mineralization and from rocks of the Nicola Group could hamper the recognition of the epidote derived from the porphyry source. However, in our study, the trace element composition of epidote might provide some insight to its genesis: porphyry-related versus metamorphic.

The major and trace element compositions of epidote from three types of samples were determined in this study: (1) polished mounts of epidote grains from till, (2) polished thin sections of intrusive rocks from the Granite Mountain batholith, and (3) polished thin sections of metamorphosed Nicola Group rocks. A total of 180 epidote grains were analyzed by LA-ICP-MS: 80 from nine till samples (Fig. 8I), 38 from nine bedrock samples of the Granite Mountain batholith (Fig. 2), and 62 from 12 bedrock samples of the Nicola Group (the locations of seven of the Nicola Group sample sites are shown in Fig. 2; the locations of the other sites are outside the map area).

Petrographic examination of the polished thin sections from the Granite Mountain batholith and the Nicola Group bedrock samples was conducted to attempt to characterize the epidote crystal forms and to relate them to genesis. Euhedral and poikilitic epidote replacing feldspar or overprinting chlorite is typical of the Nicola Group rocks and is interpreted to be metamorphic in origin. Epidote that replaces chlorite pseudomorphs after biotite with titanite inclusions, typical in the rocks of the Granite Mountain batholith, is interpreted as hydrothermal in origin. Some of this epidote shows a deformed wavy texture, suggesting deformation or, most likely, that epidote followed the texture of the preexisting biotite that was deformed. Pistachio-green epidote that occurs in veins in both Nicola Group and the Granite Mountain batholith is thought to be hydrothermal in origin. Sulfides are observed in places in association with vein epidote. Although the petrography provides an overview of the epidote texture in bedrock samples, no clear and unique link was observed between epidote texture and its composition, as described next.

The epidote supergroup includes four groups: allanite, åskagenite, epidote, and dollaseite (Armbruster et al., 2006; Mills et al., 2009; www.mindat.org, accessed January 2021). Epidote [Ca2Al2Fe3+[Si2O7][SiO4]O(OH)] and clinozoisite [Ca2Al3[Si2O7][SiO4]O(OH)] represent the Fe- and Al-rich end members, respectively, of a solid-solution series within the epidote group. The major element concentrations in all epidote grains, from all sample types, dominantly range from 10.0 to 16.0 wt % Fe2O3 and 19.3 to 24.6 wt % Al2O3, corresponding to epidote sensu stricto. In addition, Kobylinski et al. (2017) observed a predominance of epidote in the Granite Mountain batholith with lesser amounts of clinozoisite, allanite, and ferriallanite.

In terms of trace elements, the average concentrations of As, Sb, Y, total rare earth elements (ΣREEs; corresponding to the total concentrations of La to Lu), Sc, Th, Hf, and Cr in epidote are significantly higher in the Nicola Group rock samples compared to those from the Granite Mountain batholith (Table 2) and therefore are used to differentiate the two main sources of epidote in the till. Three plots are presented in Figure 10 to elucidate the differences in the epidote composition between these two rock types: As versus Sb, ΣREE versus Y, and Sc + Cr + Y versus Th + Hf.

In the As versus Sb plot (Fig. 10A), the compositional range of epidote from the Granite Mountain batholith overlaps the composition of epidote from the Nicola Group rocks. Epidote in till broadly overlaps with both rock types. Epidote from three Nicola Group rock samples collected close (<1 km) to the Burgess Creek stock (Fig. 2) contains higher As (>10 ppm) and/or Sb (>4 ppm) concentrations compared to epidote from other rock samples. In Figure 10, epidote composition from till samples are divided into two broad populations based on their geographic location: epidote from two up-ice samples collected east of the intrusion and epidote from samples collected down-ice, i.e., overlying or northwest of the intrusion (sample site locations are shown in Fig. 8I). In the As versus Sb plot, there is no distinction between the composition of epidote from up-ice and down-ice till samples (Fig. 10A).

In a ΣREE versus Y plot, epidote from the Granite Mountain batholith generally contains lower Y (≤45 ppm) and ΣREEs (≤200 ppm) than epidote from Nicola Group rocks, but there is overlap between the epidote populations (Fig. 10B). The composition of most epidote grains from till corresponds with those from the Granite Mountain batholith.

The Sc + Cr + Y versus Th + Hf plot provides additional support to partly discriminate between epidote derived from the Granite Mountain batholith versus the Nicola Group (Fig. 10C). These elements were grouped based on their general concentration ranges: 0.01–10 ppm for Th and Hf and 10–1,000 ppm for Sc, Cr, and Y. Epidote from the Granite Mountain batholith generally contains 5–180 ppm Sc + Cr + Y and 0.04–1.3 ppm Hf + Th, as opposed to Nicola Group epidote that generally contains 35–1,600 ppm Sc + Cr + Y and 0.1–10 ppm Hf + Th. Similar to plots in Figure 10A and B, there is a partial overlap between the epidote from the Granite Mountain batholith and Nicola Group rocks, and most epidote in till generally corresponds to the composition of epidote from the intrusion. For the epidote in till, there is no distinction between the epidote located up-ice and down-ice.

In addition to these trace elements, the Cu content of epidote was identified as a potential indicator of porphyry Cu mineralization. Epidote from the Granite Mountain batholith contains an average Cu concentration of 37 ppm compared to 6 ppm Cu in epidote from Nicola Group rocks (Fig. 11). Epidote from the Nicola Group rocks that contain the three highest Cu concentrations (26, 30, and 42 ppm) also has elevated As and Sb concentrations and is from samples collected <1 km from the Burgess Creek stock of the Granite Mountain batholith (Fig. 2). Epidote grains in till have a mean Cu concentration of 3 ppm (Fig. 11). Three grains with >30 ppm Cu (34, 40, and 64 ppm Cu) are from three different till samples: one collected close to mineralization (<1 km) and two collected to the northwest (down-ice) at 4 and 7 km from mineralization (Fig. 8I). We suspect that Cu could be present as nanoparticles (chalcopyrite?) within epidote, but its exact form has not been resolved.

Patterns of glacial dispersal

The distribution of elevated elemental concentrations and mineral abundances in till indicative of porphyry Cu mineralization does not outline a single glacial dispersal vector parallel to one dominant ice-flow movement. Instead, elevated values occur as amoeboid-shaped anomalies displaced in the combined direction of the three ice-flow movements: southeast, southwest, and north to northwest from the main mineralized zones. Such a dispersal pattern reflects multiple bedrock sources, which consist of a cluster of economic and subeconomic mineralized zones (i.e., not a single point source) and three distinct vectors of ice-flow movements (Plouffe et al., 2016). There is evidence for glacial dispersal by each ice-flow phase. Elevated Zn concentrations, south of Gibraltar (Fig. 6G, H), can be attributed to the erosion of the Zn zone west of the Gibraltar East and West mineralized zones, at lower elevations, and glacial dispersal to the southeast during the early phase of glaciation (Fig. 4A). Notably, Zn is not reported as a commodity for any of the mineral occurrences south of Gibraltar (British Columbia Geological Survey, 2020). Elevated jarosite grain counts in till west of Gibraltar (Fig. 8E) are most likely the result of glacial transport during the second phase of ice flow. Lastly, elevated Cu concentrations in the clay fraction of till (Fig. 6A), as well as epidote and zircon abundances (Fig. 8I-K) north to northwest of the main cluster of mineralized zones or the intrusion, are attributed to northward to northwestward glacial transport during the last ice-flow event. Amoeboid-shaped glacial dispersal patterns have been described elsewhere in Canada where multiple ice-flow movements have eroded and transported debris to form a stellate pattern (e.g., Shilts, 1993; Stea and Finck, 2001; Trommelen et al., 2013).

Till profile composition

The geochemical and mineralogical profiles in till can be interpreted in the context of glacial dispersal. At less than 1 km down-ice of mineralization (west and northwest), ore and pathfinder element concentrations and sulfide minerals (pyrite and chalcopyrite) abundances are highest in the lower part of the till (Figs. 7, 9). At 4 km south of mineralization, the geochemical and mineralogical profiles are reversed, with greater abundances of the element and mineral indicators in the upper part of the till. These trends follow schematic representations of glacial dispersal in three dimensions by Drake (1983), Miller (1984), and McClenaghan et al. (2018) in which the dispersal train typically rises in a till unit with increasing distance of glacial transport. The complete transport path of till with high metal concentrations and sulfide mineral abundances in the upper part of the till, 4 km from mineralization, is complex, resulting from erosional and depositional events of three ice-flow movements. The general uniform distribution of alteration-related oxides (Al2O3, K2O) and minerals (titanite, epidote) with depth in the till profiles (Figs. 7, 9) probably relates to the large areal extent of the hydrothermal alteration zones, which are reflected at all depths in the till.

Epidote chemistry

Epidote, an abundant mineral in the alteration zones of the vast majority of porphyry Cu deposits, can be a useful indicator for porphyry Cu exploration in glaciated terrains. In the Quesnel terrane, the greater abundance of epidote in till overlying and down-ice of intrusions with porphyry mineralization compared to surrounding country rocks can be an indicator of the presence of hydrothermal alteration (Plouffe et al., 2016; Plouffe and Ferbey, 2017). The composition of epidote grains in till can provide additional information about their source and relationship to potential mineralization. Trace element compositions of epidote in rocks vary with distance from mineralization in a single porphyry system (e.g., Cooke et al., 2014a, 2020b; Baker et al., 2020; Pacey et al., 2020; Wilkinson et al., 2020) and some trace elements of interest (e.g., As, Sb) are heterogeneously distributed in epidote (Cooke et al., 2014a; Ahmed et al., 2020). Despite this variability, Wilkinson et al. (2020) found that porphyry-related epidote is typically enriched in As and Sb derived from hydrothermal fluids, compared to metamorphic epidote (Fig. 10A). Based on these criteria, the epidote with high As (>10 ppm) and/or Sb (>4 ppm) concentrations in the Nicola Group rocks close to the Granite Mountain batholith (Figs. 2, 10A) is interpreted to be of hydrothermal origin. This interpretation is further supported by the occurrence of epidote veins in these rocks (Kobylinski et al., 2020), which formed during hydrothermal alteration as opposed to metamorphism. The epidote in the Granite Mountain batholith that plots within the metamorphic field, with low As and Sb concentrations, could be either hydrothermal in origin and affected by secondary alteration, deformation, and metamorphism or metamorphic in origin (Plouffe et al., 2021a).

Using Wilkinson et al.’s (2020) epidote classification (shown in Fig. 10A), 41% of the epidote grains from till plot in the porphyry field and 32% in the metamorphic field. The remaining grains plot either within the overlapping field or outside and above both fields. The presence of porphyry-related epidote grains in till samples collected 1.4 km up-ice and to the east of the Granite Mountain batholith indicates either that the propylitic alteration extends to the east in the up-ice region or that there is another unrecognized source of epidote with high As and Sb in the up-ice region. The results of this study suggest that the As versus Sb discrimination tool (Wilkinson et al., 2020) broadly differentiates between metamorphic and hydrothermal epidote in the till. However, grain populations need to be analyzed (80 grains from nine till samples in our study represents a minimum), and the As versus Sb discrimination should be used with caution in areas where porphyry systems were subjected to greenschist and higher-grade metamorphism.

The other trace elements investigated in this study (ΣREEs, Y, Sc, Cr, Y, Hf, and Th) (Fig. 10B, C) show broad differences between intrusion-related and metamorphic epidote. The concentrations of these elements in epidote roughly discriminate between epidote from the Granite Mountain batholith and that from the Nicola Group (i.e., approximately ≤200 ppm ΣREEs, ≤45 ppm Y, <100 ppm Sc + Cr + Y, and <2 ppm Th + Hf in epidote from intrusive rocks) (Fig. 10B, C).

Copper concentrations in epidote at many porphyry Cu deposits are reported to be highest in the peripheral zones of mineralization. For example, the median Cu concentration in epidote reaches 5 ppm in the Poderosa deposit of the Collahuasi porphyry district in Chile (Baker et al., 2020) and 28 ppm at the Nugget Hill porphyry Cu-Au prospect in the Philippines (Cooke et al., 2014a). Similarly, the average Cu concentration of epidote is 107 ppm at the El Teniente porphyry Cu deposit in Chile (Wilkinson et al., 2020). At Mount Milligan, 325 km north of Gibraltar, the highest Cu concentration in epidote near porphyry mineralization is 45 ppm (Jago et al., 2014). In contrast, at the Northparkes porphyry Cu-Au district in New South Wales, Australia, Pacey et al. (2020) detected Cu in only 9% of their LA-ICP-MS epidote analyses, and a maximum concentration of 5 ppm. At all of these deposits, Cu concentrations in epidote generally decrease with increasing distance from mineralization. In addition to these studies, the results reported here show that the Cu content of epidote is a potential indicator of fertility but with limitations. Epidote Cu concentrations herein reported vary widely in all rock types (Fig. 11), but values >30 ppm Cu are only encountered in rocks of the Granite Mountain batholith and in Nicola Group rocks close to the intrusion. Furthermore, Cooke et al. (2014a) and Ahmed et al. (2020) report heterogeneous distribution of Cu within single epidote grains. Consequently, a wide range of Cu concentrations should be expected, and multiple samples would need to be analyzed to detect this fertility signal in till. In this study, only three of the 80 epidote grains from the nine till samples, collected at <0.5, 4, and 7 km northwest (down-ice) from mineralization, contain >30 ppm Cu (Fig. 8I).

Zircon chemistry

In addition to epidote, zircon grains in till can be indicators of porphyry Cu mineralization because of their physical robustness, their widespread occurrence in most felsic to intermediate intrusive rocks, and their reported REE composition that is indicative of porphyry fertility (e.g., Ballard et al., 2002; Dilles et al., 2015; Lu et al., 2016; Lee et al., 2017, 2021a, b; Shu et al., 2019; Cooke et al., 2020a). Here, we synthesize and discuss zircon chemistry results obtained as part of our study at Gibraltar but previously published separately. The potential use of zircon composition is discussed in the context of mineral exploration in covered terrains.

The composition of zircon in the intrusion and till at Gibraltar was determined by Wolfe (2017) and Kobylinski et al. (2018, 2020), with preliminary results reported by Plouffe et al. (2019). Kobylinski et al. (2020) calculated the Ce+4/Ce+3 ratios of zircon from the Granite Mountain batholith, following the method described by Ballard et al. (2002) (Fig. 12A). Ballard et al. (2002) and others (e.g., Liang et al., 2006; Shen et al., 2015), have shown that intrusions with economic porphyry Cu mineralization typically contain zircon with higher Ce+4/Ce+3 ratios compared to zircon from barren intrusive rocks. As interpreted by Ballard et al. (2002), the higher Ce+4/Ce+3 ratios in zircon from porphyry-related magmas can be explained by the oxidized nature of the intrusion and, therefore, a greater abundance of Ce+4 in the melt, which leads to high Ce+4/Ce+3 ratios in zircon. Oxidizing conditions are necessary to keep chalcophile elements in the melt as it fractionates (Sillitoe, 2010; Richards, 2011; Cooke et al., 2014b). Under reducing conditions, chalcophile elements form sulfides, which may fractionate early and deplete the melt of economic metals (e.g., Cu, Mo). The Ce+4/Ce+3 ratios in zircon from the Granite Mountain batholith exhibit a wide range of values: 9–555 with a mean of 128 in zircon from barren tonalite and 21–1,780 with a mean of 311 in zircon from Cu-bearing tonalite (Fig. 12A; Kobylinski et al., 2020). In addition, of the 119 zircon grains from the barren tonalite, only four reach a Ce+4/Ce+3 ratio of >400 (3.4%) whereas 19 of 74 zircon grains from the Cu-bearing tonalite (25.7%) had Ce+4/Ce+3 >400 (Kobylinski et al., 2020).

The method of Ballard et al. (2002) for calculating the Ce+4/Ce+3 ratios in zircon requires the determination of the Ce abundance of the host rock to estimate the Ce content in the parental melt. Given that such values cannot be obtained from detrital zircon grains detached from their host rock, the Ce/Nd ratio has been used here as a proxy for the Ce+4/Ce+3 ratio. The ratios Ce/Nd and Ce+4/Ce+3 are positively correlated in zircon from the Granite Mountain batholith with a correlation coefficient of 0.90 following Ce+4/Ce+3 = 7.28 (Ce/Nd)1.35 (Kobylinski et al., 2020). Wolfe (2017) determined the age and tested the composition of 45 zircon grains recovered from five till samples (Fig. 8K). The average ages of all zircon grains in each of the five till samples as determined by LA-ICP-MS have low precision and vary from 204 to 170 Ma (Wolfe, 2017) revealing that the grains are not solely derived from the Late Triassic Granite Mountain batholith. In a graph of Ce/Nd ratio against crystallization temperature using Ti-in-zircon geothermometer (Ferry and Watson, 2007) (Fig. 12B), only four zircon grains returned a Ce/Nd ratio >19; three grains are from a till sample collected <1 km northwest (down-ice) from mineralization (sample 11PMA024A2, Fig. 8K) and one grain from a sample collected at the southwestern edge of the intrusion (sample 12PMA032B01, Fig. 8K). Following the correlation established by Kobylinski et al. (2020), a Ce/Nd ratio of 19 in zircon from the Granite Mountain batholith corresponds to a Ce+4/Ce+3 ratio of 388. This value is potentially indicative of a fertile intrusion, particularly in comparison to other mineralized systems, e.g., values of 300 defined by Ballard et al. (2002) from the Chuquicamata-El Abra porphyry Cu belt in Chile and a value of 120 in porphyry Cu deposits from Tibet (Liang et al., 2006) and Central Asia (Shen et al., 2015).

Detecting the fertility signal in zircon grains in till in the Gibraltar deposit region could be improved by expanding analyses to include smaller-grained fractions. Note that Plouffe et al. (2021b) suggest a method combining MLA-SEM followed by LA-ICP-MS to analyze zircon grains in the 0.125- to 0.180-mm size range. Zircon grains in Cu-bearing tonalite are more abundant in the 0.125- to 0.212-mm size range compared to zircon grains in barren tonalite, which are dominantly 0.212–0.350 mm in size (Kobylinski et al., 2020). Based on grain size alone, the analysis of zircon grains from the 0.25- to 0.50-mm fraction of till, as conducted by Wolfe (2017), might have been biased toward sampling zircon grains from barren tonalitic rocks.

These results, with a limited number of samples, demonstrate that detrital zircon grains with a positive fertility signal can be detected in till in the Gibraltar region. However, additional till samples and more zircon grains per sample should be analyzed with two REE measurements (rim and core) on single grains, given the known intra- and intergrain variability of zircon compositions (e.g., Hofmann et al., 2009, 2014; Chamberlain et al., 2014; Loucks et al., 2020).

In addition to Ce+4/Ce+3 ratios, Eu anomalies (Eu/Eu*) in zircon can be utilized as an additional fertility indicator for porphyry mineralization, such that a threshold of about >0.4 differentiates mineralized from barren intrusions (Lu et al., 2016; Lee et al., 2021a). Hydrous oxidized melt associated with porphyry mineralization contains Eu+3, which facilitates its incorporation in the zircon lattice to produce less negative Eu/Eu* values, which is enhanced by the suppression of plagioclase fractionation in a hydrous melt (Ballard et al., 2002; Dilles et al., 2015; Lee et al., 2017; Lu et al., 2019). In a study of the traceability of Eu anomalies in zircon recovered from till collected near porphyry mineralization, Lee et al. (2021b) analyzed 296 zircon grains from 12 till samples from the Highland Valley porphyry district, 247 km south of the Gibraltar deposit, and detected a positive fertility signal (≥0.4 Eu/Eu* in zircon) up to 9 km down-ice of mineralization. However, zircon Eu/Eu* values in bedrock samples of the Granite Mountain batholith are not an indicator of fertility, as Eu/Eu* values vary from 0.19 to 0.39, with an average of 0.29 in Cu-bearing tonalite versus 0.19–0.57 and with an average of 0.35 in barren tonalite (Kobylinski et al., 2020). Bouzari and Hart (2019) also reported Eu/Eu* values <0.35 in zircon of the mineralized intrusive phase of the Granite Mountain batholith. Kobylinski et al. (2020) interpreted Eu/Eu* in zircon to be largely controlled by the cocrystallization of plagioclase and titanite, as described by Loader et al. (2017) and discussed in Bouzari and Hart (2019), and not redox-related conditions of the melt. Clearly, more studies are needed to define the factors that control REE composition of zircon in intrusive rocks and how it can be related to a positive porphyry fertility signal in detrital zircon grains.

Implications for mineral exploration

Based on the geochemistry and mineralogy results of till from around the Granite Mountain batholith, porphyry Cu mineralization exposed to glacial erosion is well reflected in the local till composition. The mineralogical and geochemical compositions of the till reflect a combination of ore metals (Cu, Mo), pathfinder elements (Zn, Ag), and hydrothermal alteration (Al, K) associated with mineralization. As demonstrated in this case study at Gibraltar and also reported at other porphyry study sites (e.g., Sibbick and Kerr, 1995; Levson, 2001, 2002; Ferbey and Levson, 2009; Hashmi et al., 2015; Montsion et al., 2019; Shewchuk et al., 2020), multielement anomalies are present in till down-ice of porphyry Cu mineralization. Geochemical analyses can be completed on either the clay or the silt plus clay fractions because dispersal patterns are similar for both fractions. The influence of till texture on the trace element composition of the silt plus clay fraction, as originally described by Shilts (1975), was not observed in this study, likely because of the homogeneity of the till texture (silty sand to sandy silt) in the Gibraltar deposit region.

Mineral exploration programs in glaciated regions with prospective porphyry intrusions should target the geochemical and mineralogical footprints in till, which can extend over 6–136 km2 around and down-ice of mineralization, as found in this study. In till geochemistry and mineralogy maps of this study (Figs. 6, 8), samples with elevated indicator mineral abundances or high elemental concentrations were often collected near samples that had only background indicator mineral abundance or elemental concentrations. Such till heterogeneity is also exemplified in the depth profiles in which the anomaly (dispersal train) is not found at all depths (Figs. 7, 9). Consequently, for a reconnaissance till-sampling survey to be effective in testing for porphyry mineralization in an unexplored region, a sample density of one sample per square kilometer is recommended, which is similar to the one applied in our survey. Sampling till at various depths where feasible is recommended considering that geochemical and mineralogical anomalies can be limited to the upper or lower part of a till profile. Such sample density would provide multiple samples within a dispersal train, i.e., within a geochemical or mineralogical anomaly, and would limit the effect of an anomalous signal being obscured because of till heterogeneity. Reconnaissance surveys could be conducted using a cost-benefit approach with the less expensive analyses undertaken first (geochemistry), followed by more detailed mineralogical analyses of targeted samples, which can be completed on the mid- (2.8–3.2 SG) and high-density (>3.2 SG) fine- to medium-sand fraction of till by the optical method or on the high-density (>3.2 SG) 0.125- to 0.180-mm fraction by MLA-SEM.

The Gibraltar porphyry Cu-Mo deposit, located in south-central British Columbia, is surrounded by till geochemical and mineralogical anomalies, which extend over 6–136 km2 down-ice, forming amoeboid-shaped dispersal trains resulting from dispersal by three phases of ice flow that affected the region. These anomalies are derived from a cluster of porphyry Cu mineral occurrences that include not only the mineralized zones at the Gibraltar mine but also showings adjacent to it. These large geochemical and mineralogical anomalies in till provide much larger exploration targets than the mineralized bedrock, and they can be detected using reconnaissance till sampling. Targeted till surveys can be a particularly efficient method for exploration in regions where the geology is prospective for porphyry Cu mineralization but extensively covered by glacial sediments.

Porphyry Cu discoveries in the future will be made at increasing depths within bedrock—in some cases with mineralized zones not exposed to glacial erosion. Consequently, ore elements and minerals would be absent in the local till. However, if the alteration zone was exposed to glacial erosion, the recognition of its geochemical or mineralogical footprint in till will be conducive for detecting more deeply buried mineralization. Determining the composition of alteration minerals (e.g., epidote) or magmatic minerals (e.g., zircon) could provide indications that an intrusion may host porphyry Cu mineralization. A better understanding of the factors that control the trace element composition of minerals found in porphyry Cu systems will improve our ability to detect mineralization at increasing depths.

We acknowledge coworkers at GSC Ottawa I. Bilot for completing the XRD analyses, K.E. Venance for the electron microprobe analyses, and Claudia Moore and Miriam Wygergangs, who completed the grain size separations of the till samples. P. Schiarizza from the British Columbia Geological Survey kindly provided Nicola Group rock samples. This project was completed under the auspices of the Targeted Geoscience Initiative of the Geological Survey of Canada, Natural Resources Canada. Retiree J. Fleming, formerly with Taseko Mines Ltd., ensured our safety and provided support and cooperation while we visited the mine site. The manuscript greatly benefited from the internal GSC review completed by M.B. McClenaghan (GSC, Ottawa) and journal reviews by Ross Sherlock and Carmina Jorquera. E. Ambrose provided professional editorial services to improve the text and figures. This paper is NRCan contribution 20210117.

Alain Plouffe is a Quaternary geologist at the Geological Survey of Canada (GSC) with over 30 years of experience combining glacial geology applied to mineral exploration in the Canadian Cordillera. He holds an M.Sc. degree from Carleton University and a Ph.D. degree from the University of Montreal. He has led surficial geology mapping projects and regional till geochemical and mineralogical surveys in British Columbia, Yukon, and northwest Alberta, with direct applications for the discovery of buried mineralization. He is currently pursuing research on developing mineral exploration methods for porphyry mineralization in glaciated landscape, studying examples from the Canadian Cordillera.

Gold Open Access: This paper is published under the terms of the CC-BY-NC 3.0 license.

Supplementary data